article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
in a now classical paper in bioinformatics @xcite , hannenhalli and pevzner showed that there is a polynomial time algorithm to sort signed permutations by reversals , i.e. , turn any signed permutation into the identity by reversing subwords ( and flipping their signs ) .
this has important implications for computational phylogenetics : when comparing the sequence of genes of two related species , the shortest length of a sequence of reversals that transforms one into the other is one prominent measure of the evolutionary distance of the associated organisms .
the authors strategy , and one that was improved upon in later work ( for example , @xcite ) , is to construct the so - called `` breakpoint graph '' for the permutation to be sorted , show that a certain operation on the breakpoint graph corresponds to reversals , and then use certain numerical invariants of subgraphs to guide the sequence of moves to the identity .
this framework is now a keystone of bioinformatics algorithms , but it leaves many questions unanswered . in particular , the proposed methodologies generate just one successful sorting of the signed permutation under consideration , and it is understood that there are often many such minimum - length sorting sequences . since each is only representative of one _ possible _ evolutionary history , it would be valuable to be able to sample from all possible such sequences to obtain more sensitive statistical properties .
as of yet , there is no full understanding of the space of possible histories , so markov chain monte carlo methods are valuable for approximately uniform sampling .
such approaches present their own problems , however : it is necessary to obtain a proof of connectivity of the underlying graph of the markov chain to know that it will eventually reach every vertex ; and it is necessary to obtain bounds on the mixing time of the process to ensure that near - uniformity will be achieved in reasonable time .
indeed , some researchers have investigated these very kinds of questions : see , for example , @xcite . in order to state our results and situate it in the above discussion
, we need the following definitions .
a * bicolored * graph is a pair @xmath1 where @xmath2 is a simple graph , and @xmath3 is a coloring of its vertices .
write @xmath4 and @xmath5 .
denote by @xmath6 the vertex set of a graph , @xmath7 its edge set , and @xmath8 $ ] the induced subgraph of a set @xmath9 ; let @xmath10 denote the neighborhood of @xmath11 , i.e. , @xmath12 , and @xmath13 the closed neighborhood of @xmath14 , i.e. , @xmath15 .
consider a bicolored graph , @xmath1 with a black vertex @xmath11 .
`` pressing @xmath14 '' is the operation of transforming @xmath1 into @xmath16 , a new bicolored graph in which @xmath17 $ ] is complemented .
that is , @xmath18 , @xmath19 ( where `` @xmath20 '' denotes symmetric difference ) and @xmath21 for @xmath22 and @xmath23 for @xmath24 .
[ cols="^ " , ] the `` pressing game '' ( to use terminology from @xcite ) is played by pressing black vertices of @xmath2 iteratively with the ultimate goal of transforming @xmath2 into an all - white , empty graph .
hannenhalli and pevzner showed ( @xcite ) that `` successful '' sequences of presses in the breakpoint graph of a signed permutation , i.e. , sequences that result in an all - white empty graph , correspond bijectively to minimum - length sequences of reversals that turn the permutation into the identity . therefore , sampling from successful pressing sequences is equivalent to sampling from the minimum length sequences of reversals that sort a signed permutation . in @xcite ,
the authors make the following `` pressing game conjecture '' : every successful pressing sequence can be reached from every other one by a sequence of edits that involve at most four deletions or insertions .
if successful pressing sequences are taken to be the vertices of a graph @xmath25 , and the edges correspond to edits of at most four deletions or insertions , then the pressing game conjecture implies that @xmath25 is connected .
then a simple random walk converges to a uniform distribution on the set of all successful pressing sequences , and markov chain monte carlo can be used to analyze typical pressing sequences .
bixby , flint , and mikls @xcite proved the conjecture for paths . despite this , the current authors have doubts about the statement for general graphs . in the present manuscript
, we explore a few aspects of matrix theory over @xmath0 so as to better understand the successful pressing sequences of a graph . among other results : in corollary [ cor : rank_cor ]
, we show that the rank of the `` augmented adjacency matrix '' of a bicolored graph is the length of every successful pressing sequence of a graph ; theorems [ thm : main_thm ] and [ thm : main_thm2 ] provide a substantial collection of equivalent characterizations of successful pressing sequences ; proposition [ prop : lpqu ] gives a matrix - theoretic formulation of the relationship between successful pressing sequences ; and theorem [ thm : average ] shows that the average number of successful pressing sequences of a random ( full - rank ) bicolored graph is large .
the final section contains several open problems concerning these sequences that arose in connection with the present work .
we note that some special cases of a few of our results are announced but left largely unproven in @xcite ; the authors refer to the matrix analogue of pressing as `` clicking '' and to the condition of the existence of a successful pressing sequence that consists of all vertices as `` tightness . ''
the following result appears ( less explicitly ) in @xcite and @xcite , but we include the proof for completeness . [
prop : success_is_possible ] any graph with a black vertex in every component has a successful pressing sequence . it suffices to prove the statement for connected graphs .
let @xmath2 be a connected graph and @xmath26 be the set of black vertices with the fewest possible black neighbors .
choose some @xmath27 such that @xmath28 is maximal in @xmath26 .
when @xmath29 is pressed , we obtain @xmath30 . we claim that each component of @xmath30 is either a white isolated vertex or has at least one black vertex .
let @xmath31 be the set of neighbors of @xmath32 , let @xmath33 be the set of vertices in @xmath34 that were white in @xmath2 , and let @xmath35 .
note that @xmath2 and @xmath30 are identical except on the induced subgraphs of @xmath36 .
every vertex in @xmath37 is in a component with a vertex of @xmath34 ( in both @xmath2 and @xmath30 ) , so it suffices to show that each vertex of @xmath36 is in a component of @xmath30 with a black vertex or is an isolated white vertex .
furthermore , in @xmath30 , @xmath29 is isolated and white and the vertices of @xmath33 are black , so we need only consider the elements of @xmath38 .
pick some @xmath39 . if @xmath40 is adjacent to a black vertex outside of @xmath36 or @xmath40 is not adjacent in @xmath2 to some vertex in @xmath33 , then in @xmath30 , @xmath40 is adjacent to a black vertex .
otherwise , in @xmath2 , @xmath40 is adjacent to all vertices of @xmath33 and its black neighbors are a subset of @xmath41 . by the choice of @xmath29
, this implies that the closed neighborhoods of @xmath29 and @xmath40 are the same in @xmath2 , which implies that @xmath40 is a white isolated vertex in @xmath30 .
the * augmented adjacency matrix * @xmath42 of a bicolored graph @xmath2 on @xmath43 vertices , is the adjacency matrix of @xmath2 where the entries along main diagonal correspond to the vertices @xmath2 and are indexed by the color of the vertex ; @xmath44 if white or @xmath45 if black .
given a bicolored graph @xmath2 , we can define a ( loopy simple , uncolored ) graph @xmath46 to be the graph on the same vertex set with the same edges , but with a loop at each black vertex ( and none at white vertices ) .
a perfect matching in such a graph is a set of edges incident to every vertex exactly once , where a loop is considered to be incident to its vertex only once . a special case of the following result ( that of zero diagonal ) appears in @xcite . [ prop : matchings ]
the number of perfect matchings in the loopy graph @xmath46 corresponding to a bicolored graph @xmath1 is odd if and only if @xmath47 is invertible over @xmath0 .
it is well known that the permanent ( which is equal to the determinant in characteristic 2 ) of @xmath47 is equal to @xmath48 where @xmath49 ranges over all _ vertex circuit covers _ , i.e. , families of circuits ( closed walks ) in which each vertex appears exactly once and @xmath50 is the number of such circuits of length greater than two .
( see , for example , @xcite . )
therefore , over @xmath0 , the only terms which make a contribution to @xmath51 are those in which there are no circuits of length more than two , i.e. , every component is a loop or a single edge precisely the condition of being a perfect matching .
since @xmath52 if and only if @xmath47 is invertible , this is equivalent to there being an odd number of perfect matchings .
define the function @xmath53 on @xmath54 nonzero matrices over @xmath55 as follows ; the action of @xmath56 will amount to a slight modification of gaussian elimination , wherein row - permutations are prohibited and the pivot row is added to itself as well .
let @xmath57 denote the smallest row index of a left - most @xmath45 in @xmath58 , that is , the positive integer for which there exists a @xmath59 so that 1 .
@xmath60 2 .
@xmath61 if @xmath62 3 . if @xmath63 and @xmath64 , then @xmath65 .
note that @xmath57 and @xmath59 are uniquely determined by @xmath66 in satisfying the above requirements .
let @xmath67 be the set of row indices which have a @xmath45 in column @xmath59 , i.e. , @xmath68 .
let @xmath53 denote the @xmath54 matrix so that @xmath69 note that , for every matrix @xmath58 , there is a sequence of @xmath57 s and @xmath59 s that arise from the iterative application of @xmath56 to @xmath58 .
that is , given @xmath58 , there is an increasing sequence @xmath70 and increasing sequence @xmath71 which serve as the indices in the above definition of @xmath53 , @xmath72 , etc .
indeed , it is easy to see that the sequence must eventually result in the all - zeroes matrix , so this process terminates at some finite @xmath73 .
if , for each @xmath74 $ ] , @xmath75 , we call @xmath58 `` leading principally nonsingular ( lpn ) '' . if @xmath58 is lpn , then the sequence @xmath76 , @xmath77 , is precisely the sequence of matrices one obtains by performing gaussian elimination on @xmath58 , with the additional operation of adding pivot rows to themselves ( thus replacing them with the zero vector ) . furthermore , this elimination does not involve row permutations .
therefore , @xmath58 is row - reducible to a diagonal matrix whose leading principle submatrix of size @xmath78 is the identity matrix @xmath79 , and whose other entries are zero .
note that , if @xmath58 is symmetric , then @xmath53 is , as well .
indeed , @xmath80 therefore , suppose that the @xmath58 above is @xmath47 and is lpn .
then it is straightforward to see that @xmath53 is in fact @xmath81 where @xmath82 is obtained from @xmath2 by pressing its lowest - indexed ( black ) vertex . since @xmath47 being the all - zeroes matrix is precisely the condition that @xmath2 has no edges and all vertices are white , @xmath83 being lpn is equivalent to @xmath2 having a successful pressing sequence consisting of the vertices indexing the first @xmath84 columns of @xmath58 in increasing order .
we may conclude the following .
[ prop : row_reduction ] for a graph @xmath2 with vertex set @xmath85 $ ] , the following are equivalent : 1 .
gaussian elimination applied to @xmath47 consists of row reduction applied , in consecutive order , to the first @xmath86 rows .
the first @xmath87 leading principal minors of @xmath47 are nonzero , and the rest are zero . 3 .
@xmath88 is a successful pressing sequence for @xmath2 .
[ cor : rank_cor ] the number of vertices in any successful pressing sequence for a graph @xmath2 depends only on the graph , and is equal to @xmath84 . the preceding result justifies the following definition .
the * pressing number * of a graph is the number of presses required to transform the graph into an all - white , empty graph .
a matrix @xmath58 is said to have an @xmath89-decomposition if there exist a lower triangular matrix @xmath90 and an upper triangular matrix @xmath66 so that @xmath91 call a matrix @xmath58 `` cholesky '' ( to borrow terminology from the theory of real / complex matrices ; q.v .
@xcite ) if there exists a lower - triangular @xmath90 so that @xmath92 ; such a product is evidently a special type of @xmath89-decomposition .
the following lemma is folkloric .
[ lem : lu_is_unique ] if @xmath93 and @xmath58 is invertible , this decomposition is unique .
suppose @xmath94 .
then @xmath90 , @xmath66 , @xmath95 , and @xmath96 are invertible , so @xmath97 the left - hand side of this equation is lower - triangular and the right is upper - triangular , so they must both be diagonal .
however , since the only invertible diagonal matrix over @xmath0 is the identity , @xmath98 and @xmath99 .
[ lem : symmetric_and_lu ] if a symmetric matrix @xmath58 over @xmath0 has an @xmath89-decomposition , then it has a cholesky decomposition @xmath100 .
we proceed by induction .
the base case is trivial : @xmath101 $ ] or @xmath102 $ ] .
suppose @xmath58 is @xmath54 and the statement is true for all @xmath103 .
if the first row of @xmath58 is all zero , then gaussian elimination of @xmath58 can proceed without row permutations if and only if this is true of @xmath104 , the matrix @xmath58 with its first row and column removed .
since the existence of an @xmath89-decompostion is equivalent to gaussian elimination proceeding without the necessity of row permutations , it follows that we may apply the inductive hypothesis to @xmath104 .
suppose the first row of @xmath58 is nonzero
. then @xmath58 can be written @xmath105 \left [ \begin{array}{ccc } u_0 & c \\ o & d \end{array } \right ] \\ & = \left [ \begin{array}{ccc } l_0u_0 & l_0c \\ au_0 & ac+bd \end{array } \right ] \end{aligned}\ ] ] where @xmath106 are @xmath107 are invertible leading principal submatrices , @xmath108 is lower - triangular and @xmath109 is upper - triangular because , otherwise , the first row or column of @xmath58 would be zero . since @xmath110 is nonsingular , this decomposition of @xmath111 is unique by lemma [ lem : lu_is_unique ] .
thus , @xmath112 implies that @xmath113 . since @xmath58 is symmetric , @xmath114 whence @xmath115 .
therefore , if @xmath116 is the lower - right @xmath117 principal submatrix of @xmath58 , we may rewrite @xmath118 since the left - hand side is symmetric , the right - hand is as well , and we may apply induction : @xmath119 has an @xmath89-decomposition of the form @xmath120 , so it has an @xmath121-decomposition as well . therefore
, we may let @xmath122\ ] ] and conclude that @xmath123 .
[ thm : main_thm ] given a bicolored labeled graph @xmath2 and integer @xmath87 , the following are equivalent : 1 .
the pressing number of @xmath2 is @xmath87 .
@xmath47 has rank @xmath87 and can be written @xmath124 for some lower - triangular matrix @xmath90 and permutation matrix @xmath33 .
3 . @xmath125 and @xmath2 has a black vertex in each component that is not an isolated vertex .
4 . there is some permutation matrix @xmath33 so that the @xmath126-th leading principal minor of @xmath127 is nonzero for @xmath128 $ ] and is zero for @xmath129 if @xmath130 .
there is an ordering of the vertices @xmath131 of @xmath46 so that the induced subgraph @xmath132 $ ] has an odd number of perfect matchings for each @xmath128 $ ] , and , for each @xmath133 \setminus [ k]$ ] , @xmath132 $ ] has an even number of perfect matchings . 6 .
@xmath134 for some permutation matrix @xmath33 , lower triangular matrix @xmath90 , and upper triangular matrix @xmath66 , where @xmath135 .
we note that there are yet other equivalent conditions in terms of rank of submatrices which are given by johnson and okunev in @xcite .
@xmath136 : this follows from proposition [ prop : row_reduction ] , since the existence of an @xmath89-decomposition is equivalent to a matrix being row - reducible without performing row permutations ; conjugation by @xmath33 has the effect of placing rows and columns indexed by the pressing sequence in an initial position of the matrix .
+ @xmath137 : this is a consequence of lemma [ lem : symmetric_and_lu ] .
+ @xmath138 : the combination of proposition [ prop : success_is_possible ] , proposition [ prop : row_reduction ] , and corollary [ cor : rank_cor ] gives this equivalence .
+ @xmath139 : this follows from proposition [ prop : row_reduction ] .
+ @xmath140 : this is a consequence of proposition [ prop : matchings ] . +
all conditions are therefore equivalent by transitivity .
we may consider the special case when the labeling provides a successful pressing sequence .
[ thm : main_thm2 ] given a bicolored labeled graph @xmath2 on @xmath85 $ ] , the following are equivalent : 1 .
the vertices of @xmath2 , in the usual order , are a successful pressing sequence .
@xmath47 can be written @xmath141 for some invertible lower - triangular matrix @xmath90 .
3 . every leading principal minor of @xmath47 is nonzero for @xmath133 $ ] .
the induced subgraph @xmath142 $ ] has an odd number of perfect matchings for each @xmath133 $ ] .
@xmath143 for some invertible lower triangular matrix @xmath90 and invertible upper triangular matrix @xmath66 .
all statements are simply specializations of those from the preceding theorem , except for @xmath144 . that @xmath90 ( and not just @xmath47 ) can be taken to be full rank follows from the fact that @xmath145 if @xmath90 is invertible .
[ prop : onebicoloringpersequence ] given a graph @xmath2 with vertex set @xmath85 $ ] and a permutation @xmath146 of @xmath85 $ ] , there is exactly one bicoloring of @xmath2 for which @xmath146 is a valid pressing sequence .
we apply the part of theorem [ thm : main_thm ] that says that @xmath146 is a valid pressing sequence if and only if @xmath147 $ ] has an odd number of perfect matchings for each @xmath148 $ ] .
the proof proceeds by induction on @xmath87 . if @xmath149 , clearly the number of perfect matchings is odd if and only if @xmath150 $ ] has a loop , so @xmath151 must be black in the bicoloring .
suppose the statement is true for @xmath130 .
let @xmath152 denote the number of perfect matchings of @xmath147 $ ] , @xmath153 the number of perfect matchings of @xmath154 $ ] , and @xmath155 the number of perfect matchings of @xmath156 $ ] .
then @xmath157 if @xmath158 does not have a loop , while @xmath159 if @xmath158 does have a loop . since @xmath152 is odd by the inductive hypothesis , exactly one of @xmath155 or @xmath160 is odd , so the color of @xmath158 is uniquely determined . as with real / complex matrices , a matrix @xmath38 over @xmath0
is said to be `` orthogonal '' if @xmath161 .
the set of all @xmath162 orthogonal matrices over @xmath0 is the `` orthogonal group '' and is denoted @xmath163 . [ prop : lpqu ] suppose the bicolored graph @xmath2 with vertex set @xmath85 $ ] has the identity permutation as a pressing sequence , let @xmath164 be the augmented adjacency matrix of @xmath2 , and let @xmath165 be the cholesky decomposition guaranteed by theorem [ thm : main_thm ]
. then @xmath146 is also a pressing sequence of @xmath2 iff there exist an orthogonal matrix @xmath38 and an upper triangular matrix @xmath66 so that @xmath166 where @xmath33 is the permutation matrix encoding @xmath146 . if @xmath167 , then @xmath168 a cholesky decomposition for the matrix @xmath169 . on the other hand ,
suppose @xmath146 is a pressing sequence for @xmath2 .
then @xmath170 for some @xmath66 , whence @xmath171 let @xmath172 ; then @xmath161 , so @xmath38 is orthogonal , and @xmath173 it is worth remarking that one may take @xmath33 to be any permutation matrix representing an automorphism of @xmath2 , @xmath174 , and @xmath175 to obtain a solution to @xmath176 .
indeed , acting on a bicolored graph by an automorphism fixes its pressing sequences .
therefore , by the above proposition , one may view successful pressing sequences as a kind of @xmath0-relaxation of automorphisms . given a matrix @xmath108
, we define a new matrix @xmath177 as follows .
let the columns of @xmath108 be @xmath178 and the columns of @xmath177 , @xmath179 . if @xmath180 have been defined for @xmath130 , we define @xmath181 note that we can also define @xmath182 by @xmath183 suppose the bicolored graph @xmath2 with vertex set @xmath85 $ ] has the identity permutation as a pressing sequence , let @xmath164 be the augmented adjacency matrix of @xmath2 , and let @xmath165 be the cholesky decomposition guaranteed by theorem [ thm : main_thm2 ] .
let @xmath146 be a permutation of @xmath85 $ ] and @xmath33 is the permutation matrix encoding @xmath146 .
then @xmath146 is a valid pressing sequence for @xmath2 iff the all - ones vector @xmath184 is a ( left ) eigenvector of @xmath185 iff @xmath185 is orthogonal .
note that the computation of @xmath177 is precisely that of performing the gram - schmidt algorithm on the columns of @xmath108 , except that no `` normalization '' occurs , i.e. , one does divide by the norm of the resulting vectors .
however , over @xmath0 , there are only two possible `` norms '' : 0 and 1 .
therefore , if the norm of each of the @xmath186 produced in the computation of @xmath177 is @xmath45 for all @xmath126 , the columns of @xmath177 are the same as the output of gram - schmidt orthonormalization , whence we obtain a factorization of @xmath108 of the form @xmath187 with @xmath38 orthogonal and @xmath66 upper triangular .
( this is usually termed a `` qr - factorization '' . ) the only failure mode of this computation is if some @xmath188 is self - orthogonal . since self - orthogonality is equivalent to having an inner product of @xmath44 with @xmath184 , by proposition [ prop : lpqu ] , if @xmath146 is a successful pressing sequence , @xmath185 is orthogonal and @xmath189 ; otherwise , @xmath185 is not orthogonal and @xmath190 .
[ thm : average ] for an ( ordinary ) graph @xmath2 on @xmath43 vertices , let @xmath191 denote the average number of length-@xmath43 successful pressing sequences over all bicolorings of @xmath2 . then @xmath192 we construct a bipartite graph @xmath193 as follows
. one partition class @xmath194 consists of all permutations of @xmath85 $ ] ; the other partition class is @xmath195 , the set of all bicolorings of @xmath2 , which we assume has vertex set @xmath85 $ ] .
we place an edge between a permutation @xmath146 and a bicolored graph @xmath2 iff @xmath146 is a successful pressing sequence for @xmath2 . by proposition [ prop : onebicoloringpersequence ] ,
the @xmath193-degree of each vertex in @xmath194 is @xmath45 , so the number of edges in @xmath193 is @xmath196 . on the other hand ,
the number of edges incident to a bicolored graph is its number of length-@xmath43 successful pressing sequences .
therefore , the sum of all degrees in @xmath195 is @xmath197 .
we may conclude that @xmath192 note that , since a particular labeling of a graph admits precisely one bicoloring so that @xmath198 , the probability that a symmetric matrix ( i.e. , the adjacency matrix of a bicolored graph ) is cholesky @xmath199 .
we present a few open questions on the subject of pressing sequences in addition to the pressing game conjecture ( discussed in the introduction ) .
how hard is it in general to compute the number of successful pressing sequences of a given bicolored graph ? by the remarks following proposition [ prop : onebicoloringpersequence ] , it is perhaps the case that this enumeration problem is gi - complete , i.e. , the same difficulty as certifying graph isomorphism and counting automorphisms .
alternatively , the connection with counting perfect matchings suggests it might be # p - hard .
given that we do not know the complexity of counting pressing sequences exactly , perhaps the approximation problem is easier : is there a polynomial time algorithm for approximating within a small factor the number of successful pressing sequences of a given bicolored graph ? in studying some of these questions , the authors found a substantial , though small , number of nonisomorphic graphs which have exactly one pressing sequence
graphs we term `` uniquely pressable '' .
however , we lack a characterization of these graphs .
describe the uniquely pressable graphs .
thank you to va czabarka and kevin costello for helpful and insightful discussions .
thank you also to the south carolina honors college at the university of south carolina for supporting the present work with their science undergraduate research fellowship .
t. hartman and e. verbin .
matrix tightness : a linear - algebraic framework for sorting by transpositions . in _ proceedings of the 13th international conference on string processing and information retrieval _ , spire06 , 2006 .
h. kaplan , r. shamir , and r. e. tarjan . faster and simpler algorithm for sorting signed permutations by reversals . in _ proceedings of the eighth annual acm - siam symposium on discrete algorithms
_ , soda 97 , pages 344351 , 1997 . | we apply matrix theory over @xmath0 to understand the nature of so - called `` successful pressing sequences '' of black - and - white vertex - colored graphs .
these sequences arise in computational phylogenetics , where , by a celebrated result of hannenhalli and pevzner , the space of sortings - by - reversal of a signed permutation can be described by pressing sequences .
in particular , we offer several alternative linear - algebraic and graph - theoretic characterizations of successful pressing sequences , describe the relation between such sequences , and provide bounds on the number of them .
we also offer several open problems that arose as a result of the present work .
msc classes : 05c50 , 15b33 , 92d15 .
keywords : binary matrix , pressing sequence , adjacency matrix , matching , bicolored graph . |
understanding the process of star formation requires reliable observations of fundamental stellar properties so that theoretical models can be tested .
the copious population of young binary systems , however , complicates the problem .
the binary fraction can be high @xcite , and may depend on the density of the star forming region @xcite and spectral type of the primary @xcite , underscoring the importance of their study .
fortuitously , binary stars serve two purposes .
one , characterization of their frequency , separation distribution , and mass ratio distribution for a given star forming region ( sfr ) provides clues to the broad star forming properties ( angular momentum , density , turbulence , etc . ) of the parent molecular cloud .
two , the special class of very small separation spectroscopic binaries with periods sufficiently short to enable the measurement of the individual stellar velocities , and thus the system s mass ratio , are potential targets for the dynamical determination of individual component stellar masses ( e.g. , steffen et al . 2001 ; prato et al . 2002a ; boden et al .
2005 ; stassun et al .
knowledge of absolute masses and observable properties , such as effective temperature and luminosity , plays a key role in the improvement of pre - main sequence ( pms ) evolutionary models @xcite .
once spectroscopic binaries have been identified , it is necessary to characterize their properties .
it is often the case with low mass - ratio systems that they are identified as single@xmath1lined spectroscopic binaries when observed in visible light ( mazeh et al .
2002 ) , in which case the large difference in flux between the primary and secondary at short wavelengths prevents detection of the secondary component and thus the measurement of the mass ratio for the system . in the raleigh - jeans regime , however , flux scales much less steeply as a function of mass .
thus , by observing single - lined spectroscopic binaries with infrared ( ir ) spectroscopy we are able to improve our chances of detecting the lower - mass secondary not seen in visible light .
this technique was initially outlined in prato et al .
( 2002b ) and mazeh et al .
( 2002 , 2003 ) .
our primary motivation for observing rx j0529.3@xmath01210 was to use this ir approach to determine the mass ratio of the system by converting it into a double - lined spectroscopic binary . given that the secondary had not been detected in visible light , we anticipated a relatively low mass ratio for rx j0529.3@xmath01210 . with the advent of the _ rntgensatellit ( rosat ) _
all - sky survey , a number of researchers began a search for x - ray sources with optical counterparts of brightness consistent with membership in nearby sfrs , motivated in part by the goal of detecting the `` post - t tauri '' population postulated by herbig ( 1978 ) . in follow up observations of x - ray sources near taurus , rx j0529.3@xmath01210
was identified variously as a pms star ( neuhauser et al .
1997 ; magazz et al .
1997 ) and a post - t tauri star ( magazz et al . 1999 ) ; high - resolution visible light spectroscopy revealed its spectroscopic binary nature @xcite .
@xcite undertook a seven year campaign to characterize the orbits of all spectroscopic binaries identified in the x - ray sample of @xcite and @xcite , including rx j0529.3@xmath01210 .
table 1 summarizes the general properties of this system .
the @xmath33 dozen high - resolution visible light spectra taken of this binary suggested the presence of a secondary star ; however , a conclusive identification was not possible .
a single - lined orbital solution was determined , although , probably owing to the extremely high eccentricity of the system the uncertainties are relatively large .
this paper describes the results of using high resolution ir spectroscopy to observe rx j0529.3@xmath01210 and determine the component radial velocities .
rx j0529.3@xmath01210 is the most eccentric pre@xmath1main sequence spectroscopic binary known to date .
this provides a unique context in which to speculate on the formation of the system and on the impact of the stellar dynamics on potential planet formation . in
2 we briefly describe our observations and data reduction . our analysis and results appear in 3 .
section 4 provides a discussion , and 5 summarizes our findings .
our observations were made during six epochs between 2002 january and 2004 december with the keck ii 10-m telescope on mauna kea .
the ut dates of observation are listed in table 2 .
@xmath7-band data , at a central wavelength of @xmath31.555 @xmath5 m , were obtained using the facility near - infrared , cross - dispersed , cryogenic spectrograph nirspec ( mclean et al . 1998 ; 2000 ) .
nirspec employs a 1024 @xmath8 1024 aladdin insb array detector .
we used the 0.288@xmath9 ( 2 pixel ) @xmath8 24@xmath9 slit , yielding an oh night sky emission line determined resolution of r@xmath1025,000 .
source acquisition was accomplished with the slit viewing camera , scam , which utilizes a 256 @xmath8 256 hgcdte detector with 0.18@xmath9 pixels .
rx j0529.3@xmath01210 has a 2mass @xmath7-band magnitude of 9.40 ; integration times for individual frames were between 60 and 240 s. we nodded the telescope 10@xmath9 between two positions on the slit to allow for background subtraction between sequential spectra .
all spectroscopic reductions were made with the redspec package , software written at ucla by s. kim , l. prato , and i. mclean specifically for the analysis of nirspec data , following the procedures outlined by prato et al .
( 2002a ) .
we used the central order , 49 , which covers 1.545 to 1.567 @xmath5 m at our setting ; this order has three advantages .
one , it is rich in both atomic and molecular lines and is therefore suitable for identifying spectra of both warm and cool stars .
two , the oh night sky emission lines across order 49 are numerous and well - distributed , yielding accurate and precise dispersion solutions .
three , this order has the advantage of lacking prominent telluric absorption lines .
consequently , we did not have to divide by telluric standard star spectra .
on ut dates 2004 december 24 and 2008 january 17 images were taken of rx j0529.3@xmath01210 , point spread function ( psf ) stars , and photometric standards at the keck ii telescope with the nirc2 camera behind the adaptive optics ( ao ) system @xcite .
the plate scale was 0.01@xmath9 per pixel and integration times were 0.18 s for the l filter , 0.20 s for the k filter , and 1.0 s for the narrow k - continuum filter .
ten coadds were made per image , and two images were obtained at each location in a five - point box dither pattern .
seeing conditions were excellent for both nights . with the use of customized idl reduction and analysis routines ,
images were flat - fielded , cleaned of bad pixels , and compared with psf stars to search for higher order multiplicity and evidence for extended sources . l and
k - continuum filter data from 2008 january 17 ( ut ) were reduced with standard photometric techniques , including flat - fielding , background subtraction , aperture photometry , and airmass correction .
the spectra from the six epochs of observation are shown in figure 1 .
individual stellar radial velocities were measured by using two - dimensional cross correlation ( e. g. , zucker & mazeh 1994 ) ; observed or synthetic spectral templates were shifted in relative velocity to maximize the correlation with the observed binary spectrum , thus identifying the component radial velocities .
the observed template spectra @xcite that best matched our data were bs 8086 and gl 436 , spectral types k7 and m2.5 , consistent with @xcite .
the corresponding best rotational velocities were 20 km s@xmath11 for the primary and 25 km s@xmath11 for the secondary .
flux ratios determined by cross - correlation range from 0.50 to 0.69 and have an average value of @xmath30.6 .
@xcite employ synthetic spectra in their final radial velocity analysis of the components of the young spectroscopic binary , haro 1 - 14c .
in addition to using observed templates , we also fit the components of rx j0529.3@xmath01210 with synthetic template spectra , calculated from the updated nextgen models @xcite .
this is desirable because observed template spectra have inherent radial velocity uncertainties , whereas the synthetic template spectra have no associated velocity measurement uncertainties .
thus , by using synthetic templates , the uncertainties in the component radial velocities of a spectroscopic binary are likely attributable to factors such as the signal to noise ratio of the observed spectroscopic binary spectra and imperfect matches to model stellar atmosphere spectra . for rx j0529.3@xmath01210
we were unable to determine consistent velocity solutions using synthetic templates for both the primary and secondary stars because of the poor fit of a synthetic secondary to the observed spectra .
however , the combination of a synthetic spectrum for the primary component ( rotated to 30 km s@xmath11 ) and the gl 436 observed template for the secondary ( rotated to 25@xmath130 km s@xmath11 ) was successful .
the velocities derived from this approach appear in table 2 , columns ( 3 ) and ( 4 ) .
the velocity uncertainties associated with the primary star are 0.5 km s@xmath11 and with the secondary star are 2.0 km s@xmath11 .
analysis of the spectra with only the observed templates yields results that are consistent to within 1@xmath4 with the mixed approach presented here . following wilson ( 1941 ) , we plot the six epochs of the primary versus secondary radial velocities ( figure 2 ) , extracted from our ir spectra , and derive a mass ratio of [email protected]@xmath20.05 and a center - of - mass velocity of @xmath1219.01 @xmath2 0.87 km s@xmath11 .
we combined our radial velocities for the primary and secondary stars with the 34 measurements for the primary star presented in @xcite and solved for the best - fit orbital elements in the @xmath13 sense .
the orbital fit is a standard least - squares program using the levenberg - marquardt method taken from @xcite .
initial guesses for the solution are found by using an amoeba search routine also from @xcite .
the orbital elements are given in table 3 .
the mass ratio and the center - of - mass velocity found from the single- and double - lined velocities together are consistent with those derived from figure 2 to within 1@xmath4 , although the mass ratio calculated from the full orbital solution , [email protected] , has a large associated uncertainty . because the primary star velocity measurements dominate the data set , the orbital parameters are similar to those found in @xcite .
determination of the orbital solution from only the double - lined velocities yields results consistent with the combined single- and double - lined solution but with much larger uncertainties .
we describe future work to improve our orbital fit in 4.4 .
figure 3 shows the primary star velocities from @xcite along with ir determined radial velocities for both the primary and secondary stars plotted as a function of phase .
the phases for the ir observations were computed by taking the difference in the heliocentric julian day ( hjd ) from the last observation made by torres et al .
( 2002 ) and the hjds of our observations and dividing by our period ( table 3 ) .
the errors associated with the velocities are typically a few km s@xmath11 for the @xcite data , 0.5 km s@xmath11 for our primary star data , and 2.0 km s@xmath11 for our secondary data .
our orbital solution is also plotted .
clearly there is scatter of up to several sigma in many of the data points , particularly among the visible light data with velocities closer to the center - of - mass velocity of the system , compared to the orbital solution . @xcite
give a typical radial velocity uncertainty for all their targets of 0.5 km s@xmath11 , probably an underestimate in the case of rx j0529.3@xmath01210 .
the ir data are within 1@xmath12 @xmath4 of the orbital solution .
the sample of pms spectroscopic binary star eccentricities as a function of period , based on data from @xcite , is plotted in figure 4 and indicates that rx j0529.3@xmath01210 is the most eccentric pms spectroscopic binary known to date .
the combination of the system orientation and high eccentricity results in fairly low velocity amplitudes during the majority of the orbital period . until the high - velocity cusp of the radial velocity versus phase curve is well - sampled
, the uncertainty in the elements of this system will remain large .
near - ir j@xmath1h and h@xmath1k colors for rx j0529.3@xmath01210 were calculated from 2mass jhk magnitudes and plotted on a color - color diagram .
the location in the color - color plane of this source is consistent with that of a late k or early m dwarf .
there is no near - ir excess evident . from our keck@xmath0nirc2 data we determine a k - continuum magnitude of [email protected] , consistent with 2mass measurements ( table 1 ) , and an l-band magnitude of [email protected] from our 2008 keck ao images .
the resultant k@xmath1l color is thus [email protected] magnitudes . in the ao images from 2008 january 17 ,
rx j0529.3@xmath01210 appears to be extended in comparison with the observed single star psfs . modeling rx j0529.3@xmath01210 as a binary with the single star dn tau used as a psf yields a separation of @xmath14 at a position angle of @xmath15 and a secondary - to - primary k - band flux ratio of @xmath16 .
the reliability of the solution is likely to be low because of the difficulty in measuring separations this far below the diffraction limit .
we also note the possibility that rx j0529.3@xmath01210 could be broadened by a lower ao correction rate during the observations as compared with that of the single star . however , the consistency of the flux ratio with the h - band spectroscopic values determined by cross - correlation provides support that we are seeing evidence of the companion in rx j0529.3@xmath01210 .
we do not see any extension in the 2004 december 24 images of rx j0529.3@xmath01210 , which is consistent with the binary being near periastron , whereas in 2008 january the companion was approaching apastron .
the projected location of rx j0529.3@xmath01210 is associated with a high - density area in the co maps of @xcite for the @xmath6 ori region . however , @xcite find a mean radial velocity for the strong lithium sources identified in @xmath6 ori of 24.5 km s@xmath11 , with a dispersion of only 2.3 km s@xmath11 .
our center - of - mass velocity for rx j0529.3@xmath01210 is 18.38 km s@xmath17 , indistinguishable from that found by @xcite .
thus , on the basis of radial velocities alone , it seems unlikely that the system is associated with @xmath6 ori since rx j0529.3@xmath01210 s center - of - mass velocity is inconsistent with that of @xmath6 ori at the 3@xmath4 level . using the 2mass jhk magnitudes of rx j0529.3@xmath01210 , an effective temperature ( t@xmath18 ) for a k7/m0 of 3900 k @xcite , and the nominal distance to @xmath6 ori ( 400 pc ; barrado y navascus 2005 ) we have estimated the luminosity and placed the system on an h - r diagram .
we find l @xmath10 3.44 @xmath2 0.01 l@xmath19 , yielding an age of @xmath30.1 myr using the pms tracks of @xcite .
accounting for a companion star with luminosity equal to that of the primary results in an age of @xmath30.5 myr . using the tracks of @xcite , or absolute k magnitude and the tracks of
, we obtain similar age estimates .
@xcite estimates the age of @xmath6 ori to be between 3 and 10 myr , @xcite find an age of 5 myr , and @xcite describe a spread in ages from 1 to 10 myr .
a @xmath201 myr object is expected to be at least somewhat embedded in its natal cloud and associated with circumstellar material , yet near - ir colors from 2mass magnitudes indicate zero extinction , no veiling is detected in the spectra , and the h@xmath21 emission line equivalent width ( table 1 ) is only 2 .
we consider two possible mechanisms for disk dissipation in this system , photoevaporation from nearby hot stars and the orbital dynamics of rx j0529.3@xmath01210 itself .
the projected separation of the b8 star hd 36104 @xcite from rx j0529.3@xmath01210 is @xmath31 pc if both are assumed to be at a distance of 400 pc . @xcite
discuss the surprising lack of evidence for accretion disks around the pms stars in the central @xmath6 ori cluster and suggest that photoevaporation and/or possibly a supernova event 1@xmath12 myr ago played a role in the dispersion of disks in the local young low - mass stellar population .
rx j0529.3@xmath01210 , located just north of the central cluster in the barnard 30 dark cloud , could have experienced photoevaporation from the nearby b star at a young age , obliterating circumstellar material , although this is probably unlikely given the long survival time of proplyds in the trapezium @xcite .
given the extremely high eccentricity of this binary , however , the action of the companion star may have precluded formation of , or dissipated any , circumstellar material . with an assumed primary star mass of 0.75 m@xmath19 and the mass function from torres et al .
( 2002 ) , we find a minimum secondary star mass of 0.40 m@xmath19 , and thus a minimum total mass of 1.15 m@xmath19 . in conjunction with the 461.89 day period , this yields a periastron separation of 0.15 au and an apastron separation of 2.30 au .
it is unknown if orbital evolution may have occurred , or even whether it is possible that this unusual system formed by capture , but if rx j0529.3@xmath01210 is located in @xmath6 ori with an age of @xmath221 myr , then it seems likely that the system formed in a similar configuration to its present one .
independent of the potentially harsh @xmath6 ori environment and the orbital dynamics of rx j0529.3@xmath01210 , the system manifests spectra of similar surface gravity to dwarf stars , as well as a relatively small lithium equivalent width , in comparison to classical t tauri stars @xcite .
the lithium and h@xmath21 equivalent widths of @xcite for @xmath6 ori average greater than twice what is found for rx j0529.3@xmath01210 and are consistent with the classical and weak - lined t - tauri limits defined by @xcite .
according to these limits , rx j0529.3@xmath01210 would be a post - t tauri star and have one of the lowest lithium and h@xmath21 equivalent widths in the @xmath6 ori region .
these characteristics are inconsistent with an age of @xmath221 myr .
alternatively , rx j0529.3@xmath01210 could be an older , closer object .
@xcite describes a new candidate moving group , 32 ori , consisting of a small cluster of x - ray bright , late type stars .
the @xmath310 young stars identified in the group are located around 5@xmath23 20@xmath24 to 5@xmath23 30@xmath24 and @xmath25 , at the proposed distance of @xmath390 pc .
rx j0529.3@xmath01210 is about 9 pc from the central clump of objects in 32 ori and is one of the members used by mamajek to define the common proper motion group ( e. mamajek 2008 , private communication ) .
given the proper motion of rx j0529.3@xmath01210 , pmra @xmath10 4.1 @xmath2 5.8 mas / yr and pmdec @xmath10 @xmath130.7 @xmath2 5.8 mas / yr , compared to the proper motions of the stars @xmath6 ori ( pmra @xmath10 0.8 @xmath2 1.5 , pmdec @xmath10 @xmath12.3 @xmath2 1.5 ) and 32 ori ( pmra @xmath10 6.57 @xmath2 1.15 , pmdec @xmath10 @xmath132.45 @xmath2 0.48 ) , evidence in support of membership in the 32 ori group is compelling . again combining the 2mass jhk magnitudes of rx j0529.3@xmath01210 , an effective temperature of 3900 k , and a distance now of 90 pc , we find l @xmath10 0.17 @xmath2 0.02 l@xmath19 , giving an age for rx j0529.3@xmath01210 on the tracks of @xcite of @xmath315@xmath25 myr , consistent with the 25@xmath210 myr age derived by @xcite for the candidate group members .
@xcite lists a group radial velocity of 18 km s@xmath11 , in excellent agreement with our measured center - of - mass velocity ( table 2 ) .
morales calderon ( 2008 , in prep ) has observed the @xmath6 ori region with the _ spitzer _ space telescope and finds a 3 @xmath4 detection of rx j0529.3@xmath01210 at 24 @xmath5 m with a 20 % excess above a 3900 k photosphere . for a star with [email protected] l@xmath19
, the equilibrium temperature of a black - body grain with peak emission at 24 @xmath5 m corresponds to a distance of 4.36 au .
a more realistic treatment of the dust grain distribution would necessarily take into account the additional flux from the secondary star , yielding a larger distance from the center - of - mass of the system to the putative dust .
however , given the estimated apastron separation of 2.30 au , a lower limit of 4.36 au for the dust radius from the system center illustrates the plausibility of a circumbinary debris disk .
additional _ spitzer _
observations at shorter wavelengths have been taken and should reveal more information regarding the extent and location of the dust ( mamajek 2008 , in preparation ) .
the lack of j@xmath1h , h@xmath1k , and k@xmath1l excesses is hardly surprising . for black - body grains with a peak wavelength in the l band ,
the corresponding disk temperature occurs at a distance of @xmath30.1 au , nearly coincident with the binary periastron .
stable dust in this system is most likely to be located in a circumbinary distribution .
furthermore , if the closer distance and therefore older age for this system implied by membership in 32 ori are correct , then the presence of an evolved debris disk would not be unusual ( e.g. , * ? ? ?
finally , the fact that the rx j0529.3@xmath01210 secondary was possibly detected in our ao images strongly supports the @xmath390 pc distance . for the projected separation of 0.018 " , determined using psf fitting , and a distance of 90 pc , the corresponding distance in au is [email protected] , not too different from our estimated apastron of 2.30 au . for a distance of 400 pc the separation would be [email protected] au .
based on the currently available data , it therefore appears that rx j0529.3@xmath01210 is associated with the new 32 ori group , not with @xmath6 ori , and has an age of @xmath315 myr .
@xcite note that the appearance of the velocity correlation function for rx j0529.3@xmath01210 suggests the presence of at least one other star in the system . for a primary star mass of 0.75 m@xmath19
, the mass function implies a minimum secondary star mass of 0.40 m@xmath19 and a minimum mass ratio of 0.53 , consistent to within 1@xmath4 of the mass ratio found from the full orbital solution for the system , [email protected] . the approximate h - band flux ratio measured from the cross - correlation is [email protected] .
the k - band flux ratio found in the best psf fit to the january , 2008 ao images is [email protected] .
for a m2.5 @xmath0 k7 pair , the spectral types which provided the best correlation ( 3 ) , models of @xcite and @xcite imply an h - band flux ratio of [email protected] based on components with a primary t@xmath263900 k and a secondary t@xmath263400 k @xcite . taken together ,
these properties suggest a relatively red color for the secondary star , hindering detection in visible light .
the ir primary and secondary velocities , presented in the wilson plot ( figure 2 ) , imply a range of mass ratios , from [email protected] , well within the range we obtain by combining the ir and visible light data in a full orbital solution , @[email protected] . because the full solution includes phase information , it is generally more reliable , although for rx j0529.3@xmath01210 the sparse phase coverage during the epochs of largest velocity separation yields large uncertainties in the orbital solution . clearly more data are merited . for the case of a mass ratio relatively close to unity in this system ( i.e. [email protected] ) ,
it is necessary to explain the lack of a visible light detection of the secondary star . in the following paragraphs we outline several scenarios that would account for this .
a relatively red secondary and yet a mass ratio close to unity is possible if a third component is present , as first suggested by @xcite . if the secondary star is actually a binary pair , with a very small separation and similar component masses , then the spectral types could be consistent with the best - fitting todcor templates
. this inner binary pair would necessarily be in an orbit in a plane relatively perpendicular to our line of sight , such that the large velocity changes would not be obvious .
this scenario would imply a near - unity mass ratio yet a smaller near - ir flux ratio , and a red enough secondary system to evade visible light detection .
the drawback is that it requires a very specific geometry .
such a companion might also account for the 24 @xmath5 m excess .
a heavily spotted secondary star could mimic a lower temperature source , resulting in a best fitting secondary m2.5 template , a mass ratio close to 1.0 , and a flux ratio @xmath221 in the near - ir .
again , the reddening effect of the cooler temperature in the spot covered areas would increase the difficulty of visible light detection . for two t@xmath263900 k stars , if one has 50% of its surface covered with a 2900 k spot @xcite , the secondary / primary bolometric flux ratio would be @xmath30.7 . for objects with ages between 10 and 30
myr , some models @xcite of pms evolution indicate that targets in the mass range of the rx j0529.3@xmath01210 primary star , [email protected] m@xmath19 , are located near the transition between the convective hyashi track and the radiative henyey track . as a result ,
models show a sharp turn in the mass tracks towards higher temperatures . a similar but slightly lower mass secondary star that has not yet turned this corner off of the hyashi track , and indeed , objects with masses lower than @xmath30.5 m@xmath19 never will
, may have a similar mass as the primary star but a much lower temperature .
these models for the system behavior are highly speculative .
the absence of an unambiguous secondary star detection in visible light suggests that the mass ratio is _ not _ close to unity . a value in the 0.7 to 0.8 range is most likely based on data to date .
the formation of such a highly eccentric multiple is challenging to understand as a primordial event given current theories of star formation @xcite .
figure 4 shows that pre@xmath1main sequence spectroscopic binaries are observed over a wide range of eccentricities .
although rx j0529.3@xmath01210 represents the maximum of that range , it does not stand out particularly from the eccentricity distribution .
thus , some form of dynamical evolution may have taken place in this system , but , if so , it is not distinguished by an outlying eccentricity .
possibilities for dynamical evolution range from an improbable capture event to disk excitation of stellar eccentricity @xcite .
if indeed this system is @xmath2710 myr old , only fossil evidence , such as the detected 24 @xmath5 m excess , might remain of the primordial , presumably massive disk(s ) responsible . in a variation of the interactions proposed by @xcite ,
a three body dynamical encounter between an object from outside of the binary and a companion short period binary could have stimulated the eccentricity of the system and tightened the orbit of a close ( secondary ) pair ( 4.2 ) .
the eventual formation of a circumbinary ( or circumtriple ) debris disk could have been stimulated by dynamical disruption of the disks in the system in any one of these scenarios .
the formation of planets in binary systems has been observationally supported in recent years and is of primary significance since a majority of stars have companions @xcite . @xcite
have determined the strict criteria for classifying stable and unstable planetary orbits in binary systems , however , their analysis is restricted to circular orbits .
many of the extra - solar planets observed in binaries are in wide separation systems , where the mass ratio for the stellar components is near unity and the eccentricity of the system is low . @xcite have modeled the formation of terrestrial planets around individual stars in binaries and find that , for periastron distances of @xmath225 au , such planet formation is restricted .
considering that the periastron distance in the rx j0529.3@xmath01210 system is only 0.15 au , we conclude that the probability of even a low - mass circumstellar planet in rx j0529.3@xmath01210 is remote .
the trend determined by @xcite towards the formation of fewer circumbinary terrestrial planets in systems with apastron distances of @xmath270.2 au and non - zero eccentricities also bodes for poor circumbinary planetary stability around the stars in rx j0529.3@xmath01210 .
clearly it is imperative to improve the orbital solution for rx j0529.3@xmath01210 , particularly the measurement of the mass ratio .
visible light and especially infrared data are critical to obtain , preferably at high precisions .
rx j0529.3@xmath01210 passes through its next maximum velocity separation in late december , 2009 . densely sampled high signal - to - noise spectra taken during these epochs will yield a greatly improved precision for the orbital solution .
if rx j0529.3@xmath01210 is located at a distance of only 90 pc , an apastron separation of 2.30 au implies a maximum projected separation on the sky of @xmath30.03@xmath9 .
much of the orbit of this system lies within reach of the diffraction limit of the 85 m baseline of the keck interferometer , 0.005@xmath9 .
although the k@xmath1band magnitude of rx j0529.3@xmath01210 , 9.2 , is relatively faint for observations with this facility , planned improvements to the system may eventually enable resolved observations of this unusual binary , permitting the determination of individual component masses .
the 90 pc distance may also be confirmed by measuring the parallax , which is slightly greater than 10 mas at this distance , and is at the achievable limit of current instrumentation .
@xcite reported the orbital solution for rx j0529.3@xmath01210 based on single - lined spectroscopic data ; we have elaborated on this solution using near - ir data to identify the spectrum of the secondary star and determine the mass ratio of the system .
the ir data alone presented in a primary vs. secondary velocity plot ( figure 2 ) indicate a mass ratio of [email protected] . in concert with the visible light data , the full orbital solution yields a more uncertain mass ratio of [email protected] ( table 3 ) .
other orbital parameters are in good agreement with the results of @xcite .
this system is the most eccentric pre@xmath1main sequence double@xmath1lined binary known to date ( figure 4 ) and thus presents challenges to observation .
it is an unlikely site for the formation of circumstellar planets , or of circumbinary planets with orbits within several au of the stars .
the fact that the secondary component was not identified in visible light suggests a relatively faint and/or red companion . for an estimated primary mass of 0.75 m@xmath19
the minimum mass of the secondary is 0.40 m@xmath19 .
this implies a minimum mass ratio of 0.53 , consistent with the value in table 3 . a robust determination of the mass ratio along with the best estimate for the primary star mass will eventually provide an improved estimate for the secondary mass .
on the basis of the center - of - mass velocity , common proper motion , a lack of near - ir excess , a tentative k - band ao detection of the secondary , lithium equivalent width , and 24 @xmath5 m _ spitzer _ excess evidence for a debris disk , we argue that the rx j0529.3@xmath01210 system is a member of the 32 ori moving group , recently identified by @xcite . assuming a distance of 90 pc , we estimate an age for rx j0529.3@xmath01210 of 15@xmath25 myr , consistent with that of @xcite for 32 ori .
it is imperative to improve the orbital elements of rx j0529.3@xmath01210 in order to determine a precise mass ratio for the system . in conjunction with future interferometric observations
, this will ultimately yield component masses for this system .
given its unusual eccentricity , proximity , young age , and debris disk , such observations are likely to reveal clues to the formation of this unique binary system as well as additional data for the improvement of pre - main sequence models .
the authors are grateful to k. kilts for conducting the preliminary reduction of these data , to t. barman and c. johns - krull for helpful discussions , and to g. torres for catching an error in our original manuscript .
we thank m. morales calderon , d. barrado y navascus , and j. stauffer for sharing the results of their _ spitzer _ mips data in advance of publication .
we appreciate the input of e. mamajek , on the traits of the 32 ori moving group , and we are grateful to the anonymous referee for detailed and thoughtful recommendations which improved this paper .
this research was funded by nsf grant ast 04 - 44017 ( to lp ) and the associated reu supplement , and by nasa space grant , through northern arizona university , which has provided support for gnm , as well as nsf grant ast 06 - 07612 ( to ms ) .
this work made use of the simbad reference database , the nasa astrophysics data system , and the data products from the two micron all sky survey , which is a joint project of the university of massachusetts and the infrared processing and analysis center / california institute of technology , funded by the national aeronautics and space administration and the national science foundation .
data presented herein were obtained at the w.m .
keck observatory from telescope time allocated to the national aeronautics and space administration through the agency s scientific partnership with the california institute of technology and the university of california .
the observatory was made possible by the generous financial support of the w.m .
keck foundation .
we recognize and acknowledge the significant cultural role that the summit of mauna kea plays within the indigenous hawaiian community and are grateful for the opportunity to conduct observations from this special mountain .
lc r.a.(j2000 ) = 05:29:18.8 + dec.(j2000 ) = + 12:09:30 + v ( mag ) = 12.86 + j ( mag ) = 10.05 @xmath2 0.020 + h ( mag ) = 9.40 @xmath2 0.023 + k ( mag ) = 9.19 @xmath2 0.020 + spt = k7-m0 + h@xmath21 ew ( ) = @xmath12.0 + li ew ( ) = 0.27 ,
0.35 + v@xmath28sin@xmath29 @xmath3018 @xmath2 3 km s@xmath11 + lccrrl 2002 jan 1 & 2452275.88 & 24.7 & 12.3 & 0.183 + 2002 feb 5 & 2452310.87 & 26.9 & 13.9 & 0.258 + 2002 dec 14 & 2452622.97 & 19.5 & 14.9 & 0.933 + 2004 jan 28 & 2453032.76 & 15.4 & 19.7 & 0.818 + 2004 dec 26 & 2453365.99 & 2.3 & 42.3 & 0.538 + 2008 sep 15 & 2454725.15 & 15.3 & 17.9 & 0.498 + lc @xmath31 days + @xmath32 km s@xmath11 + @xmath33 km s@xmath11 + @xmath34 km s@xmath11 + @xmath35 + @xmath36 degrees + @xmath37 mjd + @xmath38 sin@xmath39 + @xmath40 sin@xmath41 + @xmath42 + @xmath43 sin @xmath44 km + @xmath45 sin @xmath46 km + | the young system rx j0529.3@xmath01210 was initially identified as a single@xmath1lined spectroscopic binary . using high@xmath1resolution infrared spectra , acquired with nirspec on keck ii
, we measured radial velocities for the secondary .
the method of using the infrared regime to convert single@xmath1lined spectra into double@xmath1lined spectra , and derive the mass ratio for the binary system , has been successfully used for a number of young , low - mass binaries . for rx
j0529.3@xmath01210 , a long - period ( 462 days ) and highly eccentric ( 0.88 ) binary system , we determine the mass ratio to be 0.78 @xmath2 0.05 using the infrared double - lined velocity data alone , and 0.73 @xmath2 0.23 combining visible light and infrared data in a full orbital solution . the large uncertainty in the latter is the result of the sparse sampling in the infrared and the high eccentricity : the stars do not have a large velocity separation during most of their @xmath31.3 year orbit .
a mass ratio close to unity , consistent with the high end of the 1@xmath4 uncertainty for this mass ratio value , is inconsistent with the lack of a visible light detection of the secondary component .
we outline several scenarios for a color difference in the two stars , such as one heavily spotted component , higher order multiplicity , or a unique evolutionary stage , favoring detection of only the primary star in visible light , even in a mass ratio @xmath31 system .
however , the evidence points to a lower ratio . although rx j0529.3@xmath01210 exhibits no excess at near - infrared wavelengths , a small 24 @xmath5 m excess is detected , consistent with circumbinary dust . the properties of this binary and its membership in @xmath6 ori versus a new nearby stellar moving group at @xmath390 pc are discussed .
we speculate on the origin of this unusual system and on the impact of such high eccentricity , the largest observed in a pre@xmath1main sequence double@xmath1lined system to date , on the potential for planet formation . |
the joint analysis of different types of big data should provide a better understanding of the molecular mechanisms underlying cancers @xcite , @xcite , @xcite , @xcite , @xcite and @xcite . in this work
we introduce an approach that involves the use of matrix factorizations such as the singular value decomposition ( svd ) to simultaneously analyze a finite number of matrices by iteratively identifying cross - correlations between them . the iterative singular value decomposition ( isvd )
jointly analyzes a finite number of data matrices by automatically identifying any correlations that may exist among the rows of matrices , or the rows and columns of any one of the matrices . since big data sets contain many distinct signals that associate the measured variables and samples , repeated application of the isvd on re - normalized data iteratively detects these signals as orthogonal correlations among the matrices or within a single matrix
this approach is computationally efficient .
the following notation is used in this work .
* notation : * * @xmath0 the linear space of real column vectors with @xmath1 coordinates . * @xmath2 the linear space of real @xmath3 matrices . *
@xmath4 the subset of @xmath5 real orthonormal matrices . * @xmath6 the subset of @xmath3 real non - negative diagonal matrices . * @xmath7 the subset of @xmath3 real matrices whose @xmath8 columns are normalized .
the singular value decomposition of a matrix @xmath9 is the decomposition of @xmath10 into the product of three matrices of the form @xmath11 where @xmath12 , @xmath13 , and @xmath14 .
the diagonal entries of @xmath15 are called the singular values of @xmath10 , and columns of @xmath16 and @xmath17 are called left and right singular vectors of @xmath10 respectively .
the left - singular vectors are eigenvectors of @xmath18 and right - singular vectors are eigenvectors of @xmath19 .
non - zero singular values of @xmath10 are same as the square roots of non - zero eigenvalues of @xmath18 ( which are the same as the non - zero eigenvalues of @xmath19 ) . based on the svd , the matrix @xmath10 can be written as an outer product of orthogonal rank one approximations @xmath20 where @xmath21 and @xmath22 are the @xmath23th columns of @xmath16 and @xmath17 respectively ; @xmath24 is the @xmath23th diagonal element of @xmath15 . given a set of @xmath25 matrices @xmath26 with full column rank .
the svd of stacked matrix @xmath27 will be given by @xmath28 the @xmath29th column of @xmath30 can be written as @xmath31 where @xmath32 .
if @xmath33 represent the @xmath29th ( @xmath34 ) columns of @xmath35 , then @xmath36 can be viewed as correlation detectors of the signal @xmath33 that is common to @xmath37 .
that is , we can detect the rows of @xmath37 that are mutually correlated with the common signal @xmath33 by applying thresholds to the vectors @xmath36 .
note that the above procedure will result in @xmath8 distinct signals that link the rows of @xmath37 via the common signals @xmath33 .
the analysis of the rows of @xmath37 will be based on @xmath35 .
specifically , given the rank one approximation to the data we form the residual data and apply rank one approximation to the residual data pair .
this procedure is iterated until the largest singular value of residue falls below a certain threshold that is near zero .
this algorithm is based on the svd of the stacked matrix @xmath27 .
define @xmath38 ( @xmath39 ) to be the @xmath23th column of matrix @xmath35 ( @xmath30 ) and @xmath24 be the @xmath23th diagonal entry of matrix @xmath15 in ( [ eq : svd ] ) .
the first column of @xmath30 can be written as @xmath40 where @xmath41 .
the rank one approximation to svd of @xmath27 is given as @xmath42 @xmath43 @xmath44 @xmath45 with @xmath46 @xmath47 we are interested to find the support of signals in @xmath37 .
the vectors @xmath48 in the rank one approximation of @xmath37 can be viewed as detectors of the signal in @xmath37 , respectively .
that is , the components of @xmath49 with large absolute magnitude correspond to the rows in @xmath37 that are highly correlated with @xmath50 .
note that @xmath51 is common in the matrix factorization of @xmath37 .
the indices of @xmath49 with large absolute magnitude are the signal support for @xmath37 respectively .
we apply the universal threshold proposed by donoho and johnstone ( @xcite and @xcite ) to detect components of @xmath49 with large absolute magnitude .
now we subtract @xmath52 from @xmath37 respectively and get the rank one approximation to svd of the pair @xmath53 .
it is shown in theorem @xmath54 that @xmath55 the rank one approximation to svd of @xmath53 is given as @xmath56 @xmath57 @xmath44 @xmath58 with @xmath59 @xmath60 @xmath61 similarly to the first iteration , the signals in @xmath62 are detected .
the sequential application of the isvd is continued until the maximum singular value of @xmath63 drops below a predefined threshold indicating that the noise floor of the data has been reached .
the number of iterations is at most rank of @xmath27 .
therefore , the isvd algorithm is computationally efficient and enhances signal detection by rank one approximation . to get the largest singular value and singular vector of @xmath10 we can use the matlab command @xmath64 . for further information about the algorithm of finding a few eigenvalue of a large matrix
please refer to @xcite .
* theorem 1 : * let the svd of the stack matrix be @xmath65 then , splitting apart the first column of @xmath30 , we can write @xmath66 satisfying @xmath67 let @xmath68 which latter equality holds since @xmath69 the claim is that * the nonzero eigenvalues of @xmath70 are @xmath71 , * the nonzero eigenvalues of @xmath72 are @xmath73 , * with the corresponding eigenvalues @xmath74 , so that the above representation of @xmath75 is its singular value decomposition .
* proof : * the proof is in the calculation @xmath76 and similarly @xmath77 , etc . the algorithm proceeds to find @xmath78 satisfying @xmath79 * theorem 2 : * let @xmath80 be an @xmath81 unit vector and @xmath82 be a set of orthonormal vectors .
define two matrices @xmath83 ( @xmath84 ) and @xmath85 ( @xmath86 ) as follows @xmath87 and @xmath88 where the support of @xmath89 and @xmath90 in @xmath10 are @xmath91 and @xmath92 respectively and the support of @xmath89 in @xmath85 is @xmath93 .
define @xmath75 as the stack of two matrices @xmath83 and @xmath85 @xmath94 it is claimed that * @xmath95 and @xmath92 are the eigenvalues of @xmath96 corresponding to eigenvector @xmath89 and @xmath90 respectively . * if @xmath97 then @xmath98 and @xmath99 are the eigenvectors of @xmath96 corresponding to eigenvalue @xmath97 . proof : @xmath100 @xmath101 by multiplying @xmath89 to both sides of ( [ eq1 ] ) @xmath102 similarly for @xmath90 .
suppose @xmath97 , by multiplying @xmath89 to both sides of ( [ eq1 ] ) @xmath103 @xmath104 similarly for @xmath99 .
* theorem 3 : * let @xmath105 be an @xmath81 unit vector and @xmath106 be a set of i.i.d .
white noise vectors , in which @xmath107 is a @xmath81 vector .
define the matrices @xmath10 ( @xmath108 ) as follows @xmath109 it is claimed that * @xmath1 is the approximate eigenvalues of @xmath110 corresponding to eigenvector @xmath105 for a non - trivial range of signal to noise ratio .
proof : [ eq : eq1111 ] a^ta & = s+n_1 & & s+n_m & n_m+1 & & n_p s^t+n_1^t + + s^t+n_m^t + n_m+1^t + + n_p^t + & = ss^t+n_1s^t+sn_1^t+n_1n_1^t + & + & + ss^t+n_ms^t+sn_m^t+n_mn_m^t + & + n_m+1n_m+1^t++n_pn_p^t by multiplying @xmath111 from right hand side to both sides of ( [ eq : eq1111 ] ) [ eq : eq2222 ] a^ta = s+ it is easy to see that if s or [ eq : eq22221 ] then the right hand side of ( [ eq : eq2222 ] ) can be approximated by @xmath105 .
it is noticeable that if @xmath112 , then with a high probability the inequality in ( [ eq : eq22221 ] ) is satisfied . below the convergence behaviour of @xmath113 as @xmath1 grows
is studied . by taking limit from both sides of ( [ eq : eq2222 ] ) we have [ eq : eq222222 ] a^ta = s+=s by law of large number we have [ eq : eq3333 ] pr(=0)=1 from ( [ eq : eq222222 ] ) and ( [ eq : eq3333 ] ) we can have almost surely [ eq : eq4444 ] a^ta = s therefore almost surely for any given @xmath114 there exists an @xmath115 such that for all @xmath116 [ eq : eq5555 ] |a^ta - s| < and almost surely [ eq : eq6666 ] m(s- ) < a^tas <
m(s+ ) . from chebychev s inequality [ eq : eq6666 ] p(|_i=1^mn_i|
> ) = let @xmath117 be the weakest signal to noise ratio that thresholding strategy works , @xmath118 .
[ eq : eq6666 ] p(|_i=1^mn_i| > ) = = what is the critical @xmath117 ?
critical @xmath117 can be achieved from roc curve and donoho johnson threshold . for critical @xmath117
we can find an upper bound on probability of error for any value of error .
the maximum probability of error is a function of signal to noise ratio , because the critical @xmath117 is a function of signal to noise ratio and @xmath1 is the power of signal .
the isvd algorithm is able to detect signals that are common or unique in finite number of data sets .
signals are detected sequentially based on signal strength .
very weak signals with small singular values are detectable since noise background is systematically reduced relative to signal strength . note that the isvd is computationally feasible in situations where the gsvd is not
. it may also be theoretically more meaningful than the generalized singular value decomposition ( gsvd ) @xcite , @xcite , @xcite and @xcite . a zeinalzadeh , t wenska , g okimoto , integrated analysis of multiple high - dimensional data sets by joint rank-1 matrix approximations , 54th ieee conference on decision and control ( cdc ) , pp . 38523857 , 2015 .
g okimoto , a zeinalzadeh , t wenska , m loomis and etc . ,
joint analysis of multiple high - dimensional data types using sparse matrix approximations of rank-1 with applications to ovarian and liver cancer , biodata mining 9 ( 1 ) , 24 , 2016 .
s. p. ponnapalli , m. a. saunders , c. f. van loan , orly alter , a higher - order generalized singular value decomposition for comparison of global mrna expression from multiple organisms , science , 6(12 ) , 2011 . | we develop an iterative version of the singular value decomposition ( isvd ) that jointly analyzes a finite number of data matrices to identify signals that correlate among the rows of matrices .
it will be illustrated how the supervised analysis of a big data set by another complex , multi - dimensional phenotype using the isvd algorithm could lead to signal detection . |
algorithms in invariant theory @xcite become inefficient for lie groups interesting for physics . to overcome this problem , we try to use the following well known fact .
every automorphism of a lie algebra root system ( in particular , an element of the weyl group @xmath7 ) defines an automorphism of the cartan subalgebra @xmath8 and , subsequently , can be extended to the automorphism of the whole lie algebra @xmath9 @xcite .
for example , let us consider @xmath10 algebra with the standard basis @xmath11 $ ] and define the adjoint action @xmath12 $ ] . the operator @xmath13 induces a second order inner automorphism of @xmath10 : @xmath14 the action of @xmath15 on to the cartan subalgebra @xmath16 corresponds to a weyl group @xmath17 reflection of the 1-dimensional @xmath10-root system : @xmath18 this paper is organized as follows .
section 2 illustrates a method on the construction of two casimir invariants of the lie group @xmath19 . an adjoint representation for the lie algebra @xmath6
is done in the subsection 2.1 . in 2.2 , using a simple algorithm ( @xcite , 4.3 ) , we construct a hilbert basis of invariants for a maximal torus @xmath20 .
subsections 2.3 and 2.4 describe two main steps of our method . in 2.3 ,
we construct a set of the @xmath15-operators extending the action of the weyl group for the @xmath6 root system and apply corresponding reynolds operator to the torus invariants .
derived invariants we call _ weyl blocks_. in 2.4 , we join them into linear combinations which are invariant relative to the adjoint action of @xmath19 . in section 3 , @xmath21-invariants in the algebra @xmath6 are found . in 3.1 ,
an embedding of @xmath21 into @xmath19 is used to construct all elements necessary for our method . in 3.2 ,
we list a set of invariants and prove that it is a fundamental system of @xmath21-invariants in the adjoint representation of the algebra @xmath6 .
the dimension of @xmath6 is equal to 8 and the rank is 2 and , correspondingly , two @xmath10-triples generate the algebra : @xmath23 where @xmath24 are unit matrices .
these operators with two additional @xmath25 , \qquad y_{3}:=[y_2,y_1]\ ] ] form the cartan - weyl basis of @xmath6 .
below we use the following notations for the basis elements : @xmath26 consider @xmath6 as a vector space @xmath27 with the basis @xmath28 corresponding to the generators @xmath29 .
the adjoint representation of @xmath6 @xmath30=c_{ij}{}^k { \mathrm{e } } _ k\ ] ] induces a linear transformation @xmath31 in @xmath27 by the matrix @xmath32 : @xmath33 where @xmath34 numerates rows , @xmath35 numerates columns of the matrix @xmath36 and @xmath37 is the matrix of the cartan metric on @xmath6 : @xmath38 the adjoint action of the lie group @xmath19 on its algebra @xmath6 @xmath39 can be represented as a linear transformation of @xmath27 by one - parameter groups : @xmath40)^{k}{}_j\,.\ ] ] in this section we construct polynomial @xmath19-invariants in variables @xmath41 a @xmath42-dimensional torus @xmath43 acting on the ring @xmath44=\mathbb{c}[x_1,{\ldots } , x_n]$ ] is isomorphic to a group of the diagonal matrices @xmath45 where @xmath46 are torus variables and @xmath47 have the weights @xmath48 @xmath49 maps monomials into monomials and its invariant ring @xmath44^{\mathcal{t}}$ ] is monomial .
suppose @xmath50 for the lie group @xmath0 .
then the algebra of @xmath0-invariants is a subalgebra of torus invariants @xmath51^{{\mathrm } { g}}\subset \mathbb{c}[v]^{\mathcal{t}},\ ] ] i.e. a g - invariant polynomial is a linear combination of @xmath49-invariant monomials . by definition
a weight @xmath52 of an element @xmath53 is the set of values of the corresponding root @xmath54 on the elements @xmath55 of the cartan subalgebra @xmath8 .
the values @xmath56 can be directly read from the commutation relations in the cartan - weyl basis : @xmath57=\alpha_i(h_j)x_i.\ ] ] weights of the variables ( [ variables ] ) are ones of the generators ( [ generators ] ) @xmath58 an efficient algorithm 4.3.1 from @xcite allows to find the _
hilbert basis _
@xcite , i.e. a minimal generating set , for the torus invariants .
we realize it in the following way .
let us construct the @xmath42-dimensional set @xmath59 of boxes marked by weights of the variables @xmath47 and filled in with the corresponding @xmath47 considered as initial monomials .
after that we produce new monomials degree by degree multiplying on @xmath47 monomials derived at the previous step .
if a new monomial @xmath60 with the weight @xmath61 is not divisible on derived early monomials with the same weight @xmath62 we put @xmath63 into the corresponding box .
we discard a new monomial if its weight is not in @xmath59 .
an output of the algorithm is the set of @xmath49-invariant monomials with the weight @xmath64 . for @xmath65 the two dimensional torus
@xmath66 ( [ tor ] ) is the following : @xmath67 and the hilbert basis consists from the monomials : @xmath68 the root system @xmath69 of @xmath6 algebra lies in the euclidean plane and has six roots @xmath70 ( see figure 1 ) .
a group of the reflections @xmath71 relative to hyperplanes orthogonal to the roots @xmath54 , i.e. the weyl group @xmath72 , is the permutation group @xmath73 @xcite @xmath74 the permutations @xmath75 and @xmath76 correspond to the reflections @xmath77 and @xmath78 ( relative to hyperplanes shown by dashed lines in the figure 1 ) respectively and generate @xmath73 according to formulas : @xmath79 root system ] suppose we treat @xmath6 algebra as a linear space @xmath27 according to the subsection [ sect : adjoint ] .
the reflections @xmath77 and @xmath78 can be extended to linear operators in @xmath27 analogously to @xmath10 case ( see introduction ) @xmath80 \exp[-\tilde{a}_{y_1 } ] \exp[\tilde{a}_{x_1 } ] \label{s1}\\ & s_2=\exp[\tilde{a}_{x_2 } ] \exp[-\tilde{a}_{y_2 } ] \exp[\tilde{a}_{x_2 } ] \label{s2}\end{aligned}\ ] ] as we have two @xmath10-triples ( [ generator1]),([generator2 ] ) in @xmath6 .
transformations by @xmath81 of basis elements of @xmath6 @xmath82 are shown in the following tables : @xmath83 @xmath84 where @xmath85 . as a proof of correctness of extension
we compare the action of @xmath86 on the cartan subalgebra @xmath87 and weyl group reflections of the dual roots @xmath88 in the @xmath6 root space @xmath69 : @xmath89 @xmath90 taking a cartan decomposition @xmath91 we see from the tables ( [ s1trans]),([s2trans ] ) that every one - dimensional subspace @xmath92 is transformed according to the reflection @xmath93 : @xmath94 .
let us consider matrices @xmath95 where we construct @xmath96 following to ( [ gen13]),([gen123]),([gen132 ] ) respectively .
@xmath97 are not a representation of @xmath98 @xcite .
@xmath81 generate a non - abelian matrix group of order 24 but we postpone investigation of this group .
we define a reynolds operator @xmath99 on a polynomial @xmath100 in variables ( [ variables ] ) @xmath101 where transformations @xmath102 correspond to the matrices @xmath103 according to ( [ s_transformation ] ) .
application of @xmath104 to the torus invariants ( [ sl3torinvs ] ) yields polynomials which we call _ weyl blocks _
@xmath105 ( @xmath35 numerates weyl blocks of degree @xmath106 ) .
linear combinations of the torus invariants give _ initial _ weyl blocks : @xmath107 taking products of torus invariants we derive additional blocks : @xmath108 the group @xmath4 is generated by eight one - parameter subgroups @xmath110 ( [ adgrouplineartrans ] ) corresponding to the basis elements ( [ generators ] ) .
a difference of a weyl block @xmath105 under the action of @xmath111 has a part which is linear in the parameter @xmath112 : @xmath113 a homogeneous linear combination @xmath114 of the weyl blocks with numerical coefficients @xmath115 is @xmath4-invariant if and only if the sum @xmath116 is equal to zero for all @xmath111 .
now let us take into account that @xmath117 are polynomials in the variables @xmath118 ( [ variables ] ) , @xmath119 where @xmath120 is a set of multidegrees @xmath121 of monomials @xmath122 in @xmath117 . rewriting the equation @xmath123 as a sum over @xmath121 , we get a homogeneous system of linear equations ( one equation for every monomial @xmath124 and fixed @xmath34 ) @xmath125 for the coefficients @xmath115 . for the weyl blocks ( [ initwb ] ) , ( [ prodwb ] ) we get equations : second order @xmath126 , @xmath127 third order @xmath128 , @xmath129 the equations have solutions spanned by vectors @xmath130
thus we derive two casimir invariants in the algebra @xmath6 under the adjoint action of the group @xmath4 : @xmath131 the weyl blocks @xmath132 and @xmath133 represent reductions of the casimir invariants onto the cartan subalgebra @xmath134 because the variables @xmath135 correspond to the generators @xmath136 .
we substitute @xmath137 by dual roots @xmath54 and get invariants of the weyl group @xmath138 in accordance with the chevalley theorem about the homomorphism of lie group invariants and weyl group invariants @xcite .
the generators @xmath140 ( see ( [ generator1 ] ) ) form the subalgebra @xmath141 in @xmath142 .
we decompose @xmath6 into a direct sum of irreducible @xmath10-modules looking at the commutation relations : @xmath143=y_2 & \qquad & [ y_1,y_2]=-y_3 & \qquad & [ x_1,y_2]=0 , \\ & [ h_1,y_3]=-y_3 & \qquad & [ y_1,y_3]=0 & \qquad & [ x_1,y_3]=-y_2.\end{aligned}\ ] ] @xmath144=x_3 & \qquad & [ y_1,x_3]=x_2 & \qquad & [ x_1,x_3]=0 , \\ & [ h_1,x_2]=-x_2 & \qquad & [ y_1,x_2]=0 & \qquad & [ x_1,x_2]=x_3.\end{aligned}\ ] ] it is convenient to introduce a new operator @xmath145 , @xmath146 with commutators : @xmath147=[y_1,h_0]=[x_1,h_0]=0.\ ] ] according to the above relations , @xmath142 is a direct sum of four @xmath10-modules : @xmath148 where each module @xmath149 is a disjoint union of one - dimensional subspaces @xmath150 the decomposition can be summarized by the following table .
@xmath151 \hline\rule{-4pt}{15pt } 2 & x_1 & & & \\[2 mm ] \hline \rule{-4pt}{15pt } 1 & & y_2 & x_3 & \\[2 mm ] \hline \rule{-4pt}{15pt } 0 & h_1 & & & h_0 \\[2 mm ] \hline \rule{-4pt}{15pt } -1 & & y_3 & x_2 & \\[2 mm ] \hline \rule{-4pt}{15pt } -2 & y_1 & & & \end{array}\ ] ] embedding @xmath10 into @xmath6 allows us to use many constructions from the previous section .
let us reorder @xmath142-generators ( [ generators ] ) in accordance with the table ( [ sl3_decomposition ] ) @xmath152 the group @xmath5 is generated by three one - parameter subgroups @xmath153 ( see ( [ adgrouplineartrans ] ) ) and acts on the row @xmath154 from the right .
the cartan subalgebra of @xmath10 is one - dimensional : @xmath155 . and
the weights of the basis elements ( [ sl3modulegen ] ) are given by the first component of their @xmath6-weights ( [ sl3weights ] ) except @xmath145 with the weight @xmath156 . from other hand
, @xmath10-weights are given by the value of @xmath157 in the table([sl3_decomposition ] ) .
accordingly , one - dimensional torus @xmath158 @xmath159 acts on 8-dimensional vector @xmath160 ( [ vector ] ) from the right .
the hilbert basis of the torus invariants includes new monomials @xmath161 in addition to the invariants of @xmath162 ( [ sl3torinvs ] ) where @xmath163 must be included instead of @xmath164 .
the root system @xmath165 is represented by @xmath166 and has the weyl group @xmath167 .
the operator @xmath168 ( [ s1 ] ) has the block - diagonal form @xmath169 now the reynolds operator ( [ rey ] ) includes only summation over @xmath168 and the identity operator . following the described above procedure , we get as a result : @xmath171 \hline\rule{0pt}{15pt } 1 & \mathcal{i}_1=\underline{h_0 } \\[2 mm ] \hline\rule{0pt}{15pt } 2 & \mathcal{i}_2=h_1 ^ 2 + 4\,\underline{x_1 y_1 } \\[2 mm ] & \mathcal{i}_3=\underline{x_2 y_2 + x_3 y_3 } \\[2 mm ] \hline\rule{0pt}{15pt } 3 & \mathcal{i}_4=h_1 y_2 y_3+\underline{y_1 y_2
^ 2-x_1 y_3 ^ 2 } \\[2 mm ] & \mathcal{i}_5=h_1 x_2 x_3+\underline{x_1 x_2 ^ 2-x_3 ^ 2 y_1 } \\[2 mm ] & \mathcal{i}_6=h_1(x_2 y_2 - x_3 y_3)-2\,(\underline{y_1y_2x_3+x_1x_2y_3 } ) \end{array}\ ] ] where initial weyl blocks are underlined .
we shall prove that the set of invariants @xmath172 generates the invariant ring @xmath173^{{\mathrm } { sl}(2)}$ ] , i.e. this set is a fundamental system of @xmath5-invariants in @xmath6 algebra .
this is the case if and only if the hilbert series @xmath174 of the subalgebra @xmath175 $ ] is equal to the molien series @xcite of @xmath173^{{\mathrm } { sl}(2)}$ ] . for this purpose
we calculate the molien function @xcite for the adjoint action of @xmath176 ( i.e. the real compact form of @xmath5 ) on the algebra @xmath142 .
an injective map of @xmath177 into @xmath178 @xmath179 corresponds to our choice of representation of @xmath10 in @xmath6 ( see the previous subsection [ sl2_in_sl3 ] ) .
one can instead of the adjoint action @xmath180 consider a linear representation @xmath181 of @xmath182 on @xmath183 @xmath184 where @xmath185 is the complex conjugation of @xmath186 .
after that we calculate the molien function @xmath187^{\mathrm{su}(2)},q)=\int_{\mathrm{su}(2)}\ , \frac{d\mu}{\det(\mathbb{i}-q\,\pi(\rho)\otimes\overline{\pi(\rho ) } ) } , \qquad @xmath177 , @xmath188 is an identity operator , and @xmath186 is a representation of @xmath182 . using weyl integration formula @xcite
, one can reduce the integral over @xmath177 to the integral over its maximal torus @xmath189 .
an expression of @xmath190 via the haar measure @xmath191 on @xmath189 is the following one : @xmath192 @xmath186 has a diagonal form @xmath193 and , hence , @xmath194 @xmath195 substituting ( [ haar ] ) , ( [ det ] ) in the expression ( [ eq : molienfunct ] ) we derive that the molien function is equal to the following contour integral over the unit circle : @xmath196^{\mathrm{su}(2)},q)= -\frac{1}{4\pi i}\frac{1}{(1-q)^3 } \oint \frac{zdz(1-z^2)^2}{(1-qz)^2(1-qz^2)(z - q)^2(z^2-q)}.\ ] ] there are two simple poles @xmath197 and a double pole @xmath198 ( of order @xmath199 ) .
application of the residue theorem @xcite yields the result : @xmath200^{\mathrm{su}(2)},q)=\frac{1+q^3}{(1-q)^2(1-q^2)^2(1-q^3)^2}.\ ] ] @xmath201^{\mathrm{su}(2)},q)$ ] satisfies a functional equation @xmath196^{\mathrm{su}(2)},q^{-1})=q^{9 } m(\mathbb{c}[v]^{\mathrm{su}(2)},q ) \,,\ ] ] hence , @xmath44^{\mathrm{su}(2)}$ ] being a graded cohen - macaulay algebra @xcite is the gorenstein one according to stanley @xcite .
required molien series of @xmath177-invariant ring @xmath44^{\mathrm{su}(2)}$ ] is nothing then an expansion of @xmath201^{\mathrm{su}(2)},q)$ ] in series of powers of @xmath202 .
now we turn out to the calculation of the hilbert series @xmath174 for the subalgebra @xmath175 $ ] . using slack - variables method @xcite
, we find one syzygy between the invariants @xmath203 ( [ sl2invariants ] ) , @xmath204 thus @xmath205 are algebraically independent while @xmath206 can be considered as algebraically dependent .
this implies that @xmath207 is decomposed as the direct sum of graded @xmath208-vector space @xmath209= \mathbb{c}[\mathcal{i}_1,{\ldots } , \mathcal{i}_5]\oplus \mathcal{i}_6\,\mathbb{c}[\mathcal{i}_1,{\ldots } , \mathcal{i}_5]\ ] ] with the hilbert series @xmath210 we see that the denominator of the molien function ( [ molien_result ] ) has an extra term @xmath211 in comparison with @xmath174 .
the reason is that we have calculated @xmath201^{\mathrm{su}(2)},q)$ ] in 9-dimensional vector space but @xmath6 has dimension 8 .
this completes the proof . as a corollary
we get that a hironaka decomposition @xcite of @xmath207 is done by the formula ( [ hironaka_decomposition ] ) .
@xmath212 are primary invariants and @xmath206 is a secondary invariant .
for example , a decomposition of the casimir invariants ( [ sl3cas2 ] ) is the following : @xmath213
for construction of lie group invariants we propose a method based on the weyl block structure of invariants . to produce @xmath7-blocks , we extend the weyl group action on the cartan subalgebra to the automorphism of the whole lie algebra .
corresponding set of operators defines a reynolds operator @xmath99 .
being applied to hilbert basis of torus invariants operator @xmath99 produces initial @xmath7-blocks . beside
that we construct @xmath7-blocks from all possible products of torus invariants up to the higher degree of invariants from the hilbert basis .
a homogeneous linear combination of @xmath7-blocks is invariant relative to the adjoint action of a lie group if a difference of the combination under a transformation is zero .
this condition gives a system of linear equations .
if there is a solution of the system , then we get an invariant . using the described above method
, we construct fundamental sets of invariants of the adjoint actions of the lie groups @xmath4 and @xmath5 on the lie algebra @xmath6 .
we see that the weyl block structure of invariants reflects very deep properties of lie algebras and lie groups such as generating relations , properties of root systems , irreducible representations and so on .
all this information can be useful for studying rings of invariants .
for example , let us mention that the initial @xmath7-blocks of the invariants ( [ sl2invariants ] ) taken with their coefficients satisfy the same syzygy ( [ syzygy ] ) as invariants themselves .
the author thanks vladimir gerdt and arsen khvedelidze for helpful discussions .
the work was supported in part by the rfbr ( grant no .
10 - 01 - 00200 ) and by the ministry of education and science of the russian federation ( grant no .
3802.2012.2 ) . | for an adjoint action of a lie group @xmath0 ( or its subgroup ) on lie algebra @xmath1 we suggest a method for construction of invariants . the method is easy in implementation and may shed the light on algebraical independence of invariants .
the main idea is to extent automorphisms of the cartan subalgebra to automorphisms of the whole lie algebra @xmath1 .
corresponding matrices in a linear space @xmath2 define a reynolds operator `` gathering '' invariants of torus @xmath3 into special polynomials .
a condition for a linear combination of polynomials to be @xmath0-invariant is equivalent to the existence of a solution for a certain system of linear equations on the coefficients in the combination . as an example we consider the adjoint action of the lie group @xmath4 ( and its subgroup @xmath5 ) on the lie algebra @xmath6 . |
random numbers are crucial for various tasks , among them generating cryptographic secret keys , authentication , monte - carlo simulations , digital signatures , statistical sampling , etc .
random number generators can be classified into two types : pseudo - random number generators ( prng ) @xcite and true random number generators ( trng ) .
a prng is an algorithm , computational or physical , for generating a sequence of numbers that approximates the properties of random numbers .
a physical or hardware version is typically based on stochastic noise or chaotic dynamics in a suitable physical system @xcite .
computational prngs are based on computational algorithms that generate sequences of numbers of very long periodicity , making them look like true random numbers for sufficiently short sequences .
careful observation over long periods will in principle reveal some kind of pattern or correlation , suggestive of non - randomness . as far as is known today , the inherent indeterminism or fluctuations in quantum phenomena is the only source of true randomness , an essential ingredient in quantum cryptography @xcite .
various proposed underlying physical processes for quantum random number generators ( qrngs ) include : quantum measurement of single photons @xcite , an entangled system @xcite , coherent states @xcite or vacuum states @xcite ; phase noise @xcite , spin noise @xcite , or radioactive decay or photonic emission @xcite .
in this work , we propose a novel method of qrng that is a quite different indeterministic paradigm from the above two .
it uses _ bosonic stimulation _ to randomly amplify weak coherent pulses to intense pulses that can be easily detected by a conventional apds .
bosons ( integer - spin quantum particles ) obey bose - einstein statistics , which entails that the transition probability of a boson into a given final state in enhanced by the presence of identical particles in that state .
if there are @xmath0 particles in a given quantum state , the probability that an incoming boson makes a transition into that state is proportional to @xmath1 .
this effect is called bosonic stimulation , and is responsible for coherent matter wave amplification in atomic lasers , as well as the sustenance of a particular mode in a laser cavity , whereby the presence of photons in a particular lasing mode stimulates the emission of more photons into that mode .
it provides a new way to combine quantum indeterminism with the tendency of bosons to congregate indistinguishably .
we may call this qrng a _ random bosonic stimulator_. perhaps the practical merit of our proposed qrng , apart from its harnessing a novel version of quantum indeterminism , is that it simplifies the detection module to the point where it may be accessible to an advanced undergraduate laboratory .
consider an urn with @xmath2 blue balls and @xmath3 red balls .
a ball is picked at random and replaced with @xmath4 balls of the same color or @xmath5 balls of different color .
the addition of same color balls results in positive feedback whereas that of different color balls results in negative feedback . during a run of trials , the fractional population of each urn
may initially fluctuate randomly , but eventually settles down to a randomly selected _ limiting value _
@xmath6 , giving an instance of symmetry breaking ( figure [ fig : graph ] ) .
let the initial population of two states labelled as blue " and red " be @xmath7 and @xmath8 , respectively . as the incoming balls start populating the two states , the subsequent growth in population of the modes exhibits plya urn behaviour .
the probabilistic law of evolution of the fractional population at @xmath9th instance is given by bosonic stimulation to be @xmath10 .,width=453 ] the limiting value @xmath11 for large @xmath9 itself varies randomly in the range @xmath12 $ ] from run to run , having a beta distribution @xmath13 where @xmath14 is the beta function that normalizes @xmath15 , @xmath16 , @xmath17 , @xmath18 , @xmath19 . for bosonic stimulation , the shifts @xmath20 .
if the seeding is symmetric , then so is the asymptotic distribution of the limiting values , thereby restoring symmetry , as it should be @xcite .
depending on the number of trial runs , the final state can have an arbitrarily large number of bosons . depending on whether @xmath21 ( blue dominates ) or @xmath22 (
red dominates ) , one generates a random bit 0 or 1 .
this can serve as the basis of generating random bits at a rate determined by the frequency with which each run can be repeated .
thus , the phenomenon of bosonic stimulation acts as a macroscopic qrng .
a concrete idea for realizing a random bosonic stimulator is to use a lasing medium that supports two radiation modes , for example by vertical and horizontal polarization of the same frequency @xcite .
a scheme of the proposed experiment is given in figure [ fig : rng ] .
two equal intensity , highly attenuated modes of coherent states are input into a lasing medium . to ensure that the two inputs are synchronized and of equal intensity , a calibrated mach - zehnder set - up is used with an attentuated coherent laser pulse fed into one of its input ports .
this results in an ouput consisting of two ( unentangled ) coherent pulses with half the intensity .
a quarter wave plate in one of the arms ensures that the polarization in one arm rotated to be 90@xmath23 with respect to the other .
each mode in a pulse corresponds to a ball color in the plya urn problem .
because of bosonic stimulation , the output intensity will randomly favor vertical or horizontal polarization .
let @xmath24 ( @xmath25 ) denote the intensity of the outcoming light in the horizontal ( vertical ) polarization mode .
with horizontal polarization and average photon number @xmath26 enters the experiment .
the upper arm branch is rotated to vertical polarization by the quarter - wave plate qwp .
the product state @xmath27 enters the lasing medium , where the @xmath28 and @xmath29 modes participate in bosonic stimulation . in the symmetric case ,
the detector complex generates a 0 or 1 bit depending on whether @xmath30 or the converse , where @xmath31 are intensities of the horizontal and vertical polarization output components.,width=377 ] the incoming photon emitted as a result of the coherent de - excitation of the atoms in the medium is assumed to be equally coupled to both modes .
suppose the two modes start in the state @xmath32 , where the first register corresponds to the ` blue ' mode and the second to the ` red ' mode .
further let the series of atoms in the population - inverted state be in the initial excited state @xmath33 . by giving up a photon into the blue or red mode ( which could be angular momentum states ) ,
the atom is left in the state @xmath34 or @xmath35 , assumed to be mutually orthogonal .
the joint system of the modes and atoms evolves in a manner analogous to a quantum walk , given by : @xmath36 and so on .
each term in the superposition is rendered incoherent because it is entangled with a distinct state of atoms , and thus the probability for scattering into a given urn state is quantitatively the same as the classical plya urn situation . because of the entanglement with the polarization degrees of freedom of the atoms ( eq .
( [ eq : nieuw ] ) ) , the state of the atoms bears an imprint of the final outcome of laser light .
however , as the laser atoms are not individually accessible , the random bit generated is practically unique .
from the view point of monte - carlo simulations , etc .
, only the randomness from the laser light read - out will be used . from a cryptographic perspective
, the laser system will remain physically well within the encoder unit , preventing its access to a malevolent eavesdropper .
assuming that any possible information leakage through side - channels ( like heat radiations from the laser ) are reasonably blocked out to the outside world , the assumption of practical uniqueness of the generated randomness applies here , too .
a random bit @xmath37 is generated by the detector module , depending on which mode dominates , for example , according to the recipe : @xmath38 the mean and variance of the distribution @xmath39 are given , respectively , by @xmath40 the value @xmath6 obtained will tend to peak towards the mean , with ever lower variance if one or both of the initial populations are large . if the instrument function of the detector is denoted by a normal distribution @xmath41 with fwhm @xmath42 ,
the observed distribution is the convolution @xmath43 .
it is important that we use sufficiently weak pulses obtained by attenuating coherent light sources , so that @xmath44 .
this ensures that the quantum randomness dominates over stochastic noise in the detector s reading . the distribution @xmath45 is uniform over @xmath12 $ ] , and thus the performance of the random bosonic stimulator is least affected by detector tolerance
this means that the two inputs should ideally be vacuum modes .
a challenge will be to ensure that the beam splitters are truly 50 - 50 and the coupling of the excited atoms is equal to the two modes .
however , even if this is not so , eq . ( [ eq : beta ] ) and the prescription ( [ eq : rbit ] ) can be suitably generalized to yield a random bit of uniform distribution .
suppose a photon couples to the @xmath29 mode stronger than to the @xmath28 mode ( because of an atomic or beam - splitter feature ) by a factor @xmath46 , then the new distribution can be shown to be eq .
( [ eq : beta ] ) , but with @xmath47 .
let @xmath48 be the median of the distribution @xmath15 , defined such that @xmath49 , where @xmath15 is the beta distribution ( [ eq : beta ] ) .
then , we obtain our uniformly random bit by replacing prescription ( [ eq : rbit ] ) by : @xmath50 to be precise , the above numbers assume that the input modes are pure number states .
more realistcally , taking into account that they are coherent states , we must replace eq .
( [ eq : beta ] ) by @xmath51 where @xmath52 is the poisson distribution of @xmath37 with mean @xmath53 .
furthermore , in practice we may have to let @xmath53 to range over an interval because of practical difficulties of producing the same exact degree of attentuation on each run .
it can be shown that this added complication does not affect our main results . for sufficiently low noise in each run , two bits
may also be generated per run according to the recipe : @xmath54 where @xmath55 is defined such that @xmath56 .
more generally , to generate @xmath57 bits , the noise level should be lower than @xmath58 .
we have proposed a novel qrng principle , based on bosonic stimulation , in which , while the state preparation procedure presents experimental challenges , the detection and read - out parts are easier to implement . in an actual experiment , it is possible that systematic experimental biases might introduce correlations into the sequence of bits produced , thereby degrading the randomness .
statistical analyses like the diehard tests and national institute of standards and testing ( nist ) suite of tests for randomness have to be carried out to know the quality of randomness and improve upon it .
some implementational details are worth noting .
since the required light is not for communication to a geographically distant station , any laser system that can be conveniently tuned to operator at 800 nm , where good detectors ( apds ) are available , is suitable for our purpose .
a good candidate is femtosecond titanium - sapphire lasers , for which pulse repetition rates upto a few gigahertz can be obtained .
further , the potential problem posed by the cavity s unequal coupling to the two modes can taken care of by `` biasing '' the comparator ( as clarified above ) .
care must be taken to choosing the operational / differential amplifier ( op amp ) ic designed to work as comparator , so that their frequency specifications are appropriate for the pulse rate of the laser .
even with the low resolution provided by two - bit comparators , for a titanium - sapphire laser , mentioned above , this yields a random bit - rate of the order of @xmath59 bits per second , which is comparable to bit - rates in state - of - the - art qrngs .
the present method of generating randomness is based on the essentially quantum feature of being able to turn a distinct bosonic particle into an indistinguishable part of a collective object , which is the mode or a condensate .
the notion of identity in classical logic and philosophy is as such incapable of handling this situation , requiring recourse to concepts like quasi - set theory @xcite .
we think that , as in the case of the application of quantum nonlocality to cryptography @xcite , our method of generating randomness based on bosonic indistinguishability , can perhaps be termed as another instance of _ applied philosophy _ !
here we have presumed the axiomatic indeterminism of quantum mechanics .
this assumption would be falsified if a ( nonlocal ) hidden variable theory were able to explain quantum mechanics in the future .
as this seems unlikely , a qrng still seems the best bet for a source of genuine and unique randomness .
miesner , h. j. , stamper - kurn , d. m. , andrews , m. r. , durfee , d. s. , inouye , s. and ketterle , w. science , 279 ( 1998 ) , 1005 .
the nalanda dialogs on _ individuation and identity _ : http://nalanda-dialogforum.org a. k. ekert .
67 , 661 ( 1991 ) . | we propose a method to realize a robust quantum random number generator based on bosonic stimulation . a particular implementation that employs weak coherent pulses and conventional avalanche photo - diode detectors ( apds )
is discussed . |
this work arose from the following question : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ given a convex polytope @xmath3 in @xmath0 and a point @xmath4 inside @xmath3 , does there exist a projective transformation @xmath5 such that @xmath6 is the centroid of the vertices of @xmath7 ? _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ for example , if @xmath3 is a @xmath2-simplex , then as @xmath5 one can take the projective transformation that fixes the vertices of @xmath3 and maps @xmath4 to the centroid of @xmath3 .
the question can be generalized as follows : [ quest:1 ] given two subsets @xmath8 , does there exist a projective transformation @xmath9 such that the centroids of @xmath10 and @xmath11 coincide ?
the centroid @xmath12 can be defined ( by the usual integral formula ) for any subset @xmath13 that has a positive finite @xmath14-hausdorff measure for some @xmath15 , see e. g. @xcite .
thus , any of the sets @xmath16 in question [ quest:1 ] may be , say , a convex polytope or the @xmath14-skeleton of a convex polytope .
if @xmath17 or @xmath18 , then the centroid of @xmath19 is affinely covariant : @xmath20
so that we can quotient out affine transformations when searching for @xmath5 . since @xmath21 , the `` number of equations '' becomes equal to the `` number of variables '' , and the following question poses itself .
[ quest:2 ] if @xmath22 in question [ quest:1 ] , then is @xmath23 such that @xmath24 unique up to post - composition with an affine transformation ?
for example , if @xmath25 is the vertex set of a simplex , and @xmath26 is a single point , then @xmath5 is unique in the above sense . for @xmath27
the centroid is in general not affinely covariant .
indeed , it is well - known that the centroid of the boundary of a triangle coincides with the centroid of its vertices if and only if the triangle is regular . as the centroid of the vertices is affinely covariant ,
the centroid of the boundary is not .
see @xcite for centroids of skeleta of simplices in higher dimensions .
let us discuss some restrictions we will impose on @xmath16 and @xmath5 .
first , @xmath19 is always assumed to be compact and equal to the closure of its interior .
the latter is not really a restriction , since replacing @xmath19 by the closure of @xmath28 does nt change the centroid .
second , we will always assume one set to be contained in the convex hull of the other : @xmath29 .
although this looks quite restrictive , it leaves enough room for non - trivial results .
for an idea of what can be done in the case when the convex hulls are incomparable , see section [ sec : futres ] .
third , in order for the centroid of @xmath30 to be defined , no point of @xmath19 may be sent to infinity for @xmath17 and @xmath31 must hold for @xmath18 .
the following restriction ( together with compactness of @xmath19 ) guarantees both .
[ dfn : admissible ] let @xmath32 . a projective transformation
@xmath9 is called _ admissible _ for @xmath19 , if @xmath33 .
non - admissible projective transformations are more difficult to handle ; besides , admissible transformations will often suffice .
if we allow a projective transformation to send to infinity a hyperplane that separates the points @xmath34 , then we can lose the uniqueness , see proposition [ prp : nonadm ] . the requirement that @xmath19 is equal to the closure of its interior forbids @xmath19 to have `` antennas '' .
it turns out , we heed to forbid `` horns '' in order to ensure the existence of a suitable projective transformation .
a @xmath2-dimensional compact subset @xmath13 is called _ cusp - free _ , if for every @xmath35 there is a @xmath2-simplex contained in @xmath19 with a vertex at @xmath36 .
examples of cusp - free sets are : pure @xmath2-dimensional polyhedra ( finite unions of convex @xmath2-dimensional polyhedra ) ; @xmath2-submanifolds of @xmath0 with smooth boundary ; @xmath2-submanifolds with corners . here
we present our results in the case when @xmath37 is a single point .
[ thm : onevsmany ] let @xmath38 be a finite set of points affinely spanning @xmath0 , and let @xmath39 be a point in the interior of their convex hull .
then there exists a projective transformation @xmath9 , admissible with respect to @xmath19 , such that @xmath40 if @xmath41 is any other admissible projective transformation with @xmath42 , then @xmath43 is an affine transformation .
a projective transformation modulo post - composition with affine ones is uniquely determined by the hyperplane @xmath44 that it sends to infinity .
associate to every @xmath45 the point @xmath4 that becomes the centroid of @xmath46 after @xmath45 is sent to infinity .
then , by theorem [ thm : onevsmany ] , the hyperplanes disjoint from @xmath47 are in one - to - one correspondence with the points inside the convex hull .
this correspondence is related to the polarity with respect to an algebraic set .
namely , let @xmath48 be the union of the hyperplanes dual to @xmath46 ; then the dual of @xmath4 is the polar with respect to @xmath48 of the dual of @xmath45 .
see proposition [ prp : polarity ] for more details .
on the other hand , there is a relation to the analytic center of a polytope and the karmarkar s algorithm , @xcite .
[ thm : onevsbody ] let @xmath13 be a compact cusp - free @xmath2-dimensional set , and @xmath39 . then there exists a projective transformation @xmath9 , admissible with respect to @xmath19 , such that @xmath49 for any other admissible projective transformation @xmath41 with this property , the composition @xmath43 is affine .
since projective transformations can be represented by central projections ( section [ sec : conesec ] ) , theorem [ thm : onevsbody ] can be reformulated as existence and uniqueness of a hyperplane section of a cone through a given point having this point as the centroid . representing projective transformations as composition of two polarities , we can relate theorem [ thm : onevsbody ] in the case of convex @xmath19 to the santal point : the hyperplane that must be sent to infinity is dual to the santalo point of the dual of @xmath19 .
see theorems [ thm : onevsbodyb ] and [ thm : onevsbodyc ] in section [ sec : conesecsant ] . if the point @xmath4 lies sufficiently close to a sufficiently sharp cusp of @xmath19
, then there is no projective transformation making @xmath4 to the centroid of @xmath19 .
see example [ exl : counterpointbody ] .
here we present the results in the case when one of the sets is finite but consists of more than one point .
[ thm : manyvsmany ] let @xmath50 and @xmath51 be such that @xmath52 . then there exists a projective transformation @xmath5 , admissible with respect to @xmath25 ( and hence with respect to @xmath26 ) , such that @xmath24 . in general , @xmath5 is not unique , even up to post - composition with affine transformations . for non - uniqueness ,
see examples [ exl : countermanymany1 ] and [ exl : countermanymany2 ] .
[ thm : manyvsbody ] let @xmath13 be a compact cusp - free @xmath2-dimensional set , and let @xmath53 be such that every support hyperplane of @xmath19 contains less than @xmath54 of the points @xmath34 .
then there exists a projective transformation @xmath5 such that @xmath55 in general , @xmath5 is not unique , even modulo affine transformations . for the sharpness of the assumptions and for non - uniqueness , see examples [ exl : d+1points ] and [ exl : counterbodymany ] .
interestingly enough , the assumptions leading to existence become obsolete , and the transformation turns out to be unique , if @xmath19 is a ball . since the image of a ball under an admissible projective transformation is an ellipsoid , and the ellipsoid can be mapped back to the ball by an affine transformation , the following theorem is equivalent to the existence and uniqueness of a projective transformation fitting the centroids of a ball and of a finite set .
[ thm : manyvsball ] let @xmath56 be the unit ball centered at the origin , and let @xmath57 be a finite set of points , @xmath58 .
then there exists a projective transformation fixing @xmath59 such that @xmath60 the transformation @xmath5 is unique up to post - composition with an orthogonal transformation .
this result generalizes centering via mbius transformations @xcite used for unique representation of polyhedral types .
there is also a relation to the conformal barycenter @xcite , see remark [ rem : moebius ] . here
we present the results for the case when both @xmath25 and @xmath26 are @xmath2-dimensional . in order to get some uniqueness results , we need to assume that @xmath25 and @xmath26 are convex .
the uniqueness can be guaranteed if one of the bodies lies `` deep inside '' the other .
let @xmath61 be two convex bodies in @xmath0 .
the _ hilbert diameter _ of @xmath26 with respect to @xmath25 is defined as @xmath62 where @xmath63 are points collinear with @xmath64 and @xmath4 , and @xmath65 .
the _ maximum hilbert width _ of @xmath26 with respect to @xmath25 is defined as @xmath66 where @xmath67 is a @xmath68-dimensional affine subspace , and @xmath69 are support hyperplanes to @xmath16 .
it follows immediately from definition that @xmath70 where @xmath71 denotes the polar dual of @xmath19 ( one may take the polar duals with respect to any point lying in the interior of both @xmath25 and @xmath26 )
. see figure [ fig : tang ] , where @xmath72 .
( 0,0 ) and @xmath26 and the dual collinear points on the boundaries of @xmath73 and @xmath74.,title="fig : " ] # 1#2#3#4#5 @font ( 5569,2145)(407,-1110 ) ( 1129,325)(0,0)[lb ] ( 1117,-582)(0,0)[lb ] ( 1221,737)(0,0)[lb ] ( 1326,-1050)(0,0)[lb ] ( 1478,-144)(0,0)[lb ] ( 717,-134)(0,0)[lb ] ( 5310,-185)(0,0)[lb ] ( 5387,-745)(0,0)[lb ] ( 3331,-241)(0,0)[lb ] ( 4501,-916)(0,0)[lb ] ( 4816,-601)(0,0)[lb ] ( 5401,839)(0,0)[lb ] ( 5041,254)(0,0)[lb ] note that for @xmath75 the number @xmath76 is the hyperbolic distance between @xmath77 and @xmath78 , with @xmath59 viewed as the cayley - klein model of the hyperbolic space .
thus , the maximum hyperbolic width is defined as the maximum distance between support hyperplanes .
[ thm : bodyvsbody ] let @xmath8 be two convex bodies such that @xmath52 .
then there exists a projective transformation @xmath5 such that @xmath79 in general , @xmath5 is not unique modulo affine transformations .
it is unique , if @xmath80 the bound is sharp .
it is achieved for a cross - polytope @xmath25 and a rectangular parallelepiped @xmath26 inside it , provided that at least one of the sides of the parallelepiped is sufficiently long , see fig .
[ fig : sharpbound ] .
( 0,0 ) # 1#2#3#4#5 @font ( 2544,1374)(664,-523 ) ( 2566 , 74)(0,0)[lb ] ( 3151 , 74)(0,0)[lb ] the discussion in section [ sec : comppolar ] justifies the following definition .
a point @xmath81 is called a _ santal point of a pair _ @xmath82 of convex bodies , if @xmath83 and @xmath84 have the same centroid . here
@xmath85 denotes the polar dual of @xmath86 with respect to the unit sphere centered at @xmath87 . by going to the polar duals of @xmath25 and @xmath26 ,
one derives from theorem [ thm : bodyvsbody ] criteria for existence and uniqueness of a santal point of a pair .
let @xmath82 be two convex bodies such that @xmath88 .
then the pair @xmath89 has at least one santalpoint . if the hilbert diameter of @xmath74 with respect to @xmath73 satisfies @xmath90 then the santal point of the pair @xmath89 is unique .
[ thm : bodyvsball ] let @xmath91 be a convex body such that @xmath92 in the hyperbolic metric defined by @xmath59 as a cayley - klein model . then there is a unique projective transformation @xmath5 , up to post - composition with orthogonal ones , that fixes @xmath59 and such that the centroid of @xmath30 is the center of @xmath59 .
the bound in is sharp .
[ thm : ballvsbody ] let @xmath91 be a convex body such that @xmath93 in the hyperbolic metric defined by @xmath59 as a cayley - klein model .
then there exists a unique point @xmath94 such that the centroids of the polars of @xmath19 and @xmath59 with respect to @xmath87 coincide .
the bound in is sharp .
let @xmath59 be the unit ball , and @xmath19 be contained in a concentric ball of radius @xmath95 .
then there is a unique projective transformation @xmath5 , up to post - composition with orthogonal ones , that fixes @xmath59 and such that the centroid of @xmath30 is the center of @xmath59 .
let @xmath59 be the unit ball , and @xmath19 be contained in a concentric ball of radius @xmath96 .
then there is a unique santal point of the pair @xmath97 . in section [
sec : reform ] we discuss left cosets of the affine group in the projective group and represent them by elations , central projections and compositions of polarities .
section [ sec : finsets ] deals with the case of @xmath98 , that is with finite point sets . theorems [ thm : onevsmany ] and [ thm : manyvsmany ] are proved here .
the solution is found as a critical point of a concave functional , respectively of the difference of two such functionals . in section [ sec : pointbody ] , theorem [ thm : onevsbody ] is proved and interpreted in the contexts of minimizing the volume of a cone section and of the santal point .
here the convex functional associated with a convex body is introduced .
section [ sec : manyvsbody ] deals with the case @xmath99 , @xmath100 and proves theorems [ thm : manyvsbody ] and [ thm : manyvsball ] .
section [ sec : twobodies ] deals with the case of two @xmath2-dimensional sets and proves theorems [ thm : bodyvsbody ] . finally , in section [ sec : futres ] poses some questions for future research .
the author wishes to thank arnau padrol , raman sanyal , boris springborn , and gnter ziegler for useful discussions .
affine transformations ( or affinities ) of @xmath0 are maps of the form @xmath101 , where @xmath102 .
identify @xmath0 with a subset of the projective space : @xmath103 by associating @xmath104 with the equivalence class of @xmath105 .
then projective transformations of @xmath106 restricted to @xmath0 have the form @xmath107 in particular , the group @xmath108 of affinities is a subgroup of the group @xmath109 of projectivities . [ prp : a ] every right coset of @xmath108 in @xmath109 has a unique representative of the form @xmath110 or two representatives ( @xmath87 doing the same as @xmath111 ) of the form @xmath112 two projectivities belong to the same right coset of @xmath108 if and only if they send to infinity the same hyperplane .
any hyperplane that does not pass through the origin has equation @xmath113 for a unique @xmath87 , and is therefore sent to infinity by a map of the form .
in particular , for @xmath114 the hyperplane at infinity is sent to itself .
any hyperplane through the origin is sent to infinity by a map of the form .
projective transformation is an elation with the axis @xmath115 and center @xmath1 .
projective transformation is not an elation ( not even a homology ) , but can be replaced by the map @xmath116 which , for @xmath117 , is an elation with the axis @xmath118 and center @xmath111 .
we may always assume @xmath119 for @xmath120 .
then none of the maps is admissible in the sense of definition [ dfn : admissible ] , and the map is admissible if and only if @xmath121 , where @xmath122 this allows us to reformulate questions [ quest:1 ] and [ quest:2 ] as follows . [ ref : a ] for a set @xmath13 containing the origin in the interior of the convex hull , when does there exist @xmath81 such that @xmath123 ? for two sets
@xmath8 containing the origin in their convex hulls , when does there exist @xmath121 such that @xmath124 ? under what assumptions is @xmath87 unique ? for every set @xmath125 define the conical hull over @xmath126 as @xmath127 for @xmath104 denote @xmath128 and for @xmath13 denote @xmath129 .
this associates with @xmath13 a set @xmath130 in @xmath131 .
hyperplane sections of @xmath132 are central projections of @xmath133 , and thus images of @xmath19 under projective transformations .
this leads to the following reformulation of questions [ quest:1 ] and [ quest:2 ] .
[ ref : b ] for a cone @xmath134 and a point @xmath4 different from the apex of @xmath132 , when does there exist a hyperplane @xmath135 through @xmath4 such that @xmath4 is the centroid of @xmath136 ? for two cones @xmath137 with common apex , when does there exist a hyperplane @xmath135 not passing through the apex such that the centroids of @xmath138 and @xmath139 coincide ? under what assumptions is @xmath135 unique ( in the second case , up to parallel translation ) ? to show that central projections modulo dilations correspond to projectivities modulo affinities ,
let us relate central projections with transformations @xmath140 from .
for every vector @xmath81 denote @xmath141 denote by @xmath142 the central projection and by @xmath143 the parallel projection along @xmath144 .
[ prp : b ] the map is a composition of a central and a parallel projection : @xmath145 from @xmath146 and @xmath147 it follows that @xmath148 and since @xmath149 , we have @xmath150 . note
that the hyperplane contains @xmath144 , and that every non - vertical hyperplane through @xmath144 is @xmath151 for some @xmath81 .
this establishes a bijection between the images @xmath152 and sections @xmath153 and shows that reformulation [ ref : b ] is equivalent to reformulation [ ref : a ] . similarly
to , define the polar dual of @xmath86 with respect to a point @xmath87 : @xmath154 here is a reformulation of questions [ quest:1 ] and [ quest:2 ] in the case when @xmath25 and @xmath26 are both convex bodies .
[ ref : c ] for a convex body @xmath155 , when does there exist a point @xmath156 such that the polar dual of @xmath86 with respect to @xmath87 has centroid at @xmath87 ? for two convex bodies @xmath82 , when does there exist a point @xmath157 such that the centroids of the polar duals of @xmath73 and @xmath74 with respect to @xmath87 coincide ? under what assumptions is @xmath87 unique ? again , we justify this by relating polarity with variable center to the map @xmath140 from .
[ prp : c ] for every @xmath2-dimensional convex body @xmath13 and every point @xmath158 we have @xmath159 indeed , for any @xmath158 we have @xmath160 hence @xmath161 for every @xmath13 . if @xmath19 is convex , compact , and @xmath162 , then @xmath163 , and the proposition follows .
the property @xmath164 is characteristic for the santal point of @xmath86 , ( * ? ? ?
* remark 10.8 ) .
thus , existence and uniqueness of the santal point for convex bodies implies a positive answer to questions [ quest:1 ] and [ quest:2 ] in the case of a convex body and a point . in the case of two convex bodies
@xmath73 and @xmath74 in reformulation [ ref : c ] the point @xmath87 can be called the santal point of a pair of convex bodies .
here we prove theorem [ thm : onevsmany ] using reformulation [ ref : a ] from section [ sec : hypinfty ] . without loss of generality
we may assume @xmath165 .
since @xmath166 for all @xmath87 , it follows from proposition [ prp : a ] that theorem [ thm : onevsmany ] is equivalent to the following .
[ thm : onevsmanya ] let @xmath167 be such that @xmath168 , where @xmath169 .
then there exists a unique @xmath170 such that @xmath171 the proof is based on the fact that the left hand side of is the gradient of a strictly concave function .
define @xmath172 [ lem : grad1 ] we have @xmath173 indeed , for every @xmath104 and every @xmath174 we have @xmath175 [ lem : conv1 ] the function @xmath176 is strictly concave .
basically , this follows from the strict concavity of @xmath177 on @xmath178 , as @xmath176 is a sum of logarithms of affine functions whose linear parts span @xmath179 . computing the second derivative of @xmath176 yields @xmath180 besides , @xmath181 if and only if @xmath182 for all @xmath174 .
as @xmath183 are affinely spanning @xmath0 , they are also linearly spanning it , so that all scalar products vanish only if @xmath184 .
[ lem : coerc1 ] the value @xmath185 tends to @xmath186 as @xmath87 tends to @xmath187 .
as @xmath188 , some of @xmath189 tend to @xmath1 , and their logarithms tend to @xmath186 . on the other hand , since @xmath190 is bounded , all summands in are bounded from above .
thus the whole sum tends to @xmath186 as @xmath188 . by lemma [ lem : grad1 ] ,
a point @xmath170 satisfies if and only if @xmath87 is a critical point of the function @xmath176 from . by lemma [ lem : coerc1 ]
, @xmath176 attains a maximum on @xmath191 .
the point of maximum is a critical point , and this proves the existence part of the theorem . to prove the uniqueness , use the strict concavity of @xmath176 , lemma [ lem : conv1 ] .
it implies that all critical points of @xmath176 are strict local maxima .
since @xmath191 is convex , there can not be more than one strict local maximum , and the theorem is proved .
here we repeat the argument from the previous section in the spirit of section [ sec : conesec ] .
this serves as a preparation to some of the arguments that will follow .
let @xmath192 be open rays issued from the origin , linearly spanning @xmath131 , and contained in an open half - space whose boundary goes through the origin .
denote @xmath193 .
we want to show that for every open ray @xmath194 issued from the origin and contained in the interior of @xmath132 there is a hyperplane @xmath135 that intersects all rays and whose intersection point with @xmath194 is the centroid of the intersection points with @xmath195 : @xmath196 for this we choose arbitrary points @xmath197 , @xmath198 and introduce the function @xmath199 defined in the interior of the cone @xmath200 ( since all @xmath183 lie in an open half - space , @xmath201 ) .
we compute the gradient @xmath202 and find that @xmath203 if and only if the hyperplane @xmath204 satisfies the condition . on the other hand , @xmath205 as we will see in a minute , the function @xmath176 is neither convex nor concave , which complicates the search for a critical point .
the remedy is to restrict @xmath176 to the hyperplane @xmath206 .
[ lem : critptproj ] critical points of @xmath176 restricted to @xmath206 correspond to hyperplanes @xmath135 through @xmath4 that satisfy the condition .
indeed , @xmath87 is a critical point of the restriction if and only if @xmath207 is orthogonal to @xmath206 , that is collinear with @xmath4 .
[ lem : concproj ] the restriction of the function @xmath176 to @xmath208 is strictly concave . a vector @xmath209 is tangent to the hyperplane @xmath206 if and only if @xmath210 .
since @xmath46 linearly span @xmath131 , this implies @xmath211 for @xmath212 .
lemmas [ lem : critptproj ] and [ lem : concproj ] imply the existence and uniqueness of a hyperplane @xmath135 through @xmath4 for which @xmath4 is the centroid of the intersection points with the rays @xmath195 .
this gives another proof of theorem [ thm : onevsmany ] , now in formulation [ ref : b ] .
the second derivative of the function @xmath176 has signature @xmath213 at the critical points of @xmath176 and signature @xmath214 at the non - critical points . for any @xmath215
the vector @xmath87 is isotropic for @xmath216 : @xmath217 since by lemma [ lem : concproj ] the quadratic form @xmath216 has a @xmath2-dimensional positive subspace , its signature is either @xmath213 or @xmath214 , depending on whether the isotropic vector @xmath87 belongs to the kernel or not .
it belongs to the kernel if and only if @xmath203 .
fix the points @xmath34 affinely spanning @xmath0 .
theorem [ thm : onevsmany ] says that there is a bijection between the points in the interior of @xmath218 and the hyperplanes disjoint from @xmath218 : a hyperplane @xmath45 corresponds to the point @xmath219 that becomes the centroid of @xmath46 when @xmath45 is sent to infinity by a projective transformation .
( in particular , the hyperplane at infinity corresponds to the actual centroid of @xmath46 . ) in this section we will relate this correspondence to the polarity with respect to an algebraic set . for
a homogeneous degree @xmath220 polynomial @xmath221 on a vector space @xmath222 denote by the same letter @xmath221 the corresponding @xmath220-linear symmetric form : @xmath223 let @xmath224 denote the projectivization of the vector space @xmath222 .
the _ @xmath225-st kernel _ of the @xmath220-linear symmetric form @xmath221 is @xmath226 the _
@xmath225-st polar _ of a point @xmath227 \in p(v \setminus \ker_{n-1}f)$ ] with respect to the projective algebraic set @xmath228 \in v \mid f(x ) = 0\}$ ] is the projective hyperplane @xmath229 \in v \mid p(x , \ldots , x , y ) = 0\}\ ] ] by the canonical duality , hyperplanes in @xmath222 correspond to one - dimensional subspaces of @xmath230 .
thus the polarity determines a map @xmath231 .
below we consider a polynomial @xmath221 on @xmath230 , therefore will have to do with a map @xmath232 [ prp : polarity ] let @xmath233 , and @xmath234 be @xmath235-dimensional subspaces of a vector space @xmath222 , and let @xmath236 be a linear functional on @xmath222 such that @xmath237 does nt contain any of @xmath238 .
then the following conditions are equivalent . 1 . the centroid of the points where the lines @xmath233 intersect the hyperplane @xmath239 lies on the line @xmath234 .
the polar of @xmath240 \in p(v^*)$ ] with respect to the algebraic set @xmath241 is @xmath242 .
choose arbitrary points @xmath243 and @xmath244 different from the origin .
by assumption , @xmath245 , @xmath246 .
the intersection points of the corresponding lines with @xmath239 are @xmath247 thus the first condition is equivalent to @xmath248 on the other hand , for the polynomial @xmath249 on @xmath230 we have @xmath250 thus the polar of @xmath251 is the following hyperplane in @xmath230 : @xmath252 dually , this is the @xmath235-dimensional subspace of @xmath222 spanned by @xmath253 .
thus the second condition is equivalent to @xmath254 which is equivalent to and thus to the first condition . in the reformulation
b of our problem about projective transformations ( see section [ sec : homogpoints ] ) we intersect with a hyperplane not a collection of lines , but a collection of rays .
this means that each of the points is assumed to lie in a specified half of the corresponding line , i. e. the number @xmath255 , respectively @xmath256 must have a specified sign . in other words ,
@xmath240 \in p(v^*)$ ] must lie in a specified component of the complement @xmath257 . by counting the components one can determine the multiplicity of the map , i. e. the number of classes of projective transformations that send @xmath4 to the centroid of the images of @xmath46 .
[ prp : nonadm ] let @xmath258 be in general position , that is each @xmath259 of them affinely independent .
then there are exactly @xmath260 equivalence classes of projective transformations @xmath5 modulo post - composition with affine transformations such that @xmath261 let @xmath262 and @xmath263 .
let @xmath264 be the @xmath235-dimensional subspace spanned by @xmath265 , respectively by @xmath266 .
let @xmath267 be such that @xmath268 , @xmath269 . for the points ( with @xmath265 and @xmath266 instead of @xmath183 and @xmath4 ) in the affine hyperplane @xmath270
there is a unique class of admissible projective transformations making @xmath4 to the centroid of @xmath46 if and only if @xmath271 lies in the interior of the convex hull of @xmath272 .
the latter condition says that @xmath251 belongs to a component of @xmath273 that is disjoint from @xmath274 . besides
, any two functionals from the same component give rise to the same class of projective transformations .
there are @xmath275 components in total , and @xmath274 intersects @xmath220 of them , which leads to the number in the proposition . without loss of generality @xmath276 , so that the hyperplane sent to infinity by @xmath5 can not pass through the origin . by proposition [ prp : a ] we may look for @xmath5 among the maps of the form .
the condition @xmath24 then says @xmath277 similar to the proof of theorem [ thm : onevsmanya ] , the solutions of are the critical points of the function @xmath278 defined in the interior of @xmath279 .
the assumption @xmath280 implies @xmath281 , so that the function tends to @xmath186 as @xmath87 tends to the boundary of @xmath282 , and hence attains its maximum .
the point of minimum yields a desired projective transformation . for non - uniqueness , see examples [ exl : countermanymany1 ] and [ exl : countermanymany2 ]
. the sum may tend to @xmath186 under less restrictive assumptions than @xmath280 .
for example , it does so when @xmath25 consists of the vertices of a triangle in @xmath283 , and @xmath26 of three points on the sides of the triangle . on the other hand ,
if @xmath25 is the vertex set of a tetrahedron in @xmath284 , and @xmath26 consists of three points on one edge and two points on the opposite edge , then there is no projective transformation that fits the centroids of @xmath25 and @xmath26 ( the centroid of a tetrahedron lies in the plane parallel to a pair of opposite edges and equidistant from them ) . in particular , in this case the sum does not tend to @xmath186 near the boundary of the domain .
: two subsets of the line whose centroids can be fitted together in different ways . ]
[ exl : countermanymany1 ] let @xmath285 , @xmath286 , @xmath287 .
then @xmath288 , so that one solution is the identity map .
there is another solution @xmath289 .
indeed , we have @xmath290 and thus @xmath291
. one may argue that the above example only works because @xmath25 is not in convex position . for @xmath285
this is actually true : if @xmath292 and @xmath293 , then the solution is unique . in higher dimensions
this does not help , as the following example shows .
the reason for the failure is that even if @xmath25 is in convex position , its projections are not . :
two subsets of the plane whose centroids can be fitted together in different ways . ]
[ exl : countermanymany2 ] take the following subsets of the plane : @xmath294 again , both have centroid at the origin .
their images under the projective transformation @xmath295 both have centroids at @xmath296 .
a subset @xmath134 is called a _ cone _ , if @xmath297 .
a cone @xmath132 is _ pointed _ , if @xmath298 is contained in an open halfspace whose boundary hyperplane passes through the origin .
a closed pointed cone possesses bounded sections by affine hyperplanes .
we will consider only those sections that intersect each ray of the cone , and call them _
a pointed cone is the conical hull of any of its complete sections
. for @xmath13 put @xmath299 .
due to propositions [ prp : a ] and [ prp : b ] , theorem [ thm : onevsbody ] is equivalent to the following .
[ thm : onevsbodyb ] let @xmath134 be a full - dimensional pointed cone with cusp - free affine sections .
then for every point @xmath300 there exists a unique complete affine section of @xmath132 with centroid at @xmath4 . for @xmath285
this means in particular that there is a unique chord of an angle @xmath132 that goes through a given point and has it as the midpoint .
this is a popular elementary geometry problem .
the endpoints of such a chord are found by intersecting @xmath301 with its image under rotation by @xmath302 about the point .
see figure [ fig : pointinangle ] .
propositions [ prp : a ] and [ prp : c ] imply the following reformulation of theorem [ thm : onevsbody ] in the convex case .
[ thm : onevsbodyc ] let @xmath155 be a convex @xmath2-dimensional body . then there exists a unique point @xmath156 such that the polar dual of @xmath86 with respect to @xmath87 has centroid at @xmath87 .
the point @xmath87 in theorem [ thm : onevsbodyc ] is called the santal point of @xmath86 .
its existence and uniqueness was proved in @xcite , see also @xcite .
let @xmath134 be a pointed full - dimensional cone .
for every hyperplane @xmath135 such that @xmath303 is compact and intersects all rays of @xmath132 , denote the bounded component of @xmath304 by @xmath305 .
[ prp : lever ] a hyperplane section @xmath136 of a cone @xmath132 has centroid at @xmath4 if and only if @xmath135 is a critical point of the function @xmath306 on the set of all hyperplanes through @xmath4 .
it vanishes in the first order if and only if @xmath307 is an equilibrium axis for @xmath303.,title="fig : " ] # 1#2#3#4#5 @font ( 3399,1871)(-686,-1020 ) ( 556,-722)(0,0)[lb ] ( 757,-419)(0,0)[lb ] ( -516,-243)(0,0)[lb ] consider two hyperplanes @xmath135 and @xmath308 through @xmath4 close to each other .
then we have @xmath309 where @xmath310 is the linear function whose graph is @xmath308 .
[ fig : lever ] . thus @xmath135 is critical for @xmath311 if and only if all integrals over @xmath303 of linear functions vanishing at @xmath4 vanish .
on the other hand , vanishing of @xmath312 for all linear functions @xmath221 with @xmath313 is equivalent to @xmath4 being the centroid of @xmath303 .
( think of @xmath221 as the gravity torque with respect to the axis @xmath314 . ) in 1931 , tricomi @xcite and guido ascoli @xcite showed that for every point inside a convex body there exists a hyperplane section that has this point as a centroid .
tricomi dealt only with dimension @xmath315 using the `` hairy ball theorem '' .
ascoli used a variational approach based on proposition [ prp : lever ] .
they also characterized those non - convex bodies , for which the centroid of a section depends continuously on the hyperplane , making both approaches applicable . for more details
see ( * ? ? ?
* , section 8) that also deals with a beautiful related object , dupin s `` floating body '' .
filliman @xcite studied critical sections of polytopes and gave their characterization in the case of a simplex .
our plan now is to show that for cones with cusp - free sections the function is coercive , which implies the existence of a critical point , and then to prove some sort of convexity of to ensure the uniqueness of the critical point . for a non - zero vector @xmath316 denote @xmath317
the section @xmath153 is compact and intersects each ray of @xmath132 if and only if @xmath318 , where @xmath319 note that if @xmath132 is @xmath320-dimensional , closed and pointed , then @xmath321 is also @xmath320-dimensional , closed and pointed . for every @xmath318 denote @xmath322 that is
, @xmath323 is the bounded part of @xmath132 cut off by the hyperplane @xmath151 .
theorem [ thm : onevsbodyb ] is proved by using variational properties of the function @xmath324{{\operatorname{vol}}(c_y)}\ ] ] the following arguments are a slight modification of @xcite and @xcite .
[ lem : momenta ] we have @xmath325 represent @xmath132 as the union of parallel slices @xmath326 .
the distance between the hyperplanes @xmath327 and @xmath328 equals @xmath329 , therefore @xmath330 , where @xmath331 is the @xmath2-dimensional lebesgue measure on the hyperplanes orthogonal to @xmath87 .
thus we have @xmath332 because @xmath323 is a pyramid over @xmath333 with the altitude @xmath334 .
the second and the third integrals are computed similarly .
take into account that @xmath335 . in particular
, the function equals @xmath336
[ lem : gradf ] the gradient of @xmath176 is the centroid of the section : @xmath337 let us compute the derivative of @xmath176 in the direction @xmath338 : @xmath339 now the result follows from lemma [ lem : momenta ] .
[ lem : infinitybody ] if @xmath19 is cusp - free , then the value @xmath185 tends to @xmath340 as @xmath87 tends to a point in @xmath341 .
let @xmath342 .
then there exists @xmath343 such that @xmath344 and @xmath345 .
clearly , @xmath346 .
thus , by assumption of theorem [ thm : onevsbody ] there are vectors @xmath347 such that their positive hull @xmath348 is contained in @xmath132
. then we have @xmath349 for some positive constant .
hence @xmath350 as @xmath87 tends to @xmath351 , the scalar product @xmath352 tends to @xmath1 while other scalar products remain bounded below by a positive constant ( some of them may also tend to @xmath340 ) .
hence @xmath353 .
[ prp : convexbody ] the function @xmath176 is strictly convex .
we have @xmath354 using lemma [ lem : momenta ] , we get @xmath355 due to the functional arithmetic - quadratic mean inequality @xmath356 ( which is the @xmath357 cauchy - schwarz inequality for functions @xmath221 and @xmath235 ) we have @xmath358 the hyperplane @xmath151 passes through the point @xmath4 if and only if the hyperplane @xmath359 passes through @xmath87 .
the section @xmath153 is bounded if and only if @xmath215 .
thus the hyperplane sections of @xmath132 coming into question are @xmath360 restrict the function @xmath176 defined in to @xmath361 . by lemma [ lem : gradf ]
we have @xmath362 ( this is the projection of @xmath363 to @xmath359 ; one may also evoke lagrange multipliers . )
thus @xmath4 is the centroid of @xmath364 if and only if @xmath87 is a critical point of @xmath365 .
lemma [ lem : infinitybody ] implies that @xmath176 attains a minimum on @xmath366 , which shows the existence part of theorem [ thm : onevsbodyb ] .
the uniqueness follows from proposition [ prp : convexbody ] , similarly to the proof of theorem [ thm : onevsmanya ] .
[ exl : counterpointbody ] let @xmath367 and @xmath368 .
we claim that none of the maps @xmath369 has the property @xmath370 .
we have @xmath371 .
the set @xmath372 is symmetric with respect to the @xmath36-axis , and it can be shown that for @xmath373 the image under @xmath374 of an @xmath36-symmetric set has its centroid outside the @xmath36-axis .
it follows that the only candidates for @xmath5 are the maps @xmath375 for @xmath376 to be admissible , we have to assume @xmath377 .
( 0,0 ) # 1#2#3#4#5 @font ( 5379,2274)(-1136,-298 ) ( 2836,1109)(0,0)[lb ] ( 136,974)(0,0)[lb ] ( 496,-151)(0,0)[lb ] a direct computation shows that the centroid of @xmath372 always has a negative @xmath36-coordinate .
in particular , in the limit case @xmath378 we have @xmath379 see figure [ fig : counterpointbody ] , and @xmath380 denote by @xmath381 the polar dual of @xmath19 .
propositions [ prp : c ] and [ prp : b ] imply @xmath382 therefore finding the minimum of @xmath383 is equivalent to finding the minimum over all @xmath87 of the volume of the polar dual of @xmath86 with respect to @xmath156 .
this is the second characterization of the santal point of a convex body @xmath86 , the first having been given in theorem [ thm : onevsbodyc ] .
the maximum of the product @xmath384 over all origin - symmetric convex bodies is achieved when @xmath86 is an ellipsoid .
this is the _ blaschke - santal _
inequality @xcite .
the minimum of @xmath384 is not known , but is conjectured to be achieved when @xmath86 is a cube or cross - polytope or , more generally , hanner polytopes ( _ mahler conjecture _ ) .
here we prove theorem [ thm : manyvsbody ] . by combining the functionals and we see that the classes of projective transformations satisfying are in a @xmath235-to-@xmath235 correspondence with the critical points of the function @xmath385 if @xmath386 for all @xmath174 , then , for a cusp - free @xmath19 , the integral tends to @xmath340 as @xmath87 tends to @xmath387 , while the sum remains bounded . this implies the existence if all @xmath183 lie in the interior of @xmath388 .
if some of them lie on the boundary , then we need a more delicate argument .
[ lem : infinitybodysev ] if every support hyperplane of @xmath19 contains less than @xmath54 of the points @xmath34 , then the function @xmath176 tends to @xmath340 as @xmath87 tends to a point in @xmath389 .
let @xmath390 .
as in the proof of lemma [ lem : infinitybody ] , choose a point @xmath391 such that @xmath392 .
then we have @xmath393 now , we have @xmath394 . if for all @xmath174 this inequality is strict , then all @xmath395 remain bounded as @xmath396 , so that @xmath353 .
let @xmath397 and @xmath398 for @xmath399 .
as @xmath400 , there exist @xmath401 such that @xmath402 this implies @xmath403 for all @xmath174 .
it follows that @xmath404 on the other hand , @xmath405 where @xmath406 . collecting all terms we get @xmath407 due to @xmath408 and @xmath409 , all coefficients before the logarithms are negative .
hence @xmath353 .
if the hyperplane @xmath410 contains @xmath234 of the points @xmath34 , then @xmath185 is bounded below by a sum of logarithms with coefficients @xmath411 , which are still negative provided that @xmath412 .
the restriction on the points lying on the boundary of the convex hull is necessary for the existence , as the following example shows .
[ exl : d+1points ] let @xmath19 be the union of two @xmath2-simplices whose intersection is a @xmath413-face of both ( a bipyramid ) , and let @xmath414 be the vertices of one of the simplices .
then the centroid of @xmath415 coincides with the centroid of the corresponding simplex and therefore is different from the centroid of @xmath19 .
no projective transformation can help .
alternatively , take @xmath315 points on one edge of the tetrahedron and @xmath416 points on the opposite edge .
the centroid of the points lies on a plane parallel to both edges that divides the distance between them in proportion @xmath417 .
the centroid of the tetrahedron lies on a plane equidistant from both edges . :
no support hyperplane may contain too many @xmath183 . ] in the following example the transformation @xmath5 is not unique .
[ exl : counterbodymany ] let @xmath418 be the square with vertices @xmath419 , @xmath420 , and @xmath421 both @xmath19 and @xmath422 have centroid at the origin .
the images of both sets under a projective non - affine transformation @xmath423 have centroids at @xmath424 .
compare this with examples [ exl : countermanymany1 ] and [ exl : countermanymany2 ] .
the coincidences are not accidental . here
we prove theorem [ thm : manyvsball ] .
as usual , we form the difference of functions @xmath425{{\operatorname{vol}}(c_y ) } , \quad f_2(y ) { \colonequals}-\frac1n \sum_{i=1}^n \log \langle p_i , y \rangle\ ] ] where @xmath426 is the cone over the unit ball .
the domain of @xmath176 is the interior of the dual cone @xmath427 which coincides this time with @xmath132 .
we have @xmath428 where @xmath429 is the volume of a @xmath2-dimensional unit ball , and @xmath430 is the minkowski norm of @xmath87 .
as @xmath431 is homogeneous of degree @xmath432 with respect to @xmath87 ( scaling @xmath87 by @xmath433 results in scaling the truncated cone @xmath323 by @xmath434 ) , it suffices to show that @xmath435 if @xmath436 .
the equation @xmath437 of the hyperplane @xmath151 can be rewritten as @xmath438 with @xmath439 .
it follows that for @xmath436 the hyperplane @xmath151 is tangent to the upper half of the hyperboloid @xmath440 .
the group @xmath441 of linear transformations preserving the minkowski scalar product acts transitively on the set of such hyperplanes , hence there is a transformation @xmath442 that maps @xmath151 to the hyperplane @xmath443 . since @xmath444 , we have @xmath445 . the latter is a cone of height @xmath235 over the unit ball and has volume @xmath446 . as a result
we have @xmath447 by results of section [ sec : homogpoints ] and lemma [ lem : gradf ] , @xmath448 where @xmath449 is the ray generated by @xmath183 .
therefore we have to show that the function @xmath176 has a unique , up to scaling , critical point .
note also that @xmath450 , so that it suffices to consider the restriction of @xmath176 to any subset that is represented in all equivalence classes @xmath451 .
two convenient choices are @xmath452 and @xmath453 .
[ lem : coercball ] the function @xmath176 tends to @xmath340 as @xmath87 tends to a point in @xmath454 .
let @xmath455 with @xmath456 . then @xmath457 . if @xmath458 for all @xmath174 , then the other summands in remain bounded , and the sum tends to @xmath340 . if there is an @xmath174 such that @xmath459 , then @xmath460 , so that only the @xmath174-th summand under the sum sign in tends to @xmath186 .
we then have @xmath461 where @xmath462 .
this already implies the existence of a critical point of @xmath176 .
for uniqueness we would like to use convexity , but the following example shows that @xmath176 is not always convex .
for @xmath463 put @xmath464 and consider the restriction of @xmath185 to the line @xmath465 .
there we have @xmath466 for @xmath467 this function is not convex .
the following trick helps .
[ lem : geodconv ] the function @xmath176 is geodesically strictly convex with respect to the hyperbolic metric on @xmath468 .
when restricted to @xmath469 , the function @xmath176 has the form @xmath470 every geodesic is represented by a hyperplane section of the hyperboloid @xmath440 , and has a unit speed parametrization of the form @xmath471 where @xmath472 , @xmath473 , @xmath474 .
let us study the restrictions of the @xmath174-th summand in to geodesics .
if @xmath475 , then on any geodesic it is possible to choose @xmath4 and @xmath476 so that @xmath477 .
we get @xmath478 which is strictly convex .
if @xmath479 and the geodesic does nt have @xmath480 as a limit point , then one can do the same .
if @xmath479 and the geodesic has @xmath480 as a limit point , then for any parametrization we have @xmath481 , so that @xmath482 thus along such a geodesic the function is linear .
the only possibility for the sum to be linear along a geodesic ( and thus non - strictly convex ) is that all points @xmath46 have @xmath479 and lie on that geodesic .
this is only possible for @xmath483 . by lemma [ lem : coercball ]
, the function @xmath176 has a critical point inside the cone @xmath132 . by lemma [ lem : geodconv ] , this critical point is unique up to scaling , because otherwise the restriction of @xmath176 to a geodesic would have two different critical points , which contradicts the strict geodesic convexity of @xmath176 .
[ rem : moebius ] this argument generalizes that of springborn @xcite , who considers only points on the sphere . in this case
, the function @xmath176 is the sum of hyperbolic distances to horospheres centered at the given points .
for a point @xmath64 inside the ball , the term @xmath484 equals @xmath485 , where @xmath486 is the hyperbolic distance , and @xmath487 . in the case when all @xmath183 lie on the sphere , the critical point of the function @xmath176 is the so called conformal barycenter of @xmath46 . in @xcite ,
the conformal barycenter was defined for non - atomic measures on the sphere , and the construction for discrete measures was indicated .
the `` centroid '' of points in the hyperbolic space can be defined in different ways .
one of the possibilities is to take the affine centroid of the points on the hyperboloid and centrally project back ; this point minimizes @xmath488 .
another possibility is to minimize @xmath489 as in the general definition of the riemannian center of mass @xcite .
we approach theorem [ thm : bodyvsbody ] in reformulation [ ref : b ] : take cones @xmath490 and @xmath491 over @xmath25 and @xmath26 , respectively , and see under what conditions there is an affine hyperplane @xmath135 such that @xmath492 and @xmath493 have a common centroid .
following section [ sec : volvar ] , introduce the functions @xmath494{{\operatorname{vol}}(c_{i , y } ) } , \quad i = 1 , 2\ ] ] where @xmath495 is the cone @xmath496 truncated by the hyperplane @xmath151 .
their difference @xmath497 has , according to lemma [ lem : gradf ] , the gradient @xmath498 thus the following lemma holds .
projective transformations that fit the centroids of @xmath25 and @xmath26 correspond , modulo post - composition with affine transformations , to critical points of the function @xmath176 from .
the existence of a critical point follows by the usual argument . by lemma [ lem : infinitybody ]
, the function @xmath499 tends to @xmath340 as @xmath87 tends to @xmath500 .
the function @xmath501 is continuous on @xmath502 , and therefore bounded on @xmath503 .
thus @xmath353 as @xmath504 .
as next we give an example where the projective transformation fitting the centroids is not unique .
[ exl : counterbodybody ] take a unit disk and the following rectangle inside it : @xmath505 for some @xmath506 .
both @xmath25 and @xmath26 have centroid at the origin . on the other hand , the projective transformation @xmath507 maps the disk @xmath25 to itself , and the rectangle @xmath26 to a trapezoid that , as a tedious computation shows , also has centroid at the origin .
see fig .
[ fig : countertwobodies ] .
an example of this sort is possible whenever the rectangle has a side which is longer than @xmath508 .
we will consider the restriction of @xmath176 to the section of @xmath509 by the hyperplane @xmath510 . since @xmath450 , every critical point of this restriction is a critical point of @xmath176 .
we are not able to prove that @xmath511 is convex under assumption , but we can prove that it is strictly convex at every critical point . since the indices of critical points of a function defined on a ball and tending to @xmath340 near boundary sum up to @xmath1 , this implies that the critical point is unique .
[ lem : d2f ] let @xmath316 be such that @xmath510 .
then for every vector @xmath338 with @xmath512 we have @xmath513 where @xmath514 is the moment of inertia of @xmath19 with respect to the vector @xmath209 . from we
have @xmath515 by proposition [ prp : b ] , @xmath516 is the image of @xmath517 under parallel projection along @xmath144 .
therefore @xmath518 also , @xmath519 differs from @xmath520 by a multiple of @xmath144 .
it follows that for @xmath512 we can replace @xmath517 by @xmath516 in the formula for @xmath521 .
further , for any @xmath13 and @xmath522 we have @xmath523 by substituting this into the last equation we obtain @xmath524 and the lemma follows .
[ lem : momwidth ] let @xmath13 be a convex body , and @xmath525 be a non - zero vector
. then we have @xmath526 where @xmath527 is the width of @xmath19 in the direction of @xmath209 .
the lower bound is achieved for a symmetric double cone over any @xmath413-dimensional body , the upper bound is achieved for the cylinder over any @xmath413-dimensional body . if each section of @xmath19 orthogonal to @xmath209 is replaced by a @xmath413-ball of the same radius , then the body remains convex and preserves its volume and moment of inertia in the direction @xmath209 .
thus , without loss of generality , @xmath19 is a `` rotation body '' with axis @xmath209 .
we will use the fact that moving mass away from the centroid increases the moment of inertia , and moving towards decreases the moment . by the above principle
, the steiner symmetrization with respect to @xmath528 preserves the volume but decreases the moment .
the resulting body is symmetric with respect to a hyperplane orthogonal to @xmath209 , and it is possible to move more mass towards the centroid by replacing each of the symmetric halves by a cone of the same volume with the base on the hyperplane of symmetry .
the two steps are illustrated on fig .
[ fig : momentcone ] .
the convex profiles stand for the radii of the sections orthogonal to @xmath209 ; equally colored regions correspond to sets of equal @xmath2-volume . for a double cone of width
@xmath529 we have @xmath530 which yields the lower bound in the theorem . in order to prove the upper bound , first replace @xmath19 with a truncated cone @xmath25 whose parts on either sides from the hyperplane through the centroid of @xmath19 have the same volumes as the corresponding parts of @xmath19 .
one can go from @xmath19 to @xmath25 by moving mass away from the centroid , see figure [ fig : momentcyl ] , therefore @xmath25 has a bigger moment of inertia .
it turned out unexpectedly hard to prove directly that the cylinder maximizes the moment of inertia among all truncated cones of a fixed volume , therefore we will continue to move mass .
we replace @xmath25 by a union @xmath26 of a cylynder and a truncated cone as shown on figure [ fig : momentcyl ] .
a direct computation shows that the requirement @xmath531 leads to a convex @xmath26 ( the radius of the cone decreasing as on figure [ fig : momentcyl ] ) .
also , the section of @xmath532 by a hyperplane orthogonal to @xmath209 has a smaller volume as the section of @xmath533 at the same distance from the centroid of @xmath25 .
this allows to map @xmath533 to @xmath532 so that the mass is moved away from the centroid of @xmath25 .
thus @xmath534 . the body @xmath26 can be replaced by a truncated cone @xmath535 with a bigger moment , as it was done at the first step . by iterating the procedure
, we obtain a sequence of bodies converging to a cylinder .
this implies that the cylinder maximizes the moment of inertia for given volume and width .
the ratio for the cylinder of width @xmath529 equals @xmath536 [ lem : crwidth ] let @xmath537
. then for every @xmath538 we have @xmath539 for a fixed cross - ratio , the maximum of @xmath540 is achieved when the segments are concentric , that is @xmath541 .
we have @xmath542 and the lemma follows .
[ lem : morse ] let @xmath543 be homeomorphic to the ball @xmath59 , and @xmath544 be a smooth function on the interior of @xmath545 .
assume that @xmath353 as @xmath546 and that @xmath216 is positive definite at every critical point .
then the critical point of @xmath176 is unique .
the function @xmath176 attains its minimum in @xmath545 , therefore it has at least one critical point . due to @xmath547 at all critical points , critical points are isolated . due to @xmath353 as @xmath546 ,
the critical values form a discrete subset of @xmath178 , and in particular can be ordered : @xmath548 let @xmath549 . by the morse theory ,
the set @xmath550 is homeomorphic to the union of @xmath14 open disks , with @xmath551 lying in different disks . on the other hand ,
if we choose a path @xmath552 \to d$ ] joining @xmath553 and @xmath554 and take @xmath555 } f(y)\ ] ] then @xmath553 and @xmath554 lie in the same component of @xmath550 .
this contradiction shows that the critical point is unique .
we are considering the restriction of @xmath176 to @xmath556 . from section
[ sec : existtwobodies ] we know that @xmath353 as @xmath557 .
let us show that under assumption the quadratic form @xmath558 is positive definite at all critical points .
due to lemma [ lem : d2f ] , at a critical point we have @xmath559 consider the orthogonal to @xmath209 support hyperplanes @xmath560 and @xmath561 of @xmath562 and @xmath563 .
they are images under @xmath140 of support hyperplanes of @xmath25 and @xmath26 that are either parallel or share a @xmath68-dimensional affine subspace . since the cross - ratio is projectively invariant , holds for @xmath560 and @xmath561 . by lemma [ lem : crwidth ] we have @xmath564 which , by lemma [ lem : momwidth ] , implies @xmath565 thus , at every critical point @xmath566 for all @xmath212 . by lemma [ lem : morse ] , this implies that the critical point is unique .
let us show that the bound is sharp .
take as @xmath25 a double cone of height @xmath235 over a @xmath413-dimensional subset of @xmath567 with centroid at the origin , and as @xmath26 a cylinder of height @xmath568 over a similar , but smaller , set .
for example , @xmath25 may be the standard cross - polytope , and @xmath26 a rectangular parallelepiped .
then the centroids of @xmath25 and @xmath26 coincide , so that @xmath144 is a critical point of the function @xmath176 .
the quadratic form @xmath569 takes a negative value in the direction of the axis of @xmath25 and @xmath26 .
therefore @xmath144 is not the minimum point of @xmath176 .
thus a minimum point provides a non - affine projective transformation that fits the centroids of @xmath25 and @xmath26 .
[ lem : momball ] for a @xmath2-dimensional ball @xmath570 of radius @xmath476 we have @xmath571 the moment @xmath572 does nt depend on @xmath209 .
the sum of the moments in @xmath2 pairwise orthogonal directions equals the polar moment @xmath573 .
thus we have @xmath574 where @xmath575 is the @xmath413-dimensional sphere of radius @xmath576 , and @xmath577 is the volume of the unit @xmath413-sphere . on the other hand , @xmath578 , which leads to the formula of the lemma .
similar to the proof of theorem [ thm : bodyvsbody ] , it suffices to show that @xmath579 for all @xmath87 .
the image of a ball under an admissible projective transformation is an ellipsoid .
it is easily seen that the normalized moment of inertia @xmath580 of an ellipsoid equals to the normalized moment of a ball with diameter equal to the width of the ellipsoid in direction @xmath209 : @xmath581 because of lemmas [ lem : momwidth ] and [ lem : momball ] we have @xmath582 to ensure the latter inequality for all @xmath209 , it suffices to require @xmath583 for all quadruples of parallel tangent hyperplanes to @xmath584 and @xmath152 .
this , in turn , is implied by the same inequality for concurrent tangent hyperplanes to @xmath59 and @xmath19 . without loss of generality ,
assume @xmath162 ( this may be achieved by a projective transformation that fixes @xmath585 , and the theorem is of projective nature ) .
by section [ sec : comppolar ] , @xmath586 , so that @xmath587 use the same method as in the proofs of theorems [ thm : bodyvsbody ] and [ thm : bodyvsball ] .
the assumption of the theorem implies @xmath588 which implies @xmath589 for all @xmath87 , and in particular for critical points of the function @xmath176 . due to lemmas [ lem : momwidth ] and [ lem : momball ] , this implies that at the critical points @xmath176 is strictly convex .
thus by lemma [ lem : morse ] the critical point is unique .
we restricted our attention to the cases @xmath590 because it implies the affine covariance of the centroid , which makes it possible to formulate the uniqueness problem ( projective transformations modulo affine ones ) .
for @xmath591 the centroid is affinely covariant under some additional restrictions , for example if @xmath19 is centrally symmetric or if the affine span of @xmath19 has dimension @xmath14 . by theorem [ thm : onevsbody ] ,
every point inside a cusp - free set becomes the centroid after some projective transformation .
example [ exl : counterpointbody ] gives a point inside a set lying close to a sharp cusp so that no projective transformation can make it the centroid .
weaken the cusp - freeness condition so that to preserve the existence of a projective transformation for any point in the interior .
is the following condition necessary and sufficient : the image of @xmath19 under any projective transformation that sends some support hyperplane of @xmath19 to infinity has infinite volume ?
we conjecture the following solution of the latter problem : for every support hyperplane @xmath45 of @xmath19 take a projective transformation @xmath593 that sends @xmath45 to infinity .
if @xmath594 has a finite volume , then cut off by a hyperplane parallel to @xmath45 half of the volume of @xmath594 ( remove the infinite part ) . the complement in @xmath19 to the preimages of all parts removed in this way
is the set of points that can become the centroid . in examples [ exl : counterbodymany ] and [ exl : d+1points ]
we saw that the projective transformation fitting the centroids of a finite set and of a `` bigger '' @xmath2-dimensional set does not always exist , and if , then may be not unique .
by contrast , if the @xmath2-dimensional set is a ball , then we have unconditional existence and uniqueness . our method to prove the existence was to show that the function @xmath176 tends to @xmath340 near the boundary of its domain .
this was ensured by the assumption @xmath595 .
it is possible to prove the existence of a critical point of @xmath176 under less restrictive assumptions , for example , if @xmath596 and the gradient of @xmath176 `` turns '' as we go along the boundary of domain of @xmath176 .
this means that a projective transformation exists if @xmath26 `` sticks out '' of @xmath25 in at least two places .
one could formalize and generalize this argument by using the degree of the map from the boundary of the domain of @xmath176 to the @xmath413-dimensional sphere .
j. saint - raymond .
sur le volume des corps convexes symtriques . in _
initiation seminar on analysis : g. choquet - m .
rogalski - j .
saint - raymond , 20th year : 1980/1981 _ , volume 46 of _ publ .
pierre et marie curie _ , pages exp .
paris vi , paris , 1981 . | given two subsets of @xmath0 , when does there exist a projective transformation that maps them to two sets with a common centroid ?
when is this transformation unique modulo affine transformations ?
we study these questions for @xmath1- and @xmath2-dimensional sets , obtaining several existence and uniqueness results as well as examples of non - existence or non - uniqueness .
if both sets have dimension @xmath1 , then the problem is related to the analytic center of a polytope and to polarity with respect to an algebraic set . if one set is a single point , and the other is a convex body , then it is equivalent by polar duality to the existence and uniqueness of the santal point . for a finite point set versus a ball ,
it generalizes the mbius centering of edge - circumscribed convex polytopes and is related to the conformal barycenter of douady - earle .
if both sets are @xmath2-dimensional , then we are led to define the santal point of a pair of convex bodies .
we prove that the santal point of a pair exists and is unique , if one of the bodies is contained within the other and has hilbert diameter less than a dimension - depending constant .
the bound is sharp and is obtained by a box inside a cross - polytope . |
we would like to thank alessandro culatti , giuseppe degrassi , antonio masiero , joaquim matias , luca silvestrini , alessandro vicini and fabio zwirner for many useful discussions and comments .
99 k. gaemers , g.j .
gounaris , _ z. phys . _ * c1 * , ( 1979 ) 259 . k. hagiwara , k. hikasa , r. peccei and d. zeppenfeld , _ nucl .
* b282 * ( 1987 ) 253 .
m. bilenky , j.l .
kneur , f.m .
renard , d. schildknecht , _ nucl . phys . _ * b409 * ( 1993 ) 22 ; + m. bilenky , j.l .
kneur , f.m .
renard , d. schildknecht , _ nucl .
phys . _ * b419 * ( 1994 ) 240 ; + r. l. sekulin , _ phys .
lett . _ * b338 * ( 1994 ) 369 . for a recent review of standard model and new physics effects in @xmath54 process
see for example : + proceedings of the workshop on physics at lep2 , ed .
g. altarelli , t. sjstrand and f. zwirner , cern 96 - 01 ; + proceedings of the workshop on @xmath55 collisions at tev energies , annecy , gran sasso , hamburg , february 1995-september 1995 , ed .
p. m. zerwas , desy 96 - 123d ; + f. boudjema , proceedings of the workshop on physics and experiments with linear @xmath55 colliders , eds .
f. a. harris et al . , p. 712
, world scientific , 1994 ; + f. jegerlehner , in teupitz 1994 , proceedings of physics at lep200 and beyond , _ nucl .
_ proc . suppl . * 37b * ( 1994 ) 129 ; + d. zeppenfeld , talk given at the international symposium on vector boson self - interactions , feb
. 1 - 3 , 1995 , ucla , los angeles , hep - ph/9506239 .
argyres , c.g .
papadopoulos , _ phys .
* b263 * ( 1991 ) 298 ; + e.n .
argyres , g. katsilieris , a.b .
lahanas , c.g .
papadopoulos , v.c .
spanos , _ nucl .
* b391 * ( 1993 ) 23 ; + r. casalbuoni , p. chiappetta , a. deandrea , s. de curtis , d. dominici and r. gatto , _ z. phys _ * c 60 * ( 1993 ) 315 ; + l. anichini , r. casalbuoni and s. de curtis , _ phys . lett _ * b348 * ( 1995 ) 521 ; + j. fleischer , j.l .
kneur , k. kolodziej , m. kuroda , d. schildknecht , _ nucl .
_ * b378 * ( 1992 ) 443 and _ nucl .
phys . _ * b426 * ( 1994 ) 246 ; + a.b .
lahanas , v.c .
spanos , _ phys .
lett . _ * b334 * ( 1994 ) 378 and hep - ph/9504340 ; + a. culatti , _ z. phys . _
* c65 * ( 1995 ) 537 ; + a. arhrib , j. l. kneur and g. moultaka , _ phys .
lett . _ * b376 * ( 1996 ) 127
. j. m. cornwall , d. n. levin and g. tiktopoulos , _ phys . rev _ * d10 * ( 1974 ) 1145 and erratum ibidem *
d11 * ( 1975 ) 972 . m. e. rose in elementary theory of angular momentum , wiley , new york , 1957 .
u. baur and d. zeppenfeld , _ phys .
* b201 * ( 1988 ) 146 ; + g. j. gounaris , j. layssac and f. m. renard , _ phys .
lett . _ * b332 * ( 1994 ) 146 .
a. de rjula , m. b. gavela , p. hernandez and e. masso , _ nucl .
phys . _ * b384 * ( 1992 ) 3 .
a. culatti , g. degrassi , f. feruglio , a. masiero , s. rigolin , l. silvestrini and a. vicini , in proceedings of the workshop on @xmath55 collisions at tev energies , annecy , gran sasso , hamburg , february 1995-september 1995 , ed .
p. m. zerwas , desy 96 - 123d , hep - ph/9505415 .
m. lemoine and m. veltman , _ nucl . phys _
* b164 * ( 1980 ) 445 ; + r. philippe , _ phys .
* d26 * ( 1982 ) 1588 ; + m. bhm , a. denner , t. sack , w. beenakker , f. berends and h. kuijf , _ nucl . phys _ * b304 * ( 1988 ) 463 ; + w. beenakker , k. kolodziej and t. sack , _ phys .
_ * b258 * ( 1991 ) 469 ; + j. fleischer , f. jegerlehner and m. zralek , _ z. phys .
_ * c42 * ( 1989 ) 409 ; + a. denner , _ fortschr .
* 41 * ( 1993 ) 307 .
g. altarelli and r. barbieri , _ phys .
_ * b253 * ( 1991 ) 161 ; + g. altarelli , r. barbieri and s. jadach , _ nucl .
phys . _ * b369 * ( 1992 ) 3 ; + g. altarelli , r. barbieri and f. caravaglios , _ nucl .
* b405 * ( 1993 ) 3 and proceedings of `` les rencontres de la vallee daoste , @xmath56 : results and perspectives in particle physics '' , la thuile , italy , march 7 - 13 , 1993 , ed .
m. greco , editions frontieres .
r. barbieri , f. caravaglios and m. frigeni , _ phys .
lett . _ * b279 * ( 1992 ) 169 . m. peskin and t. takeuchi , _ phys
lett . _ * 65 * ( 1990 ) 964 ; + w. marciano and j. rosner , _ phys .
lett . _ * 65 * ( 1990 ) 2963 ; + d. kennedy and p. langacker , _ phys .
* 65 * ( 1990 ) 2967 ; + b. holdom and j. terning , _ phys .
_ * b247 * ( 1990 ) 88 ; + a. dobado , d. espriu , and m.j .
herrero , _ phys . lett . _
* b255 * ( 1991 ) 405 .
abbott _ nucl .
_ * b185 * ( 1981 ) 189 ; + l.f .
abbott , m.t .
grisaru , r.k .
schaefer _ nucl .
phys . _ * b229 * ( 1983 ) 373 .
a. denner , s. dittmaier , g. weinglein , _ nucl .
- proc . suppl .
_ * 37b * ( 1994 ) 87 .
a. denner , s. dittmaier , g. weinglein , _ nucl .
phys . _ * b440 * ( 1995 ) 95 .
cornwall , _ phys .
_ * d26 * ( 1982 ) 1453 ; + g. degrassi , a. sirlin , _ phys .
* d46 * ( 1992 ) 3104;g .
degrassi , b. kniehl , a. sirlin , _ phys .
* d48 * ( 1993 ) r3963 ; + w.j .
papavassiliou , k. philippides _ phys .
* d48 * ( 1993 ) 4255 ; w.j .
papavassiliou , _ phys .
_ * d50 * ( 1994 ) 5958 . c. ahn , w. lynn , m. peskin and s. selipsky , _ nucl .
phys . _ * b309 * ( 1988 ) 221 . t. appelquist and c. bernard , _ phys
* 22 * ( 1980 ) 200 ; + a.c .
longhitano , _ phys .
rev . _ * d22 * ( 1980 ) 1166 ; + t. appelquist in gauge theories and experiments at high energies , ed . by k.c .
brower and d.g .
sutherland , scottish university summer school in physics , st .
andrews ( 1980 ) ; + a.c .
longhitano , _ nucl .
_ * b188 * ( 1981 ) 118 ; + t. appelquist and g - h .
wu , _ phys .
_ * d48 * ( 1993 ) 3235 ; + m. herrero and e. ruiz morales , _ nucl .
_ * b418 * ( 1994 ) 431 ; + m. herrero and e. ruiz morales , _ nucl .
_ * b437 * ( 1995 ) 319 ; + d. espriu and j. matias , _ phys .
_ * b341 * ( 1995 ) 332 ; + j. matias , _ nucl
* b478 * ( 1996 ) 90 ; + e. dhoker and e. fahri , _ nucl .
_ * b248 * ( 1984 ) 59 and * b248 * ( 1984 ) 77 ; + e. dhoker , _ phys . rev . lett . _
* 69 * ( 1992 ) 1316 ; + g.l .
lin , h. steger and y.p .
yao , _ phys .
_ * d44 * ( 1991 ) 2139 ; + f. feruglio , l. maiani and a. masiero , _ nucl . phys _ * b387 * ( 1992 ) 523 .
b holdom , _ phys .
_ * b258 * ( 1991 ) 156 .
a. masiero , f. feruglio , s. rigolin , r. strocchi , _ phys .
* b355 * ( 1995 ) 329 ; + t. inami , c.s .
lim , b. takeuchi , m. tanabashi , _ phys .
lett . _ * b381 * ( 1996 ) 458 . | at very large energies and in @xmath0 gauge theories , the trilinear gauge boson vertices relevant for @xmath1 scattering are related in a simple way to the gauge boson self - energies .
we derive these relations , both from the requirement of perturbative unitarity and from the ward identities of the theory .
our discussion shows that , in general , it is never possible to neglect vector boson self - energies when computing the form factors that parametrize the @xmath2 helicity amplitudes .
the exclusion of the self - energy contributions would lead to estimates of the effects wrong by orders of magnitude .
we propose a simple way of including the self - energy contributions in an appropriate definition of the form factors .
16.1 cm 23.0 cm -2.0 cm -1.5 cm 2 @=11 # 1 = by60 = # 1bsphack@filesw gtempaauxout gtempa @nobreak esphack eqnlabel#1 eqnlabel vacuum # 1
@=12 2.5 cm * sum rules for asymptotic form factors + .4 cm in @xmath1 scattering * 2.0 cm f. feruglio and s. rigolin + .1 cm _ dipartimento di fisica , universit di padova and infn , i-35131 padua , italy _ 2.0 cm 2truecm * 1 . *
it has become customary to analyse the @xmath3 pair production in @xmath4 machines by parametrizing the scattering amplitude in terms of a set of form factors characterizing the most general trilinear gauge vertex ( tgv ) involving a neutral vector boson @xmath5 ( @xmath6 ) and two on - shell @xmath3 s @xcite . by studying the angular distributions of the w it is possible to extract some informations on the form factors , thus providing new tests of the underlying electroweak theory @xcite .
any sm extension can be easily constrained , by analysing directly the contribution of the new particles and/or interactions to the different anomalous couplings . indeed ,
in any recent analysis of @xmath7 @xcite , the above mentioned form factors play a crucial role and represent an important ingredient for a meaningful comparison between theory and experiment .
for this reason any new information on the trilinear gauge vertices is welcome , particularly if it comes from the theory .
purpose of this note is to show that , at large energies , the tgv relevant to the @xmath2 process satisfy a set of relations that will be called sum rules .
these relations can be viewed as a direct consequence of perturbative unitarity , that is the physical requirement that , at a given loop - order , the bad unitarity behaviour , which potentially affects the scattering amplitudes in massive yang - mills theories , disappears thanks to a cancellation among the dangerous contributions . indeed ,
in the first part of this note we derive the sum rules by discussing the one - loop renormalization of the @xmath2 scattering , and by imposing unitarity to the amplitude obtained .
it has been known since a long time that there is a deep relation between unitarity and gauge invariance @xcite and in the second part of this work we re - derive the sum rules starting from the ward identities of a spontaneously broken @xmath8 theory .
although the sum rules presented here are valid only in the limit of very large energies , we will see how they qualitatively constrain also the region of experimental interest , namely the one between the electroweak scale @xmath9 and the scale @xmath10 associated to new physics .
an important feature of the sum rules is that they involve combinations of the three - point and the two - point functions among vector bosons .
this emphasizes the importance of including the vector boson self - energies for a correct evaluation of the form factors which parametrize the @xmath2 process .
use of form factors explicitly involving only the tgv may easily lead to wrong conclusion about the corresponding cross - section .
we will discuss these issues in a concluding paragraph .
.5 cm * 2 . *
we start the discussion by summarizing the standard parametrization of the @xmath1 helicity amplitudes @xcite to which we will refer in the following .
we denote with @xmath11 and @xmath12 the helicities of the electron ( positron ) and of the @xmath13 @xmath14 , with @xmath15 the scattering angle of the @xmath13 with respect to the @xmath16 direction in the @xmath4 center of mass frame .
the polarization amplitude reads : _ ; ( ) = e^2 _ ; ( ) d^j_0 _ , ( ) [ a1 ] where @xmath17 , @xmath18 , @xmath19 , @xmath20 and @xmath21 are the @xmath22 functions @xcite .
when @xmath23 , the reduced amplitude @xmath24 contains only the contribution from the neutrino exchange and is given by : = - _ , -1 [ a2 ] here @xmath25 is the @xmath3 velocity and @xmath26 is the weak mixing angle .
for @xmath27 , the amplitude is a sum of three contributions : = + + [ a3 ] where & = & -_||,1 + & = & + & = & [ a4 ] the coefficients @xmath28 , @xmath29 , @xmath30 and @xmath31 contain the dependence on the tgv .
the terms @xmath28 , @xmath30 and @xmath31 represent the sm , tree - level contribution and they are explicitly listed in table 1 .
[ cols="^,^,^,^ " , ] this should sound as a warning against the assumption that the amplitudes for @xmath2 are dominated by the irreducible tgv corrections .
.5 cm * 4 . * the parametrization given in eqs .
( [ a1]-[a6 ] ) is quite efficient and widely used in the literature . indeed , as will be shown in this section , it is possible to cast the results of our one - loop computation in a form which is very close to that displayed in eqs .
( [ a1]-[a6 ] ) . by inspecting eqs .
( [ a10]-[a11 ] ) , we are led to introduce an effective , @xmath32-dependent , weinberg angle defined by : ^2_eff(s)=^2| ( 1+k(s ) ) [ a30 ] then we proceed by absorbing the overall correction to the neutrino - exchange amplitude in a redefinition of the electric charge : e_eff^2(s)=e^2 ( 1+k(s)- r_w - e_6 ) [ a31 ] when expressed in terms of @xmath33 and @xmath34 , the amplitudes of eqs .
( [ a1],[a11],[a12 ] ) will coincide with those of eqs .
( [ a1],[a2],[a4 ] ) , provided one defines new coefficients @xmath29 , which , to avoid confusion , we will denote by @xmath35 : a^_(s)&= & a^_(s)+ a^ _ ( r_w+e_6 + ( s)-k(s ) ) + a^z_(s)&= & a^z_(s)+ a^z _ ( ( s)- k(s ) + r_w+e_6 ) [ a32 ] if we express the quantities @xmath36 , @xmath37 , @xmath38 , @xmath39 and @xmath40 in terms of unrenormalized self - energy corrections , as in eqs .
( [ a13 ] ) and ( [ a14 ] ) , and if we relate the coefficients @xmath29 to the unrenormalized vertex corrections , we find that eqs .
( [ a32 ] ) are equivalent to the following definition of form factors : f^v_1(s)&=&f^v_1(s)-^v(s ) + f^v_2(s)&=&f^v_2(s ) + f^v_3(s)&=&f^v_3(s)-2 ^v(s ) + f^v_5(s)&=&f^v_5(s ) [ a33 ] where the functions @xmath41 are explicitly given by : ^(s)&=&_(s)+ + ^z(s)&=&_zz(s)+ [ a34]to summarize , by replacing the electric charge @xmath42 , the weinberg angle @xmath26 and the form factors @xmath43 of eqs .
( [ a1]-[a6 ] ) with the quantities @xmath44 , @xmath45 , @xmath46 introduced above , one reproduces exactly the one - loop amplitude discussed in the present paper .
in fact , eqs .
( [ a33]-[a34 ] ) provide a concise and practical prescription to account for the self - energy effects in @xmath47 , by including them in appropriately defined form factors .
moreover , it is immediate to show that the sum rules discussed before , when expressed in terms of the form factors @xmath46 , read : _ = _ = _ = 0 .
[ a35 ] when discussing the low - energy limit of the process , an effective lagrangian description turns out to be useful .
the departures from the tree - level sm predictions are described by a set of operators ( organized in a dimensional or derivative expansion ) whose coefficients can be related to physical observables .
in particular the low - energy limit of the form factors @xmath48 can be expressed in terms of the coefficients @xmath49 characterizing the so - called electroweak chiral lagrangian @xcite : f^_1(0)&=&g^2 ( a_1-a_8 ) + f^_2(0)&=&0 + f^_3(0)&=&g^2 ( a_1-a_8+a_2-a_3-a_9 ) + f^_5(0)&=&0 + f^z_1(0)&=&-g^2(a_8+a_13)-g^2 ^ 2|(a_1+a_13 ) - a_3 + f^z_2(0)&=&0 + f^z_3(0)&=&-g^2(a_8+a_3+a_9+a_13)-g^2 ^ 2|(a_1+a_2+a_13 ) - a_3 + f^z_5(0)&=&- a_14we notice that the combinations @xmath50 depend only on the coefficients @xmath51 , @xmath52 and @xmath53 which parametrize the directions that are blind to the lep1 precision tests @xcite . 0.5truecm * 5 . * in conclusion , we have shown that , at asymptotically large energies , the trilinear gauge boson vertices in @xmath2 amplitudes obey a set of sum rules connecting them to gauge vector boson self - energies .
we have derived the sum rules by performing an explicit one - loop computation in a certain class of theories and by demanding that the resulting amplitudes satisfy the requirement of perturbative unitarity .
we have also demonstrated that these sum rules are a direct consequence of the ward identities of spontaneously broken @xmath0 gauge theories and therefore they are verified in a more general context than the one considered in the first derivation .
our discussion shows that , in general , it is never possible to neglect vector boson self - energies when computing the form factors which parametrize the @xmath2 helicity amplitudes .
the exclusion of the self - energy contributions would lead to estimates of the effects which are wrong by orders of magnitude , as we have explicitly discussed in some examples .
finally we have shown how the self - energy effects can be suitably included in the formalism by a simple redefinition of the form factors which allow to use the existing parametrizations of the considered process . |
adequate models of human language for syntactic analysis and semantic interpretation are typically of context - free complexity or beyond .
indeed , prolog - style definite clause grammars ( dcgs ) and formalisms such as patr with feature - structures and unification have the power of turing machines to recognise arbitrary recursively enumerable sets .
since recognition and analysis using such models may be computationally expensive , for applications such as speech processing in which speed is important finite - state models are often preferred .
when natural language processing and speech recognition are integrated into a single system one may have the situation of a finite - state language model being used to guide speech recognition while a unification - based formalism is used for subsequent processing of the same sentences .
rather than write these two grammars separately , which is likely to lead to problems in maintaining consistency , it would be preferable to derive the finite - state grammar automatically from the ( unification - based ) analysis grammar .
the finite - state grammar derived in this way can not in general recognise the same language as the more powerful grammar used for analysis , but , since it is being used as a front - end or filter , one would like it not to reject any string that is accepted by the analysis grammar , so we are primarily interested in ` sound approximations ' or ` approximations from above ' .
attention is restricted here to approximations of context - free grammars because context - free languages are the smallest class of formal language that can realistically be applied to the analysis of natural language .
techniques such as restriction @xcite can be used to construct context - free approximations of many unification - based formalisms , so techniques for constructing finite - state approximations of context - free grammars can then be applied to these formalisms too .
a ` finite - state calculus ' or ` finite automata toolkit ' is a set of programs for manipulating finite - state automata and the regular languages and transducers that they describe .
standard operations include intersection , union , difference , determinisation and minimisation .
recently a number of automata toolkits have been made publicly available , such as fire lite @xcite , grail @xcite , and fsa utilities @xcite .
finite - state calculus has been successfully applied both to morphology @xcite and to syntax ( constraint grammar , finite - state syntax ) .
the work described here used a finite - state calculus implemented by the author in sicstus prolog .
the use of prolog rather than c or c++ causes large overheads in the memory and time required .
however , careful account has been taken of the way prolog operates , its indexing in particular , in order to ensure that the asymptotic complexity is as good as that of the best published algorithms , with the result that for large problems the prolog implementation outperforms some of the publicly available implementations in c++ .
some versions of the calculus allow transitions to be labelled with arbitrary prolog terms , including variables , a feature that proved to be very convenient for prototyping although it does not essentially alter the power of the machinery .
( it is assumed that the string being tested consists of ground terms so no unification is performed , just matching . )
there are two main ideas behind this algorithm .
the first is to describe the finite - state approximation using formulae with regular languages and finite - state operations and to evaluate the formulae directly using the finite - state calculus .
the second is to use , in intermediate stages of the calculation , additional , auxiliary symbols which do not appear in the final result .
a similar approach has been used for compiling a two - level formalism for morphology @xcite . in this case the auxiliary symbols are dotted rules from the given context - free grammar .
a dotted rule is a grammar rule with a dot inserted somewhere on the right - hand side , e.g. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ s @xmath0 @xmath1 np vp + s @xmath0 np @xmath1 vp + s @xmath0 np vp @xmath1 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ however , since these dotted rules are to be used as terminal symbols of a regular language , it is convenient to use a more compact notation : they can be replaced by a triple made out of the nonterminal symbol on the left - hand side , an integer to determine one of the productions for that nonterminal , and an integer to denote the position of the dot on the right - hand side by counting the number of symbols to the left of the dot .
so , if ` s @xmath0 np vp ' is the fourth production for s , the dotted rules given above may be denoted by @xmath2 , @xmath3 and @xmath4 , respectively .
it will turn out to be convenient to use a slightly more complicated notation : when the dot is located after the last symbol on the right - hand side we use @xmath5 as the third element of the triple instead of the corresponding integer , so the last triple is @xmath6 instead of @xmath4 .
( note that @xmath5 is an additional symbol , not a variable . ) moreover , for epsilon - rules , where there are no symbols on the right - hand side , we treat the @xmath7 as it were a real symbol and consider there to be two corresponding dotted rules , e.g. @xmath8 and @xmath9 corresponding to ` mod @xmath0 @xmath1 @xmath7 ' and ` mod @xmath0 @xmath7 @xmath1 ' for the rule ` mod @xmath0 @xmath7 ' . using these dotted rules as auxiliary symbols we can work with regular languages over the alphabet @xmath10 where @xmath11 is the set of terminal symbols
, @xmath12 is the set of nonterminals , @xmath13 is the number of productions for nonterminal @xmath14 , and @xmath15 is the number of symbols on the right - hand side of the @xmath16th production for @xmath14 .
it will be convenient to use the symbol @xmath17 as a ` wildcard ' , so @xmath18 means @xmath19 and @xmath20 means @xmath21 .
( this last example explains why we use @xmath5 rather than @xmath15 ; it would otherwise not be possible to use the ` wildcard ' notation to denote concisely the set @xmath22 . )
we can now attempt to derive an expression for the set of strings over @xmath23 that represent a valid parse tree for the given grammar : the tree is traversed in a top - down left - to - right fashion and the daughters of a node x expanded with the @xmath16th production for x are separated by the symbols @xmath24 .
( equivalently , one can imagine the auxiliary symbols inserted in the appropriate places in the right - hand side of each production so that the grammar is then unambiguous . )
consider , for example , the following grammar : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ s @xmath0 a s b + s @xmath0 @xmath7 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
then the following is one of the strings over @xmath23 that we would like to accept , corresponding to the string @xmath25 accepted by the grammar : @xmath26 our first approximation to the set of acceptable strings is @xmath27 , i.e. strings that start with beginning to parse an s and end with having parsed an s. from this initial approximation we subtract ( that is , we intersect with the complement of ) a series of expressions representing restrictions on the set of acceptable strings : and @xmath28 , respectively , while juxtaposition denotes concatenation and the bar denotes complementation ( @xmath29 ) . ]
@xmath30 formula [ e1 ] expresses the restriction that a dotted rule of the form @xmath31 , which represents starting to parse the right - hand side of a rule , may be preceded only by nothing ( the start of the string ) or by a dotted rule that is not of the form @xmath32 ( which would represent the end of parsing the right - hand side of a rule ) .
@xmath33 formula [ e2 ] similarly expresses the restriction that a dotted rule of the form @xmath32 may be followed only by nothing or by a dotted rule that is not of the form @xmath31 . for each non - epsilon - rule with dotted rules
@xmath34 , @xmath35 , for each @xmath36 : @xmath37 where @xmath38 where @xmath39 is the @xmath40th symbol on the right - hand side of the @xmath16th production for @xmath14 .
formula [ e3 ] states that the dotted rule @xmath34 must be followed by @xmath41 ( or @xmath42 when @xmath43 ) when the next item to be parsed is the terminal @xmath44 , or by @xmath45 ( starting to parse an @xmath46 ) when the next item is the nonterminal @xmath46 .
for each non - epsilon - rule with dotted rules @xmath34 , @xmath35 , for each @xmath47 : @xmath48 where @xmath49 + @xmath50 = @xmath51 @xmath52 + @xmath50 @xmath53 + @xmath54 @xmath55 formula [ e4 ] similarly states that the dotted rule @xmath34 must be preceded by @xmath51 ( or @xmath56 when @xmath57 ) when the previous item was the terminal @xmath44 , or by @xmath54 when the previous item was the nonterminal @xmath46 . for each epsilon - rule corresponding to dotted rules @xmath58 and @xmath59 : @xmath60 @xmath61 formulae [ e5 ] and [ e6 ] state that the dotted rule @xmath58 must be followed by @xmath59 , and @xmath59 must be preceded by @xmath58 . for each non - epsilon rule with dotted rules
@xmath34 , @xmath35 , for each @xmath36 : @xmath62 and @xmath63 where @xmath64 formula [ e7 ] states that the next instance of @xmath24 that follows @xmath34 must be either @xmath58 ( a recursive application of the same rule ) or @xmath65 ( the next stage in parsing the same rule ) , and there must be such an instance .
formula [ e8 ] states similarly that the closest instance of @xmath24 that precedes @xmath65 must be either @xmath59 ( a recursive application of the same rule ) or @xmath34 ( the previous stage in parsing the same rule ) , and there must be such an instance .
when each of these sets has been subtracted from the initial approximation we can remove the auxiliary symbols ( by applying the regular operator that replaces them with @xmath7 ) to give the final finite - state approximation to the context - free grammar .
it may be admitted that the notation used for the dotted rules was partly motivated by the possibility of immediately testing the algorithm using the finite - state calculus in prolog : the regular expressions listed above can be evaluated directly using the ` wildcard ' capabilities of the finite - state calculus .
figure [ codefig ] shows the sequence of calculations that corresponds to applying the algorithm to the following grammar : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ s @xmath0 a s b + s @xmath0 @xmath7 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ with the following notational explanations it should be possible to understand the code and compare it with the description of the algorithm . * the procedure ` r(re , x ) ` evaluates the regular expression ` re ` and puts the resulting ( minimised ) automaton into a register with the name ` x ` . * ` list_fsa(x ) ` prints out the transition table for the automaton in register ` x ` .
* terminal symbols may be any prolog terms , so the terminal alphabet is implicit . here atoms are used for the terminal symbols of the grammar ( ` a ` and ` b ` ) and terms of the form ` _ /_/
_ ` are used for the triples representing dotted rules .
the terms need not be ground , so the prolog variable symbol ` _ ` is used instead of the ` wildcard ' symbol @xmath17 in the description of the algorithm . * in a regular expression : * * ` # x ` refers to the contents of register ` x ` ; * * ` $ ` represents @xmath23 , any single terminal symbol ; * * ` s ` represents a string of terminals with length equal to the number of arguments ; so ` s ` with no arguments represents the empty string @xmath7 , ` s(a ) ` represents the single terminal @xmath44 , and ` s(s/_/0 ) ` represents the dotted rules @xmath66 ; * * kleene star is ` * ` ( redefined as a postfix operator ) , and concatenation and union are ` ^ ` and ` + ` , respectively ; * * other operators provided include ` & ` ( intersection ) and ` - ` ( difference ) ; there is no operator for complementation ; instead subtraction from @xmath67 may be used , e.g. ` ( $ * ) -(#l ) ` instead of @xmath68 ; * * ` rem(re , l ) ` denotes the result of removing from the language ` re ` all terminals that match one of the expressions in the list ` l ` .
the context - free language recognised by the original context - free grammar is @xmath69 .
the result of applying the approximation algorithm is a 3-state automaton recognising the language @xmath70 .
applying the restrictions expressed by formulae [ e1][e6 ] gives an automaton whose size is at most a small constant multiple of the size of the input grammar .
this is because these restrictions apply locally : the state that the automaton is in after reading a dotted rule is a function of that dotted rule .
when restrictions [ e7][e8 ] are applied the final automaton may have size exponential in the size of the input grammar .
for example , exponential behaviour is exhibited by the following class of grammars : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ s @xmath0 a@xmath71 s
a@xmath71 + @xmath72 + s @xmath0 a@xmath73 s a@xmath73 + s @xmath0 @xmath7 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ here the final automaton has @xmath74 states .
( it records , in effect , one of three possibilities for each terminal symbol : whether it has not yet appeared , has appeared and must appear again , or has appeared and need not appear again . ) there is an important computational improvement that can be made to the algorithm as described above : instead of removing all the auxiliary symbols right at the end they can be removed progressively as soon as they are no longer required ; after formulae [ e7][e8 ] have been applied for each non - epsilon rule with dotted rules @xmath24 , those dotted rules may be removed from the finite - state language ( which typically makes the automaton smaller ) ; and the dotted rules corresponding to an epsilon production may be removed before formulae [ e7][e8 ] are applied .
( to ` remove ' a symbol means to substitute it by @xmath7 : a regular operation . ) with this important improvement the algorithm gives exact approximations for the left - linear grammars _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ s @xmath0 s a@xmath71 + @xmath72 + s @xmath0 s a@xmath73 + s @xmath0 @xmath7 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ and the right - linear grammars _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ s @xmath0 a@xmath71 s + @xmath72 + s @xmath0 a@xmath73 s + s @xmath0 @xmath7 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ in space bounded by @xmath40 and time bounded by @xmath75 .
( it is easiest to test this empirically with an implementation , though it is also possible to check the calculations by hand . )
pereira and wright s algorithm gives an intermediate unfolded recogniser of size exponential in @xmath40 for these right - linear grammars .
there are , however , both left - linear and right - linear grammars for which the number of states in the final automaton is not bounded by any polynomial function of the size of the grammar .
an examples is : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ s @xmath0 a@xmath71 s s @xmath0 a@xmath71 a@xmath71 + @xmath72 + s @xmath0 a@xmath73 s s @xmath0 a@xmath73 a@xmath73 + a@xmath71 @xmath0 a@xmath76 a@xmath71 @xmath72a@xmath71 @xmath0 a@xmath73 a@xmath71 + a@xmath76
@xmath0 a@xmath71 a@xmath76 @xmath72a@xmath76 @xmath0 a@xmath73 a@xmath76 + @xmath72 + a@xmath73 @xmath0 a@xmath71 a@xmath73 a@xmath73 @xmath0 a@xmath76 a@xmath73 @xmath72 + x @xmath0 @xmath7 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ here the grammar has size @xmath77 and the final approximation has @xmath78 states .
pereira and wright point out in the context of their algorithm that a grammar may be decomposed into ` strongly connected ' subgrammars , each of which may be approximated separately and the results composed .
the same method can be used with the finite - state calculus approach : define the relation @xmath79 over nonterminals of the grammar s.t .
@xmath80 iff @xmath81 appears on the right - hand side of a production for @xmath46 .
then the relation @xmath82 , the reflexive transitive closure of @xmath79 intersected with its inverse , is an equivalence relation .
a subgrammar consists of all the productions for nonterminals in one of the equivalence classes of @xmath83 .
calculate the approximations for each nonterminal by treating the nonterminals that belong to other equivalence classes as if they were terminals .
finally , combine the results from each subgrammar by starting with the approximation for the start symbol @xmath84 and substituting the approximations from the other subgrammars in an order consistent with the partial ordering that is induced by @xmath79 on the subgrammars .
when the algorithm was applied to the 18-rule grammar shown in figure [ gramfig ] it was not possible to complete the calculations for any ordering of the rules , even with the improvement mentioned in the previous section , as the automata became too large for the finite - state calculus on the computer that was being used .
( note that the grammar forms a single strongly connected component . )
= nom @xmath0 nom mod bla= mod @xmath0 vp @xmath0 v np + mod @xmath0 p np vp @xmath0 v s + vp @xmath0 v vp + nom @xmath0 a nom vp @xmath0 v + nom @xmath0 n vp @xmath0 vp c vp + nom @xmath0 nom mod vp @xmath0 vp mod + nom @xmath0 nom s s @xmath0 mod s + s @xmath0 np s + np @xmath0 s @xmath0 s c s + np @xmath0 d nom s @xmath0 v np vp + .... % initial approximation : r ( s(s/_/0)^($ * ) ^s(s/_/z ) , a ) .
% formulae ( 1)-(2 ) : r ( ( # a ) - ( ( $ * ) -(($ * ) ^(s(_/_/_)-s(_/_/z))+s))^s(_/_/0)^($ * ) , a ) .
r ( ( # a ) - ( $ * ) ^s(_/_/z)^(($ * ) -(s+(s(_/_/_)-s(_/_/0))^($ * ) ) ) , a ) .
% formula ( 3 ) for " s - > a s b " : r ( ( # a ) - ( $ * ) ^s(s/1/0)^(($ * ) -s(a)^s(s/1/1)^($ * ) ) , a ) .
r ( ( # a ) - ( $ * ) ^s(s/1/1)^(($ * ) -s(s/_/0)^($ * ) ) , a ) .
r ( ( # a ) - ( $ * ) ^s(s/1/2)^(($ * ) -s(b)^s(s/1/z)^($ * ) ) , a ) .
% formula ( 4 ) for " s - > a s b " : r ( ( # a ) - ( ( $ * ) -($ * ) ^s(s/1/0)^s(a))^s(s/1/1)^($ * ) , a ) .
r ( ( # a ) - ( ( $ * ) -($ * ) ^s(s/_/z))^s(vp/2/1)^($ * ) , a ) .
r ( ( # a ) - ( ( $ * ) -($ * ) ^s(s/1/2)^s(b))^s(s/1/z)^($ * ) , a ) .
% formulae ( 5)-(6 ) for " s - > " : r ( ( # a ) - ( $ * ) ^s(s/2/0)^(($ * ) -s(s/2/z)^($ * ) ) , a ) .
r ( ( # a ) - ( ( $ * ) -($ * ) ^s(s/2/0))^s(s/2/z)^($ * ) , a ) . % formula ( 7 ) for " s - > a s b " : r((#a)-($ * ) ^s(s/1/0)^(($ * ) -(($ -s(s/1/_))*)^(s(s/1/0)+s(s/1/1))^($ * ) ) , a ) .
r((#a)-($ * ) ^s(s/1/1)^(($ * ) -(($ -s(s/1/_))*)^(s(s/1/0)+s(s/1/2))^($ * ) ) , a ) .
r((#a)-($ * ) ^s(s/1/2)^(($ * ) -(($ -s(s/1/_))*)^(s(s/1/0)+s(s/1/z))^($ * ) ) , a ) .
% formula ( 8) for " s - > a s b " : r((#a)-(($ * ) -($ * ) ^(s(s/1/z)+s(s/1/0))^(($ -s(s/1/_))*))^s(s/1/1)^($ * ) , a ) . r((#a)-(($ * ) -($ * ) ^(s(s/1/z)+s(s/1/1))^(($ -s(s/1/_))*))^s(s/1/2)^($ * ) , a ) . r((#a)-(($ * ) -($ * ) ^(s(s/1/z)+s(s/1/2))^(($ -s(s/1/_))*))^s(s/1/z)^($ * ) , a ) .
% define the terminal alphabet : r ( s(s/1/0)+s(s/1/1)+s(s/1/2)+s(s/1/z)+s(s/2/0)+s(s/2/z)+s(a)+s(b ) , sigma ) .
% remove the auxiliary symbols to give final result : r ( rem((#a)&((#sigma ) * ) , [ _ /_/ _ ] ) , f ) .
list_fsa(f ) . .... however , it was found possible to simplify the calculation by omitting the application of formulae [ e7][e8 ] for some of the rules . ( the auxiliary symbols not involved in those rules could then be removed before the application of [ e7][e8 ] . ) in particular , when restrictions [ e7][e8 ] were applied only for the s and vp rules the calculations could be completed relatively quickly , as the largest intermediate automaton had only 406 states . yet the final result was still a useful approximation with 16 states .
pereira and wright s algorithm applied to the same problem gave an intermediate automaton ( the ` unfolded recogniser ' ) with 56272 states , and the final result ( after flattening and minimisation ) was a finite - state approximation with 13 states .
the two approximations are shown for comparison in figure [ prettyfig ] .
each has the property that the symbols ` d ` , ` a ` and `
n ` occur only in the combination ` d ` ` a`@xmath85 `
this fact has been used to simplify the state diagrams by treating this combination as a single terminal symbol ` dan ` ; hence the approximations are drawn with 10 and 9 states , respectively .
neither of the approximations is better than the other ; their intersection ( with 31 states ) is a better approximation than either .
the two approximations have therefore captured different aspects of the context - free language .
in general it appears that the approximations produced by the present algorithm tend to respect the necessity for certain constituents to be present , at whatever point in the string the symbols that ` trigger ' them appear , without necessarily insisting on their order , while pereira and wright s approximation tends to take greater account of the constituents whose appearance is triggered early on in the string : most of the complexity in pereira and wright s approximation of the 18-rule grammar is concerned with what is possible before the first accepting state is encountered .
rimon and herz approximate the recognition capacity of a context - free grammar by extracting ` local syntactic constraints ' in the form of the left or right short context of length @xmath40 of a terminal .
when @xmath86 this reduces to next(t ) , the set of terminals that may follow the terminal t. the effect of filtering with rimon and herz s next(t ) is similar to applying conditions [ e1][e6 ] from section [ algsect ] , but the use of auxiliary symbols causes two differences which can both be illustrated with the following grammar : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ s @xmath0 a x a @xmath87 b x b + x @xmath0 @xmath7 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ on the one hand , rimon and herz s ` next ' does not distinguish between different instances of the same terminal symbol , so any @xmath44 , and not just the first one , may be followed by another @xmath44 . on the other hand , rimon and herz s ` next '
looks beyond the empty constituent in a way that conditions [ e1][e6 ] do not , so @xmath88 is disallowed .
thus an approximation based on rimon and herz s ` next ' would be @xmath89 , and an approximation based on conditions [ e1][e6 ] would be @xmath90 .
( however , the approximation becomes exact when conditions [ e7][e8 ] are added . )
both pereira and wright and rood start with the lr(0 ) characteristic machine , which they first ` unfold ' ( with respect to ` stacks ' or ` paths ' , respectively ) and then ` flatten ' .
the characteristic machine is defined in terms of dotted rules with transitions between them that are analagous to the conditions implied by formula [ e3 ] of section [ algsect ] .
when the machine is flattened , @xmath7-transitions are added in a way that is in effect simulated by conditions [ e2 ] and [ e4 ] .
( condition [ e1 ] turns out to be implied by conditions [ e2][e4 ] . )
it can be shown that the approximation @xmath91 obtained by flattening the characteristic machine ( without unfolding it ) is as good as the approximation @xmath92 obtained by applying conditions [ e1][e6 ] ( @xmath93 ) . moreover , if no nonterminal for which there is an @xmath7-production is used more than once in the grammar , then @xmath94 .
( the grammar in figure [ gramfig ] is an example for which @xmath95 ; the approximation found in section [ ressect ] includes strings such as ` vvccvv ` which are not accepted by @xmath91 for this grammar . )
it can also be shown that @xmath92 is the same as the result of flattening the characteristic machine for the same grammar modifed so as to fulfil the afore - mentioned condition by replacing the right - hand side of every @xmath7-production with a new nonterminal for which there is a single @xmath7-production .
however , there does not seem to be a simple correspondence between conditions [ e7][e8 ] and the ` unfolding ' used by pereira and wright or rood : even some simple grammars such as ` s @xmath0 a s a @xmath87 b s b @xmath87 @xmath7 ' are approximated differently by [ e1][e8 ] than by pereira and wright s and rood s methods .
in the case of some simple examples ( such as the grammar ` s @xmath0 a s b @xmath87 @xmath7 ' used earlier ) the approximation algorithm presented in this paper gives the same result as pereira and wright s algorithm .
however , in many other cases ( such as the grammar ` s @xmath0 a s a @xmath87 b s b @xmath87 @xmath7 ' or the 18-rule grammar in the previous section ) the results are essentially different and neither of the approximations is better than the other .
the new algorithm does not share the problem of pereira and wright s algorithm that certain right - linear grammars give an intermediate automaton of exponential size , and it was possible to calculate a useful approximation fairly rapidly in the case of the 18-rule grammar in the previous section .
however , it is not yet possible to draw general conclusions about the relative efficiency of the two procedures .
nevertheless , the new algorithm seems to have the advantage of being open - ended and adaptable : in the previous section it was possible to complete a difficult calculation by relaxing the conditions of formulae [ e7][e8 ] , and it is easy to see how those conditions might also be strengthened .
for example , a more complicated version of formulae [ e7][e8 ] might check two levels of recursive application of the same rule rather than just one level and it might be useful to generalise this to @xmath40 levels of recursion in a manner analagous to rood s generalisation of pereira and wright s algorithm .
grimley evans , edmund , george kiraz , and stephen pulman . 1996 . compiling a partition - based two - level formalism .
coling-96 , 454459 .
herz , jacky , and mori rimon .
1991 . local syntactic constraints .
second international workshop on parsing technology ( iwpt-2 ) .
kaplan , ronald , and martin kay .
regular models of phonological rule systems .
_ computational linguistics _ , * 20(3 )
* : 33178 .
kempe , and , and lauri karttunen .
parallel replacement in finite state calculus .
coling-96 , 622 .
pereira , fernando , and rebecca wright .
finite - state approximation of phrase structure grammars .
proceedings of the 29th annual meeting of the association for computational linguistics , 246255 .
pereira , fernando , and rebecca wright .
finite - state approximation of phrase - structure grammars .
cmp - lg/9603002 .
raymond , darrell , and derick wood .
march 1996 . the grail papers .
university of western ontario , department of computer science , technical report tr-491 .
rimon , mori , and jacky herz .
the recognition capacity of local syntactic constraints .
acl proceedings , 5th european meeting .
rood , cathy .
efficient finite - state approximation of context free grammars .
proceedings of ecai 96 .
shieber , stuart .
1985 . using restriction to extend parsing algorithms for complex - feature - based formalisms .
proceedings of the 23nd annual meeting of the association for computational linguistics , 145152 .
van noord , gertjan .
fsa utilities : manipulation of finite - state automata implemented in prolog . first international workshop on implementing automata , university of western ontario , london ontario , 2931 august 1996 .
watson , bruce .
1996 . implementing and using finite automata toolkits .
proceedings of ecai 96 . | although adequate models of human language for syntactic analysis and semantic interpretation are of at least context - free complexity , for applications such as speech processing in which speed is important finite - state models are often preferred .
these requirements may be reconciled by using the more complex grammar to automatically derive a finite - state approximation which can then be used as a filter to guide speech recognition or to reject many hypotheses at an early stage of processing .
a method is presented here for calculating such finite - state approximations from context - free grammars .
it is essentially different from the algorithm introduced by pereira and wright , is faster in some cases , and has the advantage of being open - ended and adaptable . |
the possible existence of the fulde - ferrell - larkin - ovchinnikov ( fflo ) superfluid phases @xcite has been pointed out in the early sixties and it has given rise to much work ever since that time .
in addition to their intrinsic fundamental interest @xcite , these phases are quite relevant experimentally since they are expected to arise in superconductors with very high critical fields , which are naturally very actively searched for . on several occasions these phases have been claimed to be observed experimentally , but to date these hopes have not been firmly substantiated .
very recently anomalies in the heavy fermion compound cecoin@xmath1 have been attributed to fflo phases @xcite .
the case of two - dimensional ( 2d ) systems is of particular interest @xcite since they are experimentally quite relevant .
indeed a major strategy to observe these transitions is to eliminate orbital currents , which are responsible for the low critical fields in standard superconductors .
this can be achieved in quasi two- dimensional systems , made of widely separated conducting planes , such as organic compounds or high @xmath2 cuprate superconductors . in this case
hopping between planes is very severely restricted .
hence the orbital currents perpendicular to the planes are very weak when a strong magnetic field is applied parallel to the planes , and there is essentially no orbital pair breaking effect which opens the path to fflo phases at much higher fields . indeed experimental results in organic compounds have been claimed quite recently @xcite to be compatible with the existence of fflo phases .
naturally when the magnetic field is not exactly parallel to the planes , one finds in addition vortex - like structures and the physical situation gets even more complex @xcite .
the fflo transition in 2d systems is believed to be second order and in particular burkhardt and rainer @xcite have studied in details the transition to a planar phase , where the order parameter @xmath3 is a simple @xmath4 at the transition .
this phase has been found by larkin and ovchinnikov @xcite to be the best one in 3d at @xmath5 for a second order phase transition . and
in 3d it is also found to be the preferred one in the vicinity of the tricritical point and below @xcite , although in this case the transition turns out to be first order ( except at very low temperature ) .
however it is not clear that this is always the case since , as first explored by larkin and ovchinnikov , this order parameter is in competition with any superposition of plane waves , provided that their wavevectors have all the same modulus .
indeed we have shown very recently @xcite that , at low temperature , the transition is rather a first order one , toward an order parameter with a more complex structure .
for example at @xmath5 it is very near the linear combination of three cosines oscillating in orthogonal directions .
in this paper we explore the low temperature range in 2d and show that the second order transition is indeed toward rather more complex order parameters . a short report of our results
has already been published @xcite . a first step in this direction
is found in the recent work of shimahara @xcite who found a transition toward a superposition of three cosines .
here we show that , when the temperature is lowered toward @xmath5 , one obtains a cascade of transitions toward order parameters with an ever increasing number of plane waves .
the @xmath5 limit is singular in this respect .
this is actually clear from the beginning
. indeed if one looks at the second order term in the expansion of the free energy in powers of the order parameter , which gives the location of the fflo transition , one finds it to be a singular function of the plane wave wavevector .
this is recalled in the next section .
then we calculate the fourth order term in the free energy expansion and show that the phases which are selected by this term display the cascade of transitions mentionned above .
the general expression for the free energy difference @xmath6 between the superconducting and the normal state can be obtained in a number of ways , starting for example @xcite from eilenberger s expression in terms of the quasiclassical green s function or from the gap equation @xcite and gorkov s equations . when the result is expanded up to fourth order term in powers of the fourier components @xmath7 of the order parameter @xmath8 : @xmath9 one obtains : @xmath10 where we have momentum conservation @xmath11 in the fourth order term and @xmath12 is the single spin density of states at the fermi surface .
the explicit expression of @xmath13 in terms of the standard bcs interaction @xmath14 and of the free fermions propagator is : @xmath15 where @xmath16 and @xmath17 and @xmath18 , with @xmath19 being half the chemical potential difference between the two fermionic populations forming pairs , @xmath20 the kinetic energy measured from the fermi surface for @xmath21 and @xmath22 the matsubara frequency . performing the @xmath23 integration and the 2d angular average over @xmath24 gives : @xmath25 where we have introduced the dimensionless wavevector @xmath26 . in eq.(4 )
the summation has to be cut - off at a frequency @xmath27 in the standard bcs way .
it is more convenient to rewrite @xmath28 , by introducing physical quantities related to the @xmath29 case , as : @xmath30 with : @xmath31 we have introduced : @xmath32 which is zero on the spinodal transition line ( the line in the @xmath33 plane where the normal state becomes absolutely unstable against a transition toward a space independent order parameter ) and is positive above it . at the fflo transition we are looking at , we have @xmath34 .
the actual transition corresponds to the largest possible @xmath35 at fixed @xmath36 . from eq.(7 ) this corresponds to have the largest @xmath37 . hence from eq.(5 ) we want to minimize @xmath38 with respect to @xmath39 . at low temperature
it is more convenient to express @xmath40 as : @xmath41 \label{eq8}\end{aligned}\ ] ] where the integration contour runs actually infinitesimally above the real @xmath42 axis . at @xmath5 the integration
is easily performed to give : @xmath43 the minimum is reached for @xmath44 , in agreement with shimahara @xcite and burkhardt and rainer @xcite , and @xmath45 at this minimum . in this case , from eq.(7 ) , @xmath46 where @xmath47 is the zero temperature bcs phase gap ( this corresponds to the value @xmath48 for the spinodal transition ) .
this leads to @xmath49 for the location of the fflo transition , again in agreement with previous work .
it is worth to note that , as already mentionned in the introduction , the location @xmath50 of the minimum corresponds to a singular point for @xmath51 since we have explicitely @xmath52 for @xmath53 . while @xmath51 itself is continuous , its derivative is discontinuous for @xmath50 . for @xmath54
there is no singular behaviour and we find the value of @xmath55 giving the minimum @xmath40 by writing that its derivative with respect to @xmath56 is zero . integrating the result by parts leads to the condition : @xmath57 where we have taken the new variable @xmath58 and defined the reduced temperature @xmath59 . only the ranges @xmath60 and @xmath61 contribute to the real part of the integral .
since at low @xmath36 we have @xmath62 , this last range will only give an exponentially small contribution because of the factor @xmath63 .
since this same factor makes @xmath64 to be at most of order 3 , we can make at low @xmath36 to leading order @xmath65 and @xmath66 .
it is then seen that we must have @xmath67 , because @xmath68 makes the left hand side of eq.(10 ) much larger than unity at low @xmath36 .
this implies @xmath69 and @xmath70 .
the integral is then easily evaluated and eq.(10 ) gives finally to leading order : @xmath71 note that this result is in disagreemeent with the analysis given @xcite by l. n. bulaevskii .
the reason for this discrepancy is discussed in details in appendix a. in particular we rederive in this appendix our eq .
( [ leading ] ) from the starting equation of ref .
@xcite .
a more complete low temperature expansion can be fairly easily extracted from eq.([leading ] ) .
as previously seen , the range @xmath72 in the integration is sufficient since the other integration range gives an exponentially small term . with the change of variable @xmath73 , eq.([leading ] ) leads to : @xmath74 where we have defined @xmath75 . neglecting terms of order @xmath76 eq .
( [ leading3 ] ) can be written as : @xmath77 from eq.([leading2 ] ) the leading order for @xmath78 is @xmath79 . in the second integral in eq.([devel
] ) we can replace @xmath78 by @xmath80 since @xmath81 .
setting @xmath82 , we obtain an expansion of @xmath83 in powers of @xmath84 by expanding the first integral up to second order in powers of @xmath85 : @xmath86 valid up to the order @xmath87 . solving this equation order by order , we find the following expansion for @xmath78 : @xmath88 this expression leads to a marked improvement when it is compared to the straight numerical evaluation of the matsubara sums in eq.(6 ) .
this is seen in fig.[fig0 ] where we have plotted the optimal @xmath89 from straight numerical calculation , as well as its leading low temperature approximation and the one resulting from the expansion eq.([alowtemp ] ) .
we give also for convenience , in the lower panel , the dependence of our reduced temperature @xmath90 as a function of the ratio @xmath91 , where @xmath92 is the standard bcs critical temperature , found for @xmath93 . in the low temperature limit @xmath94 , we have merely @xmath95 and @xmath49 , hence @xmath96 and the dotted straight line in the lower panel gives this limiting behaviour @xmath97 .
the second order term in the free energy gives the same transition line for all the combinations of plane waves .
the selection among the various possible order parameters will be made by the fourth order term in the free energy expansion , as pointed out by larkin and ovchinnikov .
the selected state will correspond to the lowest fourth order term .
so we turn now to this fourth order term .
its general expression is given by eq.(2 ) with @xcite : @xmath98 where we have used already momentum conservation @xmath99 .
since all the wavevectors @xmath100 have the same modulus given by eq.(11 ) , this momentum conservation implies in our 2d situation either @xcite @xmath101 together with @xmath102 , or equivalently @xmath103 together with @xmath104 .
moreover one has the additional possibility @xmath105 together with @xmath106 .
the two first possibilities lead to a same coefficient called @xmath107 by lo where @xmath108 is the angle between @xmath109 and @xmath110 .
similarly the last possibility leads to a coefficient @xmath111 .
explicitely the fourth order term in the free energy ( last term of the r.h.s . of eq.(2 ) ) becomes @xcite : @xmath112 with , after a change of variable in the summation : @xmath113 and @xmath114 when the @xmath115 integration is performed , taking into account @xmath116 since @xmath117 , one finds : @xmath118 where the bracket means the angular average , and similarly : @xmath119.[(2i\bar{\omega } _ { n})^{2}-({\bf k}.{\bf q } _ { 2}/m ) ^{2 } ] } > \label{eq17}\end{aligned}\ ] ] on which it is clear that @xmath120 .
let us first consider @xmath121 . as in the preceding section when going from eq.(6 ) to eq.(8 ) , it is more convenient to transform the sum over matsubara frequencies into an integral on the real frequency axis .
when one furthermore performs a by parts integration over the frequency @xmath42 , one gets from eq.([eq16 ] ) : @xmath122.[\bar{q } \cos ( \theta + \alpha /2)-(1 + \omega / \bar { \mu } ) ] } \label{eq18}\end{aligned}\ ] ] where as above @xmath123 and the integration contour runs again infinitesimally above the real @xmath42 axis . here in the angular integration
we have taken the reference axis bisecting the angle between @xmath109 and @xmath124 .
this angular integration is performed by residues , taking @xmath125 as a variable . for @xmath126
this leads to : @xmath127.[\cos ( \theta + \alpha /2)-y ] } = \frac{y}{(y ^{2}- \cos^{2 } ( \alpha /2 ) ) \sqrt { y ^{2}- 1 } } \label{eq19}\end{aligned}\ ] ] where the cut in the determination of the square root has to be taken on the positive real axis , as it is clear when one considers the case of very large @xmath128 .
when this result is inserted in eq.([eq18 ] ) and @xmath129 is taken as new variable , one finds , introducing again the reduced temperature @xmath59 : @xmath130 with @xmath131 . up to now
we have made no approximation in our calculation and the result is valid at any temperature .
let us now focus on the low temperature regime @xmath132 . because of the factor @xmath133 , only the vicinity of @xmath134 will contribute .
moreover only the half circle contour around the pole @xmath135 and the range @xmath136 contribute to the real part . in this last domain
we have @xmath137 from our result for @xmath138 eq.(11 ) , so we can again simplify the hyperbolic cosine into an exponential ( although this is not in practice a good approximation numerically ) . with the further change of variable @xmath139 in the resulting integral , we find to leading order : @xmath140 } - \frac{8}{\sqrt { \pi}(1 + \cos ( \alpha /2 ) ) } \int_{0 } ^ { \infty } \ ! dv \ : \frac{\exp(-v ^{2 } ) } { v ^{2}+\beta ^{2 } } \label{jalpha}\end{aligned}\ ] ] where we have substituted explicitely the result eq.(11 ) for @xmath138 and have set @xmath141 .
the integral in this result can not be further simplified in general and is related to parabolic cylinder functions .
let us consider now some important limiting cases for this result .
first we take at fixed @xmath142 the limit @xmath143 .
this implies @xmath144 , the first term goes to zero and the integral is easily calculated in this limit , leading to : @xmath145 on the other hand if at fixed @xmath146 we take the limit @xmath147 , we have @xmath148 . the limiting behaviour of the integral is @xmath149 as can be obtained through a by parts integration , the dominant divergent contribution from the two terms cancels out and we are left with @xcite : @xmath150 which goes naturally to infinity for @xmath151 .
these two limits can be obtained more rapidly by making the proper simplifications from the start of the calculation eq.([eq18 ] ) .
the two limiting cases which we have just considered show that , at low temperature , @xmath121 has a quite remarkable behaviour . for most of the range
it is negative as it can be seen from eq.([jlargealpha ] ) and it even goes to large negative values when @xmath142 gets very small .
this tends to favor states with small angle between wavevectors , as we will see below . on the other hand for @xmath152 or very small @xmath153 is positive and very large , as it results from eq.([eq23 ] ) .
clearly at low @xmath36 the interesting range is the small @xmath142 domain where we can write @xmath154 .
surprisingly @xmath153 first starts to increase strongly from its @xmath155 value before going down to very negative values .
this can be seen simply by looking at the specific point @xmath156 for which the second term is negligible and which gives to dominant order @xmath157 , even more diverging for @xmath151 than the @xmath158 value .
the integral of @xmath121 over @xmath142 can also be analytically evaluated , and it shows that the strong positive peak at small @xmath142 dominates over the negative contribution from the rest of the range .
this quasi - singular behaviour of @xmath121 is summarized in fig.[fig1 ] where we have plotted @xmath121 for various temperatures .
we perform now the same kind of treatment for @xmath159 .
one goes again from eq.([eq17 ] ) to an integration on the real frequency axis .
however there is no integration by parts , and the angular average to be calculated is somewhat more complicated .
it can nevertheless be performed by the same method and gives explicitely , with the same variable @xmath126 as above : @xmath160.[\cos ( \theta + \alpha /2 ) ^{2}-y^{2 } ] } = \frac{1}{2y \sqrt { y ^{2}- 1 } } ( \frac{1}{y ^{2}- \cos^{2 } ( \alpha /2)}+
\frac{1}{y ^{2}- \sin^{2 } ( \alpha /2 ) } ) \label{eq24}\end{aligned}\ ] ] the second term is obtained from the first one by the change @xmath161 .
this leads to @xmath162 with : @xmath163 here we have contributions coming from half circles around the poles at @xmath164 and contributions from the two domains @xmath165 and @xmath166 . neglecting terms which are exponentially small in the low temperature limit , we can replace the hyperbolic tangent by @xmath167 in the @xmath168 domain and for the @xmath169 pole . gathering similar contributions
this gives : @xmath170 where we have used the fact that , without the @xmath171 term , the integral can be performed exactly to give @xmath172 .
we have taken advantage of this to have the factor @xmath173 which is going rapidly to zero for large @xmath129 .
this will be of use when we consider below temperature corrections .
the last term in eq.([eq26 ] ) disappears in the combination @xmath174 , so we omit it from now on . looking now for the dominant contribution at low temperature , we simplify the hyperbolic tangent in the range @xmath165 since its argument @xmath175 is large and positive .
so one finds expressions which are similar to the one encountered in the calculation of @xmath153 .
moreover , because of the relation @xmath176 , we can restrict ourselves to the case @xmath177 . in this case
the expression obtained by the replacement @xmath161 is very simple since the corresponding value of @xmath141 is always very large at low t. this leads finally to : @xmath178 where we have taken into account that the second term is only significant when @xmath142 is small , so we have made @xmath155 in its prefactor . in the limit @xmath151 with fixed @xmath142 , we have @xmath179 and one gets merely : @xmath180 again if we take @xmath181 at fixed @xmath146 , @xmath182 , the dominant divergent contribution cancels and we find : @xmath183 these two limits can be obtained again more directly . from these cases
we can guess that @xmath184 is always negative .
this is seen on fig .
[ fig2 ] where @xmath185 has been plotted .
however , in the same way as @xmath186 , it has also a singular behaviour at small @xmath142 .
while for @xmath187 it diverges as @xmath188 , it goes to the finite value @xmath189 for @xmath152 .
we note that the divergent behaviour in @xmath190 is weaker than the one in @xmath191 found for @xmath186 . similarly latexmath:[$j(0 )
\gg |\tilde{j}(0 ) dominant role and @xmath184 will only give a subdominant contribution . the expressions eq.([eq20 ] ) and eq.([eq27 ] ) found above respectively for @xmath193 and @xmath194 are the leading terms at low temperature . while containing the dominant physical behaviour
, we do not expect them to be so good quantitatively at intermediate temperature .
it is actually possible to improve these analytical results markedly in this respect at the price of a slight complication by including the first two terms in an expansion in powers of @xmath195 .
this is the similar to what we have done at the end of section ii for the second order term , and we follow here the same procedure , together with the steps we have just taken above for the calculation of @xmath193 and @xmath194 .
we first consider first @xmath193 and start from the exact eq.([eq20 ] ) .
since the contribution from the domain @xmath166 is again exponentially small in the low temperature regime , we keep only the half circle contour around @xmath196 and the @xmath197 domain .
the integral appearing in eq.([eq20 ] ) can be split in two terms by : @xmath198 therefore @xmath193 can be written as @xmath199 and we concentrate on the calculation of @xmath200 . with the same notations @xmath201 , @xmath75 and the change of variable @xmath202 , eq.([eq20 ] )
leads to : @xmath203 where we have set @xmath204 . the first term in the r.h.s .
of eq.([jo ] ) is expanded for low @xmath205 by @xmath206 and the resulting temperature correction is calculated to lowest order , replacing @xmath78 by @xmath207 .
this gives our following final expression , which has to be used together with eq.(16 ) for @xmath78 and @xmath138 : @xmath208 the second term in the r.h.s of this equation is only relevant for angles that are close to zero .
otherwise , it gives an exponentially small contribution in the low temperature limit . in the expression obtained by the change
@xmath209 , this term is exponentially small in any case and can be forgotten .
the first and the third term can be calculated more explicitly for angles @xmath142 which are not close to zero ( this is in particular the case for @xmath210 ) , which implies @xmath211 . for the first term in particular we can make use of our results in section ii for the temperature expansion where similar terms were found .
this leads finally in this regime to : @xmath212 for various temperatures , we compare in fig .
[ fig1 ] these low temperature expressions with the exact calculation of @xmath193 , from the direct numerical summation in eq.([eq16 ] ) over matsubara frequencies ( with @xmath138 obtained by the numerical minimisation of @xmath38 given by eq.([eq6 ] ) ) .
we see that , at this level of accuracy , they agree remarkably well up to rather high temperatures . ]
figure-t0.01bis.eps ] figure-t0.05bis.eps we turn now to a similar calculation for @xmath194 and take up from eq.([eq26 ] ) .
we have already pointed out that , since we can restrict ourselves to @xmath213 , the replacement @xmath214 gives a value of @xmath215 which is always large at low temperature .
therefore , in this replacement , the integral in eq.([eq26 ] ) can be calculated to lowest order in @xmath216 and gives @xmath217 . in the same way in the first term
the hyperbolic tangent can be replaced by -1 in this substitution . with the same change of variable as for @xmath193
, we have finally : @xmath218 as for the calculation of @xmath193 , we can have a more explicit result by expanding the denominator in the integral to first order in powers of @xmath216 and computing the resulting temperature correction to lowest order by replacing @xmath78 by @xmath80 .
but we will not write the cumbersome resulting formula . nevertheless , as done previously for @xmath193 , we compare in fig .
[ fig2 ] our low temperature expressions with the exact calculation of @xmath219 , @xmath138 and @xmath220 .
the discrepancy appears only above @xmath221 in the vinicity of @xmath222 and is mainly due to the fact that it becomes inaccurate to consider that @xmath223 is large in calculating @xmath224 , as we have done ( see above eq.([eq27 ] ) ) .
however our low temperature expressions are already sufficient to describe the cascade .
the second order term in the expansion of the free energy @xmath225 gives the location of the transition @xmath226 and the optimum wavevectors modulus @xmath227 entering the order parameter @xmath228 but it can not distinguish between different order parameter structures , because from eq.([freeener ] ) the fourier components @xmath229 of the order parameter @xmath230 are decoupled in the second order term .
the resulting degeneracy is lifted by the fourth order term when one goes slightly into the superfluid phase .
this is the basis of the analysis of larkin and ovchinnikov @xcite and of shimahara @xcite .
so to speak , the wavevectors directions are independent at the level of the second order term while the fourth order one provides an effective interaction between these directions .
since the expression eq.([fourthorder ] ) for the fourth order term depends only on the angle @xmath231 between the wavevectors @xmath232 and @xmath233 through @xmath234 and @xmath235 , the wavevector interaction appears therefore simply as pair interactions depending only on the relative positions of the 2d wavevectors on the circle .
we note however that @xmath235 corresponds to an interaction between two pairs of opposite wavevectors @xmath236 whereas @xmath234 gives an interaction between two single wavevectors .
we consider now the ingredients which lead to the prediction of a cascade of transitions between order parameters with increasing number of wavevectors when the temperature goes to zero .
first , since @xmath193 takes very large positive values for @xmath237 smaller than a critical angle @xmath238 , we can consider that the angle domain @xmath239 is forbidden in order to avoid a dramatic increase of the free energy @xmath225 . to be quite specific
we define @xmath240 by @xmath241
. however we could as well take the critical angle as @xmath242 which gives the minimum of @xmath193 , since @xmath240 and @xmath242 are anyway very close as it can be seen on fig.[fig1 ] ( we will indeed consider also @xmath243 in the following ) . on the other hand
it is favorable to take @xmath244 slightly above @xmath238 since it is the region where @xmath193 takes its most negative values .
we note that the behavior of @xmath194 is not as strong compared to @xmath193 , so we neglect @xmath194 in a first approximation .
next we see that @xmath245 diverges as @xmath246 at low temperature and it is necessarily present in the free energy eq.([fourthorder ] ) from the terms with @xmath247 .
however their unfavorable effect on @xmath248 is lowered in relative value if one increases the number @xmath249 of wavevectors , since we have @xmath249 terms in eq.([fourthorder ] ) containing @xmath245 compared to a total number of @xmath250 terms .
since we want to minimize @xmath248 , this leads us to increase @xmath249 and hence to decrease the angle between wavevectors as much as possible .
since the angle between wavevectors is bounded from below by @xmath238 , we expect for symmetry reasons that the optimum structure to have regularly spaced wavevectors with an interval angle slightly above @xmath238 .
the last ingredient to predict the cascade is the fact that @xmath238 decreases with temperature . as a result ,
the number @xmath249 of plane waves in the optimum structure increases when the temperature goes to zero .
this leads to a cascade of states where the @xmath5 limit is singular . in the next sections , we present our arguments in more details .
we begin by considering only the contribution @xmath251 to the free energy @xmath225 from the @xmath193 terms .
even so the full problem of finding the order parameter structure minimizing the fourth order term is not a simple one .
however it is natural to assume that the wavevectors of the @xmath249 plane waves have a regular angular separation with an angle @xmath252 between neighbouring wavevectors , so that their angular position is given on the circle by @xmath253 .
otherwise we would have a minimum corresponding to a disordered situation for the angles , which sounds quite unlikely ( note that we can not collapse an angle to zero since we work at fixed @xmath249 ; we will later on minimize with respect to @xmath249 ) .
then it is easy to show that the weight @xmath254 of the various wavevectors are all equal . indeed the fourth order term eq.([fourthorder ] ) is just a quadratic form in @xmath255 and minimizing it is formally identical , for example , to find the lowest energy for a single particle in a tight binding hamiltonian on a ring , with hopping matrix element @xmath256 between site @xmath257 and site @xmath258 ( except for the on - site term which is @xmath259 ) . the eigenvectors are plane waves and the eigenvalues are @xmath260 with @xmath261 and @xmath262 .
since @xmath263 for @xmath264 , the lowest eigenvalue corresponds to @xmath265 which means that the weight @xmath255 are all equal .
conversely if we assume from the start that the all weights @xmath255 are equal , our problem of finding the best @xmath266 s is the same as the one of finding the equilibrium position of atoms on a ring with repulsive short range interaction ( because @xmath121 is large and positive for small @xmath142 ) and attractive long range interaction ( because @xmath121 is negative for larger @xmath142 ) .
we expect the equilibrium to correspond to a crystalline structure with regularly spaced atoms .
this takes into account that @xmath121 is a long range potential ( clearly this regular spacing would not be the equilibrium if we had a strongly short range potential enforcing a specific distance between our atoms ) .
naturally in this argument we take @xmath249 large enough to fill up the ring with atoms .
so we come to the conclusion that the minimum energy corresponds to equally spaced wavevectors . finally when we minimize the total free energy eq.([freeener ] ) with respect to the weight @xmath255 we find the general result : @xmath267 where we have , in the case @xmath268 : @xmath269 we note that we still have a degeneracy of the lowest energy configuration with respect to the choice of the plane waves phases .
we now take also into account in the free energy the terms containing @xmath235 .
these terms appear when there are pairs of opposite wavevectors @xmath236 . with our assumption of regularly spaced wavevectors ,
this corresponds to even @xmath249 - in which case we have @xmath270 pairs - whereas for odd @xmath249 , there are no @xmath235 terms in the free energy .
note that @xmath194 is negative for any angle @xmath271 so that it is favorable to take pairs of opposite wavevectors .
this seems to be in favor of taking @xmath249 even and we will indeed see that as a result states with even @xmath249 will be selected . for even @xmath249 ,
the total free energy is : @xmath272 where we have used @xmath273 and @xmath274 ; @xmath275 is a shorthand for @xmath276 and so on .
the fact that @xmath194 is always negative has a direct consequence on the phases of the plane waves . in order to minimize the free energy they have to be chosen so that @xmath277 is always real .
writing @xmath278 , this implies that @xmath279 for any @xmath257 , where @xmath280 is a constant phase which can be chosen to be zero , since this merely corresponds to a global phase change for the order parameter .
this link between phases for opposite wavevectors removes only a part of the degeneracy , since the phase @xmath281 can be still arbitrarily chosen .
now we see that the contribution of the @xmath235 terms to the free energy is a quadratic form in @xmath282 with negative coefficients , and no on - site contribution .
this is quite similar to what we have found for the @xmath193 terms in the preceding subsection .
so the equal weight distribution , which minimizes the @xmath193 terms , gives also independently the minimum of the contribution of the @xmath194 terms .
this means that the inclusion of these last terms in the free energy does not change the optimum structure apart from the phase link between opposite wavevectors
. finally the total minimum free energy is still given by eq.([minnodd ] ) , where eq.([g2odd ] ) is valid for odd @xmath249 , but has to be replaced for even n by : @xmath283 in this last case the corresponding equilibrium order parameter is real , being a sum of cosines of the form @xmath284 .
we turn now to the minimization of the free energy with respect to @xmath249 .
we consider now the numerical calculation of @xmath285 for different order parameter structures , i.e. , for various values of the number of plane waves @xmath249 . from eq.([minnodd ] ) the equilibrium order parameter corresponds to the maximum @xmath286 ^{-1}$ ] . in fig .
[ fig3 ] we show for @xmath286 ^{-1}$ ] both our low temperature expansion and the exact evaluation of the matsubara sum in eqs .
( [ eq16 ] ) and ( [ eq17 ] ) , as a function of the reduced temperature @xmath216 .
we see that below @xmath221 , the low temperature analytical expressions , eqs .
( [ alowtemp ] ) , ( [ eqj ] ) and ( [ jtilde ] ) , agree remarkably well with the exact result .
they are therefore completely sufficient quantitatively to study the cascade .
[ fig4 ] displays the same quantity for lower temperatures , exhibiting the cascade of transitions .
an interesting feature of this cascade is that an order parameter with an odd number of plane waves is never the lowest energy solution , except from the @xmath287 case , already found by shimahara @xcite . the reasons for this result will appear clearly in the next section , where we will study analytically the asymptotic regime of low temperatures .
in the low temperature limit @xmath288 , we can derive an explicit expansion for the value of the zero @xmath240 of @xmath193 , as well as the location @xmath242 of its minimum .
these critical angles are crucial to study the cascade since they give essentially the minimum angle between two wavevectors . in this low temperature range
we have @xmath289 , so that @xmath290 . as can be seen in fig.[fig1 ]
, @xmath193 has two extrema : a maximum for @xmath142 around @xmath291 , as it is clear from the @xmath292 term in eq.([jalpha ] ) , and , for a slightly larger value of @xmath142 , a minimum at @xmath242 .
the zero @xmath240 is naturally in between .
the condition @xmath293 implies @xmath294 .
therefore at low @xmath205 , both @xmath295 and @xmath242 are in a domain where @xmath296 and @xmath297 . in this regime
the two terms in eq.([jalpha ] ) for @xmath193 simplify to give : ] unsurg22.eps @xmath298 where @xmath299 .
the second term in this expression simply comes from @xmath300 which is the asymptotic limit eq.([jlargealpha ] ) for @xmath301 .
let us first consider the calculation of @xmath242 .
the first term is responsible for the strong upturn near the minimum . from the derivative of eq.([asympto ] ) @xmath193 is extremal for : @xmath302 the minimum we are looking for is in the domain @xmath303 and from the above equation it is found for a large value of @xmath304 .
this allows to simplify this equation into : @xmath305 with @xmath306 .
writing this equation @xmath307 one can generate a solution by the recurrence relation @xmath308 which converges very rapidly .
for example the second iteration gives : @xmath309 \label{secondit}\ ] ] the corresponding result for @xmath310 is to leading order in the low temperature limit : @xmath311 but numerically this is not such a good result at low temperature and one has rather to perform a few iterations to get the correct answer .
for example in eq.([eqk ] ) , the exact result is @xmath312 for @xmath313 , while @xmath314 , but the second iteration eq .
( [ secondit ] ) gives @xmath315 .
then in order to find the optimum number of plane waves we notice that @xmath193 rises very rapidly below @xmath316 , so we can not have basically the angular separation @xmath317 between two plane waves less than @xmath316 .
since on the other hand it is energetically favorable to take @xmath249 as large as possible , as long as @xmath193 is negative , we can find an asymptotic estimate of the optimum value @xmath318 of @xmath249 by taking the integer value of @xmath319 , that is : @xmath320 \label{asymptn}\ ] ] following the same procedure , we can also derive an explicit expression for the angle @xmath238 corresponding to the zero of @xmath193 . from eq.([asympto ] ) , we find : @xmath321 naturally @xmath238 is close to @xmath316 since @xmath193 is rapidly increasing below @xmath316 .
we are still in the @xmath322 domain and the solution corresponds to @xmath304 large .
this simplifies the above equation into : @xmath323 where @xmath324 has been defined previously .
the corresponding recurrence relation which generates the exact solution is @xmath325 .
the second iteration gives here : @xmath326\ ] ] and @xmath238 is simply given by @xmath327 .
we can then make the same argument as above for @xmath242 , and write the following asymptotic estimate of the optimum value @xmath318 : @xmath328\ ] ] which will naturally be found to be quite near the above one eq.([asymptn ] ) , since @xmath238 and @xmath316 are quite close .
naturally we can calculate numerically exactly the optimal value @xmath329 of @xmath249 as a function of the temperature , from our above results . in the same process
we find also the critical temperatures where the @xmath329 changes .
they are given by the crossings of the @xmath285 curves for different values of @xmath249 , as it is seen on fig .
[ fig3 ] and [ fig4 ] . in fig .
[ figcompare ] these exact critical temperatures are compared with the exact calculations of @xmath330 and @xmath331 , as well as their asymptotic values . as it is seen on this figure , it happens that the optimal values of @xmath249 falls essentially just between @xmath332 and @xmath331 in the low temperature domain . ]
cascade-long.eps it can be seen on fig .
[ fig1 ] that @xmath193 switches rapidly above @xmath334 to its large angle asymptotic behaviour : @xmath335 so we may with a fair precision use this simplified expression for @xmath336 , with @xmath337 , to evaluate the sum in eq.([g2odd ] ) for @xmath333 . at low temperature
the number of plane waves n is large and the sum is dominated by the terms with small @xmath257 corresponding to small angles , for which we have @xmath338 .
hence we have for the sum coming in eq .
( [ g2odd ] ) for @xmath333 : @xmath339 for even @xmath249 , we have also to evaluate the contribution coming from the terms in eq .
( [ g2even ] ) containing @xmath340 .
again we can use for most angles the asymptotic expression eq.([jtildelim ] ) for @xmath341 .
once more the dominant contribution comes from the small @xmath257 terms , and the logarithmic leading behaviour is accurately given by the asymptotic expression of @xmath340 as : @xmath342 [ verifier le n=1]we see here explicitely that @xmath193 has a dominant role compared to @xmath194
. taking into account the @xmath245 contributions , the dominant behaviour of @xmath333 can be summarized in the limit of large plane wave number @xmath249 and low temperature as : @xmath343 these expressions correspond actually in fig .
[ fig4 ] only to the rising part of @xmath344 ^{-1}$ ] on the low temperature side .
the downturn of @xmath344 ^{-1}$ ] for higher temperature is due to contributions from the positive part of @xmath345 which are beyond our asymptotic approximation eq .
( [ asymptotic ] ) .
nevertheless we have also found above the critical temperatures for switching from a given value of @xmath249 to the next one . from fig .
[ fig4 ] it is seen that it is enough to plug these critical temperatures into the evaluation of the rising part of @xmath344 ^{-1}$ ] to obtain the free energy at the transitions .
when we substitute accordingly in these expressions eq .
( [ g21 ] ) or ( [ g22 ] ) the value @xmath346 for the optimum plane wave number we have : @xmath347 since @xmath348 from eq.([alphac ] ) , we find that @xmath349 is always positive in the low temperature range ( more precisely we have @xmath350 which is positive as soon as @xmath351 , i.e. from eq.([eqk ] ) exactly for @xmath352 ) .
this means that the transition stays always second order ( a negative sign would have implied a first order transition and a breakdown of our fourth order expansion ) .
finally it seems from eq.([g21]-[g22 ] ) that it is favourable to have an even number of plane waves in order to take advantage of the additional @xmath353 contribution .
this can be confirmed by a more careful comparison between two consecutives values of @xmath249 .
let us assume that @xmath249 is odd with @xmath346 and compare @xmath354 to @xmath355 . by going from @xmath249 to @xmath356
we gain naturally the @xmath353 term in eq.([g22 ] ) , but we increase the first term due to the unfavorable of the @xmath245 term . nevertheless : @xmath357 = \frac{1}{2 n(n-1 ) t } - \ln(n-1 ) + { \mathcal o}(1).\ ] ] where the dominant term becomes exactly with the help of eq.([eqk ] ) : @xmath357 = - x ( 1 - \frac{2}{\pi^2 } ) + \frac{5}{4 } \ln x + \frac{1}{4 } \ln \frac{32 } { \pi } \ ] ] which is always negative since its maximum , reached for @xmath358 , is -0.107 .
therefore the asymptotic evaluation of @xmath285 shows indeed explicitely that the only order parameters which appears at low temperature are those with an even number of plane waves , in agreement with our exact numerical results .
it is interesting to compare our results to the pairing ring picture explored by bowers and rajagopal @xcite ( br ) in the three dimensional case .
br have looked at the free energy expansion at zero temperature in @xmath359d , extending the work of larkin and ovchinnikov . in the case of multiple plane waves in the order parameter ,
they have pointed out a simple physical interpretation of their results .
their picture is based on the fulde and ferrell study @xcite of a single plane wave order parameter where they showed that one may separate the wavevector space in two complementary domains : the pairing region and the pair breaking region . in the case of a vanishing order parameter and considering only wavevectors at the fermi surface , the pairing region is merely an infinitely thin circle .
it is given by the intersection of the up and down spin fermi surfaces , after shifting one of them by @xmath360 .
the total opening angle of this circle is thereforer given by @xmath361 . in @xmath359d and at @xmath5 , @xmath362 which gives @xmath363 . for a non - zero order parameter amplitude @xmath364 ,
the circle broadens and becomes a ring whose width is given by @xmath365 .
therefore , the pairing region for a single plane wave is a ring drawn on the fermi surface , whose center is on the plane wave direction @xmath366 ( the restriction to the fermi surface is justified by the fact that only wavevectors close to the fermi surface are relevant for pairing ) . in the case of multiple plane waves in the order parameter
, br showed that the intersection of pairing circles from different plane waves is energetically largely unfavored , the worst case being when two circles are exactly in contact .
the maximisation of the pairing regions leads naturally to increase as much as possible the plane waves number @xmath249 with the constraint that they are no intersecting circles .
this leads to nine circles on the fermi surface .
br rather concluded that the least energy state was the one with eight plane waves for regularity reasons . in particular , this leads to a real order parameter .
we see now that our cascade in @xmath367d is a direct consequence of these two heuristic principles : the no - intersecting circles rule and the maximisation of plane waves . in @xmath367d ,
the circles are replaced by pairs of points separated by the angle @xmath368 , and the no - intersecting rule becomes now a no - entanglement rule between different pairs . but this is just equivalent to our above finding from fig .
[ fig1 ] that there is an effective short range repulsion between different wavevectors @xmath369 , with indeed the worst case corresponding to the contact since the maximum of @xmath186 is just below its zero .
the contact condition @xmath370 that we have found ( corresponding to the maximum of @xmath186 ) can be rewritten as @xmath371 which implies @xmath372 .
this is exactly the same as the contact condition of br .
next we have found explicitely that it is indeed favorable to maximize the number of plane waves , taking the short range repulsion into account .
our cascade is due to the fact that , when the temperature goes to zero , @xmath368 also drops to zero and the pairing region for a single plane wave shrinks to a single point , corresponding to the situation where the two up and down fermi circles are just in contact .
in this paper we have investigated the low temperature range for the fflo transition in two dimensions and we have shown that the order parameter is no longer a simple @xmath4 at the second order transition , in contrast with the situation found near the tricritical point . instead the transition is toward more and more complex order parameters when the temperature goes to zero , which gives rise to a cascade with , in principle , an infinite number of transitions . at the transition
these order parameters are in general a superposition of an even number of plane waves with equal weight and equal wavevector modulus , which corresponds to a real order parameter equal to a superposition of cosines .
the directions of these wavevectors are found to be equally spaced angularly , with a spacing which goes to zero when the temperature goes to zero , which is the reason for the ever increasing number of plane waves in this limit .
the singular behaviour in this limit @xmath0 is actually present in all physical quantities .
it arises because , in order to obtain the lowest energy , the two fermi circles corresponding to opposite spins come just in contact when one applies the shift corresponding to the wavevector @xmath360 of the fflo phase .
this is this situation , with these two circles just touching each other , which gives rise to the singularities .
naturally this is linked to the fact that in 2d the fermi surface is actually a line .
hence this singularity is a general feature of 2d physics and we may expect it to give rise to similar consequences in more realistic and more complex models describing actual physical systems .
naturally in this paper we have addressed this remarkable situation in the most simple physical frame .
we have considered the simplest bcs model with isotropic fermi surface , that is a circle for our 2d case . naturally if anisotropy is introduced in the dispersion relation of the electrons , or in the effective electron - electron interaction , or both
, the physics will be more complex . in the same way we expect fermi liquid effects @xcite to bring important quantitative modifications .
similarly if we consider a quasi 2d superconductor with a magnetic field not perfectly aligned with the planes , currents will be produced and orbital contributions to the free energy will arise . however in order to have the physics right in all these complex situations , it is quite clear that one has to obtain the correct limit in the simplest possible case that we have considered in the present paper .
in this section we rederive rapidly our result eq.([leading2 ] ) for the leading order of the optimal wavevector at low temperature , taking the same starting point as bulaevskii @xcite .
we start from eq.11 of ref.@xcite : @xmath373\end{aligned}\ ] ] we go to our notations by setting @xmath374 and introduce as above our reduced variables @xmath375 and @xmath59 . when we perform the angular integration as in ref .
@xcite by setting @xmath376 and introduce for the digamma function the integral representation ( eq.13 of ref .
@xcite ) : @xmath377 we obtain eq.14 of ref . @xcite ( correcting some minor misprints ) .
we can then expand the result to all orders in @xmath205 and show that all the terms in the expansion are zero .
however rather than displaying the steps of this calculation here , it is more convenient to immediately remark that the optimal @xmath378 is obtained by writing that the derivative of eq.11 with respect to @xmath378 is zero .
this provides an equivalent calculation ( we work on the derivative instead of working on the function ) , which is somewhat easier and allows us also to make in the following the direct contact with our result .
this condition on the derivative , which gives the equation for the optimal @xmath378 , is : @xmath379 taking the derivative of the above integral representation , we obtain a corresponding representation for @xmath380 on which the angular integration is easily performed ( the integrals can be found in gradstein and ryzhik @xcite ) .
we display the result by introducing the function @xmath381 and setting @xmath382 , @xmath383 and @xmath384 .
the result is : @xmath385 = 0 \label{condition}\end{aligned}\ ] ] where the integral goes from @xmath386 to @xmath387 because we have collected two terms into one .
now we can write for the first three terms the taylor expansion : @xmath388 when we take the real part , only the odd order derivatives of @xmath389 contribute because @xmath378 is near @xmath390 with @xmath391 , which makes @xmath392 and @xmath393 .
this gives : @xmath394 = 1 + \sum_{p=1}^{\infty } \frac{\epsilon ^{2p+1}}{(2p ) ! } \frac{2p}{2p+1 } f ^{(2p+1 ) } ( x_0)\end{aligned}\ ] ] similarly we can perform the expansion in the integral and expand @xmath395 by introducing the binomial coefficients @xmath396 .
all the resulting integrals can be found in gradstein and ryzhik @xcite , and expressed in terms of bernoulli numbers @xmath397 .
this leads to : @xmath398 = \sum_{p=1}^{\infty } \frac{\epsilon ^{2p+1}}{(2p+1 ) ! } f ^{(2p+1 ) } ( x_0 ) \sum_{m=1}^{p } c _ { 2p+1}^{2 m } 4^m b_{2m}\end{aligned}\ ] ] now a remarkable identity ( found for example in gradstein and ryzhik @xcite ) for bernoulli numbers states that : @xmath399 as a result , by gathering all the terms , all the coefficients of the powers of @xmath400 are zero , and it is not possible to satisfy the condition that the above derivative is zero .
the answer to this puzzle is that the contribution of the terms we have considered is not exactly zero , but exponentially small , which explains why we find it to be zero in a perturbative expansion .
this is shown now in the following .
we start again from the above condition eq.([condition ] ) found for the optimum wavevector .
we transform the integral by shifting the integration contour toward the upper complex plane by @xmath401 for the variable @xmath402 .
first we have to take care that integrant in the above integral has no singularity for @xmath403 , because proper cancellation between various terms .
hence we can also say that this integral is equal to its principal part .
next if we want a complete contour @xmath404 , we have to add to this principal part the contribution of an infinitesimal semi - circle around @xmath405 with positive imaginary part . this contribution is easily found by residues to be equal to @xmath406 , so we have : @xmath407 = - \frac{\pi}{2 } { \rm re}\ , \int_{c } \frac{dy}{\sinh^2 ( \pi y ) } [ f(x ) - f(x_0 + \epsilon ( 1 + 2iy ) ) ] \end{aligned}\ ] ] now we set @xmath408 and make use of @xmath409 which gives : @xmath410 = \frac{1}{2 } { \rm re}\ , \int_{-\infty } ^ { \infty } \frac{dy}{\cosh^2 z } ( f(x ) - f(x_0 + 2tz/\bar{q } ) ) \end{aligned}\ ] ] where the contour runs infinitesimally below the real @xmath411 axis . the first term in the last integral is just equal to @xmath412 , so our equation becomes : @xmath413 = 0\end{aligned}\ ] ] it is easily checked that this equation is identical to our eq.([leading ] ) .
note that the fact that the contour runs infinitesimally above or below the real axis is unimportant since we take the real part , and the contribution along the cut due to the square root is purely imaginary .
in conclusion we have obtained our basic equation for the second order term by taking the same starting point as ref .
the end of the argument to obtain eq.([leading2 ] ) is naturally the same as following our eq.([leading ] ) . in particular
this argument shows that the integral in the above formula is proportional to @xmath414}$ ] , so it can not be expanded in powers of @xmath205 .
p. fulde and r. a. ferrell , phys.rev . * 135 * , a550 ( 1964 ) . a. i. larkin and y. n. ovchinnikov , zhetf * 47 * , 1136 ( 1964 ) [ sov .
jetp * 20 * , 762 ( 1965 ) ] .
see in particular r. casalbuoni and g. nardulli , rev .
76 , 263 ( 2004 ) for a review more particularly focused on the interest of the fflo phases for quark matter , as might be found in neutron stars .
a. bianchi , r. movshovich , c. capan , p. g. pagliuso , and j. l. sarrao , phys .
lett . * 91 * , 187004 ( 2003 ) and c. martin , c. c. agosta , s. w. tozer , h. a. radovan , e. c. palm , t. p. murphy , j. l. sarrao , cond - mat/0309125 , and references therein for other possible experimental observations .
shimahara , phys .
b , * 50 * , 12760 ( 1994 )
. j. singleton , j. a. symington , m. s. nam , a. ardavan , m. kurmoo and p. day , j. phys .
cond . mat .
* 12 * , l641 ( 2000 ) .
m. a. tanatar , t. ishiguro , h. tanaka and h. kobayashi , phys .
b 66 , 134503 ( 2002 ) ; for recent papers on this question see for example u. klein , phys .
b 69 , 134518 ( 2004 ) ; m. houzet and a. buzdin , europhys .
lett . , * 50 * , 375 ( 2000 ) ; k. yang , a. h. macdonald , cond - mat/0404580 and references therein . h. burkhardt and d. rainer , ann.physik * 3 * , 181 ( 1994 ) .
s. matsuo , s. higashitani , y. nagato and k. nagai , j. phys .
japan , * 67 * 280 ( 1998 ) .
m. houzet , y. meurdesoif , o. coste and a. i. buzdin , physica c * 316 * , 89 ( 1999 ) .
r. combescot and c. mora , eur.p.j .
b * 28 * , 397 ( 2002 ) .
r. combescot and c. mora , cond - mat/0311042 .
c. mora and r. combescot , to be published in europhys .
shimahara , j. phys .
japan , * 67 * , 736 ( 1998 ) .
g. eilenberger , z. phys . * 182 * 427 ( 1965 ) .
n. r. werthamer in superconductivity , ed .
by r. d. parks ( dekker , new - york , 1969 ) .
l. n. bulaevskii , zhetf * 65 * , 1278 ( 1973 ) [ sov .
jetp * 38 * , 634 ( 1974 ) ] .
more precisely this is only valid to lowest order in @xmath205 . to next order the divergence in @xmath415 in the two terms in eq.([jalpha ] )
do not cancel exactly , which gives rise to the divergent behaviour seen in fig .
[ fig1 ] for the dashed - dotted curve ( and a similar behaviour occurs for @xmath353 ) .
our temperature corrections eliminate this next order problem .
j. a. bowers and k. rajagopal , phys .
d * 66 * 065002 ( 2002 ) . i. s. gradshteyn and i. m. ryzhik , table of integrals , series and products ( academic press , 1980 ) . | we explore the nature of the transition to the fulde - ferrell - larkin- ovchinnikov superfluid phases in the low temperature range in two dimensions , for the simplest isotropic bcs model .
this is done by applying the larkin - ovchinnikov approach to this second order transition .
we show that there is a succession of transitions toward ever more complex order parameters when the temperature goes to zero .
this gives rise to a cascade with , in principle , an infinite number of transitions .
except for one case , the order parameter at the transition is a real superposition of cosines with equal weights .
the directions of these wavevectors are equally spaced angularly , with a spacing which goes to zero when the temperature goes to zero .
this singular behaviour in this @xmath0 limit is deeply linked to the two - dimensional nature of the problem . |
single - particle interferometry became a remarkable tool in various fields of quantum physics and sensor technology .
interferometers for coherent atoms @xcite recently investigated the nature of time @xcite and measured inertial forces @xcite and gravitational acceleration @xcite .
molecule interferometers @xcite proved the wave nature of large particles @xcite and contributed to the understanding of quantum decoherence @xcite . in neutron interferometers
@xcite , the quantum - mechanical phase shift due to the earth s gravitational field was observed @xcite .
moreover , remarkable progress was achieved in the field of matter - wave interferometry with charged particles such as electrons and ions @xcite based on new developments concerning the beam source @xcite , the precise electron guiding @xcite , the coherent beam path separation by nanostructures @xcite and highly resolved spatial and temporal single - particle detection @xcite .
this advance opened the door for experiments in aharonov - bohm physics @xcite and coulomb - induced decoherence @xcite .
technical devices with interfering single particles can decrease the amount of destructive particle deposition in electron microscopy for the analysis of fragile biological specimen @xcite .
all these technical @xcite and fundamental applications @xcite of single - particle interferometers are based on the high phase sensitivity and are therefore susceptible to dephasing , which can be caused by external electromagnetic oscillations , mechanical vibrations or temperature drifts .
in contrast to decoherence @xcite , where actual information of the quantum state is lost to the environment , dephasing is a collective , time - dependent phase shift of the interference pattern . both , decoherence and dephasing , cause a reduction of the contrast of the time - averaged interference pattern on the detector . however , in opposite to decoherence , dephasing can in principle be corrected after the measurement , if the temporal and spatial information of the single - particle events are known
then , two - particle correlations may be used to study the dephasing process and to reveal the undisturbed interference pattern .
ever since the famous hanbury brown and twiss experiment @xcite , second - order correlations are successfully used in many research areas .
thereby , noise correlations play a key role , as they give direct access to the quantum nature of the source .
this understanding has not only set the fundament for modern quantum optics @xcite , but also helped to prove the quantum nature of fermions @xcite and bosons @xcite .
today , noise correlation analysis is widely used in modern astrophysics @xcite , quantum atom optics @xcite and particle physics @xcite . for matter - waves ,
temporal correlations have been used to analyze the counting statistics of atom lasers @xcite and to demonstrate the coherent transfer of magnetic field fluctuations onto an atom laser @xcite .
spatial correlations , on the other hand , have been used to analyze atoms in optical lattices @xcite or to study many - body states , such as the mott insulator state , in cold atom physics @xcite . in previous publications @xcite
we have demonstrated in a biprism electron interferometer @xcite , how multifrequency dephasing caused by electromagnetic @xcite and vibrational oscillations @xcite can be corrected using second - order correlation analysis in combination with the amplitude spectrum of the correlation function .
latter can be used for the identification of unknown perturbation frequencies @xcite , as according to the wiener - khintchine theorem @xcite the fourier transform of the correlation function is equal to the power spectrum of the perturbed measurement signal .
for the measurements , an interference pattern was shifted artificially by external perturbations leading to a contrast reduction of the temporally integrated pattern on the detector . using the time and position information of particle impacts at the detector ,
the numerical second - order correlation function was extracted . with this , we were able to reveal the unknown perturbation frequencies , corresponding amplitudes and the characteristics of the matter - wave , such as contrast and pattern periodicity .
the undisturbed interference pattern could be reconstructed with the parameters of the perturbation .
our method is a powerful tool to prove the wave nature of particles , even if the integrated interference pattern is vanished .
therefore , it reduces the requirements for electromagnetic shielding and vibrational damping of the experimental setup , e.g. for mobile interferometers or experiments in a perturbing environment .
furthermore , it can be used to sensitively detect electromagnetic and mechanical perturbations or for the spectroscopy of the electromagnetic and vibrational response spectrum of an interferometer @xcite .
therefore , this technique has the potential for the application in sensor technology and can in principle be applied in interferometers for electrons @xcite , ions @xcite , neutrons @xcite , atoms @xcite and molecules @xcite that generate a spatial fringe pattern in the detection plane . for the application of the correlation analysis , the devices have to be equipped with a detector with high spatial and temporal single - particle resolution , which is available for all above mentioned interferometers .
another requirement is , that the particle flight time is shorter than the cycle time of the perturbation .
otherwise , the particles traverse many periods of the perturbation and therefore the perturbation is averaged out and can not be resolved .
this article provides a comprehensive description of the applied theory , being the base for the experimental application of second - order correlations in single - particle interferometry . in the first chapter ,
we give a detailed derivation of the two - dimensional second - order correlation theory for multifrequency perturbations leading to the equations applied in former dephasing experiments @xcite .
we deduce the explicit solution for the correlation function and explain under which conditions an approximation can be applied .
the characteristics of the explicit and approximate solutions are discussed and the determination of the matter - wave properties is shown . furthermore , we calculate the analytic solution of the corresponding amplitude spectrum used for the identification of the unknown perturbation frequencies and amplitudes @xcite .
the invariance of the correlation function under time and space transformations is analyzed in detail and the consequence for the determination of the perturbation parameters is shown . in the second part of this article
, we investigate the general characteristics of numerical correlation functions .
they are typically derived from a finite set of measurement data , causing statistics and noise to play an important role .
temporal and spatial discretization , then influences not only the contrast and the amplitude spectrum of the correlation function , but also the corresponding noise levels .
this limits the maximal signal - to - noise ratio in the correlation analysis . from our theoretical study
, we identify an optimal discretization step size for best sensitivity .
using single - particle simulations of a perturbed interference pattern , we cross - check our theoretical description and show , how the correlation theory can be used to identify broad - band frequency noise .
in this chapter , the theory of second - order correlations in single - particle interferometry is derived and the properties are discussed in detail .
first , the contrast reduction of a time - averaged interference pattern dephased by a perturbation is analyzed .
afterwards , the explicit solution of the second - order correlation function is calculated and discussed under which conditions an approximate solution is suitable
. the determination of the contrast and spatial periodicity of the unperturbed interference pattern is demonstrated .
the corresponding amplitude spectra used for the identification of unknown perturbation frequencies are derived . at the end of the chapter
, the invariance of the correlation function under time and space transformations is analyzed in detail and the consequence for the determination of the perturbation phases is discussed .
in many experiments in single - particle interferometry , the interference pattern is detected using multichannel plates ( mcps ) in conjunction with a phosphor screen @xcite .
the particle impacts generate light pulses on the phosphor screen , that are temporally integrated with a charge - coupled device camera ( ccd - camera ) .
an interference pattern , that is dephased by a temporal perturbation , is then irreversibly washed - out " in the spatial signal and its contrast is reduced .
this behaviour shall be calculated in the following .
the probability distribution , that describes the particle impacts in the detection plane forming the interference pattern , is given by @xmath0 where @xmath1 assures normalization , @xmath2 and @xmath3 indicate the contrast and wave number of the unperturbed interference pattern , with the spatial periodicity @xmath4 .
the time - dependent perturbation @xmath5 is described as a superposition of @xmath6 harmonic frequencies @xmath7 @xmath8 with perturbation amplitudes ( peak phase deviations ) @xmath9 and phases @xmath10 .
the perturbation leads to a washout of the time - averaged interference pattern .
for one perturbation frequency ( @xmath11 ) , this can be easily seen by calculating the time - average of equation ( [ eq1 ] ) @xmath12 resulting in @xmath13 using bessel functions of first kind @xmath14 , the exponential function can be rewritten as @xmath15 yielding @xmath16 only for @xmath17 , the limit of the time integral is equal to one . for all other @xmath18 , it approaches zero , such that the time - averaged interference pattern becomes @xmath19 the perturbation thus leads to a reduced contrast @xmath20 given by the perturbation amplitude : @xmath21 , with @xmath22 . in figure [ fig1 ] at the top ,
the dependence of the time - averaged interference pattern ( equation ( [ eq7 ] ) ) on the peak phase deviation @xmath23 is illustrated .
the interference pattern with a spatial periodicity of @xmath24{mm}$ ] can be identified in the @xmath25-direction .
the reduced contrast @xmath20 normalized to the contrast of the unperturbed interference pattern @xmath2 is plotted at the bottom of figure [ fig1 ] . for a peak phase deviation of @xmath26{\pi}$ ] , the contrast is zero corresponding to the first zero of @xmath27 .
the contrast returns for larger peak phase deviations , but does not recover completely . additionally , the interference pattern is phase shifted by @xmath28 as the sign of @xmath27 changes from positive to negative . this behaviour is repeated for higher peak phase deviations and the contrast is further reduced
. for multifrequency perturbations ( @xmath29 ) with peak phase deviations @xmath30 , equation ( [ eq7 ] ) becomes @xmath31 here , the reduced contrast @xmath32 depends on all peak phase deviations @xmath9 , via the product of the zeroth order bessel functions .
therefore , the contrast is typically stronger reduced as compared to the single frequency case in figure [ fig1 ] . ) ) on the peak phase deviation @xmath23 for an interference pattern with a periodicity of @xmath24{mm}$ ] along the @xmath25-direction .
bottom : dependence of the relative contrast reduction @xmath33 on the peak phase deviation @xmath23 . for @xmath26{\pi}$ ] , the contrast vanishes due to the first zero of @xmath27 .
the contrast returns for larger peak phase deviations , but not completely . at each sign change of @xmath27 ,
the phase of the interference pattern is shifted by @xmath28.,scaledwidth=50.0% ] with the spatial and temporal information of the particles arriving in the detection plane , it is possible to reveal the contrast @xmath2 and spatial periodicity @xmath4 of the unperturbed interference pattern by correlation analysis .
this is possible , because in the correlation function the spatial and temporal differences are taken into account in contrast to the temporally integrated interference pattern .
then , on timescales below the perturbation frequency , the interference pattern is still visible and not influenced by the perturbation .
furthermore , the characteristics of the perturbation ( frequencies @xmath7 , peak phase deviations @xmath9 and phases @xmath10 ) can be determined from the correlation analysis . with the probability distribution in equation ( [ eq1 ] ) , the second - order correlation function reads @xmath34 with @xmath35 denoting the average over position and time @xmath36 if the acquisition time @xmath37 and length @xmath38 are large compared to the involved perturbation frequencies @xmath39 and spatial periodicity @xmath40 , equation ( [ eq9 ] ) can be solved analytically .
first , the term @xmath41 in the denominator of equation ( [ eq9 ] ) is calculated , using equation ( [ eq1 ] ) and ( [ eq10 ] ) @xmath42 in the limit of large acquisition length @xmath40 , the spatial limit becomes @xmath43 yielding @xmath44 using @xmath45 and @xmath46 in equation ( [ eq11 ] ) , the second term in the denominator becomes @xmath47 .
this is expected , because shifts in time and space should not alter the long time and position average of the probability distribution . using equation ( [ eq1 ] ) and ( [ eq10 ] ) , the numerator in equation ( [ eq9 ] ) results in @xmath48 \,\mathrm{d}y\mathrm{d}t~.\end{aligned}\ ] ] similar as before , the second and third term in equation ( [ eq14 ] ) vanish , leaving @xmath49 in the limit @xmath40 , again @xmath50 , yielding with equation ( [ eq2 ] ) @xmath51 using equation ( [ eq5 ] ) for @xmath6 perturbation frequencies @xmath52 equation ( [ eq16 ] ) can be rewritten , yielding @xmath53}}_{=\,a^+_{n_j , m_j}}\cdot\mbox{e}^{i\left(n_j+m_j\right)\omega_j t}\,\mathrm{d}t\ , + \nonumber\\ & \qquad+\mbox{e}^{iku } \int_0^t \prod_{j=1}^n \sum_{n_j , m_j}\ , \underbrace{\!j_{n_j}(\varphi_j)\,\!j_{m_j}(\varphi_j)\mbox{e}^{im_j\omega_j \tau}\mbox{e}^{i\left[n_j\left(\phi_j-\frac{\pi}{2}\right)+m_j\left(\phi_j+\frac{\pi}{2}\right)\right]}}_{=\,a^-_{n_j , m_j}}\cdot\mbox{e}^{i\left(n_j+m_j\right)\omega_j t}\,\mathrm{d}t \bigg ) \nonumber\\ & = \,f_0 ^ 2 + \lim_{t\rightarrow\infty } \frac{f_0 ^ 2}{t}\frac{k^2}{4}\,\cdot \nonumber\\ & \qquad\cdot\bigg(\mbox{e}^{-iku}\int_0^t \sum_{n_1,m_1}\cdots\sum_{n_n , m_n } a^+_{n_1,m_1}\cdots a^+_{n_n , m_n}\cdot \mbox{e}^{i\left(n_1+m_1\right)\omega_1 t}\cdots \mbox{e}^{i\left(n_n+m_n\right)\omega_n t}\,\mathrm{d}t\ , + \nonumber\\ & \qquad+\mbox{e}^{iku}\int_0^t \sum_{n_1,m_1}\cdots\sum_{n_n , m_n } a^-_{n_1,m_1}\cdots a^-_{n_n , m_n}\cdot \mbox{e}^{i\left(n_1+m_1\right)\omega_1 t}\cdots \mbox{e}^{i\left(n_n+m_n\right)\omega_n
t}\,\mathrm{d}t \bigg ) \nonumber\\ & = \,f_0 ^ 2 + f_0 ^ 2\frac{k^2}{4}\,\cdot \nonumber\\ & \qquad\cdot\bigg(\mbox{e}^{-iku } \sum_{n_1,m_1}\cdots\sum_{n_n , m_n } a^+_{n_1,m_1}\cdots a^+_{n_n , m_n}\cdot\lim_{t\rightarrow\infty } \frac{1}{t}\int_0^t \mbox{e}^{i\sum_{j=1}^n\left(n_j+m_j\right)\omega_j t}\,\mathrm{d}t\ , + \nonumber\\ & \qquad+\mbox{e}^{iku } \sum_{n_1,m_1}\cdots\sum_{n_n , m_n } a^-_{n_1,m_1}\cdots a^-_{n_n , m_n}\cdot\lim_{t\rightarrow\infty } \frac{1}{t}\int_0^t \mbox{e}^{i\sum_{j=1}^n\left(n_j+m_j\right)\omega_j t}\,\mathrm{d}t \bigg)~.\end{aligned}\ ] ] a closer look to the time integral reveals , that it can only become zero or one @xmath54 this reduces the @xmath55 sums in equation ( [ eq18 ] ) to a ( single ) sum over all integer multiplets @xmath56 for which the constraint @xmath57 is fulfilled . here
it shall be noted , that for a finite acquisition time @xmath37 , the constraint has to be changed to @xmath58 because the minimal resolvable frequency is defined by the measurement time via @xmath59 . in the following
, however , it is assumed that @xmath60 and therefore equation ( [ eq20 ] ) is used for the calculations and discussions .
trivially , equation ( [ eq20 ] ) is satisfied for all integer multiplets with @xmath61 .
however , depending on the specific values of @xmath7 , the constraint might be fulfilled by additional integer multiplets @xmath62 with @xmath63 . the constraint from equation ( [ eq20 ] ) can be expressed mathematically by introducing a function @xmath64 @xmath65 with the kernel of @xmath64 @xmath66 being the set of all integer multiplets @xmath67 for which @xmath68 and therefore the constraint in equation ( [ eq20 ] ) is fulfilled . using this definition , equation ( [ eq18 ] ) simplifies to @xmath69 using the result of equation ( [ eq13 ] ) and ( [ eq23 ] ) , the second - order correlation function in equation ( [ eq9 ] ) now becomes @xmath70 with @xmath71 } ~. \label{eq26}\end{aligned}\ ] ] to calculate a more descriptive representation of the correlation function and to demonstrate , that @xmath72 , equation ( [ eq24 ] ) , ( [ eq25 ] ) and ( [ eq26 ] ) can be further rewritten in terms of real and imaginary parts @xmath73 \nonumber\\ = 1 + & \frac{k^2}{2}\sum_{\substack{\left\{n_j , m_j\right\}\in \ker(c)\\ j=1\ldots n}}\tilde{b}_{\{n_j , m_j\}}\left(\varphi_j\right)\cos\left(ku+\tilde{\varphi}_{\{n_j , m_j\}}\right)\cdot \\ \nonumber & \cdot\biggl[\cos\biggl(\sum_{j=1}^n m_j\omega_j\tau+\phi_{\{n_j , m_j\}}\biggr)+i\cdot\sin\biggl(\sum_{j=1}^n m_j\omega_j\tau+\phi_{\{n_j , m_j\}}\biggr)\biggr ] ~,\end{aligned}\ ] ] with the product of the bessel functions @xmath74 the spatial correlation phase @xmath75 and temporal phase @xmath76 in equation ( [ eq27 ] ) are given by @xmath77 if the constraint in equation ( [ eq20 ] ) is fulfilled for a multiplet @xmath78 , it is also satisfied for @xmath79 . with @xmath80 , @xmath81 and @xmath82 ,
it can be shown , that the addend with @xmath79 of the sum in equation ( [ eq27 ] ) is complex conjugated to the addend with @xmath78 .
the addend of the zero multiplet @xmath83 is purely real valued .
therefore , the imaginary part vanishes after summing up all addends and equation ( [ eq27 ] ) becomes real @xmath84 with the amplitude @xmath85 here , the sum has to be taken over all integer multiplets @xmath86 fulfilling the constraint in equation ( [ eq20 ] ) . in principle
, the constraint is satisfied for an infinite number of multiplets each with their own contribution to the correlation function given by @xmath87 however , contributions with large values of @xmath78 are suppressed , because the bessel function ( equation ( [ eq28 ] ) ) strongly decays for @xmath88 @xcite .
this limits the number of multiplets that have to be taken into account for the correlation analysis .
the contribution to @xmath89 of each multiplet addend @xmath90 shows a periodic modulation in the correlation length @xmath91 with the same spatial periodicity @xmath92 as the unperturbed interference pattern , but shifted in @xmath91-direction by the spatial correlation phase @xmath75 . the amplitude @xmath93 ( equation ( [ eq31 ] ) ) of each multiplet addend @xmath90 depends on the correlation time @xmath94 and the specific perturbation characteristics .
it shows a periodic structure in @xmath94 with the periodicity determined by the frequency component @xmath95 with the coefficients @xmath96 .
this periodic structure is shifted in @xmath94-direction by the temporal phase @xmath76 .
the amplitude of the modulation in @xmath94-direction is given by the peak phase deviations @xmath9 , via the product of the bessel functions in equation ( [ eq28 ] ) .
after summing up all addends @xmath90 , the resulting correlation function ( equation ( [ eq30 ] ) ) shows the same spatial periodicity @xmath4 .
the overall amplitude is equal to 1 only at certain correlation times @xmath94 given by the involved perturbation frequencies ( see section [ sec2.4 ] ) . at these temporal positions ,
the contrast of the correlation function is @xmath97 and therefore directly linked to the contrast of the unperturbed interference pattern @xmath2 . for all other correlation times ,
the contrast of the correlation function is @xmath98 .
it has to be noted , that the maximum contrast of the correlation function is @xmath97 and therefore a factor of @xmath99 lower than the contrast @xmath2 of the unperturbed interference pattern .
the overall amplitude of the resulting correlation function includes the perturbation frequencies @xmath7 , their harmonic frequencies as well as their differences and sums ( intermodulation terms ) .
all frequency components are given by the argument of the cosine @xmath95 in equation ( [ eq31 ] ) .
approximately , the maximum frequency component per perturbation frequency included in the correlation function is given by @xmath100 , with @xmath101 , as larger frequency components are suppressed due to the strong decay of the bessel function in equation ( [ eq28 ] ) for @xmath88 .
therefore , the maximum frequency component of all perturbation frequencies included in the correlation function is approximately given by @xmath102 . in the following ,
an approximate solution for the correlation function is deduced , by taking into account only the trivial solution to the constraint in equation ( [ eq20 ] ) .
these are the multiplets with @xmath61 , which typically give the main contribution to the correlation function .
using @xmath103 , @xmath104 and @xmath105 , equation ( [ eq30 ] ) and ( [ eq31 ] ) become @xmath106 with @xmath107 and @xmath108 , the approximate second - order correlation function yields @xmath109 with the time - dependent amplitude @xmath110 similar to the explicit solution of the correlation function in equation ( [ eq30 ] ) and ( [ eq31 ] ) , the spatial modulation of the approximate solution is given by the spatial periodicity @xmath4 of the unperturbed interference pattern .
the approximate solution is independent of the spatial correlation phases @xmath75 and temporal phases @xmath76 ( equation ( [ eq29 ] ) ) .
therefore , the addends are not phase shifted with respect to each other in @xmath91- and @xmath94-direction .
usually , the approximate correlation function in equation ( [ eq34 ] ) can be used for the description of multifrequency perturbations . however , in the case of few perturbation frequencies , that are multiples of each other , the constraint in equation ( [ eq20 ] ) is additionally fulfilled for @xmath63 and the explicit solution of the correlation function in equation ( [ eq30 ] ) has to be applied . in the case of a single perturbation frequency ,
the constraint of equation ( [ eq20 ] ) is only satisfied for @xmath111 .
thus , the explicit and approximate solution of the correlation function are identical .
the determination of the contrast @xmath2 and spatial periodicity @xmath4 from the correlation function shall be discussed now . both can only be correctly obtained at certain correlation times . as mentioned in section
[ sec2.2 ] and [ sec2.3 ] , the explicit and approximate solution of the correlation function show a periodic modulation in @xmath91-direction having the same periodicity @xmath4 as the unperturbed interference pattern .
the overall amplitude of this modulation ( after summing up all multiplet addends ) depends on the correlation time @xmath94 resulting from the specific perturbation spectrum for both correlation functions ( equation ( [ eq30 ] ) and ( [ eq34 ] ) ) .
its maximum value of 1 is achieved at @xmath112 , with the superperiod @xmath113 given by the reciprocal value of the greatest common divisor of all perturbation frequencies @xmath114 . for each frequency @xmath7
there is an integer @xmath115 for which @xmath116 . at the temporal positions
@xmath117 , only the addends with @xmath61 sum up to a maximum value of 1 , because then the addends in equation ( [ eq30 ] ) are not phase shifted in @xmath91- and @xmath94-direction with respect to each other .
the sum of the addends with @xmath63 is equal to zero at these temporal positions .
therefore , the exact and approximate solution are identical at correlation times @xmath118 . using equation ( [ eq34 ] ) and ( [ eq35 ] )
, the correlation function then becomes @xmath119 which is suitable to obtain the contrast @xmath2 and pattern periodicity @xmath92 of the unperturbed interference pattern .
if there is no greatest common divisor @xmath120 , the superperiod @xmath121 is infinite and the only position , where the amplitudes have a maximum value of @xmath122 , is at @xmath46 .
therefore , the determination of the contrast and spatial periodicity using equation ( [ eq36 ] ) can always be applied to the correlation function at the correlation time of @xmath46 .
the properties of the explicit and approximate correlation function , discussed in section [ sec2.2 ] and [ sec2.3 ] , are illustrated below for single- and two - frequency perturbations .
the commonalities and differences of both solutions are pointed out . in the case of one perturbation frequency @xmath123 ( @xmath11 ) ,
the constraint in equation ( [ eq20 ] ) is only fulfilled for @xmath124 .
thus , the explicit and approximate solution are identical .
a correlation function of an interference pattern with @xmath125 , @xmath126{mm}$ ] , a perturbation frequency @xmath127{hz}$ ] and a peak phase deviation @xmath26{\pi}$ ] is calculated according to equation ( [ eq34 ] ) and ( [ eq35 ] ) for @xmath11 and plotted in figure [ fig2 ] at the top , showing a clear periodic structure in the correlation length @xmath91 and time @xmath94 . as seen from equation ( [ eq34 ] ) , the spatial periodicity of the unperturbed interference pattern is recovered in the correlation function as modulation along @xmath91 . at each @xmath94 , however , the contrast is reduced to @xmath128 .
as seen from equation ( [ eq35 ] ) and figure [ fig2 ] at the bottom , the amplitude @xmath129 reaches its maximum at correlation times that are multiples of the perturbation time period @xmath130{ms}$ ] . here
, @xmath129 becomes 1 and the contrast in the correlation function @xmath97 is directly linked to the contrast of the unperturbed interference pattern . according to equation ( [ eq7 ] ) and figure [ fig1 ] , the integrated interference pattern would be completely washed - out " for these perturbation parameters .
using correlation theory , however , the contrast and spatial periodicity can be unveiled as described in section [ sec2.4 ] . ,
@xmath126{mm}$ ] and a perturbation consisting of @xmath127{hz}$ ] , @xmath26{\pi}$ ] calculated with equation ( [ eq34 ] ) and ( [ eq35 ] ) for @xmath11 .
bottom : correlation amplitude depending on the correlation time @xmath94 and the perturbation parameters resulting from equation ( [ eq35 ] ) .
it has its maximum value of 1 at the temporal coordinates @xmath131{ms}$ ] , with @xmath132.,scaledwidth=50.0% ] in the case of perturbation frequencies , that are multiples of each other , the explicit solution of the correlation function has to be used .
this shall be demonstrated for two frequencies with @xmath133 . here ,
equation ( [ eq20 ] ) becomes @xmath134 and the constraint is satisfied not only for integer multiplets with @xmath135 , but also for @xmath136 , as @xmath137 .
these terms lead to additional contributions to the explicit correlation function in equation ( [ eq30 ] ) , causing a spatial and temporal phase shift due to the not vanishing phases @xmath75 and @xmath76 ( equation ( [ eq29 ] ) ) .
therefore , the approximate solution of equation ( [ eq34 ] ) is not suitable , and the explicit solution has to be used . in figure [ fig3](a ) ,
a second - order correlation function calculated with the explicit solution ( equation ( [ eq30 ] ) and ( [ eq31 ] ) ) is shown for an interference pattern with a contrast of @xmath125 and a spatial periodicity @xmath126{mm}$ ] perturbed with @xmath138{hz } , \varphi_1=\unit[0.5]{\pi } , \phi_1=\unit[0.25]{\pi}$ ] and @xmath139{hz } , \varphi_2=\unit[0.5]{\pi } , \phi_2=\unit[-0.25]{\pi}$ ] . the approximate solution ( equation ( [ eq34 ] ) and ( [ eq35 ] ) )
including only the @xmath61 terms can be seen in figure [ fig3](b ) and differs clearly from the exact solution in ( a ) .
both correlation functions , figure [ fig3](a ) and ( b ) , show a periodic modulation in @xmath91-direction having the same periodicity @xmath4 as the unperturbed interference pattern .
the superperiod @xmath140{ms}$ ] can be identified in both correlation functions originating from the greatest common divisor of the involved perturbation frequencies @xmath141{hz}$ ] . at correlation times , which are multiples of this superperiod , the explicit and approximate correlation function
are identical and the amplitudes achieve their maximum value of 1 ( see section [ sec2.4 ] ) . the difference between the exact and approximate solution , which is solely given due to the terms with @xmath63 ,
is illustrated in figure [ fig3](c ) .
its contribution to the correlation function in figure [ fig3](a ) leads to the correlation time dependent phase shift of the pattern in @xmath91-direction originating from the spatial correlation phases @xmath75 and temporal phases @xmath76 . of an interference pattern with a contrast of @xmath125 , a spatial periodicity @xmath126{mm}$ ]
perturbed with two frequencies @xmath138{hz } , \varphi_1=\unit[0.5]{\pi } , \phi_1=\unit[0.25]{\pi}$ ] and @xmath139{hz } , \varphi_2=\unit[0.5]{\pi } , \phi_2=\unit[-0.25]{\pi}$ ] calculated according to equation ( [ eq30 ] ) and ( [ eq31 ] ) .
b ) approximate solution calculated with equation ( [ eq34 ] ) and ( [ eq35 ] ) including only the @xmath61 terms .
c ) difference between explicit and approximate solution solely given by the terms with @xmath63.,scaledwidth=100.0% ] { hz } , \varphi_1=\unit[0.5]{\pi } , \phi_1=\unit[0.25]{\pi}$ ] and @xmath142 with @xmath143{\pi } , \phi_2=\unit[-0.25]{\pi}$ ] as calculated according to equation ( [ eq30 ] ) .
b ) approximate solution of the second - order correlation function calculated with equation ( [ eq34 ] ) .
c ) difference between explicit and approximate solution given by the terms with @xmath63 .
the structure in c ) is almost vanished for @xmath144 and disappeared for @xmath145 . at this transition point , the correlation functions in a ) and b ) are identical and the approximate solution is suitable for the analysis.,scaledwidth=100.0% ] as discussed in section [ sec2.3 ] , the explicit and approximate solution ( equation ( [ eq30 ] ) and ( [ eq34 ] ) ) are identical in the case of one perturbation frequency ( @xmath11 ) and also in the case of numerous frequencies @xmath7 , that are not multiples of each other . a common scenario , when both solutions differ from each other , is the case when the frequencies are multiples of each other , e.g. for @xmath6 perturbation frequencies @xmath146 with @xmath147 and @xmath148 . here , it shall be shown , that also in this case the approximate solution might be suitable to describe the correlation function .
the explicit solution turns into the approximate , if the largest contributing addend in equation ( [ eq30 ] ) for @xmath63 is small compared to the largest addend with @xmath61 .
the amplitudes of all addends are given by the peak phase deviations @xmath9 , via the product of the bessel functions in equation ( [ eq28 ] ) .
the maximum value of the bessel function @xmath149 is approximately achieved , if the order is @xmath150 for @xmath151 and @xmath152 otherwise .
therefore , the amplitude of the largest addend with @xmath61 is given by @xmath153 using equation ( [ eq20 ] ) , the constraint for the integer multiplets @xmath154 can be written as @xmath155 this constraint is generally satisfied for the multiplet with @xmath63 , resulting in the largest contributing integer multiplet @xmath156 using this multiplet and @xmath150 for the order of the bessel function , the amplitude ( equation ( [ eq28 ] ) ) of the largest addend with @xmath63 contributing to the correlation function is @xmath157 here , the variable factor @xmath158 is given by @xmath159 and determines the value of @xmath160 .
the ratio of equation ( [ eq37 ] ) and ( [ eq40 ] ) is @xmath161 a closer look to the bessel function unveils that @xmath158 decays rapidly for @xmath162 , which can be seen with the asymptotic form of the bessel function for @xmath163 @xcite @xmath164 with @xmath165 denoting the gamma function .
therefore , the explicit solution approaches rapidly to the approximate solution , once @xmath166 .
figure [ fig4 ] shows an example for a two frequency perturbation with @xmath167 . for @xmath168{\pi}$ ] , the explicit solution should closely approach the approximate once @xmath169 , as can be seen in figure [ fig4 ] . for @xmath144 , the structure in figure [ fig4](c )
is almost vanished . in this case
, the leading term of equation ( [ eq40 ] ) is about seven times smaller than @xmath170 in equation ( [ eq37 ] ) .
the structure in figure [ fig4](c ) disappears for @xmath145 . here
, @xmath160 is twenty - four times smaller than @xmath170 and has a negligible contribution to the correlation function .
therefore , the correlation functions in figure [ fig4](a ) and ( b ) for @xmath145 are identical , proving the approximate solution to be suitable for the correlation analysis . as it might be difficult to identify the involved perturbation frequencies from the correlation function ( c.f .
figure [ fig3 ] ) , the amplitude spectrum of the correlation function can be calculated and used for the determination of the perturbation characteristics @xcite . this is possible , as the fourier transform of the correlation function equals the power spectrum of the perturbed measurement signal according to the wiener - khintchine theorem @xcite .
therefore , the applied perturbation frequencies can be identified in the amplitude spectrum of @xmath89 . to determine the frequency components and their amplitudes , the temporal fourier transform @xmath171 of equation ( [ eq30 ] )
is calculated @xmath172 with the fourier transformed amplitude of equation ( [ eq31 ] ) @xmath173~.\end{aligned}\ ] ] using @xmath174 and @xmath175 , the amplitude in equation ( [ eq44 ] ) results in @xmath176 where @xmath177 are the frequency components and @xmath178 the dirac delta function .
the resulting amplitude spectrum @xmath179 calculated from equation ( [ eq43 ] ) and ( [ eq45 ] ) is given by @xmath180 as for the explicit correlation function in equation ( [ eq30 ] ) , the sum in equation ( [ eq46 ] ) is taken over the integer multiplets @xmath86 resulting from the constraint in equation ( [ eq20 ] ) .
each addend shows a periodic structure in the correlation length @xmath91 with the same spatial periodicity @xmath4 as the unperturbed interference pattern , but shifted in @xmath91-direction by the spatial correlation phase @xmath75 ( equation ( [ eq29 ] ) ) .
the amplitude of this modulation depends on the peak phase deviation @xmath9 , via the product of the bessel functions in equation ( [ eq28 ] ) modified by the cosine of the temporal phase @xmath76 ( equation ( [ eq29 ] ) ) .
the frequency position of the modulation is given by the frequency component @xmath181 with the coefficients @xmath182 .
the summed up amplitude spectrum contains all frequency components up to a maximum frequency @xmath102 resulting from the strong decay of the bessel function as discussed for the explicit solution of the correlation function ( equation ( [ eq30 ] ) ) . of a perturbation with @xmath138{hz } , \varphi_1=\unit[0.5]{\pi } , \phi_1=\unit[0.25]{\pi}$ ] and @xmath139{hz } , \varphi_2=\unit[0.5]{\pi } , \phi_2=\unit[-0.25]{\pi}$ ] calculated according to equation ( [ eq46 ] ) .
b ) approximate solution of the amplitude spectrum calculated using equation ( [ eq47 ] ) .
c ) contribution of the addends with @xmath63 to the amplitude spectrum in a).,scaledwidth=100.0% ] the amplitude spectrum of the approximate second - order correlation function in equation ( [ eq34 ] ) is calculated from equation ( [ eq46 ] ) using @xmath61 and reads @xmath183 as with the approximate solution , also the amplitude spectrum is independent from the spatial correlation phases @xmath75 and temporal phases @xmath76 .
therefore , the spatial periodicity in the amplitude spectrum is directly given by the periodicity of the unperturbed interference pattern .
the amplitude of the modulation in @xmath91-direction is now solely given by the peak phase deviations @xmath9 , via the square of the bessel functions .
the amplitude spectrum corresponding to the second - order correlation function of the two frequency perturbation shown in figure [ fig3](a ) is calculated according to equation ( [ eq46 ] ) and plotted in figure [ fig5](a ) for positive frequency components .
the amplitude spectrum of the approximate solution of equation ( [ eq47 ] ) is plotted in figure [ fig5](b ) . in figure [ fig5](c ) ,
the contribution of the addends with @xmath63 to the explicit solution is shown . for each frequency component , given by @xmath184 with @xmath185
, the modulation of the unperturbed interference pattern with the periodicity @xmath4 can be identified in the spatial direction @xmath91 . in figure [ fig5](c ) , the phase shift of the interference pattern in the spatial direction is caused by the spatial phases @xmath75 . in figure [ fig5](b ) ,
the interference pattern is not shifted in the spatial direction @xmath91 and the amplitudes of the involved frequency components only depend on the peak phase deviations @xmath9 via the square of the bessel functions in equation ( [ eq47 ] ) . as for the transition between the explicit and approximate correlation function discussed in section [ sec2.6 ] , the amplitude spectrum of the explicit correlation function ( equation ( [ eq46 ] ) ) turns into the approximate solution ( equation ( [ eq47 ] ) ) , if the largest contributing addend in equation ( [ eq40 ] ) is small compared to equation ( [ eq37 ] ) .
the identification of unknown perturbation frequencies from the correlation function is difficult , because the structure of the correlation function does not reveal the frequencies directly ( c.f .
figure [ fig3 ] ) .
the amplitude spectrum offers a better access to the frequency components included in the correlation function . in @xcite ,
a numerical algorithm is described , which allows to identify unknown perturbation frequencies .
after the determination of the perturbation frequencies @xmath7 , equation ( [ eq46 ] ) or ( [ eq47 ] ) is applied to obtain the corresponding peak phase deviations @xmath9 and phases @xmath10 .
the latter are not uniquely determined , as the correlation function is invariant under phase transformations .
however , it is possible to identify the original phases by the reconstruction of the unperturbed interference pattern with the obtained perturbation frequencies and amplitudes .
as discussed in section [ sec2.4 ] , the contrast and pattern periodicity of the unperturbed interference pattern can be unveiled from the second - order correlation function . using the amplitude spectrum , also
the perturbation frequencies , amplitudes and phases can be determined ( section [ sec2.7 ] ) .
however , the perturbation phases can not be uniquely obtained . for a single perturbation frequency , for instance
, the phase information is completely lost in the correlation function . however , for larger numbers of perturbation frequencies ( @xmath29 ) , the phase informations are not completely lost . for @xmath29 , only relative phase information can be unveiled . in this context , it is useful to understand under which phase transformation @xmath186 the correlation function stays invariant . as only time and position differences between particle events
are taken into account , the second - order correlation function is invariant under position shifts @xmath187 and time transformations @xmath188 . using equation ( [ eq1 ] ) and ( [ eq2 ] ) such a time transformation and
position shift equals transformations @xmath189 furthermore , the correlation function is invariant under simultaneous time and space reversal : @xmath190 , because only time and space differences are taken into account . according to equation ( [ eq1 ] ) and ( [ eq2 ] ) , this transformation equals a phase transformation @xmath191 hence , under above phase transformations ( equation ( [ eq49 ] ) and ( [ eq50 ] ) ) the correlation function stays unchanged .
exemplary , the invariance is calculated for a perturbation with two frequencies ( @xmath192 ) and phases @xmath193 and @xmath194 . according to equation ( [ eq49 ] ) ,
the correlation function is invariant for following phase transformations @xmath195 eliminating @xmath196 yields @xmath197 every phase duplet @xmath198 described by this equation , yields the same correlation function as the original phase duplet @xmath199 .
therefore , the absolute phases @xmath193 and @xmath194 can not be uniquely determined by the correlation analysis .
a second set of phase duplets @xmath200 with identical correlation functions can be found after evaluating the phase transformation for time and space reversal from equation ( [ eq50 ] ) @xmath201 yielding together with equation ( [ eq52 ] ) @xmath202 as result , there are two linear equations ( equation ( [ eq52 ] ) and ( [ eq54 ] ) ) , that describe the same correlation function .
it is thus not possible to determine the original phases @xmath193 and @xmath194 from the correlation analysis .
although the original phases @xmath10 can not be determined by the correlation analysis , the identification is possible by the reconstruction of the unperturbed interference pattern using the obtained perturbation parameters @xmath7 and @xmath9 @xcite @xmath203 with the spatial coordinate of the reconstructed interference pattern @xmath204 , the time - dependent perturbation @xmath205 ( equation ( [ eq2 ] ) ) and @xmath206 , @xmath207 the spatial and temporal coordinate of particle @xmath208 forming the perturbed interference pattern .
the phases @xmath10 are varied until the maximum contrast of the reconstructed interference pattern is achieved . using equation ( [ eq52 ] ) or ( [ eq54 ] ) in the reconstruction process
, the free parameter space for the phases @xmath10 can be reduced by one dimension .
in this chapter , the characteristics of the numerical second - order correlation function used for the evaluation of the experimental data are analyzed in detail .
first , the numerical correlation function is derived and investigated how the discretization affects the obtained parameters as contrast and perturbation amplitude .
a theoretical description of the influence on the correlation analysis and the effect of noise is presented .
this theory is numerically cross - checked by simulations of single - particle interference .
afterwards , the possibility to use the correlation analysis for broad - band frequency noise is demonstrated by a simulation with a gaussian distributed noise . at the end of this chapter ,
the application limits of the second - order correlation theory are discussed .
so far , the correlation function has been calculated from the analytic probability distribution of particle impacts at the detector . in reality
, however , the detector yields a discrete signal of particle impacts for a finite acquisition time @xmath37 .
the corresponding second - order correlation function @xmath89 is thus extracted from this detector signal @xmath209 consisting of @xmath6 particle arrival times @xmath207 and coordinates @xmath206 , ( @xmath210 ) @xmath211 here , @xmath212 denotes the dirac delta function , which assures normalization of @xmath209 @xmath213 the approximation of the particle impacts in the detection plane by the dirac delta function is suitable as the pulse width at the detector is much smaller than the mean time distance between particle impacts . using this definition of the detector signal , the correlation function is calculated according to equation ( [ eq9 ] ) . as before , the denominator is calculated first @xmath214 with the acquisition time @xmath37 and length @xmath38 . the second term in the denominator can be determined from equation ( [ eq58 ] ) for @xmath45 and @xmath46 , yielding @xmath215 . using equation ( [ eq56 ] ) ,
the numerator in equation ( [ eq9 ] ) becomes @xmath216 with the results of equation ( [ eq58 ] ) and ( [ eq59 ] ) , the discrete second - order correlation function is given by @xmath217 to achieve a proper correlation function , which can be compared with the analytic theory , it is necessary to implement a temporal and spatial discretization step size @xmath218 and @xmath219 in equation ( [ eq60 ] ) , yielding @xmath220 with @xmath221 denoting the number of particle pairs @xmath222 with distances @xmath223 $ ] and time separations @xmath224 $ ] . due to the finite acquisition time
@xmath37 and length @xmath38 , the probability to detect particle impacts with large temporal and spatial differences is reduced by a factor of @xmath225 and @xmath226 , respectively .
therefore , the number @xmath221 in equation ( [ eq61 ] ) has to be corrected for this effect , resulting in @xmath227 as with the analytic correlation function , the numerical correlation function in equation ( [ eq62 ] ) is normalized to one .
and @xmath218 extracted according to equation ( [ eq62 ] ) from a single - particle simulation of an interference pattern with a contrast @xmath125 and a spatial periodicity @xmath126{mm}$ ] perturbed with @xmath138{hz}$ ] and @xmath228{\pi}$ ] .
a ) the structure is smeared out " , because the discretization step size is too large .
b ) compared to a ) , the structure is clearly visible , but the noise increases with smaller @xmath219 and @xmath218 .
c ) for even smaller discretization step size , the noise grows larger.,scaledwidth=100.0% ] three correlation functions with different spatial and temporal discretization step sizes are illustrated in figure [ fig6 ] .
they are extracted using equation ( [ eq62 ] ) from a single - particle simulation of an interference pattern ( section [ sec3.3 ] ) with a contrast @xmath125 and a spatial periodicity @xmath126{mm}$ ] perturbed with @xmath138{hz}$ ] and @xmath228{\pi}$ ] .
if @xmath219 and @xmath218 are too large , the structure in the correlation function gets
smeared out " as can be seen in figure [ fig6](a ) .
in contrast , the structure is clearly visible in figure [ fig6](b ) .
however , the noise in the correlation function increases compared to figure [ fig6](a ) . this noise grows larger for even smaller discretization step size as seen in figure [ fig6](c ) .
therefore , it is necessary to understand the effects of discretization and choose @xmath219 and @xmath218 in an appropriate way . in this section
, the effects of the discretization step size on the correlation analysis and noise in the correlation function are analyzed .
therefore , the influence on the contrast and peak phase deviation due to the discretization of the correlation function is derived for one perturbation frequency .
the theory reveals how the spatial and temporal discretization step size have to be chosen .
afterwards , a theoretical description for the noise in the correlation function and the corresponding amplitude spectrum is derived .
thereby , an optimum spatial discretization step size is found , resulting in a maximum signal - to - noise ratio . at the end , an estimation of the smallest detectable peak phase deviation
is given . to study
the influence of the spatial and temporal discretization step size @xmath219 and @xmath218 , the analytic solution from equation ( [ eq34 ] ) is used . for simplicity ,
only one perturbation frequency @xmath123 is taken into account .
discretizing the correlation function into temporal and spatial intervals @xmath218 and @xmath219 , then yields @xmath229 with the @xmath230-function @xmath231 and the amplitude @xmath232 for @xmath233 and @xmath234 , equation ( [ eq63 ] ) and ( [ eq64 ] ) yield the analytic solution of the correlation function in equation ( [ eq34 ] ) and ( [ eq35 ] ) for @xmath11 . as discussed in section [ sec2.4 ] , the contrast @xmath2 and spatial periodicity @xmath4 of the unperturbed interference pattern
are determined from the correlation function at @xmath46 ( equation ( [ eq36 ] ) ) . for this correlation time , equation ( [ eq63 ] ) and ( [ eq64 ] ) become @xmath235 with @xmath236 thus , the spatial periodicity @xmath92 is not influenced by the discretization step size , but the extracted contrast @xmath237 is modified to @xmath238 this contrast is zero for multiples of @xmath239 and @xmath240 , because the @xmath230-function is zero for this values . for @xmath241 and @xmath242 ,
the @xmath230-functions in equation ( [ eq66 ] ) and ( [ eq67 ] ) approach to 1 .
full contrast @xmath2 is only achieved for @xmath243 .
the dependence of the extracted contrast @xmath237 on @xmath218 and @xmath219 is illustrated in figure [ fig7](a ) for a peak phase deviation of @xmath228{\pi}$ ] .
the dependence of the extracted contrast on the temporal discretization step size @xmath218 and the peak phase deviation @xmath23 is shown in figure [ fig7](b ) for @xmath244 .
{ \pi}\right)/k$ ] on the spatial and temporal discretization step size @xmath219 and @xmath218 for one perturbation frequency @xmath123 with a peak phase deviation of @xmath228{\pi}$ ] .
b ) dependence of @xmath245 on the temporal discretization @xmath218 and peak phase deviation @xmath23 for @xmath244.,scaledwidth=90.0% ] to determine the perturbation frequency @xmath123 and corresponding peak phase deviation @xmath23 from the correlation function , the analytic amplitude spectrum of equation ( [ eq47 ] ) is used . for discrete signals ,
the amplitude spectrum also depends on @xmath219 and @xmath218 .
this dependence is calculated from equation ( [ eq63 ] ) and ( [ eq64 ] ) with a temporal fourier transformation , yielding @xmath246 with the frequency components @xmath247 .
comparing equation ( [ eq68 ] ) with the analytic solution in equation ( [ eq47 ] ) for @xmath11 , it can be seen , that the frequency components @xmath247 are not changed , but their amplitudes @xmath248 are modified in equation ( [ eq68 ] ) due to the spatial and temporal discretization .
this dependence is calculated for the amplitude of the fundamental frequency component in equation ( [ eq68 ] ) with @xmath249 and @xmath45 , resulting in @xmath250 with the modified amplitude @xmath251 of the fundamental frequency component @xmath123 depending on @xmath219 and @xmath218 @xmath252 here , the amplitude of the fundamental frequency component @xmath253 is reduced by the factor of @xmath254 , which needs to be taken into account for the determination of @xmath23 . therefore , the peak phase deviation @xmath255 extracted from the amplitude spectrum of the correlation function also depends on the discretization step size via the square of the bessel function . for @xmath242 and @xmath241 , equation ( [ eq69 ] )
results in @xmath256 yielding the correct peak phase deviation @xmath23 .
as discussed in section [ sec3.1 ] , a real experiment requires the correlation function to be derived from a finite number of detection events .
more precisely , it is calculated from the number of correlated particle pairs @xmath221 within a given correlation window @xmath218 and @xmath219 . due to the statistical nature of the particles ,
@xmath221 is subject to poissonian noise , which is transferred onto the correlation function and the corresponding amplitude spectrum . in principle
, this limits the signal - to - noise ratio and the minimal detectable perturbation amplitude . in the following section , this effect is estimated and optimal settings for the discretization step size are found .
therefore , all fluctuating variables @xmath257 are described by their corresponding mean values @xmath258 and variances @xmath259 , with @xmath260 being the standard deviation @xmath261 for simplicity reasons the analysis is restricted to correlation times @xmath262 and positions @xmath263 . with the correlation function from equation ( [ eq62 ] ) being normalized to 1 , it can be found @xmath264 the expected standard deviation can then be calculated yielding @xmath265 with @xmath221 following a poissonian distribution , variance and mean value are directly linked @xmath266 yielding @xmath267 as expected , the variance ( noise ) of the correlation function depends on the total number of detected particles and the number of bins @xmath268 and @xmath269 in the temporal and spatial direction . before calculating , how the noise in the correlation function transfers onto its amplitude spectrum , the correlation function is split in two parts : the first part @xmath270 describing the ideal correlation function , as expected in the limit of infinite detection events and the second part @xmath271 describing the noise only @xmath272 obviously , mean values and standard deviations of these functions are given to @xmath273 following this description , the amplitude spectrum of the correlation function reads @xmath274 with @xmath275 and @xmath276 denoting the discrete fourier transforms of the time discrete signals @xmath277 and @xmath271 @xmath278 here , @xmath279 denotes the maximum correlation time up to which the correlation function is evaluated .
the noise in the spectrum is thus solely included in @xmath280 .
using parsival s theorem @xcite together with @xmath281 , it can be found @xmath282 and thus a direct link between the noise in the power spectrum and the noise in the correlation function ( equation ( [ eq75 ] ) ) @xmath283 the signal - to - noise ratio @xmath284 of the amplitude spectrum used for the determination of the peak phase deviation @xmath23 can be calculated with equation ( [ eq69 ] ) , ( [ eq70 ] ) and ( [ eq81 ] ) yielding @xmath285 with @xmath286 for @xmath241 , the @xmath230-function approaches to 1 and the signal - to - noise ratio only depends on the spatial discretization step size @xmath219 .
the function @xmath287 has a global maximum at the position @xmath288 with the maximum value of @xmath289 .
hence , the optimum spatial discretization step size , leading to a maximum of the signal - to - noise ratio , becomes @xmath290 the signal - to - noise ratio calculated with equation ( [ eq82 ] ) is plotted in figure [ fig8 ] for different spatial and temporal discretization step sizes .
the optimum spatial discretization step size is clearly visible at @xmath291 . from equation ( [ eq82 ] ) and
( [ eq83 ] ) , the optimum signal - to - noise ratio is deduced for @xmath292 and @xmath241 @xmath293 a lower limit for the identification of small peak phase deviations @xmath294 can be derived from equation ( [ eq85 ] ) by setting @xmath295 equal to 1 .
this sets the threshold at which the noise and signal have equal amplitude . with @xmath296 and equation ( [ eq85 ] ) , this yields @xmath297 for the measurement of small peak phase deviations it is thus favourable to have an interference pattern with large contrast @xmath2 and pattern periodicity @xmath4 .
furthermore , a large number of particles @xmath6 decreases the minimum detectable peak phase deviation . ) and
normalized to @xmath298 ( equation ( [ eq83 ] ) ) and @xmath289 .
the optimum spatial discretization can be identified at @xmath291.,scaledwidth=50.0% ] to numerically cross - check the theoretical calculations in section [ sec3.2 ] , a set of @xmath6 particles with temporal and spatial coordinates
@xmath299 have been simulated .
therefore , time and position coordinates are generated according to the corresponding distribution functions using the acceptance - rejection method @xcite .
for the temporal coordinates , the probability distribution of time differences @xmath300 between successive events is used , which for poisson statistics is given by @xcite @xmath301 here , @xmath302 denotes the mean count rate . following this distribution function , a set of @xmath303 time differences @xmath304
is generated .
starting with the first event at @xmath305 , the time steps for successive events are given by @xmath306 after having generated all time coordinates @xmath207 , the corresponding spatial coordinates @xmath206 are created according to the probability distribution @xmath307 with the time - dependent perturbation @xmath308 from equation ( [ eq2 ] ) .
this results in a full set of time and position coordinates . and a fixed temporal discretization step size @xmath309 .
they are extracted from a single - particle simulation of an interference pattern with @xmath125 and @xmath126{mm}$ ] perturbed with @xmath138{hz}$ ] and @xmath310{\pi}$ ] used in equation ( [ eq89 ] ) .
the spatial discretization step size is increased from a ) to c ) and therefore the structure of the correlation function is smeared out " until it is vanished in c ) .
otherwise , the noise decreases for larger @xmath311.,scaledwidth=100.0% ] such a simulation has been made for an interference pattern with a contrast of @xmath125 , a spatial periodicity of @xmath126{mm}$ ] and a single frequency perturbation @xmath312 with @xmath138{hz}$ ] and @xmath310{\pi}$ ] . with an acquisition time of @xmath313{s}$ ] and length @xmath314{mm}$ ] , @xmath315 particles have been simulated .
this parameters have been chosen in accordance to typical experimental parameters @xcite .
the correlation function was extracted from the simulated data according to equation ( [ eq62 ] ) for different spatial discretization step sizes ( @xmath219 ) , a temporal discretization step size ( @xmath316{ms}$ ] ) and a maximum correlation time of @xmath317{s}$ ] .
the temporal discretization step size is chosen to ensure that it does not reduce the signal - to - noise ratio ( c.f .
figure [ fig8 ] for @xmath318 ) .
three correlation functions for different spatial discretization step sizes @xmath219 are illustrated in figure [ fig9 ] . with increasing @xmath219 from figure [ fig9](a ) to ( c ) ,
the structure in the correlation function is smeared out " .
otherwise , the noise decreases for a larger spatial discretization step size , because the mean particle number increases ( equation ( [ eq75 ] ) ) .
the relation between this two effects leads to an optimum @xmath319 ( equation ( [ eq84 ] ) ) that provides a maximum of the signal - to - noise ratio ( see equation ( [ eq85 ] ) ) . as the correlation functions in figure [ fig9 ] ,
many correlation functions with different spatial discretisation step sizes ranging from @xmath320 to @xmath321 are extracted from the above simulation according to equation ( [ eq62 ] ) . for each correlation function , the amplitude spectrum was calculated at the spatial position @xmath322{mm}$ ] using a numerical fourier transformation .
the signal height of the peak at the @xmath323{hz}$ ] position and the standard deviation of the noise were determined in each spectrum .
the result for the extracted signal height is plotted in figure [ fig10](a ) ( blue dots ) .
the standard deviation of the noise is shown in [ fig10](b ) and the signal - to - noise ratio in figure [ fig10](c ) .
the theoretical curve in figure [ fig10](a ) ( red solid line ) was calculated using equation ( [ eq69 ] ) with the parameters of the simulation .
for @xmath324 , the signal height is significantly reduced , because the structure in the correlation functions begins to smear out " ( see figure [ fig9](b ) ) until it is totally vanished for @xmath325 ( see figure [ fig9](c ) ) . the theoretical curve in figure [ fig10](b )
is evaluated with equation ( [ eq81 ] ) indicating that the noise is reduced for larger @xmath311 due to the increasing mean particle number @xmath326 in equation ( [ eq75 ] ) . the theoretical signal - to - noise ratio in figure [ fig10](c )
is calculated according to equation ( [ eq82 ] ) and ( [ eq83 ] ) illustrating the predicted optimum at @xmath327 .
the corresponding correlation function is shown in figure [ fig9](b ) . the minimum peak phase deviation in equation ( [ eq86 ] ) can be calculated with the above values and yields @xmath328{\pi}$ ] .
{ hz}$ ] position in the amplitude spectrum of the correlation function ( blue dots ) .
the solid red line indicates the theoretical signal height calculated with equation ( [ eq69 ] ) .
b ) the standard deviation of the noise determined from the amplitude spectrum . the theoretical curve was evaluated using equation ( [ eq81 ] ) .
c ) the signal - to - noise ratio of the data points in a ) and b ) shows a good agreement to the theory calculated according to equation ( [ eq82 ] ) and ( [ eq83 ] ) .
the optimum spatial discretization can be identified at @xmath327.,scaledwidth=100.0% ] to demonstrate the possibility , using the correlation analysis to describe not only single perturbation frequencies but also broad - band noise spectra , the following single - particle simulation has been made .
the temporal coordinates @xmath207 of the particles are generated in the same way as described in section [ sec3.3 ] .
the perturbation caused by a broad - band frequency noise is given by the corresponding amplitude spectrum @xmath329 and phase spectrum @xmath330 .
this is different to the former simulation , where the time - dependent perturbation @xmath308 was given by equation ( [ eq2 ] ) . with the amplitude and phase spectrum , the time - dependent perturbation @xmath308 for the temporal coordinate @xmath207
can be calculated using a discrete fourier transformation @xmath331 with the number of frequencies in the spectrum @xmath332 .
the spatial coordinate @xmath206 is determined as before according to the probability distribution ( equation ( [ eq89 ] ) ) with the calculated phase shift of equation ( [ eq90 ] ) . for
the simulation demonstrated here , a gaussian distributed noise spectrum with uncorrelated phases is applied .
the discrete amplitude spectrum is thus given by @xmath333 and the phase spectrum @xmath330 is randomly distributed between @xmath334 and @xmath28 .
here , @xmath335 denotes the maximum peak phase deviation , @xmath336 the central frequency and @xmath337 the frequency standard deviation ( band width ) . .
it is calculated with a numerical fourier transformation from the time - dependent perturbation @xmath308 ( equation ( [ eq90 ] ) ) used for the simulation in equation ( [ eq89 ] ) .
theoretical amplitude spectrum @xmath338 ( red solid line ) resulting from the fit of equation ( [ eq92 ] ) to the correlation function extracted from the simulated data .
the determined characteristics of the gaussian distributed noise in equation ( [ eq91 ] ) are @xmath339{\pi}$ ] , @xmath340{hz}$ ] and @xmath341{hz}$].,scaledwidth=50.0% ] an interference pattern consisting of @xmath342 particles acquired in with a contrast of @xmath125 and a spatial periodicity @xmath126{mm}$ ] was perturbed by a gaussian distributed noise according to equation ( [ eq91 ] ) with @xmath343{\pi}$ ] , @xmath344{hz}$ ] and @xmath345{hz}$ ] . the frequency spectrum used for
the simulation ranges from to with a resolution of .
after the simulation , the amplitude spectrum @xmath346 of the created time - dependent perturbation ( equation ( [ eq90 ] ) ) is calculated using a numerical fourier transformation and plotted in figure [ fig11 ] ( blue solid line ) . the noise on the gaussian distribution originates from the finite acquisition time and the randomly chosen phases . the correlation function of the simulated interference pattern is extracted according to equation ( [ eq62 ] ) with a spatial and temporal discretization of @xmath347 and @xmath348{ms}$ ] , that were chosen to have a good signal in the correlation function . in figure [ fig12](a ) ,
the resulting correlation function @xmath349 is shown .
the contrast @xmath350 and spatial periodicity @xmath351{mm}$ ] are determined by fitting equation ( [ eq36 ] ) to the correlation function at the temporal position @xmath352{ms}$ ] ( see section [ sec2.4 ] ) .
the superperiod of @xmath353{ms}$ ] belongs to the central frequency of @xmath323{hz}$ ] .
the contrast of the correlation function decays on timescales of @xmath354 . in figure [ fig12](a ) ,
the contrast is almost vanished for @xmath355{ms}$ ] .
this point is shifted to higher correlation times for smaller @xmath337 until the single frequency case with @xmath138{hz}$ ] is reached . for larger @xmath337 , more frequency components with random phases
contribute to the perturbation and the resulting time - dependent perturbation @xmath308 becomes more uncorrelated between two time stamps .
therefore , the corresponding particles are also uncorrelated and the contrast in the correlation function vanishes for shorter correlation times until it is completely lost . and @xmath126{mm}$ ] perturbed by a gaussian distributed noise according to equation ( [ eq91 ] ) with @xmath343{\pi}$ ] , @xmath344{hz}$ ] , @xmath345{hz}$ ] and a frequency resolution of .
the superperiod of @xmath353{ms}$ ] belongs to the central frequency of @xmath323{hz}$ ] .
with the full width at half maximum ( fwhm ) of the applied noise spectrum being @xmath356{hz}$ ] , the contrast in the correlation function vanishes on a timescale @xmath357{ms}$ ] .
b ) theoretical correlation function resulting from the fit with equation ( [ eq92 ] ) . here
, the argument of the bessel function is given by the discrete amplitude spectrum in equation ( [ eq91]).,scaledwidth=70.0% ] to determine the characteristics of the gaussian distributed noise ( @xmath335 , @xmath336 and @xmath337 ) , equation ( [ eq34 ] ) and ( [ eq35 ] ) are used together with the amplitude spectrum of the applied perturbation @xmath329 in equation ( [ eq91 ] ) as argument of the bessel function @xmath358 the approximate correlation function can be used for the theoretical description of broad - band frequency noise as long as the number of involved frequencies is large enough , so that the constraint in equation ( [ eq20 ] ) is only fulfilled for @xmath61 .
the contrast @xmath237 and spatial periodicity @xmath359 extracted according to section [ sec2.4 ] are fixed parameters for the fit to the correlation function of the simulation in figure [ fig12](a ) .
the fit parameters are @xmath335 , @xmath336 and @xmath337 in the discrete amplitude spectrum of the gaussian distributed noise ( equation ( [ eq91 ] ) ) .
the resulting theoretical correlation function @xmath360 is illustrated in figure [ fig12](b ) and shows a good agreement with the correlation function of the simulation .
the corresponding amplitude spectrum @xmath338 resulting from the fitted theoretical correlation function is plotted in figure [ fig11 ] ( red solid line ) .
it is also in good agreement with the amplitude spectrum of the applied perturbation ( blue solid line ) .
the determined characteristics of the gaussian distributed noise in equation ( [ eq91 ] ) are @xmath339{\pi}$ ] , @xmath340{hz}$ ] and @xmath341{hz}$ ] , which are in good agreement with the original values .
{ mm}$].,scaledwidth=50.0% ] in figure [ fig13 ] , the amplitude spectra of the simulated and theoretical correlation function , @xmath361 ( blue solid line ) and @xmath362 ( red solid line ) , are plotted .
both are calculated with a numerical fourier transformation at @xmath322{mm}$ ] with a frequency resolution of .
the theoretical amplitude spectrum agrees well with the simulation .
the theoretical description of the amplitude spectrum of the correlation function is given by equation ( [ eq47 ] ) for @xmath322{mm}$ ] and positive frequencies using the amplitude spectrum of the perturbation in equation ( [ eq91 ] ) as argument of the bessel function @xmath363 the broad frequency distribution around originates from the fundamental frequencies of the applied perturbation .
they are represented in equation ( [ eq93 ] ) by the first order of the bessel function @xmath364 .
the distributions around and are generated by the sum and difference frequencies ( intermodulation terms ) of the perturbation frequencies .
additionally , the distribution around originates from the sum of three frequency components of the applied perturbation .
if the properties of the applied perturbation are not known a priori , the amplitude spectrum can be used to get a reference to the shape and frequency characteristics , because they are included in the spectrum . in figure
[ fig13 ] , the central frequency of can be identified at the position of the maximum frequency distribution in the amplitude spectrum and used as starting value for the theoretical fit function ( equation ( [ eq92 ] ) ) .
if the central frequency of the perturbation would have been , for example , the distribution around would not be present in the spectrum .
the frequency standard deviation of the frequency distribution around is broadened because of additional terms in equation ( [ eq93 ] ) , that do not correspond to the fundamental perturbation frequencies @xmath364 .
however , it can be used as maximum frequency standard deviation for the theoretical fit function .
the applied perturbation spectrum can be identified , if the frequency distributions contained in the resulting amplitude spectrum are separated .
they can overlap , if the amplitudes of the perturbation spectrum are large or for very broad spectra .
the second - order correlation analysis can not be applied under all conditions .
especially , the possibility to determine the perturbation frequency and amplitude is crucial .
therefore , the limits of the applicability of the second - order correlation analysis shall be pointed out .
if the time of flight , that the particles spend in the area of perturbation , is large as compared to the cycle time of the oscillation , it is not possible to resolve the perturbation frequency , because the particles traverse many periods of the oscillation and therefore the perturbation is averaged out . in this case , the correlation analysis can not be applied . in general ,
the second - order correlation theory can be used for periodic oscillations , even if the average particle count rate is lower than the perturbation frequency , because of the infinite coherence of such a perturbation . due to the strong decay of the bessel function @xmath365 ,
the highest order per perturbation frequency contributing to the correlation function is @xmath366 .
therefore , the maximum frequency component of all perturbation frequencies included in the correlation function is given by @xmath367 , which arises from the argument of the cosine in equation ( [ eq31 ] ) and ( [ eq35 ] ) .
for slow and random perturbations with a high peak phase deviation this product sets a lower limit for the average particle count rate to get a good agreement between experiment and theory .
single - particle interferometry is an outstanding instrument in the field of quantum physics and sensor applications . due to the high sensitivity of interferometers , they are susceptible to dephasing effects originating from electromagnetic oscillations @xcite , mechanical vibrations @xcite or temperature drifts .
compared to decoherence , dephasing is a collective shift of the particle wave function and the contrast is only reduced in the temporally integrated interference pattern .
therefore , dephasing can in principle be reversed .
using second - order correlation theory , the wave properties can be identified and the perturbation characteristics can be determined .
this was demonstrated in former publications for electromagnetic perturbations @xcite and mechanical vibrations @xcite .
this paper provides the theoretical fundament for those articles and other future applications in various fields of single - particle interferometry .
it gives a detailed description of the analytic solution to the second - order correlation function and its numerical application .
we presented the full analytic derivation of our two - dimensional second - order correlation theory for multifrequency perturbations .
the difference between the explicit and approximate solution was discussed and areas of validity were investigated .
the amplitude spectra of both solutions , that are used for the identification of the perturbation characteristics @xcite , have been calculated .
we provided the numerical solution of the correlation function and investigated the dependence of the extracted contrast and perturbation amplitude on the discretization step size .
the influence of noise on the correlation function and corresponding amplitude spectrum was investigated and an optimum spatial discretization step size was provided to achieve a maximum signal - to - noise ratio . the validity of our calculations could be demonstrated with a single - particle simulation of a perturbed interference pattern evaluated for different spatial discretization step sizes . the possibility to analyze broad - band frequency noise was shown using a simulated interference pattern perturbed by gaussian distributed noise .
our method is a powerful tool for the proof of single - particle interferences , even if they are vanished in the spatial signal . especially for mobile interferometers or experiments in perturbing environments ,
the requirements for vibrational damping and electromagnetic shielding can be reduced .
furthermore , it is suitable to analyze the characteristics of multifrequency perturbations and broad - band noise .
therefore , it has possible sensor applications , which was demonstrated for mechanical vibrations in an electron interferometer @xcite .
it can be used in principle in every interferometer generating a spatial interference pattern on a detector with high spatial and temporal single - particle resolution .
this makes our method applicable in a wide range of experiments .
we gratefully acknowledge support from the dfg through the sfb trr21 and the emmy noether program sti 615/1 - 1 .
a r acknowledges support from the evangelisches studienwerk ev villigst .
the authors thank n kerker and a pooch for helpful discussions .
01 cronin a d , schmiedmayer j and pritchard d e 2009 _ rev . mod .
phys . _ * 81 * 1051 carnal o and mlynek j 1991 _ phys .
lett . _ * 66 * 2689 keith d w , ekstrom c r , turchette q a and pritchard d e 1991 _ phys .
lett . _ * 66 * 2693 margalit y , zhou z , machluf s , rohrlich d , japha y and folman r 2015 _ science _ * 349 * 1205 arndt m and brand c 2015 _ science _ * 349 * 1168 gustavson t l , bouyer p and kasevich m a 1997 _ phys . rev .
lett . _ * 78 * 2046 peters a , chung k y and chu s 1999 _ nature _ * 400 * 849 grisenti r e , schllkopf w , toennies j p , hegerfeldt g c , khler t and stoll m 2000 _ phys .
lett . _ * 85 * 2284 brezger b , hackermller l , uttenthaler s , petschinka j , arndt m and zeilinger a 2002 _ phys .
lett . _ * 88 * 100404 gerlich s , eibenberger s , tomandl m , nimmrichter s , hornberger k , fagan p j , txen j , mayor m and arndt m 2011 _ nature comm . _
* 2 * 263 arndt m and hornberger k 2014 _ nature phys .
_ * 10 * 271 haslinger p , drre n , geyer p , rodewald j , nimmrichter s and arndt m 2013 _ nature phys . _ * 9 * 144 zurek w h 2003 _ rev
phys . _ * 75 * 715 hackermller l , hornberger k , brezger b , zeilinger a and arndt m 2004 _ nature _ * 427 * 711 hornberger k , uttenthaler s , brezger b , hackermller l , arndt m and zeilinger a 2003 _ phys . rev . lett . _ * 90 * 160401 rauch h , treimer w and bonse u 1974 _ phys . lett .
a _ * 47 * 369 colella r , overhauser a w and werner s a 1975 _ phys .
lett . _ * 34 * 1472 hasselbach f 2010 _ rep .
phys . _ * 73 * 016101 hasselbach f and maier u 1999 _ quantum coherence and decoherence - proc .
isqm - tokyo 98 _ ed . by y.y .
ono and k. fujikawa ( amsterdam : elsevier ) 299 maier u 1997 _ doctoral thesis _
, university of tbingen kuo h s , hwang i s , fu t y , lin y c , chang c c and tsong t t 2006 _ japanese j. appl .
phys . _ * 45 * 8972 ehberger d , hammer j , eisele m , krger m , noe j , hgele a and hommelhoff p 2015 _ phys . rev .
_ * 114 * 227601 hommelhoff p , kealhofer c and kasevich m a 2006 _ phys .
lett . _ * 97 * 247402 hasselbach f 1988 _ z. phys .
b _ * 71 * 443 hammer j , hoffrogge j , heinrich s and hommelhoff p 2014 _ phys
* 2 * 044015 schtz g , rembold a , pooch a , meier s , schneeweiss p , rauschenbeutel a , gnther a , chang w t , hwang i s and stibor a 2014 _ ultramicroscopy _ * 141 * 9 chang c c , kuo h s , hwang i s and tsong t t 2009 _ nanotechnology _ * 20 * 115401 cho b , ichimura t , shimizu r and oshima c 2004 _ phys . rev .
lett . _ * 92 * 246103 jagutzki o , mergel v , ullmann - pfleger k , spielberger l , spillmann u , drner r and schmidt - bcking h 2001 _ nucl .
research a _ * 477 * 244 aharonov y and bohm d 1959 _ phys .
* 115 * 485 batelaan h and tonomura a 2009 _ physics today _ * 62 * 38 schtz g , rembold a , pooch a , prochel h and stibor a 2015 _ ultramicroscopy _ * 158 * 65 sonnentag p and hasselbach f 2007 _ phys .
* 98 * 200402 scheel s and buhmann s y 2012 _ phys .
a _ * 85 * 030101(r ) putnam w p and yanik m f 2009 _ phys .
rev . a _ * 80 * 040902(r ) kruit p , hobbs r g , kim c s , yang y , manfrinato v r , hammer j , thomas s , weber p , klopfer b , kohstall c et al . 2016
_ ultramicroscopy _ * 164 * 3145 grattan l s and meggitt b t 2013 eds . _
optical fiber sensor technology : fundamentals _ , springer science & business media abbott b p et al .
2016 _ phys .
lett . _ * 116 * 061102 graham p w , hogan j m , kasevich m a and rajendran s 2013 _ phys .
* 110 * 171102 hanbury brown r and twiss r q 1956 _ nature _ * 177 * 27 glauber r j 1963 _ phys . rev . _ * 130 * 2529 kiesel h , renz a and hasselbach f 2002 _ nature _ * 418 * 392 schellekens m , hoppeler r , perrin a , gomes j v , boiron d , aspect a and westbrook c i 2005 _ science _ * 310 * 648 foellmi c 2009 _ astronomy & astrophysics _ * 507 * 1719 simon j , bakr w s , ma r , tai m e , preiss p m , and greiner m 2011 _ nature _ * 472 * 307 agakishiev g et al .
2012 _ the european physical journal a - hadrons and nuclei _ * 47 * 1 ttl a , ritter s , khl m and esslinger t 2005 _ phys .
* 090404 federsel p , rogulj c , menold t , darzs z , domokos p , gnther a and fortgh j 2017 arxiv:1702.03159 grondalski j , alsing p m and deutsch i h 1999 _ optics express _ * 5 * 249 altman e , demler e and lukin m d 2004 _ phys . rev .
a _ * 70 * 013603 flling s , gerbier f , widera a , mandel o , gericke t and bloch i 2005 _ nature _ * 434 * 481 rembold a , schtz g , chang w t , stefanov a , pooch a , hwang i s , gnther a and stibor a 2014 _ phys . rev .
a _ * 89 * 033635 gnther a , rembold a , schtz g and stibor a 2015 _ phys .
rev . a _ * 92 * 053607 rembold a , schtz g , rpke r , chang w t , hwang i s , gnther a and stibor a 2017 _ new j. phys . _
* 19 * 033009 mllenstedt g and dker h 1956 _ z. phys .
a - hadron nucl . _
* 145 * 377 wiener n 1930 _ acta math .
_ * 55 * 117258 khintchine a 1934 _ math . ann . _ * 109*(1 ) 604 siegmund o h , vallerga j v , tremsin a s , mcphate j and feller b 2007 _ nucl .
instr . meth .
research a _ * 576 * 178 schellekens m , hoppeler r , perrin a , viana gomes j , boiron d , aspect a and westbrook c i 2005 _ science _ * 310 * 648 zhou x , ranitovic p , hogle c w , eland j h d , kapteyn h c and murnane m m 2012 _ nature physics _ * 8 * 232 hasselbach f , nicklaus m 1993 _ phys . rev . a _ * 48 * 143 abramowitz m and stegun i a 1964 _ handbook of mathematical functions _ ,
courier corporation * 55 * 360 arfken g b and weber h j 1999 _ mathematical methods for physicists _ , aapt * 67 * 165 casella g , robert c p and wells m t 2004 _ lecture notes - monograph series _ 342347 haight f a 1967 _ handbook of the poisson distribution _ | interferometers with single particles are susceptible for dephasing perturbations from the environment , such as electromagnetic oscillations or mechanical vibrations . on the one hand
, this limits sensitive quantum phase measurements as it reduces the interference contrast in the signal . on the other hand
, it enables single - particle interferometers to be used as sensitive sensors for electromagnetic and mechanical perturbations .
recently , it was demonstrated experimentally , that a second - order correlation analysis of the spatial and temporal detection signal can decrease the electromagnetic shielding and vibrational damping requirements significantly .
thereby , the relevant matter - wave characteristics and the perturbation parameters could be extracted from the correlation analysis of a spatially washed - out " interference pattern and the original undisturbed interferogram could be reconstructed .
this method can be applied to all interferometers that produce a spatial fringe pattern on a detector with high spatial and temporal single - particle resolution . in this article , we present and discuss in detail the used two - dimensional second - order correlation theory for multifrequency perturbations .
the derivations of an explicit and approximate solution of the correlation function and corresponding amplitude spectra are provided .
it is explained , how the numerical correlation function is extracted from the measurement data .
thereby , the influence of the temporal and spatial discretization step size on the extracted parameters , as contrast and perturbation amplitude , is analyzed .
the influence of noise on the correlation function and corresponding amplitude spectrum is calculated and numerically cross - checked by a comparison of our theory with numerical single - particle simulations of a perturbed interference pattern .
thereby , an optimum spatial discretization step size is determined to achieve a maximum signal - to - noise ratio .
our method can also be applied for the analysis of broad - band frequency noise , dephasing the interference pattern . using gaussian distributed noise in the simulations , we demonstrate that the relevant matter - wave parameters and the applied perturbation spectrum can be revealed by our correlation analysis . |
understanding the formation and evolution processes of the individual galaxy and the growth of the large scale structures ( lsss ) in the universe is one of the major goals in extragalactic astronomy today . in the scheme of a typical @xmath9-cdm model ,
the growth of structures is governed by the gravitational growth of initial fluctuations of dark matter .
the baryonic material cools in these dark matter structures and grows through hierarchical clustering to galaxies and clusters of galaxies .
the subsequent evolution of galaxies is closely connected to their environments , which , of course , relate to the lsss , where those galaxies are located .
the ultimate goal of extragalactic surveys is to trace these evolutionary processes by well - defined statistical galaxy samples .
optical imaging is arguably the cornerstone of any extragalactic survey , since it provides identifications and positions of celestial objects for follow - up spectroscopy . in recent years
, there have been a number of deep imaging survey projects which devote significant amounts of telescope time , such as great observatories origins deep survey ( goods : giavalisco et al .
2004 ) and cosmic evolution survey ( cosmos : hst treasury project : feldmann et al .
2006 ; scoville et al .
these surveys provide multi - waveband galaxy samples at faint magnitude .
then , the photometric redshift techniques ( furusawa et al . 2000 ; bolzonella et al .
2000 ; feldmann et al .
2006 ) are frequently used to pre - select candidates for spectroscopy .
although a great deal of information can be obtained from the optical data alone , the value of the data set grows significantly as data at other wavelengths from other facilities are added .
we therefore elected to use a multi - wavelength approach for the subaru / xmm - newton deep survey ( sxds ; sekiguchi et al .
2007 , in preparation ) from the very start , ensuring that our chosen field would be accessible and suitable for observations at all wavelengths . an equatorial field , centered at r.a .
= @xmath0 , decl .
= @xmath1 is chosen to tie up with the deep x - ray observations with xmm - newton observatory ( ueda et al .
2007 ) , and also due to the accessibility by all major observatories , both existing and planned .
the major multi - wavelength observation programs on the sxds field ( hereafter sxdf ) include deep radio imaging with the vla ( simpson et al .
2006 ) , sub - mm mapping with scuba ( coppin et al . 2006 ; mortier et al . , 2005 ) , mid - infrared observations with spitzer space telescope ( lonsdale et al .
2003 ) , deep near - infrared imaging with the ukirt / wfcam ( foucaud et al . 2006 ; dye et al .
2006 ; lawrence et al .
2007 ) , and the x - ray observations with xmm - newton observatory .
importantly , the survey field of an infrared ultra deep survey ( uds ; foucaud et al .
2006 ) covering 0.77 square degrees as part of the ukidss project ( lawrence et al . 2007 ) is centered on the sxdf .
it is expected that our extensive multi - wavelength data set will provide photometric redshifts accurate to @xmath10 over a wide range of redshift , as well as detailed spectral energy distributions for the vast majority of the objects in the field .
the sxds has been undertaken as a part of the subaru telescope `` observatory projects '' , in which a large amount of observing times are devoted to carry out intensive survey programs by combining observing times rewarded to builders and observatory staff of the subaru telescope .
note that subaru deep field ( sdf ; kashikawa et al .
2004 ) is another observatory project , which targets on a field different from the sxdf .
the data set of the sdf is slightly deeper than the sxdf , but concentrates on one - fifth narrower field than the sxdf .
thus , these surveys complement each other and they can be used for a wide variety of studies . in this paper , we describe the observation details and data reduction and analysis of our optical imaging survey with the prime - focus camera on subaru telescope , as well as the resultant data products .
a detailed description of the survey strategies , scientific objectives , and multi - wavelength survey plans is given in a companion paper ( sekiguchi et al .
, 2007 - paper i , in preparation ) . in section 2 ,
we describe optical imaging observations by subaru telescope , and in section 3 we explain the data reduction procedure employed .
we present creation of multi - waveband photometric catalogs including object detection and aperture photometry in section 4 , and calibrations of photometry and astrometry in section 5 . in section 6 ,
we discuss characteristics of the catalogs such as limiting magnitude , detection completeness , and number counts of galaxies .
the summary is given in section 7 . throughout this paper ,
all the magnitudes are expressed in the ab system unless otherwise mentioned .
the optical imaging observations of the sxdf are carried out using the prime - focus camera ( suprime - cam ; miyazaki et al .
2002 ) on subaru telescope in the period from september 2002 to september 2005 .
suprime - cam is a @xmath11 mosaic of ten mit / ll buttable 2k@xmath124k ccd camera .
it has a @xmath13 field of view with a projected pixel scale of 0.202 arcsec .
the layout of the pointings is arranged as a cross shape so that each of the north - south and the east - west directions has an extent of @xmath14 deg ( figure [ fig : pseudo - img ] ) .
this corresponds to a field span of a transverse dimension of @xmath15 mpc at @xmath16 and @xmath17 mpc at @xmath18 in the comoving scale for ( @xmath19 ) .
the coverage of the comoving volume is @xmath20 mpc@xmath21 for @xmath22 .
table [ tab : coords ] lists field center coordinates of the five pointings .
hereafter , the five pointings are referred to as @xmath23 ( center ) , @xmath24 ( north ) , @xmath25 ( south ) , @xmath26 ( east ) , and @xmath27 ( west ) , with respect to their relative positions on the sky . observations of the sxdf are performed using five broad - band filters , @xmath28 , @xmath29 , @xmath30 , @xmath8 , and @xmath31 to cover the entire wavelength range observable with suprime - cam .
the transmission curves of the five filters employed in the sxds , convolved with the ccd sensitivities , reflectivity of the primary mirror , transmissions of corrector lenses , and atmospheric transmission , are given in figure [ fig : band ] . also , given in figure [ fig : band ] are the redshifted spectral energy distributions ( seds ) of an early - type galaxy at @xmath32 and @xmath33 and a late - type galaxy at @xmath32 and @xmath34 for references . with photometric data sampled in the 5 bands
, we can investigate color properties of red passively evolving galaxies to @xmath16 ( see kodama et al .
2004 ; yamada et al . 2005 ) and young star - forming galaxies at high redshift such as lyman break galaxies ( lbg ) ( ouchi et al . 2004 , 2005 ) .
the observations used in this paper span observing runs over a period of 3 years .
a total of 160 hours are allocated to the sxdf imaging observations as a part of subaru telescope `` observatory projects '' ( see also kashikawa et al .
2004 ) combined with observatory staff times .
in addition to these times , we use 14 hours of the observing times offered by the supernova cosmology project ( doi et al .
2003 ; lidman et al .
2005 ) , which add the total exposure time of the center and the west pointings in the @xmath8 band ( _ @xmath8-c _ and _ @xmath8-w _ ) . moreover , in the course of another survey program for ly@xmath35 emitting galaxies ( ouchi et al .
2005 , 2007 ) , 3 hours are devoted to the @xmath29-band imaging . in total , 133 hours are used for on - source exposures for the final data set described in this paper .
the complete log of the observations is given in table [ tab : log ] .
the coordinates of the five pointings are carefully chosen so that the resultant images uniformly cover the entire sxdf , though we analyze each of the pointings separately in the present study . also ,
to eliminate bad pixels , gaps between individual ccd chips and cosmic ray events , circular dithering pattern is used for each pointing .
this dithering pattern employs a circular motion with radii of 60 to 120 arcsec around the center position of each pointing , combined with slight offsets to the center positions of the circular motion . as an example , the dithering pattern for the @xmath30 band is plotted for each pointing in figure [ fig : dither ] .
the sxds images obtained are processed using an in - house pipeline software package ( see ouchi 2003 ) .
we use it in almost the same way that successfully used for the sdf ( kashikawa et al .
this pipeline is based on a software package developed for suprime - cam data reduction ( yagi et al .
2002 ) and several iraf tasks ( geomap , geotran , etc . ) are used in the process of a geometrical transformation and alignment of final stacked images .
we use a standard mosaic - ccd data reduction procedure , which is briefly summarized below .
first , we apply a bias correction to the raw data .
the median count of the overscan region is computed for each line , which represents a typical bias level at that line , and this median count is subtracted from the counts in all the pixels in that line .
the overscan regions are trimmed after the bias subtraction .
we assign a flag number ( -32768 ) to the saturated or detected bad pixels , and these pixels are ignored in the following reduction processes . since pixels in the outside edges of each frame are likely to be affected by noises which can not be corrected by the following flatfielding process , they are also masked with the flag number .
flat frames are constructed from the normalized object frames taken during the same observing run .
we use more than 150 object frames to create flat frames .
if the number of object frames is less than 150 frames , object frames taken in different observing runs within 3 months or twilight flat frames taken in the same run are combined to increase the signal - to - noise ratios ( s / ns ) of the flat frames for that period .
the median value of the each pixel is used to create the flat frames .
pixels which could be affected by bright objects and vignetting by the auto - guider probe are flagged out before processing the flat frames .
then , object frames are divided by the flat frames .
the peak - to - peak ratios of background levels over the entire flat fielded frames are less than @xmath36 .
we do not expect any significant systematic errors by the flatfielding that affect accuracies of the following analyses .
after the flatfielding process is done , the distortion correction is applied to each frame using the 5-th order polynomial transformation formula derived from miyazaki et al .
the changes in shapes and sizes of objects in the frames due to the distortion are negligible compared with the size of psfs .
each pixel is transformed so that the surface brightness of the pixel is conserved , since the relative flux in each pixel has been already corrected by the flatfielding process .
positional displacement of objects along the perpendicular orientation due to the atmospheric effect is also corrected at the same time .
the sky background is subtracted from each distortion - free frame , we divide an image into temporary meshes with @xmath37 pixels , corresponding to 12.9 arcsec in one direction .
the mode of counts in each mesh is adopted as the sky background levels after smoothed by a median filter with a @xmath38 kernel .
this global sky subtraction should not affect photometric accuracies of compact objects such as galactic stars and faint galaxies .
however , caution must be taken when we apply these values to extended objects such as nearby galaxies with low surface brightness .
the stacking of the images is performed for each of the pointings separately in the following way .
first , we choose 100 to 200 unsaturated stellar objects with a peak flux of @xmath39adu in each frame distributing evenly over the entire frame .
second , we determine a reference frame of all the reduced frames . for the other frames , relative positions and rotation ( @xmath40 ) and flux ratios to the reference frame are determined based on a least square method using these stellar objects commonly detected in the frames .
third , for all the frames , the positional shifts and rotations are corrected . here
, we do not use frames which have the flux ratios of less than @xmath41 of that of the reference frame .
finally , all the frames are co - added by calculating flux - weighted average values in each pixel with a 3-@xmath42 clipping process .
the clipping process effectively removes cosmic rays and bad or saturated pixels , and satellite trails etc . , and should not affect the resulting total fluxes of objects because the difference in counts among frames with various seeing sizes is mostly within the 3-@xmath42 threshold .
after the co - adding of each band images , all the five - band ( @xmath43 ) images are registered into identical positional coordinates as we intend to execute multi - waveband aperture photometry .
the positional registration is performed by the geometrical transformation by 3-rd order polynomials fitting using the positions of stellar objects common to the co - added stacked images .
the r.m.s .
deviation of residuals of transformed positions with respect to the reference positions in the @xmath30-band images is approximately 0.1 - 0.15 arcsec . for each pointing ,
the geometrically matched images are smoothed with gaussian filters iteratively so that the psf sizes of all the images become the same as that in the worst image used in this study .
the psf sizes of the resultant images are @xmath44 , and @xmath45 arcseconds in _ sxds - c , sxds - n , sxds - s , sxds - e _ , and _ sxds - w
_ , respectively .
finally , regions with low s / ns or those affected by saturation or overflow of electrons are carefully checked and assigned as flagged regions which are indicated in photometric catalogs .
the resultant psf sizes and total usable areas of each pointing is summarized in table [ tab : stacked ] .
we construct multi - waveband photometric catalogs of the objects identified in the sxdf .
object detection and photometry are performed using sextractor version 2.3.2 ( bertin & arnouts 1996 ) . before performing the object detection and photometry , the outskirts of the stacked images
are trimmed in order to eliminate any failure such as inappropriate estimation of sky levels and memory overflow during extraction of objects by sextractor due to a high noise level . after trimming the outskirts ,
each individual image has similar size of approximately @xmath46 pixels .
we use the detection criteria that at least five pixels which have values of @xmath47 of sky background r.m.s .
noise must be connected in order to be included in our object list .
positions of detected objects are translated to sky coordinates based on the astrometric calibration which will be discussed in section [ sec : astr_calib ] . after the object detection is completed in one of the five bands , which we call ` detection band ' , multi - waveband aperture photometry is carried out with sextractor for the other four bands using the same apertures at the same positions as those in the detection band .
we perform aperture photometry with fixed circular apertures ( mag_aper ) of @xmath48 and @xmath49 arcseconds in diameter , and variable elliptical apertures ( mag_auto and mag_best ) utilizing the kron s first moment scheme with apertures radii of @xmath50 ( kron 1980 ) .
measured fluxes of objects are converted into magnitude using the photometric zeropoints , which will be described in section [ sec : photo_calib ] .
the extracted magnitude is considered to be the asymptotic total magnitude of galaxies , which should sample at least @xmath51 of flux of a galaxy ( bertin & arnouts 1996 ) .
other useful parameters for objects including shape , size , and stellarity and so on , are also extracted by sextractor at the same time .
all the object parameters which are listed in the catalogs are summarized in table [ tab : catalog ] . in this procedure , @xmath28- , @xmath29- , @xmath30- , @xmath8- , and
@xmath31-selected multi - waveband catalogs are created for each pointing separately , which means that each object has measurements in the five bands ( @xmath4 ) in each catalog .
thus 25 catalogs (= 5 pointings times 5 detection bands ) in total are created .
although we have flagged out bad pixels from the image frame in the reduction procedure , there still remain ( 1 ) areas strongly affected mainly by overflow of electrons or envelopes of extended objects , and ( 2 ) areas weakly affected by large halos from very bright stars or noisy regions . in our catalogs , the objects which are located in these areas are flagged 1 and 2 , respectively .
otherwise , the objects located in the clean area have a flag value of 0 .
the basic parameters used in the object detection and aperture photometry are summarized in table [ tab : sexparam ] .
the complete configuration files for sextractor can be found at http://www.naoj.org/science/subaruproject/sxds/index.html .
note that we adopt default parameters of sextractor for the source extraction unless we do not find any problem with the output values .
the parameters adopted are suitable for extraction of moderate to faint compact galaxies , and not heavily optimized to extraction of extended objects or very faint objects .
thus the reliability of the current catalogs for those objects should be regarded with some caution .
photometric zeropoints in all the five bands are determined in the following way .
first , we compare our observation with the earlier calibrated observations for the same overlapping field , which were made by the suprime - cam instrument development team ( ouchi et al .
2001 , 2004 ) .
the southern half of our center pointing ( _ sxds - c _ ) is overlapped with the northern half of the image obtained by ouchi et al .
they determined the photometric zeropoints in the suprime - cam band system based on observations of photometric standard stars sa92 and sa95 for @xmath52 and @xmath30 bands and a spectrophotometric standard star sa95 - 42 for @xmath8 and @xmath31 bands .
we estimate the photometric zeropoints for our center pointing ( _ sxds - c _ ) by comparing magnitudes ( mag_auto ) of @xmath53 objects which are commonly detected in the two images and have fwhms of smaller than 1.2 arcsec with a magnitude range of 20.0 to 24.0 .
then , we extend these zeropoints to the surrounding four pointings by using objects in the overlapping areas in the same manner as for the determination of the zeropoint for the _ sxds - c_. in our catalogs , we list the instrumental magnitude obtained in the suprime - cam band system . hence , no transformation of magnitude considering color terms of band responses are conducted .
the color terms between the suprime - cam band system and the standard system are basically small for the reddest stars , of order 0.1 in the @xmath54 bands , and slightly large ( @xmath55 ) in the @xmath30 band . to test internal consistency of our zeropoint calibrations , we examine colors of stellar objects in our photometric catalogs and those computed by convolving typical stellar seds covering the complete ranges of spectral types ( gunn & stryker 1983 ) with suprime - cam response curves .
we find that it is necessary to apply corrections to the zeropoints of @xmath56 and @xmath57 for all the five pointings , and @xmath58 and @xmath59 for the south and east pointings ( _ @xmath8-s _ and _ @xmath8-e _ ) , respectively .
after these corrections are applied , the colors of the stellar objects in our catalogs agree with those computed for the seds from gunn & stryker ( 1983 ) within @xmath60 mag in all the five bands ( figure [ fig : ccplot_gs ] ) .
the systematic errors adjusted here may be due to relatively small number of stellar objects used in the determination of the photometric zeropoints . to ensure the accuracy of our zeropoints
, we conduct ccd photometric observations of a part of the sxdf in @xmath61 and @xmath31 bands with the uh88-inch telescope at mauna kea on october 9 and 10 , 2002 . we compare magnitudes of objects in @xmath61 and @xmath31 bands obtained by the uh88-inch observations with our catalog values .
we find that our catalog values agree within 0.05 mag r.m.s . for @xmath28 and @xmath30 bands and within 0.1 mag r.m.s . for @xmath8 and @xmath31 bands .
the relatively large difference seen in @xmath8 and @xmath31 bands are probably due to the difference in the band responses between the uh88-inch system and the suprime - cam system , and low s / ns of shallow sxds images used for comparison with the uh88-inch data . to confirm accuracy of the photometric zeropoints in the @xmath8 and @xmath31 bands in our catalogs ,
follow - up observations are made for the two bands with the 1.0-m telescope at the united states naval observatory ( usno ) .
the observations at the usno were performed on the photometric night of december 3 , 2004 in the course of photometric calibration for a high-@xmath62 supernovae search ( yasuda et al .
2007 , in preparation ) .
we find that a systematic difference between zeropoints in our catalogs and those determined by the usno observation is as small as @xmath60 mag r.m.s .
thus we conclude that the uncertainties of calibrated photometric zeropoints of the sxds images are @xmath63 mag r.m.s .
the adopted photometric zeropoints are listed in table [ tab : log ] .
astrometric calibration is performed using @xmath64200 stars with magnitude of @xmath65 derived from 2-mass point source catalogs ( psc ; cutri et al .
2003 ) , in each of the five stacked @xmath30-band images .
since images in the other bands have been geometrically transformed , the positions of objects detected on the images should coincide with those found in the @xmath30-band images ( see section [ sec : reduc ] ) .
the world coordinates for the images are calculated for the @xmath30-band images and written in the fits headers of all the stacked images with the iraf tasks ccmap and ccsetwcs .
the r.m.s .
uncertainties of the determined coordinates across the fov with respect to the psc positions are on the order of 0.2 arcsec in both ra and dec in each pointing ( table [ tab : astmt ] ) .
the positional error of the psc is reported as @xmath66 arcsec .
we examine possible systematic differences in the calculated world coordinates between different pointings by using objects in the overlapping areas of two pointings .
figure [ fig : astmt_diff1 ] and figure [ fig : astmt_diff2 ] show the differences in ra and dec of point - like objects with fwhms @xmath67 arcsec in the magnitude range @xmath68 detected both in the _ sxds - c _ image and other surrounding images . from the figures ,
the world coordinates assigned to each of the five pointing images are in good agreement with one another with accuracy of about 1 arcsec at the edges of the images .
the differences of the world coordinates between two pointings studied here are thought to be for the worst cases , since they are derived based on only the objects in the overlapping areas which are located at the edge of each image .
slopes and offsets in the residuals are seen in each panel of the figures .
these disagreements in the positions of objects between two overlapping images come from possible tilts , offsets and/or difference in geometrical scale of the world coordinate systems between the two images .
residuals of distortion in the images and the atmospheric effect , which have not been completely removed in the reduction process , may cause the errors in determination of the world coordinate systems .
nevertheless , the sxds catalogs have a good enough positional accuracy to perform follow - up spectroscopy .
the astrometrically and photometrically calibrated images of the sxds in five bands ( @xmath4 ) described in this section are released via sxds web page ( http://www.naoj.org/science/subaruproject/sxds/ ) . the data set described in this paper
is based on the data release 1 ( dr1 ) .
a complete history of the data release is summarized in the web site .
the raw image data are also available to the public via subaru telescope archives system ( stars : https://stars.naoj.org ) .
all the magnitudes listed in the catalogs are not corrected for the galactic extinction .
the magnitude attenuation in each band estimated for the central position ( @xmath69 ; j2000 ) based on schlegel et al .
( 1998 ) , assuming an extinction curve with @xmath70 is as follows : @xmath71 , and @xmath72 .
table [ tab : gal_ext ] summaries the magnitude attenuation in each band for all the pointings .
the galactic extinction above must be taken into account when studying extragalactic objects .
we examine 3-@xmath42 limiting magnitudes measured for the @xmath73-diameter apertures in all five bands ( @xmath74 column in table [ tab : log ] ) in the following manner .
first , counts in unit of adu fallen into the @xmath73-diameter apertures are measured at approximately 10,000 positions randomly selected and spreading over the entire image .
next , a histogram of the counts is produced and only the faint side of the histogram is fit by the gaussian function , as the bright - side tail is composed of not only the sky background but also photons from the objects .
the @xmath42 of the best - fit gaussian is regarded as the 1-@xmath42 sky fluctuation of the image for the @xmath73-diameter apertures .
we find that limiting magnitude in each band is @xmath75 and @xmath6 ( @xmath76 ) in the deepest images of the five pointings . to examine the homogeneity of s / ns of the sxds data across the entire field , we investigate the s / n distribution map in the @xmath30-band images .
first , we calculate the sky fluctuation per @xmath73-diameter aperture within a mesh with a size of @xmath77 pixels at each position in the same manner as in estimation of the limiting magnitudes .
the mesh is shifted with a step of 175 pixels to cover the entire field of view .
then , the fluctuations for the 2-diameter aperture are converted to s / ns for objects with an @xmath30 magnitude of 27.5 at each position .
figure [ fig : snmap ] shows the s / n distribution maps for 27.5-mag objects thus obtained for each pointing .
the contour lines superimposed on the figures represent s / n=3 positions .
first , we see that the brightest sources on the images significantly affect the s / ns around the sources . these areas are flagged in the catalogs .
second , each image shows almost an axisymmetric pattern of the s / n distribution when we ignore the low s / n areas due to the brightest sources .
this pattern reflects the vignetting of the prime focus .
third , the pattern is slightly distorted for each image , probably due to bright sources , vignetting by the auto - guider probe , difference in the quantum efficiency among ccds , and so on . however , the difference in s / ns between in such distorted areas and in clean areas with higher s / ns is on the order of 20% at the maximum .
this difference is negligible in most studies even for faint sources .
thus , we can securely utilize the data with quite homogeneous s / ns across the field of view .
completeness of the object detection as a function of magnitude for each band and for each pointing is estimated .
the detection completeness is determined based on a monte carlo simulation by adding artificial objects which have gaussian profiles with their poisson photon noise into a stacked image used to create the catalogs and then detect them again in the same manner as in creating catalogs .
the detection rate thus obtained is regarded as the detection completeness . here
the artificial objects are added to random positions in the effective area of the image .
the fwhms of the gaussian profiles are set to be the same as the psf sizes representative of the images which are derived from table [ tab : stacked ] .
the artificial objects contaminated by any neighbor sources listed in the catalogs are removed in calculation of the detection rate .
if such blended objects are not removed from the sample , the detection rate should slightly decrease .
however , since the contamination by neighbor objects leads to misidentification of the artificial objects in the object detection process , we adopt the approach which excludes the blended artificial objects , which was also employed in kashikawa et al .
( 2004 ) .
the completeness determined by this procedure is shown in figure [ fig : comp ] for all of the five bands .
for the five pointings , the curves of the detection completeness drop similarly with increasing magnitude .
we can summarize that the detection completeness in each image is at approximately @xmath78 for objects with the 5-@xmath42 magnitude and @xmath41 for objects with the 3-@xmath42 magnitude , with a slight difference depending on the band .
on the other hand , rates of the false object detection are also estimated as a function of magnitude .
we generate negative images of each pointing and in each band by multiplying all the counts of the stacked images by @xmath79 , and then perform object detection for the negative images in the same manner as in creating the catalogs . in this process ,
detected objects are considered to be spurious objects .
thus we define the false detection rate as the number of the spurious objects divided by the total number of objects in the catalogs in each magnitude bin , i.e. , @xmath80 ( figure [ fig : false ] ) .
it is seen that a contribution of the false detection is negligible , which does not exceed 0.5% if any , in the magnitude range brighter than the 3-@xmath42 limiting magnitude . in the areas where two images overlap with each other , a large fraction of objects are detected and their magnitudes are measured in both of the pointings .
we can check systematic difference in magnitude of those objects between the two catalogs . since photometric zeropoints of the catalogs are determined so that mag_auto of objects in the overlapping areas should coincide between any two catalogs ( see section [ sec : photo_calib ] ) , the magnitude difference in mag_auto is negligible .
mag_best should be identical to mag_auto for isolated objects .
therefore magnitude differences only for the @xmath73- and @xmath81-diameter aperture magnitudes are investigated in the following manner .
first , we choose only compact objects with fwhms @xmath67 arcsec which are listed in both the @xmath31-selected catalog for _ sxds - c _ and that for another pointing which overlap with each other , then measure the difference in magnitude of the objects by subtracting magnitude in the _ sxds - c _ catalog by that in the other .
then , we determine the offset in the magnitude difference , i.e. , systematic magnitude difference , by fitting the magnitude difference for a range of 21.0 to 23.5 mag by a least square method .
table [ tab : diff_magap ] lists the systematic magnitude difference thus obtained for the @xmath73- and @xmath81- diameter aperture magnitudes in each band for each pair of catalogs .
we note that @xmath73- and @xmath81-diameter aperture magnitudes indicate small systematic differences of @xmath82 mag between the catalogs .
we think that this difference is probably due to difference in the psf shape of the stacked images among the pointings . in order to see the effects of the systematic difference in the aperture magnitudes , we plot differences in colors of objects between the catalogs measured with the @xmath73-diameter aperture magnitude ( figure [ fig : colordiff ] ) . in this figure , for each of the four pairs of catalogs , the differences in colors ( @xmath83 ) are plotted .
we see the systematic difference in the @xmath73-diameter aperture colors of objects , which is no larger than @xmath84 mag , due to the difference in the aperture magnitude between the catalogs .
since the magnitudes in the catalogs are not adjusted for this difference , the magnitude difference between the catalogs should be taken into account in the case that the @xmath73- or @xmath81-diameter aperture magnitude is used .
for mag_auto and mag_best , which are considered as asymptotic total magnitude , the magnitude difference is negligible . in the same manner as mentioned above , the r.m.s .
of the magnitude differences for each band are calculated as a function of mag_auto , which shows no systematic difference . here the detection band for catalogs are chosen to be the same as the band for which the magnitude difference is investigated .
for instance , to obtain the magnitude difference in the @xmath28 band , the @xmath28-selected catalogs are used .
again , only the objects with fwhms @xmath67 arcsec are used for measurement of the magnitude difference . the r.m.s . of magnitude differences measured in the above way are shown in figure [ fig : magdiff_random ] for each band .
it is seen that the r.m.s .
steadily increases with increasing magnitude . the r.m.s . of the magnitude difference can be converted into random magnitude error by dividing it by @xmath85 .
we find that the magnitude errors thus estimated are entirely consistent with those expected from our calculation of the limiting magnitude discussed in section [ sec : limmag ] .
note that the magnitude errors listed in our catalogs are computed by sextractor assuming the simple poisson statistics for the sky background fluctuations .
the sky background noises measured in photometry apertures for the stacked images are likely to be larger than those estimated by sextractor
. therefore the magnitude errors in the catalogs are underestimates and are not suitable for immediate analysis such as sed fitting and photo-@xmath62 estimation . to examine the characteristics of the catalogs we compare the number counts of galaxies with data from previous surveys .
we calculate the number counts of galaxies by correcting our raw counts of detected objects against the detection completeness estimated in section [ sec : comp_false ] .
the detailed studies of density fluctuations of galaxies based on the number counts or luminosity functions of galaxies can be found in elsewhere ( e.g. , yamada et al . 2005 ) .
star / galaxy separation is conducted based on fwhm and @xmath73-diameter fixed aperture magnitude of the objects , for which the effect of the galactic extinction ( table [ tab : gal_ext ] ) is corrected .
the criteria for the separation is determined so that both objects which have sizes as same as psf of the images and saturated objects are effectively removed . for fainter magnitudes ( @xmath86 mag ) ,
no separation is performed since a large fraction of galaxies have small sizes similar to those of stars , and contribution to the number count by stars is small at the faint magnitudes .
the criteria for the separation adopted is summarized in table [ tab : sgsep ] .
the results of the number counts of galaxies are as follows : first , internal comparison of the number counts among the five pointings of the sxds are shown for each band in figure [ fig : nm_b ] to figure [ fig : nm_z ] . in each figure ,
raw number counts of the galaxy samples extracted by the star / galaxy separation process above are plotted with open symbols .
effective areas used to calculate the number counts per unit area are given in table [ tab : stacked ] .
the number counts of galaxies which are corrected for the detection completeness are also superimposed with filled symbols . here , the correction is performed by using the detection completeness as a function of magnitude which has been estimated for the gaussian profiles , i.e. , @xmath87 .
error bars are calculated on the assumption of simple poisson noise .
we can say that no large systematic difference is seen in the superposition of the number counts of galaxies among the five pointings in @xmath52 , and @xmath30 bands down to the 3-@xmath42 magnitudes .
similarly , in the other two bands , the number counts are consistent with one another among the five pointings within the poisson errors at magnitudes brighter than @xmath88 mag .
however , differences in the number counts at fainter magnitudes are found by a factor of 1.4 in the @xmath8 band and 1.7 in the @xmath31 band .
this might be partly due to the field - to - field variation in the number counts of galaxies .
however , since we have used only the gaussian profiles with fwhm@xmath89 arcsec in the simulation of the detection completeness , the completeness is likely to be overestimated at a faint magnitude range where most objects have extended profiles .
so , we can not conclude that the differences in the number counts at the faint end seen in the @xmath8 and @xmath31 bands indicates the field - to - field variation .
next , the number counts in the sxdf are compared with results by previous surveys ( figure [ fig : nm_mean_b ] to figure [ fig : nm_mean_z ] ) . here
, we calculate mean number counts of galaxies corrected for the detection completeness for the entire sxdf using the above sample for the five pointings , and plot them with filled circles with error bars .
the error bars are again based on the poisson noise .
the results in other surveys are derived from kashikawa et al .
( 2004 ) , capak et al .
( 2007 ) , and metcalfe ( 2007 ) and the references cited therein .
a complete list of references compared in the figures is as follows : sdf kashikawa et al .
2004 , hdfs / whdfs metcalfe et al .
2001 , arnouts99 arnouts et al . 1999 ,
sdss yasuda et al .
2001 , cadis
huang et al .
2001 , kummel&wagner01 kmmel & wagner 2001 , tyson88 tyson 1988 , arnouts01 arnouts et al .
2001 , capak04 capak et al .
2004 , smail95 smail et al .
1995 , sh93 steidel & hamilton 1993 , hogg97 hogg et al .
1997 , cosmos
capak et al .
2007 , postman98 postman et al . 1998 , lilly91 lilly , cowie , & gardner 1991 . from the comparison of
the number counts among various surveys , we understand that our sample is consistent with results in other surveys for blank fields for the following reasons .
first , the mean number counts of the sxds show a good agreement with those from previous surveys down to magnitude where the number counts start to turn off .
second , it reproduces the faint - end data points derived from hubble deep field ( hdf ) .
again , the detection completeness simulated with the gaussian profiles might cause a slightly large uncertainty of the corrected number counts at the faint end in each band .
thus , we conclude that the sxds catalogs in the present study have characteristics similar to those in previous surveys and can be applied to general statistical studies on faint objects ranging from galactic objects to high - redshift galaxies .
we would like to mention how the poisson error on the number counts is decreased by increase of the survey area .
figure [ fig : poisson ] shows a comparison of the poisson error associated with the number counts of galaxies among three surveys which are based on different survey areas , namely the hdf - north ( 5.3 arcmin@xmath90 ) , goods - north ( 170 arcmin@xmath90 ) and suprime - cam 1 fov ( 918 arcmin@xmath90 ) .
the uncertainty in the number counts due to the poisson error is shown with a pair of lines for each of the surveys . for reference ,
the number count of sxds is superimposed with a thick solid line . from the figure
, we clearly see that the poisson error decreases with increasing survey area . in particular , at the bright magnitudes of @xmath91 mag , a large uncertainty by a factor of @xmath92 in the number count
is found for the hdf - north area ( dotted lines ) .
for the same magnitude range , the uncertainty for a field of view of suprime - cam ( dot - and - dash lines ) is @xmath93 times as small as that for the hdf - north area and @xmath94 times as small as that for the goods - north area . even at the faintest magnitudes of @xmath95 mag , a significant difference in the uncertainty from one to another survey area
thus it is emphasized that the sxds data , which consists of five pointings of suprime - cam field of view , is a useful data set for studies on celestial objects which dominate a large fraction of the number counts without suffering from the poisson error and the field - to - field variation as well , e.g. , a study of evolution of the lsss .
the optical imaging observations of the sxds project are carried out using the suprime - cam on subaru telescope in @xmath96 and @xmath31 bands .
the sxdf has a contiguous area coverage of @xmath641.2 deg@xmath2 , which consists of five suprime - cam pointings .
the photometric zeropoints are determined with an absolute accuracy of no larger than 0.05 mag r.m.s . in the photometry . the r.m.s . of the astrometric accuracies across the field
is on the order of 0.2 arcsec in both ra and dec .
the systematic differences in the adopted world coordinates between different pointings are within about 1 arcsec at the outer edge of the field of view .
thus , our sxds catalogs have a good enough positional accuracies to perform follow - up spectroscopy .
multi - waveband photometric catalogs of detected objects are created for each band and for each pointing .
each of the catalogs contains more than a hundred sixty thousand of objects .
the catalogs have quite homogeneous s / ns across the field . the achieved limiting magnitudes in each band are @xmath97 and @xmath6 ( @xmath76 ) .
the detection completeness as a function of magnitude is estimated by a monte - carlo simulation assuming a gaussian profile .
the number counts of galaxies in the sxdf in each band are computed by correcting the detection completeness , which are consistent with each other among the five pointings with a slight difference at faint magnitudes .
the mean number counts of galaxies averaged over the five pointings show a good agreement with results from previous surveys down to the faint - end magnitude . with the aid of the wide coverage area of sxds data , the uncertainty of the number counts of galaxies due to the poisson error
is greatly decreased .
it is emphasized that the sxds data is extremely useful for pursuing studies on celestial objects spreading in a wide field without suffering from the poisson error and the field - to - field variation .
the sxds catalogs can be applied to studies , scientific objectives of which range from the galactic objects to the large scale structures of the universe .
the optical data , the compiled photometric catalogs , and configuration files used to create the catalogs have been released to the public and can be retrieved from the public data archives server of national astronomical observatory of japan , http://www.naoj.org/science/subaruproject/sxds/index.html .
we thank dr .
masafumi yagi for his kind suggestions in analyzing the data based on the software package developed by himself .
we thank dr .
nobunari kashikawa for fruitful discussions .
the anonymous referee would also be appreciated for his / her constructive suggestions .
we would like to convey our gratitude to the subaru telescope builders and staff for their invaluable help to complete the observing run with suprime - cam .
data reduction / analysis in this work was in part carried out on `` sb '' computer system operated by the astronomical data center ( adc ) and subaru telescope of the national astronomical observatory of japan .
use of the uh 2.2-m telescope for the observations is supported by naoj .
cutri , r. m. et al .
2003 , vizier on - line data catalog : ii/246 ( originally published in university of massachusetts and infrared processing and analysis center ( ipac / california institute of technology ) ) dye , s. et al .
2006 , , * 372 * , 1227 band .
the deviation ( in adu ) of sky background fluctuations per @xmath73-diameter aperture are measured and converted to the s / ns for objects with a magnitude of @xmath98 at each position.,title="fig : " ] band . the deviation ( in adu ) of sky background fluctuations per @xmath73-diameter aperture
are measured and converted to the s / ns for objects with a magnitude of @xmath98 at each position.,title="fig : " ] band .
the deviation ( in adu ) of sky background fluctuations per @xmath73-diameter aperture are measured and converted to the s / ns for objects with a magnitude of @xmath98 at each position.,title="fig : " ] band . the deviation ( in adu ) of sky background fluctuations per @xmath73-diameter aperture
are measured and converted to the s / ns for objects with a magnitude of @xmath98 at each position.,title="fig : " ] band .
the deviation ( in adu ) of sky background fluctuations per @xmath73-diameter aperture are measured and converted to the s / ns for objects with a magnitude of @xmath98 at each position.,title="fig : " ] cccc center & @xmath99 & @xmath100 & north is up + north & @xmath99 & @xmath101 & north is up + south & @xmath99 & @xmath102 & north is up + east & @xmath103 & @xmath100 & east is up + west & @xmath104 & @xmath100 & west is up + cccccclccccccl & 345 & 0.80 & 34.723 & 28.09 & 197,317 & 2002 , sep . 29/30 , oct . 01 + & & & & & & 2003 , nov .
17 + _ @xmath28-n _ & 330 & 0.84 & 34.701 & 28.39 & 176,372 & 2002 , nov . 02/05/27 ,
01/07 + & & & & & & 2003 , nov .
17 + _ @xmath28-s _ & 330 & 0.82 & 34.706 & 28.33 & 189,916 & 2002 , nov .
02/04/05/27 , dec .
01/07 + & & & & & & 2003 , nov .
17 + _ @xmath28-e _ & 330 & 0.82 & 34.698 & 28.06 & 179,478 & 2002 , nov . 04/05 , dec .
01/07 + & & & & & & 2003 , nov .
17 + _ @xmath28-w _ & 330 & 0.78 & 34.716 & 28.21 & 197,770 & 2002 , nov .
04/05/27 , dec .
01/07 + & & & & & & 2003 , nov .
17 + & & & & & & + _ @xmath29-c _ & 319 & 0.72 & 33.639 & 27.78 & 213,851 & 2003 , oct . 02/24/26 + & & & & & & 2004 ,
jan . 16 , oct . 09 , dec . 10/14/15
+ & & & & & & 2005 , jan . 05 + _
@xmath29-n _ & 313 & 0.80 & 33.648 & 27.65 & 203,299 & 2003 , oct .
02/26 , nov .
17 + & & & & & & 2004 , jan .
10/14/15 + & & & & & & 2005 , jan . 05 , sep .
28 + _ @xmath29-s _
& 321 & 0.82 & 33.643 & 27.75 & 186,545 & 2003 , oct . 02/24/26 + & & & & & & 2004 ,
jan . 16 , dec . 10/14/15
+ & & & & & & 2005 , jan .
28 + _ @xmath29-e _ & 291 & 0.76 & 33.639 & 27.77 & 192,951 & 2003 , oct .
02/26 + & & & & & & 2004 , jan .
16 , dec . 10/14/15 + & & & & & & 2005 , jan .
28 + _ @xmath29-w _ & 293 & 0.72 & 33.649 & 27.72 & 205,915 & 2003 , oct .
02/26 , nov .
17 + & & & & & & 2004 , jan .
10/14/15 + & & & & & & 2005 , jan . 05 , sep .
28 + & & & & & & + _ @xmath30-c _ & 248 & 0.76 & 34.315 & 27.57 & 188,079 & 2002 , sep .
01/07 + & & & & & & 2003 , nov .
@xmath30-n _ & 232 & 0.78 & 34.276 & 27.74 & 174,365 & 2002 , nov .
01/09/30 , dec .
06 + & & & & & & 2003 , nov .
17 + _ @xmath30-s _ & 232 & 0.74 & 34.219 & 27.67 & 178,416 & 2002 , nov .
01/09/30 , dec .
06 + & & & & & & 2003 , nov .
17 + _ @xmath30-e _ & 232 & 0.76 & 34.259 & 27.51 & 173,586 & 2002 , nov .
02/09/30 , dec .
06 + & & & & & & 2003 , nov .
17 + _ @xmath30-w _
& 232 & 0.82 & 34.247 & 27.53 & 186,648 & 2002 , nov .
02/09/30 , dec . 06 + & & & & & & 2003 , nov .
17 + & & & & & & + _ @xmath8-c _ & 647 & 0.78 & 34.055 & 27.62 & 181,352 & 2002 , sep . 29/30 , nov .
01/02/05/09/27/29 , dec .
06/07 + & & & & & & 2003 , oct
. 20/21 + _ @xmath8-n _
& 440 & 0.76 & 34.042 & 27.66 & 176,394 & 2002 , sep .
29/30 , nov .
01/02/09/29 , dec .
07 + & & & & & & 2003 , sep .
22 , oct . 02/21 + _
@xmath8-s _ & 309 & 0.68 & 34.046 & 27.47 & 187,791 & 2002 , sep .
29/30 , nov . 01/02/09/29
+ & & & & & & 2003 , sep .
02 + _ @xmath8-e _ & 368 & 0.74 & 33.986 & 27.49 & 175,404 & 2002 , sep .
29/30 , nov .
01/02/09/29 , dec .
07 + & & & & & & 2003 , sep .
22 , oct . 02/21 + _
@xmath8-w _ & 598 & 0.82 & 34.087 & 27.58 & 178,543 & 2002 , sep .
29/30 , nov .
01/02/05/09/27/29 , dec .
06/07 + & & & & & & 2003 , oct .
20/21 + & & & & & & + _ @xmath31-c _ & 217 & 0.70 & 33.076 & 26.57 & 183,324 & 2002 , sep . 29/30 ,
01 , nov . 04/05 + _ @xmath31-n _
& 252 & 0.74 & 32.278 & 26.64 & 163,324 & 2002 , nov .
04/05/10 + & & & & & & 2003 , sep .
21/22 , oct . 01/02 + _ @xmath31-s _ & 184 & 0.76 & 32.258 & 26.39 & 167,779 & 2002 , nov . 04/05/10 + & & & & & & 2003 , sep .
21/22 , oct . 01 + _ @xmath31-e _ & 267 & 0.74 & 32.743 & 26.49 & 160,415 & 2002 , nov . 04/05/10 + & & & & & & 2003 , sep .
21/22 , oct .
01/02 + _ @xmath31-w _ & 311 & 0.74 & 32.776 & 26.60 & 167,748 & 2002 , nov . 04/05/10
+ & & & & & & 2003 , sep .
21/22 , oct .
01/02 + cccc center ( c ) & 0.80 & 1041.9 & 968.4 + north ( n ) & 0.84 & 1011.7 & 951.8 + south ( s ) & 0.82 & 1008.4 & 967.3 + east ( e ) & 0.82 & 986.4 & 917.3 + west ( w ) & 0.82 & 1017.7 & 913.9 + total & ... & 4404.4 & 4057.0 + lcl & & sequential unique i d of the object in the catalog + x & pixels & x coordinate of detection + y & pixels & y coordinate of detection + ra & h : m : s & right ascension of detection + dec & d : m : s & declination of detection + kron_radius & pixels & kron radius @xmath105 defined in bertin & arnouts 1996 + a_image & pixels & 2nd order moment along the major axis + b_image & pixels & 2nd order moment along the minor axis + ellipticity & & isophotal weighted ellipticity of the object ( = a_image / b_image ) + theta_image & degree & position angle of the major axis , counter - clockwised , 0.0=x axis + isoarea_image & pixels & area of the lowest isophote + isophotalmag(b ) & mag & isophotal magnitude in b + isophotalmag(v ) & mag & isophotal magnitude in v + isophotalmag(r ) & mag & isophotal magnitude in rc + isophotalmag(i ) & mag & isophotal magnitude in i + isophotalmag(z ) & mag & isophotal magnitude in z + isophotcormag(b ) & mag & corrected isophotal magnitude in b defined by bertin & arnouts 1996 + isophotcormag(v ) & mag & corrected isophotal magnitude in v defined by bertin & arnouts 1996 + isophotcormag(r ) & mag & corrected isophotal magnitude in rc defined by bertin & arnouts 1996 + isophotcormag(i ) & mag & corrected isophotal magnitude in i defined by bertin & arnouts 1996 + isophotcormag(z ) & mag & corrected isophotal magnitude in z defined by bertin & arnouts 1996 + 2.0aperturemag(b ) & mag & @xmath106 fixed - aperture magnitude in b + 2.0aperturemag(v ) & mag & @xmath106 fixed - aperture magnitude in v + 2.0aperturemag(r ) & mag & @xmath106 fixed - aperture magnitude in rc + 2.0aperturemag(i ) & mag & @xmath106 fixed - aperture magnitude in i + 2.0aperturemag(z ) & mag & @xmath106 fixed - aperture magnitude in z + 3.0aperturemag(b ) & mag & @xmath107 fixed - aperture magnitude in b + 3.0aperturemag(v ) & mag & @xmath107 fixed - aperture magnitude in v + 3.0aperturemag(r ) & mag & @xmath107 fixed - aperture magnitude in rc + 3.0aperturemag(i ) & mag & @xmath107 fixed - aperture magnitude in i + 3.0aperturemag(z ) & mag & @xmath107 fixed - aperture magnitude in z + mag_auto(b ) & mag & automatic - aperture magnitude in b + mag_auto(v ) & mag & automatic - aperture magnitude in v + mag_auto(r ) & mag & automatic - aperture magnitude in rc + mag_auto(i ) & mag & automatic - aperture magnitude in i + mag_auto(z ) & mag & automatic - aperture magnitude in z + mag_best(b ) & mag & mag_auto(b ) if there are no neighbors , or isophotcormag(b ) otherwise + mag_best(v ) & mag & mag_auto(v ) if there are no neighbors , or isophotcormag(v ) otherwise + mag_best(r ) & mag & mag_auto(rc ) if there are no neighbors , or isophotcormag(r ) otherwise + mag_best(i ) & mag & mag_auto(i ) if there are no neighbors , or isophotcormag(i ) otherwise + mag_best(z ) & mag & mag_auto(z ) if there are no neighbors , or isophotcormag(z ) otherwise + err(isophotalmag)(b ) & mag & isophotal magnitude rms error in b + err(isophotalmag)(v ) & mag & isophotal magnitude rms error in v + err(isophotalmag)(r ) & mag & isophotal magnitude rms error in rc + err(isophotalmag)(i ) & mag & isophotal magnitude rms error in i + err(isophotalmag)(z ) & mag & isophotal magnitude rms error in z + err(isophotcormag)(b ) & mag & corrected isophotal magnitude rms error in b + err(isophotcormag)(v ) & mag & corrected isophotal magnitude rms error in v + err(isophotcormag)(r ) & mag & corrected isophotal magnitude rms error in rc + err(isophotcormag)(i ) & mag & corrected isophotal magnitude rms error in i + err(isophotcormag)(z ) & mag & corrected isophotal magnitude rms error in z + err(2.0aperturemag)(b ) & mag & @xmath108 fixed - aperture magnitude rms error in b + err(2.0aperturemag)(v ) & mag & @xmath108 fixed - aperture magnitude rms error in v + err(2.0aperturemag)(r ) & mag & @xmath108 fixed - aperture magnitude rms error in rc + err(2.0aperturemag)(i ) & mag & @xmath108 fixed - aperture magnitude rms error in i + err(2.0aperturemag)(z ) & mag & @xmath108 fixed - aperture magnitude rms error in z + err(3.0aperturemag)(b ) & mag & @xmath109 fixed - aperture magnitude rms error in b + err(3.0aperturemag)(v ) & mag & @xmath109 fixed - aperture magnitude rms error in v + err(3.0aperturemag)(r ) & mag & @xmath109 fixed - aperture magnitude rms error in rc + err(3.0aperturemag)(i ) & mag & @xmath109 fixed - aperture magnitude rms error in i + err(3.0aperturemag)(z ) & mag & @xmath109 fixed - aperture magnitude rms error in z + err(mag_auto)(b ) & mag & automatic - aperture magnitude rms error in b + err(mag_auto)(v ) & mag & automatic - aperture magnitude rms error in v + err(mag_auto)(r ) & mag & automatic - aperture magnitude rms error in rc + err(mag_auto)(i ) & mag & automatic - aperture magnitude rms error in i + err(mag_auto)(z ) & mag & automatic - aperture magnitude rms error in z + err(mag_best)(b ) & mag & mag_best magnitude rms error in b + err(mag_best)(v ) & mag & mag_best magnitude rms error in v + err(mag_best)(r ) & mag & mag_best magnitude rms error in rc + err(mag_best)(i ) & mag & mag_best magnitude rms error in i + err(mag_best)(z ) & mag & mag_best magnitude rms error in z + flux_max(b ) & adu & peak surface brightness in b + flux_max(v ) & adu & peak surface brightness in v + flux_max(r ) & adu & peak surface brightness in rc + flux_max(i ) & adu & peak surface brightness in i + flux_max(z ) & adu & peak surface brightness in z + fwhm_image(b ) & pixels & fwhm of profile from a gaussian fit in b + fwhm_image(v ) & pixels & fwhm of profile from a gaussian fit in v + fwhm_image(r ) & pixels & fwhm of profile from a gaussian fit in rc + fwhm_image(i ) & pixels & fwhm of profile from a gaussian fit in i + fwhm_image(z ) & pixels & fwhm of profile from a gaussian fit in z + class_star(b ) & & stellarity index in b : 0.0=galaxy and 1.0=star + class_star(v ) & & stellarity index in v : 0.0=galaxy and 1.0=star + class_star(r ) & & stellarity index in r : 0.0=galaxy and 1.0=star + class_star(i ) & & stellarity index in i : 0.0=galaxy and 1.0=star + class_star(z ) & & stellarity index in z : 0.0=galaxy and 1.0=star + flags(b ) & & extraction flags in b defined by sextractor + flags(v ) & & extraction flags in v defined by sextractor + flags(r ) & & extraction flags in rc defined by sextractor + flags(i ) & & extraction flags in i defined by sextractor + flags(z ) & & extraction flags in z defined by sextractor + flag_area & & index for affected area : 0=clean , 1=strong_halo , 2=weak_halo + lc detect_minarea & 5 + detect_thresh & 2.0 + analysis_thresh & 2.0 + filter & n + deblend_nthresh & 32 + deblend_mincont & 0.005 + back_size & 32 + back_filtersize & 3 + backphoto_type & global + backphoto_thick & 24 + satur_level & 50000.0 + gain & 2.6 + ccccccc & 0.021 & 0.091 & 0.070 & 0.056 & 0.044 & 0.031 + _
sxds - n _ & 0.019 & 0.083 & 0.064 & 0.052 & 0.040 & 0.029 + _ sxds - s _ & 0.023 & 0.100 & 0.077 & 0.062 & 0.048 & 0.034 + _ sxds - e _ & 0.021 & 0.091 & 0.070 & 0.056 & 0.044 & 0.031 + _
sxds - w _ & 0.021 & 0.091 & 0.070 & 0.057 & 0.044 & 0.031 + ccccccccccc _ c - n _ & -0.04 & -0.02 & -0.04 & -0.01 & -0.03 & -0.01 & + 0.03 & + 0.02 & -0.02 & -0.01 + _ c - s _ & -0.05 & -0.02 & -0.02 & -0.00 & -0.03 & 0.00 & + 0.03 & + 0.00 & + 0.00 & + 0.01 + _ c - e _ & -0.02 & -0.01 & -0.02 & -0.01 & -0.01 & -0.00 & + 0.05 & + 0.01 & -0.01 & -0.00 + _ c - w _ & + 0.01 & + 0.01 & -0.03 & -0.01 & + 0.01 & + 0.00 & -0.01 & -0.00 & -0.01 & -0.01 + | we present multi - waveband optical imaging data obtained from observations of the subaru / xmm - newton deep survey ( sxds ) .
the survey field , centered at r.a .
= @xmath0 , decl .
= @xmath1 , has been the focus of a wide range of multi - wavelength observing programs spanning from x - ray to radio wavelengths . a large part of the optical imaging observations
are carried out with suprime - cam on subaru telescope at mauna kea in the course of subaru telescope `` observatory projects '' .
this paper describes our optical observations , data reduction and analysis procedures employed , and the characteristics of the data products .
a total area of 1.22 deg@xmath2 is covered in five contiguous sub - fields , each of which corresponds to a single suprime - cam field of view ( @xmath3 ) , in five broad - band filters @xmath4 to the depths of @xmath5 and @xmath6 ( @xmath7 ) .
the data are reduced and compiled into five multi - waveband photometric catalogs , separately for each suprime - cam pointing .
the @xmath8-band catalogs contain about 900,000 objects , making the sxds catalogs one of the largest multi - waveband catalogs in corresponding depth and area coverage .
the sxds catalogs can be used for an extensive range of astronomical applications such as the number density of the galactic halo stars to the large scale structures at the distant universe .
the number counts of galaxies are derived and compared with those of existing deep extragalactic surveys .
the optical data , the source catalogs , and configuration files used to create the catalogs are publicly available via the sxds web page ( http://www.naoj.org/science/subaruproject/sxds/index.html ) . |
in many optimization problems , one is given a ( say , geometric ) input , and one is interested in computing the minimum of a function over this input .
such a function can be , for example , the minimum cost clustering of the input , the price of a minimum spanning tree , the radius of the smallest enclosing disk , the closest pair distance , and many other functions .
often for such optimization problems it is possible to construct a decision procedure , which given a query value can decide whether the query is smaller or larger than the minimum of the function .
naturally , one would like to use this decider to perform a binary search to compute the minimum . however , often this is inherently not possible as the set of possible solutions is a real interval .
instead one must identify a set of critical values at which the function changes . naturally , searching over these values
directly can be costly as often the number of such critical values is much larger than the desired running time .
instead one attempts to perform an implicit search over them .
one of the most powerful techniques to solve optimization problems efficiently in computational geometry , using such an implicit search , is the technique of megiddo @xcite .
it is relatively complicated , as it involves implicitly extracting values from a simulation of a parallel decision procedure ( often a parallel sorting algorithm ) .
for this reason it is inherently not possible for parametric search to lead to algorithms which run faster than @xmath3 time .
nevertheless , it is widely used in designing efficient geometric optimization algorithms , see @xcite .
luckily , in many cases one can replace parametric search by simpler techniques ( see prune - and - search below for example ) and in particular , it can be replaced by randomization , see the work by van oostrum and veltkamp @xcite .
another example of replacing parametric search by randomization is the new simplified algorithm for the frchetdistance @xcite .
surprisingly , sometimes these alternative techniques can actually lead to linear time algorithms .
[ [ linear - time - algorithms . ] ] linear time algorithms .
+ + + + + + + + + + + + + + + + + + + + + + + there seems to be three main ways to get linear time algorithms for geometric optimization problems ( exact or approximate ) : * coreset / sketch*. one can quickly extract a compact sketch of the input that contains the desired quantity ( either exactly or approximately ) . as an easy example , consider the problem of computing the axis parallel bounding box of a set of points - an easy linear scan suffices .
there is by now a robust theory of what quantities one can extract a coreset of small size for , such that one can do the ( approximate ) calculation on the coreset , where usually the coreset size depends only on the desired approximation quality .
this leads to many linear time algorithms , from shape fitting @xcite , to @xmath4-approximate @xmath0-center / median / mean clustering @xcite in constant dimension , and many other problems @xcite .
the running times of the resulting algorithms are usually @xmath5 .
the limitation of this technique is that there are problems for which there is no small sketch , from clustering when the number of clusters is large , to problems where there is no sketch at all @xcite for example , for finding the closest pair of points one needs all the given input and no sketching is possible .
* prune and search*. here one prunes away a constant fraction of the input , and continues the search recursively on the remaining input .
the paramount example of such an algorithm is the linear time median finding algorithm , however , one can interpret many of the randomized algorithms in computational geometry as pruning algorithms @xcite .
for example , linear programming in constant dimension in linear time @xcite , and its extension to -type problems @xcite . intuitively ,
-type problems include low dimensional convex programming ( a standard example is the smallest enclosing ball of a point set in constant dimension ) .
however , surprisingly , such problems also include problems that are not convex in nature for example , deciding if a set of ( axis parallel ) rectangles can be pierced by three points is an -type problem .
other examples of prune - and - search algorithms that work in linear time include computing an ear in a triangulation of a polygon @xcite , searching in sorted matrices @xcite , and ham - sandwich cuts in two dimensions @xcite
. of course , there are many other examples of using prune and search with running time that is super - linear .
* grids . *
rabin @xcite used randomization , the floor function , and hashing to compute the closest pair of a set of points in the plane , in expected linear time .
et al . _
@xcite presented a simplified version of this algorithm , and smid provides a survey of algorithms on closest pair problems @xcite . a prune and search variant of this algorithm
was suggested by khuller and matias @xcite .
of course , the usage of grids and hashing to perform approximate point - location is quite common in practice . by itself , this is already sufficient to break lower bounds in the comparison model , for example for @xmath0-center clustering @xcite . the only direct extension of rabin s algorithm the authors are aware of is the work by har - peled and mazumdar @xcite showing a linear time @xmath6-approximation to the smallest ball containing @xmath0 points ( out of @xmath7 given points ) .
there is some skepticism of algorithms using the floor function , since schnhage@xcite showed how to solve a complete problem , in polynomial time , using the floor function in the real model the trick being packing many numbers into a single word ( which can be arbitrarily long in the model , and still each operation on it takes only constant time ) .
note , that as rabin s algorithm does not do any such packing of numbers ( i.e. , its computation model is considerably more restricted ) , this criticism does not seem to be relevant in the case of this algorithm and its relatives . in this paper , we present a new technique that combines together all of the above techniques to yield linear time approximation algorithms . [
[ nets . ] ] nets .
+ + + + + given a point set @xmath8 , an @xmath9-net @xmath10 of @xmath8 is a subset of @xmath8 that represents well the structure of @xmath8 in resolution @xmath9 .
formally , we require that for any point in @xmath8 there is a net point in distance at most @xmath9 from it , and no two net points are closer than @xmath9 to each other , see section [ sec : compnets ] for a formal definition .
thus , nets provide a sketch of the point - set as far as distances that are @xmath9 or larger .
nets are a useful tool in presenting point - sets hierarchically . in particular , computing nets of different resolutions and linking between different levels , leads to a tree like data - structure that can be used to facilitate many tasks , see for example the net - tree @xcite for such a data - structure for doubling metrics .
nets can be defined in any metric space , but in euclidean space a grid can sometimes provide an equivalent representation .
in particular , net - trees can be interpreted as an extension of ( compressed ) quadtrees to more abstract settings .
computing nets is closely related to @xmath0-center clustering .
specifically , gonzalez @xcite shows how to compute an approximate net that has @xmath0 points in @xmath11 time , which is also a @xmath6-approximation to the @xmath0-center clustering .
this was later improved to @xmath12 time , for low dimensional euclidean space @xcite , if @xmath0 is sufficiently small ( using grids and hashing ) .
har - peled and mendel showed how to preprocess a point set in a metric space with constant doubling dimension , in @xmath13 time , such that an ( approximate ) @xmath9-net can be extracted in ( roughly ) linear time in the size of the net . in this paper , we consider problems of the following form : given a set @xmath8 of weighted points in @xmath14 , one wishes to solve an optimization problem whose solution is one of the pairwise distances of @xmath8 ( or `` close '' to one of these values ) .
problems of this kind include computing the optimal @xmath0-center clustering , or the length of the @xmath0thedge in the @xmath2of @xmath8 , and many others .
specifically , we are interested in problems for which there is a fast approximate decider .
that is , given a value @xmath15 we can , in linear time , decide if the desired value is ( approximately ) smaller than @xmath9 or larger than @xmath9 .
the goal is then to use this decider to approximate the optimum solution in linear time . as a first step towards a linear time approximation algorithm for such problems , we point out that one can compute nets in linear time in @xmath14 ,
see section [ sec : compnets ] . however , even if we could implicitly search over the critical values ( which we can not ) then we still would require a logarithmic number of calls to the decider which would yield a running time of @xmath3 , as the number of critical values is at least linear ( and usually polynomial ) .
so instead we use the return values of the decision procedure as we search to thin out the data so that future calls to the decision procedure become successively cheaper .
however , we still can not search over the critical values ( since there are too many of them ) and so we also introduce random sampling in order to overcome this .
[ [ outline - of - the - new - technique . ] ] outline of the new technique .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + the new algorithm works by randomly sampling a point and computing the distance to its nearest neighbor .
let this distance be @xmath9 .
next , we use the decision procedure to decide if we are in one of the following three cases .
if @xmath9 is too small then we zoom out to a resolution of @xmath9 by computing an @xmath9-net and continuing the computation on the net instead of on the original point - set .
that is , we net the point - set into a smaller point - set , such that one can solve the original problem ( approximately ) on this smaller sketch of the input
. * prune . *
if @xmath9 is too large then we remove all points whose nearest neighbor is further than @xmath9 away ( of course , this implies we should only consider problems for which such pruning does not affect the solution ) .
that is , we isolate the optimal solution by pruning away irrelevant data this is similar in nature to what is being done by prune - and - search algorithms .
the value of @xmath9 is the desired approximation .
we then continue recursively on the remaining data . in either case ,
the number of points being handled ( in expectation ) goes down by a constant factor and thus the overall expected running time is linear .
[ [ significance - of - results . ] ] significance of results .
+ + + + + + + + + + + + + + + + + + + + + + + + our basic framework is presented in a general enough manner to cover , and in many cases greatly simplify , many problems for which linear time algorithms had already been discovered . at the same time the framework provides new linear time algorithms for a large collection of problems , for which previously no linear time algorithm was known .
the framework should also lead to algorithms for many other problems which are not mentioned . at a conceptual level
the basic algorithm is simple enough ( with its basic building blocks already having efficient implementations ) to be highly practical from an implementation standpoint . perhaps more importantly , with increasing shifts toward large data sets algorithms with super linear running time
can be impractical .
additionally , our framework seems amenable to distributed implementation in frameworks like mapreduce .
indeed , every iteration of our algorithm breaks the data into grid cells , a step that is similar to the map phase .
in addition , the aggressive thinning of the data by the algorithm guarantees that after the first few iterations the algorithm is resigned to working on only a tiny fraction of the data .
[ [ significance - of - the - netting - step . ] ] significance of the netting step .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + without the net stage , our framework is already sufficient to solve the closest - pair problem , and in this specific case , the algorithm boils down to the one provided by khuller and matias @xcite .
this seems to be the only application of the new framework where the netting is not necessary . in particular
, the wide applicability of the framework seems to be the result of the netting step .
[ [ framework - and - results . ] ] framework and results .
+ + + + + + + + + + + + + + + + + + + + + + we provide a framework that classifies which optimization problems can be solved using the new algorithm .
we get the following new algorithms ( all of them have an expected linear running time , for any fixed @xmath16 ) : * @xmath0-center clustering * ( section [ sec : k : center ] ) .
we provide an algorithm that @xmath6-approximates the optimal @xmath0-center clustering of a point set in @xmath14 .
unlike the previous algorithm @xcite that was restricted to @xmath17 , the new algorithm works for any value of @xmath0 .
this new algorithm is also simpler . *
@xmath0thsmallest distance * ( section [ sec : k : th : distance ] ) . in the distance selection problem , given a set of points in @xmath14 , one would like to compute the @xmath0thsmallest distance defined by a pair of points of @xmath8 .
it is believed that such exact distance selection requires @xmath18 time in the worst case @xcite , even in the plane ( in higher dimensions the bound deteriorates ) .
we present an @xmath19 time algorithm that @xmath4-approximates the @xmath0thsmallest distance .
previously , bespamyatnikh and segal @xcite presented @xmath20 time algorithm using a well - separated pairs decomposition ( see also @xcite ) . given two sets of points @xmath8 and @xmath21 with a total of @xmath7 points , using the same approach , we can @xmath4-approximate the @xmath0thsmallest distance in a bichromatic set of distances @xmath22{0cm}{0.4cm}}\right . } } \right\}}$ ] , and in particular , we can compute exactly the closest bichromatic pair between @xmath8 and @xmath23 . intuitively , the mechanism behind distance selection underlines many optimization problems , as it ( essentially ) performs a binary search over the distances induced by a set of points . as such ,
being able to do approximate distance selection in linear time should lead to faster approximation algorithms that perform a similar search over such distances . * the @xmath0thsmallest @xmath1-nearest neighbor distance * ( section [ sec : k : smallest : m : n : n ] ) . for a set @xmath24 of @xmath7 points , and a point @xmath25 , its in @xmath8 is the @xmath1thclosest point to @xmath26 in @xmath27 . in particular
, let @xmath28 denote this distance . here , consider the set of these distances defined for each point of @xmath8 ; that is , @xmath29 .
we can approximate the @xmath0th smallest number in this set in linear time . * exact nearest neighbor distances * ( section [ sec : f : n : n ] ) .
for the special case where @xmath30 , one can turn the above approximation into an exact computation of the @xmath0thnearest neighbor distance with only a minor post processing grid computation . as an application , when @xmath31 , one can compute in linear time , exactly , the furthest nearest neighbor distance ; that is , the nearest neighbor distance of the point whose nearest neighbor is furthest away .
this measure can be useful , for example , in meshing applications where such a point is a candidate for a region where the local feature size is large and further refinement is needed .
we are unaware of any previous work directly on this problem , although one can compute this quantity by solving the all - nearest - neighbor problem , which can be done in @xmath32 time @xcite .
this is to some extent the `` antithesis '' to rabin s algorithm for the closest pair problem , and it is somewhat surprising that it can also be solved in linear time . * the @xmath0thlongest @xmath2edge * ( section [ sec : longest : mst ] ) . given a set @xmath8 of @xmath7 points in @xmath14 , we can @xmath4-approximate , in @xmath19 time , the @xmath0thlongest edge in the @xmath2of @xmath8 . * smallest ball with a monotone property * ( section [ sec : min : cluster ] ) .
consider a property defined over a set of points , @xmath8 , that is monotone ; that is , if @xmath33 has this property then @xmath23 must also have this property .
consider such a property that can be easily checked , for example , ( i ) whether the set contains @xmath0 points , or ( ii ) if the points are colored , whether all colors are present in the given point set .
given a point set @xmath8 , one can @xmath4-approximate , in @xmath34 time , the smallest radius ball @xmath35 , such that @xmath36 has the desired property .
for example , we get a linear time algorithm to approximate the smallest ball enclosing @xmath0 points of @xmath8 . the previous algorithm for this problem
@xcite was significantly more complicated .
furthermore , we can approximate the smallest ball such that all colors appear in it ( if the points are colored ) , or the smallest ball such that at least @xmath37 different colors appear in it , etc .
more generally , the kind of monotone properties supported are _
sketchable _ ; that is , properties for which there is a small summary of a point - set that enables one to decide if the property holds , and furthermore , given summaries of two disjoint point sets , the summary of the union point - set can be computed in constant time .
we believe that formalizing this notion of sketchability is a useful abstraction .
see section [ sec : sketchable ] for details . *
smallest connected component with a monotone property * ( section [ sec : min : cluster : c ] ) .
consider the connected components of the graph where two points are connected if they are distance at most @xmath9 from each other .
using our techniques , one can approximate , in linear time , the smallest @xmath9 such that there is a connected component of this graph for which a required sketchable property holds for the points in this connected component . as an application , consider ad hoc wireless networks . here
, we have a set @xmath8 of @xmath7 nodes and their locations ( say in the plane ) , and each node can broadcast in a certain radius @xmath9 ( the larger the @xmath9 the higher the energy required , so naturally we would like to minimize it ) .
assume there are two special nodes .
it is natural to ask for the minimum @xmath9 , such that there is a connected component of the above graph that contains both nodes .
that is , these two special nodes node can send message to each other , by message hopping ( with distance at most @xmath9 at each hop ) .
we can approximate this connectivity radius in linear time . * clustering for a monotone property * ( section [ sec : cluster : monotone ] ) .
imagine that we want to break the given point - set into clusters , such that the maximum diameter of a cluster is minimized ( as in @xmath0-center clustering ) , and furthermore , the points assigned to each cluster comply with some sketchable monotone property .
we present a @xmath38-approximation algorithm for these types of problems , that runs in @xmath34 time .
this includes lower bounded clustering ( i.e. , every cluster must contain at least @xmath39 points ) , for which the authors recently presented an @xmath40 approximation algorithm @xcite .
one can get a @xmath6-approximation using network flow , but the running is significantly worse @xcite . see section [ sec : cluster : monotone : examples ] for examples of clustering problems that can be approximated using this algorithm . * connectivity clustering for a monotone property * ( section [ sec : cluster : spanning ] ) .
consider the problem of computing the minimum @xmath9 , such that each connected component ( of the graph where points in distance at most @xmath9 from each other are adjacent ) has some sketchable monotone property .
we approximate the minimum @xmath9 for which this holds in linear time .
an application of this for ad hoc networks is the following we have a set @xmath8 of @xmath7 wireless clients , and some of them are base stations ; that is , they are connected to the outside world .
we would like to find the minimum @xmath9 , such that each connected component of this graph contains a base station .
* closest pair and smallest non - zero distance * ( section [ sec : closest : non : zero : dist ] ) . given a set of points in @xmath14 ,
consider the problem of finding the smallest non - zero distance defined by these points .
this problem is an extension of the closest pair distance , as there might be many identical points in the given point set .
we provide a linear time algorithm for computing this distance _ exactly _ , which follows easily from our framework .
[ [ high - probability . ] ] high probability .
+ + + + + + + + + + + + + + + + + finally , we show in section [ sec : high : prob ] how to modify our framework such that the linear running time holds with high probability .
since there is very little elbow room in the running time when committed to linear running time , this extension is quite challenging , and requires several new ideas and insights .
see section [ sec : high : prob ] for details .
[ [ paper - organization . ] ] paper organization .
+ + + + + + + + + + + + + + + + + + + we describe how to compute nets in section [ sec : prelim ] , and how to remove faraway points efficiently in section [ sec : point : removal ] .
we define the abstract framework , and describe and analyze the new approximation algorithm , in section [ sec : framework ] .
we describe the applications in section [ sec : applications ] .
we show how to modify the framework to achieve linear running time with high probability in section [ sec : high : prob ] .
we conclude in section [ sec : conclusions ] .
[ sec : prelim ] the of an interval @xmath41 \subseteq { { \rm i\!\hspace{-0.025em } r}}^{+}$ ] is @xmath42 .
[ def : net ] for a point set @xmath8 in a metric space with a metric @xmath43 , and a parameter @xmath15 , an @xmath9- of @xmath8 is a subset @xmath44 , such that for every @xmath45 , @xmath46 , we have that @xmath47 , and for all @xmath25 , we have that @xmath48 . intuitively , an @xmath9-net represents @xmath8 in resolution @xmath9 .
[ def : grid : stuff ] for a real positive number @xmath49 and a point @xmath50 , define @xmath51 to be the grid point @xmath52 .
we call @xmath49 the or of the @xmath53 . observe that @xmath53 partitions @xmath14 into cubes , which are grid . the grid cell of @xmath26 is uniquely identified by the integer point @xmath54{0.0cm}{0.38cm}}\ ! { \left\lfloor { { { \mathsf{p}}}_1/\delta } \right\rfloor } , \ldots , { \left\lfloor { { { \mathsf{p}}}_d/\delta } \right\rfloor}}\right)}$ ] . for a number @xmath55 ,
let @xmath56 denote the set of grid cells in distance @xmath57 from @xmath26 , which is the of @xmath26 .
note , that the neighborhood also includes the grid cell containing @xmath26 itself , and if @xmath58 then @xmath59 .
see figure on the right .
[ sec : compnets ] there is a simple algorithm for computing @xmath9-nets .
namely , let all the points in @xmath8 be initially unmarked . while there remains an unmarked point , @xmath26 , add @xmath26 to @xmath60 , and mark it and all other points in distance @xmath61 from @xmath26 ( i.e. we are scooping away balls of radius @xmath9 ) . by using grids and hashing one can modify this algorithm to run in linear time .
the following is implicit in previous work @xcite , and we include it here for the sake of completeness , and does not work in this settings , as the number of clusters it can handle is limited to @xmath62 .
lemma [ lemma : net ] has no such restriction . ]
it was also described by the authors in @xcite .
[ lemma : net ] given a point set @xmath63 of size @xmath7 and a parameter @xmath15 , one can compute an @xmath9-net for @xmath8 in @xmath12 time .
let @xmath64 denote the grid in @xmath14 with side length @xmath65 .
first compute for every point @xmath66 the grid cell in @xmath64 that contains @xmath26 ; that is , @xmath67 .
let @xmath68 denote the set of grid cells of @xmath64 that contain points of @xmath8 .
similarly , for every cell @xmath69 we compute the set of points of @xmath8 which it contains .
this task can be performed in linear time using hashing and bucketing assuming the floor function can be computed in constant time .
specifically , store the @xmath70 values in a hash table , and in constant time hash each point into its appropriate bin . scan the points of @xmath8 one at a time , and let @xmath26 be the current point . if @xmath26 is marked then move on to the next point . otherwise , add @xmath26 to the set of net points , @xmath60 , and mark it and each point @xmath71 such that @xmath72 .
since the cells of @xmath56 contain all such points , we only need to check the lists of points stored in these grid cells . at the end of this procedure every point is marked .
since a point can only be marked if it is in distance @xmath61 from some net point , and a net point is only created if it is unmarked when visited , this implies that @xmath60 is an @xmath9-net .
as for the running time , observe that a grid cell , @xmath73 , has its list scanned only if @xmath73 is in the neighborhood of some created net point . as @xmath74
, there are only @xmath75 cells which could contain a net point @xmath26 such that @xmath76 .
furthermore , at most one net point lies in a single cell since the diameter of a grid cell is strictly smaller than @xmath9 .
therefore each grid cell had its list scanned @xmath75 times .
since the only real work done is in scanning the cell lists and since the cell lists are disjoint , this implies an @xmath12 running time overall .
observe that the closest net point , for a point @xmath25 , must be in one of its neighborhood s grid cells . since every grid cell can contain only a single net point , it follows that in constant time per point of @xmath8 , one can compute each point s nearest net point .
we thus have the following . [ cor : valid ] for a set @xmath63 of @xmath7 points , and a parameter @xmath77 , one can compute , in linear time , an @xmath9-net of @xmath8 , and furthermore , for each net point the set of points of @xmath8 for which it is the nearest net point . in the following ,
a is a point that is assigned a positive integer weight . for any subset @xmath78 of a weighted point set @xmath8 , let @xmath79 denote the number of points in @xmath78 and
let @xmath80 denote the total weight of @xmath78 .
in particular , corollary [ cor : valid ] implies that for a weighted point set one can compute the following quantity in linear time .
[ algorithm : net ] [ algorithm : net ] given a weighted point set @xmath63 , let @xmath81 { { { \textcolor[named]{redviolet}{\texttt{\bf{net } } } } } \index{algorithm!net@{{{\textcolor[named]{redviolet}{\texttt{\bf{net } } } } } } } } } \xspace}{\!\left({r , { \mathsf{p}}}\right)}}$ ] denote an @xmath9-net of @xmath8 , where the weight of each net point @xmath26 is the total sum of the weights of the points assigned to it .
we slightly abuse notation , and also use @xmath81 { { { \textcolor[named]{redviolet}{\texttt{\bf{net } } } } } \index{algorithm!net@{{{\textcolor[named]{redviolet}{\texttt{\bf{net } } } } } } } } } \xspace}{\!\left({r , { \mathsf{p}}}\right)}}$ ] to designate the algorithm computing this net , which has linear running time .
[ sec : point : removal ] for a given point @xmath25 , let @xmath82 denote the nearest neighbor of @xmath26 in @xmath83 , and the distance to it , respectively .
the quantity @xmath84 can be computed ( naively ) in linear time by scanning the points . for a set of points @xmath8 , and a parameter @xmath9 , let @xmath85 denote the set of points ; that is , it is the set of all points @xmath25 , such that the nearest - neighbor of @xmath26 in @xmath86 is at least distance @xmath9 away ( i.e. , @xmath87 ) .
similarly , @xmath88 is the set of points ; that is , all points @xmath25 , such that @xmath89 .
[ algorithm : delete ] [ lemma : delete ] given a weighted set @xmath63 of @xmath7 points , and a distance @xmath90 , in @xmath91 time , one can compute the sets @xmath92 and @xmath93 .
let @xmath94 { { { \textcolor[named]{redviolet}{\texttt{\bf{delfar } } } } } \index{algorithm!delfar@{{{\textcolor[named]{redviolet}{\texttt{\bf{delfar } } } } } } } } } \xspace}({\ell } , { \mathsf{p}})$ ] denote the algorithm which computes these sets and returns @xmath92 . build a grid where every cell has diameter @xmath95 , for some constant @xmath96 .
clearly any point @xmath25 such that @xmath97 ( i.e. , a far point ) must be in a grid cell by itself .
therefore to determine all such `` far '' points only grid cells with singleton points in them need to be considered .
for such a point @xmath98 , to determine if @xmath99 , one checks the points stored in all grid cells in distance @xmath100 from it .
if @xmath98 has no such close neighbor , we mark @xmath98 for inclusion in @xmath85 . by the same arguments as in lemma [ lemma : net ] ,
the number of such cells is @xmath75 .
again by the arguments of lemma [ lemma : net ] every non - empty grid cell gets scanned @xmath75 times overall , and so the running time is @xmath12 . finally , we copy all the marked ( resp .
unmarked ) points to @xmath93 ( resp .
@xmath92 ) .
the algorithm @xmath94 { { { \textcolor[named]{redviolet}{\texttt{\bf{delfar } } } } } \index{algorithm!delfar@{{{\textcolor[named]{redviolet}{\texttt{\bf{delfar } } } } } } } } } \xspace}({\ell } , { \mathsf{p}})$ ] then simply returns @xmath92 .
[ sec : framework ] let @xmath8 and @xmath23 be two sets of weighted points in @xmath14 ( of the same weight ) .
the set @xmath23 is a of @xmath8 , if @xmath23 can be constructed by moving each point of @xmath8 by distance at most @xmath49 ( and not altering its weight ) .
formally , there is an onto mapping @xmath101 , such that for @xmath25 , we have that @xmath102 , and for any @xmath103 , we have that @xmath104 . note that for a ( potentially weighted ) point set @xmath21 , @xmath81 { { { \textcolor[named]{redviolet}{\texttt{\bf{net } } } } } \index{algorithm!net@{{{\textcolor[named]{redviolet}{\texttt{\bf{net } } } } } } } } } \xspace}{\!\left({\delta , { \mathsf{w}}}\right)}}$ ] is a @xmath49-translation of @xmath21 .
[ def : decider ] given a function @xmath105 , we call a procedure , @xmath106{redviolet}{\texttt{\bf{{decider}}}}}}\index{algorithm!{decider}@{{{\textcolor[named]{redviolet}{\texttt{\bf{{decider}}}}}}}}}}\xspace}$ ] , a for @xmath107 , if for any @xmath108 and @xmath77 , @xmath106{redviolet}{\texttt{\bf{{decider}}}}}}\index{algorithm!{decider}@{{{\textcolor[named]{redviolet}{\texttt{\bf{{decider}}}}}}}}}}\xspace}(r , x)$ ] returns one of the following : @xmath109 $ ] , where @xmath39 is some real number , @xmath110 , or @xmath111 .
[ def : n : d : p ] an instance of a consists of a pair @xmath112 , where @xmath113 is a set of @xmath7 distinct weighted points is a multiset can also be handled .
see remark [ rem : multi ] .
] , and @xmath114 is the of the given instance ( of size @xmath12 ) and consists of the relevant parameters for the problem . for a fixed , the task is to evaluate a function @xmath115 , defined over such input pairs , that has the following properties : : there exists an @xmath12 time @xmath116-decider for @xmath117 , for some _ constant _ @xmath116 . :
let @xmath23 be any @xmath49-translation of @xmath21 .
then @xmath118 . : if @xmath119 then @xmath120 , where @xmath121 is the set of @xmath9-close points , and @xmath122 is an updated context which can be computed in @xmath12 time .
if we are interested in a @xmath4-approximation , then @xmath123 . in this case , we require that the running time of the provided @xmath124)-decider ( i.e. , the property above ) is @xmath125 , for some constant @xmath126 .
this covers all the applications presented in this paper .
our analysis still holds even if the decider has a different running time , but the overall running time might increase by a factor of @xmath127 , see lemma [ lemma : result:1 ] @xmath128 for details . [
sec : kcenter ] as a concrete example of an , consider the problem of @xmath0-center clustering .
let @xmath21 be a set of points in @xmath14 , and let @xmath129 be an integer parameter .
find a set of @xmath130 such that the maximum distance of a point of @xmath21 to its nearest center in @xmath131 is minimized , and @xmath132 .
specifically , the function of interest , denoted by @xmath133 , is the radius of the optimal @xmath0-center clustering of @xmath21 .
we now show that satisfies the properties of an .
[ lemma : decider:1 ] the following relations between @xmath134 { { { \textcolor[named]{redviolet}{\texttt{\bf{net } } } } } \index{algorithm!net@{{{\textcolor[named]{redviolet}{\texttt{\bf{net } } } } } } } } } \xspace}{\!\left({r , { \mathsf{w}}}\right ) } } } \right|}$ ] and @xmath133 hold : if we have @xmath135 { { { \textcolor[named]{redviolet}{\texttt{\bf{net } } } } } \index{algorithm!net@{{{\textcolor[named]{redviolet}{\texttt{\bf{net } } } } } } } } } \xspace}{\!\left({r , { \mathsf{w}}}\right ) } } } \right| } \leq k$ ] then @xmath136 . if @xmath134 { { { \textcolor[named]{redviolet}{\texttt{\bf{net } } } } } \index{algorithm!net@{{{\textcolor[named]{redviolet}{\texttt{\bf{net } } } } } } } } } \xspace}{\!\left({r , { \mathsf{w}}}\right ) } } } \right|}\allowbreak > k$ ] then @xmath137 .
\(a ) observe that if @xmath134 { { { \textcolor[named]{redviolet}{\texttt{\bf{net } } } } } \index{algorithm!net@{{{\textcolor[named]{redviolet}{\texttt{\bf{net } } } } } } } } } \xspace}{\!\left({r , { \mathsf{w}}}\right ) } } } \right|}\leq k$ ] , then , by definition , the set of net points of @xmath81 { { { \textcolor[named]{redviolet}{\texttt{\bf{net } } } } } \index{algorithm!net@{{{\textcolor[named]{redviolet}{\texttt{\bf{net } } } } } } } } } \xspace}{\!\left({r , { \mathsf{w}}}\right)}}$ ] are a set of @xmath138 centers such that all the points of @xmath21 are in distance strictly less than @xmath9 from these centers .
\(b ) if @xmath134 { { { \textcolor[named]{redviolet}{\texttt{\bf{net } } } } } \index{algorithm!net@{{{\textcolor[named]{redviolet}{\texttt{\bf{net } } } } } } } } } \xspace}{\!\left({r , { \mathsf{w}}}\right ) } } } \right| } > k$ ] then @xmath21 must contain a set of @xmath139 points whose pairwise distances are all at least @xmath9 . in particular any solution to with radius @xmath140 would not be able to cover all these @xmath139 points with only @xmath0 centers .
[ lemma : k : center : wbdp ] an instance @xmath141 of satisfies the properties of definition [ def :
n : d : p ] ; that is , is a @xmath38- , for any @xmath142 .
we need to verify the required properties , see definition [ def : n : d : p ] @xmath143 .
we need to describe a decision procedure for clustering . to this end , given a distance @xmath9 , the decider first calls @xmath81 { { { \textcolor[named]{redviolet}{\texttt{\bf{net } } } } } \index{algorithm!net@{{{\textcolor[named]{redviolet}{\texttt{\bf{net } } } } } } } } } \xspace}{\!\left({r , { \mathsf{w}}}\right)}}$ ] , see algorithm [ algorithm : net ] @xmath144 .
if we have @xmath134 { { { \textcolor[named]{redviolet}{\texttt{\bf{net } } } } } \index{algorithm!net@{{{\textcolor[named]{redviolet}{\texttt{\bf{net } } } } } } } } } \xspace}{\!\left({r , { \mathsf{w}}}\right ) } } } \right| } \leq k$ ] , then by lemma [ lemma : decider:1 ] the answer `` @xmath136 '' can be returned .
otherwise , call @xmath81 { { { \textcolor[named]{redviolet}{\texttt{\bf{net } } } } } \index{algorithm!net@{{{\textcolor[named]{redviolet}{\texttt{\bf{net } } } } } } } } } \xspace}{\!\left({(2+{\varepsilon}/2)r , { \mathsf{w}}}\right)}}$ ] .
if @xmath134 { { { \textcolor[named]{redviolet}{\texttt{\bf{net } } } } } \index{algorithm!net@{{{\textcolor[named]{redviolet}{\texttt{\bf{net } } } } } } } } } \xspace}{\!\left({(2+{\varepsilon}/2)r , { \mathsf{w}}}\right ) } } } \right|}\leq k$ ] , then , by lemma [ lemma : decider:1 ] , we have @xmath145 and `` @xmath146 $ ] '' is returned . otherwise @xmath134 { { { \textcolor[named]{redviolet}{\texttt{\bf{net } } } } } \index{algorithm!net@{{{\textcolor[named]{redviolet}{\texttt{\bf{net } } } } } } } } } \xspace}{\!\left({(2+{\varepsilon}/2)r , { \mathsf{w}}}\right ) } } } \right| } > k$ ] , and lemma [ lemma : decider:1 ] implies that the answer `` @xmath147 '' can be returned by the decider . observe that if @xmath23 is a @xmath49-translation of @xmath21 , then a point and its respective center in a @xmath0-center clustering of @xmath21 each move by distance at most @xmath49 in the transition from @xmath21 to @xmath23 . as such , the distance between a point and its center changes by at most @xmath148 by this process .
this argument also works in the other direction , implying that the @xmath0-center clustering radius of @xmath21 and @xmath23 are the same , up to an additive error of @xmath148 .
it suffices to show that if @xmath136 then @xmath149 .
now , if @xmath136 then any point of @xmath21 whose neighbors are all @xmath150 away must be a center by itself in the optimal @xmath0-center solution , as otherwise it would be assigned to a center @xmath151 away . similarly , any point assigned to it would be assigned to a point @xmath152 away . therefore , for any point @xmath153
whose nearest neighbor is at least distance @xmath9 away , @xmath154 . repeating this observation implies the desired result .
the context update ( and computing @xmath121 and @xmath155 ) can by done in linear time using , see lemma [ lemma : delete ] @xmath156 .
[ sec : linear ] [ algorithm : ndp : alg ] [ figure : a : algorithm ] we now describe the general algorithm which given a @xmath116- , @xmath112 , and an associated target function , @xmath107 , computes in linear time a @xmath157 spread interval containing @xmath158 . in the following ,
let denote the given @xmath116-decider , where @xmath159 is some parameter .
also , let denote the context updater associated with the given problem , see definition [ def : n : d : p ] . both and run in linear time . the algorithm for bounding the optimum value of an is shown in figure [ figure : a : algorithm ] .
[ rem : zero ] for the sake of simplicity of exposition we assume that @xmath160 .
the case when @xmath161 can be handled with an additional check of the context in the algorithm .
however , since all our applications have @xmath160 we choose to make this simplifying assumption .
[ rem : multi ] note that the algorithm of figure [ figure : a : algorithm ] can be modified to handle inputs where @xmath21 is a multiset ( namely , two points can occupy the same location ) .
specifically , it must be ensured that the distance computed in line [ line : rad ] is not zero , as this is required for the call to .
this can be remedied ( in linear time ) by first grouping all the duplicates of the point sampled in line [ line : gen : random ] into a single weighted point ( with the sum of the weights ) before calling line [ line : rad ] .
this can be done by computing a net for a radius that is smaller than ( say ) half the smallest non - zero distance . how to compute this distance
is described in section [ sec : closest : non : zero : dist ] .
a of the algorithm is an iteration where gets called .
a is one where gets called .
note that the only other type of iteration is the one where the algorithm returns .
[ lemma : remove ] let @xmath8 be a point set .
a @xmath162-net of @xmath8 , for any @xmath142 , can contain at most half the points of @xmath163 .
consider any point @xmath26 in @xmath163 which became one of the net points .
since @xmath164 , a disk of radius @xmath165 centered at @xmath26 must contain another point @xmath98 from @xmath163 ( indeed , @xmath164 only if its distance from its nearest neighbor in @xmath8 is at most @xmath165 ) .
moreover , @xmath98 can not become a net point since it is too close to @xmath26 . now
if we place a ball of radius @xmath165 centered at each point of @xmath163 which became a net point , then these balls will be disjoint because the pairwise distance between net points is @xmath166 .
therefore each point of @xmath163 which became a net point can charge to at least one point of @xmath163 which did not make it into the net , such that no point gets charged twice .
[ lemma : time ] given an instance @xmath112 of an , the algorithm @xmath167 { { { \textcolor[named]{redviolet}{\texttt{\bf{{ndpalg } } } } } } \index{algorithm!{ndpalg}@{{{\textcolor[named]{redviolet}{\texttt{\bf{{ndpalg } } } } } } } } } } \xspace}({\mathsf{w } } , { \gamma})$ ] runs in expected @xmath12 time .
in each iteration of the while loop the only non - trivial work done is in computing @xmath168 , the two calls to , and the one call to either or .
it has already been shown that all of these can be computed in @xmath169 time .
hence the total running time for the algorithm is @xmath170 , where @xmath0 denotes the last ( incomplete ) iteration of the while loop .
so consider the beginning of iteration @xmath171 of the while loop .
let the points in @xmath172 be labeled @xmath173 in increasing order of their nearest neighbor distance in @xmath172 .
let @xmath174 be the index of the point chosen in line [ line : gen : random ] and let @xmath175 and @xmath176 be the subset of the points with index @xmath177 and index @xmath178 , respectively .
now since a point is randomly picked in line [ line : gen : random ] , with probability @xmath179 , @xmath180 $ ] .
lets call this event a .
we have @xmath181 for a successful iteration . since @xmath171 is not the last iteration of the while loop , either or
must get called . if @xmath94 { { { \textcolor[named]{redviolet}{\texttt{\bf{delfar } } } } } \index{algorithm!delfar@{{{\textcolor[named]{redviolet}{\texttt{\bf{delfar } } } } } } } } } \xspace}({\ell}_i , { \mathsf{w}}_{i-1})$ ] gets called ( i.e. line [ line : delete ] ) then by lemma [ lemma : delete ] , all of @xmath175 gets removed .
so suppose gets called ( i.e. line [ line : net ] ) . in this case
lemma [ lemma : remove ] implies that the call to removes at least @xmath182 points . therefore , for any iteration @xmath171
, at least @xmath183 points get removed .
if an iteration is successful then @xmath184 .
in particular , @xmath185{0cm}{0.4cm}}\right . } } \right ] } } \geq { \left| { { \mathsf{w}}_{i-1 } } \right|}/16 $ ] .
now , @xmath186 and as such @xmath187{0cm}{0.4cm}}\right . } } \right]}}\leq ( 15/16 ) { \left| { { \mathsf{w}}_{i-1 } } \right|}$ ] . therefore , for @xmath188 , @xmath189{0cm}{0.4cm}}\ !
{ \left| { { \mathsf{w}}_i } \right| } } \right ] } } = { \mathop{\mathbf{e}}\!{\left [ { { \rule[-.2cm]{0cm}{0.8cm}}\ ! { \mathop{\mathbf{e}}\!{\left [ { { \left| { { \mathsf{w}}_i } \right| } { \,\left|\ , { { \mathsf{w}}_{i-1 } } { \rule[-.2cm]{0cm}{0.4cm}}\right . } } \right ] } } } \right ] } } \leq { \mathop{\mathbf{e}}\!{\left [ { \frac{15}{16}{\left| { { \mathsf{w}}_{i-1 } } \right| } } \right ] } } = \frac{15}{16 } { \mathop{\mathbf{e}}\!{\left [ { { \rule[-.2cm]{0cm}{0.4cm}}\ ! { \left| { { \mathsf{w}}_{i-1 } } \right| } } \right]}}. \end{aligned}\ ] ] hence , by induction on @xmath190 , @xmath191{0.0cm}{0.39cm}}\ ! { \left| { { \mathsf{w}}_i } \right| } } \right ] } } \leq ( 15/16)^i { \left| { { \mathsf{w}}_0 } \right|}$ ] and
so , in expectation , the running time is bounded by @xmath192}}}\right ) } = o{\!\left({\sum_{i=0}^{i = k-1 } { \mathop{\mathbf{e}}\!{\left [ { { \rule[-.2cm]{0cm}{0.4cm}}{\left| { { \mathsf{w}}_i } \right| } } \right]}}}\right ) } = o{\!\left({\sum_{i=0}^{i = k-1 } ( 15/16)^i { \left| { { \mathsf{w}}_0 } \right|}}\right ) } = o{\!\left({{\left| { { \mathsf{w}}_0 } \right| } { \rule[-.2cm]{0cm}{0.4cm}}\ ! } \right)}.
\end{aligned}\ ] ] [ [ correctness . ] ] correctness .
+ + + + + + + + + + + + the formal proof of correctness is somewhat tedious , but here is the basic idea : at every iteration , either far points are being thrown away ( and this does not effect the optimal value ) , or we net the points .
however , the net radius being used is always significantly smaller than the optimal value , and throughout the algorithm execution the radii of the nets being used grow exponentially . as such , the accumulated error in the end is proportional to the radius of the last net computation before termination , which itself is also much smaller than the optimal value . before proving that returns a bounded spread interval containing @xmath193 ,
several helper lemmas will be needed . for notational ease ,
in the rest of this section , we use @xmath194 as shorthand for @xmath195 .
[ lemma : double ] suppose that is called in iteration @xmath190 of the while loop ( i.e. , line [ line : net ] in figure [ figure : a : algorithm ] @xmath196 ) .
then for any iteration @xmath197 we have , @xmath198 .
consider the beginning of iteration @xmath174 of the while loop .
the current set of points , @xmath199 , are a subset of the net points of a @xmath200-net ( it is a subset since line [ line : delete ] and line [ line : net ] might have been executed in between rounds @xmath190 and @xmath174 ) .
therefore , being a net , the distance between any two points of @xmath199 is @xmath201 , see definition [ def : net ] @xmath202 . in particular
, this means that for any point @xmath26 of @xmath199 , we have @xmath203 .
[ lemma : translation ] for @xmath204 , we have @xmath205 .
let @xmath206 be the set of indices of the net iterations up to ( and including ) the @xmath190thiteration .
similarly , let @xmath207 be the set of iterations where get called .
if was called in the @xmath174thiteration , then @xmath208 is a @xmath209-translation of @xmath199 and so by the property , @xmath210 .
on the other hand , if gets called in the @xmath174thiteration , then @xmath211 by the property .
let @xmath212 , we have that @xmath213{0cm}{0.4cm}}{\ensuremath{f}}({\mathsf{w}}_j)- { \ensuremath{f}}({\mathsf{w}}_{j-1 } ) } \right|}\\ & = \sum_{j \in i } { \left| { { \rule[-.2cm]{0cm}{0.4cm}}{\ensuremath{f}}({\mathsf{w}}_j)- { \ensuremath{f}}({\mathsf{w}}_{j-1 } ) } \right| } + \sum_{j \in \overline{i } } { \left| { { \rule[-.2cm]{0cm}{0.4cm}}{\ensuremath{f}}({\mathsf{w}}_j)- { \ensuremath{f}}({\mathsf{w}}_{j-1 } ) } \right| } \\ & \leq \sum_{j \in i } 6{\ell}_j + \sum_{j \in \overline{i } } 0 \leq 6 { \ell}_m \sum_{j=0}^\infty \frac{1}{3^j } \leq 9 { \ell}_m \leq 9 { \ell}_i , \end{aligned}\ ] ] by lemma [ lemma : double ] .
the following lemma testifies that the radii of the nets computed by the algorithm are always significantly smaller than the value we are trying to approximate .
[ lemma : induction ] for any iteration @xmath190 of the while loop such that gets called , we have @xmath214 , where @xmath215
. the proof will be by induction .
let @xmath216 be the indices of the iterations of the while loop in which gets called . for the base case , observe that in order for to get called , we must have @xmath217 .
however , since this is the first iteration in which is called it must be that @xmath218 ( since for all previous iterations must have been called ) .
so now suppose that @xmath219 for all @xmath220 .
if a call to is made in iteration @xmath221 then again @xmath222 .
thus , by lemma [ lemma : translation ] and induction , we have @xmath223 if @xmath224 . this in turn is equivalent to @xmath225 , which is true by definition .
setting @xmath226 , results in @xmath227 , and by lemma [ lemma : induction ] , for all @xmath190 that correspond to a net iteration , @xmath228 . by lemma [ lemma : translation ] , for any net iteration @xmath190 , we have @xmath229{0cm}{0.4cm}}{\ensuremath{f}}({\mathsf{w}}_i ) - { \ensuremath{f}}({\mathsf{w}}_0 ) } \right| } \leq 9{\ell}_i \leq { \ensuremath{f}}({\mathsf{w}}_0 ) / 3 .\end{aligned}\ ] ] in particular , we conclude that @xmath230 for any iteration @xmath190 .
we thus get the following .
[ cor : same:1 ] for @xmath231 , and any @xmath190 , we have : @xmath232 @xmath233 . if @xmath234 $ ] then @xmath235 $ ] @xmath236 $ ] . if @xmath237 then @xmath238 .
[ lemma : summary:1 ] for @xmath231 , given an instance @xmath239 of a @xmath116- , @xmath167 { { { \textcolor[named]{redviolet}{\texttt{\bf{{ndpalg } } } } } } \index{algorithm!{ndpalg}@{{{\textcolor[named]{redviolet}{\texttt{\bf{{ndpalg } } } } } } } } } } \xspace}({\mathsf{w } } , { \gamma})$ ] returns an interval @xmath240 $ ] containing @xmath241 , where @xmath242{0.0cm}{0.38cm}}[x',y']}\right ) } } \leq 4 \max { \!\left ( { { \varphi } , { { c_{\mathrm{net}}}}}\right)}$ ] .
consider the iteration of the while loop at which terminates , see figure [ figure : a : algorithm ] @xmath196 .
if line [ line : bounded1x ] or line [ line : bounded1y ] get executed at this iteration , then the interval @xmath243 $ ] was computed by the @xmath116-decider , and has spread @xmath244 . as such , by corollary [ cor : same:1 ] , the returned interval @xmath240 = [ x/2 , 2y]$ ] contains the optimal value , and its spread is @xmath245 .
a similar argumentation holds if line [ line : bounded3 ] gets executed .
indeed , the returned interval contains the desired value , and its spread is @xmath246 .
[ lemma : result:1 ] given an instance of a @xmath116-defined by a set of @xmath7 points in @xmath14 , one can get a @xmath116-approximation to its optimal value , in @xmath247 expected time , where @xmath248 is the running of the procedure , and @xmath249 .
if @xmath123 , for some @xmath250 , ( i.e. , is @xmath4-decider ) , then one can compute a @xmath124)-approximation to the optimal value , and the expected running time is @xmath251 .
let @xmath112 be the given @xmath116-instance , and let be the corresponding @xmath116-decider .
by lemma [ lemma : time ] and lemma [ lemma : summary:1 ] , in expected @xmath12 time , one can get a bounded spread interval @xmath252 $ ] , for some constant @xmath126 such that @xmath253 , such that @xmath254 $ ] . if @xmath255 then we are done .
otherwise , perform a binary search over this interval .
specifically , for @xmath256 , let @xmath257 and let @xmath258 .
now perform a binary search over the @xmath259 s using .
if any of the calls returns an interval @xmath260 $ ] that contains the optimal value , then @xmath261 and so @xmath262 can be returned as the desired @xmath116-approximation . otherwise , if @xmath263 or @xmath264 the binary search moves left or right , respectively . in the end
, the search ends up with an interval @xmath265 which must contain the optimal value , and again , its spread is @xmath116 , and it thus provide the required approximation . the running time is dominated by the calls to the decider procedure ( we are assuming here that @xmath266 ) . clearly , the number of calls to the decider performed by this algorithm is @xmath267 , if @xmath268 . otherwise , if @xmath123 , then @xmath269 since @xmath270 , for @xmath271 , as can be easily verified .
[ theo : result : eps ] given an instance of a @xmath116-defined by a set of @xmath7 points in @xmath14 , one can get @xmath116-approximate the optimal value , in expected @xmath272 time , assuming @xmath273
. for the case @xmath274 , given an @xmath4-decider , with running time @xmath125 , one can @xmath124)-approximate , in @xmath275 expected time , the given @xmath4- , where @xmath126 is some constant .
lemma [ lemma : result:1 ] readily implies the first half . as for the second half , lemma
[ lemma : result:1 ] implies that one can get a @xmath4-approximation in expected @xmath276 time .
however , by using the same procedure as in the proof of lemma [ lemma : result:1 ] but with exponentially decreasing values for @xmath16 in the binary search , a factor of @xmath277 can be avoided . this is a standard idea and was also used by aronov and har - peled @xcite .
let @xmath106{redviolet}{\texttt{\bf{{decider}}}}}}\index{algorithm!{decider}@{{{\textcolor[named]{redviolet}{\texttt{\bf{{decider}}}}}}}}}}\xspace}_{{\varepsilon}}$ ] be our @xmath4-decider .
if @xmath16 is set to any constant ( say 1 ) , then by lemma [ lemma : time ] and lemma [ lemma : summary:1 ] , in expected linear time , one can get a bounded spread interval @xmath252 $ ] , for some @xmath278 , such that @xmath279 $ ] .
set @xmath280 and @xmath281 .
the algorithm now proceeds in rounds , doing the following in the @xmath190thround : assume that in the beginning of the @xmath190thiteration , we know that @xmath282{0cm}{0.4cm}}\right ] } , \qquad \text { where } \qquad y_i = { \left({1+{\varepsilon}_{i-1}}\right ) } x_{i-1}. \end{aligned}\ ] ] if @xmath283 , then we found the desired approximation , and the algorithm stops .
set @xmath284 and @xmath285 .
@xmath286{redviolet}{\texttt{\bf{{decider}}}}}}\index{algorithm!{decider}@{{{\textcolor[named]{redviolet}{\texttt{\bf{{decider}}}}}}}}}}\xspace}_{{\varepsilon}_i}(m_i , { \mathsf{w } } , { \gamma})$ ] , see definition [ def : decider ] @xmath287 .
there are three possibilities : if @xmath288 `` @xmath289 '' , then set @xmath290 $ ] .
if @xmath288 `` @xmath291 '' , then set @xmath292 $ ] . if @xmath288 `` @xmath293 $ ] '' , then set @xmath294 .
@xmath295 . in each round
, the algorithm computes an interval of spread @xmath296 that contains @xmath297 , and @xmath298 since the main bottleneck in each iteration is calling @xmath106{redviolet}{\texttt{\bf{{decider}}}}}}\index{algorithm!{decider}@{{{\textcolor[named]{redviolet}{\texttt{\bf{{decider}}}}}}}}}}\xspace}_{{\varepsilon}_i}$ ] , which runs in @xmath299 time , the total running time is bounded by , @xmath300 since the sum behaves like a geometric series and the last term is @xmath125 .
[ sec : applications ] we now show that theorem [ theo : result : eps ] can be applied to a wide array of problems . in order to apply it , one need to show that the given problem meets the requirements of an .
[ sec : k : center ] since computing a @xmath0-center clustering is an problem ( lemma [ lemma : k : center : wbdp ] ) , plugging this into theorem [ theo : result : eps ] , immediately yields a constant factor approximation to @xmath0-center in linear time .
it is easy to convert such an approximation to a @xmath6-approximation using a grid , see har - peled @xcite .
thus we get the following .
[ theo : k : center:2 ] given a set @xmath8 of @xmath7 points in @xmath14 , and a parameter @xmath0 , @xmath301 , one can compute a @xmath6-approximation to the optimal @xmath0-center clustering of @xmath8 in ( expected ) linear time . a result similar to theorem [ theo :
k : center:2 ] was already known for the case @xmath302 @xcite , and the above removes this restriction .
in addition , the new algorithm and its analysis are both simpler than the previous algorithm .
[ sec : k : th : distance ] [ claim : change ] let @xmath78 and @xmath303 be subsets of @xmath304 of size @xmath7 , such that @xmath303 is obtained by taking each value in @xmath78 and increasing or decreasing it by less than @xmath49 .
let @xmath305 and @xmath306 be the @xmath0thsmallest values in @xmath78 and @xmath303 , respectively
. then @xmath307 .
suppose for contradiction that @xmath308 . if @xmath309 then @xmath303 has at least @xmath310 values strictly larger than @xmath311 which implies @xmath78 has at least @xmath310 values strictly larger than @xmath305 .
similarly if @xmath312 then @xmath303 has at least @xmath0 values strictly smaller than @xmath313 which implies @xmath78 has at least @xmath0 values strictly smaller than @xmath305 .
[ lemma : k : dist ] let @xmath8 be a weighted point set in @xmath14 , and let @xmath129 be an integer parameter .
let @xmath314 denote the multi - set of pairwise distances determined by @xmath8 .
is viewed as @xmath1 unit weight points when determining the values of @xmath314 . for simplicity
we assume that the @xmath0th distance in @xmath8 is not zero ( i.e. it is determined by two distinct points ) . ]
given an instance @xmath315 the problem asks you to output the @xmath0thsmallest distance in @xmath314 .
given such an instance , one can @xmath4-approximate the @xmath0thsmallest distance in ( expected ) time @xmath316 .
let @xmath317 be the function that returns the @xmath0th smallest distance .
we prove that this function is an @xmath4-(see definition [ def : n : d : p ] @xmath143 ) .
given @xmath9 and @xmath16 , build a grid where every cell has diameter @xmath318 , and store the points of @xmath8 in this grid .
now , for any non - empty grid cell @xmath319 , let @xmath320 be the total weight of points in @xmath321 , and register this weight with all the grid cells in distance at most @xmath9 from it ( i.e. , @xmath322 ) .
let @xmath323 denote the total weight registered with a cell @xmath319 ( which includes its own weight ) .
any point in @xmath319 determines @xmath324 distances to other points in @xmath322 .
therefore the total number of distances between points in @xmath319 and points in @xmath322 is @xmath325{0cm}{0.4cm}}{{\omega}_{{\mathsf{n}}}{\!\left({{\box}}\right ) } } - 1}\right)}$ ] .
summing this over all cells ( and dividing by @xmath6 to avoid double counting ) gives the quantity @xmath326{0cm}{0.4cm}}{{\omega}_{{\mathsf{n}}}{\!\left({{\box}}\right ) } } - 1}\right)}. \end{aligned}\ ] ] note , that the count @xmath327 includes all distances which are @xmath57 , and no distances @xmath328 .
so let the desired @xmath0thdistance be denoted by @xmath329 .
thus , if @xmath330 then @xmath331 . similarly , if @xmath332 then @xmath333 .
therefore , to build a @xmath4-decider for distance @xmath9 , we run the above counting procedure on @xmath334 and @xmath335 , and let @xmath336 and @xmath337 be the two counts computed , respectively .
there are three cases : if @xmath338 then @xmath339 , and return this result . if @xmath340 then @xmath341 , then @xmath342 $ ] , and this interval has spread @xmath343
. the decider returns that this interval contains the desired value .
the only remaining possibility is that @xmath344 , and the decider returns this .
the running time of this decision procedure is @xmath316 . since the distance between any pair of points changes by at most @xmath148 in a @xmath49-translation
, the lipschitz condition holds by claim [ claim : change ] . by assumption ,
the @xmath0thsmallest distance is determined by two distinct weighted points @xmath26 and @xmath98 .
clearly these points are not in @xmath155 since @xmath345 .
so consider any point @xmath346 .
since removing @xmath347 does not remove the distance @xmath348 from the set of remaining distances , all that is needed is to show how to update @xmath0 . clearly , @xmath347 contributes @xmath349 distances , all of value @xmath150 , to @xmath350 , and @xmath351 pairwise distances of value zero .
thus , after removing @xmath347 from @xmath21 , the new context is @xmath352 .
the algorithm of lemma [ lemma : k : dist ] also works ( with minor modifications ) if we are interested in the @xmath0thdistance between two sets of points ( i.e. , the bipartite version )
. given two sets @xmath8 and @xmath23 of points in @xmath14 , of total size @xmath7 , and parameters @xmath0 and @xmath142 , one can @xmath4-approximate , in expected @xmath19 time , the following : the @xmath0thsmallest bichromatic distance .
formally , this is the smallest @xmath0thnumber in the multiset @xmath353{0cm}{0.4cm}}\right . } } \right\}}$ ] . the closest bichromatic pair between @xmath8 and @xmath23 . [ sec : k : smallest : m : n : n ] for a set @xmath24 of @xmath7 points , and a point @xmath25 , its in @xmath8 is the @xmath1thclosest point to @xmath26 in @xmath27 .
in particular , let @xmath28 denote this distance . here
, consider the multiset of these distances defined for each point of @xmath8 ; that is , @xmath354 .
interestingly , for any @xmath1 , we can approximate the @xmath0thsmallest number in this set in linear time . in order to use
, one need to generalize the problem to the case of weighted point sets .
specifically , a point @xmath26 of weight @xmath355 is treated as @xmath355 distinct points at the same location . as such
, the set @xmath356 from the above is actually a multiset containing @xmath355 copies of the value @xmath28 .
[ theo : k : th : m : nn ] let @xmath21 be a set of @xmath7 weighted points in @xmath14 , and @xmath357 parameters .
then one can @xmath4-approximate , in expected @xmath34 time , the @xmath0thsmallest @xmath1-nearest neighbor distance in @xmath21 .
formally , the algorithm @xmath4-approximates the @xmath0thsmallest number in the multiset @xmath358 we establish the required properties , see definition [ def : n : d : p ] @xmath143 .
let @xmath359 denote the desired quantity .
we first need a @xmath4-decider for this problem . to this end , given @xmath360 and @xmath21 as input , create a grid with cells of diameter @xmath361 , and mark for each point of @xmath21 all the grid cells in distance at most @xmath9 from it .
each non - empty grid cell has a count of the total weight of the points in distance at most @xmath9 from it .
thus , each point of @xmath21 can compute the total weight of all points which are approximately in distance at most @xmath9 from it .
now , a point @xmath153 , can decide in constant time if ( approximately ) @xmath362 .
if the number of points which declare @xmath362 ( where a point @xmath26 is counted @xmath355 times ) is greater than @xmath0 then the distance @xmath9 is too large , and if it is smaller than @xmath0 than @xmath9 is too small .
being slightly more careful about the details ( as was done in the proof of lemma [ lemma : k : dist ] ) one can verify this leads to a @xmath4-decider for this problem .
clearly the lipschitz property holds in this case , as @xmath49-translation only changes inter - point distances by at most an additive term of @xmath148 .
we need to show that @xmath0 can be updated properly for a weighted point @xmath26 whose nearest neighbor is further away than @xmath363 . if @xmath364 then we are throwing away @xmath355 points , all with @xmath365 , and so @xmath0 should be update to @xmath366 when @xmath26 is removed . similarly ,
if @xmath367 then all the points @xmath26 corresponds to have @xmath368 larger than the threshold , and @xmath0 does not have to be updated . plugging this into theorem [ theo : result : eps ] implies the desired result .
theorem [ theo : k : th : m : nn ] can be easily extended to work in the bichromatic case .
that is , there are two point sets @xmath8 and @xmath23 , and we are interested in the @xmath1thnearest neighbor of a point @xmath25 in the set @xmath23 .
it is easy to verify that the same time bounds of theorem [ theo : k : th : m : nn ] hold in this case . in particular , setting @xmath369 and @xmath30 in the bichromatic case , the computed distance will be the minimum radius of the balls needed to be placed around the points of @xmath23 to cover all the points of @xmath8 ( or vice versa ) .
[ sec : f : n : n ] using theorem [ theo : k : th : m : nn ] with @xmath30 and @xmath370 , results in a @xmath6-approximation , in @xmath12 time , to the of @xmath8 .
we now show how this can be converted into an exact solution .
so we have a quantity @xmath9 that is larger than the @xmath0thnearest neighbor distance , but at most twice larger .
we build a grid with a cell diameter being @xmath371 .
clearly , any point whose nearest neighbor distance is at least the @xmath0thnearest neighbor distance must be the only point in its grid cell .
for each such point @xmath26 , we compute its distance to all the points stored in its neighborhood @xmath56 , by scanning the list of points associated with these cells .
this computes for these `` lonely '' points their exact nearest neighbor distance .
the @xmath0thsmallest nearest neighbor distance will then be the @xmath372thlargest distance computed .
clearly , every cell s list of points get scanned a constant number of times , so overall the running time is linear .
we summarize the result . for a point
set @xmath8 , the , is the maximum distance of a point of @xmath8 , from the remaining points .
formally , it is @xmath373 .
let @xmath8 be a set of @xmath7 points in @xmath14 .
for an integer @xmath0 , @xmath374 , one can compute exactly , in expected linear time , the @xmath0thnearest neighbor distance in @xmath8 . in particular , setting @xmath31 , one can compute the furthest nearest neighbor distance exactly , in expected linear time .
[ sec : longest : mst ] the net computation provides a natural way to partition the data into clusters , and leads to a fast clustering algorithm .
one alternative partition scheme is based on distance connectivity .
[ def : connectivity ] for a set of points @xmath8 and a number @xmath77 , let @xmath375 be the of @xmath8 . specifically , it is a partition of @xmath8 into connected components of the @xmath2of @xmath8 after all edges strictly longer than @xmath9 are removed from it ( alternatively , these are the connected components of the intersection graph where we replace every point of @xmath8 by a disk of radius @xmath376 centered at that point ) , see figure on the right .
consider two partitions @xmath377 of @xmath8 .
the partition @xmath378 is a refinement of @xmath379 , denoted by @xmath380 , if for any set @xmath381 , there exists a set @xmath382 such that @xmath383 .
[ lemma : longest : edge : partition ] given a set @xmath63 , and parameters @xmath9 and @xmath16 , one can compute , in @xmath316 time , a partition @xmath378 , such that @xmath384 . build a grid with every cell having diameter @xmath361 . for every point in @xmath8 mark all the cells in distance @xmath376 from it .
that takes @xmath19 time .
next , for every marked cell create a node @xmath305 , and let @xmath385 be the resulting set of nodes .
next , create a bipartite graph @xmath386 connecting every node of @xmath385 , to all the points of @xmath8 marking its corresponding grid cell .
now , compute the connected components in @xmath387 , and for each connected component extract the points of @xmath8 that belong to it . clearly , the resulting partition @xmath378 of @xmath8 , is such that any two points in distance @xmath57 are in the same set ( since for such points there must be a grid cell in distance @xmath388 from both ) .
similarly , if the distance between two subsets @xmath389 is at least @xmath390 , then they are in different connected components of @xmath378 . in particular , in order for @xmath356 and @xmath391 to be in the same component
, there must exists points @xmath108 and @xmath392 which marked the same grid cell .
however , such a grid cell would need to be in distance at most @xmath376 from both @xmath393 and @xmath394 , implying @xmath395 , a contradiction .
thus , this is the desired partition .
clearly , this takes @xmath19 time overall .
[ theo : m : s : t : k : edge ] given a set of @xmath7 points @xmath8 in @xmath14 , and parameters @xmath0 and @xmath16 , one can output , in expected @xmath316 time , a @xmath4-approximation to the @xmath0thlongest ( or @xmath0th shortest ) edge in the @xmath2of @xmath8 .
consider the cost function @xmath396 which returns the @xmath0thlongest edge of the @xmath2of @xmath8 .
it is easy to verify that @xmath396 is the minimum @xmath9 such that @xmath375 has @xmath0 connected components . for the sake of simplicity of exposition ,
we assume all the pairwise distances of points in @xmath8 are distinct .
we need to fill - in / verify the properties of definition [ def : n : d : p ] @xmath143 : the @xmath124)-decider for @xmath397 uses the algorithm of lemma [ lemma : longest : edge : partition ] .
specifically , for a specified @xmath77 , compute a partition @xmath378 , such that @xmath398 .
now , if @xmath378 has @xmath138 connected components , then @xmath399 . similarly , if @xmath378 has @xmath400 connected components , then @xmath401 . by calling on two values , and using slightly smaller @xmath16 ,
it is straightforward to get the desired @xmath4-decider , as was done in lemma [ lemma : k : dist ] @xmath402 .
let @xmath403 be the complete graph defined on @xmath404 , where @xmath404 is some @xmath49-translation @xmath8 .
for the sake of simplicity of exposition , we consider @xmath404 and @xmath8 to be multi - sets of the same cardinality ( i.e. , @xmath405 ) .
let @xmath406 ( resp .
@xmath407 ) be the edges in the @xmath2of @xmath387 ( resp .
@xmath403 ) sorted by increasing order by their weight .
( note , that some of these edges might have length @xmath408 . ) observe that the graph having @xmath409 as edges , under the weight function induced by @xmath410 , it is a forest with @xmath411 connected components . as such , we have that @xmath412 , since , by the correctness of the kruskal algorithm , the first @xmath190 edges of the @xmath2form such a forest with the maximum weight edge being minimum ( among all such forests ) . as such , we have that @xmath413 a symmetric argument implies that @xmath414 .
setting @xmath415 , we have that @xmath416 , as desired .
consider a point @xmath153 , such that @xmath417 , and @xmath418 .
this implies that all the @xmath2edges adjacent to @xmath26 are of distance @xmath150 .
assume there are @xmath174 such edges , and observe that removing these edges creates @xmath419 connected components of the @xmath2 , such that each connected component is of distance @xmath150 from each other ( indeed , otherwise , one of the removed edges should not be in the @xmath2 , as can be easily verified ) .
thus , if we delete @xmath26 , and recompute the @xmath2 , we will get @xmath174 new edges that replace the deleted edges , all of these edges are of length @xmath150 . that is , deleting @xmath26 , and decreasing @xmath0 by one
, ensures that the target value remains the same in the pruned instance .
this establish that @xmath396 is an @xmath4- , and by plugging this into theorem [ theo : result : eps ] , we get the desired result . [
sec : sketchable ] let @xmath8 be a finite ground set of elements , and let @xmath420 be a family of subsets of @xmath8 .
then @xmath421 is an if for any @xmath422 and any @xmath423 , such that @xmath383 , we have that @xmath424 .
such a set system is a , if for any set @xmath425 there exists a constant size @xmath426 such that : for any @xmath427 that are disjoint , @xmath428 can be computed from @xmath426 and @xmath429 in @xmath75 time .
there is a membership oracle for the set system based on the sketch .
that is , there is a procedure @xmath430 such that given the sketch of a subset @xmath426 , @xmath431 returns whether @xmath432 or not , in @xmath75 time .
an example for such a sketchable family , is the set system @xmath433 , where @xmath425 is in @xmath420 if @xmath434 . here
the sketch of the set is simply the number of elements in the set , and combining two sketches @xmath426 and @xmath429 is adding the numbers to get @xmath435 ( for @xmath436 )
. we will be interested in two natural problems induced by such a family : smallest cluster find the smallest set in the family with certain properties , and min - max clustering
find disjoint sets in the family such that they cover the original set , and the maximum price of these sets is minimized .
note it is not necessary for the descriptions to have @xmath75 size or the oracle to run in @xmath75 time , however , assuming otherwise will affect the running time of if it is used to solve a problem involving a sketchable family .
consider associating a positive @xmath0-dimensional vector @xmath437 with each point @xmath25 ( a vector is if it is non - zero , and all its coordinates are non - negative ) .
a is an inequality of the form @xmath438 , where the coefficients @xmath439 are all non - negative .
for such a linear inequality , consider the set system @xmath440 , where a set @xmath23 is in @xmath420 if the linear inequality holds for the vector @xmath441 . clearly , this family is a sketchable family , the sketch being the sum of the vectors associated with the points of the set ( here @xmath0 is assumed to be a constant ) .
it is easy to verify that sketchable families are closed under finite intersection .
specifically , given a collection of @xmath1 such positive inequalities , the family of sets such that their sketch vector complies with all these inequalities is a sketchable family ( of course , checking if a set , given its sketch , is in the family naively would take @xmath442 time ) .
as a concrete application , consider the scenario where every element in @xmath8 has @xmath443 attributes .
one might be interested in subsets , where each set has at least a total sum of @xmath444 unit for the first two attributes , and a total sum of @xmath6 units for the last two attributes . of course
, in general , a membership oracle for the attributes space that has the property that if @xmath445 is valid then @xmath446 is also valid , for any positive @xmath447 , would define a sketchable family . as a concrete example , consider the non - linear ( and not convex ) condition that the sum of at least two attributes is larger than @xmath444 .
clearly , this defines a sketchable family .
clearly the above definition of sketchable family is very general and widely applicable .
next , we show how can be used to approximate certain objectives over sketchable families .
[ sec : min : cluster ] we now consider the problem of minimizing the cluster size of a subset of a given point set subject to inclusion in a sketchable family .
specifically , we consider the case when the cluster is defined by a ball of radius @xmath9 or when the cluster is defined by a connected component of @xmath375 for a radius @xmath9 .
note that as a ball or component grows both the cluster size and inclusion of the set of points ( in the cluster ) in the sketchable @xmath420 are monotone properties .
this correspondence is what allows us to apply our general framework . for a set of @xmath7 points @xmath63 , and a sketchable family @xmath440 , one can @xmath4-approximate , in expected @xmath316 time , the radius of the smallest ball , @xmath35 , such that @xmath448 .
let @xmath449 be the radius of the smallest ball @xmath35 in @xmath14 such that @xmath450 .
we claim that @xmath451 is an , see definition [ def : n : d : p ] @xmath143 ) .
the and properties readily hold for @xmath451 . as for property ,
given @xmath9 and @xmath142 , construct a grid with cells of diameter @xmath452 , and register each point of @xmath8 in all the grid cells in distance at most @xmath9 from it .
if there is a ball , @xmath35 , of radius @xmath9 such that @xmath453 then the set of points registered with the grid cell containing the center of this ball will be a superset of @xmath36 and hence the set is in @xmath420 , by the upward closed property of sketchable families .
moreover , the set registered at this grid cell requires a ball of radius at most @xmath454 ( centered at any point in this grid cell ) to cover it .
thus , the decision procedure simply checks the set of points associated with each grid cell to see whether or not it is in @xmath420 .
the definition of sketchable families implies this can be done in linear time in the total sizes of these sets ( i.e. , @xmath19 ) .
furthermore , if there is no ball of radius @xmath390 whose point set is in @xmath420 then the decision procedure would fail to find such a cell . thus , this is a @xmath4-decision procedure , and its running time is @xmath19 . plugging this into the algorithm of theorem [ theo : result : eps ] implies the result .
the following is a sample of what the above theorem implies .
we can @xmath4-approximate , in @xmath34 time , the following problems for a set of @xmath7 points in @xmath14 : the smallest ball containing @xmath0 points of @xmath8 .
the points of @xmath8 are weighted and we are given a threshold @xmath39 .
compute the smallest ball containing points of weight at least @xmath39 . if the points of @xmath8 are colored by @xmath0 colors , the smallest ball containing points of @xmath8 , such that they are colored by at least @xmath37 different colors .
( thus , one can find the smallest non - monochromatic ball [ @xmath455 , and the smallest ball having all colors [ @xmath456 . )
the running time is @xmath457 , as the sketch here is a @xmath0-dimensional vector .
[ sec : min : cluster : c ] note that @xmath375 is a monotone partition as @xmath9 increases , and it is natural to ask what is the minimum @xmath9 , for which there is a connected component in @xmath375 that is in a sketchable family .
[ theo : min : cluster : mst ] for a set of @xmath7 points @xmath63 , and a sketchable family @xmath458 , one can @xmath4-approximate , in expected @xmath316 time , the minimum @xmath9 , such that there is a connected component in @xmath375 that is in @xmath420 .
the target function @xmath459 is the smallest @xmath9 such that @xmath375 contains a set that is in @xmath420 .
let @xmath460 .
the algorithm first check if any of the input points by themselves have the desired property , if so then @xmath461 and the algorithm returns @xmath408 .
( for the case that @xmath8 is a multiset , see remark [ rem : multi ] @xmath462 . ) as such
, we can assume that @xmath463 , and furthermore , by corollary [ cor : same:1 ] ( c ) , @xmath464 is always non - zero during the algorithm execution .
again , we need to verify the properties of definition [ def : n : d : p ] @xmath143 for @xmath464 : one can @xmath4-approximate the connected components of @xmath375 , using the algorithm of lemma [ lemma : longest : edge : partition ] , and for each approximate connected component , use their sketch to decide if any of them is in @xmath420 .
if so , return that @xmath9 is larger than the optimal value , otherwise return that it is too small .
clearly , this is @xmath4-decider that works in @xmath19 time .
let @xmath465 .
by the definition of @xmath451 , there exists @xmath0 , such that @xmath466 .
now , arguing as in the proof of theorem [ theo : m : s : t : k : edge ] @xmath467 , implies that @xmath451 has the desired property .
consider a point @xmath153 , such that @xmath417 , and @xmath468 .
now , @xmath26 by itself can not have the desired property , as this would imply that @xmath469 .
this implies that in @xmath470 the point @xmath26 is by itself ( and not in any cluster that is in @xmath420 ) . as such ,
throwing @xmath26 away does not change the target function .
formally , @xmath471 . plugging this into theorem [ theo : result : eps ] implies the result .
one natural application for theorem [ theo : min : cluster : mst ] , is for ad hoc wireless networks . here
, we have a set @xmath8 of @xmath7 nodes and their locations ( say in the plane ) , and each node can broadcast in a certain radius @xmath9 ( the larger the @xmath9 the higher the energy required , so naturally we would like to minimize it ) .
it is natural now to ask for the minimum @xmath9 such that one of the connected components in the resulting ad hoc network has some desired property .
for example , in @xmath19 time , we can @xmath4-approximate the smallest @xmath9 such that : one of the connected components of @xmath375 contains half the points of @xmath8 , or more generally if the points are weighted , that one of connected component contains points of total weight at least @xmath39 , for a prespecified @xmath39 .
if the points are colored , the desired connected component contains all the colors ( for example , each color represent some fraction of the data , and the cluster can recover the data if all the pieces are available ) , or at least two colors , or more generally a different requirement on each color . [
sec : cluster : monotone ] we are given a sketchable family @xmath472 , and a cost function @xmath473 .
we are interested in finding disjoint sets @xmath474 , such that @xmath475 , and @xmath476 is minimized .
we will refer to the partition realizing the minimum as the of @xmath8 .
let @xmath8 be a set of points in @xmath14 , and let @xmath472 be a sketchable family .
for a set @xmath477 , let @xmath478 be the radius of the smallest ball centered at a point of and enclosing @xmath21 .
one can @xmath38-approximate , in expected @xmath19 time , the min - max clustering under @xmath479 of @xmath8 .
that is , one can cover @xmath8 by a set of balls , and assign each point of @xmath8 to one of these balls , such that the set of points assigned to each ball is in @xmath420 , and the maximum radius of any of these balls is a @xmath38-approximation to the minimum radius used by any such cover .
[ theo : cluster : min : max ] let @xmath480 be the optimal partition with radius @xmath481 , and consider an @xmath9-net @xmath10 for @xmath482 , computed using corollary [ cor : valid ] .
consider a point @xmath483 , and let @xmath484 $ ] be the set of points of @xmath8 assigned to @xmath26 by the nearest net - point assignment .
next , consider the cluster @xmath485 that contains it .
clearly , @xmath486 , and the distance of @xmath26 from all other net points in @xmath10 is at least @xmath487 .
it follows that @xmath488 $ ] , and since @xmath477 , it follows that @xmath484 \in { \mathcal{f}}$ ] . a @xmath489-decider for this problem works by computing the @xmath490-net @xmath10 , and for each @xmath483 , checking the sketchable property for the set @xmath484 $ ] .
it is easy to verify that the properties of definition [ def : n : d : p ] @xmath143 hold in this case . in particular , throwing a far away isolated point corresponds to a cluster that already fulfill the monotone property , and it is too far away to be relevant .
namely , computing @xmath481 is an and so plugging this into theorem [ theo : result : eps ] implies the result .
[ sec : cluster : monotone : examples ] if the required sketchable property is that every cluster contains at least @xmath0 points , then theorem [ theo : cluster : min : max ] approximates the problem .
that is , one has to cover the points by balls , such that every cluster ( i.e. , points assigned to a ball ) contains at least @xmath0 points .
the price of this clustering is the radius of the largest ball used .
a @xmath6-approximation to this problem is known via the usage of flow @xcite but the running time is super quadratic .
recently , the authors showed a similar result to theorem [ theo : cluster : min : max ] with running time ( roughly ) @xmath32 @xcite .
this paper also shows that this problem can not be approximated to better than ( roughly ) @xmath491 even for points in the plane .
let @xmath8 be a set of points in @xmath14 , and let @xmath0 and @xmath142 be parameters .
one can @xmath38-approximate the lower bounded center clustering in @xmath19 time .
one can plug - in any sketchable family into theorem [ theo : cluster : min : max ] .
for example , if the points have @xmath0 colors , we can ask for the min - max radius clustering , such that every cluster contains ( i ) all colors , ( ii ) at least two different colors , or ( iii ) a different requirement on each color , etc . as another concrete example , consider that we have @xmath7 customers in the plane , and each customer is interested in @xmath0 different services ( i.e. , there is a @xmath0-dimensional vector associated with each customer specifying his / her demand ) .
there are @xmath37 types of service centers that can be established , but each such center type requires a minimum level of demand in each of these @xmath0 categories ( i.e. , each type is specified by a minimum demand @xmath0-dimensional vector , and a set of customers can be the user base for such a service center if the sum of their demands vector is larger than this specification ) .
the problem is to partition the points into clusters ( of minimum maximum radius ) , such that for every cluster there is a valid service center assigned to it .
clearly , this falls into the framework of theorem [ theo : cluster : min : max ] , and can be @xmath38-approximated in @xmath492 time .
[ sec : cluster : spanning ] one can get a similar result to theorem [ theo : cluster : min : max ] for the connectivity version of clustering of @xmath8 .
formally , a set of points @xmath493 is if @xmath21 is contained in some set of @xmath375 . given a sketchable family @xmath494 , a partition @xmath378 of @xmath8 is an if all the sets in @xmath378 are in @xmath420 , and are @xmath9-valid .
[ theo : cluster : connected ] let @xmath8 be a set of points in @xmath14 , and let @xmath472 be a sketchable family .
one can @xmath4-approximate @xmath481 , in expected @xmath19 time , where @xmath481 is the minimum value such that there is a @xmath481-connected clustering of @xmath8 .
let @xmath464 be the target function in this case .
we verify the properties of definition [ def : n : d : p ] @xmath143 : given @xmath9 , we use lemma [ lemma : longest : edge : partition ] to compute a partition @xmath378 of @xmath8 such that @xmath495 .
if the sketchable property holds for each cluster of @xmath378 , then return that @xmath9 is too large , otherwise return that it is too small . as for
the quality of this decider , observe that if the optimal partition has a cluster @xmath21 that uses points from two different clusters of @xmath378 , than @xmath21 is not @xmath9-valid , as otherwise these two points would be in the same cluster of @xmath378 ( namely , @xmath496 ) .
follows easily by arguing as in theorem [ theo : m : s : t : k : edge ] @xmath467 .
if an isolated point exists , then it can be thrown away because the cluster of original points it corresponds to , is a valid cluster that can be used in the final clustering , and it does not interact with any other clusters . plugging this into theorem [ theo : result : eps ] now implies the result .
a nice application of theorem [ theo : cluster : connected ] is for ad hoc networks .
again , we have a set @xmath8 of @xmath7 wireless clients , and some of them are base stations ; that is , they are connected to the outside world .
we would like to find the minimum @xmath9 , such that each connected component of @xmath375 contains a base station .
[ sec : closest : non : zero : dist ] given a point set @xmath8 , it was already shown in section [ sec : f : n : n ] that , with a small amount of post processing , can be used to compute the closest pair distance exactly in expected linear time . if one allows @xmath8 to contain duplicate points ( i.e. @xmath8 is a multiset ) then a couple modifications to @xmath167 { { { \textcolor[named]{redviolet}{\texttt{\bf{{ndpalg } } } } } } \index{algorithm!{ndpalg}@{{{\textcolor[named]{redviolet}{\texttt{\bf{{ndpalg } } } } } } } } } } \xspace}$ ] must be made .
first modify the algorithm so that for the selected point @xmath26 it finds the closest point to @xmath26 that has non - zero distance from it .
secondly , we modify ( see lemma [ lemma : delete ] @xmath156 ) so that if a net point corresponds to several copies of the same point then it is being treated as a single point . with these two modifications , the above algorithm works verbatim , and we get the following .
[ lemma : non : zero ] let @xmath8 be a multiset of weighted points in @xmath14 .
then one can solve the problem _ exactly _ for @xmath8 , in expected linear time .
in particular , if @xmath8 contains no duplicates then this corresponds to computing the closest pair distance .
interestingly , the algorithm of lemma [ lemma : non : zero ] is a prune - and - search algorithm , as the net stage never gets executed .
observe , that it is not hard to extend the algorithm of golin _
et al . _
@xcite to solve this variant , and the result of lemma [ lemma : non : zero ] is included in the paper only for the sake of completeness . in particular , the resulting ( simplified ) algorithm is similar to the algorithm of khuller and matias @xcite .
[ sec : high : prob ] in section [ sec : framework ] , we presented an algorithm ( , see figure [ figure : a : algorithm ] @xmath196 ) that provides a constant factor approximation to a wide variety of problems in expected linear time . here
we present an extension of that algorithm that runs in linear time with high probability .
our algorithm works by computing a quantity that is not directly related to the problem at hand ( but in linear time ! ) , and then relating this quantity to the desired quantity , enabling us to compute it in linear time .
specifically , the resulting algorithm is somewhat counterintuitive and bizarre . as such , to make the algorithm description more accessible to the reader , we introduce the basic ideas in section [ sec : h : p : introduction ] and sketch the algorithm in section [ sec : motivation ] .
we describe the algorithm in detail in section [ sec : high : prob : algorithm ] , and analyze it in section [ sec : high : prob : analysis ] .
[ sec : h : p : introduction ] [ [ notation . ] ] notation .
+ + + + + + + + + see table [ table : notation ] for a summary of the notation used in this part of the paper . as it is sufficient for our purposes , for simplicity in this section @xmath8
will always be an unweighted points set , i.e. @xmath497 . the algorithm can be easily extended to handle the weighted case .
@xmath84 & @xmath498 & distance from @xmath26 to its nearest neighbor in @xmath86 .
+ @xmath499 & & distance of @xmath26 to its @xmath190thnearest neighbor in @xmath500 . + @xmath501 & @xmath502{0cm}{0.4cm}}\right . } } \right\}}$ ] & multi - set of distances between points in @xmath503 and nearest neighbors in @xmath8 . + @xmath504 & @xmath505{0cm}{0.4cm}}\right . } } \right\}}$ ] & multi - set of distances between point in @xmath503 and their @xmath190thnearest neighbors in @xmath8 .
+ @xmath506 & & value of rank @xmath507 in @xmath501 , for some @xmath508 $ ] .
+ @xmath509 & & value of rank @xmath507 in @xmath504 , for some @xmath508 $ ] .
+ [ table : notation ] [ [ sample - in - the - middle . ] ] sample in the middle .
+ + + + + + + + + + + + + + + + + + + + + observe that if one could sample a nearest neighbor distance that lies in the interval @xmath510 $ ] for some fixed constant @xmath511 , then we are done . indeed , for each iteration of for which the sampled distance was near the middle , it is guaranteed that the algorithm either terminates or removes a constant fraction of the points . as such it is sufficient to present an algorithm which returns a value from this range with probability @xmath512 , for some sufficiently large constant @xmath73 , where @xmath513 .
for the rest of this section we therefore focus on the problem of computing such an approximate middle nearest neighbor distance , in linear time , with high probability . to turn into an algorithm which runs in linear time with high probability , we replace line [ line : gen : random ] and line [ line : rad ] in figure [ figure : a : algorithm ] @xmath196 with this new subroutine .
[ [ sampling - for - the - median . ] ] sampling for the median .
+ + + + + + + + + + + + + + + + + + + + + + + + the problem with the random sampling from @xmath514 done in , is that though , with good probability , the sampled distance lies near the middle , it does not do so with high probability .
this is the same problem one is faced when performing the standard basic randomized median selection . for median selection one known solution
is rather than randomly sampling a pivot , to instead take a sufficiently large random sample , and take the median of the random sample as the pivot .
this pivot is good with high probability @xcite , and it is this result we wish to mimic . [ [ the - challenge . ] ] the challenge .
+ + + + + + + + + + + + + + given the above discussion , the natural solution to finding a good middle value from @xmath514 , would be to sample multiple values from @xmath514 , and then take the median . by the chernoff inequality ,
if one samples at least a logarithmic number of values from a set of @xmath7 values ( @xmath514 in our case ) then , with high probability , the median of the sample will lie near the median of the entire set .
( see lemma [ lemma : high : prob ] @xmath515 for a formal statement and proof of this observation . )
the problem with this approach is that sampling @xmath516 values from @xmath514 takes @xmath517 time ( at least naively ) .
specifically , while the sampling is easy , computing the associated values requires @xmath518 time per sampled point ( note this is not a problem for median selection as there we are sampling values directly rather than points from which values are determined ) . therefore , we instead not only take a sample from which to select our median value , but also a sample from which we will determine what these values are .
namely , for each point in a @xmath519 sized random sample , compute it s nearest neighbor in a second @xmath520 sized sample , and then take the median of the computed values . though this approach takes @xmath521 time
, we need to somehow use it to compute the desired quantity , which we describe how to do next .
[ sec : motivation ] we describe the algorithm here informally the exact details are provided in section [ sec : high : prob : algorithm ] .
we take a logarithmic sized sample , @xmath503 , from @xmath8 , and we are interested in computing the median value in @xmath501 , see table [ table : notation ] . as computing @xmath501 itself takes @xmath522 time ( at least naively ) , instead we compute @xmath523 , where @xmath524 is a second @xmath525 sized sample .
consider a point @xmath526 .
first observe that since @xmath527 , @xmath528 ( for most of the points in @xmath8 this inequality is strict and would be @xmath529 ) .
however , as @xmath530 , with high probability , at least one of @xmath26 s @xmath531 closest points in @xmath8 will be in @xmath524 ( formally shown in lemma [ lemma : est : log : dist ] ) . therefore , with high probability , for each point @xmath526 we get a bounded value @xmath532 $ ] , where @xmath533 .
for the purpose of building intuition , for now lets assume that @xmath534 .
in other words , assume that one could take a logarithmic sample from @xmath535 in linear time .
as usual taking the median of these values is a good approximation for the median value in @xmath535 , so assume that we have computed @xmath536 .
we next compare @xmath537 to @xmath538 by calling an appropriate decider .
there are two possible cases .
assume that @xmath539 , where @xmath540 .
since @xmath541 , this implies that @xmath542 $ ] ( and we know the value of @xmath537 ) .
for example , this happens if the points of @xmath63 are roughly uniformly distributed .
this gives us a @xmath543 spread interval containing @xmath538 .
that is , @xmath544 , and so it suffices to sample a median value from within this range .
now if the distances we are concerned with are all larger than @xmath545 then we can compute a net with slightly smaller radius , say @xmath546 , to `` remove '' all smaller distances .
then since we are only concerned with distances smaller than @xmath537 we can throw these points into a grid with cell diameter roughly @xmath537 , and in this grid the distances we care about will be between net points that are either in adjacent cells or within the same cell .
however , since the spacing between points in the net is at least @xmath547 , there can be at most @xmath548 points per cell ( and in this paper we assume @xmath549 is a constant ) .
hence computing a nearest neighbor distance for a single point to the net , if this distance is smaller than @xmath537 , takes only @xmath550 time per point .
thus , it is now possible to quickly estimate @xmath84 , for a single sample point @xmath26 , and therefor we can compute the desired median .
[ sec : justright ] assume that @xmath551 ; that is , the computed value @xmath537 is way bigger than the value we are trying to compute .
so , consider any point @xmath25 and look at the distance to its @xmath0thnearest neighbor , i.e. @xmath552 .
now imagine placing concentric adjacent rings around @xmath26 of increasing radius , each with width @xmath553 ( the rings stop at radius @xmath552 ) . by the pigeon hole principle
one of these rings must contain no points from @xmath8 .
in particular , if one instead used the value @xmath554 then at least half the points of @xmath8 would have an empty ring of width @xmath554 around them , and the inner ball of the ring would contain at most @xmath0 points ( actually at most @xmath555 ) . since these rings occur before the @xmath0thnearest neighbors , the different connected components in the connectivity clustering @xmath556 containing these points have size @xmath138 ( see definition [ def : connectivity ] @xmath557 ) .
this is good news , as at least half the points are in small clusters of @xmath60 , and from lemma [ lemma : longest : edge : partition ] these components can be computed efficiently . in particular ,
if a point is in a component of size @xmath558 then its nearest neighbor can be computed in @xmath559 time by scanning its connected component in @xmath60 .
since a large fraction of the points are in such components we can sample from these components , quickly compute nearest neighbors , and return the median .
the problem is that these components could be too small , namely just a single point , in which case one must look outside a point s component in @xmath60 to compute its nearest neighbor .
we are not stuck yet however , as thus far we have not taken into account the distribution of @xmath8 . our assumption that @xmath560 implies
that at least half the points have their nearest neighbor closer than the width of the rings we considered , i.e. half the points are in non - trivial components in @xmath60 . by carefully fine - tuning the constants
, one can guarantee that the set of points that are in small components that are also non - trivial components is a constant fraction of the points of @xmath8 .
as such , we can find the median nearest neighbor distance for these points by random sampling , which is exactly what we needed .
[ sec : toosmall ] our purpose is to approximate @xmath561 , for some constant @xmath39 bounded away from @xmath408 and @xmath444 . to this end
, we would like to sample @xmath516 values from @xmath538 and compute the median .
since this is too slow , instead we sample from @xmath562 and compute the median .
if the points are roughly uniformly distributed , @xmath563 $ ] and we can use nets and grids to efficiently sample distances in this bounded spread interval .
otherwise a constant fraction of the points must be in small but non - trivial clusters in a connectivity clustering of the points .
as such , computing nearest neighbor distances for these points can be done efficiently by scanning their respective components , thus leading to the desired middle value .
[ algorithm : midnn ] [ figure : high : prob ] [ sec : high : prob : algorithm ] we now present the algorithm in full detail . specifically , let denote the new algorithm for computing a value in @xmath564 $ ] , for some fixed constant @xmath565 . in section [ sec
: f : n : n ] it was shown how to convert a constant factor approximation to such a value into an exact value . hence , computing a constant factor approximation to a value in @xmath566 in linear time , with high probability , suffices for our purposes .
the full algorithm is shown in figure [ figure : high : prob ] . before analyzing the algorithm
, we first define several subroutines that it employs .
[ algorithm : decider : rank ] given a point set @xmath8 and a distance @xmath9 , the decision procedure @xmath567 { { { \textcolor[named]{redviolet}{\texttt{\bf{deciderm } } } } } \index{algorithm!deciderm@{{{\textcolor[named]{redviolet}{\texttt{\bf{deciderm } } } } } } } } } \xspace}({\mathsf{p } } , r , \alpha)$ ] decides if the @xmath37thnearest neighbor distance in @xmath514 is smaller than @xmath9 , larger than @xmath9 , or roughly equal to it , where @xmath568 .
this decider is implicitly described in theorem [ theo : k : th : m : nn ] , but a simpler decider suffices , which we describe here in detail for the sake of clarity . to this end , throw the points of @xmath8 into a grid with cell diameter @xmath9 .
now , for every point @xmath25 , let @xmath569 be the number of the points in @xmath56 , not including @xmath26 itself . clearly @xmath569 can be computed in constant time ( given we
have first recorded the number of points in each non - empty grid cell ) .
now , we compute the count @xmath570 in @xmath571 time . now , if @xmath572 the procedure returns that @xmath9 is smaller than the @xmath37thnearest - neighbor distance . similarly ,
if @xmath573 , then the @xmath37thnearest - neighbor distance must be at most @xmath574 . in this case , by also computing the count for the distance @xmath574 , one will be able to either return a constant spread interval that contained the desired distance , or that this distance is strictly larger than @xmath9 . thus yielding the required decider .
[ sec : estimate ] given a set @xmath8 of @xmath7 points in @xmath14 , take two random samples from @xmath8 , a sample @xmath503 of size @xmath519 and a sample @xmath524 of size @xmath525 , and compute @xmath575 .
this can be done in @xmath12 time , by scanning all of @xmath524 for each point in @xmath503 to compute the set of distances @xmath523 .
next , compute value of rank @xmath576 in this set .
we denote this algorithm by @xmath577 { { { \textcolor[named]{redviolet}{\texttt{\bf{{est{}log{}dist } } } } } } \index{algorithm!{est{}log{}dist}@{{{\textcolor[named]{redviolet}{\texttt{\bf{{est{}log{}dist } } } } } } } } } } \xspace}({\mathsf{p}})$ ] , and it is shown in figure [ figure : med ] .
[ algorithm : est : log : dist ] [ figure : med ] [ sec : logarithmicalg ] the subroutine used by ( i.e. line [ line : log : spread ] in figure [ figure : high : prob ] @xmath578 ) handles the case when the desired middle nearest neighbor distance lies in a poly - logarithmic spread interval .
the analysis of can be found in section [ sec : logarithmic ] . in the following ,
@xmath8 is a set of unweighted points such that @xmath579\subseteq [ r , r]$ ] , for some values @xmath580 , where @xmath581 .
( note , in the following , when sampling from a set @xmath524 we treat it as a set of @xmath582 points . )
[ algorithm : spread ] [ algorithm : log : n : n ] compute @xmath583 { { { \textcolor[named]{redviolet}{\texttt{\bf{net } } } } } \index{algorithm!net@{{{\textcolor[named]{redviolet}{\texttt{\bf{net } } } } } } } } } \xspace}{\!\left({r/8 , { \mathsf{p}}}\right)}}$ ] . throw the points of @xmath524 into a grid of sidelength @xmath584 . @xmath585 random sample from @xmath524 of @xmath586 points .
let @xmath587 be an approximation to the set @xmath588 .
specifically , for each point @xmath589 look within @xmath590 for its nearest neighbor , see definition [ def : grid : stuff ] @xmath591 . if such a neighbor is found record this distance ( which might be zero if @xmath26 is of weight @xmath592 ) and if no neighbor is found set this value to @xmath593 .
let @xmath25 be the point that corresponds to the number of rank @xmath594 in @xmath587 .
return @xmath84 .
[ sec : smallalg ] the subroutine , used by ( i.e. line [ line : small ] in figure [ figure : high : prob ] ) , handles the case when the desired middle nearest neighbor distance is considerably smaller than the value sampled by .
this corresponds , intuitively , to the situation where the input is mainly clustered in tiny clusters .
the analysis of is delegated to section [ sec : small ] .
[ algorithm : small : comp ] [ algorithm : small ] the algorithm works as follows : computes a partition @xmath60 , such that @xmath595 in linear time , using the algorithm of lemma [ lemma : longest : edge : partition ] @xmath596 , where @xmath597 , where @xmath598 and @xmath513 .
identify the components of @xmath60 which are neither singletons nor contain more than @xmath599 points , where @xmath600 is a sufficiently large constant .
collect all the points in these components into a set of points @xmath503 ( this set is not empty ) . take a random sample @xmath601 of @xmath602 points from @xmath503 , where @xmath603 is a sufficiently large constant .
for each point in @xmath601 , compute its nearest neighbor in @xmath8 by scanning the non - trivial connected component in @xmath60 that contains it .
let @xmath587 be this set of numbers .
return the median value in @xmath587 .
[ sec : high : prob : analysis ] the following lemma is by now standard , and we include its proof for the sake of completeness .
[ lemma : high : prob ] let @xmath503 be a set of @xmath7 real numbers , constants @xmath604 , @xmath605 , and a parameter @xmath606 .
let @xmath601 be a random sample of size @xmath37 picked ( with repetition and uniformly ) from @xmath503 .
then , the element of rank @xmath607 in @xmath601 has rank in the interval @xmath608{0cm}{0.4cm}}\right]}$ ] in the original set @xmath503 , with probability @xmath609 .
in particular , for @xmath610 , this holds with probability @xmath611 .
let @xmath612 be the number of values in @xmath601 of rank @xmath613 in @xmath503 .
similarly , let @xmath614 be the number of values in @xmath601 of rank @xmath615 in @xmath503 .
clearly @xmath616{0.0cm}{0.38 cm } } } \right ] } } = ( 1-\delta)\alpha t$ ] and @xmath617{0.0cm}{0.38cm}}y_r } \right ] } } = ( 1+\delta)\alpha t$ ] .
observe that if @xmath618 then the element of rank @xmath607 in @xmath601 has rank in @xmath503 in the interval @xmath619{0cm}{0.4cm}}\right]}$ ] , as desired . by the chernoff inequality @xcite
, we have that @xmath620{0cm}{0.4cm}}y_l > \alpha t } \right ] } } & = { \mathop{\mathbf{pr}}\!{\left [ { y_l > \frac { \alpha t}{\mu_l } \mu_l } \right ] } } = { \mathop{\mathbf{pr}}\!{\left [ { y_l > \frac { \alpha t}{(1 - \delta)\alpha t } \mu_l } \right ] } } \leq { \mathop{\mathbf{pr}}\!{\left [ { { \rule[-.2cm]{0cm}{0.4cm}}y_l > ( 1+\delta ) \mu_l } \right ] } } \\ & \leq \exp { \!\left ( { - \frac { \delta^2}{4 } \mu_l } \right ) } \leq \exp { \!\left ( { - \frac { \delta^2 } { 8 } \alpha t } \right ) } , \end{aligned}\ ] ] as @xmath621 . arguing as above
, the other bad possibility is that @xmath620{0cm}{0.4cm}}y_r < \alpha t } \right ] } } & \leq { \mathop{\mathbf{pr}}\!{\left [ { { \rule[-.2cm]{0cm}{0.4cm}}y_r < \frac { \alpha t } { \mu_r } \mu_r } \right ] } } = { \mathop{\mathbf{pr}}\!{\left [ { { \rule[-.2cm]{0cm}{0.4cm}}y_r < \frac { \alpha t } { ( 1+\delta)\alpha t } \mu_r } \right ] } } \\ & \leq { \mathop{\mathbf{pr}}\!{\left [ { { \rule[-.2cm]{0cm}{0.4cm}}y_r < { \!\left({1-\frac{\delta}{2}}\right ) } \mu_r } \right ] } } \leq \exp{\!\left ( { - \frac{\delta^2}{8}\mu_r}\right ) } \leq \exp{\!\left ( { - \frac{\delta^2}{8}\alpha t}\right)}. \end{aligned}\ ] ] [ sec : est : log : dist : analysis ] [ lemma : est : log : dist ] let @xmath8 be a set of @xmath7 points in @xmath14 . let @xmath165 be the distance returned by executing @xmath577 { { { \textcolor[named]{redviolet}{\texttt{\bf{{est{}log{}dist } } } } } } \index{algorithm!{est{}log{}dist}@{{{\textcolor[named]{redviolet}{\texttt{\bf{{est{}log{}dist } } } } } } } } } } \xspace}({\mathsf{p}})$ ] , see figure [ figure : med ] @xmath622 . with high probability , we have the following : for all points @xmath25 , the set @xmath623 has at most @xmath624 points , where @xmath625 , @xmath524 is the random sample used by , and @xmath626 is the ball of radius @xmath627 centered at @xmath26 . let @xmath378 be a clustering of @xmath8 such that @xmath628 , as computed by the algorithm of lemma [ lemma : longest : edge : partition ] @xmath596 , for @xmath629 .
then , at least @xmath630 of the points of @xmath8 are in clusters of @xmath378 that contain at most @xmath1 points .
@xmath631 .
\(a ) fix a point @xmath25 , and consider a ball @xmath35 centered at @xmath26 whose interior contains exactly @xmath632 points of @xmath8 .
the probability that a random point from @xmath8 hits @xmath35 is @xmath633 . as such , the probability that all the points of @xmath524 miss it is at most @xmath634 , implying the claim , if @xmath635 is sufficiently large .
\(b ) by lemma [ lemma : high : prob ] , for @xmath636 sufficiently large , we have that @xmath165 is in the range @xmath637{0cm}{0.4cm}}\right]}$ ] , and this holds with high probability . that is , for at least @xmath630 points of @xmath8 have @xmath638 , and let @xmath23 be this set of points . for every point @xmath639
, we have that the ball @xmath640 contains at most @xmath1 points of @xmath8 , by ( a ) . partitioning @xmath35 into equal radius rings of width @xmath641
, one of them must be empty of points of @xmath8 .
in particular , no cluster of @xmath642 can contain points on both sides of this ring , see figure on the right .
as such , at least @xmath630 of the points of @xmath8 are in clusters of @xmath378 that contain at most @xmath1 points .
\(c ) follows immediately from ( b ) , as @xmath643 .
[ rem : easyout ] in line [ line : easyout ] in algorithm @xmath644 { { { \textcolor[named]{redviolet}{\texttt{\bf{midnn } } } } } \index{algorithm!midnn@{{{\textcolor[named]{redviolet}{\texttt{\bf{midnn } } } } } } } } } \xspace}({\mathsf{p}})$ ] we return @xmath165 as our approximate middle nearest neighbor distance if @xmath567 { { { \textcolor[named]{redviolet}{\texttt{\bf{deciderm } } } } } \index{algorithm!deciderm@{{{\textcolor[named]{redviolet}{\texttt{\bf{deciderm } } } } } } } } } \xspace}$ ] returned @xmath645 . by lemma [ lemma : est : log : dist ] ( c ) we then know that @xmath646 with high probability and so @xmath165 is a valid approximation . in the unlikely event that @xmath647 , one simply runs @xmath644 { { { \textcolor[named]{redviolet}{\texttt{\bf{midnn } } } } } \index{algorithm!midnn@{{{\textcolor[named]{redviolet}{\texttt{\bf{midnn } } } } } } } } } \xspace}({\mathsf{p}})$ ] again .
this event can be determined with the call @xmath567 { { { \textcolor[named]{redviolet}{\texttt{\bf{deciderm } } } } } \index{algorithm!deciderm@{{{\textcolor[named]{redviolet}{\texttt{\bf{deciderm } } } } } } } } } \xspace}({\mathsf{p } } , { \ell},1/8)$ ] .
[ sec : logarithmic ] during an execution of @xmath644 { { { \textcolor[named]{redviolet}{\texttt{\bf{midnn } } } } } \index{algorithm!midnn@{{{\textcolor[named]{redviolet}{\texttt{\bf{midnn } } } } } } } } } \xspace}({\mathsf{p}})$ ] , @xmath648 { { { \textcolor[named]{redviolet}{\texttt{\bf{{lowspread } } } } } } \index{algorithm!{lowspread}@{{{\textcolor[named]{redviolet}{\texttt{\bf{{lowspread } } } } } } } } } } \xspace}({\mathsf{p } } , r , r)$ ] gets called ( see algorithm [ algorithm : spread ] ) only if the interval @xmath649 $ ] contains the values @xmath650 .
let @xmath8 be a point set in @xmath14 of size @xmath7 .
given @xmath651 , such that ( i ) @xmath652 , and ( ii ) @xmath653 $ ] , then @xmath648 { { { \textcolor[named]{redviolet}{\texttt{\bf{{lowspread } } } } } } \index{algorithm!{lowspread}@{{{\textcolor[named]{redviolet}{\texttt{\bf{{lowspread } } } } } } } } } } \xspace}({\mathsf{p } } , r , r)$ ] runs in @xmath12 time , and returns a value which is a @xmath75-approximation to some nearest neighbor distance @xmath654 $ ] , with high probability , for some point @xmath25 .
set @xmath655 . using the notation from ,
see algorithm [ algorithm : spread ] @xmath656 , let @xmath583 { { { \textcolor[named]{redviolet}{\texttt{\bf{net } } } } } \index{algorithm!net@{{{\textcolor[named]{redviolet}{\texttt{\bf{net } } } } } } } } } \xspace}{\!\left({\delta , { \mathsf{p}}}\right)}}$ ] and @xmath601 the @xmath586 size sample from @xmath524 . by lemma [ lemma :
high : prob ] , with high probability @xmath657{0cm}{0.4cm}}\right]}{\rule[-.3cm]{0cm}{0.8cm}}$ ] , and assume that this is the case . since @xmath524 is an @xmath49-net of @xmath8 , by claim [ claim : change ] , this changes distances by at most @xmath148 , which implies that @xmath658{0cm}{0.4cm}}\right ] } \subseteq { \left [ { \frac{1}{2}{{\mathtt{d}}_{1/2}{\!\left({{\mathsf{p}}}\right ) } } , \ , 2 { { \mathtt{d}}_{3/4}{\!\left({{\mathsf{p}}}\right ) } } } { \rule[-.2cm]{0cm}{0.4cm}}\right ] } \subseteq { \left [ { \frac{r}{2 } , 2r } { \rule[-.2cm]{0cm}{0.4cm}}\right]}. \end{aligned}\ ] ] by assumption . in particular , this implies that all the relevant distances from the computation of this median value are in the range @xmath659 $ ] . as such , when computing the median it is allowable to not compute precisely values outside this interval , as done by the algorithm
this implies that indeed returns the value @xmath660 , which is the desired approximation . as for the running time , observe that @xmath524 is an @xmath49-net , and an open ball with this radius around any point must be empty of any other net point . in particular ,
for any point @xmath661 , the maximum number of other points that can be in @xmath590 is @xmath662 .
therefore , for a point @xmath589 , computing @xmath663 takes polylogarithmic time ( using linear time preprocessing and a grid ) .
thus , overall , computing the set of distances @xmath664 takes polylogarithmic time , since @xmath665 .
hence , the overall running time is @xmath666 . [
sec : small ] the call to @xmath667 { { { \textcolor[named]{redviolet}{\texttt{\bf{{smallcomp } } } } } } \index{algorithm!{smallcomp}@{{{\textcolor[named]{redviolet}{\texttt{\bf{{smallcomp } } } } } } } } } } \xspace}({\ell } , { \mathsf{p}})$ ] by computes a point @xmath25 , such that @xmath668{0cm}{0.4cm}}\right]}$ ] .
the procedure succeeds and its running time is @xmath12 , with high probability , where @xmath513 . computing the components of @xmath60 takes linear time , and given @xmath60 , computing @xmath503 takes linear time . for each point in the sample @xmath601 , it takes @xmath669 time to compute its nearest neighbor since one only has to scan the point s @xmath670 sized connected component in @xmath60 . since @xmath671
, computing the nearest neighbor distances and taking the median takes polylogarithmic time overall .
now , @xmath667 { { { \textcolor[named]{redviolet}{\texttt{\bf{{smallcomp } } } } } } \index{algorithm!{smallcomp}@{{{\textcolor[named]{redviolet}{\texttt{\bf{{smallcomp } } } } } } } } } } \xspace}({\ell } , { \mathsf{p}})$ ] is called by only if the following conditions holds : @xmath672 .
namely , at least half of the points of @xmath8 have their nearest neighbor in their own cluster in @xmath673 , where @xmath674 .
let @xmath675 be the set of these points .
let @xmath378 be the partition of @xmath8 computed by .
we have that @xmath676 . by lemma
[ lemma : est : log : dist ] ( b ) , at least @xmath677 of the points of @xmath8 are in clusters of @xmath378 that contain at most @xmath1 points ( note , that some of these points might be a cluster containing only a single point ) .
let @xmath678 be the set of these points . as such , @xmath679 .
namely , at least @xmath680 of the points of @xmath8 , are in small clusters and these clusters have strictly more than one point .
furthermore , @xmath681 , and we conclude that @xmath682 .
now , by lemma [ lemma : high : prob ] , with high probability , the returned value is in the range @xmath683 , which implies that the returned value is in the range @xmath684 , as desired .
the analysis in the previous section implies the following .
[ lemma : high : prob : summary ] given a set @xmath63 of @xmath7 points , @xmath644 { { { \textcolor[named]{redviolet}{\texttt{\bf{midnn } } } } } \index{algorithm!midnn@{{{\textcolor[named]{redviolet}{\texttt{\bf{midnn } } } } } } } } } \xspace}({\mathsf{p}})$ ] computes , a value @xmath393 , such that there is a value of @xmath685 in the interval @xmath686 $ ] , where @xmath96 is a constant .
the running time bound holds with high probability . as shown in section [ sec
: f : n : n ] , given a constant factor approximation to some nearest neighbor distance , we can compute the corresponding nearest neighbor distance exactly in linear time using a grid computation .
let denote the algorithm that performs this additional linear time step after running . as argued above , replacing line [ line : gen : random ] and line [ line : rad ] in figure [ figure : a : algorithm ] @xmath196 with `` @xmath687{redviolet}{\texttt{\bf{midnnexact}}}}}\index{algorithm!midnnexact@{{{\textcolor[named]{redviolet}{\texttt{\bf{midnnexact}}}}}}}}}\xspace } ( { \mathsf{p}}_{i-1 } ) $ ] '' , turns into an algorithm that runs in linear time with high probability .
we thus have the following analogue of theorem [ theo : result : eps ] . for a constant @xmath688 , given an instance of a @xmath116-defined by a set of @xmath7 points in @xmath14 , with a @xmath116-decider with ( deterministic or high - probability ) running time @xmath12
, one can get a @xmath116-approximation to its optimal value , in @xmath689 time , with high probability .
similarly , for @xmath690 , given an instance of a @xmath691-defined by a set of @xmath7 points in @xmath14 , with a @xmath4-decider having ( deterministic , or high probability ) running time @xmath125 , then one can @xmath4-approximate the optimal value for this , in @xmath125 time , with high probability .
all the applications of section [ sec : applications ] can be modified so that their running time bound holds not only in expectation , but also with high probability .
note , that the above high probability guarantee is in the size of the input . as such , in later iterations of the algorithm ,
when the input size of the subproblem is no longer polynomial in @xmath7 ( say , the subproblem size is @xmath550 ) , the high probability guarantee is no longer meaningful . fortunately , this is a minor technicality once the subproblem size drops below @xmath692 , the remaining running time of is @xmath12 , with high probability .
this follows by a simple direct argument by the chernoff inequality , after @xmath693 iterations , with high probability at least @xmath694 iterations are successful , see the proof of lemma [ lemma : time ] for details . after this number of successful iterations
, the algorithm terminates .
this implies that the remaining running time is @xmath12 , with high probability , once the subproblem size is @xmath695 . as such
, our new high - probability algorithm is needed only in the initial `` heavy '' iterations of the algorithm , until the subproblem size drops below @xmath695 .
[ sec : conclusions ] there is still a lot of further research to be done in investigating this technique .
for example , looking into the implementation of this new algorithm in both standard settings and distributed settings ( e.g. , mapreduce ) . additionally , since now one can do approximate distance selection in linear time , maybe now one can get a speed up for other algorithms that do ( not necessarily point based ) distance selection .
our framework provides a new way of looking at distance based optimization problems , in particular through the lens of nets .
we know how to compute nets efficiently for doubling metrics and it seems one can compute approximate nets in near linear time for planar graphs .
for example , it seems the new technique implies that approximate @xmath0-center clustering in planar graphs can be done in near linear time .
this provides fertile ground for future research .
another interesting open problem arising out of the high probability linear time algorithm is the following : given sets @xmath8 and @xmath503 in @xmath14 , of size @xmath7 and @xmath696 , respectively , how long does it take to approximate the values in the set @xmath501 ( i.e. , for each point of @xmath503 we approximate its nearest - neighbor in @xmath8 ) .
this problem can be easily solved in @xmath32 time .
the open question is whether it is possible solve this problem in linear time .
the algorithm in this paper implies that we can compute exactly a value ( of a specified rank ) in this set in @xmath12 time ( even if @xmath697 is of size @xmath7 ) .
we currently have an algorithm with running time @xmath698 for approximating @xmath501 .
http://www.cs.duke.edu/~pankaj[p .
k. agarwal ] , http://www.uiuc.edu/~sariel[s .
har - peled ] , and http://www.cs.uiowa.edu/~kvaradar/[k .
r. varadarajan ] . http://cs.uiuc.edu/~sariel/papers/01/fitting/[approximating extent measures of points ] . , 51(4):606635 , 2004 .
http://www.cs.duke.edu/~pankaj[p .
k. agarwal ] , http://www.uiuc.edu/~sariel[s .
har - peled ] , and k. varadarajan .
geometric approximation via coresets . in j.
e. goodman , http://www.math.nyu.edu/~pach[j .
pach ] , and e. welzl , editors , _ combinatorial and computational geometry _ , math .
research inst . pub .
cambridge , 2005 .
http://compgeom.cs.uiuc.edu/~jeffe/[j .
erickson ] .
http://compgeom.cs.uiuc.edu/~jeffe/pubs/relative.html[on the relative complexities of some geometric problems ] . in _ proc
7th canad .
_ _ _ , pages 8590 , 1995 .
http://www.uiuc.edu/~sariel[s .
har - peled ] and a. kushal .
coresets for @xmath0-median and @xmath0-means clustering ] . in _ proc .
21st annu .
_ _ _ , pages 126134 , 2005 .
http://www.uiuc.edu/~sariel[s .
har - peled ] and s. mazumdar .
http://cs.uiuc.edu/~sariel/research/papers/03/kcoreset/[coresets for @xmath0-means and @xmath0-median clustering and their applications ] . in _ proc .
36th annu .
acm sympos .
theory comput .
_ _ _ , pages 291300 , 2004 .
http://www.uiuc.edu/~sariel[s .
har - peled ] and b. raichel . the frchet distance revisited and extended . in _ proc .
27th annu .
_ _ _ , pages 448457 , 2011 .
/ papers/10/frechet3d/. http://www.uiuc.edu/~sariel[s .
har - peled ] and b. raichel . http://cs.uiuc.edu/~sariel/papers/12/aggregate/[net and prune : a linear time algorithm for euclidean distance problems ] . in _ proc .
45th annu .
acm sympos .
theory comput .
_ _ _ , pages 605614 , 2013 .
r. krauthgamer and j. r. lee .
navigating nets : simple algorithms for proximity search . in _ proc .
15th acm - siam sympos . discrete algs .
_ _ _ , pages 798807 . society for industrial and applied mathematics , 2004 .
j. salowe .
parametric search . in j.
e. goodman and http://cs.smith.edu/~orourke/[j .
orourke ] , editors , _ handbook of discrete and computational geometry _ , chapter 37 , pages 683698 .
crc press llc , boca raton , fl , 1997 .
http://www.math.tau.ac.il/~michas[m .
sharir ] and e. welzl .
a combinatorial bound for linear programming and related problems . in _ proc .
9th sympos .
aspects comput .
volume 577 of _ lect .
notes in comp .
_ , pages 569579 .
springer - verlag , 1992 . | we provide a general framework for getting expected linear time constant factor approximations ( and in many cases s ) to several well known problems in computational geometry , such as @xmath0-center clustering and farthest nearest neighbor .
the new approach is robust to variations in the input problem , and yet it is simple , elegant and practical . in particular , many of these well studied problems which fit easily into our framework , either previously had no linear time approximation algorithm , or required rather involved algorithms and analysis .
a short list of the problems we consider include farthest nearest neighbor , @xmath0-center clustering , smallest disk enclosing @xmath0 points , @xmath0thlargest distance , @xmath0thsmallest @xmath1-nearest neighbor distance , @xmath0th heaviest edge in the @xmath2and other spanning forest type problems , problems involving upward closed set systems , and more .
finally , we show how to extend our framework such that the linear running time bound holds with high probability .
net and prune : a linear time algorithm for euclidean distance problems work on this paper was partially supported by nsf af awards ccf-0915984 and ccf-1217462 .
the full updated version of this paper is available online @xcite . a preliminary version of this paper appeared in stoc 2012 @xcite .
this work is supported by the national science foundation , under author s addresses : s. har - peled , department of computer science ; university of illinois ; 201 n. goodwin avenue ; urbana , il , 61801 , usa ; ; http://illinois.edu/\string~sariel . b. raichel , department of computer science ; university of illinois ; 201 n. goodwin avenue ; urbana , il , 61801 , usa ; 2uiuc.edu ; http://illinois.edu/\string~raichel2 . |
when a temperature difference @xmath4 exceeding a critical value @xmath5 is applied across a thin , horizontal layer of fluid , convection occurs .
@xcite this phenomenon is know as rayleigh - bnard convection ( rbc ) .
the fluid flow then forms a pattern .
the onset of convection is determined by the rayleigh number @xmath15 here @xmath16 is the isobaric thermal expansion coefficient , @xmath17 the acceletation of gravity , @xmath3 the thickness of the sample , @xmath18 the thermal diffusivity , and @xmath19 the kinematic viscosity .
linear stability - analysis for a laterally infinite system shows that convection will occur with a critical wave number @xmath20 ( @xmath21 is measured in units of @xmath3 ) when @xmath2 exceeds @xmath22 .
@xcite for the laterally infinite system the patterns that evolve beyond the onset of convection and their stability are determined by @xmath2 and by the prandtl number @xmath23 weakly nonlinear theory predicts that immediately above onset , a laterally infinite uniform system should yield a pattern consisting of parallel , straight rolls.@xcite for real experimental systems the patterns are influenced by the lateral boundaries which contain the fluid . even in these finite systems
excellent approximations of the predicted ideal straight rolls of the infinite system can be found under some conditions .
@xcite however , the boundaries can also lead to patterns of different symmetry .
thus , by heating a thin cylindrical convection cell gently from the side in addition to heating primarily from below , a pattern of _ concentric _ rolls can be stabilized .
a concentric pattern can also be the result of horizontal temperature gradients near a cylindrical side wall which are intrinsic to the particular construction of the sample cell .
the concentric pattern will consist of @xmath24 rolls , where @xmath24 is an integer . of these
, the one in the middle really would be more properly viewed as a convection _ cell _ , with either up - flow or down - flow in its center .
it is often referred to as the umbilicus .
this type of pattern , know also as a target pattern , was studied extensively by koschmieder and pallas @xcite for relatively large @xmath25 .
the rolls have a mean dimensionless wave length @xmath26 where @xmath27 ( @xmath2 is the sample radius ) is the aspect ratio of the sample and where @xmath24 is the number of rolls along a radius .
the corresponding wave number is @xmath28 .
koschmieder and pallas found that the wave length of the rolls away from the center and the side wall increased with increasing @xmath4 .
they also observed a sequence of transitions with increasing @xmath4 , each of which involved the loss of a roll at the sample center . later work for smaller @xmath25 showed that at a certain value @xmath29 of @xmath30 the target pattern becomes unstable .
@xcite the instability is known as the focus instability .
@xcite how this occurs in detail depends on @xmath25 , @xmath31 , and on the nature of the confining side walls . here
we describe what happens for our particular sample ( @xmath32 ) , where the side walls fixed the phase of the pattern at the radial position @xmath33 ( @xmath34 is scaled by the cell thickness @xmath3 ) . for the work of ref.@xcite ( @xmath35 )
the observed phenomena were similar .
as @xmath8 is increased in small steps from small values , the umbilicus shrinks .
the result is a gradual decrease of the local wave numbers of the rolls away from the center .
when @xmath8 is increased beyond @xmath29 , the umbilicus breaks the cylindrical symmetry by moving off center .
this transition is continuous and the off - center patterns are stationary in time , but the radial umbilicus position is @xmath8 dependent . as @xmath8 is further increased , the umbilicus becomes even smaller , and then disappears at @xmath36 .
at that point the pattern once more approaches a state of cylindrically symmetric rolls , although perfect rotational invariance is not fully recovered . with further increase of @xmath8 , this process repeats itself , with another roll loss at @xmath37 , and so forth .
the net result is a gradual reduction with increasing @xmath8 of the wave numbers of the rolls between transitions and a discontinuous change of @xmath38 at each transition .
the transitions are hysteretic , occurring at @xmath39 when @xmath4 is decreased .
we studied target patterns for @xmath40 in a sample with @xmath41 using methanol with @xmath0 .
we developed high - resolution numerical umbilicus - detection algorithms and determined quantitatively the umbilicus displacement @xmath42 with increasing as well as decreasing @xmath8 . at small @xmath8 the results revealed the location of the focus instability , at @xmath43 .
previous measurements , @xcite for @xmath44 and a larger @xmath45 , had found the instability at lower values , near @xmath46 . over the whole @xmath8 range of our experiment the data for @xmath42 , together with measurements of an average wave number @xmath47 away from the center and the side wall , revealed eight hysteretic transitions .
we observed large umbilicus displacement only for increasing @xmath8 ; for decreasing @xmath8 the hysteresis yielded transition points @xmath14 which were sufficiently low to avoid the @xmath8-range over which the umbilicus displacement was large . for @xmath48 cross
rolls @xcite formed at the outermost roll when the system was close to a transition at @xmath13 .
this was particularly pronounced on the side opposite to the umbilicus displacement where the local wave lengths were exceptionally large when @xmath42 was large .
nonetheless , the general nature of the pattern was maintained up to our largest @xmath8 values .
for @xmath49 periodic time dependence associated with the oscillatory instability @xcite developed in the region near the cross rolls . a complete characterization of the pattern involves a knowledge of the entire wave - director field @xmath50 where @xmath51 is the angular and @xmath34 the radial position in the sample . rather than simply measuring an average wave number along a cell diameter ( as was done in previous work
) , we implemented a local wave - director analysis @xcite and determined @xmath50 .
from @xmath50 we could then determine various averaged quantities , including @xmath47 .
the results for @xmath47 , determined over a radius range that excluded the rolls near the center and the side wall , were discontinuous at the transitions .
the hysteresis loop of each transition was traced out quantitatively .
we found that the azimuthal average @xmath52 had interesting structure as a function of @xmath34 , showing different selection at the umbilicus , in the bulk of the sample , and at the cell wall .
the competition between the selection mechanisms yielded a broad maximum of @xmath53 near @xmath54 . for our experimental conditions
it did not yield any radially traveling waves , as predicted by tuckerman and barkley @xcite for the case of conducting side walls . as was found in previous work , @xcite the values of @xmath47 selected in the bulk were in good agreement with the predicted wave numbers for the far field of
infinitely extended concentric rolls when the discontinuities at the transitions were smoothed out .
we also report results for the wave numbers selected at the umbilicus and at the side wall ; but for these there seem to be no theoretical prediction . in the next section
we shall review theoretical predictions for the selection by concentric rolls and relevant previous experiments . then . in sect .
[ sec : experiment ] , we shall discuss the experimental apparatus and procedures , as well as the image - analysis methods .
section [ sec : results ] gives our results .
it consists of a discussion of the patterns observed in various @xmath8 ranges , of a presentation of the umbilicus - displacements results , and of a presentation of our wave - number results . a brief section summarizing
this work ends the paper .
theoretical predictions of the wave numbers @xmath55 selected in the far field of infinitely extended target patterns have been made by several authors .
@xcite for the infinite system wave - number adjustment can take place by expansion of the pattern to large distances .
finite laboratory systems differ from this in an important way . the side wall , under typical experimental conditions , pins the phase of the pattern and prevents this expansion .
thus , a change of the average wave number with changing @xmath8 is possible only at the umbilicus .
however , also at the umbilicus unhindered phase slip is not possible .
instead , a discontinuous and hysteretic process involving the destruction or creation of a convection cell occurs and leads to a discontinuous change of the wave - number field at the transition . the discontinuous effect on the average wave number decreases as the aspect ratio of the sample increases because the loss of a single cell in the center is a smaller perturbation for a larger sample .
thus , even for the sample with a boundary one expects , in the large-@xmath31 limit , a continuous curve @xmath55 as a function of @xmath8 .
theoretically the location of this curve is determined by the rotational symmetry of the target pattern which does not permit the mean flow that is induced by roll curvature under less symmetric conditions .
thus , the mean flow must be balanced precisely by a pressure gradient.@xcite this condition leads to a unique wave number @xmath55 .
initial predictions of the selected wave numbers were applicable only for small @xmath8 and are given by @xcite @xmath56 with @xmath57 , @xmath58 , and @xmath59 . here
@xmath60 is the critical wave number at the onset of convection . in fig .
[ fig : s_b ] we show @xmath61 as a function of @xmath25 .
for the present work we have @xmath0 and @xmath62 .
the measurements to be described below yielded the experimental value @xmath63 .
this result is shown as the circle in the figure .
it falls slightly above , but is generally in good agreement with , the prediction . for the work of ref .
@xcite ( @xmath64 ) the theory yields @xmath65 , but no corresponding result has been extracted from the data .
measurements for @xmath44 were reported in ref .
@xcite and yielded a value for @xmath61 close to zero , but a precise number was not quoted . for that case one obtains the prediction @xmath66 , but @xmath67 changes rapidly with @xmath25 and passes through zero at @xmath68 . in the limit of small @xmath8 as a function of the prandtl number @xmath25
the circle is the result of the resent work to be discussed below.,width=321 ] calculations valid at larger @xmath8 were made by buell and catton @xcite ( bc ) and by newell et al .
@xcite using a combination of analytic and numerical methods .
bc gave results for a few values of @xmath25 and @xmath2 , and we shall compare our measurements with interpolations between these predictions .
we show their results for a prandtl number close to that of our experiment in fig . [
fig : bussball ] . also shown in that figure
are the nearby stability boundaries of infinitely extended uniform straight rolls.@xcite although in the large-@xmath25 limit it is expected that the zig - zag instability will coincide with the selected wave number,@xcite it is clear from the figure that @xmath69 is still far from that limit . .
dashed line : zigzag instability @xcite .
solid line : cross - roll instability @xcite . also shown , as a dotted line , are the wave numbers predicted for the far field of infinitely extended concentric rolls @xcite .
the two solid circles correspond to our experimental observation of the cross - roll instability near the side wall.,width=321 ] to our knowledge there is as yet no prediction of the wave - number selection at the umbilicus itself .
the results to be presented here show that the wave number selected there is smaller than the one in the far field . aside from the phase
pinning mentioned above , the side walls of experimental cells have an additional influence on the pattern .
the boundary conditions imposed by the walls can yield a separate and competing selection mechanism.@xcite this mechanism is expected to depend on the conductivity of the side walls relative to that of the fluid , and it is not understood quantitatively for typical experimental conditions .
for perfectly conducting side walls , tuckermann and barkley @xcite predicted a pattern of radially - traveling waves .
the traveling nature of the pattern can be understood in terms of the competition between the selection by the curved rolls on the one hand and by the wall on the other which can lead to a wave - number gradient which in turn can lead to a non - zero time derivative of the phase of the pattern .
@xcite however , to our knowledge such radially traveling waves have not yet been observed in experiments .
@xcite for @xmath25 values not too small the wave numbers predicted for curved rolls do not depend very strongly on @xmath25 .
in this @xmath25 range the wave numbers of target patterns have been determined before , @xcite but to our knowledge they were never compared quantitatively with the predictions of bc .
koschmieder and pallas @xcite determined a weighted average of the wave lengths of all rolls except for the outermost one . in fig .
[ fig : k_others]a we show their results for @xmath70 , and for @xmath71 and 916 .
the solid line is the prediction of ref .
@xcite for @xmath72 .
aside from a lateral shift , the overall shape of a curve passing through the data is very similar to the theoretical curve .
since the data do not extrapolate to @xmath73 as @xmath8 vanishes , the shift presumably is due to experimental uncertainties of the length scales involved in the determination of @xmath21 . in fig .
[ fig : k_others]b we show results from ref .
@xcite which were for @xmath74 and @xmath75 .
these data were obtained by measuring the average wave length of 2 rolls near @xmath76 .
they seem to show considerable scatter , but actually this is due to the discreteness of the values of @xmath38 which is noticeable in the figure and which becomes more apparent for this relatively small @xmath31 where the system contains only a small number of rolls . in this system
hysteretic transitions were clearly observed .
the theoretical curve @xcite for this @xmath25 value ( solid line ) is a good smoothed representation of the data .
511 ( circles ) and 916 ( squares ) , and for @xmath70 . for ( b ) and
( c ) open ( solid ) symbols were taken with increasing ( decreasing ) @xmath8 .
( b ) : from ref .
@xcite for @xmath77 and @xmath75 .
( c ) : from ref .
@xcite for @xmath78 and @xmath79 .
the solid lines in all figures are the predictions from refs .
@xcite and @xcite for the relevant @xmath25 values.,width=264 ] finally , in fig .
[ fig : k_others]c , some of the results reported by croquette and pocheau @xcite are shown .
those measurements were for @xmath78 and @xmath80 , and were obtained by measuring the average roll width along a diameter in a region away from the center and sides where the roll width appeared constant .
the solid line is the numerical result @xcite for @xmath64 .
the agreement clearly is very satisfying .
the data do not reveal much difference when @xmath8 is increased ( open symbols ) or decreased ( solid symbols ) .
the authors report observing hysteretic transitions , and thus it is somewhat surprising that this hysteresis does not manifest itself in the wave - number measurements .
all of the previous investigations revealed that the roll adjacent to the side wall was exceptionally wide , but none of the prior investigations made any attempt to determine quantitatively the wave number selected at the wall or at the cell center .
we used a rayleigh - bnard convection apparatus described in detail elsewhere @xcite .
the top plate was a single - crystal - sapphire disk , and the bottom plate was a polished aluminum disk with a metal - film heater mounted under it to provide uniform bottom - plate heating .
temperature - controlled water was circulated over the top plate .
the bottom - plate temperature was controlled to create the desired temperature difference .
the mean temperature was maintained at @xmath81c .
for both the top and bottom plate , the temperature varied about the set temperature by less then @xmath82c .
the cell wall was made of lexan ( thermal conductivity 0.23 w / m k ) with an outer diameter of 9.6 cm and an inner diameter of @xmath83 cm .
a 0.012 cm diameter manganin wire was embedded in the cell wall to provide the side - wall heating .
this wire had a resistance of about 13 @xmath84 , and typically dissipated 0.26 w. the height of the side wall was 0.229 cm , though the actual fluid height @xmath3 depended on the compression of an o - ring which sealed the cell .
the actual fluid height was determined by measuring @xmath5 and then inferring @xmath3 from the fluid properties and the measured @xmath85c .
we found @xmath86 cm , giving an aspect ratio @xmath41 .
the fluid used was methanol which , at @xmath87c , had a prandlt number @xmath0 and a thermal conductivity of @xmath88 w / m k. the vertical diffusion time was @xmath89s .
the convection rolls were imaged using the shadowgraph method @xcite . at large @xmath8 , strong spatial variation of the refractive index caused nonlinear optical effects in the images,@xcite but the general features of the pattern could still be discerned . this effect limited our ability to do wave - number measurements for @xmath90 .
nonetheless , the location of the umbilicus could still be measured . in order to obtain concentric rolls , @xmath4 was set to zero and the side - wall heater - power was set to 0.26 w. after waiting two hours , we increased the temperature difference to @xmath91c ( @xmath92 ) and allowed the system to equilibrate for one hour .
after this the temperature difference was abruptly increased to @xmath93c ( @xmath94 ) and the system was allowed to equilibrate for two hours .
the jump of @xmath4 was necessary because defects typically formed in the interior when the temperature difference was raised gradually .
once concentric rolls were obtained , the temperature difference was adjusted in small steps ( typically between @xmath95c and @xmath96c ) and the system equilibrated for one hour before taking an image at each temperature step . the power supplied by the side - wall heater was kept constant throughout the run .
section of the pattern for @xmath8 values near the transition at @xmath97 .
they are from a run in which @xmath4 was increased with @xmath98c steps @xmath99 .
the system was equilibrated for one hour after each temperature step before taking a picture .
the images were divided by a background image , and then processed numerically so as to greatly enhance their contrast .
they are for ( a ) @xmath100 , ( b ) @xmath101 , ( c ) @xmath102 , and ( d ) @xmath103 .
, width=264 ] shadowgraph images of some of the patterns are shown in figs .
[ fig : transition_up ] and [ fig : transition_higher ] .
all images were divided by a background image taken at @xmath104 .
they were filtered by setting to black ( white ) all pixels that fell below ( above ) a certain threshold . a gaussian blur
was then applied to the chopped image .
we developed an algorithm to determine the local wave - director field @xmath105 .
this method will be described in detail elsewhere .
first the orientation field @xmath106 at each point of the image was calculated using a method similar to one introduced by cross et .
al . , @xcite but higher angular resolution was employed .
the wave - number field @xmath107 was then calculated at each point by moving orthogonal to the the roll orientation at that point and determining the number of pixels which falls within one wave length , assuming a locally periodic structure .
determinations of the wave number @xmath108 from the images require a knowledge of @xmath3 and of the horizontal distance @xmath109 between adjacent pixels .
the uncertainty of @xmath3 yielded an uncertainty of only a small fraction of a percent .
we determined @xmath109 by counting the number of pixels which spanned the sample diameter .
this yielded @xmath110 and led to a combined uncertainty of about 2 % for @xmath111 .
an extrapolation of the measured @xmath111-values with decreasing @xmath8 to @xmath112 gave @xmath113 , which is within the experimental uncertainty of the value 3.117 for the infinitely extended set of straight rolls .
we measured the location of the umbilicus by using a @xmath114 square section surrounding the center .
the location of the umbilicus was estimated by computing the maximum of the cross - correlation of the measured orientation field @xmath115 and the orientation field for perfectly centered rolls @xmath116 .
this estimate was used only to identify the rough location of the umbilicus .
once the umbilicus was identified , we used an edge - detection algorithm on the umbilicus to determine its position with a resolution of about @xmath117 m or @xmath118 .
and @xmath119 . the optical effects that distort
the roll patterns can be seen in images ( b ) and ( e ) in the rolls that are being compressed by the off - center umbilicus .
these images are from the same run as those in fig .
[ fig : transition_up ] .
the values of @xmath120 are ( a ) 5.649 , ( b ) 5.707 , ( c ) 5.821 , ( d ) 6.910 , ( e ) 7.368 , and ( f ) 7.482 .
, width=264 ] we covered the range @xmath121 and with increasing @xmath8 observed eight transitions at @xmath10 under quasi - static conditions . a closeup view of images near the third transition is shown in fig .
[ fig : transition_up ] . at each transition
a roll was lost at the umbilicus . for all but the first transition
the umbilicus first moved away from the center as @xmath8 grew .
the angle of the umbilicus displacement did not have a preferred value , as can be seen from the examples in figs .
[ fig : transition_up ] and [ fig : transition_higher ] .
this indicates that the cell was sufficiently uniform . at the transition the umbilicus collapsed , reducing the number of rolls in the system by one .
after a roll was lost , the new umbilicus immediately moved back toward the center of the sample but never quite reached it .
this process was analogous to that observed in refs .
@xcite and @xcite .
the umbilicus displacement before each transition produced a noticeable azimuthal variation of the wave number , with higher wave numbers in the direction of the displacement , and lower wave numbers opposite the displacement direction .
close to but before the sixth transition , for @xmath122 , a patch of cross rolls @xcite formed along the outermost roll on the side opposite the direction of umbilicus displacement . at that point
the local wave number @xmath111 of the concentric rolls was the smallest of the entire wave - number field , and was about 1.95 ( @xmath123 ) .
the corresponding point in the @xmath124 plane is shown as a solid circle in fig .
[ fig : bussball ] . the wave number @xmath125 of the cross rolls that formed was about 4.3 .
the value of @xmath111 is somewhat larger than the prediction @xcite @xmath126 ( @xmath127 ) for laterally infinite straight rolls at this value of @xmath8 .
this suggests that the roll curvature or the side wall reduces the stablity of the rolls against the cross - roll perturbation .
the wave number @xmath125 of the cross rolls that formed is larger than the predicted value @xcite @xmath128 .
the cross rolls disappeared when @xmath8 was increased further , the pattern lost the middle roll , and the umbilicus returned toward the center . at that point
all parts of the pattern had returned to wave numbers safely in the stable busse balloon . before the next ( seventh ) transition , near @xmath129 , cross rolls appeared again . in this case we found @xmath111 = 1.86 ( @xmath130 ) and @xmath131 .
this point is shown as well in fig .
[ fig : bussball ] .
again the value of @xmath125 is somewhat larger than the prediction @xmath132 for infinitely extended uniform straight rolls .
after this transition the cross rolls fanned out to cover the entire outer roll .
images near the two transitions are shown in fig .
[ fig : transition_higher ] .
the cross rolls remained along the outer roll , and after the eighth transition , spread to the inner rolls as @xmath8 was further increased . despite the cross rolls , there was still a discernible pattern of nearly concentric rolls , with the cross rolls superimposed on these . around @xmath133
the focus began to move off center but did not complete a ninth transition .
this is illustrated in fig .
[ fig : large_eps]a for @xmath134 .
cross rolls spread to the inner rolls , and as @xmath8 was further increased , the umbilicus returned to the center without losing a roll , as seen in fig .
[ fig : large_eps]b for @xmath135 . and ( b ) @xmath135 from the same run as the one used for figs .
[ fig : transition_up ] and [ fig : transition_higher ] ) . the pattern in image ( a ) was stationary , while in image ( b ) the cross rolls oscillated along the axis of the main rolls . at these large @xmath8 values
the images are strongly influenced by nonlinear effects in the shadowgraph method.,width=321 ] the pattern became time dependent near @xmath136 , where small patches of traveling waves developed .
these were superimposed upon and orthogonal to the concentric roll , and traveled along the axes of the concentric rolls .
the travelling waves occurred first on the third or fourth roll from the wall , and with increasing @xmath8 spread throughout the cell .
we asociate this phenomenon with the oscillatory instability predicted by clever and busse , @xcite although for our prandtl number the cross - roll instability precedes the oscillatory instability .
the cross rolls observed here had been previously seen by croquette et .
@xcite , though there are differences in the way they appeared .
croquette et .
al . observed that cross rolls moved in from the side wall only after @xmath8 became larger than about 10 .
they did not observe a relationship between the umbilicus transitions and the cross - roll formation . despite nearly identical prandlt numbers and aspect ratios ( @xmath137 , @xmath80 ) , croquette et .
al . could produce off - center patterns near transitions comparable to those shown in fig .
[ fig : transition_higher ] with no cross rolls @xcite . in our system
cross rolls permanently covered the outer roll for @xmath138 , well before croquette et .
al . first observed any .
this difference can perhaps be explained by a difference in the strength of side - wall forcing .
a future study of the influence of various levels of side - wall heating should shed some light on this issue .
sections of images obtained with decreasing @xmath8 near the transition at @xmath139 .
this is from a run where @xmath8 was first increased to 3.49 and then decreased in steps @xmath140.the values of @xmath8 are ( a ) 3.49 , ( b ) 2.94 , ( c ) 2.67 , and ( d ) 2.65 .
note how the size of the middle roll changes from image ( a ) to ( c ) , and how in ( c ) the umbilicus remains close to the center even for an @xmath8 value only just above @xmath139.,width=264 ] the nature of the patterns for decreasing @xmath8 depended on the initial state . when starting above the oscillatory instability ( say @xmath49 ) , decreasing @xmath8 broke up the concentric rolls .
wall foci evolved , displaced the umbilicus , and emitted new rolls as @xmath8 was further decreased .
however , the side - wall forcing was strong enough to re - orient the outermost rolls parallel to the sidewall for @xmath141 . by @xmath142
the outer roll had returned to being completely parallel to the sidewall .
however , defects remained in the interior of the pattern , and the system did not return to concentric rolls as @xmath8 was decreased below zero .
starting from @xmath143 with concentric time - independent rolls , new rolls were generated at the umbilicus for distinct values @xmath11 as @xmath8 was decreased quasi - statically . during these transitions the umbilicus stayed close to the center , as illustrated in fig .
[ fig : transition_down ] .
we did not determine the maximum @xmath8 at which we could start decreasing @xmath8 without loosing the concentric pattern .
away from the sample center as a function of @xmath8 for increasing @xmath8 . this data is from the same run as the images in figs .
[ fig : transition_up ] and [ fig : transition_higher ] , where we increased @xmath8 in steps of @xmath144 and waited 1 hour after each temperature step for the system to equilibrate before taking a picture . , width=264 ] with increasing @xmath8 we measured the displacement @xmath42 of the umbilicus from the center of the cell as a function of @xmath8 for @xmath145 .
results over the entire @xmath8-range are shown in fig .
[ fig : dist_umbil ] . a more detailed plot of the results for @xmath146 is shown in fig .
[ fig : dist_umbil_detail ] .
seven transitions , at @xmath147 , are apparent from these data . at the first transition , at @xmath148 , we did not resolve any discontinuity in @xmath42 .
prior to each observable transition there was a large displacement , followed by a relaxation back toward a centro - symmetric pattern when the umbilicus collapsed .
figure [ fig : dist_umbil_detail2 ] shows the data for @xmath42 over the range @xmath149 . here
we see that the pattern is , within possible systematic errors of @xmath42 , accurately centered for @xmath150 . above @xmath29
the increase of @xmath42 is consistent with an initially linear dependence , and thus we fitted the results to @xmath151 over the range @xmath152 and to @xmath153 for @xmath154 . here
@xmath155 corresponds to the small offset , well within our possible systematic errors , which is found even at small epsilon .
the parameters @xmath156 , and @xmath29 were least - squares adjusted .
the fit gave @xmath157 , @xmath158 , and @xmath159 .
it is shown as the dashed curve in fig .
[ fig : dist_umbil_detail2 ] .
the statistical error of @xmath160 indicates that a fit to a quadratic equation ( i.e. eq .
[ eq : delta(eps ) ] with @xmath161 ) should be equally good .
it yielded @xmath162 and @xmath163 .
we identify @xmath29 as the focus instability , i.e. as the first instability of the centro - symmetric pattern with our prevailing wave number and our prandtl number @xmath0 . of the results shown in fig .
[ fig : dist_umbil ] .
the thin solid line shows the umbilicus displacement for decreasing @xmath8 .
, width=264 ] of the results shown in figs .
[ fig : dist_umbil ] and [ fig : dist_umbil_detail ] .
this graph reveals the initial focus instability of the centro - symmetric pattern at @xmath29 .
the points @xmath164 and @xmath148 are the limits of the hysteresis loop associated with the transition which are revealed by the wave - number measurements to be discussed below .
the dashed curve is a fit of eq .
[ eq : delta(eps ) ] to the data.,width=264 ] when @xmath8 was decreased , transitions involving the addition of a roll at the umbilicus occurred at @xmath14 , but were not associated with large displacements of the umbilicus . experimental results for @xmath42 with decreasing @xmath8
are shown by the thin solid line in fig .
[ fig : dist_umbil_detail ] .
this difference between increasing and decreasing @xmath8 was observed also by steinberg et .
al . @xcite . at @xmath165 .
the highlighted section covers the range @xmath166 and shows the annulus over which we averaged the wave numbers to obtain the values @xmath47 shown in figs .
[ fig : eps_k_down ] , [ fig : eps_k ] , and [ fig : eps_k_detail ] below.,width=226 ] obtained with decreasing @xmath8 .
the solid line is a fit of eq .
[ eq : k_b ] to the data.,width=264 ] we measured the averaged wave numbers @xmath167 for @xmath168 for both increasing and decreasing @xmath8 .
for comparison with theoretical predictions one would like to make this measurement well away from the umbilicus ; in practice there is a limit set by the aspect ratio of the sample and by a competing selection mechanism associated with the side wall . as we shall discuss below
, the local wave numbers have a maximum at a radial position in the range @xmath169 ( @xmath170 ) .
thus we selected an annular region extending over this radial range as illustrated by the highlighted region in fig .
[ fig : image ] to compute @xmath47 . in fig .
[ fig : eps_k_down ] the results obtained with decreasing and relatively small @xmath8 are shown .
the solid line is a fit of the equation @xmath171 to the data over the range @xmath172 .
the fit yielded @xmath173 , @xmath174 , and @xmath175 . the result for @xmath176 , within the a priori estimate of the experimental uncertainties , agrees with the theoretical value @xmath73 for infinitely extended straight rolls . the result for @xmath61
is shown in fig .
[ fig : s_b ] as an open circle .
it falls slightly above the predicted value @xmath177 .
results for @xmath178 over the entire @xmath8 range are shown in fig .
[ fig : eps_k ] for both increasing ( open circles ) and decreasing ( solid circles ) @xmath8 . within experimental uncertainties they are consistent with previous measurements for similar @xmath25 .
the jumps in the wave number correspond to the loss or creation of a roll at a transition .
the hysteresis in the wave - number selection , previously observed by others , @xcite is clearly visible in our results .
it indicates partial pinning of the phase of the pattern at the umbilicus .
we believe that this phase pinning is responsible also for the difference between experiment and theory for the value of @xmath61 . on average
the data are in excellent agreement with the calculations by buell and catton @xcite which are given by the solid line through the data .
vs. @xmath8 .
squares : the reduced wave numbers @xmath179 selected at the side wall .
open symbols : increasing @xmath8 .
closed symbols : decreasing @xmath8 .
solid line through the data for @xmath180 : the prediction of refs . @xcite and @xcite .
lower solid line : the zigzag instability for @xmath137 .
, width=321 ] near the focus instability at @xmath29 and the first hysteretic transition at @xmath148 and @xmath164 .
the symbols are as in fig .
[ fig : eps_k ] , width=321 ] .
[ fig : eps_k_detail ] in order to illustrate the relationship between the first hysteretic transition at @xmath164 and @xmath148 and the focus instability at @xmath29 , we give an expanded view of the selected wave numbers at small @xmath8 in fig . [
fig : eps_k_detail ] .
one sees that the discontinuities at @xmath164 and @xmath148 are quite sharp , and that @xmath29 is located near the middle of this first hysteresis loop .
values of @xmath14 and @xmath13 for all eight transitions are given in table [ table : eps ] . 0.1 in .location of the hysteretic transitions with decreasing ( @xmath14 ) and increasing ( @xmath13 ) @xmath8 .
for @xmath14 , @xmath181 the uncertainty @xmath182 . for @xmath183 and @xmath184 , @xmath185 . for @xmath13 , @xmath186 .
[ cols="^,^,^ " , ] [ table : eps ] .
bottom : azimuthal averages of the local wave numbers .
the vertical dashed line shows the location of the side wall , and the two vertical dotted lines show the inner and outer edge of the annulus highlighted in fig.[fig : image ] and used to compute the average wave numbers @xmath47 shown in figs .
[ fig : eps_k_down ] , [ fig : eps_k ] , and [ fig : eps_k_detail ] . from top to bottom , the first three data sets are for @xmath187 , 1.297 , and 2.614 .
the lowest set is for a synthetic concentric pattern ( down - shifted in the figure by 0.1 ) with a wave - number distribution ( shown by the solid line ) given by straight - line representations of the experiment for @xmath188 .
the solid straight lines through the experimental data at large @xmath34 are fits to the data with @xmath189 and were used to find @xmath190 .
the dashed straight lines through the experimental data at small @xmath34 are fits for @xmath191 and were used to determine @xmath192.,width=321 ] selected at the origin .
open symbols : increasing @xmath8 .
closed symbols : decreasing @xmath8 .
rough solid line : @xmath190 for decreasing @xmath8 .
smooth solid lines : predictions for @xmath47 @xcite and for the zigzag and cross - roll instabilities @xcite.,width=321 ] the top part of fig .
[ fig : k_rad ] gives an example for the radial variation of the shadowgraph intensity .
in the bottom part we display the radial variation of the azimuthal average @xmath193 of @xmath111 for three @xmath8 values .
also shown for comparison ( bottom curve , down - shifted by 0.1 ) is the result of our analysis for a synthetic pattern of concentric rolls with a radial wave - number distribution shown by the solid line which represents a smooth curve through the results for @xmath188 .
our analysis yields results for @xmath53 which oscillate as a function of @xmath34 .
this is particularly noticeable for @xmath34 below the maximum of @xmath53 .
these oscillations are not present in the analysis of a synthetic concentric pattern with a uniform wave number , but as shown by the lowest set in fig .
[ fig : k_rad ] , they do appear in the analysis of a concentric pattern with a radial wave - number distribution equal to the smooth curve through the data for @xmath188 .
thus we regard them to be an artifact of the numerical procedures .
we note , however , that the amplitude of the oscillations is only about 0.5% of @xmath111 , and that a smooth curve through them is a good representation of the actual wave numbers . from fig .
[ fig : k_rad ] one sees that the experimental data for @xmath193 , as mentioned above , have a maximum in the range @xmath169 , and this range was used to compute @xmath47 .
for @xmath194 the data were fitted by straight lines as shown in the figure , and these fits were used to extrapolate the selected @xmath53 to @xmath195 .
the results for @xmath190 are shown in fig .
[ fig : eps_k ] above .
they show , as noted by others , @xcite that the outermost roll has an anomalously large wave length ( small wave number ) .
we note that @xmath190 is not influenced significantly by the transitions at the umbilicus .
interestingly , the wall - selected wave number does not extrapolate to @xmath176 as @xmath8 vanishes .
instead , it crosses the zigzag instability - line of the laterally infinite system ; but in the experiment for this finite system no instability was observed .
we are not aware of any quantitative previous measurements or predictions for @xmath190 .
the experimental results at small @xmath34 suggest that yet another unique wave number is selected by the umbilicus itself .
fits of straight lines to the data , this time with @xmath196 , could be used to estimate the selected values @xmath192 at @xmath197 .
the results are given in fig .
[ fig : k_0 ] . for comparison
we show in that figure also the prediction for the concentric - roll selection @xcite ( smooth solid line ) and the experimental results for @xmath190 ( jagged line ) .
one sees that @xmath192 is influenced strongly by the hysteretic transitions at the umbilicus . at large @xmath8 @xmath192
is on average close to the result for @xmath190 , but as @xmath8 decreases , @xmath192 is somewhat larger and tends toward an intermediate value somewhere between @xmath190 and @xmath198 as @xmath199 . also for @xmath192 there appear to be no previous measurements or predictions .
we studied rayleigh - bnard convection of a fluid with a prandtl number @xmath0 in a cylindrical container of aspect ratio @xmath200 in the presence of static side - wall forcing .
patterns of concentric rolls were obtained , and we studied the quasi - static evolution of these patterns for both increasing and decreasing @xmath8 . for increasing @xmath8 , over a range @xmath201 , the pattern underwent eight transitions where the middle roll moved off center and then disappeared as @xmath8 was increased quasi - statically .
we measured the displacement of the umbilicus for @xmath202 . above @xmath203
the pattern became time dependent with oscillating rolls traveling along the axis of the concentric rolls . for decreasing @xmath8 ,
the concentric - roll pattern was lost when the initial @xmath8 was too large . when @xmath8 was decreased from a moderate initial value , the concentric - roll pattern remained and new rolls nucleated at the umbilicus . any umbilicus displacement preceding the nucleation of new rolls was modest , in sharp contrast with the behavior of the pattern for increasing @xmath8 . using new image - analysis techniques
, we determined the wave - number field for @xmath204 . from this we computed the azimuthal average @xmath193 .
there was a very noticeable radial gradient of @xmath193 both near the side wall at @xmath33 and near the center at @xmath197 .
averaging the wave - number field over the annulus @xmath169 where @xmath193 was relatively constant , we obtained an average wave number @xmath47 .
we found that @xmath47 was discontinuous and hysteretic at the transitions where new rolls were formed or disappeared at the umbilicus .
nonetheless , a smooth curve through the data agreed well with predictions @xcite for the wave - number selection in the far field of concentric rolls .
these average wave numbers were also consistent with previous work by others . @xcite near the wall and near the center
the measured @xmath193 were extrapolated to @xmath33 and to @xmath197 to obtain the wave numbers @xmath190 and @xmath192 selected by the wall and the umbilicus .
we are not aware of previous measurements or predictions for these quantities .
we are grateful to worawat meevasana for his guidance during the early stages of this experiment , and to w. pesch for stimulating discussions .
this work was supported by us national science foundation grant dmr02 - 43336 .
perhaps the most convincing evidence for parallel straight rolls is found in the experiments of ref.@xcite where the effect of side walls on the pattern was minimized by introducing a radial ramp of the cell spacing near the cell wall which gradually reduced the convection amplitude to zero as the wall was approached .
inward - traveling concentric rolls have been reported in ref .
@xcite , but for a probably unrelated case of convection in the presence of rotation about a vertical axis with periodic modulation of the angular velocity . | we present experimental results for patterns of rayleigh - bnard convection in a cylindrical container with static side - wall forcing . the fluid used was methanol , with a prandlt number @xmath0 , and the aspect ratio was @xmath1 ( @xmath2 is the radius and @xmath3 the thickness of the fluid layer ) . in the presence of a small heat input along the side wall , a sudden jump of the temperature difference @xmath4 from below to slightly above a critical value
@xmath5 produced a stable pattern of concentric rolls ( a target pattern ) with the central roll ( the umbilicus ) at the center of the cell .
a quasi - static increase of @xmath6 beyond @xmath7 caused the umbilicus of the pattern to move off center .
as observed by others , a further quasi - static increase of @xmath8 up to @xmath9 caused a sequence of transitions at @xmath10 , each associated with the loss of one convection roll at the umbilicus .
each loss of a roll was preceded by the displacement of the umbilicus away from the center of the cell . after each transition
the umbilicus moved back toward but never quite reached the center . with decreasing @xmath8 new rolls formed at the umbilicus when @xmath8 was reduced below @xmath11 .
when decreasing @xmath8 , large umbilicus displacements did not occur .
in addition to quantitative measurements of the umbilicus displacement , we determined and analyzed the entire wave - director field of each image .
the wave numbers varied in the axial direction , with minima at the umbilicus and at the cell wall and a maximum at a radial position close to @xmath12 .
the wave numbers at the maximum showed hysteretic jumps at @xmath13 and @xmath14 , but on average agreed well with the theoretical predictions for the wave numbers selected in the far field of an infinitely extended target pattern . to our knowledge there is as yet no prediction for the wave number selected by the umbilicus itself , or by the cell wall of the finite experimental system . |
since their appearance , neutrinos have always been extremely peculiar .
their charge neutrality , near masslessness , flavor mixing , oscillation , type of dirac or majorana , and in particular parity - violating gauge - coupling have been at the center of a conceptual elaboration and an intensive experimental analysis that have played a major role in donating to mankind the beauty of the standard model for particle physics .
several experiments studying solar , atmospheric and reactor neutrinos in past several years provide strong evidences supporting the existence of neutrino oscillations @xcite .
this implies that neutrinos are not exactly massless although they chirally couple to gauge bosons
. therefore they can not be exactly two - component weyl fermions .
instead , it is tempting to think of the nature of neutrinos , they might be fermions of dirac type or majorana type @xcite .
dirac and majorana neutrinos have different electromagnetic properties @xcite , which can be used to determine which type of neutrinos exists .
the well - known example is the neutrino - less double beta decay @xcite , still to be experimentally verified .
some attention has been recently driven to study the neutrino interactions with laser beams ; here we mention a few examples .
the emission of @xmath0 pairs off electrons in a polarized ultra - intense electromagnetic ( e.g. , laser ) wave field is analyzed in ref .
the electron - positron production rate has been calculated @xcite using neutrinos in an intense laser field . in the frame of the standard model , by studying the elastic scattering of a muon neutrino on an electron in the presence of a linearly polarized laser field , multi- photon processes have been shown @xcite . in this letter , we quantitatively show and discuss linearly polarized photons acquire circular polarization by interacting with neutrinos , for the reason that neutrinos are left - handed and possess peculiar couplings to gauge bosons in a parity - violating manner . in particular we
quantitatively calculate the circular polarization that a linearly polarized laser beam develops by interacting with a neutrino beam in ground laboratories .
moreover we show that using advanced laser facility one can possibly measure the circular polarization of laser photons interacting with a dirac or majorana neutrino beam produced by neutrino experiments , e.g. , the tokai - to - kamioka ( t2k ) @xcite .
this could possibly provide an additional way to study some physics of neutrinos , for example , the nature of neutrinos , dirac or majorana type .
we recall that the similar photon - neutrino process was considered for the generation of circular polarizations of cosmic microwave background ( cmb ) photons @xcite interacting with cosmic background neutrinos ( cnb ) , instead of photons @xcite and electrons in the presence of magnetic field @xcite . there
the main calculation was done to obtain the power spectrum @xmath1 of the circular polarization of cmb photons by the forward scattering between cmb photons and cnb neutrinos .
the polarization of laser beam is characterized by means of the stokes parameters : the total intensity @xmath2 , intensities of linear polarizations @xmath3 and @xmath4 , as well as the intensity of circular polarizations @xmath5 indicating the difference between left- and right- circular polarizations intensities . the linear polarization can be represented by @xmath6 .
an arbitrary polarized state of a photon @xmath7 , propagating in the @xmath8-direction , is given by @xmath9 and @xmath10 indicate the polarization states in the @xmath11- and @xmath12-directions , and @xmath13 are initial phases .
quantum - mechanical operators in this linear bases , corresponding to stokes parameters , are given by & = & |_1_1|+|_2_2|,=|_1_1|-|_2_2| , + & = & |_1_2|+|_2_1|,=i|_2_1|-i|_1_2| .
[ i - v ] an ensemble of photons in a general mixed state is described by a normalized density matrix @xmath14 , and the dimensionless expectational values for stokes parameters are given by ii & = & tri=1,[i ] + qq & = & tr=_11-_22,[q ] + u u&=&tr=_12+_21,[u ] + vv & = & tr = i_21-i_21 , [ v ] where `` @xmath15 '' indicates the trace in the space of polarization states . as shown below , in a quantum field theory , the density matrix describing stokes parameters of polarized particle is represented in the phase space @xmath16 in addition to the space of polarization states .
( [ i]-[v ] ) determine the relations between stokes parameters and the @xmath17 density matrix @xmath18 of photon polarization states .
we express the laser field strength @xmath19 , and free gauge field @xmath20 in terms of plane wave solutions in the coulomb gauge @xcite , a_(x ) = , where @xmath21 are the polarization four - vectors and the index @xmath22 , representing two transverse polarizations of a free photon with four - momentum @xmath23 and @xmath24 , @xmath25 and @xmath26 .
the creation operators @xmath27 and annihilation operators @xmath28 satisfy the canonical commutation relation @xmath29 = ( 2\pi ) ^3 2k^0\delta_{ij}\delta^{(3)}({\bf k } - { \bf k } ' ) .
\label{comm}\ ] ] the density operator describing an ensemble of free photons in the space of energy - momentum and polarization state is given by ( x)= _ ij(x , k)a^_i(k)a_j(k ) , [ d - m ] where @xmath30 is the general density - matrix , analogous to eqs .
( [ i]-[v ] ) , in the space of polarization states with the fixed energy - momentum `` @xmath23 '' and space - time point `` @xmath11 '' . the number operator of photons @xmath31 and its expectational value with respect to the density - matrix ( [ d - m ] ) is defined by d^0_ij(k)[d^0_ij(k)]=(2)^3 ^3(0)(2k^0)_ij(x , k).[t1 ] the time - evolution of the number operator @xmath32 is governed by the heisenberg equation @xmath33,\ ] ] where @xmath34 is the interacting hamiltonian of photons with other particles in the standard model . calculating the expectational value of both sides of eq .
( [ heisen ] ) , one arrives at the following quantum boltzmann equation ( qbe ) for the number operator of photons @xcite , ( 2)^3 ^3(0)(2k^0 ) _
ij(x , k ) = i - dt[bo ] , where we only consider the interacting hamiltonian @xmath35 of photons with neutrinos . the first term on the right - handed side is a forward scattering term , and the second one is a higher - order collision term , which will be neglected . in the following ,
we attempt to calculate the stokes parameter @xmath5 of eq .
( [ v ] ) to show that a linearly polarized laser beam acquires the component of circular polarization while it interacts with a neutrino beam .
first we consider dirac neutrinos @xmath36 interacting with charged leptons @xmath37 and @xmath38 gauge bosons in the standard model [ lagw ] _ int = _
l = e , , , where we omit the unitary mixing matrix which is not relevant to the following calculations .
because neutrino masses are approximately zero , compared with their energies ( @xmath39 ) , we adopt the two - component theory for zero - mass particles to describe neutrinos ( see for example ref .
@xcite ) _ ( x ) = _ r=,[psi ] where neutrino energy @xmath40 , and helicity or chirality states @xmath41 are eigenstates of helicity operator @xmath42 and chirality operator @xmath43 , ( @xmath44 ) representing the left - handed neutrino @xmath45 and right - handed anti - neutrino @xmath46 ; ( @xmath47 ) representing the right - handed neutrino @xmath48 and left - handed anti - neutrino @xmath49 .
the annihilation and creation operators @xmath50 ( @xmath51 ) and @xmath52 ( @xmath53 ) for corresponding states satisfy following relations , @xmath54 for a given momentum @xmath55 and chirality state ( @xmath44 ) , there exists only two independent states of left - handed neutrino and right - handed anti - neutrino .
the chiral projector @xmath56 in the interacting hamiltonian ( [ lagw ] ) implies that only left - handed neutrinos ( @xmath44 ) participate the weak interaction .
the relations @xmath57 are useful for the following calculations . and @xmath58 ( @xmath55 and @xmath59 ) represent photons ( neutrinos ) momenta . along the loop , there are two qed - vertexes and two vertexes of eq .
( [ lagw ] ) , as well as a @xmath60 gauge - boson propagator @xmath61 ( dashed line ) and three charged lepton propagators @xmath62 ( solid lines ) . note that the contribution from the neutral channel diagram with a @xmath63-boson exchange ( see ref .
@xcite ) is proportional to neutrino masses @xmath64 , we consider in this letter exactly massless neutrinos so that the neutral channel diagram with a @xmath63-boson exchange does not contribute to the interacting hamiltonian ( [ h0 ] ) for studying the generation of circular polarization of laser photons .
even if we consider massive neutrinos , the contribution from the neutral current , being proportional to neutrino masses , can be neglected in comparison with the contribution from charged current.,title="fig : " ] + in the context of standard model @xmath65 for the electro - weak interactions , two feynman diagrams representing the leading - order contribution to the interaction between photons and neutrinos are shown in fig . [ cmb1 ] .
the leading - order interacting hamiltonian is given by @xmath66\left[b^\dagger_{r'}a^{\dagger}_{s'}(\mathcal{m}_1+\mathcal{m}_2)a_sb_r\right],\label{h0}\end{aligned}\ ] ] where @xmath67 s_f(l)\gamma^\beta ( 1-\gamma_5)\mathcal{u}_r(q),\label{md}\end{aligned}\ ] ] @xmath61 and @xmath62 are respectively @xmath60 gauge - boson and charged - lepton propagators . in eq .
( [ h0 ] ) , our notation @xmath68 $ ] , the same for @xmath69 and @xmath70 .
from eq . ( [ h0 ] ) for the photon - neutrino interacting hamiltonian
, we obtain the first forward scattering term of eq .
( [ bo ] ) @xmath71 & = & \int d\mathbf{q } d\mathbf{q ' } d\mathbf{p } d\mathbf{p ' } ( 2\pi)^3\delta^3(\mathbf{q ' } + \mathbf{p ' } -\mathbf{p } -\mathbf{q } ) ( \mathcal{m}_1+\mathcal{m}_2)\nonumber\\ & \times & ( 2\pi)^3[b^\dagger_{r'}b_{r}a^\dagger_{s'}a_{s}2p^0\delta_{is}\delta^3({\bf k}-{\bf p})-b^\dagger_{r'}b_{r}a^\dagger_{s'}a_{s}2p'^0\delta_{js'}\delta^3({\bf k}-{\bf p'})].\label{fws}\end{aligned}\ ] ] then , using the commutators of eqs . ( [ comm ] ) and ( [ commf ] ) , as well as the following expectation values of operators @xcite , a_1a_2 ... b_1b_2 ... & = & a_1a_2 ...
b_1b_2 ... , [ contraction1 ] + a^_s(p)a_s(p ) & = & 2p^0(2)^3 ^ 3(-)_ss ( , ) , [ contraction2 ] + b^_r(q)b_r(q ) & = & ( 2)^3 ^ 3(-)_rrn _ ( , ) , [ contraction3 ] where @xmath72 is the local matrix density and @xmath73 is the local spatial density of neutrinos in the momentum state @xmath74 . thus we define the neutrino distribution function @xmath75 and the average momentum @xmath76 of neutrinos as @xmath77 for a neutrino beam we assume @xmath78 $ ] , representing the most of neutrinos carry the momentum @xmath79
. then we arrive at @xmath80\rangle & = & \frac{i}{16}e^2g_w^2\int d\mathbf{q } \big[\rho_{s'j}(\mathbf{x},{\bf k})\delta_{is } -\rho_{is}(\mathbf{x},{\bf k})\delta_{js'}\big]n_\nu(\mathbf{x},\mathbf{q})\nonumber\\ & \times & \int\frac{d^4l}{(2\pi)^4}d_{\alpha\beta}(q - l)\bar{\mathcal{u}}_{r}(q')\gamma^\alpha ( 1-\gamma_5)s_f(l)\nonumber\\ & \times & \left[\,\ep_{s'}s_f(l+k)\,\ep_{s}+\ep_{s}\,s_f(l - k)\ep_{s'}\,\,\right ] s_f(l)\gamma^\beta ( 1-\gamma_5)\mathcal{u}_r(q).\label{fws1}\end{aligned}\ ] ] moreover , using dimensional regularization and feynman parameterizations for four - momentum integration over @xmath81 , we approximately obtain @xmath80\rangle & \approx & -\frac{1}{16}\frac{1}{4\pi^2}e^2g_w^2\int d\mathbf{q } \big[\rho_{s'j}(\mathbf{x},{\bf k})\delta_{is } -\rho_{is}(\mathbf{x},{\bf k})\delta_{js'}\big]n_\nu(\mathbf{x},\mathbf{q})\nonumber\\ & \times & \int_0 ^ 1dy\int_0^{1-y}dz\frac{(1-y - z)}{zm^2_w}\bar{\mathcal{u}}_{r}(q)(1+\gamma_5)\big[2z q\!\!\!/\epsilon_{s'}\cdot\epsilon_s\nonumber\\ & + & \,2z(\ep_{s'}\,\,q\cdot\epsilon_s\,+\ep_{s}\,q\cdot\epsilon_{s'}\,)+ ( 3y-1)k\!\!\!/\,(\ep_{s}\,\ep_{s'}\,-\ep_{s'}\,\ep_{s}\ , )
\big]\,\mathcal{u}_r(q ) , \label{fws2}\end{aligned}\ ] ] where neutrino and photon energies are much smaller than @xmath38-boson mass @xmath82 , i.e. , @xmath83 .
the time evolution of stokes parameter @xmath5 of linearly polarized laser - photons is determined by eqs .
( [ bo ] ) and ( [ fws2 ] ) , @xmath84\mathcal{u}_r(q),\label{v1}\end{aligned}\ ] ] where high - order terms suppressed at least by @xmath85 have been neglected .
using eq .
( [ sn2 ] ) , we obtain @xmath86,\label{v2}\end{aligned}\ ] ] where the fermi and fine - structure constants are @xmath87 as follows , we will consider the neutrino beam is produced by high - energy muons ( @xmath88 ) decay , the beam angle divergence @xmath89 , where @xmath90 is the muon energy ( mass ) .
this implies that near to the source of high - energy muons ( @xmath88 ) decay , the momenta of neutrinos are around the averaged one @xmath76 , given by eq.([nd - n0 ] ) . in this circumstance , we approximate eq .
( [ v2 ] ) as @xmath91,\label{v2'}\end{aligned}\ ] ] where @xmath92 indicates the direction of neutrino beam , and @xmath93 along the beam direction . on the analogy of averaged momentum @xmath76 of neutrino beam
, @xmath94 should be understood as an averaged momentum of photons in a laser beam in the direction @xmath95 , the stokes parameters @xmath96 and @xmath97 are the functions averaged over the momentum distribution of photons in a laser pulse . as shown in fig .
[ axis ] , we select the @xmath98 along the @xmath8-direction , and @xmath99 and @xmath100 spherical angles of @xmath101 with respect to the @xmath98 direction
. then eq .
( [ v2 ] ) becomes @xmath102\nonumber\\ & = & \frac{\sqrt{2}}{6\pi k^0}\alpha\,g_f\,|\bar{\mathbf{q}}|\sin^2\theta\,n_\nu(\mathbf{x},\bar{\mathbf{q } } ) q(\mathbf{x},\bf k),\label{v21'}\end{aligned}\ ] ] where in the third line we set @xmath103 in the plane of @xmath76 and @xmath104 . to discuss possible experimental relevance ,
we rewrite eq .
( [ v21 ] ) as @xmath105 where the averaged energy - flux of neutrino beam @xmath106 eq . ( [ v22 ] ) represents the ratio of circular and linear polarizations of laser beam interacting with the neutrino beam for a time interval @xmath107 . in the @xmath108-direction ( the direction of incident laser beam ) ,
polarization vectors @xmath109 in the @xmath110-direction and @xmath111 in the @xmath112-direction .
this sketch shows the relative angles @xmath99 and @xmath100 between the laser beam direction @xmath98 and neutrino beam direction @xmath101 .
, width=384 ] to end this section , we would like to point out that the purely left - handed interaction ( [ lagw ] ) is the crucial reason why linearly polarized photons acquire circular polarizations by interacting with left - handed neutrinos , in contrast they do not acquire circular polarizations by interacting with electrons in the forward scattering terms of eq .
( [ bo ] ) @xcite .
if right - handed neutrinos were involved in the weak interaction , replacing @xmath56 by @xmath113 in eq .
( [ lagw ] ) , we would have obtained a contribution from right - handed neutrinos that completely cancels the circular polarization ( [ v21 ] ) .
this can be understood by the angular - momentum conservation in the photon - neutrino interaction ( [ lagw ] ) , where neutrinos are in left - handed state @xmath114 , as a result , photons acquire the component of circular polarizations .
this point will be further illustrated by the compton scattering of photons and polarized electrons in the next section .
it is shown @xcite that the photon circular - polarization is not generated by the forward compton scattering of linearly polarized photons on unpolarized electrons .
the reason is that the contributions from left- and right - handed polarized electrons to photon circular - polarization exactly cancel each other . in the following ,
we show the circular polarization generated by a polarized compton scattering : a photon beam scattering on a polarized electron beam , whose the number - density of incident left - handed electrons is not equal to the right - handed one . for the sake of simplicity , we describe the polarized electron beam by the net number - density of left - handed electrons @xmath115 , and consider relativistic electrons @xmath116 , @xmath117 and @xmath118 , where @xmath74 , @xmath119 and @xmath94 are the momenta of incident electrons , incident and scattered photons respectively ( see fig . [ compton ] ) . + following the calculations of ref .
@xcite , we obtain the time evaluation of the stokes parameter @xmath5 for the circular polarization of scattered photons , @xmath120 ^ 2 - 2[\epsilon_1(\hat{\mathbf{p}})\cdot\epsilon_2(\mathbf{k})][\bar{\mathbf{q}}\cdot\epsilon_2(\mathbf{k } ) ] [ \bar{\mathbf{q}}\cdot\epsilon_1(\hat{\mathbf{p}})]+[\bar{\mathbf{q}}\cdot\epsilon_1(\hat{\mathbf{p}})]^2 -(1\leftrightarrow 2 ) \big\}u(\hat{\mathbf{p}}),\label{compton1}\end{aligned}\ ] ] which is proportional to @xmath121 the averaged number - density of left - handed polarized electrons , here @xmath76 is the averaged momenta of incident polarized electrons and @xmath122 .
this shows that linearly polarized photons scattering on polarized electrons acquire circular polarization , analogously with photon circular - polarization generated by linearly polarized photons scattering on left - handed neutrinos .
we turn to calculate the circular polarization of laser - photon beam due to its scattering with majorana neutrinos @xcite , @xmath123,\label{major}\end{aligned}\ ] ] which are self - conjugated dirac neutrinos ( particle and anti - particle are identical ) , i.e. , @xmath124 up to a phase @xmath125 ( @xmath126 ) , where the conjugated field @xmath127 , @xmath128 , and @xmath129 the interaction of majorana neutrinos with charged particles via the @xmath130 in the standard model @xmath65 can be written as @xcite @xmath131,\label{lagwm}\end{aligned}\ ] ] where @xmath132 is the conjugated lepton field .
the first term is due to the left - handed majorana neutrino , analogously to eq .
( [ lagw ] ) for the left - handed dirac neutrino , yielding feynman diagrams in fig .
the second term comes from the conjugated left - handed majorana neutrino @xmath133 , yielding `` conjugated '' feynman diagrams , which are all the internal lines in fig .
[ cmb1 ] are replaced by their conjugated lines . the contribution from `` conjugated '' feynman diagrams
can be obtained by substitutions @xcite : @xmath134 as a result , in majorana neutrino case , the interacting hamiltonian is @xmath66\left[b^\dagger_{r'}a^{\dagger}_{s'}(\mathcal{m}_1+\mathcal{m}_2)a_sb_r+ a^{\dagger}_{s'}b_r(\mathcal{m}'_1+\mathcal{m}'_2)a_sb^{\dagger}_{r'}\right]\nonumber\\ & = & \int d\mathbf{q } d\mathbf{q ' } d\mathbf{p } d\mathbf{p ' } ( 2\pi)^3\delta^3(\mathbf{q ' } + \mathbf{p ' } -\mathbf{p } -\mathbf{q } ) \nonumber \\ & \times & \exp[it(q'^0+p'^0-q^0-p^0)]\left[b^\dagger_{r'}a^{\dagger}_{s'}(\mathcal{m}_1+\mathcal{m}_2-\mathcal{m}'_1-\mathcal{m}'_2)a_sb_r\right],\label{h0m}\end{aligned}\ ] ] where @xmath135 is given by eq .
( [ md ] ) and the `` conjugated '' part @xmath136 s_f(-l)\gamma^\beta ( 1+\gamma_5)\mathcal{v}_{r'}(q ' ) .
\label{conj}\end{aligned}\ ] ] analogously to eqs .
( [ fws]-[fws2 ] ) , we obtain the contribution of `` conjugated '' part ( [ conj ] ) to the forward scattering term @xmath137\rangle$ ] , @xmath138n^m_{\nu}({\bf x},{\bf q})\nonumber\\ & \times & \int_0 ^ 1dy\int_0^{1-y}dz\frac{(1-y - z)}{zm^2_w}\bar{\mathcal{v}}_{r}(q)(1-\gamma_5)\big[2z q\!\!\!/\epsilon_{s'}\cdot\epsilon_s\nonumber\\ & + & \,2z(\,\ep_{s'}\,q\cdot\epsilon_s\,+\ep_{s}\,q\cdot\epsilon_{s'}\,)- ( 3y-1)k\!\!\!/\,(\ep_{s}\,\ep_{s'}\,-\,\ep_{s'}\,\ep_{s}\,)\,\big ] \mathcal{v}_r(q).\label{fws2m}\end{aligned}\ ] ] as a result , we approximately obtain the time - evolution of the circular - polarization that laser - photons acquire after their interaction with a majorana neutrino beam , @xmath139\mathcal{v}_r(q).\label{v1m}\end{aligned}\ ] ] the first term is the same as eq .
( [ v2 ] ) with the replacement @xmath140 , and the rest comes from the so - called `` conjugated '' part , which is the contribution from conjugated left - handed majorana neutrinos interacting with @xmath141 in the second term of eq .
( [ lagwm ] ) . from the identity @xmath142 eq .
( [ v1 m ] ) becomes @xmath143\mathcal{u}_r(q).\label{v2m}\end{aligned}\ ] ] by using identities @xmath144 we approximately obtain @xmath145 indicating that the photon circular - polarization generated by the photon - neutrino ( majorana ) scattering is about two times larger than that generated by the photon - neutrino ( dirac ) scattering .
this difference is due to the fact that in the standard model , see eq .
( [ lagw ] ) , there is only one species of the left - handed dirac neutrino field , whereas the left - handed majorana neutrino field has two species for it being a self - conjugated field , see eq .
( [ lagwm ] ) . as a result , in the majorana neutrino case , the interacting hamiltonian @xmath146 of eq .
( [ h0 m ] ) has an additional `` conjugated '' part , compared with the interacting hamiltonian @xmath146 of eq .
( [ h0 ] ) in the dirac neutrino case .
these two parts in eq .
( [ h0 m ] ) have the same contribution to the @xmath5-parameter , which is equal to the contribution in the dirac neutrino case .
thus , the @xmath5-contribution ( [ fv2 m ] ) from the photon and majorana neutrino interaction is twice larger than the @xmath5-contribution ( [ v22 ] ) from photon and dirac neutrino interaction .
our result ( [ v22 ] ) shows that in order to produce a large intensity of photon circular - polarization for possible measurements , the intensities of neutrino and laser - photon beams should be large enough .
in addition the interacting time @xmath107 of two beams , i.e. , the spatial dimension @xmath147 of interacting spot of two beams ( @xmath148 ) should not be too small .
this depends on the sizes of two beams at interacting point , assuming @xmath149 .
based on the figs . 1 or 25 in ref . @xcite for the tokai - to - kamioka ( t2k ) neutrino experiment ( nd280 ) , the maximal neutrino flux @xmath150
locating at energies @xmath151 , we estimate the mean energy - flux of muon neutrino beam , @xmath152 the beam angle divergence @xmath153 , where the muon energy @xmath154 . in addition , in the neutrino experiment ( nd280 ) , we estimate the spatial and temporal intervals @xmath147 and @xmath107 of interacting spot as @xmath155 and @xmath156 sec . , where @xmath157 is the beam radius near to the source of muon neutrinos and @xmath158 cm is the distance between the muon neutrino source and interacting spot with laser photons .
using optic laser beams , laser photon energy @xmath159 , we approximately obtain the ratio of eq .
( [ v22 ] ) @xmath160 where the factor @xmath161 is the order of unit .
( [ v23 ] ) represents the result for laser and neutrino beams ( pulses ) interacting once only .
suppose that by using mirrors or some laser facilities the laser beam can bent @xmath162-times in its traveling path , and all laser pulses are identical without interference , so that each laser pulse can interact with the neutrino beam @xmath162-times in its path , the ratio of eq .
( [ v23 ] ) is approximately enhanced by a factor of @xmath162 , @xmath163 since @xmath164 for back and forth laser beams in opposite directions .
suppose that mirrors that trap laser beams can reflect @xmath165 of laser light for a narrow range of wavelengths and angles .
the absorption coefficient of the mirror is then @xmath166 , and the intensity @xmath3 of linearly polarized laser beam is reduced to @xmath167 for each reflection , we approximately have @xmath168 . in general , we express our result as @xmath169 based on the t2k neutrino experiment ( nd280 ) and discussions presented in this section , to have a significant probability of observing a circularly polarized photon ( @xmath170 ) , the intensity of incident linearly polarized photons of optic laser should at least be : @xmath171 , which is not difficult to be achieved in present laser technologies .
the larger intensity @xmath3 of incident laser beam in linear polarization is , the larger intensity of photons in circular polarizations should be observed .
in addition , a laser or neutrino beam is made by sequent bunches ( pulses ) of photons or neutrinos , it should be important to synchronize two beams in such a way that the interacting probability of photons and neutrinos bunches is maximized . for a comparison with eq .
( [ v25 ] ) , using eq .
( [ compton1 ] ) we estimate the quantity @xmath172 for a linearly polarized laser beam interacting with a polarized electron beam too . considering the left - handed polarized electron beam with
averaged density @xmath121 and averaged momentum @xmath173 , we obtain @xmath174 where @xmath175 is the averaged energy of laser photons , @xmath107 indicates time interaction , the averaged velocity and energy flux of the left - handed polarized electron beam @xmath176 and @xmath177 .
in addition , ref .
@xcite discussed the possibility of using intense muon neutrino beams , such as those available at proposed muon colliders , interacting with high powered lasers to probe the neutrino mass .
the rate of photon - neutrino scattering ( @xmath178 1/year or @xmath179 ) was estimated by considering dramatically short pulsed lasers with energies of up to @xmath180 per pulse and very short pulse durations ( @xmath181 ) , which is near to the critical intensity @xmath182 for the production of electron - positron pairs . even in this
extremal powered lasers , the total probability for photon - neutrino scattering is too small to be observed .
the main reason is that the interacting rate @xmath183 is proportional to the photon - neutrino scattering cross - section @xmath184 , which is very small . to compare with the photon - neutrino scattering rate @xmath183 obtained in ref .
@xcite , we estimate the rate of generating circular polarization of laser photons presented in this letter . the quantity @xmath185 of eq .
( [ v22 ] ) represents the number of circularly polarized photons of energy @xmath186 per unit area ( @xmath187 ) , unit time ( @xmath188 ) in a laser pulse .
therefore the rate of generating circular polarization of laser photons can be estimated by @xmath189 where @xmath190 is the time duration of a laser pulse , the effective area of photon - neutrino interaction is represented by the laser - beam size @xmath191 being smaller than the neutrino - beam size @xmath147 , and the laser repetition rate @xmath192 is the number of laser pulses per second . to have more efficiency , we assume that laser and neutrino beams are synchronized and the @xmath192 is equal to the repetition rate of neutrino beam @xmath193 , which is the number of neutrino bunches per second .
in contrast with the rate of photon - neutrino scattering @xmath194 , the rate @xmath195 of eq .
( [ rate ] ) linearly depends on @xmath196 via the @xmath185 of eq .
( [ v22 ] ) .
this implies that the rate @xmath195 of eq .
( [ rate ] ) should be much larger than the photon - neutrino scattering rate @xmath183 considered in ref . @xcite .
combining eqs .
( [ v25 ] ) and ( [ rate ] ) , we obtain @xmath197 where @xmath198 is the number of photons in a laser pulse @xmath199 and the total energy of a laser pulse @xmath200 . the averaged power of linearly polarized laser beam is approximately given by @xmath201 , then eq . ( [ rate2 ] ) becomes @xmath202 as a result , with a neutrino beam @xmath203 and a linearly polarized laser beam of energy @xmath204ev and power @xmath205mw , the rate of generating circularly polarized photons @xmath206 ( @xmath207 ) .
this rate should be large enough for observations .
we show and discuss the reason why a linearly polarized photon acquires its circular polarization by interacting with a neutrino is due to the fact that the neutrino is left - handed and possesses chiral gauge - coupling to gauge bosons . calculating the ratio of linear and circular polarizations of photons interacting with either dirac or majorana neutrinos
, we obtain that this ratio in the case of dirac neutrinos is about twice less than the ratio in the case of majorana neutrinos . based on quantitative value of this ratio
, we discuss the possibility of using advanced laser facilities and the t2k neutrino experiment to measure the circular polarization of laser beam interacting with neutrino beam in ground laboratories . in ref .
@xcite , the rate of neutrino - photon scattering was calculated by the feynman diagram of one fermion- loop with exchanging a @xmath63-boson ( neutral current channel ) .
this rate is very small because the obtained cross - section of photon - neutrino is the order of @xmath208 , and it was concluded that the total probability for photon - neutrino scattering is too small to be observed ( having an event rate below 1/year ) , even at pw laser powers . in our case , the time evaluation of the @xmath5-parameter ( [ bo ] ) for the photon circular polarization can be simply written as @xmath209+\int dt [ h^0_i,[h^0_i , v]],\label{v5}\end{aligned}\ ] ] where the leading non - trivial contribution comes from the forward scattering amplitude @xmath210 $ ] , which is the order of @xmath211 , instead of @xmath212 . whereas the second term @xmath213 $ ] corresponds to the high - order contributions .
this implies that our results ( [ fv2 m ] ) and ( [ v25 ] ) of generating circular polarization of laser beam by interacting with neutrino beam might be experimentally measurable .
this could be an additional and useful way to gain some insight into the physics of neutrinos , for instance their dirac or majorana nature , moreover their mixing with the right - handed sterile neutrino @xcite .
we thanks the anonymous referee for his / her comments and suggestions . m. giovannini and k. e. kunze , phys . rev . * d78 * , 023010 ( 2008 ) , [ arxiv : astro - ph/0804.3380 ] ; + m. giovannini , ( 2002 ) , [ arxiv : hep - ph/0208152 ] ; + m. zarei , e. bavarsad , m. haghighat , r. mohammadi , i. motie , z. rezaei , phys .
rev . * d81 * , 084035 ( 2010 ) , [ arxiv : hep - th/0912.2993 ] . | it is shown that for the reason of neutrinos being left - handed and their gauge - couplings being parity - violated , linearly polarized photons acquire their circular polarization by interacting with neutrinos . calculating the ratio of linear and circular polarizations of laser photons interacting with either dirac or majorana neutrino beam
, we obtain this ratio for the dirac neutrino case , which is about twice less than the ratio for the majorana neutrino case . based on this ratio
, we discuss the possibility of using advanced laser facilities and the t2k neutrino experiment to measure the circular polarization of laser beams interacting with neutrino beams in ground laboratories .
this could be an additional and useful way to gain some insight into the physics of neutrinos , for instance their dirac or majorana nature . |
after the first discovery of cosmic acceleration from the distance measurements of the supernovae type ia ( sn ia ) @xcite , the existence of dark energy has been also supported from other observational data such as cosmic microwave background ( cmb ) @xcite and baryon acoustic oscillations ( bao ) @xcite . from the theoretical point of view , such a late - time cosmic acceleration is problematic because of the huge difference between the observed dark energy scale and the expected value of the vacuum energy appearing in particle physics @xcite . along with the cosmological constant @xmath9 ,
many alternative acceleration mechanisms have been proposed , including modifications of the matter / energy content and large - scale modifications of gravity ( see refs .
@xcite for reviews ) .
the dark energy equation of state @xmath8 is constrained by measuring the expansion rate of the universe from the observations of sn ia , cmb , and bao @xcite .
although it is possible to rule out some accelerating scenarios from the analysis of the cosmic expansion history alone , we require further precise observational data to clearly distinguish between models with subtly - varying @xmath8 .
so far the @xmath9-cold - dark - matter ( @xmath9cdm ) model has been consistent with the data , but there are many other dynamical models such as quintessence @xcite , k - essence @xcite , and @xmath10 gravity @xcite which are compatible with current observations .
the large - scale structure of the galaxy distribution provides additional important information to discriminate between different dark energy models @xcite .
the galaxy clustering occurs due to the gravitational instability of primordial matter density perturbations .
the growth rate of matter perturbations can be measured from the redshift - space distortion ( rsd ) of clustering pattern of galaxies .
this distortion is caused by the peculiar velocity of inward collapse motion of the large - scale structure , which is directly related to the growth rate of the matter density contrast @xmath0 @xcite .
hence the rsd measurement is very useful to constrain the cosmic growth history .
recent galaxy redshift surveys have provided bounds on the growth rate @xmath11 or @xmath12 in terms of the redshift @xmath13 , where @xmath3 and @xmath5 is the rms amplitude of @xmath0 at the comoving scale @xmath14 mpc ( @xmath15 is the normalized hubble constant @xmath16kmsec@xmath17mpc@xmath17 ) @xcite .
although the observational error bars of @xmath18 are not yet small , the data are consistent with the prediction of the @xmath9cdm model @xcite .
recently the rsd data were used to place constraints on modified gravity models such as @xmath10 gravity and ( extended ) galileons @xcite .
since the growth rate of matter perturbations in these models is different from that in the @xmath9cdm @xcite , the allowed parameter space is quite tightly constrained even from current observations . for the models based on general relativity ( gr ) without a direct coupling between dark energy and non - relativistic matter ,
the gravitational coupling appearing in the matter perturbation equation is equivalent to the newton s gravitational constant , as long as the dark energy perturbation is negligible relative to the matter perturbation .
nonetheless the evolution of perturbations depends on the background cosmology , so that the dynamical dark energy models with @xmath8 different from @xmath19 can be distinguished from the @xmath9cdm .
in particular , the future rsd observations may reach the level of discriminating between different dark energy models constructed in the framework of gr .
in this paper we derive an analytic formula of @xmath1 valid for dynamical dark energy models including imperfect fluids , quintessence , and k - essence .
provided that the sound speed @xmath20 is not much smaller than @xmath21 and that the variation of @xmath8 is not significant , our formula can reproduce the full numerical solutions with high accuracy .
the derivation of @xmath1 is based on the expansion of @xmath8 in terms of the dark energy density parameter @xmath22 , i.e. , @xmath23 . since @xmath1 is expressed in terms of the present values of @xmath5 and @xmath22 as well as @xmath24 ( @xmath25 ) , our formula is convenient to test for dynamical dark energy models with the observational data of the cosmic growth rate . for the models with constant @xmath8
there are 3 parameters @xmath26 , @xmath27 , and @xmath28 in the analytic expression of @xmath1 . in tracking quintessence models @xcite , where the dark energy equation of state is nearly constant during the matter era ( @xmath29 ) ,
we show that our formula of @xmath1 also contains only 3 parameters : @xmath30 , @xmath27 , and @xmath28 .
using the recent rsd data , we carry out the likelihood analysis by varying these 3 parameters to find observational bounds on @xmath26 and @xmath30 .
this paper is organized as follows . in sec .
[ cosmodsec ] we review cosmological perturbation theory in general dark energy models including imperfect fluids , quintessence , and k - essence . in sec .
[ dynamicalsec ] dynamical dark energy models are classified depending on how the equation of state @xmath8 is expanded in terms of @xmath22 . in sec .
[ analyticsec ] we derive an analytic formula for @xmath1 and in sec .
[ validitysec ] we confirm the validity of this formula in concrete examples of dark energy models . in sec .
[ obserconsec ] we perform the likelihood analysis to test for constant @xmath8 and tracking quintessence models with the recent rsd data .
[ consec ] is devoted to our main conclusions .
as the dark energy component we consider a fluid characterized by the equation of state @xmath31 , where @xmath32 is the pressure and @xmath33 is the energy density .
we also take into account non - relativistic matter ( cold dark matter and baryons ) with the energy density @xmath34 and treat it as a perfect fluid with the equation of state @xmath35 .
we deal with such a two - fluid system in the framework of gr under the assumption that dark energy is uncoupled to non - relativistic matter .
our analysis covers quintessence @xcite and k - essence @xcite models , in which the lagrangian @xmath36 of dark energy depends on a scalar field @xmath37 and a kinetic term @xmath38 . in these models we have
that @xmath39 and @xmath40 , where @xmath41 . in the flat friedmann - lematre - robertson - walker ( flrw ) background with the scale factor @xmath42 , where @xmath43 is the cosmic time , dark energy and non - relativistic matter obey , respectively , the following continuity equations & & _ x+3(1+w)_x=0 , + & & _ m+3 _ m=0 , [ rhophieq ] where a prime represents a derivative with respect to @xmath44 .
we introduce the density parameters @xmath45 and @xmath46 , where @xmath47 is the gravitational constant and @xmath48 is the hubble parameter ( a dot represents derivative with respect to @xmath43 ) . from the einstein equations it follows that & & _ x+_m=1 , [ fri1 ] + & & = -32 ( 1+w _ x ) . [ fri2 ]
the dark energy density parameter satisfies the differential equation _ x=-3w _ x(1-_x ) .
[ omexeq ] let us consider scalar metric perturbations about the flat flrw background .
we neglect the contribution of tensor and vector perturbations . in the absence of the anisotropic stress the perturbed line element in the longitudinal gauge
is given by @xcite ds^2=-(1 - 2 ) dt^2+a^2(t)(1 + 2 ) d x^2 , [ permet ] where @xmath49 is the gravitational potential .
we decompose the energy densities @xmath50 ( where @xmath51 ) and the pressure @xmath32 into the background and inhomogeneous parts , as @xmath52 and @xmath53 .
we also define the following quantities _
i , _ i ( i = x , m ) , where @xmath54 and @xmath55 are the rotational - free velocity potentials of dark energy and non - relativistic matter , respectively . in fourier space
dark energy perturbations obey the following equations of motion @xcite & & _ x+3(c_x^2-w)_x = -(1+w ) ( 3+_x ) , [ de1 ] + & & _ x+ ( 2 - 3w++ ) _
x= ( ) ^2 ( _ x- ) , [ de2 ] where @xmath56 and @xmath57 is a comoving wave number .
the perturbed equations of non - relativistic matter ( perfect fluids ) are & & _
m=-3 -_m , [ dm1 ] + & & _ m+ ( 2 + ) _ m= - ( ) ^2 .
[ dm2 ] from the einstein equations we obtain & & ( ) ^2 = 32 ( _ m _ m+ _ x _ x ) , [ ein1 ] + & & + = -32 ( ) ^2 , [ ein2 ] where @xmath58 and @xmath59 are the rest frame gauge - invariant density perturbations defined by _ m=_m+3 ( )
^2 _ m,_x=_x+3 ( ) ^2 ( 1+w ) _
x. [ hatqu ] for imperfect fluids such as quintessence and k - essence there exist non - adiabatic entropy perturbations generated from dissipative processes .
we introduce a gauge - invariant entropy perturbation @xmath60 of dark energy , as @xcite s_x=(c_x^2-c_a^2 ) _
x= ( - ) , where @xmath61 is the adiabatic sound speed defined by c_a^2= = w- . in the rest frame where the entropy perturbation is given by @xmath62 ,
the sound speed squared @xmath63 is gauge - invariant . using the definition of @xmath59 in eq .
( [ hatqu ] ) , the pressure perturbation of dark energy can be expressed as p_x = c_s^2_x+3 ( ) ^2 ( 1+w ) ( c_s^2-c_a^2 ) _ x _ x , [ delp ] whereas the sound speed squared @xmath64 in the random frame is related to @xmath65 via c_x^2=c_s^2 + 3 ( ) ^2 ( 1+w ) ( c_s^2-c_a^2 ) .
[ delp2 ] in terms of @xmath65 , eqs .
( [ de1 ] ) and ( [ de2 ] ) can be written as & & _
x+3(c_s^2-w)_x = -3(1+w ) -(1+w ) _ x , [ de3 ] + & & _ x+ ( 2 + -3c_s^2 ) _
x= ( ) ^2 ( _ x- ) .
[ de4 ] in k - essence characterized by the lagrangian density @xmath66 , the perturbed quantities can be expressed as _ x & = & ( p_,x+2xp_,xx ) x- ( p_,-2x p_,x ) , [ delrho ] + p_x & = & p_,x x+p _ , , [ delp3 ] + _ x & = & , where @xmath67 is the field perturbation and @xmath68 . from eq .
( [ delp ] ) the rest frame sound speed can be obtained by setting @xmath69 in eqs .
( [ delrho ] ) and ( [ delp3 ] ) , i.e. , @xmath70 @xcite . in quintessence characterized by the lagrangian @xmath71 ,
the sound speed squared reduces to @xmath72 .
from eqs .
( [ dm1 ] ) and ( [ dm2 ] ) it follows that @xmath73\ , .
\label{delmeq0}\ ] ] for the perturbations deep inside the hubble radius ( @xmath74 ) relevant to large - scale structures , the r.h.s . of eq .
( [ delmeq0 ] ) can be neglected relative to the l.h.s . of it , in addition to the fact that @xmath75 . if the contribution of dark energy perturbations can be neglected relative to that of matter perturbations in eq .
( [ ein1 ] ) , i.e. @xmath76 , eq .
( [ delmeq0 ] ) reads _
m+12 ( 1 - 3w _ x ) _ m -32 _ m _ m 0 , [ delmeq ] where we made use of eq .
( [ fri2 ] ) . during the deep matter era in which @xmath77 , there is a growing - mode solution @xmath78 to eq .
( [ delmeq ] ) . in this regime eq .
( [ ein1 ] ) tells us that @xmath79 and hence @xmath80 from eq .
( [ dm1 ] ) . for the dark energy density contrast @xmath81 ,
it is natural to choose the adiabatic initial condition @xcite _ x=(1+w ) _ m. [ adi ] the initial condition of @xmath82 is known by substituting eq .
( [ adi ] ) and @xmath83 into eq .
( [ de3 ] ) .
we will discuss the accuracy of the approximate equation ( [ delmeq ] ) in sec .
[ dynamicalsec ] .
.data of @xmath84 versus the redshift @xmath2 measured from the rsd . [ cols="^,^,^,^",options="header " , ] the growth rate of matter perturbations can be measured from the rsd in clustering pattern of galaxies because of radial peculiar velocities .
the perturbation @xmath85 of galaxies is related to the matter perturbation @xmath0 , as @xmath86 , where @xmath87 is a bias factor .
the galaxy power spectrum @xmath88 in the redshift space can be expressed as @xcite @xmath89 where @xmath90 is the cosine of the angle of the @xmath91 vector to the line of sight ( vector @xmath92 ) , @xmath93 and @xmath94 are the real space power spectra of galaxies and @xmath95 , respectively , and @xmath96 is the cross power spectrum of galaxy-@xmath95 fluctuations in real space . in eq .
( [ dm1 ] ) the variation of the gravitational potential is neglected relative to the growth rate of @xmath0 , so that @xmath97 under the continuity equation ( [ continuity ] ) , @xmath98 , @xmath99 , and @xmath100 in eq .
( [ pred ] ) depend on @xmath101 , @xmath102 , and @xmath103 , respectively .
we normalize the amplitude of @xmath0 at the scale @xmath104 mpc , for which we write @xmath5 . assuming that the growth of perturbations is scale - independent , the constraints on @xmath105 and @xmath106 translate into those on @xmath107 and @xmath18 .
the advantage of using @xmath18 is that the growth rate is directly known without the bias factor @xmath87 . in table
[ table : fs8constraints ] we show the current data of @xmath84 as a function of @xmath2 from the rsd measurements .
in this section , we discuss a number of dynamical dark energy models in which the field equations presented in the previous section can be applied . for imperfect fluids the rest frame sound speed @xmath20 is generally different from the adiabatic sound speed @xmath61 . for constant @xmath8
one has @xmath108 .
if @xmath65 is constant as well , the evolution of dark energy perturbations is known by solving eqs .
( [ de3 ] ) and ( [ de4 ] ) together with the background equations ( [ fri1])-([omexeq ] ) .
this is the approach taken in ref .
note also that for @xmath20 of the order of 1 the contribution of dark energy perturbations to @xmath49 in eq .
( [ ein1 ] ) is negligibly small relative to matter perturbations @xcite .
evolution of @xmath84 versus the redshift @xmath2 for @xmath72 , @xmath109 mpc@xmath17 , @xmath110 , and @xmath111 . the cases ( a ) and ( b ) correspond to @xmath112 and @xmath113 , respectively .
the solid lines show the numerically integrated solutions , whereas the bald dashed lines correspond to the solutions derived by using the approximate equation ( [ delmeq]).,width=326,height=307 ] in fig .
[ dperfig ] we plot the evolution of @xmath18 for @xmath72 and @xmath109 mpc@xmath17 with two different values of @xmath8 . the approximate equation ( [ delmeq ] )
reproduces the full numerical solution within the 0.1 % accuracy .
this means that , for @xmath72 , the contribution of dark energy perturbations to the gravitational potential is suppressed relative to that of matter perturbations . for @xmath114 and @xmath112
we find that the difference of @xmath115 between the numerical result and the approximated solution is about 4% for the mode @xmath109 mpc@xmath17 . for @xmath8 larger than @xmath116
the difference gets even larger , but such values of @xmath8 are not allowed observationally @xcite . in ref .
@xcite it was found that galaxy redshift and tomographic redshift surveys can constrain the sound speed only if @xmath65 is sufficiently small , of the order of @xmath117 .
quintessence @xcite is characterized by the lagrangian @xmath118 , where @xmath119 is the field potential . in this case
the sound speed @xmath20 is equivalent to 1 , so that the contribution of dark energy perturbations to the gravitational potential is negligible .
there are several classes of potentials which give rise to different evolution of @xmath8 .
the first class is the model with constant @xmath8 , which can be realized by the following potential @xcite v ( ) = ^-2 ( @xmath120 and @xmath121 are the today s values of @xmath37 and @xmath122 respectively , and _ 1= .
this case is identified as an imperfect fluid with @xmath72 and @xmath123constant discussed in sec .
[ impersec ] .
the second class consists of freezing models @xcite , in which the evolution of the field gradually slows down at late times .
a typical example is the inverse power - law potential @xcite v()=m^4+p ^-p , [ inversepo ] where @xmath124 is a positive constant .
for this potential there exists a so - called tracker solution @xcite with a nearly constant field equation of state @xmath125 during the matter era , which is followed by the decrease of @xmath8 . considering a homogeneous perturbation @xmath126 around @xmath30 , the field equation of state
is expressed as @xcite w = w_(0)+ _ n=1^ ( ) ^n .
[ wtrack ] expansion of @xmath8 around @xmath127 reads w & = & w_(0)+ _ x+
_ x^2 + & & + _ x^3 + o ( _ x^4 ) , [ wtrack2 ] which varies with the growth of @xmath127 .
the third class consists of thawing models @xcite , in which the field is nearly frozen by a hubble friction during the early cosmological epoch and it starts to evolve once the field mass @xmath128 drops below @xmath129 . in this case
@xmath8 is initially close to @xmath19 and then @xmath8 starts to grow at late times .
the representative potential of this class is that of the pseudo nambu - goldstone boson @xcite , i.e. , v()=^4 , where @xmath9 and @xmath130 are constants . assuming that the evolution of the scale factor can be approximated as that of the @xmath9cdm model
, the field equation of state is estimated as @xcite ( see also ref .
@xcite ) w = -1+(1+_0 ) a^3(k-1 ) ^2 , [ wtha ] where @xmath131 and @xmath132 are the today s values of @xmath8 and @xmath127 respectively , and k=,f(a ) = .
the constant @xmath133 is related to the mass of the field at the initial displacement , @xmath134 .
expansion around @xmath135 gives w=-1+w_1 _ x + 15 ( k^2 + 1)w_1 _
x^2 + ( 3k^4 + 31k^2 + 3)w_1 _ x^3 + o(_x^4 ) , [ wthaw2 ] where w_1= ( ) ^k-1 k^2 ( k-1)^2 ( k+1)^2 .
the growth of @xmath127 leads to the deviation from @xmath136 .
the equation of state for k - essence with the lagrangian density @xmath137 is given by @xmath138 .
this shows that cosmic acceleration with @xmath139 can be realized either for ( a ) @xmath140 or ( b ) @xmath141 . in the case ( a ) chiba _
@xcite showed that , for the factorized function @xmath142 , the field equation of state is given by the same form as eq .
( [ wtha ] ) with the replacement @xmath143 , where @xmath144 is expanded around @xmath145 as @xmath146 . in this case
the sound speed squared is also close to 1 , so that the situation is similar to that in thawing quintessence models . in the case ( b ) the evolution of @xmath8
depends on the functional form of @xmath66 , so it is difficult to derive a general expression of @xmath8 @xcite .
one of the examples which belongs to this class is the dilatonic ghost condensate model @xcite characterized by the lagrangian @xmath147 , where @xmath148 , @xmath149 and @xmath150 are constants ( see also ref .
@xcite ) . in this model
the fixed points during the radiation and matter eras correspond to @xmath151 and @xmath136 , i.e. , @xmath152 .
on the other hand the ( no - ghost ) accelerated fixed point corresponds to @xmath153 with the equation of state @xmath154/[1 + 3\lambda^2 f(\lambda)/8]>-1 $ ] , where @xmath155 .
hence the evolution of @xmath8 is similar to that in thawing quintessence models .
the sound speed squared in the dilatonic ghost condensate model is given by @xmath156 , so that @xmath157 during radiation and matter eras .
the late - time cosmic acceleration occurs for @xmath158 and hence @xmath159 at this fixed point .
the fact that @xmath65 is close to 0 during most of the cosmological epoch is a different signature relative to quintessence .
however , since @xmath8 is very close to @xmath19 during radiation and matter eras , the adiabatic initial condition ( [ adi ] ) shows that the dark energy perturbation @xmath81 is initially suppressed relative to the matter perturbation @xmath0 . as long as the today s value of @xmath8 is not significantly away from @xmath19 the contribution of the dark energy perturbation to @xmath49 is suppressed relative to the matter perturbation , so that the approximate equation ( [ delmeq ] ) can be trustable even in such cases .
we derive analytic solutions of @xmath18 by solving the approximate equation ( [ delmeq ] ) . recall that the equation of state for tracking and thawing models of quintessence can be expressed in terms of the field density parameter @xmath22 , see eqs .
( [ wtrack2 ] ) and ( [ wthaw2 ] ) .
we generally expand the dark energy equation of state in terms of the density parameter @xmath127 , as w = w_0+_n=1^ w_n ( _ x)^n .
[ wexpand ] since @xmath127 grows as large as 0.7 today , we expect that it may be necessary to pick up the terms higher than the first few terms in eq .
( [ wexpand ] ) . in terms of the function @xmath160 , eq .
( [ delmeq ] ) can be written as @xcite @xmath161 where we employed eq .
( [ omexeq ] ) . introducing the growth index @xmath162 as @xmath163 @xcite , eq .
( [ fome ] ) reads @xmath164 we derive the solution of eq .
( [ gamome ] ) by expanding @xmath162 in terms of @xmath127 , i.e. , @xmath165 . while @xmath122 is smaller than @xmath22 today , the former is not suitable as an expansion parameter as we would like to derive an analytic formula valid at high redshifts as well .
in fact , it is expected that future rsd surveys such as subaru / fmos will provide high - redshift data up to @xmath166 . using the expansion of @xmath8 in eq .
( [ wexpand ] ) as well , we obtain the following approximate solution @xmath167/ [ 4(5 - 6w_0)^3 ( 5 - 12w_0)(5 - 18w_0)]\,\omega_{x}^2 + { \cal o } ( \omega_{x}^3)\ , .
\label{gammaap}\end{aligned}\ ] ] the 1-st order solution is identical to the one found in ref .
@xcite by setting @xmath168 . for @xmath169 , @xmath168 , and @xmath170 it follows that @xmath171 . in this case
the second and third terms are indeed much smaller than the first one , so that @xmath162 is nearly constant . for the models with @xmath169 , @xmath172 , and @xmath170
( in which case the value of @xmath8 today is around @xmath116 ) one has @xmath173 . even in this case
the variation of @xmath162 induced by the second and third terms is small ( see also refs .
@xcite for related works ) . from the definition of @xmath130 the matter perturbation obeys the differential equation @xmath174 .
using eq .
( [ omexeq ] ) , we obtain @xmath175 in the following we derive the solution of this equation under the approximation that @xmath162 is constant .
we expand the term @xmath176 around @xmath135 , as @xmath177 in order to evaluate the r.h.s . of eq .
( [ delmeqa ] ) , we expand @xmath178 in the form @xmath179\,,\ ] ] where the coefficients @xmath180 s can be expressed by @xmath24 s , say @xmath181 .
then eq .
( [ delmeqa ] ) can be written as @xmath182\ , , \label{delmeqa2}\ ] ] where @xmath183 with @xmath184 .
the first three coefficients @xmath185 are @xmath186 integrating eq .
( [ delmeqa2 ] ) , it follows that @xmath187 \right\},\ ] ] where @xmath188 is the today s value of @xmath0 .
normalizing @xmath188 in terms of @xmath189 , we obtain the following expression @xmath190 \right\}\ , . \label{fsigma8}\ ] ] in terms of the redshift @xmath2 the energy densities of non - relativistic matter and dark energy are given , respectively , by @xmath191 and @xmath192 $ ] . the 0-th order solution to the field energy density
is obtained by substituting @xmath193 into the expression of @xmath194 , i.e. , @xmath195 .
this gives the 0-th order solution to @xmath127 , as @xmath196 $ ] .
if we expand @xmath8 up to first order with respect to @xmath127 , we can use the iterative solution @xmath197 .
this process leads to the following integrated solution of @xmath127 : @xmath198^{w_1/w_0 } } { 1-\omega_{x 0}+\omega_{x 0 } ( 1+z)^{3w_0 } [ 1-\omega_{x 0}+\omega_{x 0 } ( 1+z)^{3w_0}]^{w_1/w_0}}\ , .
\label{omegax1}\ ] ] in the presence of the terms higher than second order , we can simply carry out the similar iterative processes .
practically it is sufficient to use the 1-st order solution ( [ omegax1 ] ) for the evaluation of @xmath22 in eq .
( [ fsigma8 ] ) . the growth factor @xmath162 in eq .
( [ fsigma8 ] ) is given by the analytic formula ( [ gammaap ] ) .
since @xmath162 is expressed in terms of @xmath24 ( @xmath199 ) and @xmath22 , this means that @xmath200 depends on the free parameters @xmath201 , @xmath202 , and @xmath189 .
for the models of constant equation of state there are 3 free parameters @xmath26 , @xmath202 , @xmath189 in the expression of @xmath200 . in tracking quintessence models the coefficients @xmath24 s ( @xmath203 )
are expressed in terms of @xmath204 , see eq .
( [ wtrack2 ] ) .
hence there are only 3 free parameters @xmath26 , @xmath202 , and @xmath27 . in thawing models of quintessence
one has @xmath169 and @xmath24 s ( @xmath205 ) can be expressed in terms of @xmath206 and @xmath133 , see eqs .
( [ wthaw2 ] ) .
then @xmath84 in eq . ( [ fsigma8 ] )
is written as a function of @xmath2 with 4 free parameters : @xmath206 , @xmath202 , @xmath27 , and @xmath133 .
we study the validity of the analytic estimation given in the previous section . we discuss three different cases : ( i ) constant @xmath8 models , ( ii ) tracking models , and ( iii ) thawing models , separately . in all the numerical simulations in this section
, we identify the present epoch to be @xmath207 with @xmath110 .
evolution of @xmath84 versus @xmath2 for the constant @xmath8 model with @xmath208 and @xmath72 .
the present epoch is identified to be @xmath209 with @xmath110 .
the solid line shows the numerically integrated solution , whereas the dot - dashed , dotted , and bold dashed lines correspond to the solutions summed up to 1-st , 3-rd , and 7-th order terms of @xmath210 in eq .
( [ fsigma8]).,width=316,height=297 ] evolution of @xmath84 versus @xmath2 for the models with @xmath72 and ( a ) @xmath211 , ( b ) @xmath136 , ( c ) @xmath112 , ( d ) @xmath212 , ( e ) @xmath213 , respectively .
the solid lines correspond to the numerically integrated solutions , whereas the bald dashed lines are derived from the analytic estimation ( [ fsigma8 ] ) with the 10-th order terms of @xmath210 .
we also show the current rsd data.,width=326,height=307 ] let us first study constant @xmath8 models realized by either imperfect fluids or quintessence . unless @xmath214 the dark energy perturbation is negligible relative to the matter perturbation , so that the approximate equation ( [ delmeq ] ) is sufficiently accurate .
since @xmath215 ( @xmath203 ) , the coefficients @xmath210 and @xmath127 are given , respectively , by @xmath216 for the growth index ( [ gammaap ] ) we take into account the terms up to 2-nd order with respect to @xmath127 , i.e. , @xmath217 recall that the terms higher than 2-nd order in @xmath162 are negligibly small .
then @xmath84 is analytically known from eq .
( [ fsigma8 ] ) just by specifying the values of @xmath189 , @xmath132 , and @xmath193 . in fig .
[ worder ] we plot the evolution of @xmath84 obtained by the analytic estimation ( [ fsigma8 ] ) for @xmath208 and @xmath72 .
a number of different lines correspond to the solutions derived by taking into account the @xmath210 terms up to 1-st , 3-rd , and 7-th orders . as we pick up higher - order coefficients @xmath210 in eq .
( [ cnconst ] ) , the solutions tend to approach the numerically integrated solution of @xmath18 . in fig .
[ worder ] we find that the solution up to 7-th order terms of @xmath210 can reproduce the full numerical result in good precision . in fig .
[ fsigmafig ] we show @xmath84 versus @xmath2 for five different values of @xmath8 . in order to obtain a good convergence
we pick up the @xmath210 terms up to 10-th order .
figure [ fsigmafig ] shows that our analytic estimation ( [ fsigma8 ] ) is sufficiently trustable to reproduce the numerically integrated solutions accurately .
if we only pick up the terms inside @xmath162 up to 1-st order with respect to @xmath22 , there is small difference of @xmath84 between the analytic estimation and the numerical solutions for @xmath218 ( which occurs in the low redshift regime ) .
taking into account the 2-nd order term in eq .
( [ gammaap ] ) , this difference gets smaller . fig .
[ fsigmafig ] also displays the rsd data given in table [ table : fs8constraints ] , which will be used to place observational constraints on @xmath8 in sec .
[ obserconsec ] . from eq .
( [ wtrack2 ] ) , we see that all coefficients @xmath24 s ( @xmath203 ) in tracking models of quintessence are expressed in terms of the tracker equation of state @xmath204 . in this case
there are contributions to @xmath210 coming from the variation of @xmath8 , i.e. , non - zero values of @xmath180 .
note that @xmath180 s depend only on @xmath30 .
evolution of @xmath84 versus @xmath2 for the tracking quintessence model characterized by the inverse power - law potential @xmath219 . from top
to bottom the solid lines correspond to the numerically integrated solutions for @xmath220 , respectively , whereas the bold dashed lines are derived from the analytic solution ( [ fsigma8 ] ) with the 10-th order terms .
the rsd data are the same as those given in fig .
[ fsigmafig].,width=326,height=307 ] for the tracking quintessence with the inverse power - law potential @xmath219 ( @xmath221 ) , we compare the numerically integrated solutions of @xmath18 with those derived by the analytic expression ( [ fsigma8 ] ) . in fig .
[ tracker ] we show the evolution of @xmath84 for @xmath220 evaluated from ( [ fsigma8 ] ) as well as the numerical solutions . in eq .
( [ fsigma8 ] ) we take into account the @xmath210 s up to 10-th order terms , whereas in the analytic expression of @xmath162 in eq . ( [ gammaap ] ) the terms up to 2-nd order in @xmath127 are included .
for the evaluation of @xmath127 the 1-st order solution ( [ omegax1 ] ) with @xmath207 is used . from fig .
[ tracker ] we find that the analytic solution ( [ fsigma8 ] ) is accurate enough to reproduce the full numerical solution in high precisions .
if we take into account the @xmath210 s up to the 3-rd order terms , for example , there is some difference between the analytic and numerical results .
this difference tends to disappear by including the higher - order terms of @xmath210 . while the terms up to 10-th order are taken into account in fig .
[ tracker ] , the 7-th order solutions are sufficiently accurate .
while our analytic formula of @xmath18 is trustable , readers may think that 7-th order expansion of @xmath210 is not very convenient for practical purpose .
however , using this analytic formula is simpler than solving the perturbation equations numerically for arbitrary initial conditions .
if we take the latter approach , we need first to identify the present epoch ( say , @xmath222 ) by solving the background equations from some redshift ( @xmath223 ) .
then the perturbation equations are solved with arbitrary initial values of @xmath224 to find @xmath225 for each @xmath2 . on the other hand , with our analytic formula , the likelihood analysis in terms of 3 free parameters @xmath26 , @xmath27 , and @xmath202 can be done much easier even with the 7-th order expansion of @xmath210 .
we also would like to stress that our formula of @xmath226 includes the free parameters @xmath189 and @xmath202 _ today _ , by which the joint analysis with other data ( such as cmb ) can be conveniently performed .
evolution of @xmath84 versus @xmath2 for the freezing quintessence with the potential @xmath227 $ ] .
the cases ( a)-(c ) correspond to the full numerical solutions with the model parameters ( a ) @xmath228 , @xmath229 , ( b ) @xmath230 , @xmath231 , and ( c ) @xmath232 , @xmath233 , respectively , whereas the case ( d ) is derived by the 10-th order analytic solution ( [ fsigma8 ] ) for @xmath230 , @xmath231.,width=326,height=307 ] in thawing models of quintessence the field equation of state is given by eq .
( [ wtha ] ) . for larger @xmath133
the deviation of @xmath8 from @xmath19 occurs at later times with a sharper transition . from eq .
( [ wthaw2 ] ) we find that the higher - order terms in @xmath22 are not negligible for @xmath133 larger than the order of 1 .
in fact we have numerically confirmed that , for @xmath234 , the expansion ( [ wthaw2 ] ) does not accommodate the rapid transition of @xmath8 at late times unless the higher - order terms are fully taken into account .
this property also holds for the evolution of @xmath18 .
only when @xmath133 is smaller than the order of 1 , the analytic estimation ( [ fsigma8 ] ) can reproduce the full numerical solutions in good precision . in fig .
[ thawing ] we plot the numerical evolution of @xmath84 for the potential @xmath227 $ ] with three different values of @xmath133 .
since @xmath8 is close to @xmath19 in all these cases and the deviation from @xmath136 occurs only at late times , the evolution of @xmath18 is not very different from each other for @xmath235 . from this analysis ,
it is clear that , only when very accurate data of @xmath18 are available in the future , it will be possible to distinguish between the models with different values of @xmath133 . in fig .
[ thawing ] we also show the analytic solution derived for @xmath230 and @xmath231 as the bold dashed line ( d ) .
we take into account the @xmath210 s up to 10-th order to evaluate @xmath84 in eq .
( [ fsigma8 ] ) . compared to the full numerical solution labelled as ( b )
, there is a small difference in the high - redshift regime .
we confirm that this deviation tends to be smaller by involving the @xmath210 s higher than 10-th order . for @xmath236 the analytic estimation is more accurate even without including such higher - order terms .
in this section , we place observational bounds on two models of dark energy discussed in secs .
[ conwsec ] and [ trasec ] by using the current rsd data presented in table [ table : fs8constraints ] . for the today s value of @xmath5 we consider the prior obtained from observations of cmb , bao , and hubble constant measurement ( @xmath237 ) ,
i.e. , _ 8 ( z=0)=0.816 0.024 .
[ sigprior ] here and in what follows all the error bars correspond to the 68.3% confidence level ( cl ) . recall that we derived the analytic formula ( [ fsigma8 ] ) under the approximation that the dark energy perturbation is neglected relative to the matter perturbation . for the validity of this approximation we put the prior @xmath238 .
for the models of constant @xmath8 the today s matter density parameter constrained from sn ia , cmb , bao , and @xmath237 observations is @xcite _ m0=0.272^+0.013_-0.013 .
[ omeprior ] under the priors ( [ sigprior ] ) and ( [ omeprior ] ) we estimate the best - fit to the set of parameters @xmath239 by evaluating the likelihood distribution function , @xmath240 , with ^2=_i=1 ^ 9 ( ) ^2 . here
@xmath241 ( @xmath242 ) are the 9 data displayed in table [ table : fs8constraints ] with the error bars @xmath243 , whereas @xmath244 are the theoretical values derived from the analytic solution ( [ fsigma8 ] ) . for the evaluation of @xmath244 we pick up the @xmath210 s up to 10-th order . for the growth index @xmath162 the terms
up to 2-nd order with respect to @xmath127 are included in eq .
( [ gammaap ] ) .
for @xmath22 we use the 1-st order solution ( [ omegax1 ] ) . in tracking quintessence models analyzed in sec .
[ trasubsec ] we also take the same orders of expansions for @xmath84 , @xmath162 , and @xmath22 . we find that the best - fit model parameters are w=-0.604,_8 ( z=0)=0.840,_m0=0.285 , with reduced @xmath245 ( @xmath246 , where @xmath247 stands for the degrees of freedom ) . at 68.3% cl ,
our analysis restricts the equation - of - state parameter to the interval -1.245 < w < -0.347 , [ bouw ] whereas @xmath189 and @xmath121 are unconstrained by current data even assuming the priors ( [ sigprior ] ) and ( [ omeprior ] ) .
although the current bounds on @xmath8 are weaker than those arising from background tests ( see , e.g. , refs .
@xcite ) , we expect that upcoming rsd data from ongoing and planned galaxy redshift surveys can improve this situation in the near future . for the tracking quintessence models in which the equation of state is given by eq .
( [ wtrack ] ) we also carry out the likelihood analysis by using the analytic solution ( [ fsigma8 ] ) of @xmath18 as well as the expressions for @xmath162 and @xmath22 given in eqs .
( [ gammaap ] ) and ( [ omegax1 ] ) respectively .
while the equation of state ( [ wtrack ] ) is derived for quintessence , we do not impose the prior @xmath248 for generality .
for this kind of models , a joint analysis involving current sn ia , cmb , and bao gives the following bound on the matter density parameter @xcite : 0.273<_m0<0.293 , [ bouome2 ] which is used in our analysis as a prior for @xmath121 .
the best - fit model parameters are found to be w_(0)=-0.461,_8 ( z=0)=0.840,_m0=0.293 , with @xmath249 . at 68.3% cl
, we found -1.288 <
w_(0 ) < -0.214 , [ bouw2 ] whereas the parameters @xmath189 and @xmath121 are again unconstrained in the regions of ( [ sigprior ] ) and ( [ bouome2 ] ) .
as expected , the bounds on @xmath30 are weaker than those obtained in constant @xmath8 models [ eq . ( [ bouw ] ) ] .
we note that in tracking models the equation of state decreases at late times , which is accompanied by the decrease of @xmath84 . compared to constant @xmath8 models , this allows the possibility to fit the data better even for larger values of @xmath8 during the matter era .
in this paper we have provided an analytic formula of @xmath18 for dynamical dark energy models in the framework of gr .
this was derived by using the approximate matter perturbation equation ( [ delmeq ] ) , which is trustable as long as the contribution of the dark energy perturbation to the gravitational potential is negligible relative to that of the matter perturbation .
our formula of @xmath18 can be applied to many dark energy models including imperfect fluids , quintessence , and k - essence in which the sound speed squared @xmath65 is not very close to 0 .
our derivation of @xmath18 is based on the expansion of @xmath8 with respect to the dark energy density parameter @xmath22 , i.e. , @xmath250 .
the growth rate @xmath160 of matter perturbations is parametrized by the growth index @xmath162 , as @xmath251 .
we expanded @xmath162 in terms of @xmath22 up to 2-nd order terms .
since @xmath162 is dominated by the term @xmath252 , it is a good approximation to treat @xmath162 as a constant for the derivation of the integrated solution of @xmath0 . the @xmath210 s in eq .
( [ fsigma8 ] ) are given by eq .
( [ cndef ] ) , where @xmath253 and @xmath180 appear as the coefficients of the expansion of the terms @xmath254 and @xmath178 respectively . for the density parameter @xmath22 ,
the 1-st order solution ( [ omegax1 ] ) is usually sufficient to obtain accurate analytic solutions of @xmath84 . in sec .
[ validitysec ] we have studied the validity of the analytic formula ( [ fsigma8 ] ) in concrete models of dark energy . for constant @xmath8 models in which @xmath210 and @xmath22 are given by eq .
( [ cnconst ] ) , the analytic solution up to 7-th order terms of @xmath210 reproduces the numerically integrated solutions with good precision .
this property also holds for tracking quintessence models where the evolution of @xmath8 is given by eq .
( [ wtrack2 ] ) . in thawing quintessence and k - essence models , where @xmath8 is given by eq .
( [ wtha ] ) , the formula ( [ fsigma8 ] ) can be trustable for @xmath255 , but for @xmath133 larger than the order of 1 , we need to fully take into account the higher - order terms of @xmath210 to have good convergence of @xmath18 . in sec .
[ obserconsec ] we have discussed observational constraints on two different dark energy models by using the current rsd data . in both constant @xmath8 and tracking quintessence models
the analytic solution ( [ fsigma8 ] ) includes the three parameters @xmath27 , @xmath121 , and @xmath8 ( or @xmath30 ) . under the priors on @xmath27 and @xmath121 constrained by sn ia , cmb , bao , and @xmath237 measurements
, we derived the bounds @xmath256 ( 68%cl ) for constant @xmath8 models and @xmath257 ( 68%cl ) for tracking quintessence models .
although the upper bounds on the dark energy equation of state are still weak with current data , we expect to obtain more precise data from ongoing surveys or near - future projects such as subaru / fmos , hetdex , and j - pas .
our analytic formula of @xmath18 will be useful to place tighter bounds on dynamical dark energy models in the future .
so far , observational bounds on @xmath84 ( listed in table [ table : fs8constraints ] ) have been derived in the standard cosmological scenario without taking into account additional effects such as a possible coupling between dark matter and dark energy , irrotational flow , and so on . reflecting this observational status
, we did not assume any non - standard picture to estimate the theoretical values of @xmath18 .
however , if the standard cosmological scenario does not match with future high - precision data very well , it may be necessary to include non - standard effects mentioned above as a next step .
we leave the theoretical estimation of such effects for future work .
we thank luca amendola , hiroyuki okada , and tomonori totani for useful discussions .
a. d. f. is supported by jsps ( under the grant no .
s12135 ) and thanks tokyo university of science , for the warm hospitality received while part of the project was finalized .
s. t. is supported by the grant - in - aid for scientific research fund of the fund of the jsps no .
24540286 and scientific research on innovative areas ( no .
21111006 ) . s. t. thanks warm hospitalities during his stays in weihai , observatorio nacional in rio de janeiro , passa quatro , szczecin , and university of heidelberg .
j. s. a. is supported by cnpq under grants no .
305857/2010 - 0 and no .
485669/2011 - 0 and faperj grant no . e-26/103.239/2011 .
d. j. eisenstein _ et al .
_ [ sdss collaboration ] , of sdss luminous red galaxies , astrophys . j. * 633 * , 560 ( 2005 ) ; w. j. percival _ et al . _
[ sdss collaboration ] , mon . not .
soc . * 401 * , 2148 ( 2010 ) .
s. m. carroll , living rev .
rel . * 4 * , 1 ( 2001 ) ; t. padmanabhan , phys .
rept . * 380 * , 235 ( 2003 ) ; p. j. e. peebles and b. ratra , rev . mod
. phys . * 75 * , 559 ( 2003 ) ; e. j. copeland , m. sami and s. tsujikawa , int .
j. mod .
d * 15 * , 1753 ( 2006 ) ; t. p. sotiriou and v. faraoni , rev .
phys . * 82 * , 451 ( 2010 ) ; a. de felice and s. tsujikawa , living rev .
* 13 * , 3 ( 2010 ) ; s. tsujikawa , lect .
notes phys .
* 800 * , 99 ( 2010 ) ; arxiv:1004.1493 [ astro-ph.co ] .
y. fujii , phys .
d * 26 * , 2580 ( 1982 ) ; l. h. ford , phys . rev .
d * 35 * , 2339 ( 1987 ) ; c. wetterich , nucl .
phys b. * 302 * , 668 ( 1988 ) ; t. chiba , n. sugiyama and t. nakamura , mon . not .
soc . * 289 * , l5 ( 1997 ) ; p. g. ferreira and m. joyce , phys .
lett . * 79 * , 4740 ( 1997 ) ; r. r. caldwell , r. dave and p. j. steinhardt , phys .
lett . * 80 * , 1582 ( 1998 ) .
s. capozziello , int .
j. mod .
d * 11 * , 483 ( 2002 ) ; s. capozziello , s. carloni and a. troisi , recent res .
astrophys .
* 1 * , 625 ( 2003 ) ; s. m. carroll , v. duvvuri , m. trodden and m. s. turner , phys . rev .
d * 70 * , 043528 ( 2004 ) ; l. amendola , r. gannouji , d. polarski and s. tsujikawa , phys .
d * 75 * , 083504 ( 2007 ) ; l. amendola and s. tsujikawa , phys . lett .
b * 660 * , 125 ( 2008 ) ; w. hu and i. sawicki , phys .
d * 76 * , 064004 ( 2007 ) ; a. a. starobinsky , jetp lett . * 86 * , 157 ( 2007 ) ; s. a. appleby and r. a. battye , phys .
b * 654 * , 7 ( 2007 ) ; s. tsujikawa , phys .
d * 77 * , 023507 ( 2008 ) ; e. v. linder , phys .
d * 80 * , 123528 ( 2009 ) ; m. campista , b. santos , j. santos and j. s. alcaniz , phys .
b * 699 * , 320 ( 2011 ) ; b. santos , m. campista , j. santos and j. s. alcaniz , arxiv:1207.2478 [ astro-ph.co ] .
m. tegmark _ et al . _
[ sdss collaboration ] , phys .
d * 69 * , 103501 ( 2004 ) ; u. seljak _ et al .
_ [ sdss collaboration ] , phys . rev .
d * 71 * , 103515 ( 2005 ) ; m. tegmark _ et al .
_ [ sdss collaboration ] , phys . rev .
d * 74 * , 123507 ( 2006 ) .
s. m. carroll , i. sawicki , a. silvestri and m. trodden , new j. phys .
* 8 * , 323 ( 2006 ) ; r. bean , d. bernat , l. pogosian , a. silvestri and m. trodden , phys .
d * 75 * , 064020 ( 2007 ) ; s. tsujikawa , phys .
d * 76 * , 023514 ( 2007 ) ; l. pogosian and a. silvestri , phys
. rev .
d * 77 * , 023503 ( 2008 ) ; s. tsujikawa , r. gannouji , b. moraes and d. polarski , phys .
d * 80 * , 084044 ( 2009 ) . | the red - shift space distortions in the galaxy power spectrum can be used to measure the growth rate of matter density perturbations @xmath0 . for dynamical dark energy models in general relativity
we provide a convenient analytic formula of @xmath1 written as a function of the redshift @xmath2 , where @xmath3 ( @xmath4 is the cosmological scale factor ) and @xmath5 is the rms amplitude of over - density at the scale @xmath6@xmath7mpc . our formula can be applied to the models of imperfect fluids , quintessence , and k - essence , provided that the dark energy equation of state @xmath8 does not vary significantly and that the sound speed is not much smaller than 1 .
we also place observational constraints on dark energy models of constant @xmath8 and tracking quintessence from the recent data of red - shift space distortions . |
from a quantum field theory point of view , gravity is an interaction mediated by a spin-2 particle , the graviton @xcite .
a long - standing issue has been whether gravitons are truly massless like the gluon , or more like a light but massive neutrino ( the latter has no gauge `` protection '' , of course ) .
massive gravity ( mgr ) can be traced back to fierz and pauli s ( fp ) 1939 formulation @xcite of the free massive spin-2 field in flat space .
they uniquely fixed it by requiring that it represents @xmath0 ( rather than the generic 6 ) excitations which then also guaranteed tachyon- and ghost - freedom .
the fp action is the sum of the massless
linearized gr kinetic and @xmath1 mass , terms and @xmath2 necessarily .
but , as was realized much later @xcite , when stress - tensor sources @xmath3 are included , their induced interactions take the form @xmath4 , where @xmath5 , while @xmath6 .
this discrete discontinuity of course persists as @xmath7 , leading to the well - known 25% error in the massive model s light - bending prediction .
the possibility remained that this disease was caused by taking the linear limit before the massless one @xcite .
however , this triggers a new and frightening obstacle : at nonlinear level there ( generically ) arises a massive , _ ghost _ , 6th degree of freedom beyond the 2@xmath8 + 1=5 of fp the so - called boulware - deser ( bd ) ghost @xcite .
consequently , interest in mgr dwindled , until a presumably consistent bd - ghost - free extension was recently constructed @xcite .
these models involve , besides the dynamical metric @xmath9 , a second , fixed , ( fiducial ) background metric @xmath10 ; in tetrad formalism , @xmath11 is the background ( with inverse denoted by @xmath12 ) and @xmath13 the dynamical field .
all index manipulations will be performed using the dynamical metric and tetrad , with greek and latin indices respectively representing world and local lorentz coordinates .
the simplest example of nonlinear mgr has field equation @xmath14 where @xmath15 and @xmath16 .
the parameter @xmath17 is the fp mass when the theory is linearized around a cosmological @xmath18 background .
the correct linearization requires that @xmath19 .
the tetrads obey the symmetry constraint @xmath20m}=0 $ ] ; its curl implies the integrability condition @xmath21\sigma}=0 $ ] , where @xmath22 is the contortion and @xmath23 , @xmath24 are the spin connections . despite being ghost - free
, subsequent investigations indicated that nonlinear mgr is still problematic .
in particular , the characteristic equations in the eikonal limit were analyzed by two of the authors @xcite .
it was found that the model admits superluminal ( second order ) shock wave solutions , which ironically , is due to the very constraint that removes the bd ghost . previously superluminal behavior was also uncovered in the model s stckelberg sector and decoupling limit @xcite as well as in a spherically symmetric analysis on friedmann - lematre - robertson - walker ( flrw ) backgrounds @xcite .
one might think that since the graviton in this theory is _ massive _ , it would automatically propagate _ slower _ than light .
however this is not necessarily the case .
a simple counter - example is the nonlinear proca field ( massive photon ) that also gives rise to a mode that propagates faster than light @xcite .
after the analysis , by the other two authors , of the characteristic matrix of the theory using pde analysis la cauchy and kovalevskaya @xcite , the negative results of refs . and were combined to show the massive theory not only gives rise to superluminal shock waves , but also local acausality @xcite that can arise even in an infinitesimal neighborhood of a spacetime event , as summarized in the next section .
superluminal shock waves can be found by studying discontinuities in the first derivative of fields across a hypersurface @xmath25 , with normal @xmath26 chosen to be timelike ; this is denoted by @xmath27_\sigma=\xi_\alpha{\cal e}_\mu{}^m , \qquad [ \partial_\alpha \omega_\mu{}^m{}_n]_\sigma=\xi_\alpha \omega_\mu{}^m{}_n.\ ] ] in particular we investigate the discontinuities in the scalar constraint ( this equation does _ not _ involve higher than first derivatives , and is responsible for removing the 6th , bd ghost , excitation ) @xmath28)+\frac{1}2\ , g^{\mu\nu } \,[{g}_{\mu\nu}-\tau_{\mu\nu}],\ ] ] and in the curl of the symmetry constraint . denoting the contraction of @xmath26 on an index of any tensor by an `` @xmath29 '' ,
e.g. _ , @xmath30 , upon carrying out a shock analysis of the theory s constraints , we obtain the characteristic matrix @xcite @xmath31 & \frac 12\ , \tilde k_j \\[3 mm ] [ f\times k\ell]_i & f_{ij}-g_{ij } f^{(3 ) } \end{pmatrix}\begin{pmatrix}{\cal f}_{oo}\\[3mm]\tilde\omega^j\end{pmatrix}.\ ] ] here @xmath32 , @xmath33 and @xmath34 , where @xmath35 , and @xmath36_i:=2\ , \epsilon_{ijk } f^{k\mu } k^{j}{}_{\nu\mu}\ell^\nu_o$ ] and @xmath37 .
this gives a sufficient ( but not necessary ) condition for the _ field - dependent _ determinant to vanish : @xmath38 -\frac 12\ , \tilde k_i\ell^{ij}_{(3 ) } [ f\times k\ell]_j,\ ] ] where @xmath39 . by choosing appropriate values of the fields
, this determinant can be easily made to vanish , and so superluminal shocks are quite generic .
an easy ( and by no means even the only ) demonstration of this fact is that for configurations such that @xmath40 in some region , it follows that @xmath41 whence the condition for a vanishing determinant of the characteristic matrix reduces to , in an obvious matrix notation , @xmath42 .
thus , the non - vanishing of sums of every pair @xmath43 of eigenvalues of @xmath44 is sufficient and necessary for @xmath42 to imply @xmath45 .
however , the eigenvalues of @xmath44 are not necessarily positive , because of the difference in being `` spacelike '' with respect to the two metrics .
hence superluminal shocks can occur .
moreover , acausality is now very likely to arise since closed timelike curves can be locally embedded into spacetime because spacelike surfaces can be characteristic ones as well . to be more specific , specializing to a minkowski background ( say ) ,
consider the case when the temporal direction , together with one of the spatial coordinates ( say the @xmath46-direction ) , of the fiducial and physical tetrads do not coincide .
for example , if @xmath47 , then we have @xmath48 and @xmath49 , _
i.e. _ , a constant time hypersurface is a characteristic .
the action is invariant under simultaneous local lorentz rotations of both physical and fiducial tetrads .
because of the interchange of the time- and @xmath46-direction some of these symmetries are `` spontaneously '' broken , while @xmath50-rotation and @xmath46-boost symmetries are kept .
the @xmath51-rotation symmetry is broken ( because of the flip of the @xmath46-direction , physical and fiducial tetrads rotate oppositely so physical and fiducial tetrads configurations change ) , but rotating by @xmath52 , the new configurations thus generated become the same as the original ones .
combining with the @xmath50-rotation symmetry , the solution has still a parity symmetry with respect to the 3-direction .
the time - constant characteristic hypersurface must support propagation in some direction @xmath53 ( say ) .
thanks to the @xmath50-rotation symmetry and parity symmetry with respect to @xmath46-direction , there must also be propagations in directions @xmath54 where @xmath55 is an arbitrary constant .
thus , we can construct closed timelike curves , for instance form a loop : @xmath53 , @xmath56 , @xmath57 and @xmath58 . furthermore
, since @xmath46-boosts do not change field configurations , and preserve @xmath47 , the same method shows that closed timelike curves can be formed on the hypersurface generated by boosts .
moreover , although a @xmath59-boost does not keep the original tetrad configurations , it will preserve the condition @xmath49 , and thus the boost - hypersurface is again characteristic . because of @xmath50-rotational symmetry , the same holds if we consider a @xmath60-boost or mixture of @xmath59- and @xmath60-boosts . by appropriate boosting , any spacelike hypersurface can be characteristic .
we emphasize that we used flat background and field configurations @xmath47 purely for simplicity ; it is not essential to our acausality .
let us comment further on the difference between superluminality and acausality . in gr , we expect superluminal propagation to be also acausal , _
i.e. _ we can construct closed timelike curves in the theory .
however , in a theory that does not have coordinate invariance , superluminality does not always lead to acausality .
for example , in newtonian gravity , there is no upper speed limit , but there is no problem with acausality .
in fact , even in a theory with diffeomorphism invariance and lorentz invariance , one should remember that the mathematical structure of special relativity only requires _ existence _ of an _ upper bound _ on speed , and thus it could well be that the speed of light is very close to , but not really , the upper bound .
thus , superluminality is not always disastrous@xcite .
nonlinear mgr is manifestly _ not _ diffeomorphism invariant and the existence of superluminal shocks by itself is only an indication of _ possible _ disaster , but existence of acausality means that the theory is _ definitively _ bad .
this scenario also happens in @xmath61 gravity , which is not locally lorentz invariant and thus can only be argued to be problematic via an acausality argument or the complete absence of predictability @xcite .
our investigations showed that nonlinear mgr does indeed admit acausality , which is not only much easier to construct than in gr , but is _ local _ in nature .
that is , unlike say , the gdel solution in gr , which in the neighborhood of any point of the closed timelike curve is still perfectly well - behaved ( _ i.e. _ , locally one always moves forward in time ) , the acausality in nonlinear mgr can be constructed in an infinitesimal regions of spacetime , and thus is much more serious .
we mention that these problems are likely to persist in the bimetric extension of mgr , where the background field becomes dynamical @xcite .
ever since its conception , mgr has struggled to survive successive blows by adding successive epicycles ( for a review , see ref . ) .
our results demonstrate that its current incarnation is also untenable , due to the existence of not only superluminal shock waves , but also _ local _ acausality .
[ the background metric is essentially an external field ; such fields are well - known to generate acausalities in higher - spin theory contexts @xcite . ]
furthermore this acausality occurs quite generically , not only in the model s decoupling limit .
this means that mgr _ can not _ be a uv - complete fundamental theory of gravity .
one may argue that it can still be useful as an effective field theory .
that is , it could still be well - behaved without acausality on some specific background , separated from the problematic ones by a putative potential barrier .
however , since our acausality argument only depends on the rather weak condition that @xmath41 , this situation is very unlikely .
furthermore , such a potential barrier can not save the theory since it is no longer protected at the quantum level , especially since the natural scale of the theory is @xmath62 , where @xmath63 , the graviton mass , is necessarily very small .
indeed , for @xmath64 , the value of such an energy scale is roughly @xmath65ev , above which the effective field description is no longer applicable .
[ the absence of a supersymmetric ( @xmath66 ) extension even of fp simply because massive spin 3/2 has only @xmath67 excitations
shows that susy s virtues are likely not available here either . ]
the only possible way to remove the offending ghost - removing , but also the cause of all our problems
scalar constraint would be to rely on the existence of a partially massless version of mgr ; unfortunately , this last hope is also excluded , precisely at nonlinear level @xcite .
it is gratifying that that gr and su(3 ) ym are unique in both being exempt from the standard higgs mass coupling mechanism , and in being isolated " : not having viable neighboring " non - gauge models .
we thank m. porrati for requesting a ctc clarification , and w. siegel for a susy query .
s. deser was supported in part by nsf phy- 1266107 and doe de- fg02 - 164 92er40701 grants .
k. izumi is supported by taiwan national science council under project no .
nsc101 - 2811-m-002 - 103 .
y. c. ong was supported by the taiwan scholarship from taiwan s ministry of education . c. de rham , g. gabadadze , phys .
d**82 * * ( 2010 ) 044020 , http://arxiv.org/abs/1007.0443[[1007.0443v2 [ hep - th ] ] ] ; c. de rham , g. gabadadze , a. j. tolley , phys .
106 * ( 2011 ) 231101 , http://arxiv.org/abs/1011.1232[[1011.1232v2 [ hep - th ] ] ] .
a. gruzinov , http://arxiv.org/abs/1106.3972[[1106.3972 [ hep - th ] ] ] ; c. burrage , c. de rham , l. heisenberg and a. j. tolley , jcap * 1207 * ( 2012 ) 004 , http://arxiv.org/abs/1111.5549[[1111.5549 [ hep - th ] ] ] ; p.
de fromont , c. de rham , l. heisenberg and a. matas , http://arxiv.org/abs/1303.0274[[1303.0274 [ hep - th ] ] ] .
r. geroch , http://arxiv.org/abs/1005.1614[[1005.1614 [ gr - qc ] ] ] ; j - p .
bruneton , phys .
d 75 ( 2007 ) 085013 , http://arxiv.org/abs/gr-qc/0607055[[gr-qc/0607055 ] ] ; n. afshordi , d. j. h. chung , g. geshnizjani , phys . rev .
d 75 ( 2007 ) 083513 , http://arxiv.org/abs/hep-th/0609150[[hep-th/0609150 ] ] ; e. babichev , v. mukhanov , a. vikman , jhep 0802 ( 2008 ) 101 , http://arxiv.org/abs/0708.0561[0708.0561 [ hep - th ] ] .
k. johnson and e.c.g .
sudarshan , ann .
* 13 * ( 1961 ) 126 ; g. velo and d. zwanziger , phys.rev .
* 186 * ( 1969 ) 1337 ; m. kobayashi and a. shamaly , phys .
d * 17 * , ( 1978 ) 2179 ; prog .
phys . * 61 * ( 1979 ) 656 ( 1979 ) ; s. deser and a. waldron , nucl .
b * 631 * ( 2002 ) 369 , http://arxiv.org/abs/hep-th/0112182[[hep-th/0112182 ] ] . | massive gravity is an old idea : trading geometry for mass .
much effort has been expended on establishing a healthy model , culminating in the current ghost - free version .
we summarize here our recent findings that it is still untenable because it likely admits _ local _ acausalities : solutions with ctcs in a small neighborhood of any event . |
the consideration of exotic dark matter has become necessary in order to close the universe @xcite@xmath0 @xcite .
furthermore in in order to understand the large scale structure of the universe one has to consider matter made up of particles which were non - relativistic at the time of freeze out , i.e cold dark matter ( cdm ) .
the cobe data @xcite suggest that cdm is at least @xmath1 @xcite .
recent data from the high - z supernova search team @xcite and supernova cosmology project @xcite @xmath0 @xcite suggest the necessity of the cosmological constant @xmath2 .
in fact the situation can adequately be described by a barionic component @xmath3 along with the exotic components @xmath4 and @xmath5 . in another analysis turner @xcite
gives @xmath6 .
since the non exotic component can not exceed @xmath7 of the cdm @xcite@xmath0 @xcite , there is room for the exotic wimp s ( weakly interacting massive particles ) .
in fact the dama experiment @xcite has claimed the observation of one signal in direct detection of a wimp , which with better statistics has subsequently been interpreted as a modulation signal @xcite .
the above developments are in line with particle physics considerations .
thus , in the currently favored supersymmetric ( susy ) extensions of the standard model , the most natural wimp candidate is the lsp , i.e. the lightest supersymmetric particle . in the most favored scenarios the lsp can be simply described as a majorana fermion , a linear combination of the neutral components of the gauginos and higgsinos @xcite@xmath0 @xcite@xmath8@xcite .
since this particle is expected to be very massive , @xmath9 , and extremely non relativistic with average kinetic energy @xmath10 , it can be directly detected @xcite@xmath8@xcite mainly via the recoiling of a nucleus ( a , z ) in the elastic scattering process : @xmath11 ( @xmath12 denotes the lsp ) . in order to compute the event rate
one needs the following ingredients : \1 ) an effective lagrangian at the elementary particle ( quark ) level obtained in the framework of supersymmetry as described in refs .
@xcite , bottino _ et al . _
@xcite , @xcite , kane _ et al .
_ , castano _ et al .
_ and arnowitt _
et al_. our own susy input parameters will appear elsewhere @xcite : 2 ) a procedure in going from the quark to the nucleon level , i.e. a quark model for the nucleon .
the results , for the scalar and the isoscalar axial couplings , depend crucially on the content of the nucleon in quarks other than u and d @xcite .
\3 ) compute the relevant nuclear matrix elements @xcite@xmath8@xcite .
since the obtained rates are very low , one would like to be able to exploit the modulation of the event rates due to the earth s revolution around the sun .
one also would like to know the directional rates , by observing the nucleus in a certain direction , which correlate with the motion of the sun around the center of the galaxy .
the purpose of our present review is to focus on this last point along the lines suggested by our recent work @xcite .
combining for results of the previous section we can write @xmath13 where @xmath14 we have neglected the uninteresting pseudoscalar and tensor currents . note that , due to the majorana nature of the lsp , @xmath15 ( identically ) . with the above ingredients the differential cross section can be cast in the form @xmath16 \label{2.9}\ ] ] @xmath17 \simeq \sigma^s_{p,\chi^0 } a^2 ( \frac{\mu_r}{m_n})^2 \label{2.10}\ ] ] @xmath18 @xmath19 \label{2.10b}\ ] ] @xmath20 @xmath21 \label{2.10d}\ ] ] + @xmath22 proton cross - section,@xmath23 given by : + @xmath24 ( scalar ) , ( the isovector scalar is negligible , i.e. @xmath25 + @xmath26 ( spin ) , @xmath27 ( vector ) + where @xmath28 is the proton mass , @xmath29 , and @xmath30 is the reduced mass and @xmath31 @xmath32 where b is ( the harmonic oscillator ) size parameter
, q is the momentum transfer to the nucleus , and q is the energy transfer to the nucleus + in the above expressions @xmath33 is the nuclear form factor and @xmath34 are the spin form factors @xcite ( @xmath35 are isospin indices ) + both form factors are normalized to one at @xmath36 .
+ the nom - directional event rate is given by : @xmath37 where @xmath38 is the lsp density in our vicinity and m is the detector mass the differential non - directional rate can be written as @xmath39 where @xmath40 was given above .
the directional differential rate @xcite in the direction @xmath41 is given by : @xmath42 where h is the heaviside step function .
the factor of @xmath43 is introduced , since we have chosen to normalize our results to the usual differential rate .
we have seen that the event rate for lsp - nucleus scattering depends on the relative lsp - target velocity . in this section
we will examine the consequences of the earth s revolution around the sun ( the effect of its rotation around its axis is expected to be negligible ) i.e. the modulation effect . in practice
this has been accomplished by assuming a consistent lsp velocity dispersion , such as a maxwell distribution @xcite or its extensions with only axial symmetry @xcite , already been discussed in the lierature @xcite .
more recently other non - isothermal approaches , in the context velocity peaks and caustic rings , have been proposed , see e.g sikivie et al @xcite . in the last case one
considers the late in - fall of dark matter into our galaxy producing flows of caustic rings .
in particular the predictions of a self - similar model have been put forward as a possible scenario for dark matter density - velocity distribution , see e.g. sikivie et al @xcite . the implications of such theoretical schemes and , in particular , the modulation effect are the subject of this section . following sikivie we will consider @xmath44 caustic rings , ( i , n ) , i=(+.- ) and n=1,2, ... n ( n=20 in the model of sikivie et al ) , each of which contributes to the local density a fraction @xmath45 of the of the summed density @xmath46 of each of the @xmath47 .
it contains wimp like particles with velocity @xmath48 in units of essentially the sun s velocity ( @xmath49 ) , with respect to the galactic center .
the z - axis is chosen in the direction of the disc s rotation , i.e. in the direction of the motion of the the sun , the y - axis is perpendicular to the plane of the galaxy and the x - axis is in the radial direction .
we caution the reader that these axes are traditionally indicated by astronomers as @xmath50 respectively .
the needed quantities are taken from the work of sikivie et al @xcite , via the definitions @xmath51 and @xmath52^{1/2}$ ] ( for each flow + .- ) .
this leads to a velocity distribution of the form : @xmath53 since the axis of the ecliptic @xcite .
lies very close to the @xmath54 plane the velocity of the earth around the sun is given by @xmath55 where @xmath56 is the phase of the earth s orbital motion , @xmath57 around second of june .
one can now express the above distribution in the laboratory frame @xcite by writing @xmath58
the mean value of the non - directional event rate of eq .
( [ 2.18 ] ) , is given by @xmath59 the above expression can be more conveniently written as @xmath60 where @xmath61 there are now experiments under way aiming at measuring directional rates , i.e. the case in which the nucleus is observed in a certain direction .
the rate will depend on the direction of observation , showing a strong correlation with the direction of the sun s motion . in a favorable situation
the rate will merely be suppressed by a factor of @xmath62 relative to the non - directional rate .
this is due to the fact that one does not now integrate over the azimuthal angle of the nuclear recoiling momentum .
the directional rate will also show modulation due to the earth s motion . the mean value of the directional differential event rate of eq .
( [ 2.20 ] ) , is defined by @xmath63 it can be more conveniently expressed as @xmath64 where @xmath65 we are not going to discuss the differential rates further .
we will limit our discussion to the case of the total rates .
integrating the differential rate from @xmath66 to @xmath67 we obtain for the total non - directional ratei . in the case of caustic rings
we find : @xmath68 \label{3.55a } \end{aligned}\ ] ] where @xmath69 \label{3.39b } \end{aligned}\ ] ] with @xmath70 have been defined above , see eqs ( [ 2.10 ] ) - ( [ 2.10c ] ) .
furthermore @xmath71^{-1}$ ] @xmath72 here @xmath73 , with @xmath74 is the escape velocity from the galaxy .
the modulation is described by the parameter @xmath75 . similarly integrating eq .
( [ 3.12 ] ) we obtain for the total non - directional rate in our isothermal model as follows : @xmath76 \label{3.55b } \end{aligned}\ ] ] now @xmath77 note the difference of sign in the definition of the modulation amplitude h compared to eq .
( [ 3.55a ] ) .
the modulation can be described in terms of the parameter @xmath75 .
the effect of folding with lsp velocity on the total rate is taken into account via the quantity @xmath78 .
the meaning of @xmath78 is clear from the above discussion .
we only like to stress that it is a common practice to extract the lsp nucleon cross section from the the expected experimental event rates in order to compare with the susy predictions as a function of the lsp mass .
we will again examine the case of caustic rings and the isothermal models considered above .
we need distinguish distinguish the following cases : a ) @xmath41 has a component in the sun s direction of motion , i.e. @xmath79 , labeled by i = u ( up ) .
b ) detection in the direction specified by @xmath80 , labeled by i = d ( down ) .
thus : \1 . in the first quadrant ( azimuthal angle @xmath81 .
@xmath82 \label{3.56 } \end{aligned}\ ] ] 2 . in the second quadrant ( azimuthal angle @xmath83 @xmath84 \label{3.57 } \end{aligned}\ ] ] 3 . in the third quadrant ( azimuthal angle @xmath85 .
@xmath86 \label{3.58 } \end{aligned}\ ] ] 4 . in the fourth quadrant ( azimuthal angle @xmath87 @xmath88 \label{3.59 } \end{aligned}\ ] ] in this case we find it convenient to consider the absolute value of the difference of the rates in the two opposite directions .
integrating eq .
( [ 4.10 ] ) and restricting ourselves close to the axes we obtain @xmath89 note that now the rate is normalized to @xmath90 and not to @xmath78 and that the modulation can be described in terms of three parameters @xmath91 , l=1,2,3 .
we will discuss the effects of folding with the lsp velocity combined with the nuclear form factor and specialized in the case of the nucleus @xmath92 , which is one of the most popular targets @xcite@xmath0@xcite @xcite@xmath8 @xcitefor the enegy cutoffs : @xmath93 kev . in the case of the non isothermal model of sikivie
et al @xcite , the total rates are described in terms of the quantities @xmath94 for the unmodulated amplitude and @xmath95 @xmath96 for the modulated one .
.the quantities @xmath78 and @xmath75 entering the total non - directional rate in the case of the target @xmath97 for various lsp masses and @xmath98 in kev . also shown are the quantities @xmath99 @xmath96 and @xmath100 , entering the directional rate for no energy cutoff . for definitions see text . [ cols="^,^ , > , > , > , > , > , > , > " , ]
we have studied the parameters describing the event rates for direct detection of susy dark matter and , in particular , two characteristic signatures , which will aid the experimentalists in reducing background : the non directional modulated event rates , which are correlated with the motion of the earth and the directional event rates , which are correlated with both the velocity of the sun and the velocity of the earth . in the case of non isothermal models @xcite from table 1 we see that the maximun no longer occurs around june 2nd , but about six months later . the difference between the maximum and the minimum is about @xmath101 , smaller than that predicted by the symmetric isothermal models @xcite . in the case of the directional rate we found that the rates depend on the direction of observation .
the biggest rates are obtained , if the observation is made close to the direction of the sun s motion .
the directional rates are suppressed compared to the usual non - directional rates by the factor @xmath102 .
we find that @xmath103 , if the observation is made in the sun s direction of motion , while @xmath104 in the opposite direction .
the modulation is a bit larger than in the non - directional case , but the largest value , @xmath105 , is not obtainrd along the sun s direction of motion , but in the x - direction ( galactocentric direction ) . in the case of the isothermal models we restricted our discussion to the directional event rates .
the reduction factor is now given by the parameter @xmath106 .
we find that @xmath107 is around 0.6 for no asymmetry and around 0.7 for maximum asymmetry ( @xmath108 ) .
in other words it is not very different from the naively expected @xmath109 .
the modulation of the directional rate increases with the asymmetry parameter @xmath110 and it depends , of course , on the direction of observation . for @xmath111 it can reach values up to @xmath112 .
larger values , up to @xmath113 , are possible for large values of @xmath98 , but they occur at the expense of the total number of counts . in all cases our results indicate that @xmath78 for large reduced mass deviates from unity .
thus , if one attempts to extract the lsp - nucleon cross section from the data to compare it with the predictions of susy models , one must take @xmath78 into account .
_ acknowledgments : _ one of the authors ( jdv ) would like to acknowledge partial support of this work by @xmath114ene@xmath115 1895/95 of the greek secretariat for research and tmr no erb fmax - ct96 - 0090 of the european union .
kolb , e.w . and
turner m.s .
( 1990 ) _ the early universe _ , addison , wesley .
+ peebles , p.j.e ( 1993 ) , _ principles of physical cosmology _
, princeton university press .
+ for a recent review see e.g. jungman , g. _ et al .
_ , ( 1996 ) _ phys .
rep._*267 * , 195 .
smoot , g.f .
_ et al . _
( 1992 ) , _ astrophys . j._*396
gawiser e. and silk j. ( 1988 ) _ science _ * 280 * , 1405 ; gross , m.a.k , somerville , r.s . ,
primack , j.r . , holtzman , j. and klypin , a.a .
( 1998 ) _ mon . not .
r. astron .
soc . _ * 301 * , 81 .
riess , a.g .
_ et al _ ( 1998 ) , _ astron .
j. _ * 116 * , 1009 .
somerville , r.s . ,
primack , j.r . and
faber , s.m.(2000 ) _ astro - ph_/9806228 ; _ mon . not .
( in press)_. perlmutter , s. _ et al _ ( 1999 ) _ astrophys .
* 517*,565 ; ( 1997 ) * 483*,565 .
+ perlmutter , s. , turner , m.s . and
white , m. ( 1999 ) , _ phys . rev .
_ * 83*,670 ( 1999 )
turner , m.s .
cosmological parameters_,_astro - ph_/9904051 ; ( 1990 ) _ phys .
rep . _ * 333 - 334 * , 619 .
bennett , d.p . , _
( 1995 ) , ( macho collaboration ) , _ a binary lensing event toward the lmc : observations and dark matter implications , proc .
5th annual maryland conference , edited by s. holt _ ; alcock , c. , _ et al .
_ , ( macho collaboration ) , _ phys . rev . lett._,*74 * 2967 .
bernabei , r. _ et al _ , ( 1998 ) infn / ae-98/34 ; bernabei , r.,_et al _ ( 19996 ) , _ phys .
* b 389 * , 757 .
bernabei , r. , et al ( 1998 ) , _ phys .
* b 424 * , 195 ; ( 1999 ) _ phys .
lett._*b 450 * , 448 .
vergados , j.d .
( 1996 ) _ j. of phys .
_ * g 22 * , 253 ) .
kosmas , t.s . and vergados , j.d ( 1997 ) , _ phys . rev .
_ * d 55 * , 1752 .
drees , m. and nojiri , m.m ( 1993 ) _ phys
_ * d 48 * , 3843 ; djouadi , a. and drees , m. ( 2000 ) , _ qcd corrections to the neuttalino nucleon scattering _ , tum - hep-370 - 00;pm/00 - 14 ; dawson , s. ( 1991 ) , _ nucl . phys .
_ * b359*,283 ; spira , m. , _ et al _ ( 1995 ) , _ nucl . phys . _ * b453*,17 .
goodman , m.w . and witten , e. ( 1985 ) , _ phys . rev .
_ * d 31 * , 3059 ( 1985 ) ; griest , k. ( 1998 ) , _ phys .
* 61 * , 666 ; ( 1988 ) , _ phys .
_ * d 38 * , 2357 ; ( 1989 ) , _ phys .
rev._*d 39 * , 3802 ; ellis , j. and flores , r.a .
( 1991 ) _ phys .
* b 263 * , 259 ; ( 1993 ) , _ phys .
_ * b 300 * , 175 ; ( 1993),_nucl .
phys._*b 400 * , 25 ; ellis , j. and roszkowski , l. ( 1992 ) , _ phys .
* b 283 * , 252 .
vergados , j.d .
( 1999 ) _ phys .
_ * 83 * , 3597 .
vergados j.d .
( 2000 ) , _ phys .
_ * d 62*,023519 - 1 ; _ astro - ph_/0001190 m. gomez and j.d .
vergados , ( 2000 ) , _ direct lsp - nucleus cross section _ ( to appear ) .
bottino , a. , _ et al . _
( 1992 ) , _ mod .
lett . _ * a 7 * , 733 ( 1992 ) ; ( 1991 ) , _ phys .
* b 265 * , 57 ; ( 1997 ) , phys .
lett * b 402 * , 113 ; _ hep - ph _
@xmath116 9709222 ; _ hep - ph _ @xmath1169710296 ; edsjo , j. and gondolo , p. ( 1997 ) , _ phys .
* d 56 * , 1789 ; berezinsky , z. , _ et al . _
( 1996 ) , _ astroparticle phys . _
* 5 * , 1 ; bednyakov , v.a . , klapdor - kleingrothaus , h.v . and
kovalenko s.g . , ( 1994 ) ,
* b 329 * , 5 .
kane , g.l . , _
( 1994 ) , _ phys .
* d 49 * , 6173 ; castao , d.j . ,
piard , e.j . and
ramond , p. ( 1994 ) , _ phys .
* d 49 * , 4882 ; d.j .
castao , _ private communication _ ; + chamseddine , a.h . ,
arnowitt , r. and nath , p. ( 1982 ) , phys .
* 49 * , 970 ; nath , p , arnowitt , r and chamseddine , a.h .
( 1983 ) , _ nucl .
* b 227 * , 121 ; arnowitt , r. and nath , p. ( 1995 ) , mod .
. lett . * 10 * , 1215 ; arnowitt , r. and nath , p. ( 1995 ) , _ phys .
lett . _ * 74 * , 4952 ; arnowitt , r. and nath , p. ( 1996 ) , _ phys .
_ * d 54 * , 2394 .
arnowitt , r. and nath , p. ( 1999 ) , _ hep - ph_@xmath117 ; _ hep - ph_/9902237 ; soni , s.k . and weldon , h.a .
( 1983 ) , _ phys .
* b 126 * , 215 kapunovsky , v.s . and louis , j. ( 1993 ) _ phys .
* b 306 * , 268 ; ( 1986 ) , _ phys . lett _ * b 181 * , 279 ; nath , p. and arnowitt , r. ( 1989 ) , _ phys .
_ * d 39 * , 279 ; hagelin , j.s .
and kelly , s. ( 1990 ) , _ nucl .
phys . _ * b 342 * , 95 ; kamamura , y , murayama , h. and yamaguchi , m ( 1994 ) , _ phys
* b 324 * , 52 ; dimopoulos s. and georgi , h. ( 1981 ) , _ nucl .
* b 206 * , 387 .
haber , h.e . and
kane , g.l .
( 1985 ) , phys
117 * 75 ( 1985 ) .
cheng , t.p .
( 1998 ) , _ phys .
* d 38 * 2869 ; cheng , h .- y . ( 1989 ) , _ phys .
* b 219 * 347 .
ressell , m.t . , _
( 1993 ) , _ phys .
rev_. * d 48 * , 5519 ; ressell , m.t . and dean , d.j .
( 1997 ) , _ hep - ph_@xmath118 .
dimitrov , v.i . ,
engel , j. and pittel , s. ( 1995 ) , _ phys .
_ * d 51 * , r291 .
engel , j. ( 1991 ) , _ phys .
lett . _ * b 264 * , 114 .
nikolaev , m.a . and klapdor - kleingrothaus , h.v .
( 1993 ) , _ z. phys . _ * a 345 * , 373 ; ( 1993 ) , _ phys .
lett . _ * 329 b * , 5 ; ( 1995 ) , _ phys .
* d 50 * , 7128 .
divari , p.c . ,
kosmas , t.s .
, vergados , j.d . and skouras , l.d .
( 2000 ) , _ phys .
_ * c 61 * , 044612 - 1 .
copi , c.j . ,
heo , j. and krauss , l.m .
( 1999 ) , _ phys .
lett . _ * 461 b * , 43 ; drukier , a.k.,_et al _ , ( 1986 ) _ phys .
* d 33 * , 3495 ; + collar , j.i . , _
et al _ ( 1992 ) , _ phys .
* b 275 * , 181 .
sikivie , p. , tkachev , i. and wang , y. , ( 1995 ) _ phys .
_ * 75 * , 2911 ; ( 1997 ) , _ phys
. rev . _ * d 56 * , 1863 .
sikivie , p. ( 1998 ) , _ phys .
_ * b 432 * , 139 ; astro - ph/9810286 smith , p.f . , _
et al _ ( 1999 ) , _ phys .
_ * 307 * , 275 ( 1999 ) ; spooner , n. ( 1999 ) , _ phys.rep . _
* 307 * , 253 ; quenby , j.j . , _
et al _ ( 1996 ) _ astropart .
phys . _ * 5 * , 249 ) .
primack , j.r . ,
seckel , d. and sadoulet , b. ( 1998 ) _ ann .
nucl . part .
sci . _ * 38 * , 751 .
smith , p.f . and
lewin , j.d .
( 1990 ) _ phys .
rep . _ * 187 * , 203 .
vergados , j.d . and kosmas , t.s .
( 1998 ) , _ physics of atomic nuclei _ ,
vol . * 61 * , no * 7 * , 1066 ( from _ yadernaya fisika _ , vol.*61 * , no*7 * , 1166 .
vergados , j.d .
( 1998 ) , _ phys .
* d 58 * , 103001 - 1 . | the direct detection of dark matter is central to particle physics and cosmology .
current fashionable supersymmetric models provide a natural dark matter candidate which is the lightest supersymmetric particle ( lsp ) .
such models combined with fairly well understood physics like the quark substructure of the nucleon and the structure of the nucleus ( form factor and/or spin response function ) , permit the evaluation of the event rate for lsp - nucleus elastic scattering .
the thus obtained event rates are , however , very low or even undetectable .
so it is imperative to exploit characteristic signatures , like the modulation effect , i.e. the dependence of the event rate on the earth s annual motion , and the directional rate , i.e its dependence on the direction of the recoiling nucleus . in this paper
we do this using various velocity distributions , isothermal ( symmetric as well as only axially asymmetric ) and non isothermal ( e.g. due to caustic rings ) . |
let @xmath16 be fixed .
we say that a holomorphic function @xmath17 having power series expansion @xmath18 belongs to the _ anisotropic weighted dirichlet space _ @xmath19 if @xmath20 these spaces have been considered by a number of mathematicians , see for instance @xcite .
we refer the reader to these papers , and the references therein , for further background material , and only give a brief summary of some facts we shall need later on .
as is pointed out in these references , for @xmath21 with @xmath22 and @xmath23 , the spaces can be furnished with the equivalent norm @xmath24 where @xmath25|^2da_{\alpha_1}(z_1)+ \int_{\mathbb{d}}|\partial_{z_2}[f(0,z_2)]|^2da_{\alpha_2}(z_2)\\&+ \int_{\mathbb{d}^2}|\partial_{z_2}\partial_{z_1}f(z_1,z_2)|^2da_{\alpha_1}(z_1)da_{\alpha_2}(z_2 ) .
\label{normintegrals}\end{aligned}\ ] ] here , for @xmath26 , we have set @xmath27 , where @xmath28 denotes normalized area measure .
more compactly , we have @xmath29 for all choices of @xmath21 ,
polynomials in two complex variables form a dense subset of @xmath19 . moreover , the classical one - variable weighted dirichlet spaces @xmath30 , consisting of analytic functions @xmath31 on the unit disk @xmath32 having @xmath33 embed in @xmath19 in a natural way .
these one - variable spaces are discussed in the textbook @xcite , and again admit an equivalent integral norm , @xmath34 if @xmath11 , then @xmath19 is a banach algebra of continuous functions on the closed unit bidisk ; this can be seen from the cauchy - schwarz inequality and the convergence of the series @xmath35 . setting @xmath36 ,
we are led to the hardy space @xmath37 of the bidisk , which was studied by rudin in the 60s , see @xcite .
the parameter choice @xmath38 yields the bergman space of the bidisk .
the choice @xmath39 corresponds to the dirichlet space of the bidisk , which can be characterized by the fact that pre - composition with automorphisms of the bidisk form a set of unitary operators .
this space was considered by kaptanolu @xcite , among others .
isotropic weighted dirichlet spaces , the cases with @xmath40 , were recently studied in depth in @xcite ; we shall use @xmath41 to denote these isotropic spaces .
the anisotropic spaces @xmath19 were studied by jupiter and redett @xcite , who consider order relations and identify multipliers between different @xmath19 .
for instance , they observe that @xmath42 if @xmath43 and @xmath44 . a related fact that we shall use
frequently is that @xmath45 and similarly for @xmath46 .
a _ multiplier _ of @xmath19 is a function @xmath47 that is holomorphic and satisfies @xmath48 for every @xmath49 . in the case of the hardy and bergman spaces ,
the multipliers are precisely the bounded analytic functions , but for general @xmath19 it is not as easy to describe the multiplier space @xmath50 . for our purposes it will suffice to note that any function that is analytic on a neighborhood of the closed bidisk is a multiplier on each @xmath19 .
consider the two linear operators @xmath51 defined via @xmath52 when viewed as acting on the coefficient matrix of a function @xmath53 , the operators @xmath54 and @xmath55 become right and upwards translations , justifying the designation _
shift operators_. it is clear that the coordinate shifts commute , and , in view of , they are bounded on @xmath19 . coordinate shifts @xmath56 acting on @xmath19 furnish a natural model of multivariate operator theory .
the structure of invariant subspaces of these operators is still rather poorly understood , even in the case of the unweighted hardy space @xmath57 , where the shifts are commuting isometries .
see rudin s book @xcite for some basic results and pathologies such as the existence of an invariant subspace containing no bounded elements , as well as @xcite and the references therein for some positive results , such as a conditional version of beurling s theorem on invariant subspaces .
it is easy to exhibit invariant subspaces of @xmath19 .
for instance , fixing a function @xmath49 , we can form the _ cyclic subspace _
@xmath58_{{{\vec{\alpha}}}}=\mathrm{clos}\,\,\mathrm{span}\{z_1^kz_2^lf\colon k , l\geq 0\ } \label{e : cyclicsubspace}\ ] ] which is invariant under @xmath59 by definition ; the closure is taken with respect to the @xmath19 norm .
another class of invariant subspaces is given by _
zero - based subspaces _ : a simple example is the subspace of functions divisible by the polynomial @xmath60 .
it is a much more difficult task to obtain a concrete description of general invariant subspaces of @xmath19 and their elements , even in the simplest case of cyclic subspaces . in this paper
, we are primarily interested in identifying _ cyclic vectors _ for the coordinate shifts : functions @xmath49 such that @xmath61_{\vec{\alpha}}={{\mathfrak d}_{\vec{\alpha}}}$ ] .
this seems like a hard problem for general functions , and we restrict our attention to the case where @xmath53 itself is a polynomial in two variables . in what follows , we shall use the letter @xmath7 to indicate that we are dealing with a fixed polynomial . in that setting , we are able to give a complete characterization , extending the corresponding result in @xcite to the anisotropic setting .
the cyclicity of a function @xmath49 is intimately connected with its vanishing properties .
the constant function @xmath62 is cyclic in all @xmath19 ; this is just a reformulation of the fact that polynomials are dense .
similarly , functions that are holomorphic on a neighborhood of the bidisk , and are non - vanishing on @xmath63 are cyclic for all @xmath19 . at the other extreme ,
the polynomial @xmath64 is clearly not cyclic , as elements of @xmath65_{{{\vec{\alpha}}}}$ ] have to vanish on the set @xmath66 .
more generally , as a consequence of boundedness of point evaluation functionals on @xmath19 ( see @xcite ) , no function that vanishes on the interior of the bidisk can be cyclic .
the case of zeros on the boundary is subtler , and polynomials that vanish on the boundary of the bidisk remain cyclic provided their zero sets in the torus are not too large , relative to the parameter @xmath21 . in @xcite , a complete classification of cyclic polynomials in the isotropic spaces @xmath41 was found in terms of conditions on @xmath67 , where @xmath68 is the zero set of a polynomial @xmath69 .
earlier , neuwirth , ginsberg , and newman @xcite had shown that all polynomials that do not vanish in @xmath70 are cyclic in @xmath71 , and hence in all @xmath19 that contain @xmath72 ; see also gelca s paper @xcite .
the purpose of this paper is to extend the classification result of @xcite to the anisotropic setting .
part of the proof in that paper was based on @xmath73-capacities and cauchy integrals , and does not generalize in an obvious way to the general setting where one of the components in @xmath74 may be negative . at the same time , we show how the arguments in @xcite which relied on prior work of knese and others ( see @xcite and the references in those papers ) on polynomials having determinantal representations can be replaced by integral estimates . while the former theory is elegant , our approach is more direct , and applies in the case of negative parameters as well ; cf ( * ? ? ?
* theorem 3.1 ) , where it is assumed that @xmath75 .
we now state the main result of this paper .
[ t : maintheorem ] let @xmath7 be an irreducible polynomial , depending on both @xmath4 and @xmath5 , with no zeros in the bidisk . 1 . if @xmath76 then @xmath7 is cyclic in @xmath77 .
2 . if @xmath8 and @xmath78 , then @xmath7 is cyclic in @xmath77 if and only if @xmath79 is empty or finite .
3 . if @xmath80 then @xmath7 is cyclic in @xmath77 if and only if @xmath79 is empty . the parameter regions in theorem [ t : maintheorem ] are illustrated in figure [ alphaplanefig ] .
as is explained in @xcite , the set @xmath67 associated with an irreducible polynomial @xmath7 is either a curve or a finite number of points .
the requirement that @xmath53 be irreducible is not a serious restriction . indeed ,
since all polynomials are multipliers , a product of polynomials is cyclic precisely when all its factors are . in the formulation of the theorem
the assumption that @xmath53 depends on both variables is made to avoid complications created by one variable polynomials .
for instance , the function @xmath81 is cyclic in @xmath82 precisely when @xmath83 .
[ r : tensorrem ] a new feature that appears in the anisotropic framework is that the degrees of smoothness / roughness of the space in the two coordinate directions interact to determine whether a polynomial is cyclic or not .
for instance , according to theorem [ t : maintheorem ] , the two - variable polynomial @xmath84 is cyclic in @xmath85 , as is the one - variable polynomial @xmath86 .
by contrast , the function @xmath87 is _ not _ cyclic in @xmath85 because @xmath88 is not cyclic in @xmath89 .
in this section , we first record some facts concerning one - variable functions in @xmath30 .
both the results and the arguments used to establish them will be used in our subsequent two - variable proofs .
we then discuss two representative examples that illustrate the contents of theorem [ t : maintheorem ] .
let us consider the polynomial @xmath90 and investigate its properties as an element of @xmath19 .
its zero set in @xmath91 is rather large : it can be represented as @xmath92 nevertheless , @xmath7 is cyclic in all @xmath19 having @xmath93 .
this follows from the cyclicity of @xmath94 in the classical dirichlet space on the disk originally established by brown and shields @xcite .
[ brownshields ] if @xmath95 is a polynomial with no zeros in @xmath32 , then @xmath95 is cyclic in @xmath30 for @xmath96 .
we give an alternate proof of this theorem , using dilations of @xmath97 and the following integral estimate of forelli and rudin ( see ( * ? ? ? * theorem 1.7 ) ) : for @xmath98 and @xmath99 , @xmath100 as @xmath101 .
this estimate will be useful later on in the paper as well .
the _ radial dilation _ of a function @xmath102 is defined for @xmath103 by @xmath104 .
[ onevardilation ] let @xmath95 be a polynomial with no zeros in @xmath32 .
then @xmath105 weakly in @xmath106 as @xmath107 . if @xmath108 , then @xmath105 in norm in @xmath109 as @xmath107 .
we consider first the case @xmath110 .
it suffices to establish that @xmath111 is bounded in the dirichlet norm .
indeed , if this holds , then the family @xmath111 is relatively weakly compact , and since @xmath112 pointwise , weak convergence follows .
( see @xcite for a comprehensive discussion of convergence concepts . )
every @xmath95 can be factored into linear factors , and it is not hard to see that every such factor @xmath113 satisfies the estimate @xmath114 for @xmath115 .
thus , in showing that @xmath111 is bounded in the dirichlet norm , it suffices to treat the case where @xmath116 .
also , we may as well as assume that the zero of @xmath95 lies on @xmath117 , the other case ( when the zero lies outside @xmath118 ) being obvious .
finally , by rotation invariance , we can suppose that @xmath119 .
computing the dirichlet integral of @xmath111 , and invoking the estimate with @xmath120 and @xmath121 , we find that @xmath122 and the proof for @xmath110 is complete .
the case @xmath108 may be treated in a similar way .
alternatively one can remark that , since the inclusion @xmath123 is a compact linear map , weak convergence in @xmath106 carries over to norm convergence in @xmath109 .
recall that @xmath124 $ ] is the smallest closed invariant subspace of the dirichlet space @xmath109 that contains a given @xmath95 . by definition
, it contains all functions of the form @xmath125 , where @xmath126 is a polynomial .
it also contains all functions of the form @xmath127 , where @xmath128 is holomorphic on a neighborhood of @xmath118 , because the taylor polynomials of @xmath128 converge to @xmath128 in the multiplier norm of @xmath109 .
we therefore have @xmath129 $ ] for each @xmath103 .
we have shown that @xmath112 weakly in @xmath109 for all @xmath130 . as @xmath124 $ ]
is weakly closed in @xmath109 , it follows that @xmath131 $ ] , and hence @xmath124=d_\alpha$ ] , as desired .
we return to the two - variable setting and the product function @xmath132 .
since the @xmath19-norm restricted to functions depending on the first variable only coincides with the norm in @xmath133 , and the @xmath19-norm reduces to the norm in @xmath134 for function of @xmath5 only , theorem [ brownshields ] implies that at least one of the factors of @xmath7 is cyclic precisely when @xmath135 .
they are both cyclic if @xmath136 , and since the product of cyclic multipliers is itself cyclic , the result follows . to illustrate what happens when the zero set is infinite ,
let us consider the polynomial @xmath137 , which does not vanish in the bidisk or on @xmath138 , and has @xmath139 a curve in the distinguished boundary . the same kind of reasoning as in ( * ?
* section 3 ) reveals that @xmath7 is cyclic in @xmath19 precisely when @xmath140 .
we first recall that an equivalent criterion for a function @xmath49 to be cyclic is the existence of a sequence @xmath141 of polynomials in two variables such that @xmath142 for then the cyclic element @xmath143_{{{\vec{\alpha}}}}$ ] . by orthogonality , the expression @xmath144 is minimized by taking @xmath141 to be polynomials in @xmath145 .
next , it is elementary to see that @xmath146 for functions @xmath49 of the form @xmath147 , where @xmath148 is a function on the unit disk .
but the polynomial @xmath94 is cyclic in @xmath30 if and only if @xmath149 .
hence @xmath84 is cyclic precisely when @xmath150 .
we now consider a polynomial that does depend on both variables , but whose zero set in the distinguished boundary is minimal .
the polynomial @xmath151 is cyclic for @xmath19 whenever @xmath9 .
we note that @xmath152 , so that @xmath7 can certainly not be cyclic in @xmath19 when @xmath11 .
this example will follow from section [ sec : finite ] , where we rehash an argument from @xcite . in appendix
a , we offer an elementary proof of this fact which may be of independent interest .
we now turn to the proof of theorem [ t : maintheorem ] .
the case of a polynomial that does not vanish on the closed bidisk is trivial , so we exclude it .
our first step is to address finitely many zeros on @xmath10 .
let @xmath153 $ ] have no zeros in the open bidisk , and finitely many zeros on @xmath10 .
then @xmath7 is cyclic in @xmath19 for @xmath78 .
the proof is a very slight modification of a corresponding argument in section 3 of @xcite . without loss of generality ,
assume @xmath154 , so that in particular @xmath155 .
the basic idea is to compare our polynomial to a product of factors of the form @xmath156 , which are known to be cyclic .
write @xmath157 for some @xmath158 ; each @xmath159 with @xmath160 .
we now recall _ ojasiewicz s inequality _ , a classical result in real algebraic geometry @xcite : if @xmath53 is a real analytic function on an open set @xmath161 , and @xmath162 is compact , then there exist a constant @xmath163 and a number @xmath164 such that @xmath165 apply to the function @xmath166 , which is real - analytic on @xmath167 , and the compact set @xmath168 to see there exist @xmath169 such that @xmath170 thus , @xmath171 is bounded on @xmath10 and if we increase @xmath126 we can make this function as smooth as we like on @xmath10 . but then @xmath172 because of rapid decay of fourier coefficients of @xmath113 , so that in turn @xmath173
. now @xmath7 is a multiplier , hence @xmath174_{{{\vec{\alpha}}}}$ ] , and @xmath128 is cyclic in @xmath19 for @xmath155 , since it is a product of cyclic multipliers .
hence @xmath175_{{{\vec{\alpha}}}}$ ] contains a cyclic element for @xmath19 .
this section is devoted to proving that a polynomial with no zeros in @xmath13 and infinitely many zeros in the @xmath14-torus is cyclic in @xmath19 exactly when @xmath176 .
the proof of cyclicity when @xmath177 and the proof of non - cyclicity when @xmath178 have the interesting feature that they in some sense reduce the problem to the model polynomial @xmath179 . by the previous section
we only need to address polynomials with infinite zero set on @xmath10 ( necessarily forming a curve ) but the proof below does not use this in an essential way .
[ t : raddiltwovars ] suppose @xmath153 $ ] is irreducible , has no zeros in the bidisk , and is not a polynomial in one variable only .
then @xmath7 is cyclic in @xmath19 whenever @xmath140 .
we will use @xmath179 as a model polynomial , " analogous to @xmath88 in the one - variable setting .
the crux of the proof of theorem [ t : raddiltwovars ] is to establish the following analog of lemma [ onevardilation ] ; see also @xcite .
[ twovardilation ] let @xmath140 .
then @xmath180 weakly in @xmath19 as @xmath181 .
[ sides ] suppose a polynomial @xmath153 $ ] is irreducible , has no zeros on @xmath13 , and is not a polynomial in one variable only .
then @xmath7 has no zeros on @xmath182 . for fixed @xmath183
, @xmath184 has no zeros in @xmath32 . by hurwitz s theorem , if we send @xmath185 , then @xmath186 either has no zeros in @xmath32 , or is identically zero .
if it is identically zero , then @xmath187 divides @xmath3 , contrary to our assumptions .
thus , for each @xmath188 , @xmath184 has no zeros in @xmath32 . by symmetry
, @xmath7 has no zeros on @xmath189 either .
assuming lemma [ twovardilation ] , cyclicity of @xmath7 in @xmath19 with @xmath140 follows as before . to see this , set @xmath190 . by lemma [ sides ] , the polynomial @xmath191 does not vanish on the closed bidisk for @xmath192 .
thus each @xmath193 extends holomorphically past the closed bidisk , and hence is a multiplier , which in turn implies @xmath194_{{{\vec{\alpha}}}}$ ] for each @xmath192 . thus @xmath195 $ ] , as a limit of @xmath196 , and we are done .
before we give the proof of lemma [ twovardilation ] , we make some preliminary remarks , and a few reductions .
suppose that the given polynomial @xmath7 has bidegree @xmath197that is , degree @xmath198 in the variable @xmath4 and degree @xmath199 in the variable @xmath5 .
first of all , we can view @xmath7 as a polynomial in the variable @xmath4 , @xmath200 with coefficients @xmath201 that are polynomials in @xmath5 .
since @xmath7 does not vanish on the bidisk we infer that @xmath202 does not vanish on the disk .
in fact , @xmath202 does not vanish on the unit circle , for if it did then @xmath7 would vanish on @xmath203 .
such a zero is ruled out by the assumptions that @xmath7 is irreducible and depends on both variables via lemma [ sides ] . for each fixed @xmath5
where @xmath204 , we can therefore factor @xmath7 into linear factors in the variable @xmath4 , @xmath205 before we proceed , we need to discuss the nature of the functions that appear in the right - hand side : the @xmath206 require a particularly careful treatment . the function @xmath207 is a polynomial in one complex variable with no zeros in the closed unit disk , and hence @xmath202 is a multiplier and cyclic in every space @xmath19 .
we can of course reverse the roles of @xmath4 and @xmath5 in the factorization and write for each fixed @xmath4 ( outside a finite set ) @xmath208 and again obtain a non - vanishing polynomial @xmath209 furnishing a cyclic multiplier . since the product @xmath210 of a function @xmath49 and a multiplier @xmath211 is cyclic if and only if both factors are cyclic ( * ? ? ?
* proposition 8) , it is enough to establish cyclicity of the factor that depends on both variables .
namely , we can drop the factors @xmath202 or @xmath212 . in summary , we may assume without loss of generality that , outside of a discrete set , the polynomial @xmath7 is locally of the form @xmath213 the functions @xmath214 , @xmath215 , are more problematic as they are no longer polynomials ( or even single - valued ) , but can be represented as branches of algebraic functions . in particular , the factors @xmath216 are in general not elements of @xmath19 .
[ e : branchpoints ] for each @xmath217 , the zero set of the irreducible polynomial @xmath218 meets @xmath219 in a curve in @xmath10 . for this polynomial
, we have the local factorization @xmath220 and hence @xmath206 , @xmath221 , have a branch point at @xmath222 . to get around the difficulty posed by the nature of the functions @xmath223 , we let @xmath224 be a simply connected subdomain of @xmath32 such that each @xmath206 extends holomorphically to @xmath225 , and @xmath226 ( here @xmath227 denotes 2-dimensional lebesgue measure ) . more precisely , first notice that we allow some @xmath228 to equal zero but this necessarily occurs at the finitely many @xmath5 where @xmath229 .
we can think of @xmath230 as a root of @xmath7 in such cases .
now define @xmath231 this set is necessarily finite because @xmath7 is irreducible .
the first set in the union consists of the points where @xmath232 has a repeated root and the second set consists of the points where @xmath232 has a repeated root `` at @xmath233 . '' on @xmath234 , there are distinct @xmath235 which can be defined locally in an analytic fashion and satisfy . given @xmath236 form the ray @xmath237 ,
let @xmath238 , and finally define @xmath239 on @xmath225 , the @xmath223 can be analytically continued to single - valued analytic functions satisfying .
notice that across the boundary slits of @xmath225 the @xmath223 can be analytically continued necessarily to some @xmath206 .
near a point of @xmath240 , two of the @xmath223 tend to the same value ( the case of @xmath241 means that two of the @xmath223 tend to zero at @xmath5 as in example [ e : branchpoints ] ) .
the @xmath206 will extend analytically to any point of @xmath242 .
we may assume @xmath243 ; if this is not the case , we can replace @xmath3 by @xmath244 , where @xmath245 is a mbius transformation of the unit disk .
cyclicity of one of these two functions implies cyclicity of the other .
example [ e : branchpoints ] , where @xmath246 ( @xmath120 ) is transformed into @xmath247 ( @xmath248 ) in precisely this way .
[ l : hfunctlemma ] let @xmath249 be as above .
then @xmath250 in @xmath225 , @xmath223 has bounded multiplicity , and @xmath251 .
if @xmath252 for some @xmath253 then the polynomial @xmath7 would vanish inside @xmath13 , contradicting our assumptions .
hence @xmath250 for all @xmath253 , and for every @xmath254 .
furthermore , if for some @xmath222 it were the case that @xmath255 for more than @xmath199 values of @xmath5 , then @xmath256 would have more than @xmath199 zeros , which is impossible since @xmath7 is irreducible of bidegree @xmath197 .
next , put @xmath257 by the discussion before the lemma , this is well - defined in a neighborhood of @xmath118 minus @xmath240 since each @xmath223 analytically continues across the boundary slits of @xmath225 and to @xmath242 .
thus @xmath258 is subharmonic on @xmath259 and extends to be subharmonic on @xmath118 because the points of @xmath240 will be removable singularities .
clearly @xmath260 , @xmath261 .
according to the hopf lemma for subharmonic functions ( see ( * ? ? ?
* proposition 12.2 ) ) there is @xmath163 such that @xmath262 the last assertion of the lemma follows . finally , we shall need the following lemma .
[ l : changevar ] let @xmath263 be a domain , and suppose @xmath264 is an analytic map of multiplicity at most @xmath265 .
then , for any non - negative function @xmath266 , we have @xmath267 see ( * ? ? ?
* lemma 6.2.2 ) .
since the @xmath223 appearing in lemma [ l : hfunctlemma ] are of bounded multiplicity , we apply lemma [ l : changevar ] to @xmath268 , and will do so frequently in what follows . to begin our proof of lemma [ twovardilation ] , we set , for @xmath103 , @xmath269 where @xmath270 since pointwise convergence of @xmath271 as @xmath181 clearly holds , it suffices , as in the proof of lemma [ onevardilation ] , to show that the @xmath19-norm of @xmath196 remains bounded as @xmath272 .
[ model ] @xmath273 is bounded by a constant for @xmath274 . by the integral
is bounded by a constant times @xmath275 and the above integral is equal to a constant times @xmath276 .
we record some qualitative features of the derivatives of @xmath196 .
[ qual ] the function @xmath277 is bounded in @xmath13 independent of @xmath278 . for fixed @xmath254 , the derivatives @xmath279 and @xmath280 are of the form @xmath281 for some polynomial @xmath282 .
consequently , the derivatives @xmath283 can be written as a finite sum of terms of the form @xmath284 where * @xmath285 is bounded on @xmath286 * for some integer @xmath287 , @xmath288 , @xmath289 equals the product of @xmath290 and @xmath291 terms of the form @xmath292 for multiple choices of @xmath293 , * in case ( a ) , @xmath294 and @xmath295 is unnecessary and in case ( b ) @xmath295 is of the form @xmath296 for some @xmath293 .
assuming @xmath297 and @xmath298 , and @xmath299 , we have the estimate @xmath300 valid in @xmath301 for some constant @xmath163 . the formula for @xmath302 shows @xmath303 .
the other formulas are calculus exercises . to get the final estimate let @xmath304 . then , @xmath305 is of the form @xmath306 terms of the form @xmath307 appear in the denominator @xmath308 times .
we use the estimate @xmath309 for @xmath310 of the terms .
evidently , @xmath311 and by lemma [ l : hfunctlemma ] we have @xmath312 .
using either of these two estimates for the remaining @xmath313 terms in the denominator of @xmath305 gives the bound @xmath314 the assumptions on the parameters mean @xmath315 with both summands non - negative .
this immediately yields @xmath316 for some constant @xmath163 . combining this with the definition of @xmath295
, we get @xmath317 since @xmath318 can be bounded by a finite sum of such terms the main estimate holds . in view of our construction of the domain @xmath225
, the seminorm @xmath319 is not affected if we restrict integration to @xmath301 .
note that the integral formula for @xmath319 is valid if both @xmath22 and @xmath23 .
estimate can be used with @xmath320 if we assume @xmath321 .
certainly , since @xmath176 we can assume without loss of generality that @xmath322 .
later on we will adjust the argument in case @xmath323 .
the problematic term in @xmath319 is the integral @xmath324 by with @xmath320 , this is bounded by @xmath325 by lemma [ l : changevar ] .
the integral on the right is bounded by a constant by lemma [ model ] .
this proves @xmath319 is bounded independent of @xmath103 when @xmath321 .
suppose now that one of the parameters , say @xmath326 , satisfies @xmath327 .
we then use that @xmath49 if and only if @xmath328 and @xmath329 to switch from @xmath196 to @xmath330 , where @xmath331 is chosen so that @xmath332 . to bound the @xmath19-norm of @xmath196
, it is now enough to bound the norms @xmath333 for @xmath334 , the one - variable norms @xmath335 and @xmath336 , and the double integral @xmath337 by case ( a ) of lemma [ qual ] , @xmath338 is bounded ; this takes care of the contribution @xmath339 . by case ( b ) of lemma [ qual ]
, @xmath340 can be controlled by terms of the form @xmath341 .
but , since @xmath342 @xmath343 which is finite .
by case ( a ) of lemma [ qual ] , to control @xmath344 it suffices to control @xmath345 , which by can be bounded by @xmath346 where @xmath347 .
now this is enough to show that the integral @xmath348 .
finally , can be bounded using with @xmath349 in the same way we bounded .
this concludes the proof of lemma [ twovardilation ] and consequently theorem [ t : raddiltwovars ] .
we now complete the proof of our main theorem by proving that an irreducible polynomial that has no zeros in @xmath13 but vanishes along a curve in @xmath10 fails to be cyclic in certain @xmath19 . in the sequel we lose no generality
assuming that @xmath350 .
[ t : infinitezerosthm ] let @xmath153 $ ] be an irreducible polynomial depending on both variables with infinitely many zeros in @xmath10 . then @xmath7 is not cyclic in @xmath19 whenever @xmath8 .
it is clear that the polynomial @xmath7 is not cyclic if it vanishes in the open bidisk , so we exclude that case in what follows .
then the idea of the proof is again to argue by reduction to the polynomial @xmath179 : we show that if the polynomial @xmath7 is cyclic , then so is @xmath179 , leading to a contradiction when @xmath8 . to this end , let @xmath7 be irreducible , depending on both variables with infinitely many zeros in @xmath10 and no zeros on @xmath13 .
then , as is pointed out in ( * ? ? ?
* section 2.2 ) , we must have @xmath351 where @xmath352 .
( this is not so in the case of _ finitely many _ zeros : @xmath353 vanishes at @xmath354 for example . )
this follows from the fact that the irreducible polynomial @xmath7 will have infinitely many zeros in common with @xmath355 which implies these polynomials are constant multiples of one another by bzout s theorem .
thus , @xmath7 has no zeros in @xmath356 . by lemma [ sides ]
, @xmath7 has no zeros in @xmath357 and by reflection no zeros in @xmath358 .
this leaves only the set . as in the previous section ,
we write @xmath359 and @xmath360 note that for every @xmath361 , there is a @xmath362 such that @xmath363 this holds because each root of @xmath7 can be written in the form @xmath364 ( outside a finite set of values of @xmath4 and @xmath5 ) .
now put @xmath365 , and note that implies that @xmath366 for @xmath367 , where @xmath240 is the set defined in the previous subsection .
applying puiseux s theorem in the form of lemma [ rootglue ] in appendix b to @xmath368 we obtain bounds on the derivatives of @xmath369 on a neighborhood of the closed unit disk .
namely , for @xmath370 , @xmath371 from the equalities we deduce that @xmath372 where @xmath282 is a function holomorphic on a neighborhood of @xmath63 .
[ l : symmetrization ] let @xmath373 be a sequence of polynomials in two complex variables , set @xmath374 , and suppose @xmath375 as @xmath376 .
then , for any @xmath377 , there are @xmath378 such that for any permutation @xmath379 , we have the seminorm estimate @xmath380 assuming lemma [ l : symmetrization ] , we now present the proof of theorem [ t : infinitezerosthm ] .
suppose @xmath7 is cyclic .
then there is a sequence of polynomials @xmath381 such that @xmath382 . put @xmath383 and let @xmath384 denote the symmetric group of order @xmath199 .
define @xmath385 by lemma [ l : symmetrization ] , @xmath386 can be brought arbitrarily close to @xmath387 in @xmath19 through a suitable choice of @xmath388 .
we note that @xmath389 , where @xmath390 is holomorphic on a neighborhood of @xmath63 and hence a multiplier .
now @xmath391 can in turn be approximated in multiplier norm by polynomials , and this shows that @xmath392 is a cyclic function .
then also @xmath179 is cyclic since a product of multipliers is cyclic if and only if each factor is cyclic .
but as we have seen , @xmath179 is only cyclic when @xmath140 , and the theorem follows .
we are left with the proof of the lemma .
this will require the following one - variable estimate .
[ l : quotientderivative ] let @xmath393 and @xmath394 .
then there is a constant @xmath395 depending only on @xmath396 and @xmath397 such that for any @xmath398 , @xmath399 we assume that @xmath400 belongs to the dirichlet - type space @xmath401 ; otherwise , the inequality is trivial .
we may also assume that @xmath402 .
applying the cauchy - schwarz inequality to the expression @xmath403 , we have @xmath404 by fubini s theorem , it now suffices to show that the series @xmath405 converges .
we recall the known asymptotics of the beta function , valid for @xmath406 , @xmath407 moreover @xmath408 applying jensen s inequality , bearing in mind that @xmath393 , we get @xmath409 and it follows that converges .
we proceed inductively .
the base case @xmath410 follows directly from lemma [ l : changevar ] .
let us address the inductive step .
set @xmath411 and @xmath412 our task is to show that by choosing @xmath413 large enough , we can bring @xmath414 arbitrarily close to @xmath415 .
let @xmath331 be such that @xmath332 . to bound the @xmath19-norm of @xmath416 ( where potentially @xmath417 ) we estimate the norms @xmath418 for @xmath334 , the two one - variable norms @xmath419 and @xmath420 , and the double integral @xmath421 let us present the details for the double integral the proofs for the the one - variable integrals follow along the same lines .
computing @xmath422 we see that it comprises the following terms : * @xmath423 , * @xmath424 , where @xmath425 and @xmath426 .
* @xmath427 , where @xmath425 , we shall show that ( ii ) and ( iii ) tend to @xmath428 in @xmath429 as @xmath376 .
moreover , we shall show the integral of the modulus squared of ( i ) with respect to the measure @xmath430 tends to @xmath431 which is then controlled by the induction hypothesis .
we first turn to ( iii ) .
since @xmath432 is bounded on @xmath13 we need to estimate @xmath433 which we do by changing variables ( lemma [ l : changevar ] ) and invoking the hypothesis that @xmath434 in @xmath19 as @xmath435 .
we next continue with terms ( i ) and ( ii ) simultaneously .
since @xmath436 tends to @xmath387 , we first replace @xmath437 with @xmath438 , and we show that after performing this replacement , the resulting terms converge to @xmath428 in @xmath439 .
note that the functions @xmath440 need not be bounded .
however , each of them can be written as a sum of products of a function that is bounded on @xmath13 , and the functions @xmath441 , @xmath442 .
hence , it suffices to show that @xmath443 tends to @xmath428 as @xmath376 if @xmath444 .
note that the set of singularities @xmath445 of the functions @xmath369 , @xmath446 , is finite .
invoking the estimate at a point @xmath370 to bound @xmath447 leads us to integrals of the form @xmath448 where @xmath449 and @xmath450 .
we distinguish between two cases : @xmath222 and @xmath451 .
we first deal with the case when @xmath396 is in the boundary of @xmath32 . applying a rotation if necessary
, we can take @xmath452 and use lemma [ l : quotientderivative ] ( keeping @xmath4 fixed ) to get @xmath453 we now integrate over @xmath32 with respect to @xmath454 , and since the right - hand side tends to zero @xmath413 , we obtain the desired conclusion .
now let us consider the case when @xmath396 lies in the open unit disk .
we can cover a neighborhood of @xmath455 with a finite family of relatively compact disks , and estimate the integral over any such disk @xmath456 .
this leaves us with two types of integrals , one over @xmath457 where @xmath458 , meaning that it suffices to bound @xmath459 and the contribution that arises from integrating over @xmath456 , @xmath460 a bound on follows from ( with @xmath461 ) .
since @xmath462 is relatively compact in @xmath32 and @xmath463 is bounded , after an affine change of variables in , we can bound this integral by @xmath464 where @xmath465 is of bounded multiplicity and @xmath466 . that integral , according to lemma [ l : quotientderivative ] , can be estimated by @xmath467 since @xmath468 is relatively compact in @xmath32 we can insert the weight @xmath469 in the above integral . changing variables as in lemma
[ l : changevar ] , we bound by @xmath470 which we again have control over .
this concludes the proof of lemma [ l : symmetrization ] .
here we give a direct proof that @xmath151 is cyclic for @xmath19 whenever @xmath9 .
let @xmath53 be a function in @xmath19 orthogonal to @xmath471 for all @xmath472 .
we need to show that @xmath473 .
if we write @xmath474 then the orthogonality condition becomes @xmath475 and the condition that @xmath53 belongs to @xmath19 translates to @xmath476 our goal is thus prove that conditions and together imply that @xmath477 for all @xmath478 .
we can suppose without loss of generality that @xmath479 and @xmath480 .
define a new function @xmath128 by @xmath481 condition ensures that @xmath128 is holomorphic on @xmath13 .
also condition implies the identity @xmath482 which , after rearrangement , becomes @xmath483 consider now the substitutions @xmath484 and @xmath485 .
note that we have @xmath486 and @xmath487 . substituting these values of @xmath488 into , we find that @xmath489 thus , if we define @xmath490 by @xmath491 then @xmath492 is a well - defined entire function .
note also that @xmath493 both these series are finite , by , so the sum of the two integrals is finite , and consequently @xmath494 where @xmath495 .
this forces @xmath492 to be a polynomial .
thus , if @xmath496 , then @xmath497 for all large @xmath498 , and substituting this information back into the integrals already known to be finite , we get @xmath499 and @xmath500 which implies that @xmath501 and @xmath502 , contrary to the hypothesis that @xmath9 .
we conclude that @xmath503 .
it follows that @xmath504 and @xmath505 , whence also @xmath506 by and so , finally , @xmath477 for all @xmath478 .
the proof shows that , if @xmath9 , then the conditions @xmath507 and @xmath508 together imply that @xmath477 for all @xmath472 .
[ rootglue ] suppose @xmath509 is holomorphic in a domain @xmath510 .
if @xmath511 for a domain @xmath512 , and satisfies @xmath513 , then @xmath514 as @xmath515
. we may assume @xmath516 and @xmath517 .
losing no generality we may demand that an analytic set @xmath518 is irreducible in a neighborhood of @xmath428 ( otherwise we may decrease @xmath199 ) . applying puiseux parametrization ( see @xcite , corollary , page 171 ) we find that there is an analytic germ in a neighborhood of @xmath428 such that @xmath519 in a neighborhood of @xmath520 .
this means that @xmath521 , where the branch of square is properly chosen . from this one
can immediately derive the assertion . | consider the dirichlet - type space on the bidisk consisting of holomorphic functions @xmath0 such that @xmath1 here the parameters @xmath2 are arbitrary real numbers .
we characterize the polynomials that are cyclic for the shift operators on this space .
more precisely , we show that , given an irreducible polynomial @xmath3 depending on both @xmath4 and @xmath5 and having no zeros in the bidisk : * if @xmath6 , then @xmath7 is cyclic ; * if @xmath8 and @xmath9 , then @xmath7 is cyclic if and only if it has finitely many zeros in the two - torus @xmath10 ; * if @xmath11 , then @xmath7 is cyclic if and only if it has no zeros in @xmath10 . |
for a number of years , the more so since the discovery of high-@xmath5 superconductors@xcite , the study of strongly correlated electron systems has been a major theme of theoretical condensed matter physics .
the @xmath1 model is one of the simplest of such models@xcite . although high-@xmath5 cuprates are at least two - dimensional ( @xmath6 ) systems , it is also relevant to fully understand the one - dimensional ( @xmath0 ) model .
the hamiltonian for the @xmath1 model in one dimension can be written in the subspace of no doubly occupied sites as @xmath7 where @xmath8 and @xmath9 are the creation and annihilation operators for an electron at lattice site @xmath10 with spin @xmath11 , while @xmath12 and @xmath13 are the corresponding spin and electron number operators .
this model has been solved exactly only for @xmath14 , where it is equivalent to the @xmath15 hubbard model , and at the supersymmetric point@xcite @xmath16 . in both cases
the ground state at arbitrary density belongs to a broad class of interacting fermi systems known as luttinger liquids@xcite , which exhibit power - law decay of correlation functions with exponents characterized by a single parameter@xcite .
additionally , for very large @xmath17 , the attractive heisenberg interaction term dominates the kinetic energy and the model phase separates@xcite@xmath18@xcite .
the level of understanding of this model is derived mostly from small cluster exact diagonalizations@xcite , variational monte carlo methods@xcite and finite temperature monte carlo simulations for relatively larger system sizes@xcite .
there still lurks , however , the question of whether the thermodynamic limit has been reached or not . in this paper
we use the density matrix renormalization group ( dmrg ) method@xcite to study the ground - state properties of the @xmath1 model in the thermodynamic limit .
the text is organized as follows . in sec .
ii , we review the formulation of dmrg and its extension to arbitrary band - filling . in sec .
iii , we discuss our numerical results .
we present ground - state static spin and charge correlation functions as well as electron momentum distribution functions .
finally in sec .
iv we summarize our results .
the dmrg technique was developed by white@xcite in 1992 .
it leads to highly accurate results for systems much larger than those which can be solved by exact diagonalization .
the dmrg allows a systematic reduction of the hilbert space to the basis states most relevant to describe a given eigenstate ( @xmath19 the ground state ) of a large system .
this is to be contrasted with previous real space renormalization techniques in which the lowest states are kept .
a general iteration step of the method for open boundary conditions proceeds as follows : @xmath20 the effective hamiltonian defined for the superblock @xmath21 ( where the block 1 and @xmath22 come from previous iterations and 2 and @xmath23 are new added ones ) is diagonalized to obtain the ground - state wave function @xmath24 ( other states could be also kept ) .
@xmath25 the reduced density matrix of blocks @xmath26 : @xmath27 is constructed , where @xmath28 , the states @xmath29 ( @xmath30 ) belongs to the hilbert space of blocks 1 and 2 ( @xmath22 and @xmath23 ) .
the eigenstates of @xmath31 with the highest eigenvalues ( equivalent to the most probable states of blocks @xmath26 in the ground state of the superblock ) are kept up to a certain cutoff .
@xmath32 these states form a new reduced basis in which all the operators have to be expanded and the block @xmath26 is renamed as block 1 .
@xmath33 a new block 2 is added ( one site in our case ) and the new superblock @xmath34 is formed as the direct product of the states of all the blocks ( the blocks @xmath22 and @xmath23 are identical to blocks 1 and 2 respectively ) .
the method has been successfully applied to problems such as the haldane gap of spin-1 chains , critical exponents of spin-@xmath35 chains , the @xmath0 kondo - insulator and two - chain hubbard model@xcite . in our study of the @xmath0 @xmath1 model
we have used the infinite - size version of the above iteration scheme to reach the thermodynamic limit .
one immediately realizes , however , there is a problem in keeping the electron density fixed in the iteration process since we insert only two sites at each iteration ( which only makes the half - filling and quarter - filling cases invariant ) . to get around this ,
we construct the reduced density matrix from two ground states that bracket the desired density .
to be more specific , if our desired electron density is @xmath36 and the current superblock lattice size is @xmath37 , then we can always find two nearest integers @xmath38 and @xmath39 such that @xmath40 . assuming @xmath41 is the ground state wavefunction of @xmath38 electrons and @xmath42 is that of @xmath39 electrons , we build the reduced density matrix by the following weighting procedures : @xmath43 it is clear that the above construction ensures the desired constant band filling at every iteration .
the success of this construction inherently requires that the ground state be homogeneous
. therefore one should expect failure when the system goes into the phase separation regime .
however , phase separation can still be studied using the finite size version of dmrg@xcite which does not require translational invariance .
it is perhaps worthwhile to remark that in our computer program we have generally targeted three states since the ground state has two - fold degeneracy when the electron number is odd .
the correlation functions are calculated when the iteration has converged .
typically we start to measure them after the superblock size reaches 40 lattice sites .
more specifically we measure the correlation between operators in the middle of the superblock .
for example , the nearest - neighbor correlation is measured between sites 20 and 21 in the superblock size 40 ; while the second - nearest - neighbor correlation is measured between sites 20 and 22 in the next iteration ( i.e when superblock size becomes 42 ) .
similarly the third - nearest - neighbor correlation is measured between sites 20 and 23 in the superblock size 44 .
thus as we iterate our procedure to increase the system size we obtain correlations over increasing distances .
finally , the desired ground - state correlations are obtained in a similar way as the reduced density matrix , @xmath19 the spin - spin correlation is measured in the following way @xmath44 before we discuss the numerical results , let us prove the following theorem : _ the eigenstates of the reduced density matrix retain their good quantum numbers when the corresponding operator of the superblock is a direct sum of the operators of its subsystems_. the electron number operator @xmath45 and the total @xmath46-component of the spin operator @xmath47 are two such operators . to prove such a theorem
all we need to show is that the reduced density matrix is block diagonal in its good quantum operator subspace .
let us consider one subspace at a time , @xmath19 the electron number subspace .
then the most general form of the eigenstate of the superblock with @xmath48 electrons can be written as @xmath49 where @xmath50 stands for the basis wavefunction of block 1 with @xmath51 electrons , and @xmath52 labels other quantum numbers , while @xmath53 electrons , and @xmath54 marks other quantum numbers .
thus the reduced density matrix elements have the form @xmath55 however according to the construction of the wavefunction we have both @xmath56 and @xmath57 leading automatically to the conclusion @xmath58 .
therefore the reduced density matrix is block diagonal in the electron number subspace . in the same way one can show that the reduced density matrix is block diagonal in the @xmath59 subspace .
we emphasize that this theorem can be explicitly implemented to accelerate the diagonalization of the reduced density matrix in the subspace of its good quantum numbers .
for most of the numerical results reported here we have kept 110 states@xcite in blocks 1 and @xmath22 , the truncation error defined as @xmath60 ( where @xmath61 is the summation of the highest @xmath62 eigenvalues of the reduced density matrix ) is of the order of @xmath63 . in table
i we list some of the ground - state energies per site @xmath64 as a function of band filling @xmath36 and spin exchange coupling @xmath65 . in the same table
we also list the results obtained from the exact bethe ansatz solution at @xmath16 .
as can be seen immediately from the table , our dmrg results are highly accurate . to make it convenient for the reader we quote the coupled bethe ansatz equations@xcite determining the ground - state energy at density @xmath36 and @xmath16 : @xmath66 @xmath67)\rho ( \nu ^{\prime } ) d\nu ^{\prime } , \ ] ] where @xmath68 denotes shiba s function @xmath69 then the ground - state energy is given by @xmath70.\ ] ] since we are going to present our results in momentum space it is now opportune to discuss the way we analyse our data . first , because open boundary conditions are used in our dmrg procedure , we obviously loose translational invariance . to overcome this boundary effect
we have done an average on the real space correlation functions before we do the fourier transform into momentum space .
more specifically we obtain several real - space correlation functions of the same system by starting the measurement at different superblock sizes .
for example we start our real - space density - density correlation @xmath71 measurement after the superblocks reach the sizes of 40 , 42 , 44 , 46 and 48 sites and then take the average among them .
the final fourier transform is done in the usual way : @xmath72,\ ] ] where @xmath73 is the largest separation available in our measurement . for the results of this work we stop at a superblock size of 150 sites . in fig .
1(a ) we plot the electron momentum distribution function for quarter filled band at @xmath74 we see that this result is almost identical to the @xmath0 infinite-@xmath75 hubbard model studied by ogata and shiba@xcite .
this is not surprising considering the fact that the small @xmath3 limit of the @xmath1 model is equivalent to the strong coupling limit of the hubbard model .
here we have both @xmath76 @xmath77 and @xmath2 anormalies as indicated by arrows in the figure . unlike normal fermi liquids , however , the @xmath2 anomaly is a new feature which is related to the effect of the coupled holon and spinon representation of the normal electron@xcite .
1(b ) shows the spin - spin correlation .
there is a @xmath78 anomaly reflecting the nature of the antiferromagnetic exchange among nearest - neighbor electrons .
1(c ) is the density - density correlation function .
here one observes the @xmath79 @xmath80 anomaly only . before discussing the effect of increasing @xmath3
let us remember the two competing factors in correlation effects : @xmath20 the nearest - neighbor antiferromagnetic spin exchange @xmath3 favors the formation of spin - singlet electron pairs on nearest - neighbor sites which in turn leads to @xmath78 spin density wave ( sdw ) fluctuations .
but eventually , when @xmath3 is large enough , there emerges a phase separation exhibiting antiferromagnetic sdw fluctuations .
@xmath25 the kinetic energy @xmath81 term favors the delocalization of the electrons .
but the pauli principle as well as on - site repulsion promote uniform separation between electrons synonymous with @xmath79 charge density wave ( cdw ) fluctuations . on the other hand the formation of bound singlet pairs due to @xmath3 term enhances @xmath78 cdw fluctuations . in fig .
2(a ) it is seen that as the antiferromagnetic exchange @xmath3 increases , the momentum distribution function is modified more drastically in the region @xmath82 where it changes from decreasing to increasing as a function of @xmath83 . notably the @xmath2 feature has been washed out for larger @xmath3 . besides , the critical exponent@xcite at the fermi wavevector @xmath76 seems to decrease as @xmath3 increases . in fig
2(b ) we have plotted the spin - spin correlation function at quarter filling with @xmath84 and 2.5 .
we see that at @xmath85 there is a clear @xmath78 anomaly , however at @xmath16 this anomaly is significantly weakened . at @xmath86
the @xmath78 anomaly seems to have completely disappeared and the maximum is located at @xmath87 .
this is an indication of singlet pairing between nearest - neighbor electrons .
2(c ) shows the corresponding density - density correlation . at @xmath85
there is only a @xmath79 anomaly which indicates the system is quite uniformly distributed .
however as one increases @xmath3 the @xmath79 feature disappears and a new feature at @xmath78 develops which reflects the formation of nearest - neighbor pair . at @xmath86
the behavior near @xmath88 has a tendency to flatten out .
but the @xmath78 feature still dominates which shows singlet pairing .
3 shows similar results for the band filling @xmath89 .
again it is seen from fig .
3(a ) that @xmath3 has the largest effect on the momentum distribution function in the regime @xmath82 . fig .
3(b ) is very much the same as of fig .
2(b ) : as @xmath3 increases the maximum in the spin - spin correlation function moves to the antiferromagnetic wavevector @xmath90 . but in fig .
3(c ) the @xmath91 density - density correlation shows a clearer tendency towards phase separation near @xmath88 .
both @xmath78 and @xmath79 anomalies are seen in the density correlation at intermediate @xmath3 as predicted@xcite . as mentioned in the previous section
our dmrg procedure is not applicable to the study of the phase separated regime .
but we know from other numerical calculations there exists a @xmath92 @xmath93 indicating the onset of a phase separation .
here we wish to argue that there exists another critical @xmath4 ( that is larger than the @xmath94 $ ] ) beyond which the system forms an @xmath95 @xmath96 , i.e. , a phase with no hole inside but having a thin interface at which the boundary electrons can still evaporate into the vacuum ( hole ) .
this @xmath4 should be independent of the band filling since once the @xmath95 @xmath96 is reached there is only a single phase boundary between the hole region and the solid , and that is density independent . by definition ,
the first instability of such an @xmath95 @xmath96 comes from the ability to dissolve a single hole inside .
accordingly @xmath4 can be estimated from finite size exact diagonalizations by comparing the ground - state energy of a @xmath97-site heisenberg chain with that of 2 holes in the @xmath97-site @xmath1 model .
this has been done by ogata et al@xcite on a 16-site system and by assaad et al@xcite with path - integral monte carlo simulation .
both results seem to point towards @xmath98 .
further , yokoyama et al@xcite have hinted the existence of such a @xmath4 in their recent variational monte carlo study of @xmath1 model .
the same thing is also implied in the work of ammon el al@xcite .
we have extended the dmrg method to allow calculations at arbitrary band filling and proven a theorem regarding the reduced density matrix that could accelerate the numerical calculation .
we then used the extended dmrg procedure to study the @xmath0 @xmath1 model .
it is found that our procedure gives highly accurate ground - state energy and correlation functions .
it is shown that the @xmath2 anomaly first discussed by ogata and shiba in the electron momentum distribution disappears as @xmath3 is increased .
it is argued that there exists a density - independent @xmath4 beyond which the system forms an @xmath95 @xmath96 .
we acknowledge the support of the natural sciences and engineering research council of canada ( nserc ) , the fonds pour la formation de chercheurs et laide la recherche from the government of qubec ( fcar ) .
we are grateful to a .-
s. tremblay and l.g .
caron for encouragements and reading of the manuscript .
.ground - state energies of the @xmath0 @xmath1 model calculated using dmrg method .
exact results for the @xmath16 case are listed as a comparison .
the last number in the brackets are the estimated value accurate at that digit . [ cols="^,^,^,^",options="header " , ] current address : dept .
de physique , univ .
de moncton , moncton , n.b .
e1a 3e9 , canada . j.g . bednorz and k.a .
mller , z. phys .
b * 64 * , 189 ( 1986 ) .
anderson , science * 235 * , 1196 ( 1987 ) ; f.c .
zhang and t.m .
rice , phys .
b * 37 * , 3759 ( 1988 ) .
white and d. huse , phys .
b * 48 * , 3844 ( 1993 ) ; e. s. srensen and i. affleck , @xmath100 * 71 * , 1633 ( 1993 ) ; c.c .
yu and s.r .
white , @xmath100 * 71 * , 3866 ( 1993 ) ; r.m .
noack , s.r .
white and d.j .
scalapino , @xmath100 * 73 * , 882 ( 1994 ) .
the code written for this work did not implement the theorem of block diagonalization of the reduced density matrix so we have to directly diagonalize matrix size of @xmath101 in each iteration .
however we expect that after implementing the block diagonalization of the reduced density matrix we can keep more states in blocks 1 and @xmath22 and obtaining higher accuracy .
work in this direction is in progress . | the one - dimensional ( @xmath0 ) @xmath1 model is investigated using the density matrix renormalization group ( dmrg ) method .
we report for the first time a generalization of the dmrg method to the case of arbitrary band filling and prove a theorem with respect to the reduced density matrix that accelerates the numerical computation .
lastly , using the extended dmrg method , we present the ground state electron momentum distribution , spin and charge correlation functions .
the @xmath2 anomaly of the momentum distribution function first discussed by ogata and shiba is shown to disappear as @xmath3 increases .
we also argue that there exists a density - independent @xmath4 beyond which the system becomes an electron solid . |
the strong spin - orbit interaction is one of the current topics in condensed matter physics , as it accompanies fascinating phenomena such as the multipole order in @xmath14-electron systems@xcite and the @xmath15 mott transition in 5@xmath16 electron systems.@xcite in this context , the layered perovskite vanadium oxide sr@xmath0vo@xmath1 is attracting renewed interest as a candidate compound for a @xmath17 mott insulator showing novel spin - orbital magnetism .
it has a tetragonal k@xmath0nif@xmath1 structure with the space group @xmath18 at room temperature.@xcite the magnetic vanadium ion ( v@xmath19 , 3@xmath17 ) is octahedrally coordinated with six ligand oxygen ions , where the octahedron is elongated along the @xmath8 axis .
the anisotropy of the octahedron leads to a tetragonal crystal field that partially lifts the three - fold degeneracy of the @xmath20 orbitals by splitting them into @xmath21 and @xmath22 orbitals with @xmath21 having a higher energy .
thus , the single @xmath16 electron would occupy the unquenched @xmath22 orbitals in case of a negligible spin - orbit interaction.@xcite however , it has been pointed out that the spin - orbit coupling may be large enough to split the @xmath22 state into two kramers doublets.@xcite the kramers doublets are mapped to isospin up and down states , which consist of the electron spin and orbital angular momentum .
it was also suggested that a magnetic octupole order ( an antiferroic order of the isospins ) would be the ground state of sr@xmath0vo@xmath1 .
furthermore , an alternating scenario of a spin - orbital order with almost muted magnetic moments has been proposed for the ground state of this compound.@xcite despite the efforts of several groups to synthesize high - quality powdered samples in search for signatures of novel spin - orbital magnetism , the varying electronic properties reported in earlier studies seem to call for precautions in assessing the results in relation to the sample quality .
the magnetic susceptibility ( @xmath23 ) and heat capacity measurements provided a common ground that the compound successively undergoes three phase transitions at @xmath24 10 k , @xmath25 100 k , and @xmath3 130 k.@xcite a report on an in - house x - ray diffraction ( xrd ) experiment showed the splittings of several reflection peaks between @xmath7 and @xmath26.@xcite it was also reported that a rapid enhancement in the @xmath27 ratio ( where @xmath8 and @xmath9 are the lengths of the unit cell along the respective crystalline axes ) occurred below @xmath7 .
@xcite regarding the magnetic properties , a sudden decrease in @xmath23 was observed at @xmath7 , although the magnitude of the reduction varied from sample to sample.@xcite below @xmath13 , a clear hysteresis was observed in magnetization measurements , where the magnetism was confirmed to be a bulk property.@xcite the magnetic hysteresis implies the appearance of weak ferromagnetism below @xmath13 .
although no experimental evidence for the magnetic octupole order has been reported so far , recent inelastic neutron scattering experiments suggest that the spin - orbit coupling is strong enough to induce excitation between the crystal field levels.@xcite further experimental investigation is required to clarify the microscopic details of magnetism in sr@xmath0vo@xmath1 .
a previous study on muon spin rotation / relaxation ( @xmath2sr ) reported a spontaneous internal field at the muon site with a broad distribution and residual magnetic fluctuation in the temperature ( @xmath28 ) range of 1.8 k @xmath29.@xcite these results suggest that the spin - orbital ordering state , which was proposed by first - principles calculation studies by imai _ _ et al.__@xcite , appeared below @xmath13 .
on the basis of the theoretical studies , we can expect a severe competition between many types of spin - orbital ordering patterns due to the coexistence of ferromagnetic and antiferromagnetic correlations and the associated effect of frustration .
in such an exotic ground state , it is expected that the persistent spin dynamics and short - range correlations are found far below the magnetic ordering temperature .
then , we extended the @xmath2sr experiments down to 30 mk . in this paper , we report on the electronic properties of sr@xmath0vo@xmath1 inferred from the combined use of synchrotron radiation x - ray diffraction ( sr - xrd ) and @xmath2sr measurements of powdered samples of the same batch . from the sr - xrd results
, we found an intermediate orthorhombic phase in the @xmath28 range @xmath30 , whereas a tetragonal phase appears for @xmath5 and @xmath7 .
it is demonstrated by @xmath2sr that the anomalies at @xmath7 and @xmath26 do not accompany long - range magnetic order .
moreover , highly inhomogeneous magnetism is observed to develop below @xmath13 . a possible model of local field distribution
is introduced to discuss the observed @xmath2sr lineshape and magnetism behind it over the relevant @xmath28 region .
we prepared two batches of polycrystalline samples ( labeled sample - a and b with a net weight of @xmath31300 mg and @xmath3130 mg , respectively ) by a two step solid - state reaction.@xcite first , the precursor sr@xmath32(vo@xmath1)@xmath33(oh)@xmath0 was prepared by sintering a stoichiometric mixture of sr(no@xmath34)@xmath0 and v@xmath0o@xmath35 at 550 - 800 @xmath36c for 84 h in air .
then , sr@xmath32(vo@xmath1)@xmath33(oh)@xmath0 and sro were mixed at a molar ratio of 1:2 and sintered at 950 @xmath36c for 162 h in a h@xmath0 gas flow .
@xmath2sr measurements were conducted for sample - a , whereas @xmath23 and sr - xrd measurements were performed for both samples .
@xmath23 was measured over a @xmath28 range between 2 k and 300 k in the presence of a magnetic field of 0.2 t with a superconducting quantum interference device ( squid ) .
an sr - xrd experiment was carried out using the imaging - plate diffractometer installed on beamline 8a at the photon factory at kek ( kek - pf ) .
the incident beam was monochromatized by a si(111 ) double crystal .
x - ray diffraction patterns were measured in the @xmath28 range of 47 - 305 k by using a he blower with the x - ray wavelength of 0.077534(3 ) nm .
the lattice constants were obtained by the le bail whole - profile fitting to the sr - xrd patterns.@xcite the le bail whole - profile refinements were performed using the rietan - fp software package.@xcite a conventional @xmath2sr experiment was conducted above 4.2 k using the d1 instrument at j - parc muse and down to 30 mk using the dr spectrometer furnished with a dilution refrigerator on the m15 beamline of triumf , canada .
the muon spin depolarization function @xmath37 was monitored by measuring the decay - positron asymmetry @xmath38 where @xmath39 is the initial asymmetry , and @xmath40 is the background contribution mainly ascribed to the sample holder made of silver ( in which the muon depolarization is negligible ) .
the @xmath2sr time spectra @xmath41 was then analyzed by curve fitting , where we employed the recursive function @xmath37 derived from the appropriate models for the local field distribution ( see the next section for more detail ) .
we obtained @xmath42 for j - parc muse d1 and 0.083(3 ) for triumf m15 dr from the calibration measurements .
( color online ) ( a ) temperature dependence of the magnetic susceptibility for polycrystalline samples of sr@xmath0vo@xmath1 , sample - a ( blue ) and sample - b ( red ) .
the filled and empty circles represent data for the zero - field - cooling ( zfc ) and field - cooling ( fc ) processes , respectively .
the inset shows a magnified view near @xmath31100 k. ( b ) a magnetization curve at 2 k for sample - b.,width=264 ] figure [ fig:1](a ) shows the @xmath28 dependence of @xmath23 for two polycrystalline samples used in the present experiment .
we observed a sudden decrease in @xmath23 at @xmath7 in both results , despite their quantitative differences .
we also observed a clear difference in the @xmath28 dependence of magnetization between the zero - field - cooling ( zfc ) and field - cooling ( fc ) processes below @xmath13 .
we can see a clear magnetic hysteresis in the magnetization curve at 2 k in fig .
[ fig:1](b ) , suggesting weak ferromagnetism .
the saturated ferromagnetic moment is estimated to be @xmath43/v . the data for sample - a exhibit a small cusp at @xmath3155 k in addition to the anomalies at @xmath7 and @xmath13 .
although we could not identify the origin of the cusp at present , we note that the peak temperature is slightly higher than the liquid - solid phase transition temperature of oxygen , which is usually observed at @xmath3150 k. the inset in fig .
[ fig:1](a ) shows a magnified view near 100 k , and it clearly shows that sample - b exhibits a larger decrease in @xmath23 at @xmath7 than sample - a .
it was reported that the anomalies at @xmath7 and @xmath13 were gradually suppressed as the oxygen deficiency @xmath44 increased for the formula unit ( sr@xmath0vo@xmath45).@xcite this suggests that sample - a is slightly oxygen off - stoichiometric , although both batches of samples were prepared by the same synthetic method .
the oxygen off - stoichiometry may be introduced by the preparation with complex precursor materials and a hydrogen treatment .
these facts indicate that we need greater control of the synthesis conditions for more precise control of the oxygen content .
( color online ) the sr - xrd patterns at 305 k for sample - a ( blue ) and sample - b ( red ) .
the green tick marks indicate the peak positions of the main phase .
the asterisks indicate the peaks from the secondary phase . , width=264 ] ( color online ) typical examples of synchrotron x - ray diffraction data measured at ( a ) 305 k , ( b ) 114 k , and ( c ) 66 k. the red lines are the fitted results of the le bail whole - profile refinement .
the green tick marks represent the positions of the calculated reflections .
the blue lines denote the difference between the observed and calculated profiles.,width=283 ] ( color online ) temperature variation of the selected peaks in the synchrotron radiation x - ray diffraction patterns.,width=264 ] the structural properties of the present samples were investigated using the sr - xrd technique .
figure [ fig:2 ] shows the sr - xrd patterns at 305 k for sample - a and b. it has been confirmed that the samples were mostly single - phase sr@xmath0vo@xmath1 with small amount of a secondary phase at 2@xmath46 .
the secondary phase was identified as sr@xmath32(vo@xmath1)@xmath33(oh)@xmath0 used as the precursor during synthesis .
hereafter , we only show the sr - xrd results for sample - b because there are very few differences between the two samples .
figure [ fig:3 ] shows the diffraction patterns measured at ( a ) 305 k , ( b ) 114 k , and ( c ) 66 k with the results of the le bail whole - profile fitting method . the sr - xrd pattern at 305 k was indexed by a tetragonal structure with the @xmath18 space group .
the lattice parameters obtained were @xmath47 nm and @xmath48 nm for sample - a and @xmath49 nm and @xmath50 nm for sample - b at 305 k. it has been confirmed by the sr - xrd experiment of the present sample that a structural transition occurs at @xmath26 and @xmath7 , as shown in previous studies.@xcite figure [ fig:5](a ) shows the @xmath28 variation of the half width at half maximum ( hwhm ) for the selected diffraction peaks .
although we could not find any additional peaks in the sr - xrd patterns , the hwhm of the 200 and 420 peaks increased with decreasing temperature for @xmath30 . on the other hand , the 110 , 004 , and 002 peaks did not exhibit line broadening within the experimental accuracy .
these results might be attributed to the existence of an orthorhombic phase for @xmath30 .
the absence of broadening in the 110 peak suggests no diagonal distortion along the @xmath51 direction .
it also excludes the coexistence of two tetragonal phases , which had been suggested in a previous study.@xcite the enhancement in the hwhm for 420(200 ) peak is explained by the tiny splitting of the peak into the 420(200 ) and 240(020 ) peaks due to the expansion and contraction of the two tetragonal @xmath9 axes .
( color online ) ( a ) temperature dependence of the half width at half maximum ( hwhm ) for selected peaks .
( b ) temperature dependence of lattice constants .
solid lines are guide for the eyes.,width=264 ] the continuous line broadening and a lambda - like anomaly in the @xmath28 dependence of the specific heat@xcite strongly suggest that the phase transition at @xmath26 is second - order . because no superlattice modulation has been discovered below @xmath26 , we can provide the possible orthorhombic space group @xmath52 among the subgroups of @xmath18 .
we confirmed a clear difference between @xmath18 and @xmath52 in the reliability factor ( @xmath53 ) of the le bail analysis for @xmath30 , whereas we did not observe a large difference between them for @xmath5 and @xmath6 .
for instance , @xmath54 for @xmath18 and @xmath55 for @xmath52 at 114 k. the @xmath28 dependencies of the lattice constants are plotted in fig . [ fig:5](b ) .
we can see the reduction in the @xmath8 and @xmath9 lattice constants above @xmath26 , and the @xmath9-axis length splits into two below @xmath26 .
however , the origins of the phase transitions at @xmath26 are unknown at present .
single - crystal data are highly desirable for further understanding .
although line broadening was observed for some reflection peaks for @xmath30 as shown above , we could not identify clear splitting , broadening , or additional peaks in the relevant sr - xrd patterns below @xmath7 .
the hwhms above @xmath26 and below @xmath7 are identical within the experimental uncertainty , as shown in fig . [
fig:5](a ) .
furthermore , we obtained similar values of @xmath53 for the le bail analysis with the @xmath18 and @xmath52 structures ; as shown for @xmath5 , @xmath56 for @xmath18 and @xmath57 for @xmath52 at 66 k. these results suggest that a tetragonal @xmath18 phase appears below @xmath7 .
figures [ fig:4](a ) and ( b ) show the 002 and 004 peaks in the xrd patterns observed at various values of @xmath28 .
we found that the 002 and 004 peaks shifted to lower angles below @xmath31105 k , whereas the 200 , 110 , and 420 peaks moved to higher angles , as shown in figs .
[ fig:4](c ) , ( d ) , and ( f ) . as shown in fig .
[ fig:5 ] ( b ) , these results correspond to a steep enhancement in the @xmath8-axis length , whereas the @xmath9-axis length slightly decreases below @xmath7 .
this explains the asymmetric broadening of the 118 and 220 peaks below @xmath7 , as shown in fig .
[ fig:4](e ) because these two peaks move in opposite directions owing to the structural change .
although there are some anomalies in the structural properties , as shown above , we could not find evidence of the lowering of the lattice symmetry and superlattice modulation below @xmath7 , which are expected in the orbital order state with the superstructure of the @xmath10 and @xmath11 orbitals .
this indicates that the degeneracy of the @xmath58 orbitals is not lifted below @xmath7 .
these structural properties below @xmath7 are fully consistent with the previous results reported by zhou _
et al_.@xcite they suggested that the enhancement in the @xmath27 ratio without orthorhombic distortion might originate from the increased occupancy of the @xmath22 orbitals in place of the @xmath21 orbital .
this seems reasonable because the crystal field splitting between the @xmath21 and @xmath22 orbitals is quite small ( @xmath310.08 ev in the local density approximation calculation).@xcite thus , we suggest that the @xmath21 orbital contribution to the electronic state is not negligible in the high-@xmath28 tetragonal phase above @xmath26 owing to the narrow gap between the @xmath22 and @xmath21 states . then , the electronic state changes to the pure @xmath22 state in the low-@xmath28 tetragonal phase below @xmath7 owing to an enhancement in the energy gap , which is caused by the increase in the @xmath27 ratio to gain crystal field energy .
in the paramagnetic state of ferromagnetic and antiferromagnetic systems , the muon spin depolarization at zero external field has contributions from the static nuclear dipolar fields and the fluctuating fields from electron spins .
the nuclear spin contribution is well reproduced by the gaussian kubo - toyabe ( gkt ) function derived from the gaussian distribution of the local field ( @xmath60 ) , which is a good approximation for a dense ensemble of randomly orientated nuclear magnetic moments.@xcite the gkt function is written as @xmath61 \exp\left[-\frac{(\delta t)^2}{2}\right ] , \label{eq : gkt}\ ] ] where @xmath62 corresponds to the root - mean - square for the distribution of the local magnetic field ( where @xmath63 mhz / t is the muon gyromagnetic ratio ) . on the other hand ,
the dynamical electron spin contribution is given as @xmath64 with the relaxation rate @xmath65 .
then , the depolarization function @xmath37 is described by the product of the two contributions : @xmath66 figure [ fig:6](a ) shows the zero - field ( zf)-@xmath2sr spectra at 300 k and 80 k ( i.e. , above and below @xmath7 ) , where no sign of the emergence of the spontaneous internal magnetic field , such as oscillation and/or fast damping , is observed below @xmath7 .
this immediately leads us to the conclusion that no long - range magnetic order is present below @xmath67 , which is in qualitative agreement with the previous result.@xcite the temperature dependence of @xmath65 deduced by a curve fitting is plotted in the inset of fig .
[ fig:6](a ) .
( color online ) ( a ) zero - field ( zf)-@xmath2sr time spectra at 300 k , 80 k , and 18 k. the solid curves are the fitted results of eq .
( [ eq : pm_g ] ) .
the inset shows the temperature dependence of the relaxation rate .
( b ) @xmath2sr spectra at 18 k measured with zf and longitudinal fields ( lf = 10 mt and 100 mt ) with the results of the curve fitting with eq.([eq : lkt ] ) .
, width=264 ] the spectra in fig .
[ fig:6](a ) including the one at 18 k show that the exponential damping gradually overcomes the gkt depolarization below 40 k. figure [ fig:6](b ) shows @xmath2sr spectra at 18 k measured with zf and longitudinal fields ( lfs ) of 10 mt and 100 mt .
generally , the exponential damping due to the electron spin fluctuation does not depend much on the lf because the fluctuation rate of the dynamical field from the ordinary paramagnetic moments is much larger than the larmor frequency of the muon spin precession .
however , the depolarization is quenched by the application of a relatively weak lf , which indicates that @xmath60 is much weaker than the lf and that it is quasistatic within the time window of @xmath2sr ( @xmath68 s ) .
these features allow us to estimate an upper limit for the mean hyperfine field at the muon site , @xmath69 2 mt , and that of the fluctuation rate @xmath70 0.5 mhz .
we also performed curve fitting for the lf dependence of the @xmath2sr spectra at 18 k by replacing the exponential term in eq .
( [ eq : pm_g ] ) with the lorentzian kubo - toyabe ( lkt ) function for a finite lf : @xmath71 \nonumber \\ & & - \bigl[1 + \bigl(\frac{\delta}{\omega_0}\bigr)^2 \bigr ] \delta \int^t_0 j_0(\omega_0 t^{\prime } ) e^{-\delta t^{\prime } } d t^{\prime } , \label{eq : lkt}\end{aligned}\ ] ] where @xmath72 is the larmor frequency of the muon spin precession in the presence of an external field @xmath73 , and @xmath74 and @xmath75 denote spherical bessel functions.@xcite we note that the results of the curve fitting indicate that the entire volume of the sample exhibits a lorentzian - like field distribution .
thus , we are led to the conclusion that the dilute magnetic moments of @xmath16 electrons having an unusually quasistatic characteristic make a significant contribution to the muon depolarization for @xmath76 , although their origin is not clear at this stage . as is observed for the zf-@xmath2sr spectra in fig .
[ fig:7](a ) , a rapid depolarization develops with decreasing temperature below @xmath13 .
we show the zf- and lf-@xmath2sr spectra at 30 mk measured in the presence of various lfs up to 100 mt in fig .
[ fig:7](b ) .
the zf spectra clearly demonstrate the absence of conventional long - range magnetic order at the base temperature .
the initial fast depolarization exhibits a gradual suppression with increasing lf and is nearly quenched at lf = 100 mt .
surprisingly , we also observed a finite slope due to the residual magnetic fluctuation in these spectra at a large time , which is almost independent of the lf , even at 30 mk .
these features can be observed from the fast damped 2/3 amplitude followed by a slowly relaxing 1/3 component originating from the magnetic ordered state with slow persistent spin dynamics .
( color online ) ( a ) zf-@xmath2sr time spectra in the temperature range between 4.2 k and 10 k. ( b ) @xmath2sr time spectra measured at 30 mk in the presence of various longitudinal fields ( lfs ) up to 100 mt .
the inset shows the earlier part of the spectrum at zf .
the solid lines are the fitted results by using eq .
( [ eq : g_mo]).,width=264 ] . in a previous @xmath2sr study ,
sugiyama @xmath78 analyzed their zf-@xmath2sr spectra by using a combination of an exponentially relaxing cosine oscillation for the damped 2/3 component and an exponential function for the 1/3 tail signal.@xcite they showed that the ratio of the relaxation rate of the 1/3 tail component @xmath79 to the precession frequency @xmath80 is close to unity , @xmath81 , where @xmath82 corresponds to the field distribution normalized to the mean value of @xmath60 .
however , their fitting function is based on an unique precession signal at a spontaneous local field for the conventional magnetic ordered state , @xmath83 .
the observed broad @xmath60 distribution and residual magnetic fluctuation strongly suggest that a highly inhomogeneous magnetic state appears in the ground state of sr@xmath0vo@xmath1 . the inset of fig .
[ fig:7](b ) shows the earlier part of the zf-@xmath2sr spectrum at 30 mk . as mentioned above ,
the gkt and lkt functions are widely attributed to the @xmath2sr spectra for random magnetism in dense and diluted spin systems .
however , the zf spectrum at 30 mk lacks a decrease that usually precedes the asymptotic tail for the gkt function , although it exhibits a gaussian - like initial depolarization .
we attributed this to a broad distribution of @xmath60 , which is not characterized by a single root - mean - square width , and adopted the gaussian - broadened gaussian ( gbg ) function as a recursive function for the curve fitting.@xcite the gbg function is described as one of the generalized forms of the gkt function , where @xmath84 in eq .
( [ eq : gkt ] ) has a distribution with the mean value @xmath85 and root - mean - square width @xmath86 .
@xmath87 for zf is written as @xmath88 where @xmath89 is the ratio of the distribution width , and @xmath90 . for @xmath91 ,
the gbg function is identical to the gkt function with a deep minimum after the initial depolarization .
the depth of the minimum decreases with increasing @xmath89 and becomes almost independent of @xmath89 for @xmath92 .
thus , we imposed the condition @xmath93 in the following analysis .
we note that the gbg function does not approach the lkt function , even for @xmath94 . for the curve fitting ,
we presume that @xmath37 is primarily described by the product of the gbg function and a slow exponential decay due to the persistent spin dynamics in addition to a supplementary term for exponential damping : @xmath95 where @xmath14 is the volume fraction exhibiting inhomogeneous magnetism , and @xmath65 is the relaxation rate reflecting the spin dynamics .
the second exponential term with the relaxation rate @xmath96 is for extrinsic contributions originating from the magnetically ordered state region with relatively fast persistent dynamics , which makes the above muon spin depolarization model unreasonable .
the @xmath28 dependence of @xmath14 is plotted in the inset of fig .
[ fig:8 ] .
the inhomogeneous magnetism develops over the almost entire volume of the sample below 6 k. we obtained @xmath97 below 4.2 k , whereas @xmath89 is unity at higher temperatures .
this discrepancy might be caused by the lack of data for @xmath98 0.1 @xmath2s that is masked by the limited time resolution of the pulsed muon beam ( @xmath31100 ns ) .
we note that the dynamical lkt function and dynamical spin glass models@xcite could not reproduce the spectra over the early time range .
the absence of decrease in the @xmath2sr time spectra tells us important aspects of the magnetic ground state .
_ reproduced the shallow decrease in the gbg function by assuming a local field distribution having excess low - field sites.@xcite this implies that the magnetic ordered state is accompanied by a large modulation of the ordered moment magnitude .
they also investigated the relationship between @xmath89 and the correlation length by monte carlo simulation . on the basis of their results
, @xmath99 corresponds to a correlation length of @xmath314 nm .
_ predicted that the spin and orbital correlation functions are reduced as the system size increases.@xcite this result suggests that the correlation length of the spin and orbital degrees of freedom are finitely limited .
our result strongly supports their predictions .
( color online ) ( a ) crystal structure of sr@xmath0vo@xmath1 with the minimum regions of the electrostatic potential .
the yellow regions define the expected muon site from our hartree potential calculation by vasp code . @xcite
this figure is drawn by vesta.@xcite ( b ) temperature dependencies of @xmath85 ( left axis ) and the relaxation rate @xmath65 ( right axis ) .
the inset shows the temperature dependence of the volume fraction @xmath14 , showing the inhomogeneous magnetism .
, width=321 ] the @xmath28 dependence of @xmath85 and @xmath65 are shown in fig .
[ fig:8](b ) .
@xmath85 behaves like an order parameter below @xmath13 , approaching 10.3(6 ) @xmath2s@xmath100 at 30 mk . from this value
, we can extract a mean field at the muon site , @xmath101 mt . according to our hartree potential calculation by using the vienna ab - initio simulation package ( vasp),@xcite the potential minima for a muon are located around the ( 0.50 , 0.24 , 0.69 ) position , as shown in fig . [
fig:8](a ) .
assuming randomly oriented 1@xmath102 v @xmath16-electron moments , we obtained @xmath10352 mt/@xmath102 at the relevant muon site .
thus , we roughly estimate the ordered moment at the v site to be @xmath104 .
this value is much smaller than the dipole moment for the v@xmath19(3@xmath17 ) ion with quenching of the orbital moment , @xmath105 for @xmath106 . on the other hand
, @xmath65 slightly decreases with decreasing @xmath28 indicating the gradual suppression of the persistent dynamics .
however , we observed finite residual magnetic fluctuation even at 30 mk .
persistent spin dynamics have been observed for the magnetic ordered state of several geometrically frustrated magnetic systems .
@xcite finally , we briefly discuss the origin of the inhomogeneous magnetism . as shown in fig .
[ fig:1](a ) , the magnetization data suggest the existence of the oxygen off - stoichiometry in sample - a , which is used in the present @xmath2sr experiment .
however , we could not find any line broadenings and additional peaks in the sr - xrd pattern for sample - a , as shown in fig .
[ fig:2 ] .
this indicates that there are no severe deformation and crystallographic randomness due to the oxygen off - stoichiometry in sample - a . from these results
, we can consider that the inhomogeneous magnetism might have an intrinsic origin , although we would observe a qualitative difference in the field distribution , correlation length , and relaxation rate depending on the magnitude of the oxygen off - stoichiometry .
et al . _ also revealed that there is severe competition between many spin - orbital order patterns due to the coexistence of ferromagnetic and antiferromagnetic interactions and the associated effect of frustration in sr@xmath0vo@xmath1.@xcite in their proposed spin - orbital ordering patterns , orthogonal and not strictly orthogonal nearest - neighboring orbital configurations appeared .
such orbital configurations induce the coexistence of ferromagnetic and antiferromagnetic signs of the exchange interactions because of the goodenough - kanamori rule .
we suggest that the observed broad @xmath60 distribution , short - range correlations , and persistent spin dynamics are signatures of the severe competition between the possible spin - orbital ordered patterns in sr@xmath0vo@xmath1 .
we have conducted sr - xrd and @xmath2sr measurements of polycrystalline samples of sr@xmath0vo@xmath1 .
the sr - xrd results suggest the existence of an intermediate orthorhombic phase between @xmath3 130 k and @xmath4 100 k , whereas a tetragonal phase appears for @xmath5 and @xmath6 .
although a large enhancement in the @xmath27 ratio had been confirmed below @xmath7 , we could not find clear evidence for the superlattice modulation and the lowering of the lattice symmetry from tetragonal @xmath18 , which is easily expected for the orbital order state with a superlattice structure of @xmath10 and @xmath11 orbitals .
these results indicate that the degeneracy of the @xmath22 orbitals is not lifted , even below @xmath7 .
the @xmath2sr data revealed that the structural transition at @xmath7 does not accompany magnetic ordering . on the other hand
, we extended the @xmath2sr measurement down to 30 mk and found a highly inhomogeneous magnetic ordered state with slow persistent spin dynamics below @xmath12 10 k. in the magnetic ordered state , the @xmath2sr spectra were well fitted by the gbg function , which is known to describe the muon spin depolarization due to the inhomogeneous magnetism with short - range correlations and a vast moment size distribution .
these results suggest that the inhomogeneous magnetic state may be a signature of the severe competition between the many possible spin - orbital ordered states predicted by first - principles calculations @xcite .
the authors thank h. okabe , s. takeshita , j. sugiyama , y. kato , and m. itoh for fruitful discussions .
we also thank a. sagayama , h. lee , and the staff of triumf and j - parc muse for their technical support .
p. santini , s. carretta , g. amoretti , r. caciuffo , n. magnani , and g. h. lander , rev .
, * 81 , * 807 ( 2009 ) .
b. j. kim , h. ohsumi , t. komesu , s. sakai , t. morita , h. takagi , and t. arima , science , * 323 , * 1329 ( 2009 ) .
a. nozaki , h. yoshikawa , t. wada , h. yamauchi , and s. tanaka , phys .
b , * 43 , * 181 ( 1991 ) .
h. d. zhou , b. s. conner , l. balicas , and c. r. wiebe , phys .
lett . , * 99 , * 136403 ( 2007 ) .
g. jackeli and g. khaliullin , phys
. rev . lett . , * 103 , * 067205 ( 2009 ) .
m. v. eremin , j. deisenhofer , r. m. eremina , j. teyssier , d. van der marel , and a. loidl , phys .
b , * 84 , * 212407 ( 2011 ) .
r. viennois , e. giannini , j. teyssier , j. elia , j. deisenhofer , d. van der marel , j. phys.:conference series , * 200 , * 012219 ( 2010 ) . j. teyssier , r. viennois , e. giannini , r. m. eremina , a. gnther , j. deisenhofer , m. v. eremin , and d. van der marel , phys .
b , * 84 , * 205130 ( 2011 ) .
t. ueno , j. kim , m. takata , and t. katsufuji , j. phys .
* 83 , * 034708 ( 2014 ) .
k. nawa , h. ueda , k. yoshimura , and m. takigawa , _ proceedings of the 2012 powder metallurgy world congress & exhibition _ , 16f - t14 - 21 ( 2012 ) .
j. sugiyama , h. nozaki , i. umegaki , w. higemoto , e. j. ansaldo , j. h. brewer , h. sakurai , t .- h .
yang , and m. mnsson , phys .
b , * 89 , * 020402(r ) ( 2014 ) .
j. sugiyama , h. nozaki , i. umegaki , w. higemoto , e. j. ansaldo , j. h. brewer , h. sakurai , t .- h .
yang , and m. mnsson , j. phys.:conference series , * 551 , * 012011 ( 2014 ) .
h. d. zhou , y. j. jo , j. fiore carpino , g. j. munoz , c. r. wiebe , j. g. cheng , f. rivadulla , and d. t. adroja , phys .
b , * 81 , * 212401 ( 2010 ) .
y. imai , i. solovyev , and m. imada , phys .
lett . , * 95 , * 176405 ( 2005 ) . y. imai and m. imada , j. phys .
jpn . , * 75 , * 094713 ( 2006 ) .
a. le bail , h. duroy , and j. l. fourquet , mat .
, * 23 , * 447 , ( 1988 ) . f. izumi and k. momma , solid state phenom . , * 130 , * 15 , ( 2007 ) .
r. s. hayano , y. j. uemura , j. imazato , n. nishida , t. yamazaki , and r. kubo , phys .
b , * 20 , * 850 ( 1979 ) . y. j. uemura , t. yamazaki , d. r. harshman , m. senba , and e. j. ansaldo , phys . rev .
b , * 31 , * 546 ( 1985 ) .
d. r. noakes and g. m. kalvius , phys .
b , * 56 , * 2352 ( 1997 ) .
d. r. noakes , j. phys.:condens . matter , * 11 , * 1589 ( 1999 ) .
a. yaouanc , a. maisuradze , and p. dalmas de rotier , phys .
b , * 87 , * 134405 ( 2013 ) .
g. kresse and j. hafner , phys .
b , * 47 , * 558 ( 1993 ) . ; g. kresse and j. hafner , _ ibid . _ , * 49 , * 14251 ( 1994 ) . ; g. kresse and j. furthmller , comput . mat .
, * 6 , * 15 ( 1996 ) . ; g. kresse and j. furthmller , phys . rev .
b , * 54 , * 11169 ( 1996 ) .
s. r. dunsiger , r. f. kiefl , k. h. chow , b. d. gaulin , m. j. p. gingras , j. e. greedan , a. keren , k. kojima , g. m. luke , w. a. macfarlane , n. p. raju , j. e. sonier , y. j. uemura , and w. d. wu , phys .
b , * 54 , * 9019 ( 1996 ) .
j. a. hodges , p. bonville , a. forget , a. yaouanc , p. dalmas de rotier , g. andr , m. rams , k krlas , c. ritter , p. c. m. gubbens , c. t. kaiser , p. j. c. king , and c. baines , phys .
lett . , * 88 , * 077204 ( 2002 ) .
dalmas de rotier , v. glazkov , c. marin , a. yaouanc , p. c. m. gubbens , s. sakarya , p. bonville , a. amato , c. baines , p. j. c. king , physica b , * 374 - 375 , * 145 ( 2006 ) . p.
dalmas de rotier , a. yaouanc , l. keller , a. cervellino , b. roessli , c. baines , a. forget , c. vaju , p. c. m. gubbens , a. amato , and p. j. c. king , phys .
, * 96 , * 127202 ( 2006 ) .
r. dally , t. hogan , a. amato , h. luetkens , c. baines , j. rodriguez - rivera , m. j. graf , and s. d. wilson , phys .
lett . , * 113 , * 247601 ( 2014 ) .
k. momma and f. izumi , j. appl . crystallogr . , * 44 , * 1272 ( 2011 ) . | we report on the electronic ground state of a layered perovskite vanadium oxide sr@xmath0vo@xmath1 studied by the combined use of synchrotron radiation x - ray diffraction ( sr - xrd ) and muon spin rotation / relaxation ( @xmath2sr ) techniques , where @xmath2sr measurements were extended down to 30 mk .
we found an intermediate orthorhombic phase between @xmath3 130 k and @xmath4 100 k , whereas a tetragonal phase appears for @xmath5 and @xmath6 .
the absence of long - range magnetic order was confirmed by @xmath2sr at the reentrant tetragonal phase below @xmath7 , where the relative enhancement in the @xmath8-axis length versus that of the @xmath9-axis length was observed .
however , no clear indication of the lowering of the tetragonal lattice symmetry with superlattice modulation , which is expected in the orbital order state with superstructure of @xmath10 and @xmath11 orbitals , was observed by sr - xrd below @xmath7 .
instead , it was inferred from @xmath2sr that a magnetic state developed below @xmath12 10 k , which was characterized by the highly inhomogeneous and fluctuating local magnetic fields down to 30 mk .
we argue that the anomalous magnetic ground state below @xmath13 originates from the coexistence of ferromagnetic and antiferromagnetic correlations . |
one of the fundamental properties of fractal objects is self - similarity , which means that pieces of an object are similar to the whole object . in the last twenty years
the notion of self - similarity has successfully penetrated into algebra .
this lead to the development of many interesting constructions such as self - similar groups and semigroups , iterated monodromy groups , self - iterating lie algebras , permutational bimodules , etc .
the first examples of self - similar groups showed that these groups enjoy many fascinating properties ( torsion , intermediate growth , finite width , just - infiniteness and many others ) and provide counterexamples to several open problems in group theory .
later , nekrashevych showed that self - similar groups appear naturally in dynamical systems as iterated monodromy groups of self - coverings and provide combinatorial models for iterations of self - coverings .
self - similar groups are defined by their action on the space @xmath4 of all finite words over a finite alphabet @xmath1 one the most basic self - similar object .
a faithful action of a group @xmath5 on @xmath4 is called self - similar if for every @xmath6 and @xmath7 there exist @xmath8 and @xmath9 such that @xmath10 for all words @xmath11 .
the self - similarity of the action is reflected in the property that the action of any group element on a piece @xmath12 ( all words with the first letter @xmath13 ) of the space @xmath4 can be identified with the action of another group element on the whole space @xmath4 .
we can also imagine the set @xmath4 as the vertex set of a regular rooted tree with edges @xmath14 for @xmath15 and @xmath11 .
then every self - similar group acts by automorphisms on this tree .
alternatively , self - similar groups can be defined as groups generated by the states of invertible mealy automata which are also known as automaton groups or groups generated by automata .
all these interpretations come from different applications of self - similar groups in diverse areas of mathematics : geometric group theory , holomorphic dynamics , fractal geometry , automata theory , etc .
( see @xcite and the references therein ) .
in this paper we consider schreier graphs of self - similar actions of groups . given a group @xmath5 generated by a finite set @xmath16 and acting on a set @xmath17 , one can associate to it the ( simplicial ) schreier graph @xmath18 : the vertex set of the graph is the set @xmath17 , and two vertices @xmath13 and @xmath19 are adjacent if and only if there exists @xmath20 such that @xmath21 .
schreier graphs are generalizations of the cayley graph of a group , which corresponds to the action of a group on itself by the multiplication from the left .
every self - similar group @xmath5 preserves the length of words in its action on the space @xmath4 .
we then have a family of natural actions of @xmath5 on the sets @xmath22 of words of length @xmath23 over @xmath1 . from these actions
one gets a family of corresponding finite schreier graphs @xmath24 .
it was noticed in @xcite that for a few self - similar groups the graphs @xmath25 are substitution graphs ( see @xcite ) they can be constructed by a finite collection of vertex replacement rules and , normalized to have diameter one , they converge in the gromov - hausdorff metric to certain fractal spaces .
however , in general , the schreier graphs of self - similar groups are neither substitutional nor self - similar in any studied way ( see discussion in ( * ? ? ?
* section i.4 ) ) . nevertheless , the observation from @xcite lead to the notion of the limit space of a self - similar group introduced by nekrashevych @xcite , which usually have fractal structure .
although finite schreier graphs @xmath25 of a group do not necessary converge to the limit space , they form a sequence of combinatorial approximations to it .
any self - similar action can be extended to the set @xmath0 of right - infinite words over @xmath1 ( boundary of the tree @xmath4 ) .
therefore we can also consider the uncountable family of schreier graphs @xmath26 corresponding to the action of the group on the orbit of @xmath27 .
each schreier graph @xmath28 can be obtained as a limit of the sequence @xmath24 in the space @xmath29 of ( isomorphism classes of ) pointed graphs with pointed gromov - hausdorff topology .
the map @xmath30 sending a point @xmath27 to the isomorphism class of the pointed graph @xmath31 pushes forward the uniform probability measure on the space @xmath32 to a probability measure on the space of schreier graphs .
this measure is the so - called benjamini - schramm limit of the sequence of finite graphs @xmath24 .
therefore the family of schreier graphs @xmath2 and the limit space represent two limiting constructions associated to the action and to the sequence of finite schreier graphs @xmath24 .
structure of these schreier graphs as well as some of their properties such as spectra , expansion , growth , random weak limits , probabilistic models on them , have been studied in various works over the last ten years , see @xcite .
most of the studied self - similar groups are generated by the so - called bounded automata introduced by sidki in @xcite .
the structure of bounded automata is clearly understood , which allows one to deal fairly easily with groups generated by such automata .
the main property of bounded automaton groups is that their action is concentrated along a finite number of `` directions '' in the tree @xmath4 .
every group generated by a bounded automaton belongs to an important class of contracting self - similar groups , which appear naturally in the study of expanding ( partial ) self - coverings of topological spaces and orbispaces , as their iterated monodromy groups @xcite .
moreover , all iterated monodromy groups of post - criticaly finite polynomials are generated by bounded automata .
the limit space of an iterated monodromy group of an expanding ( partial ) self - covering @xmath33 is homeomorphic to the julia set of the map @xmath33 . in the language of limit spaces ,
groups generated by bounded automata are precisely those finitely generated self - similar groups whose limit spaces are post - critically finite self - similar sets ( see @xcite ) .
such sets play an important role in the development of analysis on fractals ( see @xcite ) .
the main goal of this paper is to investigate the ends of the schreier graphs @xmath34 of self - similar groups generated by bounded automata and the corresponding limit spaces .
the number of ends is an important asymptotic invariant of an infinite graph .
roughly speaking , each end represents a topologically distinct way to move to infinity inside the graph .
the most convenient way to define an end in an infinite graph @xmath35 is by the equivalence relation on infinite rays in @xmath35 , where two rays are declared equivalent if their tails lie in the same connected component of @xmath36 for any finite subgraph @xmath37 of @xmath35 .
any equivalence class is an _
end _ of the graph @xmath35 .
the number of ends is a quasi - isometric invariant .
the cayley graph of an infinite finitely generated group can have one , two or infinitely many ends .
two - ended groups are virtually infinite cyclic and the celebrated theorem of stallings characterizes finitely generated groups with infinite number of ends .
our main results can be summarized as follows .
* given a group generated by a bounded automaton we exhibit a constructive method that determines , for a given right - infinite word @xmath27 , the number of ends of the schreier graph @xmath28 ( section [ section_ends ] ) . *
we show that the schreier graphs @xmath28 of a group generated by a bounded automaton have either one or two ends almost surely , and there are only finitely many schreier graphs with more than two ends ( section [ section_number_ends ] , proposition [ prop_more_ends ] , theorem [ th_classification_two_ends ] ) . *
we classify bounded automata generating groups whose schreier graphs @xmath28 have almost surely two ends ( theorem [ th_classification_two_ends ] ) . in the binary case ,
these groups agree with the class of groups defined by uni in @xcite ( theorem [ th_sch_2ended_binary ] ) . * we exhibit a constructive method that describes cut - points of limit spaces of groups generated by bounded automata ( section [ section_component ] ) .
* in particular , we get a constructive method that describes cut - points of julia sets of post - critically finite polynomials ( section [ section_component ] ) . *
we show that a punctured limit space has one or two connected components almost surely ( theorem [ thm_punctured_tile_ends ] ) .
* we classify contracting self - similar groups whose limit space is homeomorphic to an interval or a circle ( corollary [ cor_limsp_interval_circle ] ) .
the paper is organized as follows .
first , we determine the number of connected components in the schreier graph @xmath38 with removed vertex @xmath39 .
the answer comes from a finite deterministic acceptor automaton over the alphabet @xmath1 ( section [ finite automaton section ] ) so that given a word @xmath40 the automaton returns the number of components in @xmath38 ( theorem [ thm_number_of_con_comp ] ) . using this automaton we determine the number of finite and infinite connected components in @xmath41 for any @xmath3 . by establishing the connection between the number of ends of the schreier graph @xmath2 and the number of infinite components in @xmath41 ( proposition [ prop_ends_limit ] )
, we determine the number of ends of @xmath2 ( theorem [ thm_number_of_ends ] ) .
if a self - similar group acts transitively on @xmath22 for all @xmath42 , the action on @xmath0 is ergodic with respect to the uniform measure on @xmath0 , and therefore the schreier graphs @xmath2 for @xmath3 have the same number of ends almost surely . for a group generated by a bounded automaton
the `` typical '' number of ends is one or two , and we show that in most cases it is one , by characterizing completely the bounded automata generating groups whose schreier graphs @xmath2 have almost surely two ends ( theorem [ th_classification_two_ends ] ) . in the binary case we show that automata giving rise to groups whose schreier graphs have almost surely two ends correspond to the adding machine or to one of the automata defined by uni in @xcite ( theorem [ th_sch_2ended_binary ] ) . in section [ section_cut - points ]
we recall the notion of the limit space of a contracting self - similar group , and study the number of connected components in a punctured limit space for groups generated by bounded automata .
we show that the number of ends in a typical schreier graph @xmath2 coincides with the number of connected components in a typical punctured neighborhood ( punctured tile ) of the limit space ( theorem [ thm_punctured_tile_ends ] ) . in particular , this number is equal to one or two .
this fact is well - known for connected julia sets of polynomials . while zdunik @xcite and smirnov @xcite proved that almost every point of a connected polynomial julia set is a bisection point only when the polynomial is conjugate to a chebyshev polynomial , we describe bounded automata whose limit spaces have this property .
moreover , we provide a constructive method to compute the number of connected components in a punctured limit space ( section [ cut - points section ] ) . finally , using the results about ends of schreier graphs , we classify contracting self - similar groups whose limit space is homeomorphic to an interval or a circle ( corollary [ cor_limsp_interval_circle ] )
. this result agrees with the description of automaton groups whose limit dynamical system is conjugate to the tent map given by nekrashevych and uni in @xcite .
in section [ section_preliminaries ] we recall all needed definitions concerning self - similar groups , automata and their schreier graphs . in section 5
we illustrate our results by performing explicit computations for three concrete examples : the basilica group , the gupta - fabrykowski group , and the iterated monodromy group of @xmath43 .
* acknowledgments . * the substantial part of this work was done while the first author was visiting the geneva university , whose support and hospitality are gratefully acknowledged .
in this section we review the basic definitions and facts concerning self - similar groups , bounded automata and their schreier graphs . for more detailed information and for further references , see @xcite .
let @xmath1 be a finite set with at least two elements . denote by @xmath44 the set of all finite words over @xmath1 ( including the empty word denoted @xmath45 ) and with @xmath22 the set of words of length @xmath23 .
the length of a word @xmath46 is denoted by @xmath47 .
we shall also consider the sets @xmath0 and @xmath48 of all right - infinite sequences @xmath49 , @xmath50 , and left - infinite sequences @xmath51 , @xmath52 , respectively with the product topology of discrete sets @xmath1 . for an infinite sequence @xmath53 ( or @xmath54 )
we use notation @xmath55 ( respectively , @xmath56 ) . for a word @xmath39 we use notations @xmath57 and @xmath58 . the _ uniform bernoulli measure _ on each space @xmath0 and @xmath48 is the product measure of uniform distributions on @xmath1 .
shift @xmath59 _ on the space @xmath0 ( respectively , on @xmath48 ) is the map which deletes the first ( respectively , the last ) letter of a right - infinite ( respectively , left - infinite ) sequence .
a faithful action of a group @xmath5 on the set @xmath60 is called @xmath61 if for every @xmath7 and @xmath6 there exist @xmath9 and @xmath8 such that @xmath62 for all @xmath63 .
the element @xmath64 is called the _ restriction _ of @xmath65 at @xmath13 and is denoted by @xmath66 .
inductively one defines the restriction @xmath67 for all @xmath68 and @xmath69 ( we are using left actions so that @xmath70 ) .
if @xmath71 then every element @xmath7 can be uniquely represented by the tuple @xmath72 , where @xmath73 is the permutation induced by @xmath65 on the set @xmath1 .
it follows from the definition that every self - similar group @xmath5 preserves the length of words under its action on the space @xmath4 , so that we have an action of the group @xmath5 on the set @xmath22 for every @xmath23 .
the set @xmath4 can be naturally identified with a rooted regular tree where the root is labeled by the empty word @xmath45 , the first level is labeled by the elements in @xmath1 and the @xmath23-th level corresponds to @xmath22 .
the set @xmath0 can be identified with the boundary of the tree .
every self - similar group acts by automorphisms on this rooted tree and by homeomorphisms on its boundary . another way to introduce self - similar groups is through input - output automata and automaton groups .
a transducer _ automaton _ is a quadruple @xmath74 , where @xmath16 is the set of states of automaton ; @xmath1 is an alphabet ; @xmath75 is the transition map ; and @xmath76 is the output map .
we will use notation @xmath16 for both the set of states and the automaton itself .
an automaton is _ finite _ if it has finitely many states and it is _ invertible _ if , for all @xmath77 , the transformation @xmath78 is a permutation of @xmath1 .
an automaton can be represented by a directed labeled graph whose vertices are identified with the states and for every state @xmath77 and every letter @xmath6 it has an arrow from @xmath79 to @xmath80 labeled by @xmath81 .
this graph contains complete information about the automaton and we will identify them . when talking about paths and cycles in automata we always mean directed paths and cycles in corresponding graph representations .
every state @xmath77 of an automaton defines a transformation on the set @xmath82 , which is again denoted by @xmath79 by abuse of notation , as follows .
given a word @xmath49 over @xmath1 , there exists a unique path in @xmath16 starting at the state @xmath79 and labeled by @xmath83 , @xmath84 , for some @xmath85 .
then @xmath86 .
we always assume that our automata are minimal , i.e. , different states define different transformations .
the state of automata that defines the identity transformation is denoted by @xmath87 .
an automaton @xmath16 is invertible when all transformations defined by its states are invertible . in this case
one can consider the group generated by these transformations under composition of functions , which is called the _ automaton group _ generated by @xmath16 and is denoted by @xmath88 .
the natural action of every automaton group on the space @xmath0 is self - similar , and vise versa , every self - similar action of a group @xmath5 can be given by the automaton with the set of states @xmath5 and arrows @xmath89 labeled by @xmath90 for all @xmath7 and @xmath6 .
a self - similar group @xmath5 is called _ contracting _ if there exists a finite set @xmath91 with the property that for every @xmath7 there exists @xmath42 such that @xmath92 for all words @xmath39 of length greater or equal to @xmath23 .
the smallest set @xmath93 with this property is called the _ nucleus _ of the group .
it is clear from definition that @xmath94 for every @xmath95 and @xmath6 , and therefore the nucleus @xmath93 can be considered as an automaton .
moreover , every state of @xmath93 has an incoming arrow , because otherwise minimality of the nucleus would be violated . also
, the nucleus is symmetric , i.e. , @xmath96 for every @xmath95 . a self - similar group @xmath5 is called _ self - replicating _ ( or recurrent )
if it acts transitively on @xmath1 , and the map @xmath97 from the stabilizer @xmath98 to the group @xmath5 is surjective for some ( every ) letter @xmath6 .
it can be shown that a self - replicating group acts transitively on @xmath22 for every @xmath99 .
it is also easy to see ( ( * ? ? ?
* proposition 2.11.3 ) ) that if a finitely generated contracting group is self - replicating then its nucleus @xmath93 is a generating set .
let @xmath5 be a group generated by a finite set @xmath16 and let @xmath100 be a subgroup of @xmath5 .
the _ ( simplicial ) schreier coset graph @xmath101 _ of the group @xmath5 is the graph whose vertices are the left cosets @xmath102 , and two vertices @xmath103 and @xmath104 are adjacent if there exists @xmath105 such that @xmath106 or @xmath107 .
let @xmath5 be a group acting on a set @xmath17 , then the corresponding ( simplicial ) _ schreier graph @xmath108 _ is the graph with the set of vertices @xmath17 , and two vertices @xmath39 and @xmath109 are adjacent if there exists @xmath77 such that @xmath110 or @xmath111 .
if the action @xmath112 is transitive , then the schreier coset graph @xmath108 is isomorphic to the schreier graph @xmath113 of the group with respect to the stabilizer @xmath114 for any @xmath115 . let @xmath5 be a self - similar group generated by a finite set @xmath16 .
the sets @xmath22 are invariant under the action of @xmath5 , and we denote the associated schreier graphs by @xmath116 . for a point @xmath3
we consider the action of the group @xmath5 on the @xmath5-orbit of @xmath27 , and the associated schreier graph is called the _ orbital schreier graph
_ denoted @xmath117 .
for every @xmath3 we have @xmath118 , where @xmath119 denotes the prefix of length @xmath23 of the infinite word @xmath27
. the connected component of the rooted graph @xmath120 around the root @xmath119 is exactly the schreier graph of @xmath5 with respect to the stabilizer of @xmath119 .
it follows immediately that the graphs @xmath120 converge to the graph @xmath121 in the pointed gromov - hausdorff topology @xcite . besides the schreier graphs
, we will also work with their subgraphs called tile graphs .
the _ tile graph _
@xmath122 is the graph with the set of vertices @xmath22 , where two vertices @xmath39 and @xmath109 are adjacent if there exists @xmath77 such that @xmath110 and @xmath123 .
the tile graph @xmath124 is thus a subgraph of the schreier graph @xmath125 . to define a tile graph for the action on the space @xmath0 , consider the same set of vertices as in @xmath2 and connect vertices @xmath39 and @xmath109 by an edge if there exists @xmath77 such that @xmath110 and @xmath126 for some finite beginning @xmath127 of the sequence @xmath39 .
the connected component of this graph containing the vertex @xmath27 is called the _ orbital tile graph _
it is clear from the construction that we also have the convergence @xmath129 in the pointed gromov - hausdorff topology .
the study of orbital tile graphs @xmath128 is based on the approximation by finite tile graphs @xmath124 .
namely , we will frequently use the following observation .
every tile graph @xmath124 can be considered as a subgraph of @xmath128 under the inclusion @xmath130 . indeed , if @xmath39 and @xmath109 are adjacent in @xmath124 then @xmath131 and @xmath132 are adjacent in @xmath128
. moreover , every edge of @xmath128 appears in the graph @xmath124 for all large enough @xmath23 .
hence the graphs @xmath124 viewed as subgraphs of @xmath128 form a cover of @xmath128 .
( sidki @xcite ) a finite invertible automaton @xmath16 is called _ bounded _
if one of the following equivalent conditions holds : 1 .
the number of paths of length @xmath23 in @xmath133 is bounded independently on @xmath23 ; 2 .
the number of left- ( equivalently , right- ) infinite paths in @xmath133 is finite ; 3 .
any two nontrivial cycles in the automaton are disjoint and not connected by a path , where a cycle is called trivial if it is a loop at the trivial state ; 4 .
the number of left- ( equivalently , right- ) infinite sequences , which are read along left- ( respectively , right- ) infinite paths in @xmath133 , is finite .
the states of a bounded automaton @xmath16 can be classified as follows : +
a state @xmath79 is _ finitary _ if there exists @xmath42 such that @xmath123 for all @xmath134 ; + a state @xmath79 is _ circuit _ if there exists a nonempty word @xmath134 such that @xmath135 , + i.e. , @xmath79 belongs to a cycle in @xmath16 ; in this case @xmath136 is finitary for every @xmath137 , @xmath138 ; + for every state @xmath79 there exists @xmath42 such that for every @xmath134
the state @xmath139 is + either finitary or circuit .
+ by passing to a power @xmath140 of the alphabet @xmath1 every bounded automaton can be brought to the _ basic form _ ( see ( * ? ? ?
* proposition 3.9.11 ) ) in which the above items hold with @xmath141 ; in particular , all cycles are loops , and @xmath142 for every finitary state @xmath79 and every @xmath6 ( here for @xmath143 we can take an integer number which is greater than the diameter of the automaton and is a multiple of the length of every simple cycle ) .
every self - similar group @xmath5 generated by a bounded automaton is contracting ( see ( * ? ? ?
* theorem 3.9.12 ) ) .
its nucleus is a bounded automaton , which contains only finitary and circuit states ( because every state of a nucleus should have an incoming arrow ) . in this section
we introduce the notion of critical and post - critical sequences that is fundamental for our analysis .
let @xmath5 be a contracting self - similar group generated by an automaton @xmath16 and we assume @xmath145 .
let us describe the vertex sets of the orbital tile graphs @xmath146 .
two right- ( or left- ) infinite sequences are called _ cofinal _ if they differ only in finitely many letters . cofinality is an equivalence relation on @xmath0 and @xmath48 .
the respective equivalence classes are called the cofinality classes and they are denoted by @xmath147 .
the following statement characterizes vertices of tile graphs in terms of the cofinal sequences .
[ cofinality ] suppose that the tile graphs @xmath148 are connected for all @xmath42 .
then for every @xmath3 the cofinality class @xmath149 is the set of vertices of the orbital tile graph @xmath150 . if @xmath151 for @xmath152 and @xmath153 for a finite beginning @xmath154 of @xmath39 then @xmath39 and @xmath109 are cofinal .
conversely , since every graph @xmath124 is connected , for every @xmath155 there exists @xmath7 such that @xmath151 and @xmath156 .
hence @xmath157 for all @xmath158 and every two cofinal sequences belong to the same orbital tile graph .
we classify infinite sequences over @xmath1 as follows .
a left - infinite sequence @xmath159 is called _ post - critical _ if there exists a left - infinite path @xmath160 in the automaton @xmath161 labeled by @xmath162 for some @xmath85 .
the set @xmath163 of all post - critical sequences is called _ post - critical_. a right - infinite sequence @xmath164 is called _ critical _ if there exists a right - infinite path @xmath165 in the automaton @xmath161 labeled by @xmath166 for some @xmath85 .
it follows that every shift @xmath167 of a critical sequence @xmath27 is again critical , and for every @xmath42 there exists @xmath134 such that @xmath168 is critical ( here we use the assumption that every path in @xmath16 can be continued to the left ) .
it is proved in ( * ? ? ?
* proposition iv.18 ) ( see also ( * ? ? ? * proposition 3.2.7 ) ) that the set of post - critical sequences coincides with the set of sequences that can be read along left - infinite paths in the nucleus of the group with removed trivial state .
the same proof works for critical sequences .
therefore the sets of critical and post - critical sequences do not depend on the chosen generating set ( as soon as it satisfies assumption that every state of the automaton @xmath16 has an incoming arrow , and @xmath169 ) .
finally , a sequence @xmath3 is called _ regular _ if the cofinality class of @xmath27 does not contain critical sequences , or , equivalently , if the shifted sequence @xmath167 is not critical for every @xmath170 . notice
that the cofinality class of a critical sequence contains sequences which are neither regular nor critical .
[ prop_properties_of_sequences ] suppose that the automaton @xmath16 is bounded .
then the sets of critical and post - critical sequences are finite .
every post - critical sequence is pre - periodic .
every cofinality class contains not more than one critical sequence .
the cofinality class of a regular sequence contains only regular sequences .
if @xmath27 is regular , then there exists a finite beginning @xmath39 of @xmath27 such that @xmath123 for every @xmath77 .
the number of right- and left - infinite paths avoiding the trivial state is finite in every bounded automaton .
thus the number of critical and post - critical sequences is finite .
the pre - periodicity of post - critical sequences and the periodicity of critical sequences follow from the cyclic structure of bounded automata .
the statement about the cofinality class of a critical sequence follows immediately , because different periodic sequences can not differ only in finitely many letters . finally , if @xmath164 is regular , then starting from any state @xmath77 and following the edges labeled by @xmath171 we will end at the trivial state .
hence there exists @xmath23 such that @xmath172 for all @xmath77 .
note that if the automaton @xmath16 is in the basic form , then every post - critical sequence is of the form @xmath173 or @xmath174 for some letters @xmath175 , and every critical sequence is of the form @xmath176 for some @xmath6 .
let @xmath5 be a group generated by a bounded automaton @xmath16 .
we assume @xmath145 and every state of @xmath16 has an incoming arrow . in what follows
we describe an inductive method ( called _ inflation of graphs _ ) to construct the tile graphs @xmath122 developed in ( * ? ? ?
* chapter v ) .
let @xmath177 be a post - critical sequence .
the vertex @xmath178 of the graphs @xmath125 and @xmath124 will be called _ post - critical_. since the post - critical set @xmath163 is finite , for all large enough @xmath23 , the post - critical vertices of @xmath125 and @xmath124 are in one - to - one correspondence with the elements of @xmath163 ( just take @xmath23 large enough so that @xmath179 whenever @xmath180 ) . hence , with a slight abuse of notations , we will consider the elements of the set @xmath163 as the vertices of the graphs @xmath125 and @xmath124 .
let @xmath181 be the set of all pairs @xmath182 for @xmath183 and @xmath175 such that there exists a left - infinite path in the automaton @xmath16 , which ends in the trivial state and is labeled by the pair @xmath184 .
* theorem v.8)[th_tile_graph_construction ] to construct the tile graph @xmath185 take @xmath186 copies of the tile graph @xmath124 , identify their sets of vertices with @xmath187 for @xmath6 , and connect two vertices @xmath188 and @xmath189 by an edge if and only if @xmath190 and @xmath191 .
the procedure of inflation of graphs given in theorem [ th_tile_graph_construction ] can be described using the graph @xmath17 with the vertex set @xmath192 and the edge set @xmath181 , which we call the _ model graph _ associated to the automaton @xmath16 .
the vertex @xmath193 of @xmath17 is called _ post - critical _ , if the sequence @xmath194 is post - critical .
note that the post - critical vertices of @xmath17 are in one - to - one correspondence with elements of the post - critical set @xmath163 .
now if we `` place '' the graph @xmath124 in the model graph instead of the vertices @xmath195 for each @xmath6 such that the post - critical vertices of @xmath124 fit with the set @xmath195 , we get the graph @xmath185 .
moreover , the post - critical vertices of @xmath17 will correspond to the post - critical vertices of @xmath185 . in order to construct the schreier graph @xmath125
we can take the tile graph @xmath124 and add an edge between post - critical vertices @xmath196 and @xmath197 if @xmath198 for some @xmath77 .
indeed , if @xmath110 and @xmath199 ( the edge that does not appear in @xmath124 ) then @xmath39 and @xmath109 are post - critical vertices , and they should be adjacent in @xmath125 .
notice that there are only finitely many added edges ( independently on @xmath23 ) and they can be described directly through the generating set @xmath16 .
namely , define the set @xmath200 as the set of all pairs @xmath201 for @xmath183 such that there exists a left - infinite path in @xmath16 labeled by the pair @xmath202 or @xmath203 .
then if we take the tile graph @xmath124 and add an edge between @xmath204 and @xmath205 for every @xmath206 , we get the schreier graph @xmath125 .
in this section we present the main results about the number of ends of schreier graphs @xmath28 of groups generated by bounded automata .
our method passes through the study of the same problem for the tile graphs @xmath128 .
first , we show that the number of ends of graphs @xmath128 can be deduced from the number of connected components in tile graphs with a vertex removed ( see proposition [ prop_ends_limit ] ) .
we use the inflation procedure to construct a finite deterministic automaton @xmath207 , which given a sequence @xmath3 determines the number of infinite connected components in the graph @xmath128 with the vertex @xmath27 removed ( see proposition [ prop_infcomp_a_ic ] ) .
then we describe all sequences @xmath3 such that @xmath128 has a given number of ends in terms of sofic subshifts associated to strongly connected components of the automaton @xmath207 .
further we deduce the number of ends of schreier graph @xmath28 ( see corollary [ remarktile ] ) .
after this , we pass to the study of the number of ends for a random schreier graph @xmath28 .
we show that picking randomly an element @xmath27 in @xmath32 , the graph @xmath28 has one or two ends ( see corollary [ cor_one_or_two_ends_a_s ] ) .
the latter case is completely described ( see theorem [ th_classification_two_ends ] ) . in what follows , except for a few special cases directly indicated , we make the following assumptions about the studied self - similar groups @xmath5 and their generating sets @xmath16 : + + 1 . _
the group @xmath5 is generated by a bounded automaton @xmath16 . _ + + 2 .
_ the tile graphs @xmath122 are connected . _
_ every state of the automaton @xmath16 has an incoming arrow , and @xmath169 . _
+ + instead of the assumption 2 it is enough to require that the group acts transitively on @xmath22 for every @xmath208 , i.e. , the schreier graphs @xmath209 are connected .
then , even if the tile graphs @xmath148 are not connected , there is a uniform bound on the number of connected components in @xmath148 ( see how the schreier graphs are constructed from the tile graphs after theorem [ th_tile_graph_construction ] ) , and one can apply the developed methods to each component .
the assumption 3 is technical , it guaranties that every directed path in the automaton @xmath16 can be continued to the left .
if the generating set @xmath16 contains a state @xmath210 , which does not contain incoming edges , then @xmath211 for every @xmath77 and @xmath6 , and hence the state @xmath210 does not interplay on the asymptotic properties of the tile or schreier graphs .
moreover , if the group is self - replicating , then the property 3 is always satisfied , when we take its nucleus @xmath93 as the generating set @xmath16 . for a graph @xmath212 and its vertex @xmath39 we denote by @xmath213 the graph obtained from @xmath212 by removing the vertex @xmath39 together with all edges adjacent to @xmath39 .
let @xmath35 be an infinite , locally finite graph .
a _ ray _ in @xmath35 is an infinite sequence of adjacent vertices @xmath214 in @xmath35 such that @xmath215 for @xmath216 .
two rays @xmath217 and @xmath218 are equivalent if for every finite subset @xmath219 infinitely many vertices of @xmath217 and @xmath218 belong to the same connected component of @xmath220 .
an _ end _ of @xmath35 is an equivalence class of rays . in what follows we use the notation : * @xmath221 is the number of ends of @xmath212 ; * @xmath222 is the number of connected components in the graph @xmath212 ; * @xmath223 is the number of infinite connected components in @xmath212 .
we will show later that the number @xmath223 can be computed for @xmath224 .
the following proposition relates this value to the number of ends of @xmath128 .
[ prop_ends_limit ] every tile graph @xmath128 for @xmath3 has finitely many ends , which is equal to @xmath225 where @xmath59 is the shift map on the space @xmath0 .
let us show that the number of infinite connected components of the graphs @xmath226 and @xmath227 is the same for every @xmath23 .
consider the natural partition of the set of vertices of @xmath128 given by @xmath228 using the graph @xmath128 we construct a new graph @xmath229 with the set of vertices @xmath230 , where two vertices @xmath39 and @xmath109 are adjacent if there exist @xmath231 such that @xmath232 and @xmath233 are adjacent in @xmath128 .
the graph @xmath229 is isomorphic to the tile graph @xmath234 under the identity map on @xmath230 .
indeed , let @xmath39 and @xmath109 be adjacent in @xmath229 .
then there exist @xmath231 and @xmath77 such that @xmath235 and @xmath236 for a finite beginning @xmath237 of @xmath39 .
it follows that @xmath238 , @xmath239 , and @xmath240 , because of self - similarity . therefore @xmath39 and @xmath109 are adjacent in @xmath234 .
conversely , suppose @xmath110 and @xmath241 for some @xmath77 and a finite beginning @xmath237 of @xmath39 .
since each element of @xmath16 has an incoming edge , there exist @xmath242 and @xmath231 such that @xmath243 and @xmath244 .
hence @xmath39 and @xmath109 are adjacent in the graph @xmath229 .
the subgraph of @xmath245 spanned by every set of vertices @xmath246 for @xmath247 is connected , because , by assumption , the tile graphs @xmath124 are connected .
hence , the number of infinite connected components in @xmath227 is equal to the number of infinite connected components in @xmath248 .
in particular , this number is bounded by the size of the generating set @xmath16 . every infinite component of @xmath249
contains at least one end .
hence the estimate @xmath250 holds for all @xmath23 .
in particular @xmath251 for the converse consider the ends @xmath252 of the graph @xmath245 .
they can be made disconnected by removing finitely many vertices .
take @xmath23 large enough so that the set @xmath253 disconnects the ends @xmath254 .
since every end belongs to an infinite component , we get at least @xmath255 infinite components of @xmath248 . in particular , the number of ends is finite and the statement follows . in particular , the number of ends of every tile graph @xmath128 is not greater than the maximal degree of vertices , i.e. , @xmath256 .
now let us show how to compute the number @xmath257 in terms of the components that contain post - critical vertices . in other terms , only components of @xmath258 with post - critical vertices give a positive contribution , in the limit , to the number of infinite components . in order to do that ,
we denote by @xmath259 the number of connected components of @xmath258 that contain a post - critical vertex .
[ prop_infcomp_limit ] let @xmath164 be a regular or a critical sequence , then @xmath259 is an eventually non - increasing sequence and @xmath260 choose @xmath23 large enough so that the subgraph @xmath124 of @xmath128 contains all edges of @xmath128 adjacent to the vertex @xmath27 .
notice that if a vertex @xmath39 of @xmath124 is adjacent to some vertex @xmath261 in @xmath262 , then @xmath199 and thus @xmath39 is post - critical .
it follows that if @xmath263 is a connected component of @xmath258 without post - critical vertices , then all the edges of the graph @xmath264 adjacent to the component @xmath263 are contained in the graph @xmath258 .
hence @xmath263 is a finite component of @xmath264 . therefore the number of infinite components of @xmath264 is not greater than the number of components of @xmath258 that contain a post - critical vertex .
it follows @xmath265 for all @xmath266 .
in fact , if @xmath267 are post - critical and belong to the same connected component of @xmath268 , then there exists a path @xmath269 connecting them such that @xmath270 , in particular the same holds for @xmath271 , @xmath272 .
this implies the monotonicity of @xmath273 for @xmath272 .
suppose now that @xmath27 is regular .
let @xmath263 be a finite component of @xmath264 .
since @xmath263 is finite , every edge inside @xmath263 appears in the graph @xmath124 for all large enough @xmath23 , and thus @xmath263 is a connected component of @xmath258 . since @xmath263 contains only regular sequences , the last statement in proposition [ prop_properties_of_sequences ] implies that for all @xmath77 and every vertex @xmath39 in @xmath263 we have @xmath274 for all large enough @xmath23 . in other words ,
the vertex @xmath275 of @xmath124 is not post - critical for every vertex @xmath39 in @xmath263 .
therefore the component @xmath263 is not counted in the number @xmath259 .
hence @xmath276 for all large enough @xmath23 .
the same arguments work if @xmath27 is critical , because every cofinality class contains not more than one critical sequence , and hence the graph @xmath264 has no critical sequences . with a slight modification the last proposition also works for a sequence @xmath27 , which is not critical but is cofinal to some critical sequence @xmath109 .
in this case , we can count the number of connected components of @xmath258 that contain post - critical vertices other than @xmath277 , and then pass to the limit to get the number of infinite components in @xmath264 .
indeed , it is enough to notice that if the graph @xmath258 contains a connected component @xmath263 with precisely one post - critical vertex @xmath277 for large enough @xmath23 , then @xmath263 is a finite component in the graph @xmath264 . under this modification
the proposition may be applied to any sequence . also to find the number of ends it is enough to know that the limit in proposition [ prop_infcomp_limit ] is valid for regular and critical sequences .
for any sequence @xmath27 cofinal to a critical sequence @xmath109 we just consider the graph @xmath278 centered at the vertex @xmath109 and apply the proposition .
it is not difficult to observe that one can use the same method to obtain the number @xmath279 of all connected components of @xmath264 . in particular
, one has @xmath280 using the iterative construction of tile graphs given in theorem [ th_tile_graph_construction ] we can provide a recursive procedure to compute the numbers @xmath259 .
we will construct a finite deterministic ( acceptor ) automaton @xmath207 with the following structure : it has a unique initial state , each arrow in @xmath207 is labeled by a letter @xmath6 , each state of @xmath207 is labeled by a partition of a subset of the post - critical set @xmath163 .
the automaton @xmath207 will have the property that , given a word @xmath134 , the final state of @xmath207 after reading @xmath39 corresponds to the partition of the post - critical vertices of the graph @xmath124 induced by the connected components of @xmath281 .
then @xmath282 is just the number of parts in this partition .
we start with the following crucial consideration for the construction of the automaton @xmath207 .
let @xmath39 be a vertex of the tile graph @xmath124 .
the components of @xmath281 partition the set of post - critical vertices of @xmath124 .
let us consider only those components that contain at least one post - critical vertex .
let @xmath283 be the set of all post - critical sequences , which represent post - critical vertices in @xmath284-th component . if the vertex @xmath39 is not post - critical , then @xmath285
otherwise , @xmath286 is a proper subset of @xmath163 ; every sequence @xmath196 in @xmath287 represents @xmath39 , i.e. , @xmath288 ( for all large enough @xmath23 the set @xmath287 consists of just one post - critical sequence , while for small values of @xmath23 the same vertex may be represented by several post - critical sequences ) . in any case
, we say that @xmath289 is the _ partition _ ( of a subset of @xmath163 ) _ induced by the vertex _
if @xmath290 does not contain post - critical vertices ( this happens when @xmath288 for every @xmath291 ) , then we say that @xmath39 induce the empty partition @xmath292 .
the set of all partitions induced by the vertices of tile graphs is denoted by @xmath293 .
the set @xmath293 can be computed algorithmically . to see this ,
let us show how , given the partition @xmath294 induced by a vertex @xmath39 and a letter @xmath6 , one can find the partition @xmath295 induced by the vertex @xmath188 .
we will use the model graph @xmath17 associated to the automaton @xmath16 in section [ section_inflation ] , which has the vertex set @xmath192 and edges @xmath181 . recall that the set @xmath163 is identified with the set of post - critical vertices of @xmath17 .
let us construct the auxiliary graph @xmath296 as follows : take the model graph @xmath17 , add an edge between @xmath193 and @xmath297 for @xmath298 and every @xmath284 , and add an edge between @xmath299 and @xmath300 for every @xmath183 and @xmath301 . put @xmath302 .
if the graph @xmath303 contains no post - critical vertices , then we define @xmath304 as the empty partition @xmath292 .
otherwise , we consider the components of @xmath303 with at least one post - critical vertex , and let @xmath305 be the set of all post - critical vertices / sequences in @xmath306-th component .
[ lemma construction aic ] @xmath304 is exactly the partition induced by the vertex @xmath188 .
let us consider the map @xmath307 given by @xmath308 , @xmath291 and @xmath8 .
this map is neither surjective , nor injective in general , nor a graph homomorphism .
however , it preserves the inflation construction of the graph @xmath185 from the graph @xmath124 .
namely , @xmath309 maps each subset @xmath310 into the subset @xmath311 , and the edges in @xmath181 onto the edges of @xmath185 obtained under construction ( see theorem [ th_tile_graph_construction ] ) .
also , by definition of post - critical vertices , the map @xmath309 sends the post - critical vertices of @xmath296 onto the post - critical vertices of @xmath185 . note
that the set @xmath312 is exactly the preimage of the vertex @xmath188 under @xmath309 .
therefore we can consider the restriction @xmath313 .
for every @xmath8 , @xmath314 all vertices in @xmath310 belong to the same component of @xmath303 , and @xmath309 maps these vertices to the same component of @xmath315 , because the subgraph of @xmath185 induced by the set of vertices @xmath311 contains the graph @xmath124 , which is connected . for each @xmath284 the vertices in @xmath316
are mapped to the same component of the graph @xmath315 , because its subgraph induced by the vertices @xmath317 contains the graph @xmath318 and @xmath319 is its connected component .
it follows that , if two post - critical vertices @xmath196 and @xmath197 can be connected by a path @xmath320 in @xmath303 , then for every @xmath284 the vertices @xmath321 and @xmath322 belong to the same component of @xmath315 , and therefore the post - critical vertices @xmath323 and @xmath324 lie in the same component .
conversely , suppose @xmath323 and @xmath324 can be connected by a path @xmath325 in @xmath315 .
we can subdivide @xmath325 as @xmath326 , where @xmath327 and each subpath @xmath254 is a path in a copy of @xmath124 inside @xmath185 .
the preimages of the end points of each @xmath328 belong to the same component in @xmath303 .
therefore @xmath196 and @xmath197 lie in the same component in @xmath303 .
the statement follows .
it follows that we can find the set @xmath293 algorithmically as follows .
note that the empty partition @xmath292 is always an element of @xmath293 ( it is induced by the unique vertex of the tile @xmath329 of zero level ) .
we start with @xmath292 and for each letter @xmath6 construct new partition @xmath304 as above .
we repeat this process for each new partition until no new partition is obtained .
since the set @xmath163 is finite , the process stops in finite time .
then @xmath293 is exactly the set of all obtained partitions .
we construct the ( acceptor ) automaton @xmath207 over the alphabet @xmath1 on the set of states @xmath293 with the the unique initial state @xmath330 .
the transition function is given by the rule : for @xmath331 and @xmath6 we put @xmath332 , where @xmath304 is defined as above . the automaton
@xmath207 has all the properties we described at the beginning of this subsection : given a word @xmath11 , the final state of @xmath207 after accepting @xmath39 is exactly the partition induced by the vertex @xmath39 .
since we are interested in the number @xmath282 of components in @xmath281 containing post - critical vertices , we can label every state of @xmath207 by the number of components in the corresponding partition .
we get the following statement .
[ thm_number_of_con_comp ] the graph @xmath333 has @xmath255 components containing a post - critical vertex if and only if the final state of the automaton @xmath207 after accepting the word @xmath39 is labeled by the number @xmath255 . in particular , for every @xmath255 the set @xmath334 of all words @xmath11 such that the graph @xmath335 has @xmath255 connected components containing a post - critical vertex is a regular language recognized by the automaton @xmath207 .
similarly , we construct a finite deterministic acceptor automaton @xmath336 for computing the number of all components in tile graphs with a vertex removed .
the states of the automaton @xmath336 will be pairs of the form @xmath337 , where @xmath331 and @xmath143 is a non - negative integer number , which will count the number of connected components without post - critical vertices .
we start with the state @xmath338 , which is the unique initial state of @xmath336 , and consequently construct new states and arrows as follows .
let @xmath339 be a state already constructed .
for each @xmath6 we take the graph @xmath303 constructed above , define @xmath304 as above , and put @xmath340 to be equal to the number of connected components in @xmath303 without post - critical vertices .
if @xmath341 is already a state of @xmath342 , then we put an arrow labeled by @xmath13 from the state @xmath337 to the state @xmath341 .
otherwise , we introduce @xmath341 as a new state and put this arrow .
we repeat this process for each new state until no new state is obtained .
since the number of all components in @xmath281 is not greater than @xmath343 , the number @xmath143 can not exceed @xmath343 , and the construction stops in finite time . the automaton @xmath336 has the following property : given a word @xmath11 , the final state of @xmath336 after accepting the word @xmath39 is exactly the pair @xmath337 , where @xmath344 is the partition induced by @xmath39 and @xmath143 is the number of components in @xmath333 without post - critical vertices .
since we are interested only in the number @xmath345 of all components of @xmath281 , we label every state @xmath337 of the automaton @xmath342 by the number @xmath346 , where @xmath255 is the number of sets in the partition @xmath344 .
we get the following statement .
[ prop_number_of_con_comp a_c ] the graph @xmath333 has @xmath255 components if and only if the final state of the automaton @xmath342 after accepting the word @xmath39 is labeled by the number @xmath255 . in particular , for every @xmath255 the set @xmath334 of all words @xmath11 such that the graph @xmath333 has @xmath255 connected components is a regular language recognized by the automaton @xmath342 .
we need the following properties of vertex labels in the automata @xmath207 and @xmath342 .
[ lemma_properties_a_ic ] 1 . in every strongly connected component of the automata @xmath207 and
@xmath336 all states are labeled by the same number .
all strongly connected components of the automaton @xmath207 without outgoing arrows are labeled by the same number .
_ suppose that there is a strongly connected component with two states labeled by different numbers .
it would imply that there exists an infinite word such that the corresponding path in the automaton passes through each of these states infinite number of times .
we get a contradiction with proposition [ prop_infcomp_limit ] , because the sequences @xmath259 and @xmath347 are eventually monotonic ( for the last one the proof is the same ) .
_ suppose there are two strongly connected components in the automaton @xmath207 without outgoing arrows which are labeled by different numbers .
let @xmath39 and @xmath109 be finite words such that starting at the initial state of @xmath207 we end at the first and the second components respectively . then for the infinite sequence @xmath348 the limit in proposition [ prop_ends_limit ] does not exist , and we get a contradiction .
proposition [ prop_infcomp_limit ] together with theorem [ thm_number_of_con_comp ] imply the following method to find the number of infinite components in @xmath264 for @xmath3 .
[ prop_infcomp_a_ic ] let @xmath164 be a regular or critical sequence .
the number of infinite connected components in @xmath264 is equal to the label of a strongly connected component of @xmath207 that is visited infinitely often , when the automaton reads the sequence @xmath27 .
the characterization of the number of infinite components in the graph @xmath264 together with proposition [ prop_ends_limit ] allows us to describe the number of ends of @xmath128 .
since every critical sequence @xmath27 is periodic , we can algorithmically find the number of ends of the graph @xmath128 using proposition [ prop_ends_limit ] and the automaton @xmath207 .
now we introduce some notations .
fix @xmath272 and let @xmath349 be the union of cofinality classes of critical sequences @xmath27 whose tile graph @xmath128 has @xmath255 ends .
similarly we define the sets @xmath350 and @xmath351 .
let * @xmath352 be the subgraph of @xmath207 spanned by the strongly connected components labeled by numbers @xmath353 ; * @xmath354 be the one - sided sofic subshift given by the graph @xmath352 , i.e. , @xmath354 is the set of all sequences that can be read along right - infinite paths in @xmath352 starting at any state ; * @xmath355 be the set of all sequences which are cofinal to some sequence from @xmath354 .
since the set @xmath354 is shift - invariant , the set @xmath355 coincides with @xmath356 .
[ thm_number_of_ends ] the tile graph @xmath128 has @xmath353 ends if and only if @xmath357 .
hence , the tile graph @xmath128 has @xmath255 ends if and only if @xmath358 .
we need to prove that for a regular sequence @xmath27 the graph @xmath128 has at least @xmath255 ends if and only if @xmath359 .
first , suppose @xmath359 .
then @xmath27 is cofinal to a sequence @xmath360 , which is also regular .
the sequences @xmath27 and @xmath361 belong to the same tile graph @xmath362 .
there exists a finite word @xmath39 such that for the sequence @xmath363 the corresponding path in @xmath207 starting at the initial state eventually lies in the subgraph @xmath352 .
then the graph @xmath364 has @xmath353 infinite components by proposition [ prop_infcomp_limit ] .
using the correspondence between infinite components of @xmath365 and of @xmath366 shown in the proof of proposition [ prop_ends_limit ] we get that the graph @xmath365 has @xmath353 infinite components , and hence @xmath367 has @xmath353 ends .
for the converse , suppose the graph @xmath245 has @xmath353 ends and the sequence @xmath27 is regular .
then @xmath368 for some @xmath23 by propositions [ prop_ends_limit ] and [ prop_infcomp_limit ] .
hence some shift @xmath167 of the sequence @xmath27 is in @xmath354 and thus @xmath359 .
example with @xmath369 in section [ section_examples ] shows that we can not expect to get a description using subshifts of finite type , and indeed the description using sofic subshifts is the best possible in these settings .
now we can describe how to derive the number of ends of schreier graphs from the number of ends of tile graphs .
[ prop_schreier_construction ] 1 .
the schreier graph @xmath2 coincides with the tile graph @xmath128 for every regular sequence @xmath3 .
2 . let @xmath3 be a critical sequence , and let @xmath370 be the set of all critical sequences @xmath371 such that @xmath372 for some @xmath7 .
the schreier graph @xmath2 is constructed by taking the disjoint union of the orbital tile graphs @xmath373 for @xmath374 and connecting two critical sequences @xmath375 by an edge whenever @xmath376 for some @xmath77 .
_ if the point @xmath27 is regular , then the set of vertices of @xmath2 is the cofinality class @xmath149 , which is the set of vertices of @xmath128 by proposition [ cofinality ] .
suppose there is an edge between @xmath39 and @xmath109 in the graph @xmath2 .
then @xmath110 for some @xmath77 .
since the sequence @xmath27 is regular , all the sequences in @xmath149 are regular , and hence there exists a finite beginning @xmath154 of @xmath39 such that @xmath126 .
hence there is an edge between @xmath39 and @xmath109 in the tile graph @xmath128 .
if the point @xmath27 is critical , then the set of vertices of @xmath2 is the union of cofinality classes @xmath377 for @xmath374 .
consider an edge @xmath376 in @xmath2 . if this is not an edge of @xmath373 for @xmath374 , then the restriction of @xmath79 on every beginning of @xmath378 is not trivial .
hence @xmath379 are critical , and this edge was added under construction .
the following corollary summarizes the relation between the number of ends of the schreier graphs @xmath2 with the number of ends of the tile graphs @xmath128 .
it justifies the fact that , for our aims , it was enough to study the number of ends and connected components in the tile graphs .
[ remarktile ] 1 .
if @xmath27 is a regular sequence , then @xmath380 .
if @xmath27 is critical , then @xmath381 where the set @xmath370 is from proposition [ prop_schreier_construction ] . using the automata @xmath336 and @xmath207
one can construct similar automata for the number of components in the schreier graphs @xmath125 with a vertex removed .
for every state of @xmath336 or @xmath207 take the corresponding partition of the post - critical set and combine components according to the edges @xmath200 described in the last paragraph in section [ subsection_boundedautomata ] .
for example , if @xmath344 is a state of @xmath207 , then we glue every two components @xmath382 and @xmath383 if @xmath206 for some @xmath384 and @xmath385 .
we get a new partition , and we label the state by the number of components in this partition .
basically , we get the same automata , but vertices may be labeled in a different way .
a case of special interest is when all schreier graphs have one end . in our settings of groups generated by bounded automata ,
our construction enables us to find a necessary and sufficient condition when all schreier graphs @xmath28 have one end .
[ all - one - ended ] all orbital schreier graphs @xmath28 for @xmath3 have one end
if and only if the following two conditions hold : 1 .
all arrows along directed cycles in the automaton @xmath16 are labeled by @xmath386 for some @xmath6 ( depending on an arrow ) ; 2 .
all strongly connected components of the automaton @xmath207 are labeled by @xmath87 ( partitions consisting of one part ) .
let us show that the first condition is equivalent to the property that for every @xmath3 the schreier graph @xmath28 and tile graph @xmath128 coincide .
if there exists a directed cycle that does not satisfy condition 1 , then there exist two different critical sequences @xmath387 that are connected in the schreier graph , i.e. , @xmath388 for some @xmath77 . in this case @xmath389 , because by proposition [ prop_properties_of_sequences ] different critical sequences are non - cofinal , and therefore belong to different tile graphs .
and vice versa , the existence of such critical sequences contradicts condition 1 .
therefore condition 1 implies @xmath390 for every critical sequence @xmath27 , and thus @xmath391 by proposition [ prop_schreier_construction ] .
theorem [ thm_number_of_ends ] implies that condition 2 is equivalent to the statement that every tile graph @xmath128 has one end .
therefore , if the conditions @xmath87 and @xmath392 hold , then any schreier graph coincides with the corresponding tile graph which has one end .
conversely , proposition [ prop_schreier_construction ] implies that if the schreier graph @xmath393 for a critical sequence @xmath27 does not coincide with the corresponding tile graph @xmath245 , then the number of ends of @xmath28 is greater than one .
( the graph @xmath28 is a disjoint union of more than one infinite tile graphs @xmath373 , @xmath374 connected by a finite number of edges . ) therefore , if all schreier graphs @xmath28 have one end , then they should coincide with tile graphs ( condition 1 holds ) and tile graphs have one end ( condition 2 holds ) .
[ remarkanoi ] the hanoi towers group @xmath394 @xcite is an example of a group generated by a bounded automaton for which all orbital schreier graphs @xmath28 have one end . on the other hand ,
this group is not indicable ( since its abelianization is finite ) but can be projected onto the infinite dihedral group @xcite .
this implies that it contains a normal subgroup @xmath395 such that the schreier coset graph associated with @xmath395 has two ends .
clearly , for what said above , @xmath395 does not coincide with the stabilizer of @xmath27 for any @xmath158 .
the structure of the automaton @xmath207 allows to get results about the measure of infinite sequences @xmath3 for which the tile graphs @xmath264 have a given number of infinite components .
we recall that the space @xmath32 is endowed with the uniform measure .
[ rem_subword_and_inf_comp ] it is useful to notice that we can construct a finite word @xmath396 such that starting at any state of the automaton @xmath207 and following the word @xmath109 we end in some strongly connected component without outgoing edges . if these strongly connected components correspond to the partition of the post - critical set on @xmath255 parts , then it follows that @xmath397 for all words @xmath398 with @xmath399 . in other words , @xmath400 for every @xmath39 that contains @xmath109 as a subword . by proposition [ prop_infcomp_limit ] we get the description of sequences which correspond to infinite components using the automaton @xmath207 ( but only for regular and critical sequences ) .
[ cor_inf_comp_almost_all_seq ] the number of infinite connected components of the graph @xmath264 is almost surely the same for all sequences @xmath3 .
this number coincides with the label of the strongly connected components of the automaton @xmath207 without outgoing arrows .
the measure of non - regular sequences is zero . for regular sequences
@xmath3 we can use the automaton @xmath207 to find the number @xmath257 .
then the corollary follows from lemma [ lemma_properties_a_ic ] item 2 and the standard fact that the measure of all sequences that are read along paths in a strongly connected component with an outgoing arrow is zero ( for example , this fact follows from the observation that the adjacency matrix of such a component has spectral radius less than @xmath186 ) .
another explanation comes from remark [ rem_subword_and_inf_comp ] and the fact that the set of all sequences @xmath3 that contain a fixed word as a subword is of full measure .
the corollary does not hold for the number of all connected components of @xmath264 , see examples in section [ section_examples ] .
however , given any number @xmath255 we can use the automaton @xmath336 to compute the measure of the set @xmath334 of all sequences @xmath3 such that the graph @xmath264 has @xmath255 components . as shown in the previous proof only strongly connected components without outgoing arrows contribute the set of sequences with a non - zero measure .
let @xmath401 be the collection of all strongly connected components of @xmath336 without outgoing arrows and labeled by the number @xmath255 .
let @xmath402 be the set of finite words @xmath11 with the property that starting at the initial state and following arrows labeled by @xmath39 we end at a component from @xmath401 , and any prefix of @xmath39 does not satisfy this property .
then the measure of @xmath334 is equal to the sum @xmath403 .
since the automaton @xmath336 is finite , this measure is always a rational number and can be computed algorithmically .
[ section_ends_a_s ] corollary [ cor_inf_comp_almost_all_seq ] together with proposition [ prop_ends_limit ] imply that the tile graphs @xmath128 ( and thus the schreier graphs @xmath2 ) have almost surely the same number of ends , and that this number is equal to the label of the strongly connected components of @xmath207 without outgoing arrows . as was mentioned in introduction
, this fact actually holds for any finitely generated self - similar group , which acts transitively on the levels @xmath22 for all @xmath42 ( see proposition 6.10 in @xcite ) . in our setting of bounded automata we get a stronger description of the sequences @xmath27 for which the tile graph @xmath128 has non - typical number of ends .
[ prop_more_ends ] there are only finitely many schreier graphs @xmath2 and tile graphs @xmath128 with more than two ends .
let us prove that the graph @xmath264 can have more than two infinite components only for finitely many sequences @xmath3 .
suppose not and choose sequences @xmath404 such that @xmath405 , where we take @xmath143 larger than the number of partitions of the post - critical set @xmath163 .
choose level @xmath23 large enough so that all words @xmath406 are different and @xmath407 for all @xmath284 ( it is possible by proposition [ prop_infcomp_limit ] ) .
notice that since the graph @xmath124 is connected , the deletion of different vertices @xmath408 produces different partitions of @xmath163 .
indeed , if @xmath409 with @xmath410 is the partition we got after removing some vertex @xmath39 , then some @xmath411 sets @xmath319 will be in the same component of the graph @xmath412 for any other vertex @xmath109 ( these @xmath411 sets will be connected through the vertex @xmath39 ) .
we get a contradiction with the choice of number @xmath143 .
it follows that there are only finitely many tile graphs with more than two ends .
this also holds for schreier graphs by proposition [ prop_schreier_construction ] .
[ cor_>2ends_pre_periodic ] the schreier graphs @xmath2 and tile graphs @xmath128 can have more than two ends only for pre - periodic sequences @xmath27 .
since the graph @xmath264 can have more than two infinite components only for finitely many sequences @xmath27 , we get that , in the limit in proposition [ prop_ends_limit ] , the sequence @xmath167 attains a finite number of values .
hence @xmath27 is pre - periodic .
example with @xmath369 in section [ section_examples ] shows that the schreier graph @xmath2 and the tile graph @xmath128 may have more than two ends even for regular sequences @xmath27 .
[ cor_one_or_two_ends_a_s ] the tile graphs @xmath128 and schreier graphs @xmath2 have the same number of ends for almost all sequences @xmath3 , and this number is equal to one or two . in this section
we describe bounded automata for which schreier graphs @xmath28 and tile graphs @xmath128 have almost surely two ends .
notice that in this case the post - critical set @xmath163 can not consist of one element ( actually , every finitely generated self - similar group with @xmath413 is finite and can not act transitively on @xmath22 for all @xmath23 ) .
[ lemma_as_two_ends_|p|=2 ] if the schreier graphs @xmath28 ( equivalently , the tile graphs @xmath128 ) have two ends for almost all @xmath3 , then @xmath414 .
we pass to a power of the alphabet so that every post - critical sequence is of the form @xmath173 or @xmath174 for some letters @xmath175 and different post - critical sequences end with different letters .
in particular , every subset @xmath415 for @xmath6 of the model graph @xmath17 contains at most one post - critical vertex of @xmath17 .
we again pass to a power of the alphabet so that for every nontrivial element @xmath77 there exists a letter @xmath6 such that @xmath416 and @xmath142
. then every post - critical sequence @xmath291 appears in some edge @xmath417 of the model graph .
indeed , if the pair @xmath202 is read along a left - infinite path in the automaton @xmath133 that ends in a nontrivial state @xmath79 , then the pair @xmath418 belongs to the edge set @xmath181 of the graph @xmath17 .
now suppose that tile graphs have almost surely two ends .
then the strongly connected components without outgoing arrows in the automaton @xmath207 correspond to the partitions of the post - critical set @xmath163 on two parts ( see corollary [ cor_inf_comp_almost_all_seq ] ) . in particular , there is no state corresponding to the partition of @xmath163 with one part , because such a partition would form a strongly connected component without outgoing arrows ( see the construction of @xmath207 ) .
we will use the fact that all paths in the automaton @xmath207 starting at any partition @xmath419 end in partitions of @xmath163 on two parts ( we can not get more parts ) .
let us construct an auxiliary graph @xmath420 as follows : take the model graph @xmath17 and for each @xmath15 add edges between all vertices in the subset @xmath421 .
we will prove that the graph @xmath420 is an `` interval '' , i.e. , there are two vertices of degree one and the other vertices have degree two , and that two end vertices of @xmath420 are the only post - critical vertices . first , let us show that there are only two subsets @xmath421 for @xmath6 such that the graph @xmath422 is connected .
suppose that there are three such subsets @xmath421 , @xmath423 , @xmath424 .
fix any partition @xmath419 that corresponds to some state of the automaton @xmath207 . consider the arrow in the automaton @xmath207 starting at @xmath425 and labeled by @xmath13 .
this arrow ends in the partition @xmath426 with two parts .
recall how we construct the partition @xmath427 using the graph @xmath428 , and notice that @xmath429 coincides with @xmath422 . using the assumption that the graph @xmath422 is connected , we get that one of the sets @xmath430 is a subset of @xmath421 .
since @xmath421 contains at most one post - critical vertex , the part @xmath430 consists of precisely one element ( post - critical vertex ) , which we denote by @xmath431 , i.e. , here @xmath432 . by the same reason the
subsets @xmath423 and @xmath424 also contain some post - critical vertices @xmath433 and @xmath434 .
notice that the last letters of the sequences @xmath435 are @xmath436 respectively .
we can suppose that the sequences @xmath437 and @xmath438 are different from the sequence @xmath434 ( over three post - critical sequences there are always two with this property ) .
consider the arrow in the automaton @xmath207 starting at the partition @xmath439 and labeled by @xmath440 .
this arrow should end in the partition of @xmath163 on two parts .
since @xmath437 and @xmath434 are different , the post - critical vertex @xmath434 of @xmath428 belongs to the subset @xmath441 .
further , since @xmath434 is the unique post - critical vertex in @xmath424 , there should be no edges connecting the subset @xmath442 with its outside in the graph @xmath428 ( otherwise all post - critical vertices will be in the same component ) .
hence the only edges of the graph @xmath420 going outside the subset @xmath424 should be at the vertex @xmath443 . applying the same arguments to the partition @xmath444 , we get that this unique vertex should be @xmath445 .
hence @xmath446 and we get a contradiction .
so let @xmath421 and @xmath423 be the two subsets such that their complements in the graph @xmath420 are connected .
let @xmath447 and @xmath433 be the post - critical vertices in @xmath421 and @xmath423 respectively . by the same arguments as above , the subset @xmath421 has a unique vertex which is adjacent to a vertex from @xmath422 , and this vertex is of the form @xmath448 or @xmath449 .
the same holds for the subset @xmath423 . every other component @xmath424 contains precisely two vertices @xmath443 and @xmath445 , which have edges going outside the component @xmath424
. however every post - critical sequence appears in one of such edges ( see our assumption in the second paragraph of the proof ) .
hence the post - critical set contains precisely two elements and the structure of the graph @xmath420 follows .
[ cor_one_end_pcset_3_points ] if the post - critical set @xmath163 contains at least three sequences , then the schreier graphs @xmath28 and tile graphs @xmath128 have almost surely one end .
the following example shows that almost all schreier graphs may have two ends for a contracting group generated by a non - bounded automaton , i.e. , by an automaton with infinite post - critical set .
consider the self - similar group @xmath5 over @xmath450 generated by the transformation @xmath447 , which is given by the recursion @xmath451 ( see example 7.6 in @xcite ) .
the group @xmath5 is self - replicating and contracting with nucleus @xmath452 , but the generating automaton is not bounded and the post - critical set is infinite .
every schreier graph @xmath2 with respect to the generating set @xmath453 is a line and has two ends .
[ th_classification_two_ends ] almost all schreier graphs @xmath2 ( equivalently , tile graphs @xmath128 ) have two ends if and only if the automaton @xmath16 brought to the basic form ( see section [ subsection_boundedautomata ] ) is one of the following . 1
. the automaton @xmath16 consists of the adding machine , its inverse , and the trivial element , where the adding machine is an element of type i with a transitive action on @xmath1 ( see figure [ fig_bounded_pcs2 ] , where all edges not shown in the figure go to the identity state , and the letters @xmath13 and @xmath19 are different ) .
2 . there exists an order on the alphabet @xmath454 such that one of the following cases holds . 1
. the automaton @xmath16 consists of elements of types ii and ii@xmath455 ( see figure [ fig_bounded_pcs2 ] ) ; every pair @xmath456 is an orbit of the action of some element of type ii and all nontrivial orbits of such elements on @xmath1 are of this form ; also every pair @xmath457 is an orbit of the action of some element of type ii@xmath455 and all nontrivial orbits of such elements on @xmath1 are of this form ( in particular , @xmath186 is an odd number ) .
the automaton @xmath16 consists of elements of types ii , iii , and iii@xmath455 ; every pair @xmath456 is an orbit of the action of some element of type ii or @xmath458 and all nontrivial orbits of such elements on @xmath1 are of this form ; also every pair @xmath457 is an orbit of the action of some element of type iii@xmath455 and all nontrivial orbits of such elements on @xmath1 are of this form ( in particular , @xmath186 is an even number ) .
moreover , in this case , all schreier graphs @xmath2 are lines except for two schreier graphs @xmath459 and @xmath460 in case 2 ( a ) , and one schreier graph @xmath459 in case 2 ( b ) , which are rays .
recall the definition of the basic form of a bounded automaton from section [ subsection_boundedautomata ] . if a bounded automaton is in the basic form , its post - critical set has size two , and every state has an incoming arrow , then it is not hard to see that the automaton can contain only the states of six types shown in figure [ fig_bounded_pcs2 ] .
we will be using the fact proved in lemma [ lemma_as_two_ends_|p|=2 ] that the modified model graph @xmath420 is an interval .
there are two cases that we need to treat a little bit differently depending on whether both post - critical sequences are periodic or not .
consider the case when both post - critical sequences are periodic , here @xmath461 . in this case
the automaton @xmath16 can contain only the states of types i , i@xmath455 , ii , and ii@xmath455 .
suppose there is a state @xmath447 of type i. it contributes the edges @xmath462 to the graph @xmath420 for every @xmath463 .
if there exists a nontrivial orbit of the action of @xmath447 on @xmath1 , which does not contain @xmath13 , then it contributes a cycle to the graph @xmath420 .
if there exists a fixed point @xmath464 , then under construction of the automaton @xmath207 starting at the partition @xmath465 and following the arrow labeled by @xmath440 we get a partition with one part .
hence the element @xmath447 should act transitively on @xmath1 ( it is the adding machine ) .
every other element of type i should have the same action on @xmath1 , and hence coincide with @xmath447 , otherwise we would got a vertex in the graph @xmath420 of degree @xmath466 .
every element @xmath433 of type i@xmath455 contributes the edges @xmath467 to the graph @xmath420 .
it follows that the action of @xmath433 on @xmath1 is the inverse of the action of @xmath447 ( otherwise we would got a vertex of @xmath420 of degree @xmath468 ) , and hence @xmath433 is the inverse of @xmath447 .
if the automaton @xmath16 additionally contains a state of type ii or ii@xmath455 , then there is an edge @xmath469 or @xmath470 in the graph @xmath420 for some different letters @xmath471 .
we get a vertex of degree @xmath466 , contradiction . hence , in this case , the automaton @xmath16 consists of the adding machine , its inverse , and the identity state .
suppose @xmath16 does not contain states of types i and i@xmath455 . since the post - critical set is equal to @xmath472 the automaton @xmath16 contains states @xmath447 and @xmath433 of types ii and ii@xmath455 respectively .
these elements contribute edges @xmath473 and @xmath474 to the graph @xmath420 .
since the graph @xmath420 should be an interval , these edges should consequently connect all components @xmath475 for @xmath463 ( see figure [ fig_modelline_case 2a ] ) .
it follows that there exists an order on the alphabet such that item @xmath476 holds .
consider the case @xmath477 . in this case
the automaton @xmath16 can consist only of states of types ii , iii and iii@xmath455 .
each state of type ii or iii contributes edges @xmath473 to the graph @xmath420 .
each state of type iii@xmath455 contributes edges @xmath478 .
these edges should consequently connect all components @xmath475 for @xmath463 ( see figure [ fig_modelline_case 2b ] ) .
it follows that there exists an order on the alphabet such that item @xmath479 holds . for the converse
, one can directly check the following facts . in item 1 , every schreier graph @xmath480 is a line . in case 2 ( a ) , the schreier graphs @xmath459 and @xmath460 are rays , while all the other schreier graphs are lines . in case 2 ( b ) , the schreier graph @xmath459 is a ray , while all the other schreier graphs are lines .
dihedral group the grigorchuk group is a nontrivial example satisfying the conditions of the theorem .
it is generated by the automaton @xmath16 shown in figure [ fig_grigorchuk_aut ] .
after passing to the alphabet @xmath481 , the automaton @xmath16 consists of the trivial state and the elements @xmath482 , which are given by the following recursions : @xmath483 we see that this automaton satisfies case 2 ( b ) of the theorem when we choose the order @xmath484 on @xmath1 .
the schreier graph @xmath485 is a ray , while the other orbital schreier graphs are lines . in what follows below
, we give an algebraic characterization of the automaton groups acting on the binary tree , whose orbital schreier graphs have two ends .
it turns out that such groups are those whose nuclei are given by the automata defined by uni in @xcite . in order to show this correspondence we sketch the construction of the groups @xmath486 as in @xcite .
let @xmath487 and @xmath488 be the abelian groups @xmath489 and @xmath490 respectively .
we think of @xmath487 as the field of @xmath392 elements and of @xmath488 as the @xmath255-dimensional vector space over this field .
let @xmath491 be an automorphism of @xmath488 and @xmath492 a surjective homomorphism .
we define the action of elements of @xmath487 and @xmath488 on the binary tree @xmath493 as follows : the nontrivial element @xmath494 only changes the first letter of input words , i.e. , @xmath495 ; the action of @xmath496 is given by the recursive rule @xmath497 .
the automorphism group generated by the action of @xmath487 and @xmath488 is denoted by @xmath486 .
notice that the group @xmath486 is generated by a bounded automaton over the binary alphabet @xmath498 , which we denote by @xmath499 .
if the action of @xmath488 is faithful , the group @xmath486 can be given by an invertible polynomial over the field with two elements , which corresponds to the action of @xmath500 on @xmath488 .
examples include the infinite dihedral group @xmath501 given by polynomial @xmath502 and the gigorchuk group given by polynomial @xmath503 .
the following result puts in relation the groups @xmath486 and the groups whose schreier graphs have almost surely two ends . [ th_sch_2ended_binary ]
let @xmath5 be a group generated by a bounded automaton over the binary alphabet @xmath498 .
almost all schreier graphs @xmath504 have two ends if and only if the nucleus of @xmath5 either consists of the adding machine , its inverse and the identity element , or is equal to one of the automata @xmath505 up to switching the letters @xmath506 of the alphabet .
suppose that the schreier graphs @xmath504 have two ends for almost all sequences @xmath3 and let @xmath93 be the nucleus of the group @xmath5 . by lemma [ lemma_as_two_ends_|p|=2 ]
the post - critical set @xmath163 of the group contains exactly two elements .
if both post - critical elements are periodic , then the nucleus @xmath93 consists of the adding machine , its inverse and the identity element ( see the proof of theorem [ th_classification_two_ends ] ) .
let us consider the case @xmath477 .
we can assume that @xmath507 and @xmath508 so that @xmath509 .
it follows that every arrow along a cycle in @xmath93 is labeled by @xmath510 except for a loop at the trivial state labeled by @xmath511 .
the nucleus @xmath93 contains only one nontrivial finitary element , namely @xmath495 , since otherwise there would be a post - critical sequence with preperiod of length two .
put @xmath512 and @xmath513 .
the set @xmath488 consists exactly of those elements from @xmath5 that belong to cycles . for every @xmath496 we have @xmath514 and @xmath515 , @xmath516 .
it follows that all nontrivial elements of @xmath488 have order two .
let us show that @xmath488 is a subgroup of @xmath5 .
for any @xmath517 there exists @xmath255 such that @xmath518 for @xmath519 .
hence @xmath520 .
therefore @xmath521 belongs to the nucleus @xmath93 and thus @xmath522 .
it follows that @xmath488 is a group , which is isomorphic to @xmath523 for certain @xmath143 .
the map @xmath524 is a homomorphism from @xmath488 to @xmath488 and is bijective , because elements of @xmath488 form cycles in the nucleus .
the map @xmath525 is a surjective homomorphism from @xmath488 to @xmath487 .
we have proved that the nucleus @xmath93 is exactly the automaton @xmath499 . on the other hand ,
let us consider one of the groups @xmath486 .
each element @xmath496 has a cyclic @xmath500-orbit . in the language of automata
this means that @xmath433 belongs to a cycle in the automaton @xmath499 .
we have @xmath515 and @xmath516 for every @xmath496 .
it follows that if @xmath526 and @xmath527 then @xmath109 and @xmath39 are of the form @xmath528 , @xmath529 or @xmath530 , @xmath531 , @xmath532 . therefore the schreier graphs @xmath533 are intervals and the orbital schreier graphs @xmath504 have two ends for almost all sequences @xmath3 .
in this section we first recall the construction of the limit space and tiles of a self - similar group ( see @xcite for more details ) .
then we show how to describe the cut - points of limit spaces and tiles of self - similar groups generated by bounded automata .
let @xmath5 be a contracting self - similar group with nucleus @xmath93 .
[ limitspacedefi ] the _ limit space _ @xmath534 of the group @xmath5 is the quotient of the space @xmath48 by the equivalence relation , where two sequences @xmath51 and @xmath535 are equivalent if there exists a left - infinite path in the nucleus @xmath93 labeled by the pair @xmath162 . the limit space @xmath534 is compact , metrizable , finite - dimensional space . if the group @xmath5 is finitely generated and self - replicating , then the space @xmath534 is path - connected and locally path - connected ( see ( * ? ? ?
* corollary 3.5.3 ) ) . the shift map on the space @xmath536 induces a continuous surjective map @xmath537 .
the limit space @xmath534 comes together with a natural borel measure @xmath538 defined as the push - forward of the uniform bernoulli measure on @xmath48 .
the dynamical system @xmath539 is conjugate to the one - sided bernoulli @xmath186-shift ( see @xcite ) .
the _ limit @xmath5-space _ @xmath540 of the group @xmath5 is the quotient of the space @xmath541 equipped with the product topology of discrete sets by the equivalence relation , where two sequences @xmath542 and @xmath543 of @xmath541 are equivalent if there exists a left - infinite path in the nucleus @xmath93 that ends in the state @xmath544 and is labeled by the pair @xmath162 .
the space @xmath540 is metrizable and locally compact .
the group @xmath5 acts properly and cocompactly on the space @xmath540 by multiplication from the right .
the quotient of @xmath540 by the action of @xmath5 is the space @xmath534 .
the image of @xmath545 in the space @xmath540 is called the _ tile _ @xmath546 of the group @xmath5 .
the image of @xmath547 for @xmath134 is called the _ tile _ @xmath548 of @xmath23-th level .
alternatively , the tile @xmath546 can be described as the quotient of @xmath48 by the equivalence relation , where two sequences @xmath51 and @xmath535 are equivalent if and only if there exists a path in the nucleus @xmath93 that ends in the trivial state and is labeled by the pair @xmath549 .
the push - forward of the uniform measure on @xmath48 defines a measure on @xmath546 .
the tile @xmath546 covers the limit @xmath5-space @xmath540 under the action of @xmath5 .
the tile @xmath546 decomposes in the union @xmath550 of the tiles of @xmath23-th level for every @xmath23 .
all tiles @xmath548 are compact and homeomorphic to @xmath546 . two tiles @xmath548 and @xmath551 of the same level @xmath155 have nonempty intersection if and only if there exists @xmath95 such that @xmath552 and @xmath553 ( see ( * ? ? ?
* proposition 3.3.5 ) ) .
this is precisely how we connect vertices in the tile graph @xmath554 with respect to the nucleus .
hence the graphs @xmath555 can be used to approximate the tile @xmath546 , which justifies the term `` tile '' graph .
the tile @xmath546 is connected if and only if all the tile graphs @xmath556 are connected ( see ( * ? ? ?
* proposition 3.3.10 ) ) ; in this case also @xmath546 is path - connected and locally path - connected .
a contracting self - similar group @xmath5 satisfies the _ open set condition _ if for any element @xmath65 of the nucleus @xmath93 there exists a word @xmath557 such that @xmath156 , i.e. , in the nucleus @xmath93 there is a path from any state to the trivial state .
if a group satisfies the open set condition , then the tile @xmath546 is the closure of its interior , and any two different tiles of the same level have disjoint interiors ; otherwise for large enough @xmath23 there exists a tile @xmath548 for @xmath134 which is covered by other tiles of @xmath23-th level ( see ( * ? ? ?
* proposition 3.3.7 ) ) .
recall that the post - critical set @xmath163 of the group is defined as the set of all sequences that can be read along left - infinite paths in @xmath558 .
therefore , under the open set condition , the boundary @xmath559 of the tile @xmath546 consists precisely of points represented by the post - critical sequences . under the open set condition
, the limit space @xmath534 can be obtained from the tile @xmath546 by gluing some of its boundary points .
namely , we need to glue two points represented by ( post - critical ) sequences @xmath51 and @xmath535 for every path in @xmath558 labeled by @xmath162 .
every self - similar group generated by a bounded automaton is contracting as shown in @xcite , and we can consider the associated limit spaces and tiles . note that every bounded automaton satisfies the open set condition .
the limit spaces of groups generated by bounded automata are related to important classes of fractals : post - critically finite and finitely - ramified self - similar sets ( see ( * ? ? ?
* chapter iv ) ) .
namely , for a contracting self - similar group @xmath5 with nucleus @xmath93 the following statements are equivalent : every two tiles of the same level have finite intersection ( the limit space @xmath534 is finitely - ramified ) ; the post - critical set @xmath163 is finite ( the limit space @xmath534 is post - critically finite ) ; the nucleus @xmath93 is a bounded automaton ( or the generating automaton of the group is bounded ) . under the open set condition , the above statements are also equivalent to the finiteness of the tile boundary @xmath559 .
* iterated monodromy groups .
* let @xmath560 be a complex rational function of degree @xmath561 with finite post - critical set @xmath562 .
then @xmath33 defines a @xmath563-fold partial self - covering @xmath564 of the space @xmath565 .
take a base point @xmath566 and let @xmath567 be the tree of preimages @xmath568 , @xmath170 , where every vertex @xmath569 is connected by an edge to @xmath570 .
the fundamental group @xmath571 acts by automorphisms on @xmath567 through the monodromy action on every level @xmath568 .
the quotient of @xmath571 by the kernel of its action on @xmath567 is called the _ iterated monodromy group _ @xmath572 of the map @xmath33 .
the group @xmath572 is contracting self - similar group and the limit space @xmath534 of the group @xmath572 is homeomorphic to the julia set @xmath573 of the function @xmath33 ( see ( * ? ? ?
* section 6.4 ) for more details ) .
moreover , the limit dynamical system @xmath574 is conjugated to the dynamical system @xmath575 , where @xmath576 is the unique @xmath33-invariant probability measure of maximal entropy on the julia set @xmath573 ( see @xcite ) . in this section
we show how the number of connected components in the orbital schreier and tile graphs with a vertex removed is related to the number of connected components in the limit space and tile with a point removed .
this allows us to get a description of cut - points using a finite acceptor automata as in proposition [ prop_infcomp_a_ic ] .
let @xmath5 be a self - similar group generated by a bounded automaton .
we assume that the tile @xmath546 is connected .
then the nucleus @xmath93 of the group is a bounded automaton , every state of @xmath93 has an incoming arrow , and all the tile graphs @xmath556 are connected .
hence we are in the settings of section [ section_ends ] , and we can apply its results to the tile graphs @xmath124 . since the limit space @xmath534 is obtained from the tile @xmath546 by gluing finitely many specific boundary points ( the post - critical set @xmath163 is finite ) , it is sufficient to consider the problem for the tile @xmath546 , in analogy to what we made before for the schreier and tile graphs .
the tile @xmath546 decomposes into the union @xmath577 , where each tile @xmath578 is homeomorphic to @xmath546 under the shift map .
it follows that , if we take a copy @xmath579 of the tile @xmath546 for each @xmath6 and glue every two points @xmath580 and @xmath581 with the property that there exists a path in the nucleus @xmath93 that ends in the trivial state and is labeled by @xmath582 , and the sequences @xmath51 and @xmath535 represent the points @xmath583 and @xmath584 respectively , then we get a space homeomorphic to the tile @xmath546 .
this is an analog of the construction of tile graphs @xmath124 given in theorem [ th_tile_graph_construction ] .
the edges of the model graph @xmath17 now indicate which points of the copies @xmath579 should be glued .
we consider the tile @xmath546 as its own topological space ( with the induced topology from the space @xmath540 ) , and every tile @xmath548 for @xmath11 as a subset of @xmath546 with induced topology .
hence the boundary of @xmath546 is empty , but the points represented by post - critical sequences we still call the _ boundary points _ of the tile .
every point in the intersection of different tiles @xmath585 of the same level @xmath586 we call _
critical_. these points are precisely the boundary points of the tiles @xmath548 for @xmath11 , and they are represented by sequences of the form @xmath587 for @xmath291 and @xmath11 . in particular , the number of critical points is countable , and hence they are of measure zero .
all other points of @xmath546 we call _
regular_. note that if a regular point @xmath588 is represented by a sequence @xmath51 , then @xmath588 is an interior point of @xmath589 for all @xmath23 .
since each tile @xmath589 is homeomorphic to @xmath546 , the cut - points of @xmath546 also provide information about its local cut - points .
in what follows , let @xmath590 denote the number of connected components in @xmath591 for a point @xmath592 and @xmath593 be the number of components in @xmath591 that contain a boundary point of @xmath546 .
we will show that the numbers @xmath590 and @xmath593 can be computed using a method similar to the one developed in section [ section_ends ] to find the number of components in the tile graphs with a vertex removed .
we start with the following result which is an analog of propositions [ prop_infcomp_limit ] .
[ prop_tile_regular_point ]
let @xmath592 be a regular point represented by a sequence @xmath159 .
then @xmath594 the interior @xmath595 of the tile @xmath548 is the complement to the subset of finitely many points that also belong to other tiles of the same level .
therefore @xmath596 is the union of all tiles @xmath551 for @xmath597 .
since the point @xmath588 is regular , we can choose @xmath23 large enough so that the tile @xmath589 does not contain the boundary points of @xmath546 contained in @xmath591 , and every tile @xmath548 for @xmath134 contains at most one boundary point of @xmath546 . since @xmath588 belongs to the interior @xmath598 of the tile @xmath589 , if two boundary points of @xmath546 lie in the same connected component of @xmath599 , they lie in the same connected component of @xmath591 . therefore the value of the first limit is not less than @xmath593 .
conversely , if two boundary points of @xmath546 lie in the same component of @xmath591 , then for sufficiently large @xmath23 these two points lie in the same components of @xmath599 .
since the number of boundary points is finite , the first equality follows .
for the second equality recall that two tiles @xmath548 and @xmath551 for @xmath600 have nonempty intersection if and only if the vertices @xmath39 and @xmath109 are connected by an edge in the graph @xmath124 .
it follows that if two vertices @xmath39 and @xmath109 belong to the same component in @xmath601 , then the tiles @xmath548 and @xmath551 belong to the same component in @xmath591 .
therefore the value of the second limit is not less than @xmath593 .
conversely , since the point @xmath588 is regular , one can choose @xmath23 large enough so that for any pair of boundary points of @xmath546 that belong to the same connected component in @xmath591 , these points also belong to the same component in @xmath602 .
the second equality follows .
propositions [ prop_infcomp_limit ] and [ prop_tile_regular_point ] establish the connection between the number of components in a punctured tile and the number of infinite components in a punctured tile graph . to describe the limit in proposition [ prop_infcomp_limit ] we constructed the automaton @xmath207 , which returns the number @xmath603 by reading the word @xmath604 from left to right , so that we can apply it to right - infinite sequences .
similarly one can construct a finite automaton @xmath605 , which returns @xmath603 by reading the word @xmath606 from right to left ( the reversion of a regular language is a regular language ) so that we can apply it to left - infinite sequences .
then we can describe the limit in proposition [ prop_tile_regular_point ] in the same way as proposition [ prop_infcomp_a_ic ] describes the limit in proposition [ prop_infcomp_limit ] .
also we can construct a finite deterministic ( acceptor ) automaton @xmath607 with the following property .
the states of @xmath607 are labeled by tuples of the form @xmath608 this tuple indicates the following . for any word @xmath11 let us consider the partition of the set of boundary points of @xmath546 induced by the components of @xmath609 and
let @xmath319 be the set of all post - critical sequences representing the points from the same component .
let @xmath23 be the number of component in @xmath609 without a boundary point of @xmath546 .
similarly , we consider the boundary points of @xmath548 and let @xmath610 be the set of all post - critical sequences @xmath611 such that the points represents by @xmath612 belong to the same component in @xmath613 .
further , we define @xmath614 if the points represented by sequences from @xmath319 and @xmath615 belong to the same connected component .
then the final state of @xmath607 after accepting @xmath39 is labeled exactly by the constructed tuple .
if we are interested just in the number of all components in @xmath609 , we replace the label of each state by the number of components @xmath319 plus @xmath23 .
the following statement follows . for every integer
@xmath255 the set of all words @xmath11 with the property that @xmath616 has @xmath255 connected components is a regular language recognized by the automaton @xmath607 .
this result enables us to provide , in analogy to theorem [ thm_number_of_ends ] , a constructive method which , given a representation @xmath51 of a point @xmath592 , determines the number of components in @xmath591 .
in particular , we get a description of cut - points of @xmath546 .
we distinguish two cases .
_ regular points . _
if a sequence @xmath159 represents a regular point @xmath592 , then @xmath588 is an interior point of the tile @xmath589 for all @xmath208 .
hence , for a regular point @xmath588 , every connected component of @xmath591 intersects the tile @xmath617 .
since the number of boundary points of each tile @xmath548 is not greater than @xmath618 , it follows that @xmath590 coincides with the number of components in the partition of the boundary of the tile @xmath589 in @xmath591 for large enough @xmath23 ( in particular , @xmath619 ) .
the last problem can be subdivided on two subproblems .
first , `` outside '' subproblem : find how the boundary of the tile @xmath589 decomposes in @xmath599 .
the automaton @xmath607 provides an answer to this problem when we trace the second label @xmath304 of states .
second , `` inside '' subproblem : find how the boundary @xmath620 decomposes in @xmath621 .
since each tile @xmath548 is homeomorphic to the tile @xmath546 under the shift map , the automaton @xmath607 provides an answer to this problem when we trace the first label @xmath344 of states ( as well as the automaton @xmath605 ) . by combining the two partitions of @xmath620 given by the two subproblems
, we get the number of connected components in @xmath591
. _ critical points .
_ let @xmath588 be a critical point .
suppose @xmath588 is represented by a post - critical sequence @xmath622 with periodic @xmath623 and @xmath624 .
note that the structure of bounded automata implies that periodic post - critical sequences represent regular points of the tile .
therefore we can apply the previous case to find the number @xmath625 and the partition of the boundary of @xmath546 in @xmath626 , where @xmath627 stands for the point represented by @xmath628 .
now consider the components of @xmath591 .
the only difference with the regular case is that @xmath588 may be not an interior point of @xmath548 for all sufficiently long words @xmath39 .
however , it is an interior point of a union of finitely many tiles , and we will be able to apply the same arguments as above . consider vertices @xmath629 of the model graph @xmath17 adjacent to the vertex @xmath630 .
notice that since @xmath628 is periodic , all @xmath611 are periodic .
the sequences @xmath631 are precisely all sequences that represent the point @xmath588 .
hence the point @xmath588 is an interior point of the set @xmath632 .
we can find how the boundary of the set @xmath633 decomposes in @xmath634 .
since every tile @xmath635 is homeomorphic to the tile @xmath546 via the shift map , we can find @xmath636 , and deduce the decomposition of the boundary of @xmath635 in @xmath637 . combining these partitions
, we can find @xmath590 and the corresponding decomposition of @xmath559 ( in particular , @xmath638 ) . in this way we can find @xmath590 for every point @xmath588 represented by a post - critical sequence .
now let @xmath588 be a critical point represented by a sequence @xmath639 with @xmath291 , @xmath8 , @xmath396 , and @xmath640 .
recall that there is a finite automaton , which reads the word @xmath109 and returns the decomposition of the boundary of @xmath551 in @xmath641 .
using the fact that @xmath642 and @xmath643 are homeomorphic , we can find the number @xmath644 in the same way as we did above for the sequence @xmath645 . since the post - critical set is finite , one can construct a finite automaton , which given a finite word @xmath39 and a post - critical sequence @xmath291 returns the number @xmath646 .
we can use the results from sections [ section_ends_a_s ] and [ section_two_ends_a_s ] to get information about cut - points of the limit space and tile up to measure zero .
[ thm_punctured_tile_ends ] 1 .
the number of connected components in @xmath591 is the same for almost all points @xmath588 , and is equal to one or two .
moreover , @xmath647 almost surely if and only if the schreier graphs @xmath28 ( equivalently , the tile graphs @xmath128 ) have two ends for almost all @xmath3 , and in this case the tile @xmath546 is homeomorphic to an interval .
the number of connected components in @xmath648 is the same for almost all points @xmath588 , and is equal to one or two .
moreover , @xmath649 almost surely if and only if the nucleus @xmath93 of the group satisfies case 2 of theorem [ th_classification_two_ends ] . by corollary [ cor_one_or_two_ends_a_s ]
we have to consider only two cases .
if the tile graphs @xmath128 have one end for almost all @xmath3 , then by remark [ rem_subword_and_inf_comp ] there exists a word @xmath11 such that @xmath650 for all @xmath651 with @xmath652 .
the set of all sequences of the form @xmath653 for @xmath654 and @xmath655 is of full measure .
then by proposition [ prop_tile_regular_point ] we get that all boundary points of the tile @xmath546 belong to the same component in @xmath591 for almost all points @xmath588 .
since every tile @xmath548 is homeomorphic to @xmath546 , we get that the boundary of every tile @xmath548 belongs to the same component in @xmath656 for almost all points @xmath588 .
it follows that @xmath657 almost surely .
since the limit space @xmath534 can be constructed from @xmath546 by gluing finitely many of its points , in this case we get @xmath658 almost surely .
if the tile graphs @xmath128 have two ends almost surely , then we are in the settings of theorem [ th_classification_two_ends ] .
it is direct to check that in both cases of this theorem the tile @xmath546 is homeomorphic to an interval ( because the tile graphs @xmath124 are intervals ) , and hence @xmath647 for almost all points @xmath588 . in case 1
the limit space is homeomorphic to a circle and therefore @xmath658 almost surely . in case 2
the limit space is homeomorphic to an interval and therefore @xmath649 almost surely .
[ cor_limsp_interval_circle ] 1 .
the tile @xmath546 of a contracting self - similar group with open set condition is homeomorphic to an interval if and only if the nucleus of the group satisfies theorem [ th_classification_two_ends ] .
the limit space @xmath534 of a contracting self - similar group with connected tiles and open set condition is homeomorphic to a circle if and only if the nucleus of the group satisfies case 1 of theorem [ th_classification_two_ends ] , i.e. , it consists of the adding machine , its inverse , and the identity state .
the limit space @xmath534 of a contracting self - similar group with connected tiles and open set condition is homeomorphic to an interval if and only if the nucleus of the group satisfies case 2 of theorem [ th_classification_two_ends ] .
let @xmath5 be a contracting self - similar groups with open set condition , and let the tile @xmath546 of the group @xmath5 be homeomorphic to an interval .
then the group @xmath5 has connected tiles and all tiles @xmath548 are homeomorphic to an interval .
the boundary of tiles is finite , hence the group @xmath5 is generated by a bounded automaton ( here we use the open set condition ) , and we are under the settings of this section . since @xmath647 almost surely , the schreier graphs @xmath28 with respect to the nucleus @xmath93 have almost surely two ends , and hence @xmath93 satisfies theorem [ th_classification_two_ends ] .
it is left to prove the statements about limit spaces .
a small connected neighborhood of any point of a circle or of an interval is homeomorphic to an interval .
hence , if the limit space @xmath534 is a circle or an interval , the tile @xmath546 is homeomorphic to an interval
. therefore we are in the settings of theorem [ th_classification_two_ends ] .
as was mentioned above , in case 1 of theorem [ th_classification_two_ends ] the limit space @xmath534 is homeomorphic to a circle , and in case 2 it is homeomorphic to an interval .
the last corollary together with theorem [ th_sch_2ended_binary ] agree with the following result of nekrashevych and uni : the limit dynamical system @xmath659 of a contracting self - similar group @xmath5 is topologically conjugate to the tent map if and only if @xmath5 is equivalent as a self - similar group to one of the automata @xmath505 .
the basilica group @xmath5 is generated by the automaton shown in figure [ fig_basilicaautomaton ] .
this group is the iterated monodromy group of @xmath660 .
it is torsion - free , has exponential growth , and is the first example of amenable but not subexponentially amenable group ( see @xcite ) .
the orbital schreier graphs @xmath2 of this group have polynomial growth of degree @xmath392 ( see ( * ? ? ?
* chapter vi ) ) .
the structure of schreier graphs @xmath2 was investigated in @xcite .
in particular , it was shown that there are uncountably many pairwise non - isomorphic graphs @xmath2 and the number of ends was described .
let us show how to get the result about ends using the developed method .
the alphabet is @xmath498 and the post - critical set @xmath163 consists of three elements @xmath661 , @xmath662 , @xmath663 .
the model graph is shown in figure [ fig_basilicaautomaton ] .
the automata @xmath342 and @xmath207 are shown in figure [ fig_basilica_ac_aic ] .
we get that each tile graph @xmath128 has one or two ends , and we denote by @xmath664 and @xmath665 the corresponding sets of sequences @xmath27 . for the critical sequences @xmath666 the tile graph
@xmath128 has two ends , while for the other critical sequences @xmath667 and @xmath668 the tile graph @xmath128 has one end . using the automaton @xmath207 the sets @xmath664 and @xmath665 can be described by theorem [ thm_number_of_ends ] as follows : @xmath669 almost every tile graph @xmath128 has one end , the set @xmath665 is uncountable but of measure zero .
every graph @xmath264 has one , two , or three connected components , and we denote by @xmath670 , @xmath671 , and @xmath672 the corresponding sets of sequences .
using the automaton @xmath336 these sets can be described precisely as follows : @xmath673 the set @xmath672 is uncountable but of measure zero , while the sets @xmath670 and @xmath671 are of measure @xmath674 .
each graph @xmath264 has one or two infinite components .
the corresponding sets @xmath675 and @xmath676 can be described using the automaton @xmath207 as follows : @xmath677 the set @xmath676 is uncountable but of measure zero .
the finite schreier graph @xmath125 differs from the finite tile graph @xmath124 by two edges @xmath678 and @xmath679 . assuming these edges one can relabel the states of the automaton @xmath336 so that it returns the number of components in @xmath38 . in this way
we get that @xmath680 if the word @xmath39 starts with @xmath681 or @xmath682 ; in the other cases @xmath683 . in particular , the schreier graph @xmath125 has @xmath684 cut - vertices .
the orbital schreier graph @xmath2 coincides with the the tile graph @xmath128 except when @xmath27 is critical .
the critical sequences @xmath685 , @xmath667 , and @xmath668 lie in the same orbit and the corresponding schreier graph consists of three tile graphs @xmath686 , @xmath687 , @xmath688 with two new edges @xmath689 and @xmath690 .
it follows that this graph has four ends .
the limit space @xmath534 of the group @xmath5 is homeomorphic to the julia set of @xmath660 shown in figure [ fig_basilica_graph limitspace ] .
the tile @xmath546 can be obtained from the limit space by cutting the limit space in the way shown in the figure , or , vise versa , the limit space can be obtained from the tile by gluing points represented by post - critical sequences @xmath691 , @xmath692 , @xmath693 .
every point @xmath592 divides the tile @xmath546 into one , two , or three connected components .
put @xmath694 .
then the sets @xmath695 , @xmath696 , and @xmath697 of sequences from @xmath48 , which represent the corresponding cut - points , can be described as follows : @xmath698 the set @xmath697 of three - section points is countable , the set @xmath696 of bisection points is uncountable and of measure zero , and the tile @xmath591 is connected for almost all points @xmath588 .
every point @xmath699 divides the limit space @xmath534 into one or two connected components .
the corresponding sets @xmath700 and @xmath701 can be described as follows : @xmath702 the set @xmath701 of bisection points is uncountable and of measure zero , and the limit space @xmath648 is connected for almost all points @xmath588 .
the gupta - fabrykowski group @xmath5 is generated by the automaton shown in figure [ fig_guptafabrautomaton ] .
it was constructed in @xcite as an example of a group of intermediate growth .
also this group is the iterated monodromy group of @xmath703 ( see ( * ? ? ?
* example 6.12.4 ) ) .
the schreier graphs @xmath2 of this group were studied in @xcite , where their spectrum and growth were computed ( they have polynomial growth of degree @xmath704 ) .
the alphabet is @xmath450 and the post - critical set @xmath163 consists of two elements @xmath705 and @xmath706 .
the model graph is shown in figure [ fig_guptafabrautomaton ] .
the automata @xmath342 and @xmath207 are shown in figure [ fig_guptafabr_ac_aic ] .
every schreier graph @xmath2 coincides with the tile graph @xmath128 . we get that every tile graph @xmath128 has one or two ends , and we denote by @xmath664 and @xmath665 the corresponding sets of sequences . for the only critical sequence @xmath707 the tile graph @xmath128 has one end . using the automaton @xmath207 the sets @xmath664 and @xmath665 can be described by theorem [ thm_number_of_ends ] as follows : @xmath708 almost every tile graph has one end , the set @xmath665 is uncountable but of measure zero .
every graph @xmath264 has one or two connected components , and we denote by @xmath670 and @xmath671 the corresponding sets of sequences . using the automaton @xmath336 these sets can be described precisely as follows : @xmath709 the sets @xmath670 and @xmath671 have measure @xmath710 and @xmath711 respectively .
every graph @xmath264 has one or two infinite components .
the corresponding sets @xmath675 and @xmath676 can be described using the automaton @xmath207 as follows : @xmath712 the set @xmath676 is uncountable but of measure zero .
the limit space @xmath534 and the tile @xmath546 of the group @xmath5 are homeomorphic to the julia set of the map @xmath703 shown in figure [ fig_guptafabr_graph limitspace ] .
every point @xmath699 divides the limit space into one , two , or three connected components .
the sets @xmath695 , @xmath696 , and @xmath697 of sequences from @xmath48 , which represent the corresponding points , can be described as follows : @xmath713 the set @xmath697 of three - section points is countable , the set @xmath696 of bisection points is uncountable and of measure zero , and the limit space @xmath648 is connected for almost all points @xmath588 .
the iterated monodromy group of @xmath43 is generated by the automaton shown in figure [ fig_imgz2iautomaton ] .
this group is one more example of a group of intermediate growth ( see @xcite ) .
the algebraic properties of @xmath369 were studied in @xcite .
the schreier graphs @xmath2 of this group have polynomial growth of degree @xmath714 , where @xmath715 is the real root of @xmath716 ( see ( * ? ? ?
* chapter vi ) ) .
the alphabet is @xmath498 and the post - critical set @xmath163 consists of three elements @xmath717 , @xmath718 , and @xmath719 .
the model graph is shown in figure [ fig_imgz2iautomaton ] .
the automata @xmath342 and @xmath207 are shown in figure [ fig_imgz2i_aic ] .
every schreier graph @xmath2 coincides with the tile graph @xmath128 and it is a tree .
we get that every tile graph @xmath128 has one , two , or three ends , and we denote by @xmath664 , @xmath665 , and @xmath720 the corresponding sets of sequences . using the automaton @xmath207 the sets @xmath664 , @xmath665 , @xmath720 can be described by theorem [ thm_number_of_ends ] as follows . for
the both critical sequences @xmath668 and @xmath667 the tile graph @xmath128 has one end . denote by @xmath721 the right one - sided sofic subshift given by the subgraph emphasized in figure [ fig_imgz2i_aic ] .
then @xmath722 ( we can not describe these sets in the way we did with the previous examples , because the subshift @xmath721 is not of finite type ) .
almost every tile graph has one end , the set @xmath665 is uncountable but of measure zero , and there is one graph , namely @xmath686 , with three ends .
this example shows that corollary [ cor_>2ends_pre_periodic ] may hold for regular sequences ( here @xmath685 is regular ) .
every graph @xmath264 has one , two , or three connected components , and we denote by @xmath670 , @xmath671 , and @xmath672 the corresponding sets of sequences .
using the automaton @xmath336 these sets can be described precisely as follows : @xmath723 the set @xmath672 is of measure zero , and the sets @xmath670 and @xmath671 have measure @xmath724 and @xmath725 respectively .
every graph @xmath264 has one , two , or three infinite components .
the corresponding sets @xmath675 and @xmath676 can be described using the automaton @xmath207 as follows : @xmath726 the set @xmath676 is uncountable but of measure zero .
the limit space @xmath534 and the tile @xmath546 of the group @xmath369 are homeomorphic to the julia set of the map @xmath43 shown in figure [ fig_img_graph limitspace ] .
every point @xmath699 divides the limit space into one , two , or three connected components .
the sets @xmath695 , @xmath696 , and @xmath697 of sequences from @xmath48 , which represent the corresponding points , can be described as follows : @xmath727 where @xmath728 is the left one - sided sofic subshift given by the subgraph emphasized in figure [ fig_imgz2i_aic ] .
the set @xmath697 of three - section points is countable , the set @xmath696 of bisection points is uncountable and of measure zero , and the limit space @xmath648 is connected for almost all points @xmath588 . | every self - similar group acts on the space @xmath0 of infinite words over some alphabet @xmath1 .
we study the schreier graphs @xmath2 for @xmath3 of the action of self - similar groups generated by bounded automata on the space @xmath0 . using sofic subshifts
we determine the number of ends for every schreier graph @xmath2 .
almost all schreier graphs @xmath2 with respect to the uniform measure on @xmath0 have one or two ends , and we characterize bounded automata whose schreier graphs have two ends almost surely . the connection with ( local ) cut - points of limit spaces of self - similar groups is established . *
keywords * : self - similar group , schreier graph , end of graph , bounded automaton , limit space , tile , cut - point * mathematics subject classification 2010 * : 20f65 , 05c63 , 05c25 + |
one of the major thrusts in modern cosmology is an accurate census of star formation and star - forming galaxies in the universe .
this endeavor forms the backbone for a slew of methods ( observational , analytical , and numerical ) to study the process of galaxy formation and evolution over cosmic time . to date
, star - forming galaxies have been selected and studied mainly in two observational windows : the rest - frame ultraviolet ( uv ) , and rest - frame radio and far - infrared ( fir ) . for galaxies at high redshift these bands
are shifted into the optical and radio / sub - mm , allowing observations from the ground .
still , the problem of translating the observed emission to star formation rate ( sfr ) involves large uncertainty .
this is partly because each band traces only a minor portion of the total energy output of stars .
moreover , the optical / uv band is significantly affected by dust obscuration , thus requiring order of magnitude corrections , while the sub - mm and radio bands lack sensitivity , and therefore uncover only the most prodigiously star - forming galaxies .
the main result that has emerged from star formation surveys over the past few years is exemplified in the so - called madau diagram .
namely , the sfr volume density , @xmath10 , rises steeply to @xmath11 , and seemingly peaks at @xmath12 .
there is still some debate about the how steep the rise is , with values ranging from @xmath13 @xcite to @xmath14 ( e.g. @xcite ) .
the evolution beyond @xmath15 is even less clear since optical / uv observations indicate a decline @xcite , while recent sub - mm observations argue for a flat @xmath10 to higher redshift , @xmath16 @xcite .
consistency with this trend can be obtained by invoking large dust corrections in the optical / uv @xcite . for general reviews of star formation surveys
we refer the reader to @xcite , @xcite , and @xcite . despite the significant progress in this field , our current understanding of star formation and its redshift evolution
is still limited by the biases and shortcomings of current optical / uv , sub - mm , and radio selection techniques . in particular , despite the fact that the optical / uv band is sensitive to galaxies with modest star formation rates ( down to a fraction of a m@xmath3 yr@xmath5 ) at high redshift , these surveys may miss the most dusty , and vigorously star - forming galaxies .
moreover , it is not clear if the simple , locally - calibrated prescriptions for correcting the observed _ un - obscured _ sfr for dust extinction ( e.g. @xcite ) , hold at high redshift ; even if they do , these prescriptions involve an order of magnitude correction .
finally , the optical / uv surveys are magnitude limited , and therefore miss the faintest sources .
sub - mm surveys have uncovered a population of highly dust - extincted galaxies , which are usually optically faint , and have star formation rates of several hundred m@xmath3 yr@xmath5 ( e.g. @xcite ) .
however , unlike optical / uv surveys , sub - mm surveys are severely sensitivity limited , and only detect galaxies with @xmath17 l@xmath3 .
more importantly , current sub - mm bolometer arrays ( such as scuba ) have large beams on the sky ( @xmath18 arcsec ) making it difficult to unambiguously identify optical counterparts ( which are usually faint to begin with ) , and hence measure the redshifts @xcite ; in fact , of the @xmath19 sub - mm galaxies identified to date , only a handful have a measured redshift . finally , translating the observed sub - mm emission to a sfr
requires significant assumptions about the temperature of the dust , and the dust emission spectrum ( e.g. @xcite ) .
surveys at decimeter radio wavelengths also suffer from low sensitivity , but the high astrometric accuracy afforded by synthesis arrays such as the vla allows a sub - arcsec localization of the radio - selected galaxies . as a result , it is easier to identify the optical counterparts of these sources .
recently , this approach has been used to pre - select sources for targeted sub - mm observations resulting in an increase in the sub - mm detection rate @xcite and redshift determination @xcite . however , this method is biased toward finding luminous ( high sfr ) sources since it requires an initial radio detection .
an additional problem with radio , even more than with sub - mm , selection is contamination by active galactic nuclei ( agn ) .
an examination of the x - ray properties of radio and sub - mm selected galaxies reveals that of the order of @xmath1 can have a significant agn component @xcite . the most significant problem with current star formation studies , however , is that the link between the optical and sub - mm / radio samples is still not well understood .
the hubble deep field provides a clear illustration : the brightest sub - mm source does not appear to have an optical counterpart @xcite , and only recently a detection has been claimed in the near - ir ( @xmath20 mag ; @xcite ) . along the same line
, sub - mm observations of the optically - selected lyman break galaxies have resulted in very few detections @xcite , and even the brightest lyman break galaxies appear to be faint in the sub - mm band @xcite . in addition , there is considerable diversity in the properties of optical counterparts to sub - mm sources , ranging from galaxies which are faint in both the optical and near - ir ( nir ) to those which are bright in both bands @xcite . as a result of the unclear overlap , and the sensitivity and dust problems in the sub - mm and optical surveys , the fractions of global star formation in the optical and sub - mm / radio selected galaxies
is not well constrained .
it is therefore not clear if the majority of star formation takes place in ultra - luminous galaxies with very high star formation rates , or in the more abundant lower luminosity galaxies with star formation rates of a few m@xmath3 yr@xmath5 .
given the difficulty with redshift identification of sub - mm galaxies , the redshift distribution of dusty star forming galaxies remains highly uncertain .
one way to alleviate some of these problems is to study a sample of galaxies that is immune to the selection biases of current optical / uv and sub - mm / radio surveys , and which may draw a more representative sample of the underlying distribution of star - forming galaxies .
the host galaxies of @xmath0-ray bursts ( grbs ) may provide one such sample .
the main advantages of the sample of grb host galaxies are : ( i ) the galaxies are selected with no regard to their emission properties in any wavelength regime , ( ii ) the dust - penetrating power of the @xmath0-ray emission results in a sample that is completely unbiased with respect to the global dust properties of the hosts , ( iii ) grbs can be observed to very high redshifts with existing missions ( @xmath21 ; @xcite ) , and as a result volume corrections for the star formation rates inferred from their hosts are trivial , ( iv ) the redshift of the galaxy can be determined via absorption spectroscopy of the optical afterglow , or x - ray spectroscopy allowing a redshift measurement of arbitrarily faint galaxies ( the current record - holder is the host of grb990510 with @xmath22 mag and @xmath23 ; @xcite ) , and ( v ) since there is excellent circumstantial evidence linking grbs to massive stars ( e.g. @xcite , the sample of grb hosts is expected to trace global star formation @xcite . of course , the sample of grb hosts is not immune from its own problems and potential biases .
the main problem is the relatively small size of the sample in comparison to both the optical and sub - mm samples ( although the number of grb hosts with a known redshift exceeds the number of sub - mm galaxies with a measured redshift ) . as a result ,
at the present it is not possible to assess the sfr density that is implied by grb hosts , or its redshift evolution .
a bias towards sub - solar metallicity for grb progenitors ( and hence their environments ) has been discussed @xcite , but it appears that very massive stars ( e.g. @xmath24 m@xmath3 ) should produce black holes even at solar metallicity .
the impact of metallicity on additional aspects of grb formation ( e.g. angular momentum , loss of hydrogen envelope ) is not clear at present .
finally , given the observed dispersion in metallicity within galaxies ( e.g. @xcite ) , it is likely that even if grbs require low metallicity progenitors , this does not imply that the galaxy as a whole has a lower than average metallicity . to date ,
grb host galaxies have mainly been studied in the optical and nir bands .
with the exception of one source ( grb020124 ; @xcite ) , every grb localized to a sub - arcsecond position has been associated with a star - forming galaxy @xcite .
these galaxies range from @xmath25 mag , have a median redshift , @xmath26 , and are generally typical of star - forming galaxies at similar redshifts in terms of morphology and luminosity @xcite , with star formation rates from optical spectroscopy of @xmath27 m@xmath3 yr@xmath5 . at the same time , there are hints for higher than average ratios of [ neiii]3869 to [ oii]3727 , possibly indicating the presence of massive stars @xcite .
only two host galaxies have been detected so far in the radio ( grb980703 ; @xcite ) and sub - mm ( grb010222 ; @xcite ) . here
we present sub - mm and radio observations of a sample of @xmath28 grb host galaxies , ranging in redshift from about @xmath29 to @xmath30 ( [ sec : obs ] ) ; one of the 20 sources is detected with high significance in both the sub - mm and radio bands , and an additional source is detected in the sub - mm ( [ sec : res ] ) .
we compare the detected sub - mm and radio host galaxies to local and high-@xmath31 ultra - luminous galaxies in [ sec : sed ] , and derive the sfrs in [ sec : sfr ] .
we then compare the inferred sfrs of the detected host galaxies , and the ensemble of undetected hosts , to optical estimates in [ sec : opt ] .
finally , we compare the optical properties of the grb host galaxies to those of sub - mm and radio selected star - forming galaxies ( [ sec : comp ] ) .
at the time we conducted our survey , the sample of grb host galaxies numbered 25 , twenty of which had measured redshifts .
these host galaxies were localized primarily based on optical afterglows , but also using the radio and x - ray afterglow emission .
of the 25 host galaxies we observed eight in both the radio and sub - mm , seven in the radio , and five in the sub - mm .
the galaxies were drawn from the list of 25 hosts at random , constrained primarily by the availability of observing time .
thus , the sample presented here does not suffer from any obvious selection biases , with the exception of detectable afterglow emission in at least one band .
sub - mm observations of grb afterglows , and a small number of host galaxies have been undertaken in the past . starting in 1997 , @xcite and @xcite
have searched for sub - mm emission from the afterglow of thirteen grbs .
while they did not detect any afterglow emission , these authors used their observations to place constraints on emission from eight host galaxies , with typical @xmath32 rms values of 1.2 mjy .
since these were target - of - opportunity observations , they were not always undertaken in favorable observing conditions .
more recently , @xcite reported targeted sub - mm observations of the host galaxies of four optically - dark grbs ( i.e. grbs lacking an optical afterglow ) .
they focused on these particular sources since one explanation for the lack of optical emission is obscuration by dust , which presumably points to a dusty host .
none of the hosts were detected , with the possible exception of grb000210 ( see [ sec:0210 ] ) , leading the authors to conclude that the hosts of dark bursts are not necessarily heavily dust obscured .
thus , the observations presented here provide the most comprehensive and bias - free search for sub - mm emission from grb host galaxies , and the first comprehensive search for radio emission .
observations in the sub - mm band were carried out using the sub - millimeter common user bolometer array ( scuba ; holland et al .
1999 ) on the james clerk maxwell telescope ( jcmt ) .
we observed the positions of thirteen well - localized grb afterglows with the long ( 850 @xmath33 m ) and short ( 450 @xmath33 m ) arrays .
the observations , summarized in table [ tab : submm ] , were conducted in photometry mode with the standard nine - jiggle pattern using the central bolometer in each of the two arrays to observe the source . in the case of grb000301c
we used an off - center bolometer in each array due to high noise levels in the central bolometer . to account for variations in the sky brightness , we used a standard chopping of the secondary mirror between the on - source position and a position 60 arcsec away in azimuth , at a frequency of 7.8125 hz .
in addition , we used a two - position beam switch ( nodding ) , in which the beam is moved off - source in each exposure to measure the sky .
measurements of the sky opacity ( sky - dips ) were taken approximately every two hours , and the focus and array noise were checked at least twice during each shift .
the pointing was checked approximately once per hour using several sources throughout each shift , and was generally found to vary by @xmath34 arcsec ( i.e. less than one quarter of the beam size ) .
all observations were performed in band 2 and 3 weather with @xmath35 .
the data were initially reduced with the scuba data reduction facility ( surf ) following the standard reduction procedure .
the off - position pointings were subtracted from the on - position pointings to account for chopping and nodding of the telescope .
noisy bolometers were removed to facilitate a more accurate sky subtraction ( see below ) , and the data were then flat - fielded to account for the small differences in bolometer response .
extinction correction was performed using a linear interpolation between skydips taken before and after each set of on - source scans .
in addition to the sky subtraction offered by the nodding and chopping , short - term sky contributions were subtracted by using all low - noise off - source bolometers ( sky bolometers ) .
this procedure takes advantage of the fact that the sky contribution is correlated across the array . as a result
, the flux in the sky bolometers can be used to assess the sky contribution to the flux in the on - source bolometer .
this procedure is especially crucial when observing weak sources , since the measured flux may be dominated by the sky .
we implemented the sky subtraction using surf and our own routine using matlab .
we found that in general the surf sky subtraction under - estimated the sky contribution , and as a result over - estimated the source fluxes .
we therefore used the results of our own analysis routine . for this purpose
we calculated the median value of the two ( three ) outer rings of bolometers in the 850 @xmath33 m ( 450 @xmath33 m ) array , after removing noisy bolometers ( defined as those whose standard deviation over a whole scan deviated by more than 5@xmath36 from the median standard deviation of all sky bolometers ) .
following the sky subtraction , we calculated the mean and standard deviation of the mean ( sdom ) for each source in a given observing shift .
noisy data were eliminated in two ways .
first , the data were binned into 25 equal time bins , and the sdom was calculated step - wise , i.e. at each step the data from an additional bin were added and the mean and sdom were re - calculated . in an ideal situation where the data quality remains approximately constant , the sdom should progressively decrease as more data are accumulated .
however , if the quality of the data worsens ( due to deteriorating weather conditions for example ) the sdom will increase .
we therefore removed time bins in which the sdom increased .
following this procedure , we recursively eliminated individual noisy data points using a sigma cutoff level based on the number of data points ( chauvenet s criterion ; @xcite ) until the mean converged on a constant value . typically , two or three iterations were required , with only a few data points rejected each time . typically , only a few percent of the data were rejected by the two procedures .
finally , flux conversion factors ( fcfs ) were applied to the resulting voltage measurements to convert the signal to jy . using photometry observations of mars and uranus , and/or secondary calibrators ( oh231.8 + 4.2 , irc+10216 , and crl618 )
, we found the fcf to vary between @xmath37 jy / v at 850 @xmath33 m , consistent with the typical value of @xmath38 . at 450 @xmath33 m ,
the fcfs varied between @xmath39 jy / v .
_ very large array ( vla ) : _ we observed @xmath40 grb afterglow positions with the vla from april 2001 to february 2002 .
all sources were observed at 8.46 ghz in the standard continuum mode with @xmath41 mhz bands .
in addition , was observed at 1.43 and 4.86 ghz , and grb0010222 was observed at 4.86 ghz . in table
[ tab : rad ] we provide a summary of the observations . in principle , since the median spectrum of faint radio sources between 1.4 and 8.5 ghz is @xmath42 @xcite , the ideal vla frequency for our observations ( taking into account the sensitivity at each frequency ) is 1.43 ghz .
however , we chose to observe primarily at 8.46 ghz for the following reason . the majority of our observations were taken in the bnc , c , cnd , and d configurations , in which the typical synthesized beam size is @xmath43 arcsec at 1.43 ghz , compared to @xmath44 arcsec at 8.46 ghz .
the large synthesized beam at 1.43 ghz , combined with the larger field of view and higher intrinsic brightness of radio sources at this frequency , would result in a significant decrease in sensitivity due to source confusion .
thus , we were forced to observe at higher frequencies , in which the reduced confusion noise more than compensates for the typical steep spectrum .
we chose 8.46 ghz rather than 4.86 ghz since the combination of @xmath1 higher sensitivity and @xmath45 lower confusion noise , provide a more significant impact than the typical @xmath46 decrease in intrinsic brightness .
the 1.43 ghz observations of were undertaken in the a configuration , where confusion does not play a limiting role . for flux calibration we used the extragalactic sources 3c48 ( j0137 + 331 ) , 3c147 ( j0542 + 498 ) , and 3c286 ( j1331 + 305 ) , while the phases were monitored using calibrator sources within @xmath47 of the survey sources .
we used the astronomical image processing system ( aips ) for data reduction and analysis . for each source we co - added all the observations prior to producing an image , to increase the final signal - to - noise .
_ australia telescope compact array ( atca ) : _ we observed the positions of four grb afterglows during april 2002 , in the 6a configuration at @xmath48 and @xmath49 mhz . using the 6-km baseline resulted in a significant decrease in confusion noise , thus allowing observations at the most advantageous frequencies .
the observations are summarized in table [ tab : rad ] .
we used j1934@xmath50638 for flux calibration , while the phase was monitored using calibrator sources within @xmath47 of the survey sources .
the data were reduced and analyzed using the multichannel image reconstruction , image analysis and display ( miriad ) package , and aips .
the flux measurements at the position of each grb are given in tables [ tab : submm ] and [ tab : rad ] , and are plotted in figure [ fig : fluxes ] . of the @xmath28 sources , only was detected in both the radio and sub - mm with @xmath51 ( [ sec:0418 ] ) .
one additional source , grb000210 , is detected with @xmath51 when combining our observations with those of @xcite .
two hosts have radio fluxes with @xmath52 ( grb000301c and grb000926 ) , but as we show below this is due in part to emission from the afterglow .
the typical @xmath53 thresholds are about @xmath54 mjy , @xmath28 @xmath33jy , and @xmath55 @xmath33jy in the scuba , vla , and atca observations , respectively . in figure
[ fig : fluxes ] we plot all sources with @xmath51 as detections , and the rest as @xmath53 upper limits . in addition , for the sources observed with the atca we plot both the 1.4 ghz upper limits , and the inferred upper limits at 8.46 ghz assuming a typical radio spectrum , @xmath42 @xcite .
one obvious source for the observed radio and sub - mm fluxes ( other than the putative host galaxies ) is emission from the afterglows . to assess the possibility that our observations are contaminated by flux from the afterglows we note that the observations have been undertaken at least a year after the grb explosion . on this timescale
, the sub - mm emission from the afterglow is expected to be much lower than the detection threshold of our observations .
in fact , the brightest sub - mm afterglows to date have only reached a flux of a few mjy ( at 350 ghz ) , and typically exhibited a fading rate of @xmath56 after about one day following the burst @xcite .
thus , on the timescale of our observations , the expected sub - mm flux from the afterglows is only @xmath57 @xmath33jy , well below the detection threshold .
the radio emission from grb afterglows is more long - lived , and hence posses a more serious problem . however , on the typical timescale of the radio observations the 8.46 ghz flux is expected to be at most a few @xmath33jy ( e.g. @xcite ) . in the following sections we discuss the individual detections in the radio and sub - mm , and also provide an estimate for the radio emission from each afterglow .
a source at the position of is detected at four of the five observing frequencies with @xmath51 .
the scuba source , which we designate smm12252 + 2006 , has a flux density of @xmath58 mjy , and @xmath59 mjy .
these values imply a spectral index , @xmath60 ( @xmath61 ) , consistent with a thermal dust spectrum as expected if the emission is due to obscured star formation .
the radio source ( vla122519.26 + 200611.1 ) , is located at @xmath62(j2000)= , @xmath63(j2000)= , with an uncertainty of 0.1 arcsec in both coordinates .
this position is offset from the position of the radio afterglow of grb000418 ( berger et al .
2001 ) by @xmath64 arcsec and @xmath65 arcsec ( figure [ fig : lmap ] ) . in comparison ,
the offset measured from keck and _ hubble space telescope _
images is smaller , @xmath66 arcsec and @xmath67 arcsec .
vla122519.26 + 200611.1 has an observed spectral slope @xmath68 , flatter than the typical value for faint radio galaxies , @xmath69 @xcite , and similar to the value measured for the host of grb980703 ( @xmath70 ; @xcite ) .
the source appears to be slightly extended at 1.43 and 8.46 ghz , with a size of about @xmath71 arcsec ( @xmath72 kpc at @xmath73 ) .
the expected afterglow fluxes at 4.86 and 8.46 ghz at the time of our observations are about @xmath74 and @xmath75 @xmath33jy , respectively @xcite . at 1.43 ghz
the afterglow contribution is expected to be about @xmath75 @xmath33jy based on the 4.86 ghz flux and the afterglow spectrum @xmath76 .
thus , despite the contribution from the afterglow , the radio detections of the host galaxy are still significant at better than @xmath77 level . correcting for the afterglow contribution
we find an actual spectral slope @xmath78 , consistent with the median @xmath69 for 8.46 ghz radio sources with a similar flux @xcite . as with all scuba detections ,
source confusion arising from the large beam ( @xmath79 arcsec at 350 ghz and @xmath80 arcsec at 670 ghz ) raises the possibility that smm12252 + 2006 is not associated with the host galaxy of .
fortunately , the detection of the radio source , which is located @xmath81 arcsec away from the position of the radio afterglow of , indicates that smm12252 + 2006 and vla122519.26 + 200611.1 are in fact the same source the host galaxy of . besides the positional coincidence of the vla and scuba sources , we gain further confidence of the association based on the spectral index between the two bands , @xmath82 .
this spectral index is redshift dependent as a result of the different spectral slopes in the two regimes @xcite .
we find @xmath83 , in good agreement with the @xcite value of @xmath84 ( for the redshift of grb000418 , @xmath73 ) .
we also detect another source , slightly extended ( @xmath85 arcsec ) , approximately 1.4 arcsec east and 2.7 arcsec south of the host of ( designated vla122519.36 + 200608.4 ) , with @xmath86 @xmath33jy and @xmath87 @xmath33jy ( see figure [ fig : lmap ] ) .
this source appears to be linked by a bridge of radio emission ( with @xmath88 at both frequencies ) to the host of .
the physical separation between the two sources , assuming both are at the same redshift , @xmath73 , is 25 kpc .
there is no obvious optical counterpart to this source in _ hubble space telescope _ images down to about @xmath89 mag . based purely on radio source counts at 8.46 ghz @xcite ,
the expected number of sources with @xmath90 @xmath33jy in a 3 arcsec radius circle is only about @xmath91 .
thus , the coincidence of two such faint sources within 3 arcsec is highly suggestive of an interacting system , rather than chance superposition .
interacting radio galaxies with separations of about @xmath28 kpc , and joined by a bridge of radio continuum emission have been observed locally @xcite .
in addition , optical surveys
( e.g. @xcite ) show that a few percent of galaxies with an absolute @xmath92-band magnitude similar to that of the host of , have companions within about @xmath93 kpc .
the fraction of interacting systems is possibly much higher , @xmath94 , in ultra - luminous systems ( such as the host of ) , both locally @xcite and at high redshift ( e.g. @xcite ) .
we note that with a separation of only 3 arcsec , the host of grb000418 and the companion galaxy fall within the scuba beam .
thus , it is possible that smm12252 + 2006 is in fact a superposition of both radio sources .
this will change the value of @xmath95 to about @xmath96 .
the host galaxy of grb980703 has been detected in deep radio observations at 1.43 , 4.86 , and 8.46 ghz @xcite .
the galaxy has a flux @xmath97 @xmath33jy , and a radio spectral slope @xmath98 .
in addition , the radio emission is unresolved with a maximum angular size of @xmath99 arcsec ( @xmath100 kpc ) . based on the @xcite value of @xmath101 ( for the redshift of grb980703 , @xmath102 ) ,
the expected flux at 350 ghz is @xmath103 mjy .
the observed ( @xmath53 ) flux limit @xmath104 mjy , is consistent with the expected value .
grb010222 has been detected in scuba and iram observations with a persistent flux of about 3.5 mjy at 350 ghz and 1 mjy at 250 ghz @xcite .
the persistent emission , as well as the steep spectral slope , @xmath105 , indicated that while the detected emission was partially due to the afterglow of grb010222 , it was dominated by the host galaxy . in fact , accounting for the expected afterglow emission , we find that the host galaxy has a flux , @xmath106 mjy . the radio flux predicted from the sub - mm emission @xcite is @xmath107 @xmath33jy ( for @xmath108 , the redshift of grb010222 ) , which corresponds to @xmath109 @xmath33jy , and @xmath110 @xmath33jy ( assuming @xmath111 ) .
therefore , our measured values , @xmath112 @xmath33jy and @xmath113 @xmath33jy are consistent with the observed sub - mm emission .
the expected afterglow fluxes at 4.86 and 8.46 ghz are @xmath114 and @xmath115 @xmath33jy , respectively , significantly lower than the measured values .
thus , the observed flux mainly arises from the host .
recently , @xcite measured a flux of @xmath116 mjy for grb000210 , in good agreement with our value of @xmath117 mjy .
a weighted - average of the two measurements gives @xmath118 mjy , similar to the sub - mm flux from the host galaxies of grb000418 and grb010222 .
the radio flux at the position of grb000210 is @xmath119 @xmath33jy .
based on a redshift of 0.846 @xcite and the sub - mm detection , the expected radio flux from this source @xcite is @xmath120 @xmath33jy , consistent with the measured value .
the expected flux of the afterglow at the time of the radio observations is less than 1 @xmath33jy @xcite .
following the localization of grb980329 , @xcite observed the afterglow position with scuba and claimed the detection of a source with a 350 ghz flux of about @xmath121 mjy on 1998 , apr .
subsequent observations indicated a fading trend , with a decline to @xmath122 mjy on apr . 6.2 , and @xmath123 mjy ( @xmath53 ) on apr .
11.2 . based on a comparison to the radio flux of the afterglow
, @xcite concluded that the detected sub - mm flux was in excess of the emission from the afterglow itself , and therefore requires an additional component , most likely a host galaxy .
recently , @xcite re - analyzed the scuba data and showed that the initial sub - mm flux was in fact only about @xmath124 mjy , and perfectly consistent with the afterglow flux . as a result
, it is not clear that an additional persistent component is required .
we also re - analyzed the data from apr .
1998 using the method described in
[ sec : submm ] .
we find the following fluxes : @xmath125 mjy ( apr . 5 ) , @xmath126 mjy ( apr . 6 ) , @xmath127 mjy ( apr .
7 ) , @xmath128 mjy ( apr . 8) , and @xmath129 mjy ( apr .
a comparison to the results in @xcite reveals that , with the exception of the last epoch , they over - estimate the fluxes by about @xmath130 mjy .
our observations of grb980329 from september and october of 2001 reveal a flux , @xmath131 mjy , indicating that the flattening to a value of about @xmath132 mjy in the late epochs of the apr .
1998 observations may indicate emission from the host galaxy .
the radio observations are similarly inconclusive , with @xmath133 @xmath33jy .
we estimate that the flux of the afterglow at 8.46 ghz at the time of our observations is only @xmath134 @xmath33jy @xcite .
since the redshift of grb980329 is not known , we can not assess the expected ratio of the radio and sub - mm fluxes .
this source is detected in the vla observations with a flux of @xmath135 @xmath33jy ( @xmath136 ) .
the expected flux from the afterglow at the time of the observations , @xmath137 days after the burst , is @xmath75 @xmath33jy @xcite .
thus , the observed emission exceeds the afterglow flux by @xmath138 . in the calculations below we use a host flux of @xmath139 @xmath33jy .
the vla observations of this grb position reveal a source with @xmath140 @xmath33jy ( @xmath141 ) .
the flux of the afterglow at the time of the observations is about @xmath74 @xmath33jy ( berger et al . 2000 ) .
thus , the excess emission is significant at the @xmath142 level .
the sub - mm emission predicted based on the @xcite relation is @xmath143 mjy ( for @xmath144 , the redshift of grb000301c ) .
this value is in agreement with the measured flux of @xmath145 mjy .
in figure [ fig : sed ] we plot the radio - to - uv spectral energy distributions ( seds ) of the detected host galaxies of grb980703 , grb000418 , and grb010222 , as well as that of arp220 , a proto - typical local ultra - luminous ir galaxy ( ulirg ; @xcite ) , and ero j164502 + 4626.4 ( hur10 ) , a high-@xmath31 analog of arp220 @xcite .
the luminosities are plotted as a function of rest - frame frequencies , to facilitate a direct comparison .
the detected grb hosts are brighter than arp220 ( @xmath146 l@xmath3 , sfr@xmath147 m@xmath3 yr@xmath5 ) , and are similar in luminosity to hur10 ( @xmath148 l@xmath3 , sfr@xmath149 m@xmath3 yr@xmath5 ; @xcite ) . as such , we expect the host galaxies to have star formation rates of a @xmath150 m@xmath3 yr@xmath5 , and luminosities in excess of @xmath151 l@xmath3 .
on the other hand , the optical / nir properties of the detected grb hosts are distinctly different than those of arp220 and hur10 ( as well as other local and high-@xmath31 ulirgs ) .
in particular , from figure [ fig : sed ] it is clear that , while the grb host galaxies are similar to hur10 and arp220 in the radio and sub - mm bands , their optical / nir colors ( as defined for example by @xmath9 ) are much bluer . moreover , while there is a dispersion of a factor of few in the radio and sub - mm bands between the grb hosts , hur10 , and arp220 , the dispersion in the optical / nir luminosity is about two orders of magnitude .
this indicates that there is no simple correlation between the optical / nir properties of grb hosts ( and possibly other galaxies ) and their fir and radio luminosities . in the following sections we expound on both points
to evaluate the star formation rates that are implied by the sub - mm and radio measurements , we use the following expression for the observed flux as a function of sfr @xcite : @xmath152 } { { \rm exp}(0.00083\nu_0)-1 } \ } \frac{(1+z){\rm sfr}}{d_l^2}\,\,{\rm jy}. \label{eqn : sfr}\ ] ] here , @xmath153 ghz , sfr is the star formation rate in m@xmath3 yr@xmath5 , @xmath154 is the luminosity distance in mpc , and @xmath155 is a scaling factor ( of order unity ) which accounts for the difference in the conversion between synchrotron flux and sfr in the milky way and other galaxies . the first term on the right - hand - side accounts for the fact that non - thermal synchrotron emission arising from supernova remnants is proportional to the sfr , while the second term is the contribution of free - free emission from hii regions .
these two flux terms dominate in the radio regime .
the last term in equation . [ eqn : sfr ] is the dust spectrum , which dominates in the sub - mm and fir regimes . in this case , the parameters that have been chosen to characterize the spectrum are a dust temperature , @xmath156 k , and a dust emissivity , @xmath157 , based on a sample of 23 ir - selected starburst galaxies with @xmath158 l@xmath3 @xcite .
we note that other authors ( e.g. @xcite ) favor a lower dust temperature , @xmath159 k , which would result in somewhat different inferred star formation rates . to calculate @xmath154 we use the cosmological parameters @xmath160 , @xmath161 , and @xmath162 km s@xmath5 mpc@xmath5 .
we also use the typical value @xmath69 for the radio measurements @xcite . in figure
[ fig : fluxes ] we plot contours of constant sfr overlaid on the sub - mm and radio flux measurements .
our radio observations are sensitive to galaxies with @xmath163 m@xmath3 yr@xmath5 at @xmath164 , and @xmath165 m@xmath3 yr@xmath5 at @xmath166 .
the sub - mm flux , on the other hand , is relatively constant for a given sfr , independent of @xmath31 .
this is due to the large positive k - correction resulting from the steep thermal dust spectrum .
therefore , at the typical limit of our sub - mm observations we are sensitive to galaxies with sfr@xmath167 m@xmath3 yr@xmath5 . for the host galaxies that are detected with @xmath51 in the sub - mm and radio , as well as those detected in the past ( i.e. grb980703 and grb010222 ) we calculate the following star formation rates : grb000418
@xmath168 m@xmath3 yr@xmath5 , @xmath169 m@xmath3 yr@xmath5 ; grb000210 @xmath170 m@xmath3 yr@xmath5 ; grb010222 @xmath171 m@xmath3 ; grb980703 @xmath172 m@xmath3 yr@xmath5 . here
@xmath173 and @xmath174 are the sfrs derived from the sub - mm and radio fluxes , respectively .
the detections and upper limits from this survey , combined with the detections and upper limits discussed in the literature @xcite indicate that about @xmath1 of all grbs explode in galaxies with star formation rates of @xmath175 m@xmath3 yr@xmath5 . at the same time
, it is clear that @xmath176 of grb host galaxies have more modest star formation rates ,
@xmath177 m@xmath3 yr@xmath5 . despite the fact that the majority of the survey sources are not detected , we can ask the question of whether the grb host galaxies exhibit a significant sub - mm and/or radio emission _ on average_. the weighted average emission from the non - detected sources ( @xmath178 ) is @xmath179 mjy , and @xmath180 @xmath33jy .
this average radio flux is possibly contaminated by flux from the afterglows at the level of about @xmath114 @xmath33jy , so we use @xmath181 @xmath33jy ( @xmath182 ) . therefore , as an ensemble , the grb host galaxies exhibit radio emission , but no significant sub - mm emission . using the median redshift , @xmath183 , for the non - detected sample
, the average radio flux implies an average @xmath184 m@xmath3 yr@xmath5 , while the sub - mm @xmath53 upper limit on @xmath185 is about @xmath186 m@xmath3 yr@xmath5 .
the average sub - mm flux can be compared to @xmath187 mjy for the non - detected sub - mm sources in a sample of radio pre - selected , optically faint ( @xmath188 mag ) galaxies @xcite , @xmath189 mjy for lyman break galaxies @xcite , or @xmath190 mjy for optically - selected starbursts in the hubble deep field @xcite .
thus , it appears that grb host galaxies trace a somewhat fainter population of sub - mm galaxies compared to the radio pre - selected sample , but similar to the lyman break and hdf samples .
this is not surprising given that the radio pre - selection is naturally biased in favor of luminous sources .
we can further extend this analysis by calculating the average sub - mm and radio fluxes in several redshift bins .
here we include both detections and upper limits . from the sub - mm ( radio ) observations we find : @xmath191 mjy ( @xmath192 @xmath33jy ) for @xmath193 , @xmath194 mjy ( @xmath195 @xmath33jy ) for @xmath196 , and @xmath197 mjy ( @xmath198 @xmath33jy ) for @xmath199 .
these average fluxes are marked in figure [ fig : fluxes ] . in the sub - mm
there is a clear increase in the average flux from @xmath200 to @xmath12 , and a flattening or decrease beyond @xmath15 . in the radio , on the other hand , the average flux is about the same in all three redshift bins .
the average radio fluxes translate into the following star formation rates : for @xmath200 the inferred average sfr is @xmath201 m@xmath3 yr@xmath5 , for @xmath202 it is @xmath19 m@xmath3 yr@xmath5 , and for @xmath199 it is @xmath203 m@xmath3 yr@xmath5 ( with @xmath204 significance in each bin ) . the sub - mm observations on the other hand , indicate a rise from a value of @xmath205 m@xmath3 yr@xmath5 for @xmath200 to @xmath206 m@xmath3 yr@xmath5 for @xmath202 , followed by a decline to @xmath207 m@xmath3 yr@xmath5 for @xmath199 .
the typical _ un - obscured _ star formation rates inferred from optical spectroscopy are of the order of @xmath208 m@xmath3 yr@xmath5 ( e.g. @xcite ) . in particular
, the host galaxy of grb980703 has an optical sfr of about @xmath75 m@xmath3 yr@xmath5 @xcite , compared to about @xmath209 m@xmath3 yr@xmath5 from the radio observations .
similarly , the host of has an optical sfr of about @xmath210 m@xmath3 yr@xmath5 @xcite , compared to about @xmath211 m@xmath3 yr@xmath5 based on the radio and sub - mm detections , while the host of grb000210 has an optical sfr of @xmath212 m@xmath3 yr@xmath5 compared to about @xmath213 m@xmath3 yr@xmath5 from the sub - mm observations .
finally , the average radio sfr for the non - detected sources , @xmath214 m@xmath3 yr@xmath5 , significantly exceeds the average optical sfr .
the discrepancy between the optical and radio / sub - mm star formation rates indicates that the majority of the star formation in the grb host galaxies that are detected in the sub - mm and radio is obscured by dust .
it is possible that the same is true for the sample as a whole , but this relies on the less secure average sfr in the non - detected hosts .
the significant dust obscuration is not surprising given that a similar trend has been noted in high-@xmath31 starburst galaxies , for which the typical dust corrections ( based on the uv slope technique ) are an order of magnitude @xcite . in this case
we find similar correction factors .
we can also assess the level of obscuration by comparing the uv luminosity at @xmath215 , @xmath216 , to the bolometric dust luminosity , @xmath217 .
the ratio of these two quantities provides a rough measure of the obscuration , while the sum provides a rough measure of the total star formation rate @xcite .
we estimate @xmath216 using the following host magnitudes : @xmath218 mag ( grb980703 ; @xcite ) , @xmath219 mag ( grb0000210 ; piro et al .
2002 ) , @xmath220 mag ( grb000418 ; @xcite ) , and @xmath221 mag ( grb010222 ; @xcite ) , and assuming a spectrum @xmath222 .
we calculate the following values of @xmath216 : @xmath223 l@xmath3 ( grb980703 ) , @xmath224 l@xmath3 ( grb000210 ) , @xmath225 l@xmath3 ( grb000418 ) , and @xmath226 l@xmath3 ( grb010222 ) .
we estimate @xmath217 from the radio and sub - mm observations , using the conversion factors of @xcite .
the resulting values are : @xmath227 l@xmath3 ( grb980703 ) , @xmath228 l@xmath3 ( grb000210 ) , @xmath229 l@xmath3 ( grb000418 ) , and @xmath230 l@xmath3 ( grb010222 ) .
thus , @xmath231 evaluates to : @xmath28 ( grb980703 ) , @xmath232 ( grb000210 ) , @xmath233 ( grb000418 ) , and @xmath234 ( grb010222 ) .
these results , as well as the sample of starbursts and ulirgs at @xmath164 taken from @xcite are plotted in figure [ fig : lbol ] .
we note that the grb hosts are within the scatter of the @xmath164 sample , with a general trend of increasing value of @xmath231 ( i.e. inceasing obscuration ) with increasing @xmath235 ( i.e. inceasing sfr ) . at the same time
, the particular lines of sight to the grbs within the sub - mm / radio bright host galaxies do not appear to be heavily obscured .
for example , an extinction of @xmath236 mag has been inferred for grb980703 @xcite , @xmath237 mag has been found for grb000418 @xcite , and @xmath238 mag has been found for grb010222 .
the optically - dark grb000210 suffered more significant extinction , @xmath239 mag .
in addition , the small offset of grb980703 relative to its radio host galaxy ( @xmath240 arcsec ; @xmath241 kpc at the redshift of the burst ) , combined with the negligible extinction , indicates that while the burst probably exploded in a region of intense star formation , it either managed to destroy a large amount of dust in its vicinity , or the dust distribution is patchy .
it is beyond the scope of this paper to evaluate the potential of dust destruction by grbs ( see e.g. @xcite ) , but it is clear that the grbs that exploded in the detected sub - mm and radio host galaxies , did not occur in the most heavily obscured star formation sites .
as we noted in [ sec : sed ] , the optical / nir colors of the detected grb host galaxies are bluer than those of arp220 ( @xmath242 mag ) and hur10 ( @xmath243 mag ; @xcite ) . in this section
we compare the @xmath9 color of grb hosts to the @xmath9 colors of radio pre - selected sub - mm galaxies @xcite and sub - mm selected galaxies with a known optical counterpart and a redshift @xcite . in figure
[ fig : color ] we plot @xmath9 color versus redshift for grb hosts and radio pre - selected sub - mm galaxies . before comparing the two populations
, we note that the mean @xmath9 color and redshift for the entire grb sample are @xmath244 mag and @xmath245 , respectively , and for the hosts that are detected in the sub - mm and radio they are @xmath246 mag and @xmath247 , respectively .
thus , there is no clear correlation between the optical / nir colors of the grb hosts and their sub - mm / radio luminosity . for the sub - mm sample the mean @xmath9 color and redshift are @xmath248 mag and @xmath249 , respectively .
to facilitate a more direct comparison with the grb sample we calculate the mean values for the sub - mm sample in the same redshift range as the grb hosts : @xmath250 mag and @xmath251 . moreover ,
if we examine only the host galaxies that were detected in the radio and sub - mm with high significance we find @xmath9 colors of : 2.2 mag ( grb000418 ) , 2.8 mag ( grb980703 ) , 2.1 mag ( grb010222 ) , and 2.6 ( grb000210 ) .
the bluest sub - mm galaxies , on the other hand , have @xmath252 mag . in general , the grb hosts that are detected with @xmath51 in the sub - mm and radio have a distribution of @xmath9 that is indistinguishable from the general grb host galaxy sample .
the obvious difference in @xmath9 color indicates that the grb and radio / sub - mm selections result in a somewhat different set of galaxies .
the red colors of the sub - mm selected galaxies are not surprising since these sources are expected to be dust obscured .
on the other hand , the mean color of the grb hosts is bluer by about 2.5 mag ( @xmath253 significance ) compared to sub - mm galaxies in the same redshift range , indicating a bias towards less dust obscuration . a more patchy dust distribution , or intrinsically bluer colors .
it is possible that there is a bias toward less dust obscuration in the general grb host sample because the bursts that explode in dusty galaxies would have obscured optical afterglows , and hence no accurate localization .
however , this is not a likely explanation since the grbs which exploded in the sub - mm and radio bright hosts are not significantly dust obscured ( [ sec : opt ] ) .
moreover , it does not appear that the hosts of dark grbs are brighter in the sub - mm as expected if the dust obscuration is global @xcite .
finally , the localization of afterglows in the radio and x - rays allows the selection of host galaxies even if they are dusty .
in particular , the only two grbs in which significant obscuration of the optical afterglow has been inferred ( grb970828 : @xcite ; grb000210 : piro et al .
2002 ) , have been localized thanks to accurate positions from the radio and x - ray afterglows , and have host galaxies with @xmath9 colors of 3.7 and 2.6 mag , not significantly redder than the general population of grb hosts .
therefore , a bias against dust obscured host galaxies is not the reason for the bluer color of the sample .
an alternative explanation is that the distribution of dust in grb hosts is different than in the radio pre - selected and sub - mm selected galaxies
. this may be in terms of a spatially patchy distribution , which will allow more of the uv light to escape , or a different distribution of grain sizes ( i.e. a different extinction law ) , possibly due to a different average metallicity .
however , in both cases it is not clear why there should be a correlation between the dust properties of the galaxy and the occurence of a grb .
finally , it is possible that grb host galaxies are preferentially in an earlier stage of the star formation ( or starburst ) process . in this case
, a larger fraction of the shorter - lived massive stars would still be shining , and the overall color of the galaxy would be bluer relative to a galaxy with an older population of stars . one way to examine the age of the stellar population
is to fit population synthesis models to the broad - band optical / nir spectra of the host galaxies .
this approach has recently been used by @xcite who find some evidence that the age of the stellar population in some grb host galaxies ( including the host of grb980703 ) is relatively young , of the order of @xmath254 myr .
this result is also expected
if grbs arise from massive stars , as indicated by recent observations ( e.g. @xcite ) , since in this case grbs would preferentially select galaxies with younger star formation episodes .
regardless of the exact reason for the preferential selection of bluer galaxies relative to the radio pre - selected sub - mm population , two results seem clear : ( i ) the grb host galaxies detected in the sub - mm and radio are likely drawn from a population that is generally missed in current sub - mm surveys , and ( ii ) grb host galaxies may not be a completely bias - free sample .
the first point is particularly interesting in light of the fact that optical observations of these host galaxies do not identify them as particularly exceptional in terms of sfr .
therefore , while similar galaxies are not necessarily missed in optical surveys , their star formation rates are likely under - estimated .
we presented the most comprehensive scuba , vla , and atca observations of grb host galaxies to date .
the host galaxy of is the only source detected with high significance in both the sub - mm and radio , while the host galaxy of grb000210 is detected with @xmath255 in the sub - mm when we combine our observations with those of @xcite .
when taken in conjunction with the previous detections of grb980703 in the radio @xcite and grb010222 in the sub - mm @xcite , these observations point to a @xmath256 detection rate in the radio / sub - mm .
this detection rate confirms predictions for the number of sub - mm bright grb hosts , with @xmath257 mjy , based on current models of the star formation history assuming a large fraction of obscured star formation @xcite .
the host galaxies detected in the sub - mm and radio have star formation rates from about @xmath258 to @xmath259 m@xmath3 yr@xmath5 , while statistically the non - detected sources have an _ average _ sfr of about @xmath6 m@xmath3 yr@xmath5 . these star formation rates exceed the optically - inferred values by over an order of magnitude , pointing to significant dust obscuration within the grb host galaxies detected in the sub - mm and radio , and possibly the sample as a whole . still , the optical afterglows of the bursts that exploded in the sub - mm / radio bright host galaxies did not suffer significant extinction , indicating that : ( i ) the grbs did not explode in regions where dust obscuration is significant , or ( ii ) the uv and x - ray emission from the afterglow destroys a significant amount of dust in the local vicinity of the burst .
we have also shown that grb host galaxies , even those detected in the sub - mm / radio , have bluer @xmath9 colors compared to galaxies selected in the sub - mm or radio bands in the same redshift range .
this is not the result of an observational bias against dusty galaxies in the grb host sample since the afterglows of grbs which exploded in the radio / sub - mm bright hosts were not significantly obscured .
more likely , this is the result of younger stellar populations in these galaxies , or possibly a patchy dust distribution .
if the reason is younger stellar population then this provides additional circumstantial evidence in favor of massive ( and hence short - lived ) stars as the likely progenitors of grbs .
a potential bias of the grb host galaxy sample is that the popular `` collapsar '' model of grbs calls for high mass , low metallicity stellar progenitors @xcite .
this may result in preferential selection of low metallicity ( and hence less dusty ) host galaxies .
however , it appears that grb progenitors can even have solar metallicity , and that a very low metallicity is unfavored by the required initial conditions for a grb explosion .
moreover , studies of the milky way ( see @xcite for a recent review ) , local galaxies ( e.g @xcite ) , and high-@xmath31 galaxies
( e.g. @xcite ) , indicate that there are considerable variations in metallicity within galaxies
. this may be especially true if several independent episodes of star formation have occured within the galaxy .
thus , even if there is a bias towards low metallicity for grb progenitors ( and hence their immediate environments ) it is not obvious that this introduces a bias in the host galaxy sample .
nonetheless , while the observations presented in this paper clearly indicate the potential of grb selection of high-@xmath31 galaxies for the study of star formation , a much larger sample is required to complement existing optical and sub - mm surveys .
this may become possible in the near future with the upcoming launch ( sep .
2003 ) of swift . with an anticipated rapid ( @xmath260 minute ) and accurate localization of about 150 bursts per year
, the grb - selected sample will probably increase to several hundred galaxies over the next few years .
the rapid localization would most likely result in a large fraction of redshift measurements thanks to the bright optical afterglows .
in addition to the localization of a large number of grb hosts , the study of these galaxies ( as well as those in other samples ) would greatly benefit from the advent of new facilities , such as sirtf , alma , evla , and the ska . in figure
[ fig : sirtf ] we again plot the rest - frame seds of arp220 and the sub - mm / radio bright grb hosts .
overplotted on these seds are the @xmath32 sensitivities of sirtf , alma , and the evla for 200-sec exposures at redshifts 1 and 3 , as well as the sensitivities of current instruments ( vla and scuba ) .
the contributions of these new facilities to star formation studies are threefold : ( i ) increased sensitivity , ( ii ) increased resolution , and ( iii ) increased frequency coverage .
these improvements will serve to ameliorate the main limitations of present radio , sub - mm , and ir observations ( [ sec : intro ] ) , by allowing the detection of more representative star forming galaxies at high redshift , in addition to a better constraint on the total dust bolometric luminosity and accurate localizations , which would facilitate follow - ups at optical wavelengths . in conjunction with increasingly larger samples of galaxies selected in the optical , the radio / sub - mm / ir , and by grbs
, the future of star formation studies is poised for great advances and new discoveries .
we thank k. adelberger , a. blain , and a. shapley for helpful discussions , and g. moriarty - schieven for help with the data reduction .
we also thank s. chapman for providing us with the optical / nir colors and redshifts of radio pre - selected sub - mm galaxies prior to publication . , j. r. 1982 , an introduction to error analysis .
the study of uncertainties in physical measurements , : a series of books in physics , oxford : university press , and mill valley : university science books , 1982 ) .
lclrrr grb970228 & 0.695 & nov . 1 , 2001 & @xmath261 & @xmath262 & + & & nov . 2 , 2001 & @xmath263 & @xmath264 & @xmath265 + grb970508 & 0.835 & sep . 9 , 2001 & @xmath266 & @xmath267 & + & & sep . 10 , 2001 & @xmath268 & @xmath269 & + & & sep . 12 , 2001 & @xmath270 & @xmath271 & @xmath272 + grb971214 & 3.418 & nov . 2 , 2001 & @xmath273 & @xmath274 & @xmath273 + grb980329 & & sep . 13 , 2001 & @xmath275 & @xmath276 & + & & oct . 29 , 2001 & @xmath277 & @xmath278 & @xmath279 + grb980613 & 1.096 & nov . 1 , 2001 & @xmath280 & @xmath281 & + & & nov . 2 , 2001 & @xmath282 & @xmath283 & + & & dec . 7 , 2001 & @xmath284 & @xmath285 & @xmath286 + grb980703 & 0.966 & sep . 10 , 2001 & @xmath287 & @xmath288 & + & & sep . 12 , 2001 & @xmath289 & @xmath290 & @xmath291 + grb991208 & 0.706 & dec . 6 , 2001 & @xmath292 & @xmath293 & + & & dec . 7 , 2001 & @xmath294 & @xmath295 & @xmath296 + grb991216 & 1.020 & oct . 31 , 2001 & @xmath297 & @xmath298 & + & & nov . 3 , 2001 & @xmath299 & @xmath300 & + & & nov . 4 , 2001 & @xmath301 & @xmath302 & @xmath303 + grb000210 & 0.846 & sep . 12 , 2001 & @xmath304 & @xmath305 & + & & sep . 13 , 2001 & @xmath306 & @xmath307 & + & & sep . 14 , 2001 & @xmath308 & @xmath309 & @xmath310 + grb000301c & 2.034 & dec . 29 , 2001 & @xmath311 & @xmath312 & + & & dec . 30 , 2001 & @xmath313 & @xmath314 & @xmath315 + grb000418 & 1.119 & oct . 30 , 2001 & @xmath316 & @xmath317 & + & & oct . 31 , 2001 & @xmath318 & @xmath319 & + & & nov . 1 , 2001 & @xmath320 & @xmath321 & @xmath322 + grb000911 & 1.058 & sep . 13 , 2001 & @xmath323 & @xmath324 & + & & sep . 14 , 2001 & @xmath325 & @xmath326 & + & & oct . 31 , 2001 & @xmath327 & @xmath328 & + & & nov . 3 , 2001 & @xmath329 & @xmath330 & + & & nov . 4 , 2001 & @xmath331 & @xmath332 & @xmath333 + grb011211 & 2.140 & dec . 29 , 2001 & @xmath334 & @xmath335 & + & & dec . 30 , 2001 & @xmath336 & @xmath337 & + & & dec . 31 , 2001 & @xmath338 & @xmath339 & @xmath340 lcclcr 0.1 in grb970828 & 0.958 & vla & jun . 47 , 2001 & 8.46 & @xmath341 + grb980329 & & vla & jul . 22 sep . 10 , 2001 & 8.46 & @xmath342 + grb980613 & 1.096 & vla & may 1826 , 2001 & 8.46 & @xmath343
+ grb981226 & & vla & jul .
24 oct . 15 , 2001 & 8.46 & @xmath344 + grb991208 & 0.706 & vla & apr . 14 jul . 20 , 2001 & 8.46 & @xmath345 + grb991216 & 1.020 & vla & jun . 8 jul . 13 , 2001 & 8.46 & @xmath346 + grb000210 & 0.846 & vla & sep . 16 oct . 12 , 2001 & 8.46 & @xmath347 + grb000301c & 2.034 & vla & jun .
15 jul . 22 , 2001 & 8.46 & @xmath348 + grb000418 & 1.119 & vla & jan . 14 feb . 27 , 2002 & 1.43 & @xmath349 + & & vla & dec . 8 , 2001 jan . 10 , 2002 & 4.86 & @xmath350 + & & vla & may 28 jun . 3 , 2001 & 8.46 & @xmath351 + grb000911 & 1.058 &
vla & mar .
21 apr . 2 , 2001 & 8.46 & @xmath352 + grb000926 & 2.037 & vla & jun .
11 jul . 12 , 2001 & 8.46 & @xmath353 + grb010222 & 1.477 & vla & sep . 29 oct .
13 , 2001 & 4.86 & @xmath354 + & & vla & jun . 24 aug . 27 , 2001 & 8.46 & @xmath355 + grb990510 & 1.619 & atca & apr . 28 , 2002 & 1.39 & @xmath356 + grb990705 & 0.840 & atca & apr .
2122 , 2002 & 1.39 & @xmath357 + grb000131 & 4.5 & atca & apr .
28 , 2002 & 1.39 & @xmath358 + grb000210 & 0.846 & atca & apr . 2728 , 2002 & 1.39 & @xmath359 lccc 0.1 in grb970228 & @xmath360 & & 1 + grb970508 & @xmath361 & & 1 + grb970828 & & @xmath362 & 1.2 + grb971214 & @xmath363 & & 3 + grb980329@xmath364 & @xmath365 & @xmath366 & + grb980613 & @xmath367 & @xmath368 & + grb980703 & @xmath361 & @xmath369 & 10 + grb981226@xmath370 & & @xmath371 & + grb990510 & & @xmath372 & + grb990705 & & @xmath373 & + grb991208 & @xmath374 & @xmath375 & 20 + grb991216 & @xmath376 & @xmath377 & + grb000131 & & @xmath378 & + grb000210 & @xmath379 & @xmath380 & 3 + grb000301c & @xmath381 & @xmath382 & + grb000418 & @xmath383 & @xmath384 & 55 + grb000911 & @xmath385 & @xmath386 & 2 + grb000926 & & @xmath387 & + grb010222 & @xmath388 & @xmath389 & 1.5 + grb011211 & @xmath390 & & | we present the first comprehensive search for submillimeter and radio emission from the host galaxies of twenty well - localized @xmath0-ray bursts ( grbs ) . with the exception of a single source ,
all observations were undertaken months to years after the grb explosions to ensure negligible contamination from the afterglows .
we detect the host galaxy of grb000418 in both the sub - mm and radio , and the host galaxy of grb000210 only in the sub - mm .
these observations , in conjunction with the previous detections of the host galaxies of grb980703 and grb010222 , indicate that about @xmath1 of grb host galaxies are ultra - luminous ( @xmath2 l@xmath3 ) and have star formation rates of about @xmath4 m@xmath3 yr@xmath5 . as an ensemble ,
the non - detected hosts have a star formation rate of about @xmath6 m@xmath3 yr@xmath5 ( @xmath7 ) based on their radio emission .
the detected and ensemble star formation rates exceed the optically - determined values by an order of magnitude , indicating significant dust obscuration . in the same vein ,
the ratio of bolometric dust luminosity to uv luminosity for the hosts detected in the sub - mm and radio ranges from @xmath8 , and follows the known trend of increasing obscuration with increasing bolometric luminosity .
we also show that , both as a sample and individually , the grb host galaxies have bluer @xmath9 colors as compared with galaxies selected in the sub - mm in the same redshift range .
this possibly indicates that the stellar populations in the grb hosts are on average younger , supporting the massive stellar progenitor scenario for grbs , but it is also possible that grb hosts are on average less dusty . beyond the specific results presented in this paper , the sub - mm and radio observations serve as an observational proof - of - concept in anticipation of the upcoming launch of the swift grb mission and sirtf .
these new facilities will possibly bring grb host galaxies into the forefront of star formation studies . |
effect of spin ordering on the symmetry and lattice parameters of magnetic crystals , termed as magnetoelasticity , has recently attracted remarkable interest in multiferroic materials and complex magnets .
this is because magnetoelasticity is closely related to the magnetocapacintace and other magnetoelectric phenomena @xcite .
the dynamical coupling between lattice vibrations and spin - wave excitations plays also a crucial role in the electromagnon excitations of manganese oxides @xcite . for several chromium spinel oxides , the materials in the focus of the present study ,
multiferroic ordering has been reported at low temperatures and strong magnetic field induced changes were also found in their dielectric response @xcite . in order to reveal the nature and origin of spin - phonon coupling governing magneto - elasticity , optical probe of lattice vibrations has been shown to be an efficient tool for a broad class of chromium spinel oxides and chalcogenides with non - magnetic a - site ions @xcite . at high temperatures acr@xmath0o@xmath1 chromium spinel oxides
have the normal cubic spinel structure corresponding to the space group of fd@xmath3 m as shown in fig . [ fig01](a ) . within this structure
a@xmath2 ions form a diamond lattice with tetrahedral oxygen environment , while magnetic cr@xmath4 ions surrounded by octahedral oxygen cages form a pyrochlore sublattice . in the local crystal field of oxygen ions ,
the 3d orbitals of a cr@xmath4 ion are split into a low - lying @xmath5 triplet and a higher - energy @xmath6 doublet , while the orbitals of an a@xmath2 ion are split into a lower @xmath7 doublet and a higher @xmath8 triplet as illustrated in fig .
[ fig01](b ) .
the presence of magnetic moment and orbital degeneracy on the a - site ions can change the magnetic and structural properties in chromium spinel oxides in two ways as compared to those members of the acr@xmath0o@xmath1 family where cr@xmath4 is the only magnetic ion .
first , the antiferromagnetic j@xmath9 exchange interaction between neighboring cr@xmath4 ions leads to a highly frustrated ground state on the pyrochlore sublattice .
this frustration can only be released by the so - called spin - jahn - teller effect for compounds with nonmagnetic a - site ions , where lattice distortions induce differences between the exchange coefficients originally uniform for each cr@xmath4-cr@xmath4 pairs .
this removes the spin degeneracy of the ground state and consequently long - range magnetic order can develop as was observed at low temperatures in case of zncr@xmath0o@xmath1 @xcite and cdcr@xmath0o@xmath1 @xcite .
in contrast , if a - site ions are magnetic , the coupling of s=3/2 cr@xmath4 spins to the a - site spins on the bipartite diamond lattice removes the magnetic frustration and the magnetic ordering occurs at much higher temperatures comparable to the energy scale of the exchange interactions .
second , if a - site ions have orbital degeneracy in the cubic spinel structure , it is lifted by a cooperative jahn - teller distortion resulting to a tetragonal structure with space group i4@xmath10/amd already in the high - temperature paramagnetic phase @xcite and the onset of magnetic order may result in further reduction of the lattice symmetry through a spin - lattice coupling . on the other hand , chromium oxide spinels with magnetic a sites but without orbital degeneracy
are reported to nearly maintain the cubic structure even in their low - temperature magnetic state @xcite . besides spinel chromites , the impact of orbital degeneracy on the magnetic and structural properties is clearly manifested in vanadium oxide spinels , where both diamond and pyrochlore sublattices can host orbital degeneracy .
the most peculiar example is fev@xmath0o@xmath1 @xcite , which exhibits successive structural transitions due to the interplay between the jahn - teller effect on the two sublattices and the influence of the spin order . in the present paper
we perform a systematic optical study of acr@xmath0o@xmath1 chromium spinels having magnetic ions also on the a site ( a = mn , fe , co , ni and cu ) . in sequence of increasing number of 3d electrons on the a - site ions , a = fe , ni and
cu compounds are jahn - teller active ions , that is they show orbital degeneracy , while mncr@xmath0o@xmath1 and cocr@xmath0o@xmath1 spinels have no orbital degrees of freedom ( see fig .
[ fig01](b ) ) .
our main purpose is to reveal the key parameters responsible for the lowering of the lattice symmetry associated with their magnetic ordering , i.e. the origin of magnetoelasticity in these compounds , by an infrared optical study of their lattice dynamics .
we also investigate the intra - atomic _ 3d_-_3d _ transitions of the a - site ions to reveal their electronic state of these magnetic ions .
kaplan and menyuk @xcite proposed that for j@xmath11 ( exchange coupling between spins of neighboring a@xmath2 and cr@xmath4 ions ) sufficiently strong relative to j@xmath9 , the magnetic ground state becomes unique and non - collinear magnetic orders with three different magnetic sublattices can develop .
this has been verified experimentally for all these compounds .
cocr@xmath0o@xmath1 exhibits ferrimagnetic order below t@xmath12=93k and an incommensurate conical spin - order develops under t@xmath13=27k , which becomes commensurate to the lattice at t@xmath14=13k @xcite . in the conical magnetic phase the presence of spontaneous electric polarization and ferrotoroidicity were reported @xcite . in mncr@xmath0o@xmath1
the ferrimagnetic transition occurs at t@xmath12=48k , while the onset of the long - range conical order takes place at t@xmath13=14k @xcite .
similarly , fecr@xmath0o@xmath1 becomes a ferrimagnet below t@xmath12=93k and the conical order sets in at t@xmath13=35k @xcite .
ferroelectric polarization was observed in this material below t@xmath12 @xcite .
the magnetic structures of nicr@xmath0o@xmath1 and cucr@xmath0o@xmath1 are rather different from the former compounds .
according to magnetization measurements nicr@xmath0o@xmath1 becomes a collinear ferrimagnet at t@xmath12=74k , where the simultaneous lowering of the lattice symmetry to orthorhombic was verified by x - ray scattering measurements @xcite .
below t@xmath13=31k this compound exhibits a yafet - kittel - type canted ferrimagnetic order @xcite . in cucr@xmath0o@xmath1
the yafet - kittel - type magnetic order emerges right from the paramagnetic state at t@xmath12=152k @xcite .
o@xmath1 spinels the cr@xmath4 ions ( light blue spheres ) and the a@xmath2 ions ( grey spheres ) are located on a pyrochlore and a diamond lattice , respectively .
the oxygens ( red spheres ) form octahedral environment around cr@xmath4 ions , while a@xmath2 ions are surrounded by tetrahedral oxygen cages .
( b ) electron configuration of cr@xmath4 ion and a - site ions with increasing electron number on the 3d shell . here
, the splitting due to the local crystal field of the ions in the cubic phase and the effect of hund coupling are only taken into account . ]
single crystals of acr@xmath0o@xmath1 for a = mn , fe , co with typical linear dimensions of 300 - 600@xmath15 were synthesized by chemical vapour transport , while cucr@xmath0o@xmath1 was made by flux deposition method as described elsewhere @xcite .
we adopted flux deposition method of b. m. wanklyn et al .
@xcite using solely pbo flux in order to grow nicr@xmath0o@xmath1 .
unpolarized normal incidence reflectivity spectra were studied on the ( 111 ) surface for a = mn , fe , co , ni and on the ( 010 ) surface of cucr@xmath0o@xmath1 over the temperature range of t=10 - 335k with a bruker ifs 66v / s fourier - transform ir spectrometer coupled to a microscope ( bruker hyperion ) .
the lattice vibrations were found to be sensitive to polishing induced mechanical strain , thus , reflectivity spectra were measured on as - grown surfaces at low energies , @xmath16=100 - 7700@xmath17 .
the surface roughness of mncr@xmath0o@xmath1 and nicr@xmath0o@xmath1 samples demanded to use of polished surfaces in order to avoid light scattering at higher energies .
correspondingly , reflectivity spectra were also recorded over @xmath16=1700 - 24000@xmath17 on polished surfaces .
reflectivity spectrum of each compound was measured up to @xmath16=250000@xmath17 at room temperature @xcite .
optical conductivity spectra were obtained from the reflectivity data by kramers - kronig transformation .
the low - energy part of the measured reflectivity spectra was extrapolated to zero photon energy as a constant value , while the high - energy part was assumed to follow the free electron model above @xmath16=1000000@xmath17 .
the advantage of the optical conductivity spectra versus the reflectivity spectra is that phonon excitations appear as sharp peaks in the former , while close resonances are more difficult to resolve in the reflectivity data . in the analysis of the optical phonon modes we fitted the corresponding low - energy part of the optical conductivity spectra with a sum of lorentzian peaks according to @xmath18 where @xmath19 , @xmath20 and @xmath21 stand for the oscillator strength , the resonance frequency and the damping rate of the modes , @xmath22 is the vacuum permittivity and @xmath23 is the dielectric constant at high energies .
the measured reflectivity and the corresponding optical conductivity spectra are plotted over common energy scales for acr@xmath0o@xmath1 spinels ( a = mn , fe , co , ni and cu ) in fig .
[ fig02 ] at t=10k . at the high - energy side the upturn of the optical conductivity corresponds to a band gap of @xmath242.5ev characteristic to these chromium spinel oxides .
the gap value is not significantly influenced by the variation of the a - site ions , therefore , the lowest - energy interband transition is mainly related to the o 2p @xmath25 cr 3d charge transfer excitations . below the optical gap
weak excitations exist in the visible and near infrared region . since these structures are sensitive to the change of the a - site ion they are assigned as intra - atomic _ 3d_-_3d _ transitions of the tetrahedrally coordinated a cations in agreement with former optical studies @xcite .
a broad and low - intensity hump at e=1.6ev is the only structure common in each spectra .
this feature is due to the _
3d_-_3d _ transitions of the cr@xmath4 ions .
finally , in the low - energy range ( @xmath16=100 - 700@xmath17 ) optical phonon excitations characteristic to the lattice structure are present . in the following sections dependence of the phonon excitations and
the _ 3d_-_3d _ transitions on the a - site cations and the temperature are studied . according to powder x - ray diffraction measurements the lattice symmetry for a = mn , fe
, co corresponds to the cubic space group @xmath26 at room temperature , while in the case of a = ni , cu the symmetry is already tetragonal with the space group @xmath27 @xcite .
o@xmath1 at t=300k and 10k contain four distinct phonon modes .
neither the ferrimagnetic transition at t@xmath12=51k nor the onset of conical spin order at t@xmath13=14k are accompanied with an observable splitting of these @xmath28 phonon modes .
note that the conductivity scales for panel ( a ) and ( b ) differ by more than an order of magnitude . ]
o@xmath1 at t=300k and 10k .
similarly to mncr@xmath0o@xmath1 , no splitting of the four @xmath29phonon modes can be traced either below the ferrimagnetic transition at t@xmath30=93k or upon the conical spin ordering at t@xmath31=26k . ]
the factor group analysis for cubic spinels ( fd@xmath3 m space group , m3 m ( o@xmath32 ) point group ) yields 16 optical phonon modes in the @xmath33 point : @xmath34 where abbreviations ir , r and s refer to infrared - active , raman - active and silent modes , respectively .
thus , in the cubic phase four @xmath28 modes are expected to appear in the optical spectra , all of which have three - fold degeneracy . when the lattice symmetry is reduced to tetragonal each of these @xmath35 lattice vibrations are split into an @xmath36 singlet and an @xmath37 doublet mode .
furthermore , the two originally silent @xmath38 phonon modes can split into an infrared - active @xmath37 doublet and a silent @xmath39 singlet .
if the symmetry of the lattice further lowered the degeneracy of each mode is completely lifted .
optical conductivity spectra of mncr@xmath0o@xmath1 and cocr@xmath0o@xmath1 , spinels with no orbital degrees of freedom , are shown in fig .
[ fig04 ] and fig .
[ figa01 ] , respectively .
all four infrared active @xmath28 optical phonon modes expected in the cubic phase are observed at room temperature for both compounds . as the temperature is decreased to t=10k the lifetime of each phonon mode
is increased by @xmath4030% .
the detailed temperature dependence of the resonance frequencies was determined by fitting the spectra according to eq .
[ eq : oscmodel ] and the corresponding results are summarized in fig . [ fig09 ] .
all the modes show weak hardening towards lower temperatures followed by a tiny softening associated with the ferrimagnetic transition .
additionally , in mncr@xmath0o@xmath1 the first and second mode ( located at @xmath40195@xmath17 and @xmath40376@xmath17 , respectively ) are shifted toward higher energies upon the conical spin ordering at @xmath41=14k .
although in cocr@xmath0o@xmath1 an anomalous broadening was observed for the third and fourth mode below t@xmath12=93k , no clear sign of phonon splitting can be discerned in either of the two compounds within the resolution of our experiment , @xmath42 .
these results are compatible with a very weak , if any , distortion of the cubic lattice in mncr@xmath0o@xmath1 and cocr@xmath0o@xmath1 even at t=10k . in fecr@xmath0o@xmath1 and nicr@xmath0o@xmath1 spinel oxides with jahn - teller active a@xmath2 ions ,
the presence of four @xmath28 phonon modes above room temperature , respectively followed in fig .
[ figa02 ] and fig .
[ fig07 ] , indicates the stability of the high - symmetry cubic spinel structure . however , upon the jahn - teller transition of fecr@xmath0o@xmath1 at t@xmath43=135k the lowest- and the highest - energy mode , while in nicr@xmath0o@xmath1 below t@xmath43=320k the two high - energy modes are split indicating structural transition to a tetragonal state .
o@xmath1 at various temperatures in the vicinity of the jahn - teller transition at t@xmath43=135k , the ferrimagnetic ordering at t@xmath30=93k and the onset of the conical order at t@xmath31=35k .
lowering of the crystal symmetry can be traced as a temperature induced splitting of the lowest- and the highest - energy phonon modes both at t@xmath43 and t@xmath12 , while splitting of the other two @xmath28 modes can not be resolved . ] besides the splitting of the third and fourth mode in nicr@xmath0o@xmath1 , a new excitation appears below the jahn - teller transition at @xmath44=545@xmath17 ( see fig . [ fig07](c ) ) .
this fifth mode has low oscillator strength , short lifetime and can not be deduced from either of the cubic @xmath28 phonons .
however , in the tetragonal phase the originally silent @xmath45 phonon mode can split into an infrared - active @xmath37 doublet and a silent @xmath46 singlet beside the splitting of the @xmath28 modes to @xmath36 @xmath47 @xmath37 .
we assign this weak mode around @xmath44=545@xmath17 to such an @xmath37 mode originating from one of the two @xmath45 modes silent in the cubic symmetry .
another mode with similarly small intensity emerges just below t@xmath43 in the vicinity of the third @xmath28 mode .
we assign it as the other @xmath37 mode activated by the tetragonal splitting .
the fact that the third @xmath28 is split into two branches at t@xmath43 and becomes three non - degenerate modes with comparable oscillator strength below t@xmath12 highly support this assignment , i.e. it shows the independent origin of this weak mode .
o@xmath0 at various temperatures over the region of the jahn - teller transition at t@xmath48k , the ferrimagnetic ordering at t@xmath49k and the onset of the canted magnetic structure at t@xmath50k .
splitting of the third and the fourth @xmath28 modes can be observed both at t@xmath43 and t@xmath12 .
scale of low energy part is magnified ten times larger .
( c ) between these modes the activation of an @xmath51 mode , which originates from a @xmath45 mode being silent in the cubic phase , can be followed .
rigorous analysis shows the activation of another @xmath51 mode upon @xmath52 , which is located among the branches of the third @xmath28 mode .
( d ) oscillator strength for this @xmath51 mode and for the branches of the third @xmath28 mode as a function of temperature . ]
o@xmath1 at various temperatures .
at room temperature all t@xmath53 phonon modes are split due to the jahn - teller transition at t@xmath54854k . for each mode
the remaining degeneracy is lifted upon the magnetic ordering at t@xmath55k . ]
the optical conductivity spectra of cucr@xmath0o@xmath1 are shown in fig .
[ figa03 ] .
each @xmath28 mode is already split to @xmath36 @xmath47 @xmath37 at room temperature in accordance with its cubic to tetragonal phase transition at t@xmath43=854k .
the remaining twofold degeneracy of each @xmath37 is lifted upon the ferrimagnetic transition .
in contrast to nicr@xmath0o@xmath1 , no activation of silent modes can be followed . a summary about the temperature dependence of the phonon frequencies
are plotted in fig .
[ fig08 ] for the spinel oxides with orbital degeneracy , a = fe , ni and cu .
the ferrimagnetic phase transition completely removes the threefold degeneracy of some of the cubic @xmath28 modes in all the three compounds , therefore , the cubic symmetry of the lattice is lower than tetragonal in the ground state of these compounds .
the modes whose degeneracy is partly lifted by the jahn - teller distortion exhibit further splitting at the subsequent magnetic transition at t@xmath12 except for the highest - energy mode in nicr@xmath0o@xmath1 .
the magnetically induced splitting is as large as @xmath56 / @xmath16 @xmath5710@xmath58 for the lower - energy vibrations .
the sensitivity of the phonon modes to the magnetic ordering depends on the chemical composition ; in fecr@xmath0o@xmath1 the first and the fourth modes , in nicr@xmath0o@xmath1 the third mode , while in cucr@xmath0o@xmath1 all modes are split . furthermore , the ferroelectric phase transition in fecr@xmath0o@xmath1 is accompanied by a sudden hardening of the lower - energy modes .
the partial activation of two silent @xmath45 vibrations was only observed in nicr@xmath0o@xmath1 . among these new @xmath51
modes the vibration around @xmath44=545@xmath17 is sensitive to the magnetic ordering as it shows a splitting in the vicinity of t@xmath12 . as already reported in the literature @xcite , the lowest - energy mode in this compound exhibit an anomalous softening with decreasing temperature .
such unusual temperature dependence of the low - energy vibrations was also found in fev@xmath0o@xmath1 @xcite .
weak in - gap excitations observed over the mid - infrared visible range in acr@xmath0o@xmath1 spinels are interpreted as transitions within the _ 3d _ levels of the a - site cations split by the local crystal field , the coulomb interaction and the spin - orbit coupling @xcite . in acr@xmath0o@xmath1 compounds
these parity - forbidden transitions become allowed due to the lack of inversion symmetry in the local tetrahedral environment which provides the possibility to investigate the electronic structure of the a site ions by optical spectroscopy .
in the subsequent section the optical conductivity spectra are analysed in terms of ligand field theoretical calculations .
we observed _
3d_-_3d _ excitations in the optical conductivity spectra of a = cu , ni , co and a = fe chromium spinel oxides , while no such transition was found in mncr@xmath0o@xmath1 ( see fig . [ fig02 ] ) .
the absence of this excitation in the latter compound indicates that high - spin state is realized for the a - site since there is no spin - allowed _
3d_-_3d _ transition within the half - filled d - band of the mn@xmath2 ion in the presence of strong hund - coupling as shown in fig .
[ fig01](b ) .
ohgushi et al . have discussed the case of a = fe and co in the frame of ligand field theory including spin - orbit coupling @xcite .
the main results of their calculations are reproduced in the first two panel of fig .
[ fig3d3d01 ] .
the transition in fecr@xmath0o@xmath1 at around @xmath400.49ev , was assigned as a single electron transition through the crystal field gap , @xmath59e(e@xmath60t@xmath61 ) @xmath25 @xmath59t@xmath0(e@xmath62t@xmath63 ) , while the two absorbtion peaks in cocr@xmath0o@xmath1 at around @xmath400.84ev and @xmath401.95ev correspond to the @xmath64a@xmath0(e@xmath64t@xmath61 ) @xmath25 @xmath64t@xmath0(e@xmath60t@xmath63 ) and @xmath64a@xmath0(e@xmath64t@xmath61 ) @xmath25 @xmath59t@xmath10(e@xmath60t@xmath63 ) transitions , respectively . from these excitations
the crystal field splitting corresponding to the cubic state and the racah parameter , b , characteristic to the coulomb interaction can be determined for co@xmath2 : @xmath65e=0.84ev and b=0.083ev , respectively . in the low - temperature optical conductivity spectrum of nicr@xmath0o@xmath1 two distinct structures
are visible around @xmath400.95ev and @xmath401.94ev , which can be assigned as _
3d_-_3d _ transitions of the tetrahedrally coordinated ni@xmath2 .
since the fine structure of these peaks can not be resolved , cubic ( t@xmath66 ) site symmetry was assumed in our model and the tetragonal distortion was neglected @xcite .
the energy levels of the eight 3d electron of the ni@xmath2 , relevant for the single particle excitations , is depicted in fig .
[ fig3d3d01](c ) .
the two peaks in the optical conductivity spectrum were fitted with two lorentzian corresponding to the two electric - dipole - allowed transitions : @xmath60t@xmath10(e@xmath64t@xmath63 ) @xmath25 @xmath60t@xmath0(e@xmath60t@xmath67 ) and @xmath60t@xmath10(e@xmath64t@xmath63 ) @xmath25 @xmath60t@xmath10(e@xmath60t@xmath67 ) ( see fig . [ fig12](a ) ) .
the cubic crystal field splitting and the racah parameter for the ni@xmath2 ion were deduced from the transition energies : @xmath65e=1.20ev and b=0.082ev , respectively .
the spin - orbit and the unresolved tetragonal splitting may be responsible for the large width of the resonances peaks .
o@xmath1 and cucr@xmath0o@xmath1 as shown in panel ( a ) and ( b ) , respectively .
labels correspond to notation introduced in fig .
[ fig3d3d01 ] . besides the oscillators
, there is a common baseline corresponding to the tail of the band edge . ]
in contrast to nicr@xmath0o@xmath1 , the crystal lattice of the cucr@xmath0o@xmath1 has strong tetragonal distortion at low temperature @xcite . perhaps , the low - symmetry of the lattice is responsible for the large number of peaks observed in the near infrared and visible part of the optical conductivity spectrum . as in other acr@xmath0o@xmath1 compounds the excitations around e=1.6ev
are assigned to the intra - atomic _ 3d_-_3d _ transitions of cr@xmath4 split by the tetragonal distortion .
compared to the sister compounds the strength of this transition is remarkably enhanced .
the two lower - energy peaks at around @xmath400.56ev and @xmath401.06ev are related to the _
3d_-_3d _ transitions of cu@xmath2 . due to the strong jahn - teller distortion in the ligand field calculations
tetragonal d@xmath68 local symmetry was assumed and the spin - orbit coupling was also taken into account .
the energy levels and the possible optical transitions are shown in fig .
[ fig3d3d01](d ) . in the d@xmath68 symmetry
the @xmath62b@xmath0 @xmath25 @xmath62b@xmath10 transition is forbidden , and only the spin - orbit interaction makes it weakly allowed , therefore , its oscillator strength was assumed to be zero . the optical conductivity spectrum was fitted with three lorentzian oscillator as shown in fig . [
fig12](b ) .
the obtained values for the cubic and the tetragonal crystal field splitting , and the spin - orbit coupling are @xmath65e=0.69ev , @xmath65e@xmath69=0.56ev and @xmath70=0.1ev , respectively , if the tetragonal splitting within the cubic @xmath62e term , @xmath65e@xmath71 is neglected .
the different parameters of the electronic structure of the a - site ions , namely the cubic and tetragonal crystal field splittings , @xmath65e and @xmath65e@xmath72 , respectively , the racah parameter , b and the strength of the spin - orbit coupling , @xmath70 are summarized in table [ table : cfparameters ] for the investigated acr@xmath0o@xmath1 spinel oxides .
these fundamental energy scales may serve as a starting point for a microscopic theory describing the magneto - elasticity of the acr@xmath0o@xmath1 spinels and provide in general an important input to theories describing their magnetic states . .__cubic
( @xmath65e ) and tetragonal ( @xmath65e@xmath72 ) crystal field splittings , the racah parameter ( b ) and the strength of the spin - orbit coupling ( @xmath70 ) as determined in the present study for various spinel oxides with magnetic a - site ions . _ _ [ cols="^,^,^,^,^",options="header " , ]
from the viewpoint of the dynamical properties of the lattice , acr@xmath0o@xmath1 chromium spinels ( a = mn , fe , co , ni and cu ) can be classified into two distinct groups .
the main results of our optical study about these classes are summarized in fig .
[ fig09 ] and fig .
[ fig08 ] . in the first group
the a - site ions have no orbital degeneracy ( a = mn and co ) and neither of these compounds show observable phonon mode splitting upon the magnetic phase transitions from room temperature down to t@xmath7310k .
anomalous broadening of the third and fourth mode of cocr@xmath0o@xmath1 below t@xmath12=93k is observed , which may indicate a mode splitting slightly below the spectral resolution of our experiment ( @xmath400.5@xmath17 ) . in mncr@xmath0o@xmath1
we could not even follow any broadening of the modes associated with the magnetic transitions .
therefore , we conclude that the magnetoelasticity is weak in these compounds and the cubic symmetry of the lattice is nearly preserved even in their magnetically ordered ground state . for materials in the second class where a - site ions are jahn - teller active ,
i.e. they have orbital degeneracy in the cubic phase we observed phonon - mode splitting associated with the cooperative jahn - teller distortion .
the sequence of the splittings shows that a cubic to tetragonal distortion takes place at t@xmath43 .
phonon modes in cucr@xmath0o@xmath1 exhibit a relative splitting of @xmath5730% , @xmath578% , @xmath579% and @xmath577% in ascending order of phonon energies already at room temperture @xcite . for comparison ,
splitting of the third and fourth phonon modes in nicr@xmath0o@xmath1 is @xmath574% and @xmath571% , while the lowest- and highest - energy phonons in fecr@xmath0o@xmath1 are split by @xmath7420% and @xmath571% , respectively .
most of the modes whose degeneracy is partly lifted by the cooperative jahn - teller distortion or those activated upon the splitting of silent @xmath38 lattice vibrations , as is the case in nicr@xmath0o@xmath1 , are further split upon the ferrimagnetic ordering .
the magnitude of the magnetically induced splitting is in the range of @xmath5710% for low - energy and @xmath571% for high - energy phonon modes , the large splitting of low - energy modes implies strong spin - lattice coupling for these lattice vibrations resulting in the strong magnetoelastic effects .
recently , the static orthorhombic distortion of the crystalline lattice upon the magnetic phase transition has also been observed by means of high - resolution x - ray diffraction in nicr@xmath0o@xmath1 and cucr@xmath0o@xmath1 @xcite . in conclusions
, we found that in acr@xmath0o@xmath1 spinel compounds where the frustration of the pyrochlore sublattice is lifted by long - range magnetic ordering due to the exchange interaction between the spins of the a@xmath2 and cr@xmath4 ions the criterion of strong magnetoelasticity is the orbital degeneracy of the a - site ion in the high - symmetry cubic phase .
this observation is valid irrespective to the details of the magnetic order , i.e. we found strong magnetoelasticity for the compounds fecr@xmath0o@xmath1 , nicr@xmath0o@xmath1 and cucr@xmath0o@xmath1 independently whether they have conical or canted spin order in their ground state . a theoretical model reproducing the effect of orbital occupancy on the magnetoelasticity of these compounds will be published elsewhere @xcite .
moreover , based on the analysis of the intraatomic _
3d_-_3d _ transition of the a - site ions , we could determine the following fundamental energies : crystal - field splitting in the cubic and tetragonal phases , the racah parameters representing the strength of the intraatomic coulomb repulsion and the spin - orbit interaction in some cases .
this work was supported by hungarian research funds otka pd75615 , bolyai 00256/08/11 , tmop-4.2.2.b-10/1 - 2010 - 0009 and it was partly supported by grant - in - aid for scientific research ( s ) no .
24224009 , and first program by the japan society for the promotion of science ( jsps ) .
financial support by the deutsche forschungsgemeinschaft through sfb 484 is acknowledged . | dynamical properties of the lattice structure was studied by optical spectroscopy in acr@xmath0o@xmath1 chromium spinel oxide magnetic semiconductors over a broad temperature region of t=10 - 335k . the systematic change of the a - site ions ( a = mn , fe , co , ni and cu ) showed that the occupancy of 3d orbitals on the a - site , has strong impact on the lattice dynamics . for compounds with orbital degeneracy ( fecr@xmath0o@xmath1 , nicr@xmath0o@xmath1 and
cucr@xmath0o@xmath1 ) , clear splitting of infrared - active phonon modes and/or activation of silent vibrational modes have been observed upon the jahn - teller transition and at the onset of the subsequent long - range magnetic order .
although mncr@xmath0o@xmath1 and cocr@xmath0o@xmath1 show multiferroic and magnetoelectric character , no considerable magnetoelasticity was found in spinel compounds without orbital degeneracy as they closely preserve the high - temperature cubic spinel structure even in their magnetic ground state . besides lattice vibrations ,
intra - atomic 3d-3d transitions of the a@xmath2 ions were also investigated to determine the crystal field and racah parameters and the strength of the spin - orbit coupling . |
[ s1.3].motivation[s1.3.1].some points of history[s1.3.2].some early applications[s2].signal processing[s2.1].filters in communications engineering[s2.2].algorithms for signals and for wavelets[s2.2.1].pyramid algorithms[s2.2.2].subdivision algorithms[s2.2.3].wavelet packet algorithms[s2.2.4].lifting algorithms : sweldens and more[s2.2bis].factorization theorems for matrix functions [ snew5aug2.3.4].matrix completion[s2.2bis.4].connections between matrix functions and signal processingappendix a : topics for further research[s3].connection between the discrete signals and the wavelets [ s3.1].wavelet geometry in @xmath0 [ s3.4.1].cycles [ s3.4.2].the ruelle - lawton wavelet transfer operator [ s4].other topics in wavelets theory [ s4.1].invariants [ s4.1.1].invariants for wavelets : global theory [ s4.1.2].invariants for wavelet filters : local theory [ s4.2].function classes [ s4.2.1].function classes for wavelets [ s4.2.2].function classes for filters [ s4.3].wavelet sets [ s4.4].spectral pairs appendix b : duality principles in analysisacknowledgementsreferences while this series of four lectures will be on the subject of wavelets , the emphasis will be on some interconnections between topics in the mathematics of wavelets and other areas , both within mathematics and outside .
connections to operator theory , to quantum theory , and especially to signal processing will be studied .
concepts such as high - pass and low - pass filters have become synonymous with wavelet tools , but they have also had a significance from the very start of signal processing , for example early telephone signals over transatlantic cables .
this was long before the much more recent advances in wavelets which started in the mid-1980 s ( as a resumption , in fact , of ideas going back to alfred haar @xcite much earlier ) . since the mid-1980 s wavelet mathematics has served to some extent as a clearing house for ideas from diverse areas from mathematics , from engineering , as well as from other areas of science , such as quantum theory and optics .
this makes the interdisciplinary communication difficult , as the lingo differs from field to field ; even to the degree that the same term might have a different name to some wavelet practitioners from what is has to others . in recognition of this fact , chapter 1 in the recent wavelet book @xcite samples a little dictionary of relevant terms .
parts of it are reproduced here : * * multiresolution : * _ real world : _ a set of band - pass - filtered component images , assembled into a mosaic of resolution bands , each resolution tied to a finer one and a coarser one . + _ mathematics : _ used in wavelet analysis and fractal analysis , multiresolutions are systems of closed subspaces in a hilbert space , such as @xmath2 , with the subspaces nested , each subspace representing a resolution , and the relative complement subspaces representing the detail which is added in getting to the next finer resolution subspace . *
* matrix function : * a function from the circle , or the one - torus , taking values in a group of @xmath3-by-@xmath3 complex matrices . *
* wavelet : * a function @xmath4 , or a finite system of functions @xmath5 , such that for some scale number @xmath3 and a lattice of translation points on @xmath6 , say @xmath7 , a basis for @xmath2 can be built consisting of the functions @xmath8 , @xmath9 .
+ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ then dulcet music swelled + concordant with the life - strings of the soul ; + it throbbed in sweet and languid beatings there , + catching new life from transitory death ; + like the vague sighings of a wind at even + that wakes the wavelets of the slumbering sea__@xmath10 + shelley , _ queen mab _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ * * subband filter : * _ engineering : _ signals are viewed as functions of time and frequency , the frequency function resulting from a transform of the time function ; the frequency variable is broken up into bands , and up - sampling and down - sampling are combined with a filtering of the frequencies in making the connection from one band to the next .
+ _ wavelets : _ scaling is used in passing from one resolution @xmath11 to the next ; if a scale @xmath3 is used from @xmath11 to the next finer resolution , then scaling by @xmath12 takes @xmath11 to a coarser resolution @xmath13 represented by a subspace of @xmath11 , but there is a set of functions which serve as multipliers when relating @xmath11 to @xmath13 , and they are called subband filters . * * cascades : * _ real world : _ a system of successive refinements which pass from a scale to a finer one , and so on ; used for example in graphics algorithms : starting with control points , a refinement and masking coefficients are used in a cascade algorithm yielding a cascade of masking points and a cascade approximation to a picture .
+ _ wavelets : _ in one dimension the scaling is by a number and a fixed simple function , for example of the form @xmath14{$\scriptstyle 0 $ } } \put(1,-0.0625){\makebox(0,0)[t]{$\scriptstyle 1 $ } } \end{picture}$ ] is chosen as the initial step for the cascades ; when the masking coefficients are chosen the cascade approximation leads to a scaling function .
* * scaling function : * a function , or a distribution , @xmath15 , defined on the real line @xmath6 which has the property that , for some integer @xmath16 , the coarser version @xmath17 is in the closure ( relative to some metric ) of the linear span of the set of translated functions @xmath18 , @xmath19 , @xmath20 , @xmath21 . * * logic gates : * _ in computation _ the classical logic gates are realized as computers , for example as electronic switching circuits with two - level voltages , say high and low .
several gates have two input voltages and one output , each one allowing switching between high and low : the output of the and gate is high if and only if both inputs are high .
the xor gate has high output if and only if one of the inputs , but not more than one , is high . *
* qubits : * _ in physics and in computation : _ qubits are the quantum analogue of the classical bits @xmath22 and @xmath23 which are the letters of classical computers , the qubits are formed of two - level quantum systems , electrons in a magnetic field or polarized photons , and they are represented in dirac s formalism @xmath24 and @xmath25 ; quantum theory allows superpositions , so states @xmath26 , @xmath27 , @xmath28 , are also admitted , and computation in the quantum realm allows a continuum of states , as opposed to just the two classical bits . + _ mathematics : _ a chosen and distinguished basis for the two - dimensional hilbert space @xmath29 consisting of orthogonal unit vectors , denoted @xmath24 , @xmath25 . *
* universality : * _ classical computing : _ the property of a set of logic gates that they suffice for the implementation of every program ; or of a single gate that , taken together with the not gate , it suffices for the implementation of every program . + _ quantum computing : _ the property of a set @xmath30 of basic quantum gates that every ( invertible ) gate can be written as a sequence of steps using only gates from @xmath30 .
usually @xmath30 may be chosen to consist of one - qubit gates and a distinguished tensor gate @xmath31 .
an example of a choice for @xmath31 is cnot . an alternative universal one is the toffoli gate .
+ _ mathematics : _ the property of a set @xmath30 of basic unitary matrices that for every @xmath32 and every @xmath33 , there is a factorization @xmath34 , @xmath35 , with the understanding that the factors @xmath36 are inserted in a chosen tensor configuration of the quantum register @xmath37 .
note that the factors @xmath36 , the number @xmath38 , and the configuration of the @xmath36 s all depend on @xmath32 and the gate @xmath33 to be studied . the quantum wavelet algorithm ( [ eqwaveletmatrix ] ) is an example of such a matrix @xmath39 . *
* chaos : * a small variation or disturbance in the initial states or input of some system giving rise to a disproportionate , or exponentially growing , deviation in the resulting output trajectory , or output data .
the term is used more generally , denoting rather drastic forms of instability ; and it is measured by the use of statistical devices , or averaging methods .
* @xmath40 * * : * * the _ general linear group _ of all complex @xmath41 invertible matrices . * @xmath42 * * : * * @xmath43 where @xmath44 denotes the adjoint matrix , i.e. , @xmath45 .
* * transfer operator ( transition operator ) : * _ in probability : _ an operator which transforms signals @xmath46 from input @xmath47 to output @xmath48 .
the signals are represented as functions on some set @xmath49 . in the simplest case , the operator is linear and given in terms of conditional probabilities @xmath50 .
the number @xmath50 may represent the probability of a transition from @xmath51 to @xmath52 where @xmath52 and @xmath51 are points in the set @xmath49 .
then @xmath53 + _ in computation : _ let @xmath54 and @xmath55 be functions on a set @xmath49 , both taking values in @xmath56 .
let @xmath55 be the initial state of the bit , and @xmath54 the final state of the bit .
if the process is governed by a probability distribution @xmath57 , then the transition probabilities @xmath58 are conditional probabilities : i.e. , @xmath50 is the probability of a final bit value @xmath52 given an initial value @xmath51 , and we have @xmath59 + _ in wavelet theory : _ let @xmath60 , and let @xmath61 be a positive function on @xmath62 , for example @xmath63 where @xmath64 is some low - pass wavelet filter with @xmath3 bands .
( positivity is only in the sense @xmath65 , nonnegative , and the function @xmath61 may vanish on a subset of @xmath66 . )
then define a function @xmath67 on @xmath68 as follows : @xmath69 we arrive at the transfer operator @xmath70 , i.e. , the operator transforming functions on @xmath66 as follows : @xmath71 * * coherence : * _ in mathematics and physics : _ the vectors @xmath72 that make up a tight frame , one which is not an orthonormal basis , are said to be subjected to _
coherence_. so coherent vector systems in hilbert space are viewed as bases which generalize the more standard concept of orthonormal bases from harmonic analysis .
a striking feature of the wavelets with compact support , which are based on scaling , is that the varieties of the two kinds of bases can be well understood geometrically .
for example , the collapse of the wavelet orthogonality relations , degenerating into coherent vectors , happens on a subvariety of a lower dimension .
+ more generally , coherent vectors in mathematical physics often arise with a continuous index , even if the hilbert space is separable , i.e. , has a countable orthonormal basis .
this is illustrated by a vector system @xmath73 , which should be thought of as a continuous analogue , i.e. , a version where a sum gets replaced with an integral @xmath74 for more details , see also section 3.3 of @xcite and chapter 3 of @xcite .
+ in quantum mechanics , one talks , for example , about coherent states in connection with wavefunctions of the harmonic oscillator .
combinations of stationary wavefunctions from different energy eigenvalues vary periodically in time , and the question is which of the continuously varying wavefunctions one may use to expand an unknown function in without encountering overcompleteness of the basis
. the methods of `` coherent states '' are methods for using these kinds of functions ( which fit some problems elegantly ) while avoiding the difficulties of overcompleteness .
the term `` coherent '' applies when you succeed in avoiding those difficulties by some means or other .
of course , for students who have just learned about the classic complete orthonormal basis of stationary eigenfunctions , `` coherent state '' methods at first may seem like a daring relaxation of the rules of orthogonality , so that the term seems to stand for total freedom ! wavelet theory is the art of finding a special kind of basis in hilbert space .
let @xmath75 be a hilbert space over @xmath76 and denote the inner product @xmath77 .
for us , it is assumed linear in the second variable . if @xmath78 , then@xmath79 if @xmath80 , then@xmath81 let @xmath82 . if @xmath83 , then@xmath84 functions @xmath85 have fourier series : setting @xmath86 , @xmath87 and@xmath88 similarly if @xmath89 , then@xmath90 and@xmath91 let @xmath92 be an index set .
we shall only need to consider the case when @xmath92 is countable .
let @xmath93 be a family of nonzero vectors in a hilbert space @xmath75 .
we say it is an _
orthonormal basis _ ( onb ) if@xmath94 and if @xmath95 if only ( [ eq1.2.9 ] ) is assumed , but not ( [ eq1.2.8 ] ) , we say that @xmath93 is a ( normalized ) _
tight frame_. we say that it is a _ frame _ with _ frame constants _
@xmath96 if@xmath97 introducing the rank - one operators @xmath98 of dirac s terminology , see @xcite , we see that @xmath99 is an onb if and only if the @xmath100 s are projections and@xmath101 it is a ( normalized ) tight frame if and only if ( [ eq1.2.10 ] ) holds but with no further restriction on the rank - one operators @xmath100 .
it is a frame with frame constants @xmath102 and @xmath103 if the operator @xmath104 satisfies@xmath105 in the order of hermitian operators .
( we say that operators @xmath106 , @xmath107 , satisfy @xmath108 if @xmath109 holds for all @xmath110 ) .
wavelets in @xmath111 are generated by simple operations on one or more functions @xmath4 in @xmath112 , the operations come in pairs , say scaling and translation , or phase - modulation and translations .
if @xmath113 we set@xmath114 we stress the discrete wavelet transform .
but the first line in the two tables below is the continuous one .
it is the only treatment we give to the continuous wavelet transform , and the corresponding _ coherent vector decompositions_. but , as is stressed in @xcite , @xcite , and @xcite , the continuous version came first .
\multicolumn{3}{c}{\parbox{0.9\textwidth } { summary of and variations on the resolution of the identity operator $ 1 $ in $ l^{2}$ or in $ \ell^{2}$ , for $ \psi$ and $ \tilde{\psi}$ where $ \psi_{r , s}\left ( x\right ) = r^{-\frac{1}{2}}\psi\left ( \frac{x - s}{r}\right ) $ , $ c_{\psi}=\int_{\mathbb{r}}\frac{d\omega}{\left| \omega\right| } \left| \smash{\hat{\psi}}\left ( \omega\right ) \right| ^{2}<\infty$ , similarly for $ \tilde{\psi}$ and $ c_{\psi,\tilde{\psi}}=\int_{\mathbb{r}}\frac{d\omega}{\left| \omega\right| } \overline{\hat{\psi}\left ( \omega\right ) } \ , \hat{\tilde{\psi}}\left ( \omega\right ) $ : \rule[-12pt]{0pt}{12pt}}}\\ $ ] n=2@xmath116c_^-1_^2^ |
_ r , s_r , s| @xmath117 = 1_l^2@xmath118 c_,^-1_^2^ | _ r , s | @xmath117 = 1_l^2@xmath119 _ j_k | _ j , k | = 1_l^2@xmath120n2@xmath121 ^ 2@xmath122 ^ 2@xmath123 \multicolumn{2}{c}{\parbox{0.9\textwidth } { consult chapter 3 of \cite{kai94 } for the continuous resolution\index{resolution } , and section 2.2 of \cite{brjo02b } for the discrete resolution\index{resolution}. if $ h , k$ are vectors in a hilbert\index{hilbert!space@--- space } space $ \mathcal{h}$ , then the operator $ a=\left| h\right\rangle \left\langle k\right| $ is defined by the identity $ \left\langle \,u\mid av\,\right\rangle = \left\langle \,u\mid h\,\right\rangle \left\langle \,k\mid v\,\right\rangle $ for all $ u , v\in\mathcal{h}$. then the assertions in the first table amount to:\rule[-12pt]{0pt}{12pt}}}\\ $ ] c_^-1_^2 | _ r ,
sf| ^2 @xmath117 = f_l^2 ^ 2fl^2 ( ) @xmath125c_,^-1_^2 f_r , s g @xmath117 = fgf , gl^2 ( ) @xmath126 _ j_k _ | _ j , kf| ^2 @xmath117 = f_l^2 ^ 2fl^2 ( ) @xmath125 _ j_k _ f_j , k g @xmath117 = fgf , gl^2 ( ) @xmath126_i=0_^n-1^ s_i^c^2=c^2 c^2@xmath118_i=0_^n-1^ s_i^cd = cdc , d^2@xmath127 a function @xmath4 satisfying the resolution identity is called a _ coherent vector _ in mathematical physics . the representation theory for the @xmath128-group ,
i.e. , the matrix group @xmath129 , serves as its underpinning .
then the tables above illustrate how the @xmath130 wavelet system arises from a discretization of the following unitary representation of @xmath131 : @xmath132 acting on @xmath2 .
this unitary representation also explains the discretization step in passing from the first line to the second in the tables above .
the functions @xmath133 which make up a wavelet system result from the choice of a suitable coherent vector @xmath134 , and then setting @xmath135 even though this representation lies at the historical origin of the subject of wavelets ( see @xcite ) , the @xmath128-group seems to be now largely forgotten in the next generation of the wavelet community . but
chapters 13 of @xcite still serve as a beautiful presentation of this ( now much ignored ) side of the subject .
it also serves as a link to mathematical physics and to classical analysis .
since the representation @xmath136 in ( [ equnitaryrep ] ) on @xmath2 , when a unitary @xmath136 is defined from ( [ equnitaryrep ] ) setting @xmath137 , @xmath138 , @xmath139 , leaves invariant the hardy space @xmath140 formula ( [ eqcoherentvec ] ) suggests that it would be simpler to look for wavelets in @xmath141 .
after all , it is a smaller space , and it is natural to try to use the causality features of @xmath141 implied by the support condition in ( [ eqhardyspace ] ) .
moreover , in the world of the fourier transform , the two operations of the formulas ( [ equnitaryrep ] ) and ( [ eqcoherentvec ] ) take the simpler forms @xmath142 so in the early nineties , this was an open problem in the theory , i.e. , whether or not there are wavelets in the hardy space ; but it received a beautiful answer in @xcite .
auscher showed that there are no wavelet functions @xmath4 in @xmath141 which satisfy the following mild regularity properties : @xmath143 comparison of formulas ( [ equnitaryrep ] ) and ( [ eqcoherentvec ] ) shows that the traditional discrete wavelet transform may be viewed as the restriction to a subgroup @xmath144 of a classical unitary representation of @xmath131 .
the unitary representations of @xmath131 are completely understood : the set of irreducible unitary representations consists of two infinite - dimensional inequivalent subrepresentations of the representation ( [ equnitaryrep ] ) on @xmath2 , together with the one - dimensional representations @xmath145 parameterized by @xmath146 .
( the two subrepresentations of ( [ equnitaryrep ] ) are obtained by restricting to @xmath147 with @xmath148 and @xmath149 , respectively .
) however , the subgroup @xmath144 of @xmath131 has a rich variety of inequivalent infinite - dimensional representations that do not arise as restrictions of ( [ equnitaryrep ] ) , or of any representation of @xmath131 .
the group @xmath144 considered in ( [ eqcoherentvec ] ) is a semidirect product ( as is @xmath131 ) : it is of the form @xmath150{\left .
\vphantom { \begin{pmatrix } a & b \\ 0 & 1 \end{pmatrix } \biggm| a = n^{j},\ ; b=\sum_{i\in\mathbb{z}}n_{i}n^{i},\ ; j\in\mathbb{z},\ ; n_{i}\in\mathbb{z } , } \text { where the } \sum_{i}\text { summation is finite}\,\right\ } . }
\label{eqsemidirectprod}\end{gathered}\ ] ] ( in the jargon of pure algebra , the nonabelian group @xmath151 is the semidirect product of the two abelian groups @xmath7 and @xmath152 , with a naturally defined action of @xmath7 on @xmath152 . ) the papers @xcite , @xcite , @xcite , @xcite , @xcite , and @xcite show that it is possible to use these nonclassical representations of @xmath144 for the construction of unexpected classes of wavelets , the wavelet sets being the most notable ones .
recall that a subset @xmath153 of finite measure is a _ wavelet set _ if @xmath154 is such that , for some @xmath60 , @xmath155 , the functions @xmath156 form an orthonormal basis for @xmath2 . until the work of larson and others , see @xcite and @xcite , it was not even clear that wavelet sets @xmath49 could exist in the case @xmath157 .
the paper @xcite develops and extends the representation theory for the subgroups @xmath151 independently of the ambient group @xmath131 and shows that each @xmath151 has continuous series of representations which account for the wavelet sets .
the role of the representations of the groups @xmath151 and their generalizations for the study of wavelets was first stressed in @xcite .
there is a different transform which is analogous to the wavelet transform of ( [ equnitaryrep])([eqcoherentvec ] ) , but yet different in a number of respects .
it is the gabor transform , and it has a history of its own .
both are special cases of the following construction : let @xmath131 be a nonabelian matrix group with center @xmath158 , and let @xmath136 be a unitary irreducible representation of @xmath131 on the hilbert space @xmath2 .
when @xmath134 is given , we may define a transform @xmath159 it turns out that there are classes of matrix groups , such as the @xmath160 group , or the @xmath161-dimensional group of upper triangular matrices , which have transforms @xmath162 admitting effective discretizations .
this means that it is possible to find a vector @xmath134 , and a discrete subgroup @xmath163 , such that the restriction to @xmath164 of the transform @xmath162 in ( [ eqgabtrans ] ) is injective from @xmath2 into functions on @xmath164 .
there are many books on transform theory , and here we are only making the connection to wavelet theory .
the book @xcite contains much more detail on the group - theoretic approach to these continuous and discrete coherent vector transforms .
one of our coordinates for the landscape of multiresolution wavelets takes the form of a geometric index .
in fact , it involves a traditional operator - theoretic index with values in @xmath7 . when it is identified with a winding number or a counting of homotopy classes , it serves also as a fredholm index of an associated toeplitz operator .
an orthogonal dyadic wavelet basis has its wavelet function @xmath4 satisfying the normalization @xmath165 , i.e. , @xmath4 is a vector of norm one in the hilbert space @xmath2 . in the lingo of quantum theory
, @xmath4 is therefore a pure state , and the @xmath52-coordinate is an observable called the position .
the integral @xmath166 is the expected value of the position . if @xmath167 denotes the standard haar function in ( [ eqhaarwavelet ] ) , then clearly @xmath168 .
also note the translation formula @xmath169 .
we showed in corollary 2.4.11 of @xcite , completely generally , that the other orthonormal wavelets @xmath4 have expected values in the set @xmath170 .
hence , after @xmath4 is translated by an integer , you can not distinguish it from the haar wavelet @xmath167 in ( [ eqhaarwavelet ] ) by looking only at the expected value of its position coordinate .
the translation integer @xmath38 turns out to be a winding number .
our result holds more generally when the definition of @xmath171 is adapted to a wider wavelet context , as we showed in chapter 6 of @xcite ; but in all cases , there is a winding number which produces the above - mentioned integer translate @xmath38 .
the issue of connectedness for various classes of wavelets is a general question which has been addressed previously in the wavelet literature ; see , e.g. , @xcite , @xcite , @xcite , and @xcite . here
we bring homotopy to bear on the question , and we identify the connected components when the compact support is fixed and given .
we show among other things that for a fixed @xmath172-class a homotopy may take place within a variety of wavelets which is specified by a slightly bigger support than the initially given one .
an important point of our present discussion , beyond the mere fact of compact support , is the size of the support of the wavelets in question .
consider two wavelets @xmath102 and @xmath103 of a certain support size .
then our first results in this section also specify the paths @xmath173 , if any , which connect @xmath102 and @xmath103 , and in particular the size of the support of the wavelets corresponding to @xmath173 . in @xcite , we treat connectivity in the wider context of noncompactly supported wavelets , following at the outset @xcite , which considers scale number @xmath174 , and wavelets @xmath4 satisfying@xmath175 garrigs considers , for @xmath176 , the class @xmath177 of wavelets @xmath4 such that@xmath178 and there is an @xmath179 such that@xmath180 i.e. , the wavelet is supposed to have some degree of smoothness in the sense of sobolev .
we now turn to the group of functions @xmath181 , where @xmath182 denotes the group of all complex @xmath3-by-@xmath3 matrices .
the functions will not be assumed continuous in general .
the continuous functions will be designated @xmath183 .
each function in @xmath183 has a @xmath184-class , also called a winding number ; see @xcite .
the functions in @xmath183 with finite fourier expansion will be called _ fourier polynomials _ , also if they are functions which take values in @xmath182 .
[ prohom.1]let @xmath185 be a fourier polynomial , and assume that @xmath186 .
then @xmath136 is homotopic in @xmath183 to@xmath187 where @xmath67 is the one - dimensional projection onto the first coordinate slot in @xmath188 , and if @xmath136 has the form@xmath189 then @xmath136 may be homotopically deformed to @xmath11 in @xmath183 through fourier polynomials of degree at most @xmath190 .
this proposition remains true if the word `` fourier polynomial '' is replaced by `` polynomial '' and @xmath191 for @xmath192 . in that case @xmath193 and @xmath136 may be homotopically deformed to @xmath11 in the loop semigroup of polynomial unitaries in @xmath183 through polynomials of degree at most @xmath194 .
multiplying @xmath136 by @xmath195 , we obtain a polynomial @xmath196 of degree @xmath197 mapping @xmath66 into @xmath182 . then @xmath198 . by proposition 3.3 of @xcite
, there exist @xmath199 one - dimensional projections @xmath200 in @xmath201 and a unitary @xmath202 such that@xmath203 ( see [ exe6dauswe-11 ] for a related , but different , decomposition . ) now , deforming each of the @xmath204 s continuously through one - dimensional projections to the projection @xmath205 onto the first coordinate direction , and deforming @xmath206 in @xmath182 into @xmath207 , we see that @xmath208 can be deformed into @xmath209 thus @xmath210 itself is deformed into@xmath211 but writing @xmath212 as a sum of @xmath213 one - dimensional projections @xmath214 , we have that the unitary that @xmath210 is deformed into is@xmath215 and next deforming each of the @xmath216 in this decomposition into @xmath205 , we see that @xmath210 is deformed into@xmath217 the crude estimate @xmath190 on the degree of the fourier polynomials occurring during the deformation is straightforward .
[ remhom.2]we do not know if proposition is true if @xmath183 is replaced by @xmath220 .
it is known from lemma 11.2.12 of @xcite that if @xmath221 is a polynomial of degree @xmath23 in @xmath222 , then @xmath102 can be homotopically deformed through first - order polynomials in @xmath223 to a unitary of the form @xmath224 for some projection @xmath67 , and hence proposition for @xmath220 would follow if any polynomial @xmath225 could be factored into first - order polynomials .
it is also clear , since any element @xmath221 can be homotopically deformed into @xmath226 in @xmath220 , that if @xmath102 is a fourier polynomial , then @xmath102 can be homotopically deformed into @xmath226 through fourier polynomials .
this follows by compactness and the stone
weierstra theorem . for our purposes in wavelet theory , though , we would need a computable upper bound for the degree of the fourier polynomials . 1 . [ homcorr(1)]matrix functions , @xmath227 , satisfying the normalization@xmath228 2 .
[ homcorr(2)]high- and low - pass wavelet filters @xmath229 , @xmath230 , satisfying @xmath231 and@xmath232 3 .
[ homcorr(3)]scaling functions @xmath15 together with wavelet generators @xmath72 .
we did not specify the continuity and regularity requirements of the functions @xmath102 , @xmath233 , @xmath15 , @xmath234 above .
this will be done differently in different contexts and the classes clearly depend on these added requirements .
we will now restrict to the case that the functions @xmath15 and @xmath72 have compact support in @xmath235 , i.e. , that @xmath102 and @xmath229 are polynomials in @xmath222 .
thus @xmath236 is a polynomial function with@xmath237 one defines @xmath239 by @xmath240 ( cf .
( [ eqint.1 ] ) ) and then @xmath64 by @xmath241 or one uses@xmath242 directly .
then the high - pass filters @xmath229 , @xmath243 , can be derived from ( [ eqhom.21 ] ) below .
if we are in the generic case ( [ eqmul.12 ] ) , we may also recover the fourier coefficients @xmath244 of @xmath233 by @xmath245 where @xmath246 .
in particular it follows in this generic case that if the scaling and wavelet functions have compact support and the filters are lipschitz , then the filters are fourier polynomials .
is this true also in the nongeneric tight frame case ? now , if @xmath247 , define : @xmath248[r]{$\displaystyle\operatorname{mf}\left ( d\right ) $ } & = \begin{minipage}[t]{0.895\displayboxwidth}\raggedright the set of polynomial functions in $ z\in\mathbb{t}$ in $ c\left ( \mathbb { t},\mathrm{u}_{n}\left(\mathbb{c}\right)\right ) $ of degree\index{degree!of a fourier polynomial@--- of a fourier polynomial } at most $ d$ satisfying ( \ref{eqhom.12});\end{minipage}\label{eqhom.28}\\[\smallskipamount ] \llap{$\bullet\mkern6mu$ } \makebox[\hangboxwidth][r]{$\displaystyle\operatorname{wf}\left ( d\right ) $ } & = \begin{minipage}[t]{0.895\displayboxwidth}\raggedright the set of $ n$-tuples of wavelet filter\index{filter!wavelet}s $ \left ( m_{0},\dots , m_{n-1 } \right ) $ such that all $ m_{i}$ are polynomials in $ z\in\mathbb { t}$ of degree\index{degree!of a fourier polynomial@--- of a fourier polynomial } at most $ d$ satisfying ( \ref{eqhom.13 } ) and ( \ref{eqhom.14});\end{minipage}\label{eqhom.29}\\[\smallskipamount ] \llap{$\bullet\mkern6mu$ } \makebox[\hangboxwidth][r]{$\displaystyle\operatorname{sf}\left ( d\right ) $ } & = \begin{minipage}[t]{0.895\displayboxwidth}\raggedright the set of $ n$-tuples $ \left ( \varphi,\psi_{1},\dots,\psi_{n-1}\right ) $ of scaling\index{scaling!function@--- function } functions/ wavelet\index{wavelet!function@--- function } functions with support in $ \left [ 0,d\right]$.\end{minipage } \label{eqhom.30}\ ] ] the spaces @xmath249 , @xmath250 , and @xmath251 may be equipped with the obvious topologies , coming in the first two cases from , for example , the @xmath252-norm over @xmath222 , and in the last case either from the @xmath253-norm or , as will be more relevant , the tempered - distribution topology . by virtue of proposition 3.2 in @xcite , @xmath249 has the structure of a compact algebraic variety , and so by ( [ eqhom.17 ] ) below , @xmath254 is a compact algebraic variety .
it is clear from ( [ eqhom.17 ] ) that the map @xmath255 maps @xmath256 into @xmath257 , and that @xmath258 maps @xmath259 into @xmath249 .
furthermore , it is clear from ( [ eqint.1b ] ) and ( [ eqhom.21 ] ) that @xmath260 maps @xmath261 into @xmath251 , and conversely @xmath262 maps @xmath251 into @xmath263 .
now , let a subindex @xmath22 denote the subsets of these various spaces such that the condition @xmath264 holds .
it is known that the set of points such that ( [ eqhom.23 ] ) does not hold is a lower - dimensional subvariety of the various varieties , see section 6 of @xcite , and hence @xmath265 , @xmath266 , and @xmath267 contain the generic points in @xmath249 , @xmath254 , and @xmath251 .
[ thmhom.3]let @xmath268 .
equip the space @xmath269 of scaling functions / wavelet functions with support in @xmath270 $ ] with the tempered - distribution topology .
then @xmath269 is homeomorphic to a compact algebraic variety .
furthermore , for two elements @xmath271 , the following conditions are equivalent : 1 .
[ thmhom.3(1)]the elements @xmath272 and @xmath273 can be connected to each other by a continuous path in @xmath274 ; 2 .
[ thmhom.3(2)]@xmath275 ; 3 .
[ thmhom.3(3)]the elements @xmath272 and @xmath273 can be connected to each other by a continuous path in some @xmath276 .
in addition to the general background material in the present section , the reader may find a more detailed treatment of some of the current research trends in wavelet analysis in the following papers : @xcite ( a book review ) , @xcite ( a survey ) , and the research papers @xcite , @xcite , @xcite , @xcite , @xcite , and @xcite . as a mathematical subject
, the theory of wavelets draws on tools from mathematics itself , such as harmonic analysis and numerical analysis .
but in addition there are exciting links to areas outside mathematics .
the connections to electrical and computer engineering , and to image compression and signal processing in particular , are especially fascinating .
these interconnections of research disciplines may be illustrated with the two subjects ( 1 ) wavelets and ( 2 ) subband filtering [ from signal processing ] .
while they are quite different , and have distinct and independent lives , and even have different aims , and different histories , they have in recent years found common ground .
it is a truly amazing success story .
advances in one area have helped the other : subband filters are absolutely essential in wavelet algorithms , and in numerical recipes used in subdivision schemes , for example , and especially in jpeg 2000an important and extraordinarily successful image - compression code .
jpeg uses nonlinear approximations and harmonic analysis in spaces of signals of bounded variation .
similarly , new wavelet approximation techniques have given rise to the kind of data - compression which is now used by the fbi [ via a patent held by two mathematicians ] in digitizing fingerprints in the u.s .
it is the happy marriage of the two disciplines , signal processing and wavelets , that enriches the union of the subjects , and the applications , to an extraordinary degree .
while the use of high - pass and low - pass filters has a long history in signal processing , dating back more than fifty years , it is only relatively recently , say the mid-1980 s , that the connections to wavelets have been made .
multiresolutions from optics are the bread and butter of wavelet algorithms , and they in turn thrive on methods from signal processing , in the quadrature mirror filter construction , for example . the effectiveness of multiresolutions in data compression is related to the fact that multiresolutions are modelled on the familiar positional number system : the digital , or dyadic , representation of numbers .
wavelets are created from scales of closed subspaces of the hilbert space @xmath111 with a scale of subspaces corresponding to the progression of bits in a number representation . while oversimplified here , this is the key to the use of wavelet algorithms in digital representation of signals and images .
the digits in the classical number representation in fact are quite analogous to the frequency subbands that are used _ both _ in signal processing and in wavelets . the two functions@xmath278{ccc}\varphi\left ( x\right ) = \left\ { \renewcommand{\arraystretch}{1.125}\begin{array } [ c]{ll}1\phantom{- } & 0\leq x<1\\ 0 &
\text{elsewhere}\end{array } \right . & \text{and } & \psi\left ( x\right ) = \left\ { \renewcommand{\arraystretch}{1.125}\begin{array } [ c]{ll}1 & 0\leq x<\frac{1}{2}\\ -1 & \frac{1}{2}\leq x<1\\ 0 & \text{elsewhere}\end{array } \right . \\ & & \\ \setlength{\unitlength}{36pt}\begin{picture}(2.25,3)(-0.5,-1.5)\put(-0.5,0){\vector(1,0){2.25}}\put(0,-1.5){\vector(0,1){3}}\thicklines\put(-0.125,0){\line(1,0){0.125}}\put(0,0){\line(0,1){1}}\put(0,1){\line(1,0){1}}\put(1,1){\line(0,-1){1}}\put(1,0){\line(1,0){0.125}}\put(0.5,1.0625){\makebox(0,0)[b]{$\varphi$}}\end{picture } & & \setlength{\unitlength}{36pt}\begin{picture}(2.25,3)(-0.5,-1.5)\put(-0.5,0){\vector(1,0){2.25}}\put(0,-1.5){\vector(0,1){3}}\thicklines\put(-0.125,0){\line(1,0){0.125}}\put(0,0){\line(0,1){1}}\put(0,1){\line(1,0){0.5}}\put(0.5,1){\line(0,-1){2}}\put(0.5,-1){\line(1,0){0.5}}\put(1,-1){\line(0,1){1}}\put(1,0){\line(1,0){0.125}}\put(0.5,1.0625){\makebox(0,0)[b]{$\psi$}}\end{picture}\\ \text{father function\quad } & & \text{mother function\quad}\\ \text{(a)\quad } & & \text{(b)\quad}\end{array } \label{eq0}\ ] ] capture in a glance the refinement identities@xmath279 the two functions are clearly orthogonal in the inner product of @xmath253 , and the two closed subspaces @xmath280 and @xmath281 generated by the respective integral translates@xmath282 satisfy@xmath283 where @xmath136 is the dyadic scaling operator @xmath284 .
the factor @xmath285 is put in to make @xmath136 a unitary operator in the hilbert space @xmath111 .
this version of haar s system naturally invites the question of what other pairs of functions @xmath15 and @xmath4 with corresponding orthogonal subspaces @xmath280 and @xmath281 there are such that the same invariance conditions ( [ eq2 ] ) hold .
the invariance conditions hold if there are coefficients @xmath286 and @xmath287 such that the scaling identity@xmath288 is solved by the father function , called @xmath15 , and the mother function @xmath4 is given by@xmath289 a fundamental question is the converse one : give simple conditions on two sequences @xmath290 and @xmath291 which guarantee the existence of @xmath111-solutions @xmath15 and @xmath4 which satisfy the orthogonality relations for the translates ( [ eq1 ] ) . how do we then get an orthogonal basis from this ? the identities for haar s functions @xmath15 and @xmath4 of ( [ eq0])(a ) and ( [ eq0])(b ) above make it clear that the answer lies in a similar tiling and matching game which is implicit in the more general identities ( [ eq3 ] ) and ( [ eq4 ] ) .
clearly we might ask the same question for other scaling numbers , for example @xmath292 or @xmath293 in place of @xmath294 .
actually a direct analogue of the visual interpretation from ( [ eq0 ] ) makes it clear that there are no nonzero locally integrable solutions to the simple variants of ( [ eq3]),@xmath295 there _ are _ nontrivial solutions to ( [ eq5 ] ) and ( [ eq6 ] ) , to be sure , but they are versions of the cantor devil s staircase functions , which are prototypes of functions which are not locally integrable .
( 478,644 ) ( 0,526 ) ( 120,526 ) ( 240,526 ) ( 360,526 ) ( 0,406 ) ( 120,406 ) ( 298,344 ) ( 269,458)(0,0)[b]@xmath296 ( 388,479)(0,0)[b]@xmath15 ( 392,327)(0,0)[b](a ) : father function ( 0,199 ) ( 120,199 ) ( 240,199 ) ( 360,199 ) ( 0,79 ) ( 120,79 ) ( 298,17 ) ( 269,131)(0,0)[b]@xmath296 ( 409,152)(0,0)[b]@xmath4 ( 392,0)(0,0)[b](b ) : mother function since the haar example is based on the fitting of copies of a fixed box inside an expanded one , it would almost seem unlikely that the system ( [ eq3])([eq4 ] ) admits finite sequences @xmath290 and @xmath291 such that the corresponding solutions @xmath15 and @xmath4 are continuous or differentiable functions of compact support .
the discovery in the mid-1980 s of compactly supported differentiable solutions , see @xcite , was paralleled by applications in seismology , acoustics @xcite , and optics @xcite , as discussed in @xcite , and once the solutions were found , other applications followed at a rapid pace : see , for example , the ten books in benedetto s review @xcite .
it is the solution @xmath4 in ( [ eq4 ] ) that the fuss is about , the mother function ; the other one , @xmath15 , the father function , is only there before the birth of the wavelet .
the most famous of them are named after daubechies , and look like the graphs in figure [ figdaubechies ] . with the multiresolution idea ,
we arrive at the closed subspaces@xmath297 as noted in ( [ eq1])([eq2 ] ) , where @xmath136 is some scaling operator .
there are extremely effective iterative algorithms for solving the scaling identity ( [ eq3 ] ) : see , for example , example 2.5.3 , pp
. 124125 , of @xcite@xmath298 , @xcite , and @xcite , and figure [ figdaubechies ] .
a key step in the algorithms involves a clever choice of the kind of resolution pictured in ( [ eq12 ] ) , but digitally encoded .
the orthogonality relations can be encoded in the numbers @xmath290 and @xmath291 of ( [ eq3])([eq4 ] ) , and we arrive at the doubly indexed functions @xmath299 it is then not difficult to establish the combined orthogonality relations@xmath300 plus the fact that the functions in ( [ eq8 ] ) form an orthogonal basis for @xmath111 .
this provides a painless representation of @xmath111-functions @xmath301 where the coefficients @xmath302 are@xmath303 what is more significant is that the resolution structure of closed subspaces of @xmath111@xmath304 facilitates powerful algorithms for the representation of the numbers @xmath302 in ( [ eq11 ] ) . amazingly , the two sets of numbers @xmath305 and @xmath291 which were used in ( [ eq3])([eq4 ] ) , and which produced the magic basis ( [ eq8 ] ) , the wavelets , are the same magic numbers which encode the quadrature mirror filters of signal processing of communications engineering . on the face of it , those signals from communication engineering really seem to be quite unrelated to the issues from wavelets the signals are just sequences , time is discrete , while wavelets concern @xmath111 and problems in mathematical analysis that are highly non - discrete .
dual filters , or more generally , subband filters , were invented in engineering well before the wavelet craze in mathematics of recent decades .
these dual filters in engineering have long been used in technology , even more generally than merely for the context of quadrature mirror filters ( qmf s ) , and it turns out that other popular dual wavelet bases for @xmath111 can be constructed from the more general filter systems ; but the best of the wavelet bases are the ones that yield the strongest form of orthogonality , which is ( [ eq9 ] ) , and they are the ones that come from the qmf s .
the qmf s in turn are the ones that yield perfect reconstruction of signals that are passed through filters of the analysis - synthesis algorithms of signal processing .
they are also the algorithms whose iteration corresponds to the resolution sytems ( [ eq12 ] ) from wavelet theory . while fourier invented his transform for the purpose of solving the heat equation , i.e. , the partial differential equation for heat conduction , the wavelet transform ( [ eq10])([eq11 ] ) does not diagonalize the differential operators in the same way .
its effectiveness is more at the level of computation ; it turns integral operators into sparse matrices , i.e. , matrices which have many zeros in the off - diagonal entry slots .
again , the resolution ( [ eq12 ] ) is key to how this matrix encoding is done in practice .
the first wavelet was discovered by alfred haar long ago , but its use was limited since it was based on step - functions , and the step - functions jump from one step to the next .
the implementation of haar s wavelet in the approximation problem for continuous functions was therefore rather bad , and for differentiable functions it is atrocious , and so haar s method was forgotten for many years .
and yet it had in it the one idea which proved so powerful in the recent rebirth ( since the 1980 s ) of wavelet analysis : the idea of a _
multiresolution_. you see it in its simplest form by noticing that a box function @xmath103 of ( [ eqhaarscaling ] ) may be scaled down by a half such that two copies @xmath306 and @xmath307 of the smaller box then fit precisely inside @xmath103 .
see ( [ eqhaarscaling ] ) .
@xmath308{c}\setlength{\unitlength}{30bp}\begin{picture}(3,1.559)(-0.5,-0.309 ) \put(-0.5,0){\vector(1,0){3 } } \thicklines \put(-0.125,0){\line(1,0){2.25 } } \put(0,1){\line(1,0){2 } } \multiput(0,0)(1,0){3}{\line(0,1){1 } } \put(1,-0.0425){\makebox(0,0)[t]{$b$ } } \put(0.3657,1.09375){\makebox(0,0)[bl]{$b'$ } } \put(1.3657,1.09375){\makebox(0,0)[bl]{$b''$ } } \put(0.5,0.5){\makebox(0,0){$\varphi$ } } \put(-0.0625,0.0625){\makebox(0,0)[br]{$\scriptstyle 0 $ } } \put(1.0625,0.0625){\makebox(0,0)[bl]{$\scriptstyle 1 $ } } \put(2.0625,0.0625){\makebox(0,0)[bl]{$\scriptstyle 2 $ } } \put(2.4375,-0.125){\makebox(0,0)[t]{$x$ } } \end{picture}\end{array } \label{eqhaarscaling } \\
\begin{array } [ c]{c}\setlength{\unitlength}{30bp}\begin{picture}(3,1.61)(-0.5,-0.79 ) \put(-0.5,0){\vector(1,0){3 } } \thicklines \put(-0.125,0){\line(1,0){0.125 } } \put(1,0){\line(1,0){1.125 } } \multiput(0,0)(0.5,0){2}{\line(0,1){1 } } \multiput(0,1)(0.5,-2){2}{\line(1,0){0.5 } } \multiput(0.5,-1)(0.5,0){2}{\line(0,1){1 } } \put(0.25,0.5){\makebox(0,0){$\psi$ } } \put(0,-0.0625){\makebox(0,0)[t]{$\scriptstyle 0 $ } } \put(0.53125,-0.0625){\makebox(0,0)[tl]{$\scriptstyle \frac{1}{2}$ } } \put(1.0625,-0.0625){\makebox(0,0)[tl]{$\scriptstyle 1 $ } } \put(2.4375,-0.125){\makebox(0,0)[t]{$x$ } } \end{picture}\end{array } \label{eqhaarwavelet}\ ] ] this process may be continued if you scale by powers of @xmath309 in both directions , i.e. , by @xmath310 for integral @xmath38 , @xmath311 .
so for every @xmath312 , there is a finer resolution , and if you take an up- and a shifted mirror image down - version of the dyadic scaling as in ( [ eqhaarwavelet ] ) , and allow all linear combinations , you will notice that arbitrary functions @xmath313 on the line @xmath314 , with reasonable integrability properties , admit a representation @xmath315 where the summation is over all pairs of integers @xmath316 , with @xmath38 representing scaling and @xmath32 translation .
the very simple idea of turning this construction into a multiresolution ( multi for the variety of scales in ( [ eqinttut.1 ] ) ) leads not only to an algorithm for the analysis / synthesis problem , @xmath317 in ( [ eqinttut.1 ] ) , but also to a construction of the single functions @xmath4 which solve the problem in ( [ eqinttut.1 ] ) , and which can be chosen differentiable , and yet with support contained in a fixed finite interval .
these two features , the algorithm and the finite support ( called _ compact _ support ) , are crucial for computations : computers do algorithms , but they do not do infinite intervals well .
computers do summations and algebra well , but they do not do integrals and differential equations , unless the calculus problems are discretized and turned into algorithms . in the discussion to follow , the multiresolution analysis viewpoint is dominant , which increases the role of algorithms ; for example , the so - called pyramid algorithm for analyzing signals , or shapes , using wavelets , is an outgrowth of multiresolutions . returning to ( [ eqhaarscaling ] ) and ( [ eqhaarwavelet ] )
, we see that the scaling function @xmath15 itself may be expanded in the wavelet basis which is defined from @xmath4 , and we arrive at the infinite series @xmath318 which is pointwise convergent for @xmath319 .
( it is a special case of the expansion ( [ eqinttut.1 ] ) when @xmath320 . ) in view of the picture ( ) below , ( [ eqinfiniteseries ] ) gives an alternative meaning to the traditional concept of a _ telescoping _ infinite sum . if , for example , @xmath321 , then the representation ( [ eqinfiniteseries ] ) yields @xmath322 , while for @xmath323 , @xmath324 . more generally , if @xmath325 , and @xmath326 , then @xmath327 so the function @xmath15 is itself in the space @xmath328 , and @xmath15 represents the _
initial resolution_. the tail terms in ( [ eqinfiniteseries ] ) corresponding to @xmath329 represent the _
coarser resolution_. the finite sum @xmath330 represents the _ missing detail _ of @xmath15 as a `` bump signal '' . while the sum on the left - hand side in ( [ eqtail ] ) is _ infinite _ , i.e. , the summation index @xmath38 is in the range @xmath331 , the expression @xmath332 on the right - hand side is merely a coarser scaled version of the original function @xmath15 from the subspace @xmath333 which specifies the initial resolution .
infinite sums are _ analysis problems _ while a scale operation is a single simple _
algorithmic step_. and
so we have encountered a first ( easy ) instance of the magic of a resolution algorithm ; i.e. , an instance of a transcendental step ( the analysis problem ) which is converted into a programmable operation , here the operation of scaling .
( other more powerful uses of the scaling operation may be found in the recent book @xcite by yves meyer , especially ch . 5 , and @xcite .
) the sketch below allows you to visualize more clearly this resolution versus detail concept which is so central to the wavelet algorithms , also for general wavelets which otherwise may be computationally more difficult than the haar wavelet .
@xmath334{haarbump.eps } } \put(98,146){$\varphi\!\left ( x\right ) $ } \put(104,39){$\frac{1}{2}\psi\!\left ( \frac{x}{2}\right ) $ } \put(208,56){$\frac{1}{4}\psi\!\left ( \frac{x}{4}\right ) $ } \put(312,65){$\frac{1}{8}\psi\!\left ( \frac{x}{8}\right ) $ } \put(308,102){$\frac{1}{16}\psi\!\left ( \frac{x}{16}\right ) $ } \put(413,73){\includegraphics[bb=0 0 36 46,width=18bp]{eye.eps } } \end{picture } \\ \text{\footnotesize the wavelet\index{wavelet!decomposition@--- decomposition } decomposition of haar\index{haar1@haar!scaling function@--- scaling function } 's bump function $ \varphi $ in ( \ref{eqhaarscaling } ) and ( \ref{eqinfiniteseries})}\end{gathered}\ ] ] using the sketch we see for example that the simple step function @xmath335{haardeco.eps } } \put(14,87){$a$ } \put(96,49){$b$ } \end{picture } \nonumber\end{gathered}\ ] ] has the wavelet decomposition into a sum of a _ coarser resolution _ and an _ intermediate detail _ as follows : @xmath336 thus the details are measured as differences .
this is a general feature that is valid for other functions and other wavelet resolutions .
see , for instance ,
[ exe1sep7 - 4 ] below .
while the haar wavelet is built from flat pieces , and the orthogonality properties amount to a visual tiling of the graphs of the two functions @xmath15 and @xmath4 , this is not so for the daubechies wavelet nor the other compactly supported smooth wavelets . by the balian
low theorem @xcite , a time - frequency wavelet can not be simultaneously localized in the two dual variables : if @xmath4 is a time - frequency gabor wavelet , then the two quantities @xmath337 and @xmath338 can not both be finite . since @xmath339 , this amounts to poor differentiability properties of well - localized gabor wavelets , i.e. , wavelets built using the two operations translation and frequency modulation over a lattice . but with the multiresolution viewpoint
, we can understand the first of daubechies s scaling functions as a one - sided differentiable solution @xmath15 to@xmath340 where the four real coefficients satisfy@xmath341 the system ( [ eq1.3.2 ] ) is easily solved:@xmath342 and daubechies showed that ( [ eq1.3.1 ] ) has a solution @xmath15 which is supported in the interval @xmath343 , $ ] is one - sided differentiable , and satisfies the conditions@xmath344 the first applications served as motivating ideas as well : optics , seismic measurements , dynamics , turbulence , data compression ; see the book @xcite actually , it is two books : the first one ( primarily by kahane ) is classical fourier analysis , and the second one ( primarily by p .-
lemari - rieusset ) is the wavelet book .
it will help you , among other things , to get a better feel for the french connection , the belgian connection , and the diverse and early impulses from applications in the subject . enjoy !
if we idealize and view time as discrete , a copy of @xmath7 , then a signal is a sequence @xmath345 of numbers .
a filter is an operator which calculates weighted averages@xmath346 but working instead with functions of @xmath347 , this is multiplication , @xmath348 , where @xmath349 and @xmath350 are the usual fourier representation of the corresponding generating functions .
similarly , down - sampling @xmath351 and up - sampling @xmath352 as operators on sequences take the form@xmath353 and@xmath354 since the operators @xmath351 and @xmath352 are clearly dual to one another on the hilbert space @xmath355 of sequences ( i.e. , time - signals ) , we get the corresponding duality for @xmath356 , i.e.,@xmath357 where @xmath358 denotes the normalized haar measure on @xmath66 , or equivalently the following identity for @xmath359-periodic functions:@xmath360 ( 13.166667,5)(1.5,-2.5 ) ( 1.5,0)(1,0)2 ( 1.5,0)(2,1)input ( 1.5,-1)(2,1)@xmath361 ( 3.5,0)(2,3)1.5 ( 3.5,0)(2,3)1 ( 5,2.25)(1,0)0.875 ( 3.5,0)(2,-3)1.5 ( 3.5,0)(2,-3)1 ( 5,-2.25)(1,0)0.875 ( 5.75,1.75)(1,1)@xmath362 ( 3.25,0.875)(1,1)[r]@xmath64 ( 3.25,-1.875)(1,1)[r]@xmath363 ( 5.75,-2.75)(1,1)@xmath362 ( 6.625,2.25)(1,0)2.25 ( 6.625,2.25)(1,0)1.25 ( 6.625,-2.25)(1,0)2.25 ( 6.625,-2.25)(1,0)1.25 ( 5.75,1)(1,1)@xmath364 ( 5.75,-2)(1,1)@xmath365 ( 12.666667,0)(1,0)2 ( 12.666667,0)(2,1)output ( 12.666667,-1)(2,1)@xmath361 ( 10.5,2.25)(2,-3)1.166667 ( 10.5,2.25)(2,-3)1 ( 9.625,2.25)(1,0)0.875 ( 10.5,-2.25)(2,3)1.166667 ( 10.5,-2.25)(2,3)1 ( 9.625,-2.25)(1,0)0.875 ( 8.75,1.75)(1,1)@xmath366 ( 11.25,0.875)(1,1)[l]@xmath367 ( 11.25,-1.875)(1,1)[l]@xmath368 ( 8.75,-2.75)(1,1)@xmath366 ( 8.75,1)(1,1)@xmath369 ( 8.75,-2)(1,1)@xmath370 ( 11.666667,-0.5)(1,1)@xmath371 quadrature mirror filters with @xmath3 frequency subbands @xmath372 give perfect reconstruction when signals are analyzed into subbands and then reconstructed via the up - sampling and corresponding dual filters . in engineering formalism this is expressed in the diagram in fig .
[ figperfectreconstruction ] , for @xmath174 , and @xmath64 , resp .
@xmath363 , are called low - pass , resp .
high - pass , filters .
in operator language , this takes the form@xmath373 where @xmath364 and @xmath365 are the operators in fig .
[ figperfectreconstruction ] , with dual operators @xmath374 and @xmath375 . the quadrature conditions may be expressed as@xmath376 and@xmath377 in operator theory there is tradition for working instead with the operators @xmath378 . when viewed as operators on @xmath379 they are therefore isometries with orthogonal ranges , and they satisfy@xmath380 with @xmath381 now representing the identity operator acting on @xmath379 . the relations on the @xmath382-operators are known as the cuntz relations because of their use in @xmath383-algebra theory ; see @xcite . in the present application
they take the form@xmath384 and@xmath385 and the cuntz relations are equivalent to the conditions@xmath386 and@xmath387 the last conditions are known in engineering as the quadrature conditions for the subband filters @xmath372 , with @xmath64 denoting the low - pass filter
. the low - pass and band - pass conditions on the functions @xmath388 are perhaps more familiar in the additive notation given by the substitution @xmath389 .
then the functions @xmath388 are viewed as @xmath359-periodic , and @xmath390 while@xmath391 with both of the indices @xmath392 , @xmath38 ranging over @xmath393 .
the coefficients of the functions @xmath394 are called _ impulse response coefficients _ in communications engineering , and when used in wavelets and in subdivision algorithms , they are called _ masking coefficients_. in the finite case , the @xmath394 s are also called fir for finite impulse response . the model illustrated in fig .
[ figperfectreconstruction ] is used in filter design in either hardware or software : 1 .
[ s2.1(1)]try filters @xmath64 , @xmath363 in fig .
[ figperfectreconstruction ] , and approximate the output to the input ; 2 .
[ s2.1(2)]choose a specific structure in which the filter will be realized and then quantize the coefficients , length and numerical values ; 3 .
[ s2.1(3)]verify by simulation that the resulting design meets given performance specifications .
once filters are constructed , we saw that they are also providing us with wavelet algorithms . when the steps of fig .
[ figperfectreconstruction ] are iterated , we arrive at wavelet subdivision algorithms .
relative to a given resolution ( pictured as a closed subspace @xmath395 , say , in @xmath111 ) , signals , i.e. , functions in @xmath253 , decompose into coarser ones and intermediate details .
relative to the subspaces @xmath281 and @xmath395 ,
this amounts to @xmath396{c}\uparrow\\ \text{given}\\ \text{resolution}\end{array } } { { } _ { \phantom{1}}\mathcal{v}_{1}}=\underset{\begin{array } [ c]{c}\uparrow\\ \text{coarser}\\ \text{resolution}\end{array } } { { } _ { \phantom{0}}\mathcal{v}_{0}}+\underset{\begin{array } [ c]{c}\uparrow\\ \text{intermediate}\\ \text{detail}\end{array } } { { } _ { \phantom{0.}}\mathcal{w}_{0}.}\label{eq2.1.1}\ ] ] ideally , we wish the decomposition in ( [ eq2.1.1 ] ) to be orthogonal in the sense that@xmath397 since the subdivisions involve translations by discrete steps , we specialize the resolution such that both of the spaces @xmath280 and @xmath281 are invariant under translations by points in @xmath7 , i.e. , such that@xmath398 leaves both of the subspaces @xmath280 and @xmath281 invariant . the _ multiresolution analysis _ case ( mra ) corresponds to the setup when @xmath280 is singly generated , i.e. , there is a function @xmath399 such that the closed linear span of@xmath400 is all of @xmath280 .
if @xmath174 , then there is then also a @xmath401 such that the closed linear span of @xmath402 is all of @xmath281 . if @xmath157 , we may need functions @xmath403 in @xmath281 such that @xmath404 has a closed span equal to @xmath281 . * the pyramid algorithm and the cuntz relations .
* since the two hilbert spaces @xmath356 and @xmath405 are isomorphic via the fourier series representation , it follows that the system @xmath406 is equivalent to a system @xmath407 acting on @xmath408 .
specifically , @xmath409 , @xmath410 , where @xmath411 . for @xmath412 in @xmath408 , and functions @xmath313 on @xmath6 , set @xmath413 for the present , let @xmath414 be the low - pass and high - pass wavelet filters , and let @xmath15 , @xmath4 be the corresponding scaling function , resp .
, wavelet function , also called father function , resp . , mother function .
now introduce the corresponding operators @xmath415 and their cousins @xmath416 .
the adjoints @xmath417 are also called _
then @xmath418 define @xmath419 by @xmath420 then @xmath61 maps @xmath421 isometrically onto @xmath280 in the orthogonal case and @xmath422 further @xmath423 embedding @xmath421 into @xmath424 as @xmath425 , extend @xmath61 to @xmath424 by putting @xmath426 then the extended @xmath61 maps @xmath424 isometrically onto @xmath427 and @xmath428 for all @xmath429 , where the left @xmath61 is the one from and the right is the extension of @xmath61 to @xmath424 .
the subdivision relations are equivalent to the system @xmath435 where the coefficients @xmath239 , @xmath436 are those of the quantum wavelet algorithm , i.e. , the coefficients in the `` large '' unitary matrix .
thus the quantum algorithm does the wavelet decomposition within a fixed resolution subspace .
the scaling function @xmath15 defines a resolution subspace @xmath437 .
then , or equivalently , represents the orthogonal decomposition of functions in @xmath280 into an orthogonal sum of a function with coarser resolution and a function in the intermediate detail subspace .
let @xmath64 , @xmath363 be a dyadic wavelet filter , and let @xmath438 be the corresponding matrix function , @xmath439 .
if the low - pass filter @xmath440 , then a choice for @xmath441 is @xmath442 .
we then have @xmath443 where @xmath444 , and the following @xmath445 scalar matrix can be checked to be unitary : @xmath446 except for the scalar entries in the two extreme left and right columns , all the other entries of the big combined matrix @xmath447 are taken from the cyclic arrangements of the @xmath448 matrices of coefficients @xmath449 in the expansion of @xmath450 . for the case of @xmath451
this amounts to the simple @xmath452 wavelet matrix @xmath453{c } \setlength{\unitlength}{1bp } \begin{picture}(168,130)(12,14 ) \put(96,80){\makebox(0,0){$\displaystyle \left ( \begin{array}{c|c|c|c|c } \begin{matrix } a_{1}\\ b_{1 } \end{matrix } & a_{1 } & 0 & 0 & \begin{matrix } a_{0}\\ b_{0 } \end{matrix } \\\hline \begin{matrix } 0\\ 0 \end{matrix } & a_{0 } & a_{1 } & 0 & \begin{matrix } 0\\ 0 \end{matrix } \\\hline \begin{matrix } 0\\ 0 \end{matrix } & 0 & a_{0 } & a_{1 } & \begin{matrix } 0\\ 0 \end{matrix } \\\hline \begin{matrix } a_{3}\\ b_{3 } \end{matrix } & 0 & 0 & a_{0 } & \begin{matrix } a_{2}\\ b_{2 } \end{matrix } \end{array } \right ) $ } } \put(24,126){\includegraphics[bb=0 0 192 25,width=144bp]{cycloop1.eps } } \put(22,134){\makebox(0,0)[r]{$a_{0}$ } } \put(24,14){\includegraphics[bb=0 0 192 25,width=144bp]{cycloop2.eps } } \put(170,26){\makebox(0,0)[l]{$a_{1}$ , } } \end{picture } \end{array } \label{eqwaveletmatrix}\ ] ] which is the one that produces the sequence of quantum gates .
the quantum algorithm of a wavelet filter is thus represented by a @xmath445 unitary matrix @xmath447 acting on the quantum qubit register @xmath454 , i.e. , it acts on a configuration of @xmath455 qubits .
the realization of a wavelet algorithm in the quantum realm thus amounts to spelling out the steps in factoring @xmath447 into a product of qubit gates . by shor s theorem , we know that this can be done , and @xmath447 may be built out of one - qubit gates and cnot gates following the ideas sketched above . the reader may find more discussion of the matrix @xmath447 in section 3 of @xcite .
the generalization of classical and quantum wavelet resolution algorithms from @xmath174 to @xmath157 is immediate : then @xmath456 , @xmath457 and the transformation rules@xmath458 permute the set of onb s in @xmath408 and define a unitary commuting with the @xmath3-shift .
hence , the standard formulas from @xcite , @xcite , and @xcite for the quantum computing algorithm naturally generalize to the case @xmath157 via ( [ eqintapr23.star3 ] ) . instead of @xmath38-registers @xmath459 over @xmath29 , we will now have to work rather with @xmath460 .
[ exe2oct5 - 21 ] * the pyramid algorithm and the cuntz relations revisited .
* by fourier equivalence of @xmath356 and @xmath405 via the fourier series , it follows that the system @xmath461 is equivalent to a system @xmath462 acting on @xmath408 .
specifically , @xmath409 , @xmath463 , where @xmath411 . for @xmath412 in @xmath408 , and functions @xmath313 on @xmath6 , set @xmath464 and @xmath465 let @xmath466 be low - pass and high - pass wavelet filters , and let @xmath15 , @xmath403 be the corresponding scaling function , resp . , wavelet functions .
now introduce the corresponding operators @xmath415 , and their cousins @xmath416 .
the adjoints @xmath417 are also called _
filters_. @xmath468{ccccrrrcl}\multicolumn{9}{c}{\parbox{0.9\textwidth } { \centering\small discrete\index{wavelet!discrete } vs.\ continuous\index{wavelet!continuous } wavelets , i.e. , $ \ell\sp 2 $ vs.\
$ l\sp 2\left ( \mathbb{r}\right ) $ : \rule[-12pt]{0pt}{12pt}}}\\ \mathstrut & & & & & & & & \\ $ ] \ { 0 } @xmath469@xmath469@xmath469@xmath470_2@xmath469_1@xmath469_0@xmath471@xmath472l^2 ( ) @xmath473{0pt}{6pt } & & $ ] @xmath474@xmath469@xmath475@xmath476{\raisebox{4pt}{\makebox[0pt]{\hss$\uparrow$\hss}}\raisebox{-4pt}{\makebox[0pt]{\hss$|$\hss}}}$ } & \multicolumn{1}{c}{$\makebox[6pt]{\raisebox{4pt}{\makebox[0pt]{\hss$\uparrow$\hss}}\raisebox{-4pt}{\makebox[0pt]{\hss$|$\hss}}}$ } & \multicolumn{1}{c}{$\makebox[6pt]{\raisebox{4pt}{\makebox[0pt]{\hss$\uparrow$\hss}}\raisebox{-4pt}{\makebox[0pt]{\hss$|$\hss}}}\rlap{$w$}$ } & & \\ \rule[-6pt]{0pt}{6pt}$]\ { 0 } @xmath469@xmath469@xmath477@xmath470@xmath478@xmath479{\hfill}\llap{$s_{0}^{2}\ell^{2}$}\raisebox { 12pt}{$\nearrow\vphantom{\raisebox{2pt}{$\nearrow$}}$\hskip-6pt } & \makebox[36pt]{\hfill}\llap{$s_{0}\ell^{2}$}\raisebox{12pt}{$\nearrow \vphantom{\raisebox{2pt}{$\nearrow$}}$\hskip-6pt } & $ ] ^2@xmath480 - 6pt@xmath481 the haar wavelet is supported in @xmath482 , and if @xmath483 and @xmath312 , then the modified function @xmath484 is supported in the smaller interval @xmath485 . when @xmath392 is fixed , these intervals are contained in @xmath482 for @xmath486 .
this is not the case for the other wavelet functions .
for one thing , the non - haar wavelets @xmath4 have support intervals of length more than one , and this forces periodicity considerations ; see @xcite .
for this reason , coifman and wickerhauser @xcite invented the concept of wavelet packets .
they are built from functions with prescribed smoothness , and yet they have localization properties that rival those of the ( discontinuous ) haar wavelet .
there are powerful but nontrivial theorems on restriction algorithms for wavelets @xmath487 from @xmath2 to @xmath488 .
we refer the reader to @xcite and @xcite for the details of this construction .
the underlying idea of alfred haar has found a recent renaissance in the work of wickerhauser @xcite on _ wavelet packets_. the idea there , which is also motivated by the walsh function algorithm , is to replace the refinement equation ( [ eqint.1b ] ) by a related recursive system as follows : let @xmath489 , @xmath490 , for example @xmath491 , @xmath312 , be a given low - pass / high - pass system , @xmath174 .
then consider the following _ refinement system _ on @xmath6 : @xmath492 clearly the function @xmath493 can be identified with the traditional scaling function @xmath15 of ( [ eqint.1 ] ) .
a theorem of coifman and wickerhauser ( theorem 8.1 , @xcite ) states that if @xmath494 is a partition of @xmath495 into subsets of the form @xmath496 then the function system @xmath497 is an orthonormal basis for @xmath2 .
although it is not spelled out in @xcite , this construction of bases in @xmath2 divides itself into the two cases , the true orthonormal basis ( onb ) , and the weaker property of forming a function system which is only a tight frame . as in the wavelet case , to get the @xmath494-system to really be an onb for @xmath2 , we must assume the transfer operator @xmath498 to have _ perron
frobenius spectrum _ on @xmath499 .
this means that the intersection of the point spectrum of @xmath498 with @xmath66 is the singleton @xmath500 , and that @xmath501 [ s2.2.2]subdivision algorithms + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the algorithms for wavelets and wavelet packets involve the pyramid idea as well as subdivision .
each subdivision produces a multiplication of subdivision points .
if the scaling is by @xmath3 , then @xmath392 subdivisions multiply the number of subdivision points by @xmath502 .
if the scaling is by a @xmath503 integral matrix @xmath504 , then the multiplicative factor is @xmath505 in the number of subdivision points placed in @xmath506 .
if @xmath61 is a continuous function on @xmath66 , the _ transfer operator _ or _ kneading operator _
@xmath70 @xmath509 with the alias @xmath510 in the fourier transformed space , has an adjoint which is the _ subdivision operator _ or chopping operator @xmath511 on functions @xmath361 on @xmath66 , with the alias @xmath512 on sequences .
we will analyze the duality between @xmath70 and @xmath513 and their spectra .
specializing to @xmath514 , we note that @xmath70 is then the transfer operator of orthogonal type wavelets . in the following , @xmath61
is assumed only to satisfy @xmath515 and @xmath516 .
other conditions are discussed in @xcite . in the engineering terminology of [ exe1sep7 - 4 ] , the operation ( [ eqk&c.2 ] ) is composed of a local filter with the numbers @xmath517 as coefficients , followed by the down - sampling @xmath351 , while ( [ eqk&c.4 ] ) is composed of up - sampling @xmath352 , followed by an application of a dual filter . in signal processing , @xmath351 is referred to as `` decimation '' even if @xmath3 is not @xmath518 .
the operator @xmath30 ( @xmath519 ) is called the subdivision operator , or the _ woodcutter operator _ , because of its use in computer graphics .
iterations of @xmath30 will generate a shape which ( in the case of one real dimension ) takes the form of the graph of a function @xmath313 on @xmath6 . if @xmath520 is given , and if the differences@xmath521 are small , for example if @xmath522 then we say that @xmath361 represents _ control points _ , or a control polygon , and the function @xmath313 is the limit of the _ subdivision scheme_. it follows that the subdivision operator @xmath30 on the sequence spaces , especially on @xmath523 , governs _
pointwise approximation _ to refinable limit functions .
the dual version of @xmath30 , i.e. , @xmath524 (= the transfer operator ) governs the corresponding _
mean approximation _ problem , i.e. , approximation relative to the @xmath112-norm . in scholium 4.1.2 of @xcite
, we consider the eigenvalue problem@xmath525 and @xmath526 in some suitably defined space of sequences .
the formula ( [ eqspbmay14.poundbis ] ) for the limit of a given subdivision scheme @xmath30 makes it clear that the case ( [ eqspbmay14.poundpound ] ) must be excluded . for
if ( [ eqspbmay14.poundpound ] ) holds , for some @xmath527 , and some sequence @xmath361 of control points , then there is not a corresponding regular function @xmath313 on @xmath6 with its values given on the finer grids @xmath528 , @xmath529 , by@xmath530 we show in example 4.1.3 of @xcite that there are no such control points @xmath361 in @xmath531 .
hence the stability of the algorithm !
the main difference between the algorithms of wavelets and those of wavelet packets is that for the wavelets the path in the pyramid is to one side only : a given resolution is split into a coarser one and the intermediate detail .
the intermediate detail may further be broken down into frequency bands . with the operators @xmath532 acting on @xmath356 , the coarser subspace after @xmath392 steps is modelled on @xmath533 , and the projection onto this subspace is @xmath534 where @xmath535 is the isometry of @xmath536 defined by the low - pass filter @xmath64 .
but in the construction of the wavelet packet , the subspace resulting by running the algorithm @xmath392 times is @xmath537 , and the projection onto this subspace is @xmath538 if @xmath539 , the wavelet function @xmath540 is computed from the iteration @xmath541 corresponding to the representation@xmath542 where @xmath543 are unique from the euclidean algorithm . the discussion centers around the matrix functions @xmath544 .
[ exe6dauswe-11 ] * the case @xmath545 . *
recall that we call a finite sum @xmath546 , @xmath547 a fourier polynomial both if the coefficients @xmath548 are numbers , and if they are matrices .
the matrix - valued fourier polynomials @xmath549 such that @xmath550 form a subgroup of @xmath551 which we denote @xmath552
. for every @xmath450 in @xmath552 there are @xmath553 , @xmath554 , and scalar - valued fourier polynomials @xmath555 such that @xmath556 see .
this is the first step in the daubechies sweldens lifting algorithm for the discrete wavelet transform .
thus the case @xmath557 gives a constructive lifting algorithm for wavelets , and such an algorithm has not been established in the @xmath551 case . the decomposition could also be compared with , which was mentioned in connection with the proof of .
[ exe6dec3-last ] recall the correspondence between matrix functions and wavelet filters : if @xmath544 is a matrix function , then the corresponding dyadic wavelet filters are @xmath558 it follows that the two matrix functions @xmath102 and @xmath103 satisfy @xmath559 for some @xmath560 in the ring @xmath561 of fourier polynomials if and only if @xmath562 and @xmath563 .
similarly note that the two matrix functions @xmath102 and @xmath103 satisfy @xmath564 for some @xmath565 if and only if @xmath566 and @xmath567 .
* the conclusion is that the wavelet algorithm for a general wavelet filter corresponding to a matrix function , say @xmath102 , may be broken down in a sequence of zig - zag steps acting alternately on the high - pass and the low - pass signal components .
we mentioned that for matrix functions corresponding to finite impulse response ( fir ) filters which are unitary , we need only the constant matrix ( which is chosen such as to achieve the high - pass and low - pass conditions ) and factors of the form@xmath568{c|c}z & 0\\\hline 0 & 1 \end{array } _ { \mathstrut}^{\mathstrut } \right)\ ] ] where @xmath57 is a rank - one projection in @xmath188 and @xmath3 is the scaling number of the subdivision .
unfortunately , no such factorization theorem is available for the non - unitary fir filters .
but the matrix functions take values in the non - singular complex @xmath41 matrices .
the sweldens daubechies factorization and the lifting algorithm serve as a substitute .
there are still the general non - unimodular fir - matrix functions where factorizations are so far a bit of a mystery .
the matrix functions are called _ polyphase matrices _ in the engineering literature .
the following summary serves as a classification theorem for the orthogonal wavelets of compact support : the wavelets correspond to fir polyphase matrices which are unitary . 1 .
[ genalg(1)]pick @xmath38 one - dimensional orthogonal projections @xmath570 in @xmath571 and define the unitary - valued matrix function @xmath572 on @xmath66 by@xmath573 where@xmath574{cc}1 & 1\\ 1 & -1 \end{array } \right ) .\label{eqgen.14}\ ] ] then each @xmath575 has the form@xmath576{cc}\lambda_{j } & \sqrt{\lambda_{j}\left ( 1-\lambda_{j}\right ) } e^{i\theta_{j}}\\ \sqrt{\lambda_{j}\left ( 1-\lambda_{j}\right ) } e^{-i\theta_{j } } & 1-\lambda_{j}\end{array } \right ) , \label{eqgen.15}\ ] ] where @xmath577 $ ] and @xmath578 .
( see proposition 3.3 of @xcite . ) 2 .
[ genalg(2)]define the filters @xmath579 and @xmath580 by @xmath581 with @xmath174 .
[ genalg(3)]define @xmath582 by @xmath583 if the condition @xmath584 fails , then the algorithm stops .
4 . [ genalg(4)]if the condition ( [ eqmul.12 ] ) holds , one may alternatively define @xmath15 by the cascade algorithm @xmath585{ll}1 , & 0\leq x<1,\\ 0 , & x\in\mathbb { r}\setminus\left [ 0,1\right\rangle , \end{array } \right .
\label{eqintapr25.4}\\ m_{a}&\colon\psi\longmapsto\sqrt{n}\sum_{n}a_{n}\psi\left ( nx - n\right ) .
\label{eqintmar28.4}\ ] ] 5 .
[ genalg(5)]the wavelet function @xmath4 is then defined by @xmath586 where @xmath587 are the fourier coefficients of @xmath229,@xmath588 and @xmath589 ; this is the most general wavelet function with support in @xmath569 $ ] .
[ genalg(6)]all other wavelet functions with compact support can be obtained from the ones in @xmath590 by integer translation .
[ [ bio ] ] [ s2.2bis.1]the case of polynomial functions [ the polyphase matrix , joint work with ola bratteli ] + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + one problem occurring in the biorthogonal context which does not have an analogue in the orthogonal setting stems from the fact that the duality relations @xmath591 do not give any absolute restrictions on the size of @xmath229 and @xmath592 , e.g. , a bound on the inner product of two vectors in @xmath188 does not give a bound on the size of the vectors if they are not equal .
this is reflected in the bi - cuntz relations defined by @xmath229 , @xmath593 .
let us now define@xmath594 for @xmath347 , @xmath85 . instead of the usual cuntz relations , the @xmath415 , @xmath595 now satisfy@xmath596 if @xmath597 are the matrix - valued functions associated to @xmath229 and @xmath593 by @xmath598 we compute@xmath599 in the sense that @xmath600 is contained in the commutative algebra of multiplication operators on @xmath356 defined by @xmath601 , and @xmath602 .
correspondingly,@xmath603 so all the operators @xmath600 , @xmath604 are contained in the abelian algebra @xmath601
. we may introduce operators @xmath30 , @xmath605 from @xmath606 into @xmath356 by@xmath607 and then @xmath608 maps @xmath356 into ( [ eqbio.21 ] ) , etc . , and the relations ( [ eqbio.17])([eqbio.20 ] ) take the form@xmath609{ll}s^{\ast}\tilde{s}=1 , & \text{where } 1\text { is the identity in } m_{n}\left ( \mathbb{c}\right ) \otimes c\left ( \mathbb{t}\right ) , \\
s\tilde{s}^{\ast}=1 , & \text{where } 1\text { is the identity in } c\left ( \mathbb{t}\right ) , \end{array } \right .
\label{eqbio.23}\\ & \left\ { \begin{array } [ c]{l}s^{\ast}s = aa^{\ast},\\ \tilde{s}^{\ast}\tilde{s}=\tilde{a}\tilde{a}^{\ast}. \end{array } \right .
\label{eqbio.24}\ ] ] these relations say that all combinations of products of @xmath30 and @xmath608 with @xmath605 and @xmath610 lie in the algebra @xmath611 .
but in addition @xmath102 and @xmath612 are matrix - valued functions on @xmath66 , so @xmath613 and hence@xmath614 and all the matrix - valued functions commute .
[ schbio.1]given _ any _ bijective operator @xmath30 from @xmath615 into @xmath356 one may define @xmath616 and the bi - cuntz relations are satisfied .
if , more specifically , @xmath30 is given by and , then operators @xmath617 exist such that the bi - cuntz relations are satisfied if and only if the operator @xmath618 defined by ( [ eqbio.10 ] ) is invertible , in which case one must use @xmath619 , , and to define @xmath620 . let us now connect the filters to the wavelets .
we have already defined the scaling functions @xmath15 , @xmath621 and wavelet functions @xmath622 , @xmath623 , @xmath624 .
the expansions for @xmath15 and @xmath621 converge uniformly on compacts , thus @xmath582 and @xmath625 are continuous functions on @xmath6 . to decide that these functions are in @xmath111 one again forms@xmath626 and @xmath627 similarly , and one deduces again from the nonlinear intertwining relation@xmath628 that@xmath629 in the standard case of the good old orthogonal wavelets in @xmath111 of @xmath3 subbands
, we will look for functions @xmath403 in @xmath253 such that , if @xmath38 and @xmath32 run independently over all the integers @xmath7 , i.e. , @xmath630 , then the countably infinite system of functions@xmath631 is an _
orthonormal basis _ in the hilbert space @xmath253 .
the second half of the word orthonormal refers to the restricting requirement that all the functions @xmath403 satisfy@xmath632 or stated more briefly , @xmath633 or yet more briefly , @xmath634 from familiar properties of the lebesgue measure on @xmath6 , it then follows that all the functions@xmath635 satisfy the normalization , i.e. , that@xmath636 the functions ( [ eqchexatut.3 ] ) are said to be _ orthogonal _
if@xmath637 whenever @xmath638 .
we say that the two triple indices are different if @xmath639 or @xmath640 or @xmath641 . if , for example , @xmath642 and @xmath643 , then when the same function is translated by different amounts @xmath32 and @xmath644 , the two resulting functions are required to be orthogonal .
it is an elementary geometric fact from the theory of hilbert space that if the functions in ( [ eqchexatut.3 ] ) form an orthonormal basis , then for every function @xmath89 , i.e. , every measurable function @xmath313 on @xmath6 such that @xmath645 we have the identity@xmath646 where the triple summation in ( [ eqchexatut.6 ] ) is over all configurations @xmath647 , @xmath316 .
it is convenient to rewrite ( [ eqchexatut.6 ] ) in the following more compact form:@xmath648 surprisingly , it turns out that ( [ eqchexatut.7 ] ) may hold even if the functions @xmath649 of ( [ eqchexatut.3 ] ) do not form an orthonormal basis .
it may happen that one of the initial functions @xmath650 , @xmath10 , or @xmath651 satisfies @xmath652 , and yet that ( [ eqchexatut.7 ] ) holds for all @xmath89 . these more general systems are still called wavelets , but since they are special , they are referred to as _ tight frames _ , as opposed to orthonormal bases . in either case , we will talk about a _ wavelet expansion _ of the form@xmath653 it follows that the sum on the right - hand side in ( [ eqchexatut.8 ] ) converges in the norm of @xmath111 for all functions @xmath313 in @xmath111 if ( [ eqchexatut.7 ] ) holds .
but there is a yet more general form of wavelets , called _
biorthogonal_. the conditions on the functions @xmath403 are then much less restrictive than the orthogonality axioms . hence these wavelets are more flexible and adapt better to a variety of applications , for example , to data compression , or to computer graphics .
but the biorthogonality conditions are also a little more technical to state .
we say that some given functions @xmath72
, @xmath243 , in @xmath111 are part of a biorthogonal wavelet system if there is a second system of functions @xmath654 , @xmath243 , in @xmath111 , such that every @xmath655 admits a representation@xmath656 and@xmath657 in the standard normalized case where @xmath658 , then you will notice that condition ( [ eqchexatut.7 ] ) turns into@xmath659 for all @xmath89 .
the orthogonal wavelets correspond to matrix functions @xmath660 , while the wider class of biorthogonal wavelets corresponds to the much bigger group of matrix functions @xmath661 , via the associated wavelet filters .
you may ask , why bother with the more technical - looking biorthogonal systems ?
it turns out that they are forced on us by the engineers .
they tell us that the real world is not nearly as orthogonal as the mathematicians would like to make it out to be .
there is a paucity of symmetric orthogonal wavelets , and symmetry ( `` linear phase '' ) is prized by engineers and workers in image processing , where the more general wavelet families and their duality play a crucial role . now
what if we could change the biorthogonal wavelets into the orthogonal ones , and still keep the essential spectral properties intact ?
then everyone will be happy .
this last chapter shows that it is possible , and even in a fairly algorithmic fashion , one that is amenable to computations .
wavelet filters may be understood as matrix functions , i.e. , functions from the one - torus @xmath662 into some group of invertible matrices . if the scale number is @xmath3 , then there are three such matrix groups which are especially relevant for wavelet analysis : @xmath663{\parbox[t]{0.22\textwidth}{\raggedright $ \mathrm{u}_{n}\left ( \mathbb{c}\right ) $ : all unitary $ n\times n$ complex matrices\index{matrix!unitary } } } \subset \framebox[0.24\textwidth]{\parbox{0.22\textwidth}{\raggedright $ \mathrm{gl}_{n}\left ( \mathbb{c}\right ) $ : all invertible $ n\times n$ complex matrices\index{matrix!invertible } } } \supset \framebox[0.24\textwidth]{\parbox[t]{0.22\textwidth}{\raggedright $ \mathrm{sl}_{n}\left ( \mathbb{c}\right )
$ : all $ n\times n$ complex matrices $ a$ with $ \det a=1$.}}\ ] ] it is possible to reduce some questions in the @xmath664 case to better understood results for @xmath42 ; see chapter 6 of @xcite .
the @xmath665 case is especially interesting in view of daubechies sweldens lifting for dyadic wavelets ; see [ exe6dauswe-11 ] .
* definitions : * a function , or a distribution , @xmath15 satisfying ( [ eqint.1 ] ) is said to be _
refinable _ , the equation ( [ eqint.1 ] ) is called the _ refinement equation _ , or also , as noted above , the `` scaling identity '' , and @xmath15 is called the scaling function .
the coefficients @xmath239 of ( [ eqint.1 ] ) are called the _ masking coefficients_. we will mainly concentrate on the case when the set @xmath666 is finite .
but in general , a function @xmath667 is said to be refinable with scale number @xmath3 if @xmath668 is in the @xmath669-closed linear span of the translates @xmath670 ; see , e.g. , @xcite
. since there are refinement operations which are more general than scaling see for example , there are variations of which are correspondingly more general , with regard to both the refinement steps that are used and the dimension of the spaces . the term `` scaling identity '' is usually , but not always , reserved for , while more general refinements lead to `` refinement equations '' .
however , often goes under both names .
the vector versions of the identities get the prefix `` multi- '' , for example _ multiscaling _ and _ multiwavelet_. if @xmath64 satisfies a condition for obtaining orthogonal wavelets , @xmath671 together with the normalization@xmath672 then ( [ eqint.1 ] ) has a solution @xmath15 in @xmath111 which can be obtained by taking the inverse fourier transform of the product expansion@xmath673 ( here and later we use the convention that if @xmath674 is a function of @xmath347 , then @xmath675 . ) that ( [ eqintnew.6 ] ) gives a solution @xmath15 of ( [ eqint.1 ] ) follows from the relation @xmath676 we mentioned that there is a direct connection between @xmath677 and the scaling function @xmath15 on @xmath6 given in ( [ eqintnew.2 ] ) , ( [ eqint.1 ] ) , and ( [ eqintnew.6 ] ) .
there is a similar correspondence between the high - pass filters @xmath229 and the wavelet generators @xmath678 . in the _ biorthogonal _ case
, there is a second system @xmath679 and the two systems @xmath680 then form a dual wavelet basis , or dual wavelet frame for @xmath111 in the sense of @xcite , chapter 5 .
we considered this biorthogonal case in more detail in [ bio ] above .
much more detail can be found in chapter 6 of @xcite .
the idea of constructing maximally smooth wavelets when some side conditions are specified has been central to much of the activity in wavelet analysis and its applications since the mid-1980 s . as a supplement to @xcite ,
the survey article @xcite is enjoyable reading .
the paper @xcite treats the issue in a more specialized setting and is focussed on the moment method .
some of the early applications to data compression and image coding are done very nicely in @xcite , @xcite , and @xcite .
an interesting , related but different , algebraic and geometric approach to the problem is offered in @xcite .
we now turn to an interesting variation of this setup , which includes higher dimensions , i.e. , when the hilbert space is @xmath681 , @xmath682 . staying for the moment with @xmath507 , and @xmath3 fixed
, we will take the viewpoint of what is called _ resolutions _ , but here understood in a broad sense of closed subspaces : a closed linear subspace @xmath333 is said to be an @xmath3-resolution if it is invariant under the unitary operator @xmath683 i.e. , if @xmath136 maps @xmath684 into a proper subspace of itself .
the subspace @xmath684 is said to be _ translation invariant _
if @xmath685 if there is a function @xmath15 such that @xmath686 is the closed linear span of @xmath687 then clearly @xmath684 is translation invariant . the translation - invariant resolution subspaces @xmath684
are actively studied and reasonably well understood . if @xmath684 is of the form @xmath688 in ( [ eqintoct16.8 ] ) , then we say that it is _ singly generated _ , and that @xmath15 is a scaling function of scale @xmath3 .
[ [ s2.2bis.3.2generalized-multiresolutions-joint-work-with-l.baggett-k.merrill-and-j.packer ] ] [ s2.2bis.3.2]generalized multiresolutions [ joint work with l. baggett , k. merrill , and j. packer ] + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the case when the resolution subspace @xmath684 is not singly generated is also interesting , and these resolution subspaces are frequently called _ generalized multiresolution subspaces _ ( gmra ) .
there is much current and very active research on them ; see , for example , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , and @xcite .
the case when @xmath684 is not singly generated as a resolution subspace of scale @xmath157 , i.e. , when @xmath684 is not of the form ( [ eqintoct16.8 ] ) , occurs in the study of _ wavelet sets_. a wavelet set in @xmath506 is defined relative to an expansive @xmath503 matrix @xmath504 over @xmath7 .
a subset @xmath689 is said to be an @xmath504-wavelet set if there is a single wavelet function @xmath690 such that @xmath691 .
specifically , the condition states that the family @xmath692 is an orthonormal basis for @xmath693 .
this can be checked to be equivalent to the combined set of two tiling properties for @xmath49 as a subset of @xmath506 : we define tiling by the requirement that the sets in the family have overlap at most of measure zero relative to lebesgue measure on @xmath506 .
similarly , the union@xmath696 is understood to be only up to measure zero .
it is easy to see that compactly supported wavelets in @xmath697 are mra wavelets , while most wavelets @xmath698 from wavelet sets @xmath49 are not .
thess wavelets are typically ( but not always ) frequency localized .
the main difference between the gmra ( stands for generalized multiresolution analysis ) wavelets and the more traditional mra ones may be understood in terms of multiplicity . both come from a fixed resolution subspace @xmath699 which is invariant under the translations @xmath700 where @xmath701 hence @xmath702 is a unitary representation of @xmath703 on the hilbert space @xmath280 . as a result of stone s theorem
, we find that there are subsets@xmath704 of @xmath705 such that the spectral measure of the ( restricted ) representation has multiplicity @xmath706 on the subset @xmath707 , @xmath708 .
it can be checked that the projection - valued spectral measure is absolutely continuous . moreover
, there is an intertwining unitary operator@xmath709 such that@xmath710 holds for all @xmath711 and @xmath712 .
we may then consider the functions @xmath713 ( @xmath714 ) defined by @xmath715 it was proved by baggett and merrill @xcite that @xmath716 generates a normalized tight frame for @xmath280 : specifically , that@xmath717 holds for all @xmath711 .
treating @xmath718 as a vector - valued function , denoted simply by @xmath15 , we see that there is a matrix function@xmath719 such that@xmath720[0pt]{$\scriptstyle\nearrow$\raisebox{5pt}[0pt][0pt]{\small for transpose}}}}t\right ) = h\left ( e^{it}\right ) \hat{\varphi}\left ( t\right ) , \label{eq2.2bis.3.7bis}\ ] ] where @xmath721 , and @xmath722 . but this method takes the hilbert space @xmath693 as its starting point , and then proceeds to the construction of wavelet filters in the form ( [ eq2.2bis.3.7bis ] ) .
our current joint work with baggett , merrill , and packer reverses this .
it begins with a matrix function @xmath144 defined on @xmath705 , and then offers subband conditions on the matrix function which allow the construction of a gmra for @xmath697 with generator @xmath723 given by ( [ eq2.2bis.3.7bis ] ) .
so the hilbert space @xmath693 shows up only at the end of the construction , in the conclusions of the theorems . in using the polyphase matrices , one may only have the first few rows , and be faced with the problem of completing to get the entire function @xmath102 from a torus into the matrices of the desired size .
the case when only the first row is given , say corresponding to a specified low - pass filter , is treated in @xcite and @xcite , and we refer the reader to the references given there , especially @xcite , @xcite , @xcite , and @xcite .
the wavelet transfer operator is used in a variety of wavelet applications not covered here , or only touched upon tangentially : stability of refinable functions , regularity , approximation order , unitary matrix extension principles , to mention only a few .
the reader is referred to the following papers for more details on these subjects : @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , and @xcite . for the sake of illustration ,
let us take @xmath726 , and scaling number @xmath727 , i.e. , the case of dyadic framelets .
naturally , the notion of tight frame is weaker than that of an orthonormal basis ( onb ) , and it is shown in @xcite that when a system of wavelet filters @xmath233 , @xmath728 is given ( @xmath729 must be low - pass ) , then the orthogonality condition on the @xmath233 s always gets us a framelet in @xmath730 , i.e. , the functions @xmath234 corresponding to the high - pass filters @xmath233 , @xmath731 , generate a tight frame for @xmath730 , also called a framelet .
the correspondence @xmath233 to @xmath234 is called the uep in @xcite .
the orthogonality condition for @xmath233 , @xmath728 , referred to in the uep is simply this : form an @xmath732-by-@xmath309 matrix - valued function @xmath733 by using @xmath734 , @xmath728 in the first column , and the translates of the @xmath233 s by a half period , i.e. , @xmath735 , @xmath728 in the second .
the condition on this matrix function @xmath733 is that the two columns are orthogonal and have unit norm in @xmath736 for all @xmath52 .
note that we still get the unitary matrix functions acting on these systems , in the way we outlined above .
but there is redundancy as the unitary matrices are @xmath732-by-@xmath732 .
the reader is refered to @xcite for further details .
we emphasize that several of these , and other related topics , invite the kind of probabilistic tools that we have stressed here . but a more systematic discussion is outside the scope of this brief set of notes .
we only hope to offer a modest introduction to a variety of more specialized topics .
[ snew5aug2.3.4(a ) ] the orthogonality condition for @xmath233 , @xmath728 , may be stated in terms of the operators @xmath737 from equation ( [ eq2.9 ] ) , @xmath727 . for each @xmath738 , define an operator on @xmath739 as in ( [ eq2.9 ] ) .
then the arguments from section [ s2 ] show that the orthogonality condition for @xmath233 , @xmath728 , i.e. , the uep condition , is equivalent to the operator identity ( [ eq2.8 ] ) where the summation now runs from @xmath22 to @xmath740 .
operator systems @xmath737 satisfying ( [ eq2.8 ] ) are called row - isometries .
[ snew5aug2.3.4(b ) ] there are two properties of the low - pass filter @xmath729 which we have glossed over .
first , @xmath729 must be such that the corresponding scaling function @xmath15 is in @xmath730 . without an added condition on @xmath729
, @xmath15 might only be a distribution .
secondly , when the dyadic scaling in @xmath730 is restricted to the resolution subspace @xmath741 , the corresponding unitary part must be zero .
these two issues are addressed in @xcite , @xcite , and @xcite .
the two groups of matrix functions @xmath742 and @xmath743 , i.e. , the continuous functions from the torus into the respective groups , enter wavelet analysis via the associated wavelet filters @xmath744 . 1 . [ intmultcorr(1)]matrix functions , @xmath745 , 2 .
[ intmultcorr(2)]high- and low - pass wavelet filters @xmath229 , @xmath746 , @xmath747 , and 3 .
[ intmultcorr(3)]wavelet generators @xmath72 , @xmath748 , @xmath749 , together with scaling functions @xmath750 , @xmath751 . in particular,@xmath752 { \frac{1}{n}\sum_{w^{n}=z}m_{i}\left ( w\right ) w^{-j } , } & & z\in\mathbb{t},\qquad\label{eqintfeb22.11}\\ \qquad\left ( a^{-1}\right ) _ { i , j } & = \frac{1}{n}\sum_{w^{n}=z}\overline{\tilde{m}_{j}\left ( w\right ) } \,w^{i } , & & z\in\mathbb{t}.\qquad\label{eqintfeb22.12}\ ] ] the dependence of the @xmath111-functions in ( [ intmultcorr(3 ) ] ) on the group elements @xmath102 from ( [ intmultcorr(1 ) ] ) gives rise to homotopy properties .
the standard orthogonal wavelets represent the special case when @xmath753 , or equivalently , @xmath754 , @xmath347 .
hence , the matrix functions are unitary in this case .
the scaling / wavelet functions @xmath755 with support on a fixed compact interval , say @xmath270 $ ] , @xmath756 , can be parameterized with a finite number of parameters since the unitary - valued function @xmath236 in ( [ eqintfeb22.11 ] ) then is a polynomial in @xmath222 of degree at most @xmath757 .
it is well - known folklore from computer - generated pictures that the shape of the scaling / wavelet functions depends continuously on these parameters ; see figures 1.11.7 in @xcite and @xcite .
the scaling function @xmath758 of ( [ eqint.1 ] ) is illustrated there , in the case @xmath174 , and for orthogonal @xmath7-translates , i.e. , the case ( [ eqintnew.4 ] ) .
these pictures illustrate the dependence of @xmath15 on the masking coefficients @xmath759 in the case of @xcite:@xmath760 where @xmath761 these formulas arise from an independent pair of rotations by angles @xmath762 and @xmath763 of two `` spin vectors '' , i.e. , by taking the matrix function @xmath102 in ( [ eqintfeb22.11 ] ) unitary , @xmath764 , and setting @xmath765 with@xmath766{cc}1 & 1\\ 1 & -1 \end{array } \right ) , \label{eqa.8bis}\\[3\jot ] q_{\theta}&=\left ( \begin{array } [ c]{cc}\cos^{2}\theta & \cos\theta\sin\theta\\ \cos\theta\sin\theta & \sin^{2}\theta \end{array } \right ) \nonumber \\ & = \frac{1}{2}\left ( \left ( \begin{array } [ c]{cc}1 & 0\\ 0 & 1 \end{array } \right ) + \left ( \begin{array } [ c]{cc}\cos2\theta & \sin2\theta\\ \sin2\theta & -\cos2\theta \end{array }
\right ) \right ) , \label{eqa.9}\ ] ] and the orthogonal complement to the one - dimensional projection @xmath767,@xmath768{cc}1 & 0\\ 0 & 1 \end{array } \right ) + \left ( \begin{array } [ c]{cc}\cos2\theta & \sin2\theta\\ \sin2\theta & -\cos2\theta \end{array }
\right ) \right ) , } \label{eqa.10}\ ] ] with the coefficients @xmath769 , @xmath770 , @xmath771 , @xmath772 , @xmath773 , @xmath774 given by ( [ eqa.11 ] ) , the algorithmic approach to graphing the solution @xmath15 to the scaling identity ( [ eqint.1 ] ) is as follows ( see @xcite , @xcite for details ) : the relation ( [ eqint.1 ] ) for @xmath174 is interpreted as giving the values of the left - hand @xmath15 by an operation performed on those of the @xmath15 on the right , and a binary digit inversion transforms this into the form @xmath775 where @xmath776 is the @xmath777 matrix @xmath778 constructed from the coefficients in ( [ eqint.1 ] ) , and @xmath779 and @xmath780 are the vector functions @xmath781 the signal processing aspect can be understood from the description of subband filters in the analusis and synthesis of time signals , or more general signals for images . in either case , we have two subband systems @xmath782 and @xmath783 where the functions@xmath784 are the generating functions defined from the filter coefficients @xmath785 and @xmath786 , @xmath787 .
originally we had anticipated adding two more chapters to these tutorials , but time and space prevented this .
instead we include the table of contents for this additional material .
the details for the remaining chapters will be published elsewhere .
but as the items in the list of contents suggest , there are still many exciting open problems in the subject that the reader may wish to pursue on his / her own .
we feel that the following list of topics offers at least an outline of several directions that the reader , could take in his / her own study and research on wavelet - related mathematics .
[ [ s3.2intertwining - operators - between - sequence - spaces - l2-and - l2rn ] ] [ s3.2]intertwining operators between sequence spaces @xmath1 and @xmath0 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [ [ s3.4dependence - of - the - wavelet - functions - on - the - matrix - functions - which - define - the - wavelet - filters ] ] [ s3.4]dependence of the wavelet functions on the matrix functions which define the wavelet filters ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ several versions of spectral duality are presented .
on the two sides we present ( 1 ) a basis condition , with the basis functions indexed by a frequency variable , and giving an orthonormal basis ; and ( 2 ) a geometric notion which takes the form of a tiling , or a iterated function system ( ifs ) .
our initial motivation derives from the fuglede conjecture , see @xcite : for a subset @xmath788 of @xmath789 of finite positive measure , the hilbert space @xmath790 admits an orthonormal basis of complex exponentials , i.e. , @xmath788 admits a fourier basis with some frequencies @xmath791 from @xmath789 , if and only if @xmath788 tiles @xmath789 ( in the measurable category ) where the tiling uses only a set @xmath792 of vectors in @xmath789 .
if some @xmath788 has a fourier basis indexed by a set @xmath791 , we say that @xmath793 is a spectral pair .
we recall from @xcite that if @xmath788 is an @xmath32-cube , then the sets @xmath791 in ( 1 ) are precisely the sets @xmath792 in ( 2 ) .
this begins with work of jorgensen and steen pedersen @xcite where the admissible sets @xmath794 are characterized .
later it was shown , @xcite and @xcite that the identity @xmath795 holds for all @xmath32 .
the proofs are based on general fourier duality , but they do not reveal the nature of this common set @xmath794 .
a complete list is known only for @xmath451 , @xmath309 , and @xmath161 , see @xcite .
we then turn to the scaling ifs s built from the @xmath32-cube with a given expansive integral matrix @xmath102 .
each @xmath102 gives rise to a fractal in the small , and a dual discrete iteration in the large . in a different paper @xcite , jorgensen and pedersen
characterize those ifs fractal limits which admit fourier duality .
the surprise is that there is a rich class of fractals that do have fourier duality , but the middle third cantor set does not .
we say that an affine ifs , built on affine maps in @xmath789 defined by a given expansive integral matrix @xmath102 and a finite set of translation vectors , admits fourier duality if the set of points @xmath791 , arising from the iteration of the @xmath102-affine maps in the large , forms an orthonormal fourier basis ( onb ) for the corresponding fractal @xmath358 in the small , i.e. , for the iteration limit built using the inverse contractive maps , i.e. , iterations of the dual affine system on the inverse matrix @xmath796 . by `` fractal in the small ''
, we mean the hutchinson measure @xmath358 and its compact support , see @xcite .
( the best known example of this is the middle - third cantor set , and the measure @xmath358 whose distribution function is corresponding devil s staircase . ) in other words , the condition is that the complex exponentials indexed by @xmath791 form an onb for @xmath797 .
such duality systems are indexed by complex hadamard matrices @xmath144 , see @xcite and @xcite ; and the duality issue is connected to the spectral theory of an associated ruelle transfer operator , see @xcite .
these matrices @xmath144 are the same hadamard matrices which index a certain family of quasiperiodic spectral pairs @xmath793 studied in @xcite and @xcite
. they also are used in a recent construction of terence tao @xcite of a euclidean spectral pair @xmath793 in @xmath798 for which @xmath788 does not a tile @xmath798 with any set of translation vectors @xmath792 in @xmath798 ; see also @xcite .
we finally report on joint research with dorin dutkay @xcite , @xcite , @xcite , @xcite where we show that all the affine ifs s , and more general limit systems from dynamics and probability theory , admit wavelet constructions , i.e. , admit orthonormal bases of wavelet functions in hilbert spaces which are constructed directly from the geometric data .
a substantial part of the picture involves the construction of limit sets and limit measures , a part of geometric measure theory .
* acknowledgements : * [ ack]we are happy to thank the organizing committee at the national university of singapore for all their dedicated work in planning and organizing a successful conference , of which this tutorial is a part .
we especially thank professors wai shing tang , judith packer , zuowei shen , and the head of the department of mathematics of the nus for all their work in making my visit to singapore possible .
we thank the institute for mathematical sciences at the nus , and the us national science foundation , for partial financial support in the preparation of these lecture notes .
we discussed various parts of the mathematics with our colleagues , professors larry baggett , david larson , ola bratteli , kathy merrill , judy packer , and we thank them for their encouragements and suggestions .
the typesetting and graphics were expertly done at the university of iowa by brian treadway .
we also thank brian treadway for a number of corrections , and for very helpful suggestions , leading to improvements of the presentation .
baggett , p.e.t .
jorgensen , k.d .
merrill , and j.a .
packer , _ construction of parseval wavelets from redundant filter systems _ , preprint , 2003 , submitted to j. amer .
, arxiv : math.ca/0405301 .
baggett , p.e.t .
jorgensen , k.d .
merrill , and j.a .
packer , _ an analogue of bratteli - jorgensen loop group actions for gmra s _ , in wavelets , frames , and operator theory ( college park , md , 2003 ) , ed . c. heil , p.e.t .
jorgensen , and d.r .
larson , contemp .
345 , american mathematical society , providence , 2004 , pp . 1125 .
baggett and d.r .
larson ( eds . ) , _ the functional and harmonic analysis of wavelets and frames : proceedings of the ams special session on the functional and harmonic analysis of wavelets held in san antonio , tx , january 1314 , 1999 _ , contemp .
247 , american mathematical society , providence , 1999 .
baggett and k.d .
merrill , _ abstract harmonic analysis and wavelets in @xmath799 _ , in the functional and harmonic analysis of wavelets and frames ( san antonio , 1999 ) , ed .
baggett and d.r .
larson , contemp .
247 , american mathematical society , providence , 1999 , pp . 1727 .
o. bratteli and p.e.t .
jorgensen , _ convergence of the cascade algorithm at irregular scaling functions _
, in the functional and harmonic analysis of wavelets and frames ( san antonio , 1999 ) , ed .
baggett and d.r .
larson , contemp .
247 , american mathematical society , providence , 1999 , pp . 93130 .
to3em , _ wavelet filters and infinite - dimensional unitary groups _ , in wavelet analysis and applications ( guangzhou , china , 1999 ) , ed .
d. deng , d. huang , r .- q .
jia , w. lin , and j. wang , ams / ip studies in advanced mathematics , vol .
25 , american mathematical society , providence , international press , boston , 2002 , pp . 3565 .
coifman and m.v .
wickerhauser , _ wavelets and adapted waveform analysis : a toolkit for signal processing and numerical analysis _ , in different perspectives on wavelets ( san antonio , tx , 1993 ) ,
i. daubechies , proc .
47 , american mathematical society , providence , 1993 , pp . 119153 .
d. esteban and c. galand , _ application of quadrature mirror filters to split band voice coding systems _ , in ieee international conference on acoustics , speech , and signal processing ( washington , dc , may 1977 ) , institute of electrical and electronics engineers , piscataway , nj , 1977 , pp .
191195 .
a. fijany and c.p .
williams , _ quantum wavelet transforms : fast algorithms and complete circuits _ , in quantum computing and quantum communications ( palm springs , ca , 1998 ) , ed .
williams , lecture notes in computer science , vol .
1509 , springer , berlin , 1999 , pp .
freedman , _ poly - locality in quantum computing _ , found .
comput . math .
* 2 * ( 2002 ) , 145154 .
b. fuglede , _ commuting self - adjoint partial differential operators and a group - theoretic problem _ , j. funct .
16 * ( 1974 ) , 101121 .
d. han , d. r. larson , m. papadakis , and th .
stavropoulos , _ multiresolution analyses of abstract hilbert spaces and wandering subspaces _ , in the functional and harmonic analysis of wavelets and frames ( san antonio , 1999 ) , ed . l.w .
baggett and d.r .
larson , contemp .
247 , american mathematical society , providence , 1999 , pp . 259284 .
heller , j.m .
shapiro , and r.o .
wells , _ optimally smooth symmetric quadrature mirror filters for image coding _ , in wavelet applications ii ( orlando , 1995 ) , ed .
szu , proceedings of spie , vol .
2491 , society of photo - optical instrumentation engineers , bellingham , wa , 1995 , pp .
119130 .
heller , v. strela , g. strang , p. topiwala , c. heil , and l.s .
hills , _ multiwavelet filter banks for data compression _ , in ieee international symposium on circuits and systems , 1995 ( iscas 95 ) , institute of electrical and electronics engineers , new york , 1995 , vol . 3 , pp .
17961799 .
to3em , invited featured book review of _ an introduction to wavelet analysis _ by david f. walnut , applied and numerical harmonic analysis , birkhuser , 2002 , bull .
( n.s . ) * 40 * ( 2003 ) , 421427 .
a. klappenecker , _ wavelets and wavelet packets on quantum computers _ , in wavelet applications in signal and image processing vii ( denver , 1999 ) , ed .
unser , a. aldroubi , and a.f .
laine , proceedings of spie , vol .
3813 , society of photo - optical instrumentation engineers , bellingham , wa , 1999 , pp .
703713 .
m. lang and p.n .
heller , _ the design of maximally smooth wavelets _ , in ieee international conference on acoustics , speech , and signal processing , 1996 ( icassp-96 ) , institute of electrical and electronics engineers , new york , 1996 , vol . 3 , pp .
14631466 .
to3em , _ wavelets and functions with bounded variation from image processing to pure mathematics _ , atti accad .
lincei cl .
fis . mat .
lincei ( 9 ) mat .
* 11 * ( 2000 ) , 77105 , special issue : _ mathematics towards the third millennium _
( papers from the international conference held in rome , may 2729 , 1999 ) .
v. perrier and m.v .
wickerhauser , _ multiplication of short wavelet series using connection coefficients _ , in advances in wavelets ( hong kong , 1997 ) , ed .
lau , springer - verlag , singapore , 1999 , pp .
77101 .
riemenschneider and z. shen , _ box splines , cardinal series , and wavelets _ , in approximation theory and functional analysis ( college station , texas , 1990 ) , ed .
chui , academic press , boston , 1991 , pp .
133149 .
g. strang , v. strela , and d .- x .
zhou , _ compactly supported refinable functions with infinite masks _ , in the functional and harmonic analysis of wavelets and frames ( san antonio , 1999 ) , ed .
baggett and d.r .
larson , contemp .
247 , american mathematical society , providence , 1999 , pp . 285296 .
wickerhauser , _ best - adapted wavelet packet bases _ , in different perspectives on wavelets ( san antonio , tx , 1993 ) , ed .
i. daubechies , proc .
47 , american mathematical society , providence , 1993 , pp . 155171 | [ abs]abstract : some connections between operator theory and wavelet analysis : since the mid eighties , it has become clear that key tools in wavelet analysis rely crucially on operator theory . while isolated variations of wavelets , and wavelet constructions had previously been known , since haar in 1910 , it was the advent of multiresolutions , and sub - band filtering techniques which provided the tools for our ability to now easily create efficient algorithms , ready for a rich variety of applications to practical tasks .
part of the underpinning for this development in wavelet analysis is operator theory .
this will be presented in the lectures , and we will also point to a number of developments in operator theory which in turn derive from wavelet problems , but which are of independent interest in mathematics . some of the material will build on chapters in a new wavelet book , co - authored by the speaker and ola bratteli , see http://www.math.uiowa.edu/jorgen/0.1em .
[ subsubsection ] [ subsubsection ] [ subsubsection ] [ subsubsection ] [ subsubsection ] [ subsubsection ] [ subsubsection ] l |
the system is composed of a set of elements interconnected forming a scale - free undirected random network .
every node can present two states , active and inactive , and the dynamics that governs their state is similar to the standard sis model .
nodes flip from one state to the other following a poisson process with probability densities @xmath70 for activation and @xmath71 for deactivation . in particular
, a node @xmath72 flips from the active to the inactive state at a constant rate @xmath12 , while its activation rate is proportional to the number of active nodes to which it is connected , @xmath73 . in the latter sum , @xmath74 if node @xmath75 is active at time @xmath76 and @xmath77 otherwise ; similarly , @xmath78 if and only if nodes @xmath72 and @xmath75 are connected .
we now show that @xmath79 is , after the system has reached the steady - state , approximately time - independent and proportional to the degree of the node @xmath80 @xcite ; we will refer to it as @xmath11 . to derive its expression
, we first apply a coarse - graining approximation by assuming that all nodes with the same degree @xmath5 are equivalent and statistically independent . under this approximation ,
the state of the system is characterized by the fraction of nodes in the active state in every degree class , @xmath81 where @xmath82 stands for the number of nodes with degree @xmath5 in the network .
since we are assuming all nodes with the same degree to be statistically independent , @xmath83 yields the probability that a randomly chosen node with degree @xmath5 is active .
hence , the probability that a randomly chosen link emerging from a node with degree @xmath5 reaches an active node can be computed as @xmath84 , i.e. , the probability that the link reaches a node with degree @xmath9 , @xmath85 , and that such node is active , @xmath86 , summed for all @xmath9 . in the case of uncorrelated random networks , @xmath87 , so the probability for a link to connect to an active node becomes degree independent , @xmath88 .
moreover , in the steady - state it takes an almost constant value @xmath89 ( see fig . [
fig : wprevalence ] ) , which allows us to compute the average number of active neighbors of a node with degree @xmath5 as @xmath90 .
hence , @xmath91 in addition to the sis - like dynamics , we let nodes produce while they are in the active state with a probability density @xmath0 , which need not be exponential .
it is therefore important to note that , if @xmath92 gives the probability for the time interval between two consecutive production events to fall into @xmath93 ( for a node that is active along all the process ) , the probability density for the interval between the activation time and the first production event @xmath20 is , in general , given by @xmath94 where @xmath95 is the average production interval and @xmath96 is the survival probability distribution ( the probability that the interval is greater than @xmath76 ) .
we will use the same notation for the activation and deactivation probabilities as well : @xmath97 will refer to the survival probability of @xmath98 . to justify eq .
, let us consider a sequence of @xmath99 events @xmath100 distributed with probability density @xmath0 ( so the whole process takes a time @xmath101 ) and calculate @xmath102 , i.e. , the probability that an activation event falls at a distance between @xmath76 and @xmath103 from the next production event . assuming that the activation event is uniformly distributed along the interval @xmath104 , the probability for it to fall in any interval of length @xmath105 is simply @xmath106 . in our case , @xmath107 , where @xmath108 is the number of intervals greater than @xmath76 , since there is an interval of length @xmath109 placed at distance @xmath76 from the next production event in every production interval greater than @xmath76 .
therefore , @xmath110
in the latter section we have reduced the study of the system to that of a single node that activates with rate @xmath11 given by eq . , deactivates with rate @xmath12 and produces in the active state with probability densities @xmath0 and @xmath20 .
we now compute the probability @xmath112 that two consecutive production events take place in the interval @xmath113 taking into account the fact that the state of the node can flip any number of times between the two events ( as long as it is in the active state during both productions ) .
it is easy to see that if it does not flip to the inactive state during the process , @xmath114 is given by @xmath115 ( the probability that it does not deactivate in a time less than @xmath76 and that it produces in the given interval ) .
similarly , we can compute the probability that it deactivates only once between both productions ; however , in this case we have to integrate for all possible deactivation and activation times ( let us call them @xmath116 and @xmath117 , respectively ) , yielding @xmath118 . following the same procedure ,
we can write the total probability density as the series @xmath119 we do not need to solve the series , since we are only interested in the first two moments of the distribution to obtain @xmath34 . to this end , we make use of the laplace transform formalism , which is particularly convenient for two reasons . on the one hand ,
the @xmath120-th moment of a distribution @xmath111 can be obtained as the @xmath120-th derivative of its transform @xmath121 , @xmath122 . on the other hand ,
the transform of a convolution is @xmath123 . since any term in eq .
is a chain of convolutions , it is not hard to write down @xmath121 , @xmath124 where in the last step we have made use of the fact that @xmath125 .
we can write the latter result in terms of @xmath0 only by plugging the expressions for @xmath70 and @xmath71 into it and using @xmath126 @xmath127 and @xmath128 the final expression for @xmath121 reads @xmath129^{2}}{\frac{s + \lambda_{eff}(k)}{\lambda_{eff}(k ) } - \frac{\lambda_{i}}{s + \lambda_{i } } \left [ 1 - \frac{1 - \hat{\psi } ( s + \lambda_{i})}{\langle t
_ { p}\rangle \left ( s + \lambda_{i } \right ) } \right]}.\ ] ] up to this point , we have solved the general problem for an unspecified production probability density @xmath0 . in this paper
we have considered two different cases , which will be studied separately in two subsections . in the first stage of this work ,
we have considered a weibull probability density function with scale parameter @xmath130 and shape parameter @xmath131 , @xmath132 whose intrinsic burstiness coefficient is @xmath133 .
plugging its laplace transform @xmath134 into eq .
yields a rather cumbersome expression , @xmath135^{2}}{4 s \left ( s + \lambda_{i } \right ) \left ( s + \lambda_{i } + \lambda_{eff}(k ) \right ) + 4 \lambda_{i } \lambda_{eff}(k ) \beta - 2 \lambda_{i } \lambda_{eff}(k ) \beta e^{\frac{\beta}{2 \left ( s + \lambda_{i } \right ) } } \sqrt{\frac{2 \pi \beta}{s + \lambda_{i } } } \text{erfc } \left ( \sqrt{\frac{\beta}{2 \left ( s + \lambda_{i } \right ) } } \right)}. \end{aligned}\ ] ] we can now obtain the first and second moments of @xmath111 deriving the latter equation and therefore write the expressions for the mean @xmath136 , the standard deviation @xmath137 and @xmath138 in terms of the normalized parameters @xmath139 and @xmath140 : @xmath141 @xmath142 hence , @xmath143 we can repeat the same procedure as in the latter subsection for an exponential production probability density , @xmath144 whose intrinsic burstiness coefficient is @xmath145 .
its laplace transform is given by @xmath146 so eq . becomes @xmath147 evaluating its first and second derivatives at @xmath35 and using normalized parameters @xmath148 and @xmath140 gives @xmath149 @xmath150 and @xmath151
let us focus on the case of a single node that produces with an arbitrary @xmath0 and consider several limiting cases .
in order to define a timescale , we set @xmath152 throughout all this section without loss of generality ( so the ratios @xmath153 and @xmath154 to which we will refer become simply @xmath12 and @xmath155 ) .
in the limit of @xmath156 , the fraction of production intervals interrupted by a deactivation goes to zero , so the resulting probability density function @xmath157 .
this result can also be derived by taking the limit of eq .
as @xmath156 , since @xmath158 .
notice , however , that this limit is slightly different to the limit @xmath159 and fixed @xmath160 ( which gives @xmath161 ) ; the reason is that , in the present case , when taking the limit @xmath162 for a fixed @xmath154 , the ratio @xmath163 obviously remains constant , i.e. the interrupted intervals become less frequent but not longer than uninterrupted ones . on the other hand , taking the limit @xmath159 for a fixed @xmath160 gives @xmath164 .
therefore , even though interrupted intervals become less frequent , they become much longer than uninterrupted intervals as well , hence yielding a bursty behavior .
that is why , when working with the normalized parameters @xmath165 and @xmath166 , @xmath40 must be increased too in order to recover the original burstiness coefficient @xmath167 ( as shown in fig . [
fig : theo_reduction1 ] ) . for intermediate values of the deactivation rate @xmath168
, the fraction of production events interrupted by a deactivation may be significant ; there is , on average , one production event per unit time during the active state , while there are @xmath12 interruptions per unit time .
therefore , approximately one out of every @xmath169 production intervals is interrupted .
in addition , the average time the node remains in the inactive state is @xmath170 , so if @xmath171 , @xmath169 production events buffer during the active state according to @xmath0 and are followed by a long period of inactivity .
hence , the resulting distribution becomes burstier than the original . in the opposite case ( @xmath172 ) ,
the production process becomes poisson regardless of @xmath0 ; since @xmath173 , the probability that the node produces during an active period ( which we will denote by @xmath174 ) is very small , and we can regard the process as a succession of active and inactive states of average length @xmath175 , each of which with an independent production probability @xmath174 . under this approximation , we can easily compute the probability that two consecutive production events take place in a time interval @xmath176 greater than @xmath76 as the probability for the node not to produce in any of @xmath177 intervals , @xmath178 furthermore , @xmath174 can be estimated as well for any @xmath0 .
its exact expression is given by @xmath179 since we are considering the limit of @xmath173 , the probability that a production interval is greater than @xmath180 is approximately one , so @xmath181 in the range of small values of @xmath76 in which the exponential term in the integrand is not close to zero .
hence , we can approximate eq .
as @xmath182 we therefore see that the production process becomes exponentially distributed with mean @xmath183 which is in accordance with eqs .
( [ eq : mu_reg],[eq : mu_ind ] ) .
this result can also be obtained by taking the limit of @xmath121 as @xmath172 assuming @xmath184 and @xmath185 from all the stated above , we can conclude that the network dynamics can be used to control the production of nodes between three regimes of special interest : the endogenous production behavior , a bursty process and a poisson process ( see table [ tab : behaviours ] ) .
furthermore , these results are completely general and independent of the functional form of @xmath0 ( as long as eq
. holds ) .
.[tab : behaviours]different behaviors exhibited by a node following the dynamics presented in the paper . [ cols="^,^,^",options="header " , ]
the iai dynamics is simulated using the continuous - time gillespie algorithm .
in addition , at every deactivation event the deactivated node s production is updated by generating random intervals with distributions @xmath20 for the first one and @xmath0 for the rest of them .
the initial condition for all simulations is @xmath186 ( all nodes active ) , but we only measure nodes @xmath39 for simulation times @xmath187 , since the system reaches a stationary state before that value ( see fig . [
fig : wprevalence ] ) .
we stop simulations when all nodes have produced at least @xmath188 times , so enough statistics are guaranteed . in figs . 2 and 3 in the paper ,
the simulation parameters have been chosen in such a way that a wide heterogeneity of behaviours is exhibited ; in figs .
2a and 3b , every simulation corresponds to a different value of @xmath189 between @xmath190 and @xmath191 , so @xmath6 takes values approximately between @xmath192 and @xmath193 .
since the degree of nodes is between @xmath194 and @xmath195 , @xmath63 is approximately in the intervals @xmath196 for @xmath197 and @xmath198 for @xmath199 . from fig .
1 it can be seen that for @xmath200 , @xmath167 lies about the middle of the interval in which @xmath39 can be varied ; we have set @xmath56 .
a similar criterion has been applied when setting @xmath69 .
likewise , in figs .
2b and 3c @xmath201 is fixed , so @xmath202 and hence @xmath203 .
given that @xmath62 and @xmath42 are varied over 4 orders of magnitude , a wide range of @xmath39 values is obtained .
in this section we present some results not included in the paper . in fig .
[ fig : wprevalence ] we show the quantity @xmath204 derived in section [ sec : dynamics ] .
as expected , this quantity fluctuates around a constant value that can be easily measured from simulations . to see how the network topology affects the induction of burstiness , we have measured @xmath34 for different networks . in fig .
[ fig : heterogeneity ] we show the results for four sf networks with @xmath206 , @xmath207 and @xmath208 , but different exponents . in fig . [
fig : assortativity ] , the results correspond to four sf random graphs with the same degree distribution but different assortativities .
our results show a weak dependence on degree heterogeneity ; indeed , the higher the heterogeneity , the faster the decay of @xmath34 with @xmath5 , although the effect is not really significant .
similarly , degree correlations slightly affect the burstiness of high degree nodes , since it is greater for disassortative networks . yet , assortativity does not play an important role at burstiness induction either .
@xmath34 for sf graphs with different assortativity @xmath209 .
top - left : @xmath210 .
top - right : @xmath211 .
bottom - left : @xmath212 .
bottom - right : @xmath213.,title="fig:",scaledwidth=45.0% ] @xmath34 for sf graphs with different assortativity @xmath209 .
top - left : @xmath210 .
top - right : @xmath211 .
bottom - left : @xmath212 .
bottom - right : @xmath213.,title="fig:",scaledwidth=45.0% ] @xmath34 for sf graphs with different assortativity @xmath209 .
top - left : @xmath210 .
top - right : @xmath211 .
bottom - left : @xmath212 .
bottom - right : @xmath213.,title="fig:",scaledwidth=45.0% ] @xmath34 for sf graphs with different assortativity @xmath209 .
top - left : @xmath210 .
top - right : @xmath211 .
bottom - left : @xmath212 .
bottom - right : @xmath213.,title="fig:",scaledwidth=45.0% ] @xmath34 for sf graphs with different assortativity @xmath209 .
top - left : @xmath210 .
top - right : @xmath211 .
bottom - left : @xmath212 .
bottom - right : @xmath213.,title="fig:",scaledwidth=45.0% ] @xmath34 for sf graphs with different assortativity @xmath209 .
top - left : @xmath210 .
top - right : @xmath211 .
bottom - left : @xmath212 .
bottom - right : @xmath213.,title="fig:",scaledwidth=45.0% ] @xmath34 for sf graphs with different assortativity @xmath209 .
top - left : @xmath210 .
top - right : @xmath211 .
bottom - left : @xmath212 .
bottom - right : @xmath213.,title="fig:",scaledwidth=45.0% ] @xmath34 for sf graphs with different assortativity @xmath209 .
top - left : @xmath210 .
top - right : @xmath211 .
bottom - left : @xmath212 .
bottom - right : @xmath213.,title="fig:",scaledwidth=45.0% ] since the burstiness coefficient @xmath34 alone does not provide any details about the distribution other than the ratio between @xmath214 and @xmath215 , we have measured the cumulative inter - event time distribution function for the burstiest node in the network , as shown in fig .
[ fig : burstiest ] ( left ) .
we have also measured the burstiness of the same distribution by computing the normalized conditional average interval @xmath216 as a function of @xmath217 ( see fig . [
fig : burstiest ] , right ) . for a poisson process
, the conditional average interval @xmath218 is independent of @xmath76 and equal to @xmath219 , while for a regular - like distribution , this should be a decreasing function ( the more time elapsed since the last event , the less time is expected until the next one ) . in our case , we observe exactly the opposite situation , which evidences the counter - intuitive behaviour of bursty distributions ; the more time elapsed since the last event , the more time expected until the next one .
we thus see that the network clearly induces burstiness on the activity of nodes .
left : cumulative inter - event time distribution function of the node with highest @xmath39 in the network ( for @xmath69 ) .
right : normalized conditional average interval @xmath216 as a function of @xmath217 .
we see that it clearly increases above the constant value of 1 expected for a non - bursty poisson process.,title="fig : " ] left : cumulative inter - event time distribution function of the node with highest @xmath39 in the network ( for @xmath69 ) . right :
normalized conditional average interval @xmath216 as a function of @xmath217 .
we see that it clearly increases above the constant value of 1 expected for a non - bursty poisson process.,title="fig : " ] | we prove that complex networks of interactions have the capacity to regulate and buffer unpredictable fluctuations in production events .
we show that non - bursty network - driven activation dynamics can effectively regulate the level of burstiness in the production of nodes , which can be enhanced or reduced .
burstiness can be induced even when the endogenous inter - event time distribution of nodes production is non - bursty .
we found that hubs tend to be less controllable than low degree nodes , which are more susceptible to the networked regulatory effects .
our results have important implications for the analysis and engineering of bursty activity in a range of systems , from telecommunication networks to transcription and translation of genes into proteins in cells .
while human perception tends to appreciate regularity in the happening of events , evidence reveals that in many systems events cluster in bursts , _
i.e. _ , accumulations of a large number of rapidly occurring events during short time intervals separated by silent periods .
bursts have been experimentally and empirically observed in profusion and many models have been proposed to explain and generate them .
for example , traces of human activity are to a great extent imprinted with burstiness .
in particular , bursts and heavy - tails pervade the diversity of human communication channels , from written text @xcite , to the frequency of letter @xcite and e - mail correspondence @xcite , and cell phone calls @xcite . in a different context , key processes in single - cell biology , such as transcription and translation ,
are particularly prone to progress in bursts .
for instance , mrna synthesis proceeds in short but intense outbreaks beginning when coding genes transition from an inactive to an active state @xcite .
these bursts have been suggested to be able to affect the expression of essential genes and even the switching and rhythms of cellular states @xcite .
this may raise questions about how cells are able to maintain function in the face of such unpredictable fluctuations explained by different single gene burst generation mechanisms @xcite , with the hypothesis that those bursts could be buffered by local or bulk degradation mechanisms @xcite .
it is interesting to explore whether large - scale system mechanisms , like the interconnection of genes in regulatory networks of interactions , have the capacity to produce similar buffering effects . in this work ,
we show indeed that endogenous burstiness in the production of nodes can be both enhanced or reduced by stochastic network - driven non - bursty interactions , which activate / inactivate nodes dynamically . we observe a strong anti - correlation of these regulation phenomena with node degree , such that the burstiness of lower degree nodes is always more susceptible to being raised by dynamic network influence , while hubs tend to experience a decreased burstiness in their endogenous production , which helps to maintain a stable bulk level .
conversely , the production sequence of individual nodes of all degrees can present bursts even when their endogenous production pattern follows a non - bursty inter - event time probability distribution , and we observe again a strong anti - correlation of the induced burstiness with node degree .
note that here we are reversing the topical question about the macroscopic collective effects of burstiness on dynamical processes running on complex networks .
the bursty activity of nodes , both in models of opinion formation @xcite and epidemic spreading @xcite , has been observed to induce a slowing down of the progression of the dynamical process .
here , we focus instead on the regulatory effects of networked dynamical interactions on the endogenous bursty activity of individual nodes .
we consider nodes interconnected forming a complex network that can present two states , active and inactive . during the active state
, each node produces events according to a certain endogenous inter - event time probability distribution function @xmath0 .
the activation dynamics of nodes is governed by a propagating process in the network .
we chose an inactive - active - inactive ( iai ) dynamical process akin to the standard susceptible - infected - susceptible epidemic spreading model @xcite , so that active nodes inactivate at rate @xmath1 and inactive ones become active upon contact with active ones at rate @xmath2 per active contact . in all cases ,
these processes are assumed to follow poisson statistics .
an endemic state of active nodes can be sustained whenever the ratio @xmath3 is above a critical value that , in general , depends on network topology @xcite .
once at steady state , the effective activation rate of a node depends on the topological configuration of the network , with a temporal average value on uncorrelated networks given by @xcite @xmath4 here @xmath5 is the node degree and @xmath6 is a temporal average of a degree - weighted prevalence depending on @xmath7 the degree distribution of nodes in the network , the average degree @xmath8 , and the fraction of nodes with degree @xmath9 in the active state @xmath10 .
this temporal average fluctuates around a constant value in the stationary state of the endemic phase ( see appendix [ sec : dynamics ] ) , so that @xmath11 is proportional to degree .
this result allows us to focus the problem on the study of a single node that changes state following poisson processes with rates @xmath11 for activation and @xmath12 for inactivation and that , when active , produces events according to a general distribution @xmath0 with average @xmath13 .
for such a node , and any @xmath0 , we calculated analytically the effective inter - event time probability density function of production events @xmath14 ( see appendix [ sec : phi ] ) . here , we report the result in the laplace space @xmath15 , @xmath16^{2}}{\left < t_p \right >
\left ( s + \lambda_{i } \right)^{2 } } } { \frac{s + \lambda_{eff}(k)}{\lambda_{eff}(k ) \lambda_{i } } - \frac{\left [ 1 - \frac{1 - \hat{\psi } ( s + \lambda_{i})}{\left < t_p \right > \left ( s + \lambda_{i } \right ) } \right]}{s + \lambda_{i}}}\ ] ] where we have used that the probability density for the interval between the activation time and the first production event is given by @xmath17 , where @xmath18 is the survival function .
the rationale for this choice is that the activation event can be regarded as taking place between two consecutive production events .
since we do not have any other information , we assume that the time interval between those two events is greater than the time elapsed between the activation event and the first production . as a consequence ,
its probability density must be proportional to @xmath19 , and the denominator simply normalises the distribution @xcite .
hence , effective inter - event times of production events are ruled by the interplay of three key factors : the endogenous production statistics @xmath0 , the first production event statistics @xmath20 , and the activation dynamics through the effective activation rate that depends linearly on a node s degree @xmath11 . to measure the burstiness of production
, we use the burstiness coefficient @xmath21 as defined in @xcite , where @xmath22 is the coefficient of variation @xmath23 and @xmath24 and @xmath25 are the mean and standard deviation of the time between consecutive production events . with this definition
, @xmath26 corresponds to a strongly bursty production , @xmath27 to a neutral one following a poisson statistics , and @xmath28 to a periodic signal . as an example , we now particularize to the case of a weibull distribution for production events @xmath29 with scale parameter @xmath30 controlling the spread of the distribution , and shape parameter @xmath31 implying a heterogeneous non - poisonian distribution such that all nodes have an endogenous production of events that clusterize in time . the endogenous burstiness for a node continuously producing according to eq .
is @xmath32 , independent of @xmath33 .
the effective value of the burstiness of a node with interrupted production due to the network - driven activation / inactivation dynamics depends on its degree , @xmath34 . to compute it , we need to calculate the degree - dependent average and standard deviation of production inter - event times , which can be easily evaluated through simple derivatives of eq .
( [ eq : phi ] ) evaluated at @xmath35 ( see appendix [ sec : phi_weibull ] ) .
the degree dependent coefficient of variation reads @xmath36 where we have redefined @xmath37 and @xmath38 . as a validation , we display in fig . [
fig : theo_reduction1]a the analytical result for @xmath39 based on eq .
along with the simulation of a single node whose state changes following poisson processes with rates @xmath40 for activation and @xmath41 for inactivation , and that intrinsically produces events according to eq . with rate @xmath42 .
the agreement between the analytical surface and the simulation points is excellent .
both results prove that the level of endogenous burstiness for continuous production @xmath43 can be both increased and decreased , as shown also in the projection of the analytical surface on the @xmath44 plane in fig .
[ fig : theo_reduction1]b .
the effective burstiness ranges always between zero burstiness ( exponential inter - event time distribution ) in the limit @xmath45 and the maximum of @xmath46 when @xmath47 . for high values of @xmath40 ,
@xmath39 is dominated by the interplay of the endogenous production statistics @xmath0 and the first production event statistics @xmath20 .
the range of @xmath42 values associated to effective burstiness around @xmath48 widens with increasing @xmath40 and the endogenous production level is recovered only when both @xmath42 and @xmath40 are increased simultaneously .
in contrast , the effective burstiness raises for low values of @xmath40 due to the effect of increased inactivation periods , such that for each @xmath40 the level of effective burstiness increases with @xmath42 .
these results are qualitatively the same for any distribution @xmath0 ( see appendix [ sec : limits ] for a detailed analysis ) .
* regulation of burstiness . * * a*. comparison between simulation data ( dots ) and the analytical calculation ( solid surface and line ) of the effective burstiness @xmath49 , @xmath50 in the blue region and @xmath51 in the green one .
* b*. projection of burstiness surface in * a * in the @xmath52 plane .
colors in the contour plot code for different values of @xmath39 . ] in line with these arguments , and from the proportional dependence of @xmath11 with degree eq . , one can expect that the effective burstiness of higher degree nodes is reduced while lower degree nodes can be modulated . to confirm this
, we measured @xmath34 from simulations on a network with @xmath53 nodes ( simulation details in appendix [ sec : simulation ] ) .
although our results are valid in the variety of uncorrelated network topologies , we used a scale - free network with characteristic exponent @xmath54 . in fig .
[ fig : theo_reduction2]a , we show simulation results of @xmath34 for different values of @xmath55 . for all activation rates , a strong inverse dependence of effective burstiness on degree can indeed be observed . for high degree nodes
, @xmath34 is noticeably below @xmath48 .
this is due to the fact that hubs do not remain in the inactive state for much time since they are usually connected to many active nodes that constantly reactivate them . on the other hand
, @xmath55 still allows to control the duration of inactive periods for low degree nodes . for low values of @xmath55 ( but above the minimum required to sustain the activity ) , low degree nodes remain a long time in the inactive state , which raises their effective burstiness well above @xmath48 , while for high values of @xmath55 they behave as high degree nodes approaching a minimum value of @xmath34 independent of @xmath5 , a basal level below @xmath48 .
therefore , we conclude that the bursty production of nodes can be regulated by network - driven stochastic activation dynamics that splits the originally uniform endogenous burstiness in a range of values anti - correlated with degree .
fixing @xmath55 , the production shape parameter @xmath42 can be varied to regulate the effective level of burstiness , fig .
[ fig : theo_reduction2]b , with similar qualitative behavior . in both cases ,
the disparity of effective @xmath34 values reaches a maximum at some intermediate parameter levels .
however , @xmath42 does not have any limitation to be increased ( in our computational experiments it is varied over four orders of magnitude ) so that high - degree nodes can increase their effective burstiness above the endogenous level .
* network - driven regulation of burstiness .
* * a*. @xmath34 for different values of the activation rate @xmath55 and fixed @xmath56 , chosen so that a wide heterogeneity of behaviors can be observed and @xmath48 lies about the middle of the interval in which @xmath34 can be varied . * b*. @xmath34 for fixed @xmath57 and different values of the production scale parameter @xmath42 . ] finally , we also prove that network - driven activation can induce burstiness even when the endogenous production of nodes is not bursty .
we let nodes in the active state to produce events with an exponential inter - event time probability distribution function of rate @xmath58 , @xmath59 , meaning that the endogenous production of nodes is not bursty , @xmath60 . again , calculating the first two moments of eq .
( see appendix [ sec : phi_poisson ] ) , we obtain @xmath61 where @xmath62 and @xmath63 are the original values rescaled by @xmath1 .
this analytical expression matches almost perfectly with numerical simulations of a single node whose state changes following poisson processes with rates @xmath40 for activation and @xmath41 for inactivation , and that intrinsically produces events according to an exponential inter - event time probability distribution with rate @xmath62 , fig .
[ fig : theo_induction]a . for an exponential distribution ,
@xmath64 so that @xmath65 has no role as expected for a poisson process . hence , the effective burstiness can only be equal or above the endogenous value @xmath60 .
this induced burstiness is explained as a consequence of the fact that the node may remain in the inactive state for a period of time considerably longer than the average endogenous inter - event time @xmath66 @xcite .
* network - driven induction of burstiness . * in all graphs , solid surface and lines correspond to analytical results and dots to simulations . * a * comparison between simulation data and the analytical calculation of the effective burstiness @xmath67 , @xmath68 . * b*. @xmath34 for fixed @xmath69 and different values of the activation rate @xmath55 . * c*. @xmath34 for fixed @xmath57 and different values of the production rate @xmath62 . ]
we also simulated exponential production in the same scale - free network with @xmath53 nodes and characteristic exponent @xmath54 used in the regulation study .
we used different values of @xmath55 with @xmath69 , fig .
[ fig : theo_induction]b , and different values of @xmath62 over four orders of magnitude with @xmath57 , fig .
[ fig : theo_induction]c . as for bursty production
, we find again that @xmath34 is always a decreasing function of @xmath5 ( but always above @xmath60 ) . due to their sustained activity , hubs are less susceptible to display induced burstiness production .
in contrast , nodes with low degree stay in the inactive state for longer periods and are more prone to present a raised effective burstiness .
qualitatively , both parameters @xmath62 and @xmath55 can control similarly the induction of bustiness , affecting all nodes more efficiently for higher values of @xmath62 and lower values of @xmath55 . again , the disparity of induced burstiness values reaches a maximum at some intermediate parameter levels
. different dynamical processes occurring on a network can interplay in unexpected ways .
we have seen here that the time series of production events of interacting nodes are affected by network - driven activation / inactivation dynamics . in a broad range of parameters , hubs in a network
tend to decrease burstiness in their endogenous pattern of production , hence helping to maintain a stable bulk level , at the expenses of low degree nodes being more pliable to the regulatory effect of the network that tends to amplify their fluctuations . at the same time , lower degree nodes are more prone to burstiness induction .
these results indicate that a heterogenous network structure protects the functioning of some nodes , the hubs , making low degree nodes better targets for engineering actions to produce local modifications of production without critically affecting the behavior of the whole system . taken together
, our findings suggest that heterogeneous network interconnectivity may be a strategy in itself developed to protect complex systems against unpredictable functional fluctuations .
however , further research should be conducted to determine the effects of different activation / inactivation dynamics on node s burstiness .
this work was supported by a james s. mcdonnell foundation scholar award in complex systems ; the icrea academia prize , funded by the _ generalitat de catalunya _ ; mineco projects no .
fis2013 - 47282-c2 - 1 , fis2010 - 21781-c02 - 02 , and bfu2010 - 21847-c02 - 02 ; _ generalitat de catalunya _ grant no . 2014sgr608 ; micinn ramn y cajal program , and ec fet - proactive project multiplex ( grant 317532 ) . |
mass - loss is one of the most important and uncertain parameters in the evolution of a massive star .
there are several methods for determining mass - loss rates ( e.g. , via radio continuum flux , radiative transfer and polarization variation in close binaries ) .
it has become increasingly recognized ( e.g. , @xcite ) that smooth - wind models , based on density - squared diagnostics , overestimate clumped wind mass - loss rates by up to an order of magnitude .
another important parameter for massive stars is the wind acceleration . for most radiatively - driven stellar - wind models ,
a velocity law @xmath3 with @xmath4 is assumed . here ,
@xmath5 and @xmath6 are the terminal wind - velocity and stellar radius , respectively . in this model , usually the initial wind velocity @xmath7
is neglected , since it is thought to be @xmath81% of @xmath5 and the wind - collision x - rays are formed relatively far out in the wind . on the other hand ,
some optical observations have revealed a high value of @xmath9 for wolf - rayet ( w - r ) stars ( e.g. , 20 @xmath10 ; @xcite , which lead to values @xmath11 1 for normal w - r radii ) .
lccllllrr obs . &
observation & observation & @xmath12&@xmath13 & @xmath14&@xmath15&orbital & _ d _ + & start [ ut ] & end [ ut ] & [ ks ] & [ cps]&[cps]&[cps]&phase & [ au ] + a & 2008 - 04 - 09 05:33:13 & 2008 - 04 - 09 17:20:18 & 21.6 & [email protected]&[email protected]&[email protected] & 2.904 & 13.79 + b & 2008 - 12 - 12 10:27:51 & 2008 - 12 - 13 13:15:20 & 52.8 & [email protected]&[email protected]&[email protected] & 2.989 & 3.12 + c & 2009 - 01 - 04 08:36:00 & 2009 - 01 - 05 10:12:18 & 47.3 & [email protected]&[email protected]&[email protected] & 2.997 & 1.73 + d & 2009 - 01 - 13 12:59:45 & 2009 - 01 - 15 12:00:13 & 89.4 & [email protected]&[email protected]&[email protected] & 3.000 & 1.53 + ( 85mm,85mm)figure1.eps while the mass - loss rate and the acceleration parameter @xmath9 have been measured using the radio / ir continuum flux or line spectral analysis at optical / ir wavelengths , x - ray wavelength could be an independent window to approach these parameters . colliding
wind binary is a good target , having variable x - ray spectra with orbital phase .
the x - ray is emitted from the wind - shocked region , which is strongly dependent on the ram - pressure balance between the two hypersonic winds .
the shocked plasmas , which have temperatures of 10@xmath1710@xmath18 k , are frequently observed , and the high absorption column of 10@xmath1910@xmath20 h @xmath21 are reported ( cf .
@xcite ) . the temperature should reflect on the wind velocity , and the absorption column indicates the dense w - r wind @xcite , i.e. mass - loss rate of the w - r star ( cf .
the x - ray luminosity is highly dependent on the separation between the stars of the binary , the mass - loss rates , and wind velocities ( @xcite ; @xcite ) . if we know the orbital parameters of the binary , the x - ray luminosity at each orbital phase should depend on the mass - loss rates and wind - acceleration parameters .
wr 140 ( hd 193793 ) is considered as the textbook example of a colliding wind binary .
the star has been classified as a [email protected] binary system whose stellar masses are @xmath22 = 16 @xmath23 and @xmath24 = 41 @xmath23 by the optical spectroscopic monitoring @xcite .
its orbit and distance have been well determined with @xmath25 = 2896.35 days , _ i _
= 119.6@xmath26 , _ e _ = 0.8964 and _ d _ = 1.67 kpc by detailed ir astrometric study @xcite .
radio non - thermal ( synchrotron ) emission from wr 140 was reported that may arise from a wind - wind collision zone ( e.g. , @xcite ; @xcite ; @xcite ) .
@xcite proposed a radio , x - ray and @xmath27-ray non - thermal emission model . in their model
, there are some cases where inverse compton emission dominates the high energy x - ray emission .
@xcite predicted that _ suzaku _ would detect such non - thermal x - ray emission .
wr 140 is a bright x - ray source ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) . during the previous periastron passage , _
rossi - xte _ showed a drop in x - ray flux before periastron @xcite .
we pursued this via _ suzaku _ monitoring during the following periastron passage in 2009 .
these observations cover the energy range 0.120 kev .
the _ suzaku _ x - ray observatory @xcite is equipped with two kinds of instruments ; the xrt ( x - ray telescope , @xcite ) @xmath0 xis ( x - ray imaging spectrometer : @xcite ) system , sensitive to x - rays between 0.312 kev , and the hxd ( hard x - ray detector : @xcite ) sensitive to x - rays above 10 kev .
_ suzaku _ observed wr 140 four times around periastron passage in 2009 january .
sequence numbers of the data are 403030010 , 403031010 , 403032010 and 403033010 . logs of these four observations , labeled as a , b , c and d , are summarized in table [ tb : t2 - 1 ] and figure [ fg : f2 - 1 ] .
the total exposure of these observation was @xmath8210 ks .
the xis is composed of four x - ray ccd ( xis03 ) arrays with 1024@xmath281024 pixel formats , each of which is mounted at the focal plane of an individual xrt .
the xrt+xis system covers a field of view of @xmath818@xmath2818 .
the xis1 has a back - side illuminated ( bi ) ccd chip while the remaining active sensors ( xis0 and xis3 ) have front - side illuminated ( fi ) chips . the bi and fi chips are superior to each other in the soft and hard band responses respectively . during the observations ,
the xiss were operated in the normal clocking mode with the default frame time ( 8 s ) .
wr 140 was placed at the hxd nominal position .
the hxd is a non - imaging instrument that consists of a 4@xmath284 array of detectors ( well units ) and 20 surrounding crystal scintillators for active shielding .
each well unit consists of four si pin diodes ( pin ) , sensitive to x - rays between 1070 kev , and four gso / bgo phoswitch counters ( gso ) for detecting photons with energies in the range 40600 kev .
we used only pin data to constrain the emission above 10 kev due to the high background of the gso instrument .
( 170mm,85mm)figure2.eps ( 170mm,170mm)figure3.eps we reduced the data using heasoft version 6.16.0 and calibration versions ( caldbver ) xis20090402 , xrt20080709 and hxd20090402 .
we screened out xis and hxd events obtained during ( 1 ) the south atlantic anomaly ( saa ) passage , ( 2 ) the low geomagnetic cut - off rigidity ( 6 gv for the xis and 8 gv for the hxd ) , ( 3 ) low elevation angles from the earth rim ( @xmath2910@xmath30 for the xis and @xmath295@xmath30 for the hxd ) and the sun - lit earth rim ( @xmath2920@xmath30 ) , and ( 4 ) telemetry saturation .
we used xspec version 12.5.1 for spectral analysis .
( 160mm,70mm)figure4.eps ( 140mm,60mm)figure5.eps figure [ fg : f1 ] shows the xis images in the soft ( 0.41.6 kev ) band and the hard ( 710 kev ) band . in all the observations , the xiss detected an x - ray source at the position of wr 140 . in order to capture the spectral characteristics of wr 140
, we extracted the source events from a circle with 4 arcmin radius centered on the x - ray peak , which gives a maximum signal - to - noise ratio in the 0.410 kev band .
we extracted background events from the entire xis field of view , excluding the region within 5 arcmin from the x - ray peak and calibration source regions .
wr 140 did not show any significant x - ray variation during each observation , so that we generated time - averaged spectra for each observation .
we generated xis detector responses ( rmf ) and xrt effective area tables ( arf ) with xisrmfgen and xissimarfgen @xcite and averaged two fi spectra and responses with the heasoft mathpha , addrmf and addarf .
the fi spectra of observations a and b , c , d were grouped to a minimum of 100 and 200 counts per bin with ftool grppha , respectively .
similarly , the bi spectra of observations a , b , c and d were grouped to a minimum of 100 , 200 , 150 and 200 counts per bin , respectively . in spectral fits
, we ignored xis bins around the si k - edge between 1.82.0 kev because of the calibration uncertainty .
we noticed that the o - ly@xmath31 , ne - ly@xmath31 and mg - ly@xmath31 lines in the bi spectra of observations a and b deviate significantly from those lines in the astrophysical plasma emission database ( aped ) .
the fi spectra did not show such deviation , and are consistent within @xmath85 ev with the values in the database .
this is suggesting that the energy gain of the bi chip is not calibrated sufficiently .
center energies of these lines , measured with fits by gaussian functions , deviate by @xmath325@xmath163 ev and @xmath3220@xmath163 ev for observations a and b , respectively .
we therefore offset those spectra by the best - fit values .
we did not make gain offsets for the spectra of observations c and d , since they did not show those emission lines .
we subtracted from the pipeline - processed hxd / pin spectra non - x - ray background ( nxb ) simulated with the tuned nxb model ( lcfitdt , @xcite ) and the typical cosmic x - ray background ( cxb ) measured with the _ heao-1 _
we generated the pin detector response with hxdarfgen .
the pin spectra also were binned with grppha .
the xis spectra of wr 140 ( figure [ fg : f3 - 2 ] ) show emission lines from highly ionized ions of neon , magnesium , silicon , sulfur and iron .
earlier observations of wr 140 ( _ ginga _ : @xcite , _ asca _ : @xcite , _ chandra _ : @xcite , _ xmm - newton _ : @xcite ) also found these lines in the spectra , suggesting optically thin plasma thermalized by wind - wind collision .
the low energy cut - off the signature of absorption by neutral matter or weakly ionized plasma drastically increased from observation a to d. high density and/or large atomic number are needed for large absorption . in wr 140 ,
the mass - loss rate of the o - star is an order of magnitude smaller than that of the w - r star .
in addition , the w - r wind have higher metal abundances than the o - star wind .
therefore , it is well known that the absorption by the w - r star is dominant ( e.g. , @xcite ) . the spectral variation like this
was reported by @xcite .
we call this dominant emission `` variable component '' ( see @xmath33 [ subsec : variable_comp ] ) . during observations c and d
when soft x - rays from the wind - wind collision are suppressed by absorption from the wind of the w - r star , we detected another emission component below 2 kev .
this component did not change apparently between these observations .
we call this soft emission `` cool component '' ( see @xmath33 [ subsec : cool_comp ] ) .
the hxd / pin detected significant x - ray emission in the 1550 kev range .
the flux seems to change from observation b to d ( see figure [ fg : f3 - 2 ] ) , which seems to correlate with the variation below 10 kev .
however , we should pay attention to other hard x - ray source igr j20216 + 4359 , which is located at 17.35 arcmin away from wr 140 @xcite .
the location is shown in figure [ fg : f1 ] & [ fg:3 - 3 ] .
because of a potential for the contamination from igr j20216 + 4359 in the field of view of pin / hxd , xis and pin spectra are analyzed separately .
we found the cool component in the spectra of observations c and d ( see @xmath33 [ subsec : spectral_chara ] ) . in this subsection , we report the analysis results of the cool component .
first , we check the spatial structure of the cool component .
then , we compare the spectral features of diffuse emission around wr 140 and the cool component .
finally , we fit the spectra of the cool component .
wr 140 is located in the cygnus region that is pervaded by diffuse x - ray emission from a large ring - like structure , whose diameter is @xmath34 .
this diffuse structure is known as the cygnus superbubble ( csb ) .
we then figure out the spatial structure of the cool component before the spectral analysis .
figure [ fg:3 - 3 ] shows _ suzaku _ images in the 0.41.6 kev and 710 kev band , combined from observations c and d. in the soft band , the diffuse emission with some degree of inhomogeneity is seen .
the north and east areas were brighter than the south area by a factor of two in count rate .
some transient sources were identified in observations c and d. ( 70mm,70mm)figure6.eps figure [ fg:3 - 3 - 2 ] shows the radial profiles of the _ suzaku _ xis images for observation d. the images are taken with xis-0 , 1 , and 3 , and summed up . in figure
[ fg:3 - 3 - 2 ] , the dashed line in each panel shows the background level determined by the fitting of the data between 200600 arcsec form wr 140 with a constant model .
the bottom panel in figure [ fg:3 - 3 - 2 ] shows the ratio of the soft band count rate and the hard band count rate , after background subtraction .
the ratio is nearly flat along the radial axis .
the point spread functions of the xiss do not significantly show the energy dependency@xcite .
since the emission from the point source wr 140 dominated the xis count rate in the hard band , the constant ratio indicates that the cool component is also point - like and associated with wr 140 . in this subsubsection
, we compare the spectral features of the background diffuse emission around wr 140 and the cool component . since wr 140 lies in the north area of the diffuse structure , we adopted the north and east regions of wr 140 as the background region for observations c and d. we excluded the transient sources from the background region , as shown in figure [ fg:3 - 3 ] . in order to improve the signal - to - noise ratio of the spectra in the 0.41.6 kev band , we extracted source events within 2 arcmin from the x - ray peak .
we co - added the xis-0 , 1 and 3 spectra for observations c and d and ignored spectral bins above 1.6 kev .
figure [ fg:3 - 3 - 3 ] shows the area - normalized spectra of the source and the background region .
nxb subtraction were not made for both of the spectra .
the spectra of the background region was fairly complex , which had o - ly@xmath31 , ne - ly@xmath31 and mg - ly@xmath31 lines and the structures of he - like o and ne .
on the other hand , the source spectrum shows large excess from the background .
especially , the structures , which are shown with arrows in figure [ fg:3 - 3 - 3 ] , are unique to the source spectrum , and can not be made by an inappropriate background subtraction .
at least the 1.2 kev peaked structure in these unique structures could not been reproduced by collisional equilibrium ( ce ) plasma model .
the detail is discussed in @xmath33 [ subsubsec : cool_fit ] .
l l c + components & parameters & + absorption & @xmath35 ( 10@xmath36 @xmath37)&3.8@xmath38 + vrnei & @xmath39(kev)&0.021@xmath40 + & @xmath41(kev)&@xmath420.58 + & n ( solar ) & @xmath2947 + & o ( solar ) & 9.1@xmath43 + & ne ( solar ) & 17.1@xmath44 + & mg ( solar ) & 3.4@xmath45 + & @xmath46(10@xmath47 s @xmath48)&2.3@xmath49 + & norm ( 10@xmath50 ) & 7.5@xmath51 + & @xmath52 ( 10@xmath53 @xmath54 @xmath37 @xmath55 ) & 1.1 + & @xmath56 ( 10@xmath57 @xmath54 @xmath55 ) & 2.0 + & @xmath58/dof&98/83 + ( 160mm,70mm)figure7.eps the observed surface brightness of background region was @xmath59 ( 0.52 kev ) .
this value was obtained from a fit of the nxb subtracted bi spectrum for observation d by a thin thermal plasma model ( apec : @xcite ) .
we adopted the ancillary response file for a circular flat field sky with a 20 arcmin radius .
this observed surface brightness is not inconsistent with the reported flux within an order of magnitude @xcite . in this subsubsection , we fit the spectra of the cool component in the 0.41.6 kev band with two models .
first , we fitted the fi and bi background subtracted spectra using a model of a one - temperature ce plasma ( vapec ) emission suffering absorption by cold matter with the interstellar medium ( ism ) abundances ( tbabs : @xcite ) .
as for the ce plasma , the elemental abundance were fixed at the best - fit values for the variable component ( see @xmath33 [ subsubsec : globalfitab ] and table [ tb : t3 - 2 ] ) except for n , ne and mg .
the fitting result ( figure [ fg:3 - 4 ] _ left _ ) yields a plasma temperature of @xmath60 kev and a column density @xmath61 10@xmath36 @xmath21 ( @xmath58/dof @xmath62 163/50 ) .
( 85mm,85mm)figure8.eps this model still leaves line - shaped residuals at 1.21@xmath63 kev . in the emission line database aped
, the 1.21 kev line corresponds to ne - ly@xmath9 line .
however , in the ce plasma , the ne - ly@xmath9 line flux must be weaker than the ne - ly@xmath31 line flux at 1.02 kev . in order to get the line flux ratio of ne - ly@xmath31 to ly@xmath9
, we fit the spectra using the gaussian lines model with an absorption . the absorption fixed at @xmath647 @xmath28 10@xmath65 derived from the fitting result of the ce plasma model .
figure [ fg : f6 - 3 ] shows the confidence contour map of the normalizations of two gaussian lines ( ne - ly@xmath31 and ly@xmath9 ) and indicates that the residual at 1.2 kev is significant at the 3@xmath66 level . the best - fit ratio of ly@xmath9/ly@xmath31 is about unity that is much larger than the ratio for the ce plasma , @xmath67 . the high ly@xmath9/ly@xmath31 ratio is often obtained from a plasma in a recombining phase . from the above consideration , there is a possibility that the residuals at 0.5 , 0.73 , 0.87 , 1.21 and 1.37 kev result from radiative recombination continuum ( rrc ) edges of c@xmath68 , o@xmath69 , o@xmath70 , ne@xmath71 , and ne@xmath72 , respectively . we therefore introduce non - equilibrium recombining collisional plasma model ( vrnei ) .
this model is based on the assumption that the plasma changes from the collisional equilibrium state , with the initial temperature @xmath41 , to the non - equilibrium recombining state , with the temperature @xmath39 .
the elemental abundances except for n , o , ne and mg were fixed at the best - fit values for the colliding wind emission ( see @xmath33 [ subsubsec : globalfitab ] and table [ tb : t3 - 2 ] ) . the reduced @xmath58 was significantly improved .
the best - fit results are shown in table [ tb : t3 - 4 ] .
the temperature @xmath39 is as low as @xmath821 ev .
l c + element & abundance ratio + he&1 + c&0.4 + n&0 + o&7.2 @xmath28 10@xmath73 + ne&(6.0@xmath74 ) @xmath28 10@xmath75 + mg&(5.4@xmath76 ) @xmath28 10@xmath77 + si&5.9 @xmath28 10@xmath77 + s&2.4 @xmath28 10@xmath77 + ar&5.3 @xmath28 10@xmath50 + ca&3.2 @xmath28 10@xmath50 + fe&(4.16@xmath78 ) @xmath28 10@xmath77 + ni&(1.73@xmath79 ) @xmath28 10@xmath50 + ( 160mm,160mm)figure9.eps l l c c c c + & observation & a & b & c & d + + & orbital phase&2.904 & 2.989 & 2.997 & 3.000 + + + & @xmath80 ( 10@xmath36 @xmath37 ) & 8.51@xmath81&@xmath62 a&@xmath62 a ( fixed ) & @xmath62 a ( fixed ) + + & @xmath82 ( 10@xmath83 @xmath37 ) & &4.35@xmath84&& + & @xmath85 ( @xmath86 ) & 0.645@xmath87&0.3515@xmath88&& + & @xmath46(10@xmath89 s @xmath48)&8.2@xmath90&4.1@xmath91&& + & norm ( 10@xmath75 ) & 9.97@xmath92&14.4@xmath93&& + & @xmath52 ( 10@xmath94 @xmath54 @xmath37 @xmath55 ) & 1.28&0.45&& + & @xmath56 ( 10@xmath95 @xmath54 @xmath55)&2.08 & 4.67 & & + & @xmath96 ( 10@xmath95 @xmath54 @xmath55)&3.84 & 10.6 & &
+ + & @xmath97 ( 10@xmath98 @xmath37 ) & @xmath99 0.007&1.68@xmath100&12.6@xmath74&23.4@xmath101 + & @xmath102 ( @xmath86 ) & 3.29@xmath103&3.48@xmath104&2.97@xmath105&3.03@xmath106 + & @xmath46(10@xmath89 s @xmath48)&8.2&4.1&2.7@xmath107&@xmath108 2.5 + & norm ( 10@xmath73)&1.27@xmath63&3.45@xmath109&3.56@xmath110&1.89@xmath111 + & @xmath52 ( 10@xmath94 @xmath54 @xmath37 @xmath55 ) & 2.85&5.63&2.58&1.01 + & @xmath56 ( 10@xmath95 @xmath54 @xmath55 ) & 1.67 & 5.10 & 5.78 & 2.51 + & @xmath96 ( 10@xmath95 @xmath54 @xmath55 ) & 2.59 & 9.08 & 11.7 & 4.13 + &
@xmath58/dof & & 556/472&297/253 + the emission from variable component constitutes a considerable fraction of the x - ray emission from wr 140 ( see figure [ fg : f3 - 2 ] ) . for the spectral analysis
, we then adopted the source and background regions used in @xmath33 [ subsec : spectral_chara ] .
as for observations c and d , we ignored the soft band below @xmath82 kev because there is the cool component ( see figure [ fg : f3 - 3 ] ) . in this subsubsection , we made the fi and bi spectra of observations a and b in the 0.510 kev band .
we executed simultaneous fitting of the fi and bi spectra of observations a and b. we first tried to fit these spectra using a one - temperature plasma emission model with either a single absorption component or a combination of absorption components with two different optical depths .
we next tried a two - temperature plasma emission model with a single common absorption .
both the attempts did not produce statistically acceptable solutions .
this is perhaps because one - temperature plasma models do not reproduce emission lines in the soft band around 1 kev and iron k line at @xmath86.7 kev simultaneously , and those spectra , in particular in observation b , are too flat to be reproduced with a single absorption component .
we therefore fit the xis spectra with a two - temperature ( warm and hot ) plasma model with independent absorption components ( varabs ) [ subsubsec : fekprofile ] , had little effect on the overall fit , so that we did not change the red - shift parameter in the spectral model . ] . for the emission components , we first used the ce plasma emission code , vapec @xcite in which we can allow to vary each metal abundance .
next , we used the plane - parallel shock code , vpshock . in both of the fits
, we fixed the c / he abundance ratio at 0.4 by number @xcite and the other elements except for ne , mg , fe and ni at the best - fit abundances in @xcite .
the latter fit gave slightly better reduced @xmath58 values ( 2tvapec : @xmath58/dof @xmath81.71 , 2tvpshock : @xmath58/dof @xmath81.61 ) .
the top panels in figure [ fg : f3 - 3 ] show the best - fit 2tvpshock models and the spectra .
we use the vpshock model in the following analysis .
the x - ray emission should be absorbed both in the w - r wind and ism .
we therefore assume two absorption components .
one is the absorption component ( varabs ) for the w - r wind , whose elemental abundances are tied to those of the plasma ( vpshock ) model .
another is an interstellar absorption component , with elemental abundances fixed at ism abundances ( tbabs : @xcite ) .
we also assumed that the warm component in observation a suffers only interstellar absorption .
we tied elemental abundances of the both plasmas for observations a and b , since the elemental abundance of the plasma should not change between observations a and b. we also tied the ionization parameters @xmath112 for both hot and warm components .
the best - fit results are given in tables [ tb : t3 - 2 ] and [ tb : t3 - 1 ] .
the best - fit model for the spectrum of observation a gives an absorption column of @xmath1138.51 @xmath28 10@xmath65 .
the observation a is correspond to orbital phase @xmath114 , in the definition of @xcite .
the column density is close to that obtained with an earlier observation with _
( @xmath648.0 @xmath28 10@xmath65 at orbital phase @xmath115 ; @xcite ) . the interstellar extinction toward wr 140 has been measured at the optical wavelength as @xmath116 2.95 mag @xcite .
based on the @xmath117 , the expected ism absorption column ( @xmath118 ) is either 6.5 @xmath28 10@xmath36 ( using the correlation of @xcite ) or 4.6 @xmath28 10@xmath65(using the correlation of @xcite ) . in these correlations , we assumed that the gas - to - dust ratio of wr 140 is the same as that in ism .
the observed absorption at the phase a differs from the expected absorptions only by a factor of two .
the absorption appearing in the data of observation a may be mainly made by the ism toward wr 140 . for simplicity ,
we regard the x - ray absorption in the spectrum of observation a as the interstellar absorption hereafter .
no conclusion in this paper changes with this simplification .
we fit the spectra of observations c and d independently .
we executed simultaneous fitting of fi and bi spectra . since the soft band flux below @xmath84 kev
is strongly suppressed by circumstellar absorption , we did not include the warm component in the fits to either the spectrum of observation c or that of observation d. the elemental abundance were fixed at the best - fit values of observations a and b ( see table [ tb : t3 - 2 ] ) .
the bottom panels in figure [ fg : f3 - 3 ] and table [ tb : t3 - 1 ] show the best - fit results .
l lcccc + obs . &
i.d.&a & b & c & d + + + @xmath119&&3.0@xmath74&2.98@xmath120&3.21@xmath121&2.98@xmath122 + norm&&0.04@xmath63&0.12@xmath123&0.17@xmath124&0.06@xmath63 + + @xmath125(@xmath126)&fe@xmath127 k@xmath31&6697@xmath128 @xmath129&6678@xmath130 @xmath131&6668@xmath132 @xmath131&6661@xmath133 @xmath131 + _
v_(10@xmath134)&&@xmath328.6@xmath135 @xmath136&@xmath320.9@xmath137 @xmath136 & @xmath04.5@xmath138 @xmath139&@xmath0@xmath140 @xmath136 + @xmath141 ( @xmath126)&&@xmath295&@xmath2919&@xmath2915&@xmath2931 + @xmath142&&5.8@xmath143&18.1@xmath143&16.1@xmath143&9.4@xmath101 + @xmath144 ( @xmath126)&fe@xmath145 ly@xmath31&6973&6973&6973&6973 + @xmath146&&@xmath290.5&1.0@xmath147&0.51@xmath147&@xmath290.2 + @xmath148 ( @xmath126)&ni@xmath149 k@xmath31&7797&7797&7797&7797 + @xmath150&&@xmath290.9&1.3@xmath143&1.1@xmath151&0.9@xmath152 + @xmath153 ( @xmath126)&fe@xmath127 k@xmath9&7897&7897&7897&7897 + @xmath154&&@xmath290.8&@xmath290.9&0.8@xmath151&@xmath290.3 + + @xmath58/dof&&128/96&278/238&192/158&151/159 + mn k@xmath31(@xmath126)&&5904@xmath155&5889@xmath156&5896@xmath157&5898@xmath158 + ( 85mm,85mm)figure10.eps the xis spectra showed emission lines of fe@xmath127 k@xmath31 ( @xmath86.7 kev ) , fe@xmath145 ly@xmath31 ( @xmath86.9 kev ) , ni@xmath149 k@xmath31 ( @xmath87.8 kev ) and fe@xmath127 k@xmath9 ( @xmath87.9 kev ) . in order to investigate the variation of their line profiles , we measured the center energies , widths and fluxes of these lines using the fi data which has better gain calibration and higher sensitivity in the hard energy band .
we fitted the spectra in the 59 kev band with a continuum model and four gaussian models for the lines .
for the continuum emission , we tried the two different models , thermal bremsstrahlung and power - law .
both fits gave consistent results .
in fact , the best - fit parameters of the lines for two different continuum models are found to be within the margin of error .
we therefore show only the best - fit parameters of a power - law model for the continuum and four gaussian models for the lines .
we fixed the line - center energies , except for that of fe@xmath127 k@xmath31 line , at the rest frame values , due to poor photon statistics .
figure [ fg : fe ] and table [ tb : t3 - 6 ] show the best - fit results .
the advantage of suzaku - xis is that the energy - scale is calibrated very well with the accuracy of roughly 0.1% ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) .
therefore , we adopted the systematic errors of 0.1% ( 7 ev ) in our result , as are shown in table [ tb : t3 - 6 ] .
( 85mm,85mm)figure11.eps as the binary system approached periastron , the central energy of fe@xmath127 k@xmath31 line gradually shifted from 6,697@xmath159 ev ( observation a ) to 6,661@xmath133 ev ( observation d ) . in order to measure the shift of the central energy of the fe@xmath127 k@xmath31 line , we determined the line center energy in the rest frame for the ce plasma model ( @xmath1603 kev ) . here
, we used the fakeit of xspec to take account of the influences of not only dominant iron lines but also other weak lines derived from other elements .
the line center energy was @xmath86,676 ev .
we have to pay attention to the fact that the line center energy could shift with variations of the flux ratios between the resonance , intercombination and forbidden lines , which are caused by non - equilibrium ionization .
the ionization timescale @xmath46 changed by a factor of three from observation a to d ( see table [ tb : t3 - 1 ] ) . in order to check the influences of the difference of the ionization timescale , we estimated the center energies based on the fakeit with the non - equilibrium ionization models ( table [ tb : t3 - 1 ] ) .
the center energies for non - equilibrium ionization models had no difference and were exactly similar to that for the ce plasma model in the result .
therefore , we regarded the observed shifts in the line center energy as the doppler shift .
the radial velocity changed from @xmath32860 to @xmath0770 km s@xmath161 from observation a to that of d. the radial velocities in table [ tb : t3 - 6 ] were calculated by adopting the rest frame energy as 6,676 ev .
this is consistent with the earlier result of @xmath162 observations obtained by @xcite ( orbital phase @xmath62 1.986 and 2.032 ) .
they reported that fe@xmath127 line was shifted from @xmath32600 km s@xmath161 to@xmath0648@xmath163 during the periastron passage .
such a variation was also seen in he@xmath164 line in the near - infrared band during the periastron passage in 2001 @xcite .
in addition , @xcite reported similar shifts of c@xmath165 and he@xmath164 lines at optical wavelength , which are thought to be emitted from wind - wind collision layer .
our observed variation of line center energy is consistent with such doppler motions .
this indicates that the emission in all wavelengths comes from a common shock cone .
l l c c + & observation & a & b + + & orbital phase&2.904 & 2.989 + + + & @xmath119 & 1.7@xmath166 & = a + & norm ( 10@xmath75)&1.5@xmath167&1.8@xmath168 + & @xmath52 ( 10@xmath169 @xmath54 @xmath37 @xmath55 ) & 7.3&9.0 + & @xmath56 ( 10@xmath170 @xmath54 @xmath55 ) & 6.2 & 7.6 + & @xmath58/dof & + figure [ fg :
f3 - 3 ] shows the xis and hxd / pin spectra and the xis best - fit models of the variable component . as explained in @xmath33 [ subsec : spectral_chara ] , the combined spectra of nxb and cxb were used as the hxd / pin background spectra . in both a and b observations
, we found the hard - tail excess emission in the hxd / pin band .
the excess count rates after deduction of the emission from variable component at observations a and b are 0.022 and 0.021 counts per second in the 1550 kev band , respectively . on the other hand , in observations c and d , there are no significant excess emission in the hxd / pin band spectra .
hereafter , we fit the excesses from the warm and hot components in observations a and b , respectively .
we introduce a power - law model for the hard band excess .
we simultaneously fit hxd / pin spectra of observations a and b , tying the power - law index , since the hxd / pin spectra are limited in photon statistics .
since the spectra do not constrain the soft band shape of the hard band excess , we assumed the same absorption to the excess component as that to the hot component in each phase .
the best - fit model yields a power - law index @xmath171 .
however , the hard band excess could be plausibly reproduced with a bremsstrahlung model with @xmath172 @xmath4210 kev , too .
the results are shown in figure [ fg : pin ] and table [ tb : t3 - 5 ] .
we have to consider the contamination from the other hard x - ray source igr j20216 + 4359 , because this source is located in the field of view of hxd / pin .
we discuss this contamination in detail in @xmath33 [ subsec : hard - tail ] .
( 80mm,80mm)figure12.eps ( 85mm,85mm)figure13.eps
the warm , hot and hard band excess components vary with orbital phase .
the cool component is detected only from observations c and d. the cool component appears nearly constant .
the cool component was discovered for the first time with this observation .
the simple analysis of one - temperature ce plasma emission for the cool component failed to reproduce the observed spectra with the residuals at 1.21 kev and other energies remaining in the spectral fitting .
the energy of the residuals coincides fairly well with the free - bound absorption features of the highly ionized ions of o , ne and mg elements .
the fit is then significantly improved with the recombining collisional plasma model , which introduces the radiative recombination continua ( rrc ) .
( 85mm,120mm)figure14.eps the ly@xmath9/ly@xmath31 line flux ratio of ne in the cool component is also found to be significantly higher than that expected for the ce plasma ( figure [ fg : f6 - 3 ] ) .
the high flux ratio can be understood with a transition process from a recombination of free electrons .
both the high ly@xmath9/ly@xmath31 ratio and the rrc features indicate that the cool component is interpreted as a collisional plasma in a recombining phase .
the small absorption column of an order of @xmath173 h @xmath21 ( see table [ tb : t3 - 4 ] ) constrains the location of the cool component ( figure [ fg:3 - 4 ] ) .
the spectra of the hot components are heavily absorbed below 1.6 kev at observations c and d. the large difference in the absorptions between both components indicates that the cool component is not deeply embedded in the wind - wind collision layer at observations c and d. moreover , the column density for the cool component is similar to or is rather smaller than that for the warm and hot components at observation a , in which the distance of colliding wind region to the w - r star is about 10 au .
this suggests that the cool plasma is located far away from the w - r star by 10 au or more . in conclusion ,
the cool component that we discovered can be interpreted as a collisional plasma in a recombining phase .
a potential scenario is as follows .
the plasma was heated in the past by a wind - wind collision shock .
the plasma then escaped from the dense wind region , and now , is recombining electrons and emitting the radiative recombination continuum along with the collisional plasma emission .
( 170mm,170mm)figure15.eps the initial temperature @xmath41 of the cool component was derived from the fitting results with the recombining collisional plasma model ( table [ tb : t3 - 4 ] ) .
the range of @xmath41 contains the temperature of the warm and hot components , which were formed by the wind - wind collision . on the one hand ,
the plasma temperature @xmath39 is now 0.021 kev ( table [ tb : t3 - 4 ] ) .
this temperature of the cool component is between the temperature of the wind - wind collision plasma and that of the dust detected in the infrared wavelengths @xcite .
the observed cool component might be emitted from the area in a _
phase , where the hot gas , compressed by the wind - wind collision , is cooling down and is forming dust .
the cooling of the shock heated plasma in wind - wind collision layer should be common in all the w - r binaries . if so , is the recombining plasma with less absorption detected commonly in them ?
the rrc structures in the x - ray spectra were reported from the two bright colliding wind binaries @xmath174 velorum ( wc8 @xmath0 o7.5 : @xcite ) and @xmath175 muscae ( wc5 @xmath0 o6 @xmath0 o9.5 : @xcite ) . at least , the emission component with the rrc structure of @xmath174 velorum is less absorbed .
@xmath174 velorum and @xmath175 muscae are both wc - subtype w - r binaries . on the other hand ,
no detection of the rrc structure was reported from the other subtype wn .
systematic and deep search of the structure from w - r binaries is critical to answer the question . from table [ tb : t3 - 1 ]
, we can see that the emission measure of the hot component varied .
generally , in colliding wind binaries , the emission measure in colliding wind region is expected to vary along with inversely proportional relation with the binary separation _ d _ @xcite . in order to check whether wr 140 varied with this relation , we plotted the emission measure of the hot component in figure [ fg : f3 - 5 ] .
the emission measure of the warm component and fe@xmath127 k@xmath31 line flux are also plotted in the same figure . as the phase approaches toward periastron , to observation c and then to d
, the emission measure of the hot component deviated from the expected 1/_d _ law .
no numerical simulation ever published predicts a rapid flux decrease at periastron , although some theorists simulated the variation of the x - ray luminosity for the colliding - wind binary ( e.g. , @xcite ; @xcite ; @xcite ) .
we therefore took account of not only the separation _
d _ , but also another parameter . according to the equation ( 10 ) in @xcite , the intrinsic x - ray luminosity of the colliding wind zone can be written as @xmath176 where the wind momentum flux ratio , @xmath177 . here ,
@xmath178 and @xmath179 show the distances to the shock region from the o star and from the w - r star , respectively .
the geometry is sketched in figure [ fg : f4 - 2 ] .
note that _ a _ is not constant and changes depending on not only the binary separation _ d _ but also the wind acceleration parameter @xmath9 and the ratio of mass - loss rates . at periastron ,
the momentum flux of the o - type wind should become significantly smaller , since its wind has less space to reach terminal velocity before entering the shock region , while the w - r wind should be less affected . in figure
[ fg : f4 - 3 ] , we show the maximum pre - shock stellar wind velocities of the w - r and o stars at a given binary separation .
we adopted a simple beta law for the wind acceleration : @xmath3 .
here , @xmath5 and @xmath180 are the terminal wind velocity and the stellar radius , respectively .
we used the value of @xmath181 2860 @xmath182 @xcite , @xmath183 3100 @xmath182 @xcite , @xmath184 and @xmath185 ( cf .
@xcite ) . many studies ( e.g. , @xcite ; @xcite ) indicate that the @xmath9 value of o star winds should be near 1 , while w - r stars can have larger values @xmath1861 .
we therefore fixed the @xmath9 value of o star wind to unity in the following simulations .
the actual @xmath9 value for the w - r star is not so important since the stagnation point is far from the w - r star . near the shock region
, the w - r wind reaches its terminal velocity for any assumed @xmath9 , while the o - star wind lags far behind . as a result ,
the wind momentum flux ratio , _ a _ , rapidly becomes large near periastron .
this means that the stagnation point approaches the o star . in figure
[ fg : f4 - 6 ] , we plot the normalized x - ray luminosity as a function of the binary separation @xmath187 .
the luminosity is normalized at the separation of observation a ( @xmath188 14 au ) . in this figure
, we adopted the x - ray luminosity ratio of the hot components .
the bolometric x - ray luminosity ratio of the warm component to the hot one is about one in observations a and b ( table [ tb : t3 - 1 ] ) .
we assumed that the ratios in observations c and d remain unchanged .
the observed luminosity peak appears near observation b. the luminosity drops at observations c and d and indicates the wind momentum ratio can not be constant ; @xmath189 .
( 80mm,80mm)figure16.eps we calculate equation ( [ eq1 ] ) for three assumptions of the mass - loss rate ratio @xmath190 @xmath62 0.1 , 0.04 , and 0.02 .
the @xmath9 model can reproduce the luminosity during observations a and b. however , it is truncated near periastron for all the assumptions since the wind momentum of the w - r star overwhelms that of the o star .
this implies that the stagnation point almost arrives at the surface of the o star near observations c and d. the observed ( intrinsic , absorption - free ) x - ray luminosity during observations c and d is then mostly produced near the surface of the o star .
one may think the radiative inhibition because of the stronger radiation field of the o star ( radiative breaking effect : @xcite ) .
@xcite took the `` eclipse spectra '' ratio of the iue spectrum very close to periastron ( phase 0.01 ) to a composite from phases 0.5 to 0.95 ( figure 7 in @xcite ) .
the spectra shows red - shifted absorption at 1300 @xmath191 in the si@xmath192 13941403 doublet and 1400 @xmath191 in the c@xmath192 15481551 doublet . at this phase
the o star behind the w - r star was seen and the w - r wind was being looked through towards the stagnation point . by correcting the velocity with the inclination of 54 degrees , the wind velocity was radiatively braked and was estimated as about 2300 @xmath191 before the collision with the o star wind , i.e. about 80% of its terminal velocity when it meets the o stellar wind .
if the radiation breaking dominates the momentum balance in the collision layer , the wc star winds lose its momentum and the cone angle of the collision layer will be enlarged , making the x - ray luminosity increased .
the observed flux drop at the periastron indicates that the radiative wind breaking is not a major effect to control the momentum balance in the wind - wind collision layer ( i.e. , x - ray luminosity ) . in this condition ,
the mass - loss ratio @xmath193 of 0.04 is good to explain the observed luminosities at observaion b. if we adopted this mass - loss ratio and @xmath194 , which reproduced the observed column densities at observations c and d ( for the detail , refer to @xmath33 [ subsec : mass - loss ] ) , we obtain @xmath195 .
this mass - loss rate for the o star is close to that recently reported by a number of analyses ( @xmath196 ; @xcite , @xmath197 ; @xcite ) .
we estimated the mass - loss rate of the w - r star from the observed column density @xmath198 .
the mass - loss rate for a spherically symmetric wind is expressed as @xmath199 @xmath62 4@xmath200 @xmath201 , in which _
r _ is the distance from the center of the w - r star . here , @xmath202 is the wind mass density .
in addition the he - equivalent column density can be expressed as @xmath203 from the x - ray emitting region to infinite distance along the line of sight `` @xmath204 '' ( cf .
@xcite ) . here , @xmath205 is the proton mass and @xmath31 is the ratio of helium mass to total mass in the w - r wind .
we therefore apply for the mass - loss rate of the w - r star , whose column density gives close agreement with the observed he - equivalent column density .
the computed he - equivalent column density was obtained as follows .
we first calculated the stagnation point in consideration of the effect of non - constant wind momenta , as shown in @xmath33 [ subsec : flux ] .
then , we assumed that the x - ray is emitted from the stagnation point .
finally , we calculated the absorption column density of the w - r stellar wind . here , we used @xmath206 , which was calculated by utilizing the elemental abundances of table [ tb : t3 - 2 ] .
we adopted a simple beta law model for the wind acceleration : @xmath3 .
we used the same values of @xmath5 , @xmath180 and @xmath9 for the w - r and o stars as those in @xmath207 .
an orbital inclination angle @xmath208 was adopted @xcite .
figure [ fg : f3 - 6 ] shows a plot of the best - fit column density from table [ tb : t3 - 1 ] .
we also plotted our calculated column density through the w - r wind , in the case of @xmath209 . here , we tested the three cases with the mass - loss rate ratios of @xmath190 @xmath62 0.1 , 0.04 , and 0.02 .
the calculated column densities for the @xmath210 1 , 3 and 5 were much the same . in observation
a , we set the absorption column density caused by the w - r wind to zero .
the absorption column density by the wind increased as the binary approached periastron . in the case of @xmath190 @xmath62 0.04 , and 0.02
, the stagnation points at observations c and d were at the o star surface .
we assumed that the wind - wind collision layer is concentrated in a small region at the stagnation point .
the good correspondence between the model and the data , in observations c and d , indicates our assumption is reasonable at least in the phases . on the other hand , at observation b
, there is a discrepancy by a factor of two between the data and the model .
the layer at observation b is largely extended away from the stagnation point and the o star . during observation b ,
the x - rays emitted from the wind - wind collision layer will suffer absorption at different column density according to the different line of sight . in the calculation of the wind absorption , we assumed values for the stellar radius of the w - r star @xmath211 , the inclination angle @xmath212 of the binary system and the c / he abundance ratio .
if the stellar radius of the w - r star was in fact , for example , a half of our assumed value , i.e. , @xmath213 , the absorption column at periastron would be lower by roughly 10% than the above - mentioned value .
if the inclination angle was a five - degree larger than our assumed value , the absorption column at periastron would be larger by roughly 10% . if the assumed c / he abundance ratio was a half of it , i.e. , c / he @xmath214 , the absorption column at periastron would be larger by roughly 20% . or , if the ratio was doubled , i.e. , c / he @xmath215 , the column at periastron would be 20% lower .
as such these three parameters are sensitive to @xmath198 . accordingly the mass - loss rate derived from @xmath198 may have some considerable uncertainty .
@xcite derived the mass - loss rate @xmath216 , which assumed a distance of 1.3 kpc and made no allowance for clumping , from the radio flux at the quiescent phase . allowing for the clumping having a filling factor of 0.1 and the revised distance ,
the mass - loss rate @xmath217 was determined as @xmath218 . by using x - ray absorption , we obtained @xmath219 at observations c and d near the periastron .
our value can coincide with that by using the radio flux .
we have detected a signal in the high - energy band above 10 kev ( figure [ fg : f3 - 3 ] ) .
however , the seyfert 2 galaxy , igr j20216 + 4359 , is located in the field of view of pin / hxd .
the detected hard - tail signal may be contaminated with the emission from igr j20216 + 4359 .
igr j20216 + 4359 was discovered by @xcite during their eleven months observation . according to @xcite , igr j20216 + 4359
is highly absorbed ( @xmath220 @xmath37 ) , when they fixed @xmath119 to 1.7 , which is a typical value of a seyfert 2 galaxy . by assuming the flux and the spectral shape are the same as those in @xcite through our observations , we estimated the count rate of the contamination from igr j20216 + 4359 . the model predicted count rate with hxd / pin is dependent on the off - axis angle of igr j20216 + 4359 .
the count rates at observations a , b , c and d are 0.013 , 0.017 , 0.016 and 0.016 counts per second in the 1550 kev band , respectively .
the contaminations from igr j20216 + 4359 accounts for about 60 and 80% of the excess count rates at observations a and b , respectively . on the other hand , there are no excess count rates at observations c and d. the count rates from off - axis igr j20216 + 4359 at observations c and d were 2.0 and 3.2 times larger than the observed count rates in the 1550 kev band , respectively .
this suggests that the flux of igr j20216 + 4359 were varied . since the flux of igr j20216 + 4359 was varied , we can not constrain the origin of the hard band excess in observations a and b. therefore , we regard the flux detected at observation b is an upper limit of the hard band excess from wr 140 ( @xmath221 , see table [ tb : t3 - 5 ] ) . _ astro - h _ and
_ nustar _ , which have hard x - ray imaging instruments with high sensitivity , will be useful to verify the existence of the hard band excess . in the colliding - wind region , first - order diffusive shock acceleration results in the production of a power - law spectrum with index 2 for electrons ( e.g. , @xcite , @xcite , @xcite and @xcite ) .
the best - fit photon index ( @xmath222 1.7 ) indeed approximates a photon index for non - thermal electron energy distribution . @xcite
discuss the photon index of the non - thermal emission from wr 140 .
they expected that the photon index is 1.7 above the energy of 1 mev .
our fitting results are consistent with it , although our results have a large error . according to @xcite ,
the luminosity ratio of inverse compton to synchrotron radiation can be written as @xmath223 , where the magnetic field @xmath224 is expressed in g , @xmath225 is the distance from the colliding - wind zone to the o star in au , and @xmath226 is the o star luminosity in @xmath227 units .
for example , during observation b , if we suppose @xmath228 ( see table [ tb : t3 - 5 ] ) , @xmath229 for a distance of 1.67 kpc ( [email protected] ; @xcite ) , @xmath230 @xcite , and @xmath231 , the total value of the magnetic field for observation b is @xmath2321 g. because of the difficulty of measuring the magnetic field of the w - r stellar surface , the value of the magnetic field around the w - r star will be helpful to study the evolution of the massive star .
we have observed the colliding - wind binary wr 140 around the 2009 periastron passage with the _ suzaku _ satellite .
the following are the salient results .
* we discovered a cool component .
the absorption to this component is smaller than that to the other components .
we infer that this component extends out farther than the other components does .
arguably this component may represent a transitional phase from the compressed hot gas to dust formation . *
the luminosity of the hot component is not inversely proportional to the binary separation .
this discrepancy may be explained if the o - star wind collides with the w - r star wind before it has reached its terminal velocity , leading to a reduction in its wind momentum flux .
this interpretation needs to be verified by future theoretical modelings . * as wr 140 approaches periastron , the column density of the hot component increases .
the column densities at near periastron are reproduced well with a simple model , in which the absorption is occurred in the w - r wind and the x - ray emitting region is fairly compact .
* we detected a hard x - ray signal above 10 kev in the hxd / pin data .
however , we could not eliminate the possibility of contamination from the background source igr j20216 + 4359 , 17.3 arcmin north - east of wr 140 .
the pin data gives the upper - limit for the x - ray luminosity of wr 140 , 8@xmath233 .
we thank the referee for their help in improving the quality of this paper .
we also thank m. sakano for his invaluable help .
this research has made use of data and/or software provided by the high energy astrophysics science archive research center ( heasarc ) , which is a service of the astrophysics science division at nasa / gsfc and the high energy astrophysics division of the smithsonian astrophysical observatory . y.s . acknowledges financial support from the japan society for the promotion of science .
y.t . and y.m .
acknowledge support from the grants - in - aid for scientific research ( numbers 20540237 , 21018009 and 23540280 ) by the ministry of education , culture , sports , science and technology . k.h .
is grateful for financial support by the nasa s astrobiology institute ( rtop 344 - 53 - 51 ) to the goddard center for astrobiology .
is grateful for financial aid to nserc ( canada ) and fqrnt ( qubec ) .
p.m.w . is grateful to the institute for astronomy for continued hospitality and access to the facilities of the royal observatory .
j.m.p . would like to thank the royal society for funding a university research fellowship .
this research has made use of nasa s astrophysics data system .
antokhin , i. i. , owocki , s. p. , & brown , j. c. 2004 , , 611 , 434 bell , a. r. 1978 , , 182 , 147 bikmaev , i. f. , burenin , r. a. , revnivtsev , m. g. , et al .
2008 , astronomy letters , 34 , 653 boldt , e. 1987 , observational cosmology , 124 , 611 corcoran , m. f. , pollock , a. m. t. , hamaguchi , k. , & russell , c. 2011 , arxiv:1101.1422 de becker , m. , pittard , j. m. , williams , p. , & wr 140 consortium 2011 , bulletin de la societe royale des sciences de liege , 80 , 653 dougherty , s. m. , beasley , a. j. , claussen , m. j. , zauderer , b. a. , & bolingbroke , n. j. 2005 , , 623 , 447 eichler , d. , & usov , v. 1993 , , 402 , 271 fahed , r. , et al .
2011 , , 418 , 2 fukazawa , y. , et al .
2009 , , 61 , 17 gorenstein , p. 1975
, , 198 , 95 hamaguchi , k. , et al .
2007 , , 663 , 522 hillier , d. j. , & miller , d. l. 1999 , , 519 , 354 ishisaki , y. , et al .
2007 , , 59 , 113 kokubun , m. , et al .
2007 , , 59 , s53 koyama , k. , kawada , m. , takano , s. , & ikeuchi , s. 1990 , , 42 , l1 koyama , k. , maeda , y. , tsuru , t. , nagase , f. , & skinner , s. 1994 , , 46 , l93 koyama , k. , et al .
2007 , , 59 , s23 koyama , k. , hyodo , y. , inui , t. , et al .
2007 , , 59 , 245 lpine , s. , & moffat , a. f. j. 1999 , , 514 , 909 luo , d. , mccray , r. , & mac low , m .- m . 1990 , , 362 , 267 maeda , y. , koyama , k. , yokogawa , j. , & skinner , s. 1999 , , 510 , 967 marchenko , s. v. , et al .
2003 , , 596 , 1295 mitsuda , k. , et al .
2007 , , 59 , s1 monnier , j. d. , zhao , m. , pedretti , e. , et al .
2011 , , 742 , l1 morris , p. w. , brownsberger , k. r. , conti , p. s. , massey , p. , & vacca , w. d. 1993 , , 412 , 324 myasnikov , a. v. , & zhekov , s. a. 1993 , , 260 , 221 nugis , t. , crowther , p. a. , & willis , a. j. 1998 , , 333 , 956 oskinova , l. m. 2005 , , 361 , 679 ota , n. , fukazawa , y. , fabian , a. c. , et al .
2007 , , 59 , 351 ozawa , m. , uchiyama , h. , matsumoto , h. , et al .
2009 , , 61 , 1 pittard , j. m. , & dougherty , s. m. 2006 , , 372 , 801 pollock , a. m. t. 1987 , , 171 , 135 pollock , a. m. t. , haberl , f. , & corcoran , m. f. 1995 , wolf - rayet stars : binaries ; colliding winds ; evolution , 163 , 512 pollock , a. m. t. , corcoran , m. f. , & stevens , i. r. 2002 , interacting winds from massive stars , 260 , 537 pollock , a. m. t. , corcoran , m. f. , stevens , i. r. , & williams , p. m. 2005 , , 629 , 482 protassov , r. , van dyk , d. a. , connors , a. , kashyap , v. l. , & siemiginowska , a. 2002 , , 571 , 545 puls , j. , kudritzki , r .-
p . , herrero , a. , et al .
1996 , , 305 , 171 puls , j. , markova , n. , scuderi , s. , et al .
2006 , , 454 , 625 schild , h. , et al .
2004 , , 422 , 177 serlemitsos , p. j. , et al .
2007 , , 59 , s9 setia gunawan , d. y. a. , van der hucht , k. a. , williams , p. m. , et al .
2001 , , 376 , 460 schild , h. , gdel , m. , mewe , r. , et al .
2004 , , 422 , 177 smith , r. k. , brickhouse , n. s. , liedahl , d. a. , & raymond , j. c. 2001 , , 556 , l91 stevens , i. r. 1993 , , 265 , 601 stevens , i. r. , blondin , j. m. , & pollock , a. m. t. 1992 , , 386 , 265 stevens , i. r. , & pollock , a. m. t. 1994 , , 269 , 226 sugawara , y. , tsuboi , y. , & maeda , y. 2008 , , 490 , 259 takahashi , t. , et al .
2007 , , 59 , s35 tamura , t. , hayashida , k. , ueda , s. , & nagai , m. 2011 , , 63 , 1009 usov , v. v. 1991 , , 252 , 49 usov , v. v. 1992 , , 389 , 635 varricatt , w. p. , williams , p. m. , & ashok , n. m. 2004 , , 351 , 1307 vuong , m. h. , montmerle , t. , grosso , n. , et al .
2003 , , 408 , 581 white , r. l. , & becker , r. h. 1995 , , 451 , 352 white , r. l. , & chen , w. 1995 , wolf - rayet stars : binaries ; colliding winds ; evolution , 163 , 438 williams , p. m. , van der hucht , k. a. , pollock , a. m. t. , florkowski , d. r. , van der woerd , h. , & wamsteker , w. m. 1990 , , 243 , 662 williams , p. m. , et al .
2009 , , 395 , 1749 wilms , j. , allen , a. , & mccray , r. 2000 , , 542 , 914 zhekov , s. a. , & skinner , s. l. 2000 , , 538 , 808 | _ suzaku _ observations of the wolf - rayet binary wr 140 ( [email protected] ) were made at four different times around periastron passage in 2009 january .
the spectra changed in shape and flux with the phase . as periastron approached , the column density of the low - energy absorption increased , which indicates that the emission from the wind - wind collision plasma was absorbed by the dense w - r wind .
the spectra can be mostly fitted with two different components : a warm component with @xmath10.30.6 kev and a dominant hot component with @xmath23 kev .
the emission measure of the dominant , hot component is not inversely proportional to the distance between the two stars .
this can be explained by the o star wind colliding before it has reached its terminal velocity , leading to a reduction in its wind momentum flux . at phases closer to periastron
, we discovered a cool plasma component in a recombining phase , which is less absorbed .
this component may be a relic of the wind - wind collision plasma , which was cooled down by radiation , and may represent a transitional stage in dust formation . |
geometrically frustrated magnetoelectric ( me ) multiferroics , where complex magnetic long - range order ( lro ) induces the electric polarization , may exhibit spectacularly strong me coupling.@xcite the applied magnetic field can influence the magnetic lro and can thus through me coupling change the electric polarization .
vice versa , the magnetic order may be manipulated by external electric field , acting on the electric polarization.@xcite magnetic frustration is necessary for the realization of complex magnetic structures@xcite without the inversion symmetry , which allow for the establishment of the electric polarization.@xcite since the geometric frustration arises from the arrangement of the magnetic ions within the crystal lattice,@xcite the understanding of the relation between structural and magnetic properties is vital for the design of materials with an enhanced me effect.@xcite an established approach , which has largely contributed to the knowledge about the me phenomenon , is based on the investigation of the isostructural relatives of already known multiferroics.@xcite isostructural compounds may exhibit very different me responses , depending on the leading lattice parameter involved in the me mechanism .
the latter has two main origins ; ( i ) changes of the angle between atoms in the superexchange bridge that affects the strength of the corresponding interactions,@xcite and ( ii ) changes of the local structure in the vicinity of the magnetic ion , determining its local crystal - field anisotropy.@xcite as a result , significantly different magnetic states can be realized , from which only a few may be multiferroic .
the prominent examples of such experimental approach are studies of @xmath7mno@xmath8 , where the radius of the rare - earth ion @xmath7 influences the mn - o - mn angle and thus selects between non - ferroelectric collinear and multiferroic spiral magnetic ground states.@xcite [ ! ] o@xmath9 , @xmath10=cl ( or br ) .
the @xmath11 layers are separated by cl@xmath12 ions ( light green for cl@xmath5 and dark green for cl@xmath0 ) , which determine the interlayer distance .
te@xmath6 ions are gray ( large spheres ) , o@xmath13 are red ( small spheres ) , while [ feo@xmath14 are shown as yellow octahedra.,title="fig:",scaledwidth=50.0% ] here we focus on the fete@xmath0o@xmath1cl@xmath15br@xmath16 family of layered materials with monoclinic unit cell ( space group @xmath17 at room temperature ) , where @xmath18 can be varied between 0 and 1 ( fig.[crystruct]).@xcite the layers , oriented perpendicular to @xmath19 , consist of [ fe@xmath20o@xmath21@xmath22 iron tetramers connected via te@xmath6 ions and are separated by halogen ions located at two crystallographically non - equivalent positions.@xcite up to now , the majority of studies@xcite - including dielectric and magnetic structure investigations - have been restricted to @xmath18=0 , namely fete@xmath0o@xmath1br ( ftob ) , which was found to be multiferroic at low temperatures.@xcite the density - functional - theory calculations yield a surprising result that the intra - tetramer fe - o - fe and the inter - tetramer fe - o - te - o - fe exchange interactions are comparable.@xcite consequently , the system acts as an assembly of alternating antiferromagnetic fe@xmath3 ( @xmath4=5/2 ) spin chains running along the @xmath23 axis coupled by weaker interactions , which are geometrically frustrated.@xcite moreover , te@xmath6 ions possess lone - pair electrons , which can be easily polarized and can thus lead to the electric polarization , if the inversion symmetry of the @xmath17 structure is broken.@xcite in ftob , two magnetic transitions occur , at @xmath24=11k and @xmath25=10.5k , separating the high- and the low - temperature incommensurate ( ht - ic and lt - ic ) magnetic lro phases.@xcite both phases exhibit elliptical amplitude - modulated magnetic lro , where a substantial part of the fe@xmath3 magnetic moments is fluctuating down to the lowest experimentally accessible temperatures.@xcite below @xmath25 , in the lt - ic phase , spontaneous electric polarization develops,@xcite indicating a strong me coupling , which is evidenced also by a magnetic field dependence of @xmath25.@xcite the me coupling mechanism was explained by phase shifts between exchange coupled spin modulation waves.@xcite however , since the crystal - lattice distortions accompanying the multiferroic transition are extremely small,@xcite complete understanding of this exotic me mechanism on the microscopic level is not yet achieved .
for instance , it is still not clear to what extent the lone - pair electrons are involved in the me coupling or why electric polarization develops only below @xmath25 even though symmetry restrictions are broken already in the ht - ic phase.@xcite to address these questions we focus on the isostructural fete@xmath0o@xmath1cl ( ftoc ) compound ( @xmath18=1 ) , where the smaller size of the halogen ion ( cl@xmath12 has by @xmath260.15 smaller radius compared to br@xmath12 ) leads to substantially reduced interlayer distance.@xcite on the other hand , the interatomic distances within each layer are very similar to those in ftob , implying that the dominant intralayer exchange interactions in both systems are comparable .
this agrees with the curie - weiss temperature @xmath27 = 124k , which is close to 98k measured in ftob.@xcite moreover , the magnetic transition temperatures , @xmath24=12.6k and @xmath25=11k,@xcite are also very similar to the values found in ftob.@xcite interestingly , far - infrared study of ftoc could not detect any link between electrodynamic response and magnetic ordering at low temperatures,@xcite implying that , in contrast to ftob , ftoc may not be multiferroic .
here we present a comprehensive study of the ftoc single crystals including specific - heat , muon spin relaxation ( @xmath28sr ) , nuclear magnetic resonance ( nmr ) , neutron diffraction , dielectric and synchrotron x - ray diffraction measurements , allowing us a full characterization of magnetic ordering , dielectric response and the coupling between them .
results provide important new data on the fete@xmath0o@xmath9 family and improve the understanding of the irregular me mechanism active in these systems .
high - quality powder and single crystals of ftoc were grown by the standard chemical - vapor - phase method.@xcite specific - heat measurements were performed between 4 and 16k on a quantum design physical property measuring system at joef stefan institute ( jsi ) , slovenia .
the @xmath28sr spectra were measured on powder samples in zero field as well as in a weak transverse field of 3mt between 2 and 50k on the general purpose surface - muon instrument at the paul scherrer institute ( psi ) , switzerland .
additional zero - field experiments between 60mk and 5k were conducted on the musr instrument at the isis facility , rutherford appleton laboratory , united kingdom .
the @xmath29cl nmr was measured using a home - built spectrometer at jsi between 4 and 300k in the magnetic field of 9.4 t applied along @xmath19 , i.e. , the reference larmor frequency is @xmath30=39.192mhz .
single - crystal neutron diffraction experiments with the wave length 2.317 were performed on a plate - like crystal of the size 12@xmath318@xmath312mm@xmath32 between 2 and 20k on the single - crystal diffractometer trics at psi .
magnetic structure was refined from integrated intensities of the magnetic reflections using magopt program based on the crysfml library.@xcite the complex dielectric constant , @xmath33=@xmath34@xmath35 , was measured as a function of temperature and frequency by using an hp4282a precision lcr meter at jsi .
the dielectric constant was scanned at a few frequencies between 20hz and 1mhz on cooling or heating the sample with the typical cooling / heating rates of 10k / h in various dc bias electric fields ranging from 0 to 3kv / cm .
the excitation electric ac field of 100 - 400v / cm was applied along the @xmath19 , @xmath36 , and @xmath23 axes . zero - field ac dielectric measurements and ac dielectric measurements in the dc electric bias field were performed in an oxford continuous - flow liquid - helium cryostat .
single - crystal x - ray synchrotron diffraction data were acquired at the bm01a swiss - norwegian beamline of esrf , grenoble , france .
data sets ( @xmath26 780 reflections per temperature point ) were collected between 4.5 and 35k at the wavelength of 0.64 , using a closed - cycle he cryostat mounted on a six - circle kappa diffractometer ( kuma ) , while the interatomic distances were refined using the shelx97 program.@xcite
specific - heat measurements reveal two @xmath37-type anomalies , which in the absence of the external magnetic field ( @xmath38=0 ) appear at @xmath24=12.5k and @xmath25=10.9k ( fig.[cp ] ) , i.e. , in accordance with the previous study.@xcite as in the isostructural ftob system , the two anomalies reflect the corresponding transitions from the paramagnetic to the high- and to the low - temperature ( ht and lt ) magnetic phases.@xcite when magnetic field @xmath38 is applied along the @xmath19 axis , the lower anomaly , at @xmath25 , broadens and shifts with increasing @xmath38 to lower temperatures .
the observed @xmath39 dependence is much more pronounced than in ftob ( inset in fig.[cp ] ) and thus implies that the low - temperature magnetic ordering in ftoc is more sensitive to the applied magnetic field , making the compound more appropriate to study a potential me response . on the other hand ,
the anomaly at @xmath24 gradually shifts to higher temperatures with increasing field , reaching @xmath24=13.20(3)k at 8 t , i.e. , reminiscent of the behavior in ftob .
[ ! ] axis .
inset : derived field dependence of the magnetic transitions ( solid symbols
the colors match the colors of the corresponding experimental curves ) in comparison to their behavior in ftob ( open symbols ) .
lines are guides for the eyes.,title="fig:",scaledwidth=50.0% ] although , @xmath24 and @xmath25 in ftoc show similar field dependence as in ftob , the magnetic ordering is yet unknown . to address this issue , we first conducted local - probe @xmath28sr experiments .
the muons are positively charged particles with a large magnetic moment , making them highly sensitive probes of local magnetic fields .
hence , @xmath28sr can easily distinguish between fluctuating and static magnetism as well as between lro and static magnetic disorder.@xcite [ ! ]
sr signal in zero field at 1.6k.,title="fig:",scaledwidth=50.0% ] to trace the evolution of the static magnetic correlations , we first performed @xmath28sr measurements in a weak transverse magnetic field ( tf ) @xmath40=3mt . in the paramagnetic phase , i.e. , well above @xmath24 ,
the static field @xmath38=@xmath40 exceeds other ( nuclear ) static fields , so the muon precession frequency @xmath41=@xmath42 is exactly determined by @[email protected] / t , reflecting in an oscillating @xmath28sr signal @xmath45=@xmath46 [ fig.[musr - spec](a ) ] .
when static spin correlations start to develop , the amplitude of these oscillations is suppressed by the internal field @xmath47 , developing at the @xmath48 site and largely exceeding @xmath40 , i.e. , @xmath47@xmath49@xmath40 .
below @xmath24 , this clearly reflects in the development of the so - called 1/3 tail , which in powder samples originates from the muons in the magnetically ordered phase whose initial polarization is parallel to the internal field .
consequently , the tf signal for @[email protected]@xmath28s can be modeled as@xcite @xmath52 here @xmath53 denotes the volume fraction of the sample , where magnetic correlations are significant ( @xmath47@xmath49@xmath40 ) , @xmath37 is the corresponding relaxation rate governed by local - field fluctuations , and @xmath54 is a stretch exponent , which is for quickly fluctuating electronic fields , e.g. , in paramagnetic phase , typically close to 1 , while it can be substantially reduced when relaxation rates are distributed . [ ! ] , ( b ) the longitudinal relaxation rate , ( c ) the @xmath28sr frequencies @xmath55 ( @xmath56=1,2 ) and the stretch exponent @xmath54 , determined from the weak transverse - field ( tf ) and the zero - field ( zf ) data .
inset in ( b ) shows longitudinal relaxation for zf experiment in a dilution refrigerator , whereas solid lines are guides for the eyes.,title="fig:",scaledwidth=50.0% ] below 20k , @xmath53 increases and finally reaches @xmath261 at @xmath24 [ fig.[musr - par](a ) ] . at the same time
, @xmath37 increases in a diverging manner when approaching @xmath24 [ fig.[musr - par](b ) ] , while @xmath54 reduces from @xmath261 to 0.45(2 ) at 12k [ fig[musr - par](c ) ] .
the observed behavior indicates the growth of short - range correlations already above the magnetic transition and the establishment of the magnetic lro below @xmath24 , where @xmath47 is substantially stronger than @xmath40 and thus suppresses the amplitude of the corresponding oscillation of the @xmath28sr signal . to probe the magnetic lro we switch to the zero - field ( zf ) experiment . on cooling below 20k
the @xmath28sr signal exhibits a monotonic stretched exponential decay,@xcite in agreement with the tf results . at @xmath24 , strongly damped high - frequency oscillations start to develop [ fig.[musr - spec](b ) ] , proving the presence of static internal fields of the electronic origin .
strong damping implies a broad local - field distribution , which is clearly manifested also in the fourier transform of the time - dependent part of the signal @xmath451/3 [ fig.[musr - spec](c ) ] .
the response is reminiscent of that of isostructural ftob with the ic magnetic ordering.@xcite since the @xmath28sr study of ftob revealed two muon stopping sites,@xcite we likewise model our zf @xmath28sr spectra by the two - component empirical model@xcite @xmath57 \nonumber \\ & + ( 1-\frac{2}{3}v_m)\text{e}^{-(\lambda_l t)^\alpha}.\end{aligned}\ ] ] this expression accounts for the relaxation at the two muon stopping sites ( @xmath56=1,2 ) with the corresponding fractions @xmath58 ( @xmath59=1 ) and the mean oscillation frequencies @xmath55=@xmath60 that are determined by average local magnetic fields @xmath61 .
the longitudinal relaxation rate @xmath62 accounts for the spin fluctuations responsible for the stretch - exponential decay of the @xmath28sr tail , while in addition to spin fluctuations , transverse relaxation rate @xmath63 can effectively detect also finite local - field distributions . by fitting the complete temperature dependence of the zf signal to eq .
[ fig.[musr - spec](b ) ] , we find that below 20k @xmath53 gradually changes from 0 to @xmath261 , where it settles at @xmath64@xmath65@xmath24 [ fig[musr - par](a ) ] .
similarly , the stretch exponent @xmath54 reduces from @xmath261 to 0.3 [ fig[musr - par](c ) ] .
this indicates a transition from the paramagnetic phase to a magnetic lro state with a broad relaxation - rate distribution , i.e. , in agreement with tf results .
moreover , the magnetic transition is also reflected in @xmath66 , displaying a sharp @xmath37-type anomaly at @xmath24 [ fig[musr - par](b ) ] .
we note that the temperature dependence of @xmath54 , was correlated with the fit of the tf signal , in order to ensure the comparability of the derived parameters .
the derived @xmath67 , on the other hand , displays the evolution of the magnetic order parameter in the magnetic lro phase , as @xmath61 is directly related to the ordered magnetic moment .
we stress that @xmath68 values at low temperatures ( @xmath64@xmath69@xmath24 ) are in agreement with those determined for ftob,@xcite signifying further similarities between the two magnetic orders .
finally , this correspondence reflects in very fast relaxation rates @xmath70 and @xmath71 that stretch between 15 and 60@xmath28s@xmath72 , accounting for the broad local - field distribution [ fig.[musr - spec](c ) ] , which was found also in ftob.@xcite the fact that @xmath62@xmath730 endures even at 1.6k , indicating that spin fluctuations are still active deep in the magnetically ordered state , prompted us to further cool the sample .
we performed additional zf @xmath28sr measurements down to 60mk [ inset in fig.[musr - par](b ) ] . since @xmath54=0.3 was found to be constant below 8k [ fig.[musr - par](c ) ] , we kept this value fixed in the analysis in the whole low - temperature region between 2k and 60mk . as in ftob
, @xmath66 saturates below 1k and settles at a finite value of @xmath260.020(3)@xmath28s@xmath72 , implying the existence of the persistent spin dynamics - a generic property of ic amplitude modulated magnetic order.@xcite the limiting zero - temperature relaxation rate is notably higher than in ftob , where it is 0.010(5)@xmath28s@xmath72 , suggesting that the density of gapless magnetic excitations is in ftoc higher . [ ! ] cl nmr spectrum in the magnetic field of 9.4 t , applied along @xmath19 .
the superscripts s and c denote satellite and central cl@xmath74 ( @xmath56=1,2 ) nmr transitions , respectively , while different colors are used for different magnetic phases .
temperature dependences of ( b ) the position and ( c ) the linewidth of the central cl@xmath0 paramagnetic signal .
the dotted line denotes the temperature where @xmath75 has a maximum .
solid line in ( b ) is the magnetic susceptibility measured in field of 0.01 t applied along @xmath19.,title="fig:",scaledwidth=50.0% ] nmr is another highly sensitive local - probe technique that probes magnetic properties .
in addition , when the investigated nucleus has a spin @xmath76@xmath511/2 , it is characterized by a finite quadrupole moment ( q ) , through which it couples to electronic - field gradient ( efg ) , making it sensitive to the tiniest changes of the crystal structure . to simultaneously probe both , magnetic as well as structural ( potentially dielectric ) responses , we performed nmr measurements on the @xmath29cl nuclei ( @xmath76=3/2 ) in the magnetic field of 9.4 t applied along @xmath19 in the frequency range @xmath77mhz around the @xmath29cl larmor frequency , @xmath30 .
the cl ions are located in - between the crystal layers at two crystallographically non - equivalent sites , namely cl@xmath5 and cl@xmath0 ( fig.[crystruct ] ) , which have no special site symmetries and where their efgs should be significantly different , as found for br@xmath5 and br@xmath0 in isostructural ftob.@xcite finally , we note that the @xmath78cl isotope has @xmath265 times smaller absolute sensitivity and almost 20% lower resonant frequency ( by 6.57mhz at 9.4 t ) than @xmath29cl , hence all signals shown in fig.[nmr ] are ascribed to the @xmath29cl isotope . at room temperature in the paramagnetic phase we find a sharp line shifted from the larmor frequency by @xmath75=@xmath41@xmath79=211(5)khz , whereas three very weak additional lines are resolved when temperature is lowered [ fig.[nmr](a ) ] . on cooling , all lines with @xmath75@xmath510 shift to higher frequencies , mimicking the behavior of the bulk magnetic susceptibility @xmath80 .
for instance , the shift of the dominant line increases and develops a broad maximum @xmath75=310(5)khz at 50k , which is followed by a reduction , @xmath75=286(5)khz at @xmath24 [ fig.[nmr](b ) ] . in contrast , the corresponding linewidth does not change significantly down to 50k and exhibits a pronounced broadening only below 30k , i.e. , still far above @xmath24 .
this is in line with @xmath28sr results and is reminiscent of the behavior in several frustrated and low - dimensional magnetic systems,@xcite where short - range magnetic correlations typically emerge at temperatures that are significantly elevated compared to the magnetic ordering temperature.@xcite in particular , such correlations introduce a distribution of local magnetic fields that effectively broadens the line and potentially influences the line shift .
we note that in some highly frustrated ( kagome ) systems the broadening of the nmr line with decreasing temperature has been ascribed to a broad distribution of local susceptibilities induced by nonmagnetic defects.@xcite however , in these materials the concentration of defects is high , typically affecting @xmath265% of the magnetic sites , i.e. , orders of magnitude more than in our crystals , where we found no trace of defects / impurities by any experimental technique .
hence , the defect scenario in ftoc is highly unlikely .
finally , we stress that short - range correlations were detected also in ftob by neutron diffraction as well as by @xmath28sr at @xmath64@xmath8150k.@xcite below @xmath24 all spectral lines abruptly broaden and develop a box - like shape , indicative of a distribution of local magnetic fields or efg encountered in the ic systems,@xcite as found in ftob.@xcite the signal at @xmath75=1mhz is now almost twice as broad as the rest of the lines [ fig[nmr](a ) ] .
in addition , on cooling below 12k this signal retains the box - like shape , while the rest of the lines develop a triangular form .
hence it must correspond to a nucleus , which has a different local environment , either due to a different efg tensor and/or because of a peculiar distribution of local magnetic fields.@xcite at 4k the spectrum changes even more and seems to be composed of four broad overlapping lines , centered at the paramagnetic positions .
this must be related to the @xmath25 transition , even though its presence above 7 t is not clear from the specific heat data . to clarify which line corresponds to
which cl site , we point out that in ftob br@xmath0 has much smaller efg than br@xmath5,@xcite implying that the signal with the smallest @xmath82 corresponds to cl@xmath0 .
we thus ascribe the sharp line at @xmath75=210khz to the central transition ( 1/2@xmath831/2 ) of the cl@xmath0 site , while the three weak lines must be the satellite transitions ( @xmath843/2@xmath83@xmath841/2 ) of cl@xmath0 and cl@xmath5 .
moreover , below @xmath24 the line with the shift of @xmath75=1mhz shows distinctly different behavior from the rest , indicating that it corresponds to cl@xmath5 , while the two with @xmath75=@xmath85 and 1.5mhz correspond to cl@xmath0 .
we note that the second satellite as well as the central lines for cl@xmath5 are out of our experimental frequency range .
the room temperature @xmath29cl shifts ( @xmath75 ) for cl@xmath0 are approximately 10 times smaller compared to the shifts of @xmath86br for br@xmath0 in ftob .
since @xmath87@xmath88@xmath89,@xcite @xmath75 scales with @xmath90 , implying that it is dominated by the second order quadrupole shift.@xcite strikingly , scaling @xmath75 with the magnetic susceptibility reveals that @xmath29cl hyperfine shifts for cl@xmath0 are also by an order of magnitude smaller compared to @xmath86br shifts for br@xmath0 in ftob .
this , on the other hand , indicates different strengths of the magnetic interactions and probably reflects the difference in the core polarization hyperfine fields of the br@xmath12 and the cl@xmath12 ions . to determine the magnetic ordering in ftoc we conducted neutron diffraction experiments on a single - crystal sample .
distinct magnetic reflections are found for ic magnetic wave vector * q*@xmath91=(@xmath920.4650 ) at 5k , i.e. , almost exactly matching the ( @xmath920.4630 ) wave vector in ftob . to confirm the two magnetic transitions , we first measured the temperature evolution of the strongest magnetic reflection ( 3.5 0.465 2 ) in the @xmath93-scan geometry [ fig.[nd - par](a ) ] .
we point out that in the vicinity of @xmath24 the gaussian profile , corresponding to the magnetic lro , is superimposed on top of a broad diffuse lorentzian contribution , indicative of short - range magnetic correlations . [ ! ] position for this reflection .
solid lines are guides for the eyes.,title="fig:",scaledwidth=50.0% ] to derive the temperature dependences of the peak position and its intensity , we thus fitted the data with pseudo - voigt function , simultaneously accounting for the lro ( gaussian ) and the diffuse ( lorentzian ) magnetic scattering contributions .
our analysis [ fig.[nd - par](b ) ] clearly reveals the emergence of lro contribution at @xmath24=12.5k , while the diffuse scattering can be noticed already at 13k , i.e. , above @xmath24 .
the diffuse fraction is suppressed at @xmath25=11k , i.e. , at the second magnetic transition , where the intensity of the lro contribution exhibits an anomaly [ fig.[nd - par](b ) ] and the magnetic reflection starts to shift with further decreasing temperature [ fig.[nd - par](c ) ] .
similar behavior was observed in ftob , where the diffuse scattering below @xmath24 was ascribed to a finite magnetic correlation length along @xmath19,@xcite while the shifts of the magnetic reflections were identified as a sign of the exchange striction.@xcite since both magnetic phases exhibit ic lro , we address them as ht - ic ( between @xmath24 and @xmath25 ) and lt - ic ( below @xmath25 ) phases , in analogy with ftob . [ ! ] o@xmath9 ( @xmath10=br , cl ) , compared at fe@xmath94 and fe@xmath95 sites for the lt - ic phase in ( a ) @xmath96 and ( b ) @xmath97 projections , and for the ht - ic phase in ( c ) @xmath96 and ( d ) @xmath97 projections . for clarity , magnetic unit cell along @xmath36 is flattened and the sizes of the magnetic moments in the ht - ic phase are magnified by a factor 3.,title="fig:",scaledwidth=48.0% ] we next measured intensities of 320 magnetic reflections at 6k , which allow for the refinement of the magnetic structure in the lt - ic phase . as in ftob , the representation analysis shows that the magnetic wave vector * q*@xmath91=(@xmath920.4650 ) breaks the inversion symmetry , leaving two possible one - dimensional irreducible representations of the little ( magnetic ) group , which couple magnetic moments at the fe sites related by a 2@xmath98 twofold screw axis.@xcite in the most general case , the magnetic moment at a particular fe site is defined as @xmath99 here , the vector @xmath100 defines the origin of the @xmath56-th unit cell , @xmath101=1,2 identifies the crystallographically non - equivalent fe - sites , and @xmath102=1 - 4 denotes the four fe positions within the crystallographic unit cell ( for details see table[tab1 ] ) . the complex vector @xmath103 is determined by its real and imaginary components , @xmath104 and @xmath105 , which define the amplitude and the orientation of the magnetic moments , i.e. , the envelope of the magnetic cycloid / spiral , while @xmath106 denotes its phase shift . the magnetic wave vector * q * is in units of ( 2@xmath107,2@xmath108,2@xmath109 ) and @xmath106 in 2@xmath110 .
since the refinements considering only one irreducible representation did not converge , we took as a starting point the elliptical ic structure model used to describe the lt - ic phase in ftob.@xcite to avoid overparameterization of the problem we assume the same complex vector @xmath103@xmath111@xmath112 for all crystallographically equivalent fe sites .
as anticipated , the refinement yields a stable solution , which has strong sinusoidal modulation with @xmath113@xmath263@xmath114 and the dominant components ( * s*@xmath115 ) of the magnetic moments aligned close to the ( 1 @xmath85 0 ) orientation [ fig.[structure ] , table[tab1 ] ] , as in ftob .
however , the perpendicular components ( spanning the elliptical envelopes ) have a leading @xmath23 component , making the normal of the ellipse almost perpendicular to the one found in ftob ( fig.[structure ] ) , i.e. , @xmath116=75(4)@xmath2 . to test the reliability of our solution and its deviation from the ftob magnetic structure
, we performed additional refinement , assuming the same @xmath112 as in ftob and refining only magnetic phases .
the cost of the obtained solution has doubled , which indicates that the normal of the ellipse is indeed rotated compared to ftob . [ ! ] .components of vectors * s*@xmath117=(@xmath118,@xmath119,@xmath118 ) for @xmath120=re , im , defining the elliptical envelops for two independent magnetic atoms ( fe@xmath5 and fe@xmath0 ) for the best magnetic structure model at 6k , and eight magnetic phases @xmath106 in units of 2@xmath110 , i.e. , one for each of the magnetic fe@xmath121 atoms in the unit cell ( @xmath101=1,2 , @xmath102=1 - 4 ) .
the sites fe@xmath122-fe@xmath123 are obtained from fe@xmath94 [ @xmath124 , @xmath125 , @xmath126 and fe@xmath127-fe@xmath128 from fe@xmath95 [ @xmath129 , @xmath130 , @xmath131 by symmetry elements @xmath56 , @xmath132 and @xmath133 , respectively . the orientation of the moments is given in the @xmath134 coordinate system , while @xmath135@xmath884@xmath136 .
[ tab1 ] [ cols="^,^,^,^,^",options="header " , ] next , we collected 140 magnetic reflections at 11.2k , i.e. , in the ht - ic phase .
the best solution is again almost completely sinusoidal model , with magnetic moments pointing along the @xmath36 axis ( table[tab2 ] ) .
the shorter component of the elliptical envelope points here exactly along the @xmath23 axis , keeping the normal of the ellipse perpendicular to the one in ftob also in the ht - ic phase.@xcite we note that due to the limited amount of data and weak magnetic reflections , we assumed further simplification @xmath137@xmath111@xmath135
. once more , the refinements with a single irreducible representation do not converge .
finally , in contrast to ftob no systematic changes of magnetic phases @xmath106 were found , implying that electric polarization in ftoc might occur already in the ht - ic phase . [ ! ]
= 10kv / cm applied along @xmath19 and @xmath23 axes and changes of ( b ) the corresponding dielectric constants measured in ac field oscillating with frequency @xmath138=10khz .
inset : an anomaly in the @xmath34 for @xmath139 measured at two different @xmath138 demonstrating its frequency independence.,title="fig:",scaledwidth=50.0% ] to test the ferroelectric response of ftoc , we measured temperature dependences of the dielectric constant @xmath140 and the electric polarization , @xmath141 , on the single crystal samples in electric fields , @xmath142 , applied along each of the three crystallographic axes ( @xmath19,@xmath36,@xmath23 ) . on cooling from room temperature , @xmath143 ( @xmath144 )
first displays a clear anomaly at @xmath145=50.2(1)k [ inset in fig.[pol](a ) ] , which is close to the maximum in the magnetic susceptibility@xcite that corresponds to the onset of short - range magnetic correlations , also denoted by the broadening of the nmr line .
no further dielectric anomalies were found until @xmath24 , where @xmath146 exhibits a clear step , which is followed by a pronounced @xmath37-type anomaly at @xmath25 [ fig.[pol]b ] .
in contrast , @xmath147 reveals no anomaly at @xmath24 , but has a broad ( @xmath263k ) @xmath37-type anomaly centered at @xmath26(@xmath148k ) . in spite of the fact that @xmath146 exhibits anomaly already at @xmath24 , finite net electric polarization for @xmath149 emerges only below @xmath25 ( fig.[pol ] ) . the response for @xmath150 , however , is not so clear , as the corresponding @xmath141 builds up already 0.5 - 1k above @xmath25 .
nevertheless , below @xmath25 , @xmath151 for @xmath150 and @xmath149 exhibits a typical order - parameter - like behavior , reaching a saturation at @xmath268k .
the maximum values for both field orientations are the same @xmath141=2.8(1)@xmath28c / m@xmath152 and amount to roughly one third of the value acquired in ftob .
we note that for @xmath153 @xmath141=0 was found in the entire temperature region . finally , we inspected the response of the crystal lattice by performing synchrotron x - ray diffraction on a single - crystal sample . to refine the crystal structure successfully , high - temperature ( paramagnetic ) crystal symmetry ( space group @xmath1412@xmath5/@xmath23 )
was considered at all temperatures , hence all the deduced crystallographic changes are centrosymmetric and provide only indirect evidences of non - centrosymmetric ferroelectric distortions .
the changes of the lattice parameters at @xmath154 are very pronounced [ fig.[x - ray - all](a),(b ) ] , in contrast to ftob , where these effects were minute .
[ ! ] o@xmath9 ( @xmath10=br , cl ) compounds .
the error bars are smaller than the size of the symbols
. additional temperature dependences of ( c ) the interatomic distances and ( d ) the angles , exhibiting most pronounced changes in the vicinity of @xmath24 for the fete@xmath0o@xmath1cl ( ftoc).,title="fig:",scaledwidth=48.0% ] the most noticeable anomalies at @xmath25 are found for @xmath155 , @xmath156 and @xmath23 , which are all related to the interlayer spacing . on the other hand
, the temperature - independent @xmath36 appears to be insensitive to the magnetic ordering .
we stress that the crystal layers in this compound are formed due to te@xmath6 lone - pair cations.@xcite consequently , all changes in ftoc that are associated with the interlayer distance most likely reflect the shift of the te@xmath6 lone - pair electrons , which is also the most probable origin of the electric polarization.@xcite our refinement further enables the extraction of the interatomic distances and thus allows us to identify the atoms involved in the me effect .
we find that all significant changes close to @xmath25 are associated with the o@xmath5 atom . in fig.[x - ray - all](c ) we show temperature dependences of the distances between the o@xmath5 and the neighboring fe and te atoms that it binds ( fig.[o1-struc ] ) .
all other changes of the interatomic distances within the crystal layer are much smaller ( not shown ) , as they are below our experimental sensitivity .
the fact that the o@xmath5-fe@xmath0 and the o@xmath5-te@xmath8 distances exhibit significantly larger changes than the o@xmath5-fe@xmath5 one implies that the o@xmath5 atom is actually rotating around the fe@xmath5 one ( fig.[o1-struc ] ) .
this reflects also in the angles between the o@xmath5 and the neighboring o and fe atoms shown in fig.[x - ray - all](d ) .
in particular , the changes of the o@xmath5-fe@xmath0-o@xmath157 and the fe@xmath5-o@xmath5-fe@xmath0 angles are larger than that of the o@xmath5-fe@xmath5-o@xmath8 one , which further implies that the axis of the o@xmath5 rotation is close to the fe@xmath5-o@xmath8 direction .
since changes of other interatomic angles appear to be much smaller ( not shown ) , the observed rotation most likely applies to the o@xmath5 atom alone and does not involve the whole fe@xmath5-o octahedron .
we thus suggest that this rotation is a direct response to the lone - pair redistribution on the te@xmath8 ion . [ ! ] surrounding and assumed direction ( marked by black arrow ) of the o@xmath5 shift on heating through the @xmath25 transition.,title="fig:",scaledwidth=48.0% ]
at first sight , the only difference between the ftoc system investigated here and its isostructural relative ftob is the species of the halogen ion .
since halogen ions are located in - between the feteo layers , they mostly affect the interlayer distance .
the dominant intralayer magnetic exchange interactions in the two systems are thus expected to be comparable and should lead to similar magnetic orders .
nevertheless , the difference in the interlayer distance may significantly influence the me coupling , as te@xmath6 lone - pair electrons , which are involved in the fe - o - te - o - fe magnetic exchange pathway and most probably carry the electric polarization,@xcite protrude into the interlayer space.@xcite following , we summarize our main findings and discuss how they comply with the above assumptions .
first , we stress that we find a finite ferroelectric polarization and clear evidences of the me effect , in contrast to the prior far - infrared study@xcite that could not detect any link between electrodynamic response and magnetic ordering .
the observed dielectric response may be explained in the following way . at @xmath24 ,
small electric dipoles develop along the @xmath23 axis , inducing a concurrent step in @xmath143 [ fig.[pol](b ) ] .
however , since @xmath151 for @xmath149 is flat down to @xmath25 [ fig.[pol](a ) ] , these dipoles are most likely ordered antiparallelly , leading to the antiferroelectric polarization @xmath158 . when @xmath150 is applied , @xmath158 bends towards @xmath19 , resembling a finite net polarization already in the ht - ic phase [ fig.[pol](a ) ] . at @xmath25 ,
a `` metaelectric '' transition occurs , where antiferroelectrically ordered dipoles flop and form a ferroelectric arrangement pointing along @xmath23 , @xmath159 , reflecting as a rapid increase of @xmath141 for @xmath149 and a sharp anomaly in @xmath143 . on further cooling ,
@xmath151 for @xmath160 almost exactly follows the response for @xmath149 , suggesting that spontaneously induced electric polarization @xmath159 rotates towards @xmath19 when @xmath160 is applied . to address the magnetic behavior of the fete@xmath0o@xmath9 system , we first highlight both characteristic temperatures , i.e. , the curie - weiss ( @xmath27 ) and the nel ( @xmath24 ) temperature , which are in ftoc @xmath2625% higher than in ftob.@xcite in the case of a non - frustrated layered system , @xmath161 is related to both , intralayer as well as interlayer , exchange interactions,@xcite whereas @xmath27 is determined solely by dominant intralayer interactions . on the contrary , in frustrated systems
, @xmath161 is determined by main competing interactions@xcite and should thus scale proportionally to @xmath27 .
the fact that the substitution of the halogen ion , i.e. , the change of the interlayer distance , induces similar relative changes of both characteristic temperatures suggests that in ftoc / ftob frustration dominates and overcomes the layer - type effects .
this is in agreement with antiferromagnetic resonance and density functional theory study , revealing a quasi one - dimensional character of the system.@xcite nonetheless , the reduction of the interlayer distance may still cause higher - order effects and thus influence the magnetic anisotropy and/or the me coupling .
indeed , the two systems exhibit distinctly different magnetic anisotropies , as their flat elliptical envelopes are almost perpendicular to each other in both lro phases [ fig.[structure ] and tables[tab1],[tab2]].@xcite this further reflects in the magnetic - field dependence of @xmath25 , which is in both systems reduced when field is applied along @xmath19 ( fig.[cp]),@xcite but the effect is more significant in ftoc .
the reason for the observed response is that the magnetic field prefers antiferromagnetically ordered magnetic moments ( forming a cycloid ) to point perpendicular to it , making the lt - ic ordering with the dominant orientation of the magnetic moments along ( 1@xmath850 ) unfavorable.@xcite the larger decrease of @xmath25 in ftoc , on the other hand , complies with the orientation of the ht - ic elliptical envelope that is in ftoc perpendicular to @xmath19 , whereas in ftob the normal of the ellipse points towards @xmath23 .
such a pronounced anisotropic behavior is rather surprising , as the magnetic fe@xmath3 ion in the high - spin @xmath4=5/2 state has an almost isotropic orbital ground state ( @xmath162=0 ) , for which crystal - field anisotropy is expected to be small .
distinctly different anisotropies of ftoc and ftob , therefore , imply that the two systems are highly sensitive to the weak fe@xmath3 magnetic anisotropy .
this is most likely due to the geometric frustration , responsible for the complex magnetic ground state , which is probably very close to being degenerated and thus also very sensitive to small perturbations . due to the me coupling ,
the magnetic anisotropies are most likely responsible also for the discrepancy between the dielectric responses of the two systems .
in particular , the measured electric polarization in ftoc is in the lt - ic phase equal for @xmath150 and @xmath149 , whereas in ftob the response for @xmath150 is almost an order of magnitude weaker than for @xmath149 .
this may indicate that in ftoc the spontaneously induced net electric polarization @xmath163 can be easily rotated by the external electric field , while in ftob it is almost completely insensitive to it .
similarly , finite @xmath141 for @xmath150 above @xmath25 may result from a potentially existing antiferroelectric component @xmath158 , which is bend towards @xmath19 due to the applied electric field .
our results thus hint that the electric polarization in ftoc is significantly less rigid than in ftob and can be easily rotated within the @xmath97 plane by the external electric field , potentially affecting also the magnetic ordering .
this further complies with the magnetic - field dependence of @xmath25 , which indicates that anisotropies in ftoc are weaker than in ftob . finally , we look at the me coupling at the microscopic scale . in contrast to ftob ,
crystal - lattice changes at @xmath25 are in ftoc much more distinct and clearly indicate the shift of the o@xmath5 atom .
we stress here that o@xmath5 is bridging @xmath164 fe - o - fe as well as the @xmath165 fe - o - te - o - fe exchange interactions,@xcite from which @xmath164 was identified as the strongest interaction in ftob , while @xmath165 includes te@xmath8 the only te@xmath6 ion showing a noticeable shift in ftob.@xcite moreover , @xmath165 was found to have very strong influence on the electric polarization.@xcite since o@xmath5 is involved in a strong exchange interaction as well as it is attached to the te@xmath6 ion , which has easily polarizable lone - pair electrons , the crystallographic changes at @xmath25 ( fig.[x - ray - all ] ) suggest that the o@xmath5 atom is the source of the me effect in ftoc / ftob .
a broad set of experimental methods was used to investigate magnetic and dielectric response of the ftoc system .
we find that the system is multiferroic and is in many aspects very similar to its isostructural relative ftob .
in particular , it exhibits two ic magnetically ordered states with very similar elliptical cycloidal magnetic orders and nearly the same dominant orientations as in ftob . however , the cycloidal planes are substantially tilted , i.e. , the one in ftoc is almost perpendicular to the one in ftob .
since the only clear crystallographic difference between the two systems involves the interlayer distance , the observed response is ascribed to the geometric frustration .
the latter most likely amplifies the effects of otherwise subtle structural differences , which are probably responsible for the variation of the magnetic anisotropies as well as for the different responses of the electric polarization to the applied electric field . on the microscopic level
, however , the me mechanism seems to involve o@xmath5 , which bridges the main exchange interaction as well as it is attached to the te@xmath6 ion , which has easily polarizable lone - pair electrons that perturb into the interlayer spacing .
hence , lone - pair electrons may be affected by distortions of the exchange bridges through magnetostriction process as well as by a variation of the interlayer distance , which can thus both influence the response of the electric polarization .
finally , the observed persistent spin dynamics yet again confirms that this phenomenon is intrinsic to amplitude - modulated magnetic structures .
we acknowledge the financial support of the slovenian research agency ( project j1 - 2118 and z1 - 5443 ) and the swiss national science foundation ( project
200021 - 129899 ) .
@xmath28sr measurements at the isis facility have been supported by the european commission under the 7th framework program through the `` research infrastructures '' action of the `` capacities '' program , nmi3-ii grant no .
contract no .
the neutron diffraction work was performed at sinq , paul scherrer institute , villigen , switzerland .
cheong and m. mostovoy , nat .
* 6 * , 13 ( 2007 ) .
d. khomskii , physics * 2 * , 20 ( 2009 ) .
w. eerenstein , n. d. mathur , and j. f. scott , nature ( london ) * 442 * , 759 ( 2006 ) .
wang , j .-
liu , and z.f .
ren , adv . phys . * 58 * , 321 ( 2009 ) . , edited by c. lacroix , p. mendels , and f. mila ( springer - verlag , berlin , 2011 ) .
t. goto , t. kimura , g. lawes , a. p. ramirez , and y. tokura , phys .
* 92 * , 257201 ( 2004 ) .
n. hur , s. park , p. a. sharma , s. guha , and s - w .
cheong , phys .
lett . * 93 * , 107207 ( 2004 ) .
h. kimura , s. kobayashi , y. fukuda , t. osawa , y. kamada , y. noda , i. kagomiya , and k. kohn , j. phys .
. jpn . * 76 * , 074706 ( 2007 ) .
n. hollmann , z. hu , t. willers , l. bohat , p. becker , a. tanaka , h. h. hsieh , h .- j .
lin , c. t. chen , and l. h. tjeng , phys . rev . *
b 82 * , 184429 ( 2010 ) .
drechsler , n. tristan , r. klingeler , b. bchner , j. richter , j mlek , o. volkova , a. vasiliev , m. schmitt , a ormeci , c. loison , w. schnelle and h. rosner , j. phys . : condens .
matter * 19 * , 145230 ( 2007 ) .
r. becker , m. johnsson , r. k. kremer , h .- h .
klauss , and p. lemmens , j. am . chem . soc . * 128 * , 15469 ( 2006 ) . m. pregelj , o. zaharko , a. zorko , z. kutnjak , p. jegli , p. j. brown , m. jagodi , z. jaglii , h. berger , and d. aron , phys
lett . * 103 * , 147202 ( 2009 ) .
m. pregelj , a. zorko , o. zaharko , z. kutnjak , m. jagodi , z. jaglii , h. berger , m. de souza , c. balz , m. lang , and and d. aron , phys .
* b 82 * , 144438 ( 2010 ) . m. pregelj , a. zorko , o. zaharko , d. aron , m. komelj , a. d. hillier , and h. berger , phys . rev .
lett . * 109 * , 227202 ( 2012 ) .
m. pregelj , p. jegli , a. zorko , o. zaharko , t. apih , a. gradiek , m. komelj , h. berger , and d. aron , phys . rev . *
b 87 * , 144408 ( 2013 ) .
m. pregelj , h. o. jeschke , h. feldner , r. valent , a. honecker , t. saha - dasgupta , h. das , s. yoshii , t. morioka , h. nojiri , h. berger , a. zorko , o. zaharko , and d. aron , phys . rev . *
b 86 * , 054402 ( 2012 ) .
r. seshadri , and n. a. hill , chem .
* 13 * , 2892 ( 2001 ) .
f. pfuner , l. degiorgi , h. berger and l forr , j. phys . : condens .
matter * 21 * , 375401 ( 2009 ) .
a. youanc and p. dalmas de rotier , muon spin rotation , relaxation and resonance ( oxford university , new york , 2011 ) .
a. zorko , m. pregelj , h. berger , and d. aron , j. appl .
107 , 09d906 ( 2010 ) .
a. u. b. wolter , f. lipps , m. schpers , s .- l .
drechsler , s. nishimoto , r. vogel , v. kataev , b. bchner , h. rosner , m. schmitt , m. uhlarz , y. skourski , j. wosnitza , s. sllow , and k. c. rule , phys . rev .
* b 85 * , 014407 ( 2012 ) . s. dommange , m. mambrini , b. normand , and f. mila , phys . rev . * b 68 * , 224416 ( 2003 ) , a. olariu , p. mendels , f. bert , f. duc , j. c. trombe , m. a. de vries , and a. harrison , phys .
lett . * 100 * , 087202 ( 2008 ) . , i. rousochatzakis , s. r. manmana , a. m. luchli , b. normand , and f. mila , phys . rev .
* b 79 * , 214415 ( 2009 ) .
r. blinc , phys . rep . * 79 * , 331 ( 1981 ) .
c. brevard and p. granger , _ handbook of high resolution multinuclear nmr _ , ( wiley , new york , 1981 ) .
c. p. slichter , _ principle of magnetic resonance _
( springer - verlag , berlin , heidelberg , 1990 ) .
v. pokrovsky and g. uimin in _ magnetic properties of layered transition metal compounds _ , edited by l. de jongh ( kluwer , dordrecht , 1990 ) , p. 89 . c. yasuda , s. todo , k. hukushima , f. alet , m. keller , m. troyer , and h. takayama , phys .
lett . * 94 * , 217201 ( 2005 ) . | the layered fete@xmath0o@xmath1cl compound was studied by specific - heat , muon spin relaxation , nuclear magnetic resonance , dielectric , as well as neutron and synchrotron x - ray diffraction measurements , and the results were compared to isostructural fete@xmath0o@xmath1br .
we find that the low - temperature ordered state , similarly as in fete@xmath0o@xmath1br , is multiferroic the elliptical amplitude - modulated magnetic cycloid and the electric polarization simultaneously develop below 11k .
however , compared to fete@xmath0o@xmath1br , the magnetic elliptical envelop rotates by 75(4)@xmath2 and the orientation of the electric polarization is much more sensitive to the applied electric field .
we propose that the observed differences between the two isostructural compounds arise from geometric frustration , which enhances the effects of otherwise subtle fe@xmath3 ( @xmath4=5/2 ) magnetic anisotropies .
finally , x - ray diffraction results imply that , on the microscopic scale , the magnetoelectric coupling is driven by shifts of the o@xmath5 atoms , as a response to the polarization of the te@xmath6 lone - pair electrons involved in the fe - o - te - o - fe exchange bridges . |
the maximum magnetic field @xmath5 which can be screened by a superconductor in the vortex - free meissner state has attracted much attention , both from the fundamental and applied perspectives @xcite .
this problem is particularly important for the nb resonator cavities @xcite which have extremely high quality factors @xmath6 up to the breakdown fields @xmath7 mt at which the screening current density @xmath8 approaches the depairing limit @xcite , @xmath9 , where @xmath10 mt is the thermodynamic critical field of nb and @xmath11 nm is the london penetration depth .
the lack of radiation losses and vortex dissipation in the nb cavities ( clean nb has the highest lower critical field @xmath12 mt among type - ii superconductors ) enables one to probe the high - field nonlinear quasiparticle conductivity @xcite which can be tuned by alloying the surface with impurities @xcite .
the screening field limit of nb could be exceeded by using s - wave superconductors with higher @xmath13 and the critical temperature @xmath14 , but such materials are prone to the dissipative penetration of vortices at @xmath15 . to address this problem
it was proposed to coat the nb cavities with multilayers of thin superconductors ( s ) with high @xmath16 separated by dielectric ( i ) layers @xcite ( see fig .
this approach is based on the lack of thermodynamically stable parallel vortices in decoupled s screens of thickness @xmath17 , which also manifests itself in a strongly enhanced @xmath18 in thin films predicted by abrikosov @xcite and observed in refs . .
the maximum field @xmath19 screened by @xmath20 superconducting layers of thickness @xmath21 is limited by the superheating field @xmath0 of s - layers @xcite , for example , @xmath22 mt for nb@xmath4sn at @xmath23 . here
the meissner screening currents at @xmath24 become unstable with respect to infinitesimal perturbations of electromagnetic field and the order parameter , while the magnetic barrier for penetration of vortices vanishes @xcite .
this paper addresses the limits to which the maximum screening field can be increased by s layers with given @xmath25 , @xmath26 and @xmath0 deposited onto a thick nb substrate with given @xmath27 and @xmath28 .
it is shown that : 1 .
the maximum @xmath19 can be reached at an optimum multilayer thickness @xmath29 which depends on the materials parameters of s layers and the substrate ; 2 .
the optimized s - i - s multilayer can screen the field exceeding both superheating fields , @xmath0 and @xmath28 ; 3 .
s - i - s multilayers arrest thermomagnetic avalanches caused by local penetration of vortices at defects and do not let them develop into global flux jumps , which otherwise quench the cavity at fields well below @xmath19 .
distribution of the magnetic field @xmath30 in a multilayer of thickness @xmath31 .
the dashed lines show thin dielectric layers @xmath32 separating the superconducting layers .
, width=230 ]
distribution of a low - frequency @xmath33 rf magnetic field @xmath30 in a multilayer can be described by the london equations @xmath34 at @xmath35 and @xmath36 at @xmath37 , where the prime denotes differentiation with respect to @xmath38 . given that @xmath30 in a multilayer calculated numerically from the eilenberger equations is close to the result of the london theory @xcite , i first disregard the nonlinear meissner effect @xcite which makes @xmath26 dependent on @xmath8 at @xmath39 .
solutions of the london equations are ( see also refs . ):
@xmath40 , \quad 0<x < d , \label{e1 } \\
h_2(x)=bhe^{(d - x)/\lambda_0 } , \qquad x > d , \label{e2 } \\
c=\frac{k}{k+g^2},\qquad\quad b=\frac{(1+k)g}{k+g^2}. \label{const}\end{gathered}\ ] ] here @xmath30 and the rf electric field @xmath41 are continuous at @xmath42 , @xmath43 , and @xmath44 , both for a single s film and for a stack of s - i layers with @xmath45 . for a s - i bilayer ,
the low - field surface resistance @xmath46 is determined by the total joule rf power , @xmath47 , where @xmath48 and @xmath49 are the quasiparticle conductivities in s layers and the substrate , respectively , and @xmath50 accounts for dielectric losses .
using eqs .
( [ e1])-([const ] ) , and @xmath51 in the local dirty limit@xcite yields @xcite @xmath52 \nonumber \\
+ \frac{r_{s0}g^2(1+k)^2}{(k+g^2)^2}+\frac{4\mu_0 ^ 2d_i\omega^3\epsilon''\lambda^2\lambda_0 ^ 2g^2}{(\lambda+\lambda_0)^2(k+g^2)^2 } , \label{rs}\end{gathered}\ ] ] where @xmath53 and @xmath54 include the residual resistances @xcite , and @xmath55 is the complex dielectric permeability of a low - loss i layer with @xmath56 . for @xmath57 , the constant @xmath58 in eqs .
( [ e1])-([const ] ) is positive so the current density @xmath59 at the surface is lowered by the counterflow induced by the substrate @xcite .
the meissner state in s layers and the substrate is stable with respect to infinitesimal perturbations @xcite if the current densities are smaller than the respective depairing limits , @xmath60 and @xmath61 .
these conditions define the field region of the vortex - free state : @xmath62 shown in fig .
2 is the @xmath63 diagram where the meissner region is stable below both curves defined by eqs .
( [ ineq ] ) . if @xmath64 , thin s layers with @xmath65 are stable but do not screen the magnetic field which is thus limited by @xmath28 of the substrate . as @xmath31 increases
, @xmath66 increases until @xmath31 reaches the crossing point @xmath67 at which the maximum field @xmath19 exceeds both @xmath0 and @xmath28 .
the latter results from the stabilizing effect of counterflow on @xmath68 , where @xmath69 given by eq .
( [ const ] ) decreases as @xmath31 increases . as a result
, @xmath66 decreases with @xmath31 at @xmath70 , approaching the superheating field of the s layer at @xmath71 .
the optimal thickness @xmath29 is defined by the equalities in eq .
( [ ineq ] ) which give a quadratic equation for @xmath72 .
hence , @xmath73 where @xmath74 .
substituting @xmath29 back to one of eqs .
( [ ineq ] ) yields the maximum screening field @xmath75^{1/2}. \label{hm}\ ] ] at @xmath76 and @xmath77 , both current densities at @xmath78 and @xmath42 reach the depairing limits for the respective materials , but @xmath19 in eq .
( [ hm ] ) exceeds both @xmath0 and @xmath28 if @xmath79 , because it is the current density @xmath80 but not the magnetic field which makes the meissner state unstable @xcite .
the meissner region lying below both curves in fig .
2 is similar to that was calculated numerically in the london model assuming that the energy barrier for the vortex parallel to the surface vanishes @xcite . the london approach@xcite used in refs .
has the well - known limitations , particularly at the fields @xmath81 for which the bean - livingston barrier is only a few vortex core diameters away from the surface . as a result
, the london model can not be used for the calculations of the superheating field , so the calculations of refs . only give the right order of magnitude for @xmath19 but
can hardly be used for quantitative evaluation of @xmath19 . equations ( [ ineq ] ) obtained from the condition of instability of meissner currents have no ambiguity resulting from the evaluation of the surface barrier in the london model with the rigid core cutoff @xcite which fails at @xmath81 because of strong deformation of the vortex core at the surface @xcite .
calculations based on the ginzburg - landau ( gl ) theory have shown that the instability at @xmath24 is caused by lateral perturbations with the wavelength @xmath82 decaying over the length @xmath83 perpendicular to the surface , where @xmath84 is the coherence length @xcite . at @xmath85 ,
the theory gives @xmath86 where @xmath87 decreases as the gl parameter @xmath88 increases @xcite , so that @xmath89 at @xmath90 and @xmath91 at @xmath92 . at @xmath23 ,
only the limiting value @xmath93 for @xmath92 has been calculated @xcite .
magnetic phase diagram of a multilayer .
meissner state is stable below both solid lines calculated from eqs .
( [ ineq ] ) for @xmath94 and @xmath95 . here @xmath96 at @xmath97 , and the crossing point @xmath67 defines the optimum thickness @xmath29 at which @xmath66 is maximum . ,
width=238 ] taking @xmath98 nm and @xmath99 mt in eq .
( [ hm ] ) for a dirty nb@xmath4sn ( ref . )
on nb with @xmath100 mt and @xmath101 nm , yields @xmath102 mt at @xmath103 nm .
here the meissner state persists up to the field @xmath104 higher than @xmath0 of nb@xmath4sn , consistent with the london model of ref . .
for nbn films @xcite with @xmath105 nm and @xmath106 nm , we obtain @xmath107 nm and @xmath108 288 mt , while for pnictides@xcite with @xmath109 nm and @xmath110 t , eqs .
( [ dm])-([hm ] ) give @xmath111 mt and @xmath112 mt at @xmath113 nm . the case of the two - band superconductor mgb@xmath114 with @xmath115 mt is more complicated as the rf dissipation can occur at a smaller field @xmath116 at which the screening current decouples two weakly coupled @xmath48 and @xmath117 bands @xcite . for the typical ratio of coherence lengths , @xmath118 , the band decoupling field
@xmath119 mt is consistent with the rf breakdown field of 42 mt at 4 k observed on mgb@xmath114/al@xmath114o@xmath4 bilayers on the nb substrate @xcite .
the enhancement of @xmath19 does not necessarily require i layers because eqs .
( [ dm])-([hm ] ) also describe a dirty layer at the surface where @xmath26 is increased due to a shorter mean free path @xmath120 . for instance , a dirty nb layer with @xmath121 nm has @xmath122 nm and @xmath123 nm . using @xmath124 for @xmath125 , yields @xmath126 nm and @xmath127 mt , the same as for the above case of nbn and some @xmath128 higher than @xmath100 mt of pure nb . in the limit of @xmath129 , eq . ( [ hm ] ) gives @xmath130 mt .
to see how the nonlinear meissner effect can change eq .
( [ hm ] ) , @xmath19 is also calculated from the gl theory at @xmath92 for which the @xmath131 component of the magnetic vector potential @xmath132 satisfies the equation @xcite @xmath133 consider a dirty layer with constant @xmath26 and @xmath84 at @xmath35 , a substrate with constant @xmath27 and @xmath134 , and the same @xmath14 in both regions .
the first integrals of eq .
( [ gl ] ) are : @xmath135 at @xmath35 and @xmath136 at @xmath37 .
continuity of @xmath132 and @xmath137 at @xmath42 , the boundary condition @xmath138 , and the gl pairbreaking conditions @xmath139 and @xmath140 under which the meissner superflow becomes unstable @xcite yield @xmath141 at @xmath78 , and @xmath142 ^ 2 - 1/3\xi_0 ^ 2+\xi^2/18\xi_0 ^ 4=c_1 $ ] and @xmath143 ^ 2=5/18\xi_0 ^ 2 $ ] at @xmath42 .
here @xmath144 , and @xmath145 is the magnetic flux quantum . excluding @xmath146 and @xmath147 and solving for @xmath19 reproduces eq .
( [ hm ] ) in which @xmath148 , and @xmath149 here @xmath150 , and both @xmath14 and @xmath151 are independent of the mean free path , according to the anderson theorem @xcite .
two last terms in the parenthesis of eq .
( [ hsm ] ) resulting from the nonlinear meissner effect give a negligible contribution to @xmath19 for a dirty layer with @xmath152 .
the maximum screening field @xmath19 at which the messier state becomes absolutely unstable with respect to infinitesimal perturbations can hardly be reached under realistic operating conditions which require that the accelerating resonator cavities remain stable with respect to penetration of vortices , strong transient electromagnetic perturbations of charged beams , and local field enhancement at surface defects . in the multilayer approach
@xcite i layers are instrumental to assure the necessary stability margin with respect to local penetration of vortices at surface defects , which can otherwise trigger thermomagnetic flux jumps @xcite particularly at low temperatures @xmath23 and extremely high screening currents @xmath39 at which the cavities operate .
misinterpretation of this issue has lead to speculations that neither i layers nor the enhancement of @xmath153 is necessary , so a few @xmath154 m thick nb@xmath4sn film coating of the nb cavities could just be protected against penetration of vortices by the bean - livingstron surface barrier @xcite .
this assumption contradicts a vast body of experimental data on magnetization of high-@xmath155 type - ii superconductors for which inevitable materials or topographic defects at the surface cause premature local penetration of vortices at @xmath156 , or even @xmath157 due to grain boundary weak links @xcite .
the maximum screening field @xmath19 at which the meissner state is stable with respect to penetration of vortices can be evaluated from eqs .
( [ dm])-([hm ] ) with effective @xmath0 and @xmath28 depending on the operating conditions .
for instance , in a multilayer with @xmath158 mt at @xmath76 , a vortex entering through a defect in the s layer can not penetrate further into the bulk nb .
let a surface defect cause local penetration of vortices as the current density , @xmath159 reaches a fraction @xmath160 of @xmath161 in s layer with @xmath17 and @xmath162 .
if @xmath163 is not too close to @xmath161 , the london theory shows @xcite that the energy barrier @xmath164 per unit length of a vortex in the s layer at @xmath165 coincides with the bulk bean - livingston surface barrier @xcite at @xmath166 and @xmath167 , where @xmath168 . the criterion @xmath169 assures a reasonable protection against penetration of vortices caused by low - angle grain boundaries in polycrystalline nb@xmath4sn or pnictides @xcite , or local field enhancement at typical topographic defects in the nb cavities @xcite . at @xmath170 and @xmath171 mt , eqs .
( [ dm])-([hm ] ) give @xmath172 mt at @xmath173 nm for nb@xmath4sn .
if nb can withstand the field @xmath174 mt observed on the best cavities@xcite , the maximum screening field could reach @xmath175 mt at @xmath176 nm .
increasing @xmath177 by the materials refinements of nb@xmath4sn could push the peak fields over 300 mt .
pnictides with @xmath178 t , such as ba@xmath179k@xmath180fe@xmath114as@xmath114 ( ref . ) could provide @xmath181 mt at @xmath182 nm , @xmath183 , and @xmath184 mt . propagating vortex loop ( a ) turning into the vortex - antivortex pair ( b ) in s layer with a surface defect ( black dot ) which lowers the field threshold of vortex penetration.,width=6 ] to inhibit dissipative penetration of vortices , s film can be subdivided by i layers into @xmath20 thinner s layers with @xmath185 . at @xmath186 ,
even if a vortex penetrates at a defect in the first s layer , it could not propagate into the next s layer and further in the bulk nb where it can cause a thermomagnetic avalanche @xcite .
as @xmath187 reaches the critical value @xmath188 at a week spot , a vortex line can not penetrate parallel to the surface but first originates at a defect as a small semi - loop as depicted in fig .
the vortex semi - loop expands under the action of the perpendicular meissner current until it hits the i layer where most part of the dissipative vortex core disappears in a loss - free flux channel connecting two short vortices of opposite polarity .
because of the magnetic flux compression in the i layer and the substrate , the energy of a perpendicular vortex @xmath189 diverges with the lateral film size @xmath190 , while the energy of the vortex - antivortex pair @xmath191 grows with the distance @xmath192 , where @xmath193 ( ref . ) .
this vortex - antivortex pair expands during the positive rf cycle and contracts and annihilates as @xmath187 changes sign .
the upper limit of the pair size @xmath194 can be estimated neglecting the long - range vortex - antivortex attraction described by the last term in the dynamic equation , @xmath195 , where @xmath196 is the viscous drag coefficient , and @xmath197 is the normal state resistivity . hence , @xmath198 at @xmath199 and @xmath200 , giving @xmath201 m and the rf power @xcite @xmath202w for nb@xmath4sn at @xmath203 m , @xmath204 nm , @xmath183 , @xmath205 , @xmath206 , @xmath207 ghz and @xmath208 .
taking into account attraction of antiparallel vortices in s layer reduces @xmath194 and @xmath209 .
penetration of vortex semi - loops at a defect appears more realistic than the model of a long vortex parallel to the surface for which the ill - defined notion of @xmath210 associated with the magnetic flux trapped in i layer @xcite is irrelevant to the rf dissipation @xcite .
localization of the rf power in a thin s layer inhibits expansion of multiple vortex loops in the bulk and stops dendritic thermomagnetic avalanches @xcite that are particularly pronounced at the extremely high screening current densities @xmath39 and low temperatures @xmath23 in the materials like nb@xmath4sn , nbn or pnictides with low @xmath211 and thermal conductivity @xmath212 .
the multilayer thus significantly reduces vortex dissipation as compared to the bulk nb@xmath4sn , yet a thin nb@xmath4sn layer with @xmath213 nm may only slightly increase the thermal impedance of the cavity wall , @xmath214 , where @xmath215 is the kapitza thermal resistance . for @xmath216 mm , @xmath217 w / mk , @xmath218 kw / m@xmath219k , the nb@xmath4sn multilayer with @xmath220 nm , @xmath221 w / mk , and al@xmath114o@xmath4 dielectric layers with @xmath222 nm and @xmath223 w / mk ( ref . )
increases @xmath224 by only @xmath225 .
a thicker nb@xmath4sn film with @xmath226 m doubles @xmath224 and reduces the field of thermal quench @xcite , in addition to the smaller @xmath227 mt of bulk nb@xmath4sn with @xmath228 nm ( ref . ) as compared to @xmath229 mt .
theoretically , i layers provide the strongest pinning of propagating vortices and stop them more efficiently than holes in thin films which have been used to terminate dendritic flux avalanches @xcite
. the optimum number of s layers for particular materials is determined by a balance between reduced vortex dissipation and suppression of superconductivity at the s - i interfaces . here
@xmath0 of ideal s layers with @xmath230 remains the same as in the bulk @xcite , contrary to the assertion @xcite that @xmath0 is reduced at small @xmath25 .
this claim was based on the artifacts of the london model discussed above and on taking into account only one right vortex image in fig .
4 of ref . instead of summing up an infinite chain of vortex - antivortex image dipoles which ensure that vortex currents do not cross the film surface .
if this effect is properly taken into account @xcite @xmath0 in a thin film ( @xmath231 ) is the same as in a thick film ( @xmath232 ) .
moreover , the gl simulations of ref . for a nb@xmath4sn film on nb show that @xmath233 reaches the maximum @xmath234 at @xmath235 but remains larger than @xmath236 in the whole region @xmath237 for which the numerical results were presented .
high - field rf performance of the nb cavities can be boosted by depositing not only materials with higher @xmath0 but also alloyed nb - i - nb multilayers which can increase @xmath19 and benefit from a significant raise of @xmath238 with @xmath5 in a wide field region @xcite .
a polycrystalline nb multilayer may be tuned by alloying and heat treatment to reduce the residual resistance @xcite , and is also less prone to the current - blocking grain boundaries than the a-15 or pnictide compounds @xcite .
enhancement of the vortex penetration field by a dirty nb / al@xmath114o@xmath4 bilayer deposited onto the nb cavity was observed in ref . . in conclusion
, optimized multilayers can significantly increase the meissner screening field while inhibiting dissipative penetration of vortices .
implementation of such multilayer coatings could potentially double the accelerating field gradients of superconducting resonators as compared to the existent high - performance nb cavities .
t. tajima , n.f .
haberkorn , l. civale , r.k .
schulze , h. inoue , j. guo , v.a .
dolgashev , d. martin , s. tantawi , c. yoneda , b. moeckly , c. yung , t. proslier , m. pellin , a. matsumoto , and e. watanabe , aip conf . proc . * 1435 * , 297 ( 2012 ) .
beringer , c. clavero , t. tan , x.x .
roach , and r.a .
lukaszew , ieee trans .
. supercond .
* 23 * , 75006 ( 2013 ) ; w.m .
roach , d.b .
beringer , z. li , c. clavero , and r.a lukaszew , ieee trans .
. supercond .
* 23 * , 86002 ( 2013 ) .
j. matricon and d. saint - james , phys .
a * 24 * , 241 ( 1967 ) ; j. chapman , siam j. appl .
math . * 55 * , 1233 ( 1995 ) ; a.j .
dolgert , s.j .
di bartolo , and a.t .
dorsey , phys .
b * 53 * , 5650 ( 1996 ) ; * 56 * , 2883 ( 1997 ) ; m.k .
transtrum , g. catelani , and j.p .
sethna , phys .
b * 83 * , 094505 ( 2011 ) .
s.k . yip and j.a .
sauls , phys .
69 * , 2264 ( 1992 ) ; d. xu , s.k . yip , and j.a .
sauls , phys .
b * 51 * , 16233 ( 1995 ) ; m .- r .
hirschfeld , and p. wlfle , phys .
lett . * 81 * , 5640 ( 1998 ) ; phys .
b * 61 * , 648 ( 2000 ) ; n. groll , a. gurevich , and i. chiorescu , phys.rev .
b * 81 * , 020504(r ) ( 2010 ) .
@xmath46 of ref .
reduces to eq .
( [ const ] ) at @xmath239 and @xmath240 , but is invalid at @xmath241 as the boundary condition of the continuity of @xmath242 at @xmath42 was not taken into account @xcite .
u. welp , c. chaparro , a.e .
koshelev , w.k .
kwok , a. rydh , n.d .
zhigadlo , j. karpinski , and s. weyeneth , phys .
b * * 8**3 , 100513(r ) ( 2011 ) ; s.l .
budko , m. sturza , d.y .
chung , m.g .
kanatzidis , and p.c .
canfield , phys .
b * * 8**7 , 100509(r ) ( 2013 ) . c. martin , r. t. gordon , m.a .
tanatar , h. kim , n. ni , s.l .
budko , p.c .
canfield , h. luo , h.h .
wen , z. wang , a.b .
vorontsov , v.g .
kogan , and r. prozorov , phys .
b * 80 * , 020501(r ) ( 2009 ) ; y.m .
dai , b. xu , b. shen , h.h .
wen , x.g .
qiu , and r.p.s.m .
lobo , europhys .
lett . * 104 * , 47006 ( 2013 ) .
m. perpeet , m.a .
hein , g.mller , h. piel , j. pouryamout , and w. diete . j. appl
. phys . * 82 * , 5021 ( 1997 ) ; a. andreone , a. cassinese , a. di chiara , m. iavarone , f. palomba , a. ruosi , j. apple . phys . *
82 * , 1736 ( 1997 ) ; and r. vaglio.m .
deambrosis , g. keppel , v. ramazzo , c. roncolato , r.g .
sharma , and v. palmieri , physica c * 441 * , 108 ( 2006 ) .
oates , a.c .
anderson , c.c .
chin , j.s .
derov , g. dresselhaus , and m.s .
dresselhaus , phys .
b * 43 * , 7655 ( 1991 ) ; m. s. pambianchi , s. m. anlage , e. s. hellman , e. h. hartford , jr .
, m. bruns , and s. y. lee , appl .
64 * , 244 ( 1994 ) ; b. komiyama , z. wang , and m. tonouchi , appl .
. lett . * 68 * , 562 ( 1996 ) .
i. aranson , a. gurevich , and v. vinokur , phys .
lett . * 87 * , 0670031 ( 2001 ) ; i.s .
aranson , a. gurevich , m.s .
welling , r.j .
wijngaarden , v.k .
vlasko - vlasov , v.m .
vinokur , and u. welp , phys .
lett . * 94 * , 0370021 ( 2005 ) ; j. i. vestgarden , d. shantsev , y. galperin , and t.h .
johansen , phys .
b. * 84 * , 054537 ( 2011 ) ; p. mikheenko , a.j .
qviller , j.i .
vestgarden , s. chaudhuri , i.j .
maasilta , y.m .
galperin , and t.h .
johansen , appl .
lett . * 102 * , 022601 ( 2013 ) ; | it is shown that a multilayer comprised of alternating thin superconducting and insulating layers on a thick substrate can fully screen the applied magnetic field exceeding the superheating fields @xmath0 of both the superconducting layers and the substrate , the maximum meissner field is achieved at an optimum multilayer thickness .
for instance , a dirty layer of thickness @xmath1 m at the nb surface could increase @xmath2 mt of a clean nb up to @xmath3 mt . optimized multilayers of nb@xmath4sn , nbn , some of the iron pnictides , or alloyed nb deposited onto the surface of the nb resonator cavities could potentially double the rf breakdown field , pushing the peak accelerating electric fields above 100 mv / m while protecting the cavity from dendritic thermomagnetic avalanches caused by local penetration of vortices . |
based on the measurements of @xmath8 and @xmath9 channels the atlas and cms collaborations have independently provided compelling evidence for a bosonic resonance around 125 - 126 gev @xcite .
this is a great triumph for particle physics , but it also leads to a host of new questions about the nature of the boson .
so far there exist large uncertainties in determining the rates of the two channels , and meanwhile , observation of the boson through other signals such as @xmath10 and @xmath11 channels is still far away from becoming significant@xcite .
so although the preliminary data of the lhc indicate that the boson closely resembles the higgs boson in the standard model ( sm ) , the deficiencies of the sm itself such as gauge hierarchy problem suggest new physics explanation of the boson . obviously , in order to decide the right underlying theory , the lhc should exhaust its potential to measure the decay channels of the boson as accurately as possible in its high luminosity phase . among the decay modes of the higgs - like boson @xmath12 ,
the diphoton channel plays a very important role in determining its mass , spin and parity@xcite . at the same time , since the diphoton channel is mediated by loops of charged particles , it also acts as a sensitive probe to new physics .
in fact , this feature has been widely utilized to explain the results of the atlas and cms collaborations on the inclusive diphoton signal in 2012@xcite , which were @xmath13 and @xmath14 respectively for the signal rate normalized by its sm prediction @xcite . in this note , we concentrate on another decay mode @xmath15 . in the sm ,
the branching ratio of this decay is about two thirds of that for the diphoton decay , and just like the diphoton signal , it can provide a clean final - state topology in determining the properties of the boson , such as its mass , spin and parity@xcite .
moreover , since new charged particles affecting the diphoton decay can also contribute to the @xmath5 decay , the two decay modes should be correlated , and therefore studying them in a joint way can reveal more details about the underlying physics .
albeit the advantages , in contrast to the diphoton decay which has been intensively studied , the @xmath5 decay was paid little attention in the past . for example , since the discovery of the boson only several works have been devoted to this decay in new physics models such as the type - ii seesaw model@xcite , the georgi - machacek model@xcite , the extensions of the sm by charged scalars in different @xmath16 representations@xcite and the sm with extra colored scalars@xcite , and until very recently have the cms and atlas collaborations set an upper limit on the ratio @xmath17 @xcite . note that although the @xmath5 signal suffers from a large irreducible background at the lhc @xcite , the higgs event from the process @xmath18 can be easily reconstructed at the next generation linear collider with the center of mass energy around 250 gev @xcite , which is very helpful in suppressing the background for such a signal .
so there is a good prospect to precisely measure this decay in the future . in the following ,
we focus on the @xmath19 decay channel of the sm - like higgs boson @xmath12 in low energy supersymmetric models such as the constrained mimimal supersymmetric standard model ( cmssm)@xcite , the minimal supersymmetric standard model ( mssm)@xcite , the next - to - minimal supersymmetric standard model ( nmssm)@xcite and the nearly minimal supersymmetric standard model ( nmssm)@xcite .
we investigate the @xmath2 signal of the higgs production at the lhc and future @xmath4 linear colliders , and especially , we study its correlations with the @xmath3 signal
. as we will show below , the @xmath5 signal rate may be either enhanced or suppressed in susy , and its correlation behavior is so model - dependent that it may be utilized to distinguish the models in high luminosity phase of the lhc .
this work is organized as follows . in section
ii we introduce the basic features of the susy models and present some formulae relevant to our calculation . in section iii
we first discuss the effects of new charged susy particles on the partial decay widths of @xmath20 , then we study in a comparative way the @xmath2 and @xmath3 signal rates of the higgs production at different colliders .
finally , we draw the conclusions in section iv .
various couplings used in the calculation are given in the appendix .
in a low energy supersymmetric gauge theory , the explicit form of its lagrangian is determined by the gauge symmetry , superpotential and also soft breaking terms . as for the four models considered in this work
, their differences mainly come from the superpotential , which is the source for the yukawa interactions of fermions and self interactions of scalars .
@xmath21 the mssm@xcite contain two higgs doublets @xmath22 and it predict five physical higgs bosons , of which two are cp - even , one is cp - odd and two are charged .
its superpotential takes following form @xmath23 where @xmath24 denotes the yukawa interaction , and its form is given by @xmath25 after considering appropriate soft breaking terms , one can write down the higgs potential as @xmath26 where @xmath27 , @xmath28 and @xmath29 are all soft parameters with mass dimension , terms proportional to @xmath30 come from the @xmath31-term of the superpotential , and the last term comes from gauge symmetry ( so called @xmath32-term ) .
this potential indicates that , after the electroweak symmetry breaking , the @xmath33-parameter is related to the higgs vacuum expectation value ( vev ) and so it should be @xmath34 .
but on the other hand , since @xmath33 as the only parameter with mass dimension appears in the superpotential , its value should naturally take susy - preserving scale .
such a tremendously large scale gap is usually referred as the @xmath33-problem @xcite .
the theoretical framework of the cmssm@xcite is exactly same as that of the mssm , and the only difference between them comes from the fact that in the general mssm , all soft breaking parameters are independent@xcite , while in the cmssm they are correlated . explicitly speaking
, the cmssm assumes following universal soft breaking parameters at susy breaking scale ( usually chosen at the grand unification scale)@xcite @xmath35 with @xmath36 , @xmath37 and @xmath38 denoting gaugino mass , scalar mass and trilinear interaction coefficient respectively , and evolves the four parameters down to weak scale to get all the soft breaking parameters of the mssm . in this sense
, the parameter space of the cmssm should be considered as a subset of that for the mssm , and so is its phenomenology .
@xmath39 in order to solve the @xmath33-problem in the mssm , various singlet extensions of the mssm were proposed in history , and among them the most well known models include the nmssm and the nmssm .
the superpotentials of these two models are respectively given by @xcite @xmath40 where @xmath41 , @xmath42 and @xmath43 are dimensionless parameters of order 1 , and the dimensionful parameter @xmath44 may be naturally fixed at weak scale in certain basic frameworks where the parameter is generated at a high loop level@xcite
. one attractive feature of both the models comes from the fact that , after the real scalar component of @xmath45 develops a vev @xmath46 , an effective @xmath33 parameter is generated by @xmath47 , and its value may be as low as about @xmath48 without conflicting with current experiments @xcite ( in contrast , the @xmath33 parameter in the mssm must be larger than about @xmath49@xcite ) .
another attractive feature of the models is that the z boson mass may be obtained with less fine tuning than the mssm@xcite . in the susy models , after the minimization of the higgs potential , the @xmath50 boson mass is given by@xcite @xmath51 where @xmath52 and @xmath53 represent the soft susy breaking masses of the higgs fields , and @xmath54 and @xmath55 arise from the radiative corrections to the higgs potential with dominant contribution to @xmath54 given by @xmath56 these two equations indicate that , if the individual terms on the right hand side of eq .
( [ minimization ] ) are comparable in magnitude so that the observed value of @xmath57 is obtained without resorting to large cancelations , relatively light stops and @xmath33 of @xmath34 are preferred .
as far as the nmssm and the nmssm are concerned , due to the higgs self interactions , the squared mass of the sm - like higgs boson gets an additional contribution @xmath58 ( compared with its mssm expression ) , and further it can be enhanced by the doublet - singlet mixing @xcite .
consequently , predicting a @xmath59 higgs boson does not necessarily require heavy scalar top quarks@xcite .
this is very helpful in reducing the tuning .
for example , it has been shown that the fine tuning parameter @xmath60 , which is defined by @xmath61 with @xmath62 denoting susy parameter at gut scale @xcite ) , may be as low as 4 in the two models , while in the mssm it usually exceeds 100 @xcite .
about the nmssm , one should note that the tadpole term in eq.([nmssm ] ) only affects the higgs masses and the minima of the scalar potential , so the interactions in the higgs and neutralino sectors of the nmssm are identical to those in the nmssm with @xmath63 .
this enables us to modify the package @xmath64 @xcite and use it to study the phenomenology of the nmssm @xcite .
also note that the singlino mass vanishes at tree level and the lightest neutralino as the dark matter candidate acquires its mass through the mixing of the singlino with higgsinos and gauginos . in this case , the dark matter is light and singlino dominated , and it must annihilate through exchanging a resonant light cp - odd higgs boson to get the correct relic density@xcite . as a result , the sm - like higgs boson will decay dominantly into light neutralinos or other light higgs bosons so that the branching fractions of the visible decay channels like @xmath65 are suppressed@xcite .
this is strongly disfavored by current lhc data as shown in@xcite . in this work ,
we only take the nmssm as an example to show its peculiar behaviors in higgs physics ( in comparison with other new susy models ) .
* formula in calculation : * in order to study the @xmath66 decay and its correlation with the @xmath67 decay in susy , we define following normalized rates at the lhc and the international linear collider ( ilc)@xcite as @xmath68 where the higgs production at the lhc is dominated by the gluon fusion process , while at the ilc with @xmath69 , it is dominated by the @xmath70 associated production . here
@xmath71 and @xmath72 are the couplings of the higgs boson to gluons and @xmath50s respectively , and @xmath73 , @xmath74 and @xmath75 are the partial widths for the decays @xmath0 , @xmath76 and @xmath77 respectively . in getting these formulae
, we neglect susy radiative corrections to the signals .
those corrections are expected to be few percent given that heavy sparticles are preferred by current lhc experiments . in susy ,
the decays @xmath78 and @xmath79 get new contributions from the loops mediated by charged higgs bosons , sfermions ( including stops , sbottoms and staus ) as well as charginos .
consequently , the formula of @xmath80 and @xmath81 are modified by @xmath82 where @xmath83 ( @xmath84 ) denote the contribution from particle @xmath85 mediated loops , and their explicit expressions are listed in the appendix .
note that our expressions differ from those presented in @xcite in two aspects .
one is that we have an overall minus sign for the new contributions @xmath86 , @xmath87 and @xmath88 , and an additional factor 2 for the sfermion contributions .
this sign difference was also observed recently in @xcite .
the other difference is that we have included in a neat way the contributions from the loops with two particles ( such as @xmath89 and @xmath90 or @xmath91 and @xmath92 ) running in them .
such contributions were considered to be negligibly small @xcite , but our results indicate that sometimes they may play a role . also note that in the susy package @xmath93 @xcite , the decay @xmath15 is not calculated . in the package @xmath64 @xcite , this decay is calculated only by considering the contributions from the sm particles and the charged higgs boson .
we improve these packages by inserting our codes for @xmath15 .
in our calculation , we first perform a random scan over the parameter space of each model by considering various experimental constraints
. then for the surviving samples we investigate the @xmath94 and @xmath1 decays . since for each unconstrained susy model , there are too many free parameters involved in the calculation we make some assumptions to simplify our analysis .
our treatment of the mssm and the nmssm is as follows * firstly , we note that the first two generation squarks change little the properties of the higgs boson , and the lhc search for susy particles implies that they should be very heavy .
so in our scan , we fix all soft masses and the trilinear parameters in this sector to be 2 tev .
we checked that our conclusions are not affected by such specific choice .
* secondly , since the third generation squarks affect the higgs sector significantly , we set free all soft parameters in this sector except that we assume @xmath95 and @xmath96 to reduce the number of free parameters .
* thirdly , considering that the muon anomalous magnetic moment is sensitive to the spectrum of scalar muons and the decay @xmath94 may get significant contribution from scalar tau ( stau ) sector , we assume @xmath97 , @xmath98 and @xmath99 , and treat @xmath100 and @xmath101 as free parameters .
we checked that for our considered cases , the decay @xmath15 is insensitive to @xmath102 . * finally , since our results are insensitive to gluino mass , we fix it at 2 tev .
we also assume the grand unification relation @xmath103 for electroweak gaugino masses .
to sum up , for the mssm we scan the parameters in the following regions @xmath104 note that in actual calculation @xmath41 and @xmath105 in the nmssm are usually treated as independent input parameters , and for any given value of @xmath105 the phenomenology of the nmssm is identical to that of the mssm with @xmath106 in the limit @xmath107 @xcite .
this enables us to use the package nmssmtools @xcite , which calculates various observables and also considers various experimental constraints in the framework of the nmssm , to study the phenomenology of the mssm ( note that the validity of this method has been justified by the authors of the nmssmtools @xcite ) . in our calculation
we use the package nmssmtools-3.2.4 to perform the scan for the mssm by setting @xmath108 and @xmath109 . here
the value of @xmath110 is actually irrelevant to our calculation for the mssm as long as it is negative and satisfies @xmath111 ( in order to guarantee the squared masses of the singlet scalars to be positive ) @xcite . for the nmssm
, we use the package nmssmtools-3.2.4 to scan the region in eq.([mssmscan ] ) and also following ranges for additional parameters @xmath112 note that in our scan , we only consider a relatively large @xmath41 .
the reason is in the nmssm , the properties of the sm - like higgs boson are expected to deviate significantly from the mssm prediction only for a sizable @xmath41 @xcite , and in particular , as far as @xmath113 is concerned , the higgs mass at tree level is maximized at @xmath114 , instead of at large @xmath115 in the mssm . as for the nmssm , our assumptions are same as those for the mssm except that , in order to explain the muon anomalous magnetic moment at @xmath116 level ,
we assume all soft susy breaking parameters in slepton sector to be 100 gev @xcite . such a slepton mass is allowed by the lep ii bounds , which are 99.9 gev for the first two generations and 93.2 gev for the third generation @xcite .
although the lhc also gave a bound on slepton mass by searching for the decay @xmath117 @xcite , it is not applicable to the nmssm .
the reason is that in the nmssm , the lsp is singlino - like and the nlsp is usually a bino - like neutralino @xmath118 . as a result ,
the dominant decay chain of a slepton is @xmath119@xcite .
other parameters in this model are scanned in the following ranges @xmath120 where @xmath121 is the soft breaking mass for the singlet higgs field . in our calculation
, we adapt the code of the nmssmtools to the nmssm case as done in@xcite . for the cmssm
, we use the package nmspec@xcite to scan following parameter space @xmath122 and we take the sign of @xmath33 to be positive . similar to the mssm scan , we set @xmath123 and @xmath124 at the gut scale ( note that the validity of the nmspec to study the phenomenology of the cmssm was emphasized by the authors of the package @xcite ) .
since in the cmssm different soft breaking parameters at electroweak scale are correlated , it is expected that its phenomenology cmssm is only a subset of the mssm . in our scan
we have considered various constraints on the models , which are from vacuum stability , the lep and lhc searches for susy particles and higgs bosons , the electroweak observables @xmath125 and @xmath126 , b physics observables such as the branching ratio of @xmath127 and the mass difference @xmath128 , the dark matter relic density and its direct detection experiments . when imposing the constraint from a certain observable which has experimental central value , we require susy to explain the observable at @xmath116 level . these constraints are described in detail in @xcite and have been implemented in the package nmssmtools-3.2.4 @xcite . in particular ,
the dark matter relic density is calculated by the package @xcite , which now acts as an important component of the nmssmtools .
compared with the constraints in @xcite , we have following improvement in this work * we require @xmath129 .
this mass range is favored by the lhc search for higgs boson after considering theoretical uncertainties@xcite . *
we utilize the latest result of the xenon100 experiment to limit the models ( at @xmath130 confidence level)@xcite .
we calculate the scattering cross section of the dark matter with nucleon with the formula presented in @xcite , and we set @xmath131 with @xmath132 denoting the strange quark content in nucleon . * we consider the constraints from the recent lhcb measurement of @xmath133 , which is @xmath134@xcite .
the agreement of the measurement with its sm prediction strongly limits the combination @xmath135 in the mssm@xcite .
* we also consider the constraint from the cms search for non - sm higgs boson from the channel @xmath136 @xcite .
this search , like @xmath137 , is very power in limiting the @xmath138 plane in the mssm . finally , we emphasize that in our scan we require the mssm , nmssm and nmssm to explain muon g-2 at @xmath116 level , i.e. , @xmath139 @xcite . as for the cmssm
, it has long been noticed that there exists a tension to predict a 125 gev sm - like higgs boson and meanwhile to explain the muon g-2 @xcite . in our calculation
we consider the latest lhc bounds on @xmath140 plane of cmssm and find that under such latest bounds the cmssm can not explain the muon g-2 at @xmath116 level .
so we do not require the cmssm to explain the muon g-2 at @xmath116 level in our analysis
. plane . here
@xmath141 denotes the normalized @xmath2 signal rate in the sm - like higgs boson production at the lhc , and the black line represents its upper limit set by the cms collaboration @xcite.,width=642 ] in fig.[fig1 ] we project the surviving samples on the plane of the @xmath5 signal rate at the lhc versus the sm - like higgs boson mass in the four susy models .
we also show the cms bound on the rate in the figure @xcite . from the left panel
we see that compared with its sm prediction , the @xmath5 rate in the mssm and the nmssm can be either enhanced or suppressed with the maximal deviation reaching @xmath142 and @xmath143 respectively .
in contrast , as shown in the right panel , the @xmath5 rate is always slightly suppressed ( less than @xmath144 ) in the cmssm and severely suppressed ( more than @xmath130 ) in the nmssm .
we checked that for the cmssm and the mssm , the suppression is mainly due to the increase of @xmath145 partial width @xcite .
for the nmssm , however , it is due to the open up of new decays @xmath146 ( @xmath147 and @xmath148 denote the dark matter and lightest cp - odd higgs boson respectively ) , which significantly enlarges the total width of the higgs boson and leads to severe suppression for all visible decay channels .
for the samples shown in fig.[fig1 ] , we also compare their predictions on the @xmath6 and @xmath149 signal rates of the higgs boson with the corresponding experimental data .
we find that for most of the samples in the mssm and the nmssm and for all the samples in the cmssm , their theoretical predictions on the @xmath6 and @xmath149 rates agree with the corresponding data at @xmath150 level , while for the samples in the nmssm , their predictions always lie outside the @xmath151 regions ( see fig.1 and fig.2 in @xcite ) .
moreover , we checked that the branching ratio of @xmath145 in the mssm , nmssm and cmssm varies in the ranges [ @xmath152 , @xmath153 , [ @xmath154 , @xmath155 and [ @xmath143 , @xmath156 , respectively ( in the sm its value is about @xmath152 for @xmath157 ) , and the signal strength for the process @xmath158 normalized by its sm value varies from 0.97 to 1.12 , 0.55 to 1.05 and 1.00 to 1.02 , respectively .
considering that so far the only way to detect the @xmath145 decay at the lhc is through the @xmath159 associated production , whose signal strength @xmath160 is @xmath161 from atlas result @xcite and @xmath162 from cms result @xcite , one can conclude that such alterations of @xmath10 signal rates are allowed by the current experimental data at @xmath116 level . in the mssm and nmssm .
the magenta bullets and blue triangles represent the samples with @xmath163 and @xmath164 respectively.,width=604 ] next we focus on the mssm and the nmssm . in fig.2
we exhibit the contributions of different sparticles to the amplitude of the @xmath66 decay .
since the sbottoms and charged higgs bosons have little effect on the amplitude , we do not show their contributions .
this figure shows following features : * in the mssm , the potentially largest contribution comes from the stau loops , which can alter the sm amplitude by about @xmath165 .
in contrast , the stop contribution is small , which usually changes the amplitude by less than @xmath166 .
this feature can be well understood by the formulae listed in appendix .
explicitly speaking , in order to get a significant sfermion contribution , one necessary condition is @xmath167 should be as large as possible , where @xmath168 denotes the chiral flipping coupling of higgs to sfermions , @xmath169 is the chiral mixing angle and @xmath170 represents the lighter sfermion mass . as far as the stop sector is concerned , a relatively light @xmath171 is always accompanied by a heavy @xmath172 in order to predict @xmath173 . although @xmath174 in this case may be very large , the chiral mixing angle @xmath175 is usually small and consequently , the stop contribution can never get significantly enhanced . in the stau sector , however , both the parameters @xmath176 and @xmath177 are unlimited , and one can choose light staus and an appropriate @xmath178 to maximize the contribution . in this process , the value of @xmath179 and the splitting between @xmath176 and @xmath177 play an important role .
it is worth noting that a light stau with mass close to dark matter may co - annihilate with the dark matter , which is helpful to avoid the overabundance of the dark matter in today s universe@xcite .
* in the mssm , the chargino contribution is small and can only reach @xmath166 and @xmath180 for @xmath181 case and @xmath182 case respectively .
the reason is that in the mssm , the @xmath183 coupling is induced by the @xmath184 interaction ( @xmath185 and @xmath186 and @xmath187 denote higgsino and wino respectively ) , and this coupling strength is maximized when both the higgsino and wino components of @xmath188 are sizable .
we checked that , for most of the @xmath182 samples , @xmath189 and @xmath190 ( a large @xmath33 is needed for the stau contribution to enhance the diphoton rate @xcite ) so that @xmath188 is basically wino - like .
consequently , its coupling to the higgs boson is weak .
* in the nmssm , the largest susy contribution to the @xmath5 decay comes from the chargino loops with the correction reaching @xmath191 in optimal case , while the magnitude of the stau contribution is always smaller than @xmath192 .
this is because in the nmssm with a large @xmath41 , @xmath33 is preferred to vary from 100 gev to 250 gev and @xmath115 is usually smaller than 10 @xcite . as a result , the @xmath193 coupling is relatively large , while the @xmath194 coupling can not be pushed up by the moderate @xmath195 . due to the singlet component of @xmath12 in the nmssm ,
the @xmath196 coupling can be greatly suppressed ( reaching @xmath197 by our results ) so that the total width of @xmath12 is reduced by about @xmath198 in extreme case@xcite . consequently
, even when the susy contributions to the decay width are small , @xmath199 can still be quite large .
as we mentioned before , such a suppressed @xmath196 coupling can reduce the normalized strength of the process @xmath158 down to 0.55 , which , however , is still compatible with the current lhc data due to the large uncertainty of the measured signal strength ( @xmath200 from atlas result @xcite and @xmath201 from cms result @xcite ) .
channel with that to @xmath202 channel.,width=661 ] in fig.[fig3 ] , we show the correlation of the amplitudes for @xmath203 and @xmath1 .
this figure indicates that in both the mssm and the nmssm , the top squark contribution to the amplitude of @xmath203 correlates roughly in a linear way with that of @xmath76 , and so is the chargino contribution . fig.[fig3 ] also indicates that the correlation is spoiled for the stau contribution in the mssm .
we checked that this is because @xmath204 in the mssm can vary over a broad range and the dependence of the two amplitudes on @xmath204 are quite different .
for example , in the case @xmath205 , @xmath206 and both @xmath207 and @xmath208 couplings approach zero by accidental cancelation . as a result ,
the stau contribution to the decay @xmath15 is suppressed .
in contrast , the contribution to the decay @xmath76 is maximized since it is proportional to @xmath209 . on the other hand , if @xmath210 so that @xmath211 , the contributions are suppressed for both the decays because the dominant contribution to @xmath212 coupling and that to @xmath213 coupling are both proportional to @xmath209 . versus @xmath214 .
, width=566 ] considering @xmath199 is mainly determined by the partial width of @xmath66 and the total width of the sm - like higgs boson , we present in fig.[fig4 ] the ratio of @xmath215 versus the ratio of @xmath214 for the two models .
the left panel indicates that for almost all mssm samples the @xmath19 partial width and the total width of the sm - like higgs boson is larger than the corresponding sm predictions .
these features originate from the constructive contributions of the susy particles to @xmath15 and the enhanced width of @xmath216 respectively .
interestingly , the largest increase of @xmath80 occurs when @xmath217 .
the right panel indicates that , in order to enhance the @xmath2 signal in the nmssm with a large @xmath41 ( note that in this model , @xmath199 correlates roughly in a linear way with @xmath218 , see fig.[fig5 ] ) , the sm - like higgs boson tends to have sizable singlet component to suppress the total width . in this case , @xmath80 is suppressed too , but we have @xmath219 .
moreover , as mentioned before , @xmath220 can be slightly enhanced by the chargino contribution . and
@xmath221 in the four models.,width=661 ] now we investigate the correlation of the @xmath5 rate with the @xmath3 rate in different susy models , which is shown in fig.[fig5 ] . from this figure
we have following conclusions : * in the mssm , although the partial width of @xmath15 can be enhanced by @xmath142 ( see fig.[fig4 ] ) , due to the increase of the higgs total width and also the suppression of the @xmath222 coupling @xcite , the maximal value of @xmath199 is only 1.1 ( in comparison , @xmath218 may be as large as 2 . ) , and only when @xmath223 can @xmath224 be possible . among the sparticle contributions to @xmath199 and @xmath225 , the stau loops play the dominant role .
the difference between the two signals comes from their dependence on @xmath204 , i.e. @xmath226 @xmath218 is maximized while @xmath199 is suppressed .
our numerical results also indicate that , for the surviving samples of the mssm , the branching ratio of the invisible decay @xmath227 is usually smaller than @xmath191 and @xmath166 for @xmath181 and @xmath228 case , respectively , and the branching ratio of @xmath145 varies from @xmath152 to @xmath229 . * in the nmssm with a large @xmath41 , the sparticle corrections to the amplitudes of @xmath203 and @xmath76 are usually below @xmath165 , and the main mechanism to alter @xmath199 and @xmath218 is through the suppression of the @xmath196 and @xmath230 couplings by the singlet component of @xmath12 . as a result , @xmath199 and @xmath218
are highly correlated and both of them vary from 0.2 to 2 .
we checked that the branching ratios of the exotic decays @xmath146 may reach @xmath231 and @xmath232 , respectively ( these extreme cases correspond to some of the squared points in fig.5 of @xcite ) , and the branching ratio of the decay @xmath145 varies in a large range , from @xmath154 to @xmath233 .
* in the cmssm and nmssm , @xmath199 and @xmath218 are slightly and strongly suppressed respectively . as discussed before ,
in the cmssm the suppression is due to the increase of @xmath77 partial width , while in nmssm it is due to the open up of the exotic decay channels @xmath234 .
note that the open up of the exotic decays @xmath146 will generally lead to the suppression of visible signal rates such as @xmath218 and @xmath235 , and as analyzed in @xcite , the latest higgs data require that the total branching ratio of the exotic decays should be less than @xmath236 at @xmath237 c.l .. this conclusion again indicates that the nmssm and also some samples of the nmssm are disfavored by the current higgs data ( about this conclusion , one may also see the squared points in fig.5 of @xcite ) . for the decay @xmath238 in the nmssm
, we checked that @xmath239 varies from about @xmath240 to @xmath241 , and @xmath148 mainly decays to @xmath10 ( with a branching ratio at about @xmath130 ) and @xmath242 ( with the branching ratio at about @xmath243 ) .
so in this case , the decay product of the higgs boson is four @xmath244-jets , or four @xmath245 leptons or two @xmath244-jets plus two @xmath245 leptons , which is an interesting but challenging signal in higgs search at the lhc@xcite .
since @xmath239 is usually heavier than @xmath246 , the constraint from the decay @xmath247 @xcite is irrelevant here .
also note that generally speaking , the constrained model such as cmssm tends to predict strong correlations among observables due to the unified nature of its parameters .
this is clearly shown in fig.[fig5 ] for the cmssm in comparison with the other three models . and
@xmath248 in the four models.,width=661 ] since the @xmath4 collider at @xmath249 provides a clean environment to detect the decays @xmath0 and @xmath250 , we also investigate their rates at the ilc defined in eq.([ilc1 ] ) and eq.([ilc2 ] ) .
the corresponding results are shown in fig.[fig6 ] .
this figure exhibits following features * in the mssm , a suppressed @xmath5 signal ( compared with its sm prediction ) tends to correspond to an enhanced @xmath10 signal , and an enhanced @xmath5 signal requires the @xmath10 signal rates to be roughly at its sm prediction . in any case , the enhancement factor for the two signals are less than 1.2 .
note that there exist a few cases where the @xmath10 signal rate is slightly suppressed .
* in the nmssm with a large @xmath41 , the normalized @xmath10 signal rate is less than 1.1 , and in some cases it may be significantly suppressed .
in contrast , the @xmath5 signal rate can be either greatly enhanced or severely suppressed . in the enhancement case , the @xmath10 signal rate is usually less than its sm prediction , and the greater enhancement corresponds to the stronger suppression .
* in the cmssm , both the signal rates are roughly equal to their sm predictions . in the nmssm ,
however , both the rates are strongly suppressed .
moreover , we checked that the @xmath3 signal rate at the ilc has similar dependence on the @xmath10 rate for the four models .
in this work , we investigate the rare decay of the sm - like higgs boson , @xmath20 , and study its correlation with @xmath202 in the mssm , the nmssm , the nmssm and the cmssm .
we perform a scan over the parameter space of each model by considering various experimental constraints and present our results on various planes .
we have following observations : * in the susy models , the sparticle correction to the rare decay @xmath203 is usually several times smaller than that to @xmath1 .
* in the mssm , the net susy contribution to the amplitude of @xmath20 is constructive with the corresponding sm amplitude and can enhance the sm prediction by at most @xmath165 . as a result
, the @xmath5 signal rates at the lhc and the ilc can be enhanced by @xmath142 at most . as a comparison
, the @xmath3 rate can be enhanced by a factor of 2 due to the large stau contributions . * in the cmssm , due to the slightly enhanced total width of the sm - like higgs bson , the @xmath2 signal rates at the lhc and the ilc are both slightly below their sm predictions , and so is the @xmath8 signal .
* in the nmssm with a large @xmath41 , the susy corrections to the amplitudes for the decays @xmath15 and @xmath251 are at most @xmath165 , and to get significant deviation of the two rates from their sm values , the total width of the sm - like higgs boson must be moderately suppressed by the singlet component of @xmath12 . in this model ,
the two rates are highly correlated and vary from 0.2 to 2 . * in the nmssm , the signal rates of @xmath20 and @xmath202 are both greatly suppressed due to the open up of the exotic decay @xmath252 . finally , we note that some strategies have been proposed in the literature to discriminate the considered models , e.g. , via the correlations between the higgs couplings @xcite , via the enhanced higgs pair productions at the lhc @xcite and the ilc @xcite , or via the direct dark matter detection @xcite . compared with these existing strategies , the loop - induced @xmath5 and @xmath3 decay modes of the higgs boson seem to be more sensitive to the nature of the models ( some non - susy models predict rather different correlation behavior @xcite ) .
so the correlation between @xmath5 and @xmath3 rates analyzed in this work may play a complementary role to discriminate new physics models in the future . _
note added : _ after we finished this work , both the atlas and cms collaborations updated their higgs search results @xcite . among these new results ,
the cms data on the diphoton signal rate is changed significantly @xcite .
since in our analysis we did not use the diphoton data as a constraint ( instead we just displayed the predictions for the diphoton signal rate in different susy models ) , our results and conclusions are not affected by the new data .
we thank kun yao and jingya zhu for helpful discussions .
this work was supported in part by the national natural science foundation of china ( nnsfc ) under grant no .
10775039 , 11075045 , 11275245 , 10821504 and 11135003 , by the project of knowledge innovation program ( pkip ) of chinese academy of sciences under grant no .
in susy , the decays @xmath78 and @xmath79 get new contributions from the loops mediated by charged higgs bosons , sfermions ( including stops , sbottoms and staus ) and charginos , and as a result , the formula of @xmath80 and @xmath81 are modified by @xmath253 the expressions of @xmath254 are relatively simple and are given by @xmath255 where @xmath256 , @xmath257 denotes the higgs coupling with particles @xmath258 and @xmath259 and @xmath260 are loop functions with scalars , fermions and gauge bosons running in the loop .
the explicit expressions of @xmath257 and @xmath261 functions are given by @xmath262,\\ & & a_{1/2}(x ) = 2 \
, x^2 \left[x^{-1}+ ( x^{-1}-1)f(x^{-1})\right],\\ & & a_1(x)=-x^2\left[2x^{-2}+3x^{-1}+3(2x^{-1}-1)f(x^{-1})\right],\end{aligned}\ ] ] where s is the @xmath263 ( @xmath264 ) rotation matrix of mssm ( nmssm ) higgs mass matrix under the basis @xmath265 , h in @xmath266 denotes the row index of the sm - like higgs , @xmath267 denotes the sm - like higgs coupling to sfermion interaction states , @xmath268 denote the rotation matrices of the chargino mass matrix , and @xmath269 is defined by @xmath270 . as for @xmath271 , due to @xmath272 and the existence of @xmath273 ( @xmath274 ) couplings , their expressions are rather complex @xmath275 where @xmath276 and the coupling coefficients of @xmath12 and @xmath50 are given by @xmath277 in above formula , we have defined some new functions as @xmath278 with @xmath279 denoting three points loop functions introduced in @xcite , and @xmath280 , @xmath281 , @xmath282 , and @xmath283 . for the special cases considered here , @xmath279
are given by @xmath284 where @xmath285 note that for the special case @xmath286 , these functions can be further simplified to get their analytic expressions .
the other functions relevant to our calculation are defined by @xmath287 i_1(x , y),\end{aligned}\ ] ] with @xmath288 + \frac{x^2 y}{(x - y)^2}[g(x^{-1})-g(y^{-1})],\\ i_2(x , y ) & = & - \frac{x y}{2(x - y ) } [ f(x^{-1})-f(y^{-1})],\\ g(x ) & = & \sqrt{x^{-1 } -1 } \arcsin \sqrt{x}.\end{aligned}\ ] ] a. g.akeroyd and s. moretti , phys .
d * 86 * , 035015 ( 2012 ) ; a. kobakhidze , arxiv:1208.5180 ; m. carena , s. gori , i. low , n. r. shah and c. e. m. wagner , arxiv:1211.6136 ; e. j. chun _
et al . _ ,
jhep * 1211 * , 106 ( 2012 ) ; r. sato _
et al . _ ,
b * 716 * , 441 ( 2012 ) ; w. -c . huang and a. urbano , arxiv:1212.1399 ; m. chala , arxiv:1210.6208 ; e. o. iltan , arxiv:1212.5695 .
a. joglekar _
_ , arxiv:1207.4235 ; e. bertuzzo _ et al . _ , arxiv:1209.6359 ; l. g. almeida _ et al .
_ , jhep * 1211 * , 085 ( 2012 ) ; h. davoudiasl _ et al._,arxiv:1211.3449 ; b. batell _
et al . _ , arxiv:1211.2449 ; h. m. lee _
et al . _ , arxiv:1209.1955 ; m. b. voloshin , phys .
d * 86 * , 093016 ( 2012 ) ; n. bonne and g. moreau , phys .
b * 717 * , 409 ( 2012 ) ; s. dawson and e. furlan , phys .
d * 86 * , 015021 ( 2012 ) ; m. a. ajaib , i. gogoladze , q. shafi , arxiv:1207.7068 ; s. dawson _ et al .
_ , arxiv:1210.6663
. m. carena _ et al .
_ jhep * 1203 * , 014 ( 2012 ) ; jhep * 1207 * , 175 ( 2012 ) ; u. ellwanger , jhep * 1203 * , 044 ( 2012 ) ; u. ellwanger , c. hugonie , arxiv:1203.5048 ; k. hagiwara , j. s. lee , j. nakamura , jhep * 1210 * , 002 ( 2012 ) ; r. benbrik
_ et al . _ , arxiv:1207.1096 ; t. cheng , arxiv:1207.6392 ; b. kyae , j .- c .
park , arxiv:1207.3126 ; h. an , t. liu , l .- t .
wang , arxiv:1207.2473 ; j. ke _ et al .
_ , arxiv:1207.0990 ; g. belanger _
et al . _ , arxiv:1208.4952 ; n. liu _ et al . _ , arxiv:1208.3413 ; m. drees , arxiv:1210.6507 ; s. f. king _
et al . _ , arxiv:1211.5074 ; k. choi _ et al . _ , arxiv:1211.0875 ; m. berg _
et al . _ , arxiv:1212.5009 ; l. aparicio _ et al . _ , arxiv:1212.4808 ; c. balazs , s. k. gupta , arxiv:1212.1708 ; k. cheung , c .- t .
lu , t .- c .
yuan , arxiv:1212.1288 ; z. kang , y. liu and g. ning , arxiv:1301.2204 ; e. j. chun and p. sharma , arxiv:1301.1437 .
a. drozd _ et al .
_ , arxiv:1211.3580 ; p. m. ferreira_et al .
_ , arxiv:1211.3131 ; s. chang _ et al .
_ , arxiv:1210.3439 ; s. bar - shalom _ et al .
_ , arxiv:1208.3195 ; h. s. cheon and s. k. kang , arxiv:1207.1083 ; a. arhrib _ et al .
_ , arxiv:1205.5536 ; j. chang _
et al . _ , arxiv:1206.5853 ; l. wang and x. -f .
han , phys .
d * 87 * , 015015 ( 2013 ) ; n. chen and h. -j .
he , jhep * 1204 * , 062 ( 2012 ) ; b. swiezewska and m. krawczyk , arxiv:1212.4100 . j. s. gainer , w. -y .
keung , i. low and p. schwaller , phys .
d * 86 * , 033010 ( 2012 ) [ arxiv:1112.1405 [ hep - ph ] ] .
s. y. choi , m. m. muhlleitner and p. m. zerwas , phys .
b * 718 * , 1031 ( 2013 ) [ arxiv:1209.5268 [ hep - ph ] ] .
technical design report , the ilc baseline design , https://forge.linearcollider.org/dist/20121210-ca-tdr2.pdf ; technical design report , ilc r&d in the technical design phase , https://forge.linearcollider.org/dist/20121210-ca-tdr1.pdf .
m. maniatis , int .
phys . a * 25 * ( 2010 ) 3505
; j. r. ellis _ et al . _ phys . rev .
d * 39 * , 844 ( 1989 ) ; m. drees , int .
phys . a * 4 * , 3635 ( 1989 ) ; s. f. king , p. l. white , phys .
d * 52 * , 4183 ( 1995 ) ; b. ananthanarayan , p.n .
pandita , phys .
b * 353 * , 70 ( 1995 ) ; b. a. dobrescu , k. t. matchev , jhep * 0009 * , 031 ( 2000 ) ; r. dermisek , j. f. gunion , phys . rev .
lett . * 95 * , 041801 ( 2005 ) ; g. hiller , phys .
d * 70 * , 034018 ( 2004 ) ; f. domingo , u. ellwanger , jhep * 0712 * , 090 ( 2007 ) ; z. heng _ et al .
_ , phys .
d * 77 * , 095012 ( 2008 ) ; r. n. hodgkinson , a. pilaftsis , phys .
d * 76 * , 015007 ( 2007 ) ; w. wang _
phys . lett .
b * 680 * , 167 ( 2009 ) ; j. m. yang , arxiv:1102.4942 ; u. ellwanger and c. hugonie , arxiv : hep - ph/0612133 , arxiv : hep - ph/9909260 ; u. ellwanger , arxiv:1108.0157 .
a. h. chamseddine , r. arnowitt and p. nath , phys . rev
* 49 * ( 1982 ) 970 ; r. barbieri , s. ferrara and c.a savoy , phys .
b * 119 * ( 1982 ) 343 ; l. hall , j. lykken and s. weinberg , phys .
d * 27 * ( 1983 ) 2359 ; n. ohta , prog .
* 70 * ( 1983 ) 542 ; a. menon , et al .
d * 70 * , 035005 ( 2004 ) ; v. barger , et al .
b * 630 * , 85 ( 2005 ) ; c. balazs , et .
al . , jhep * 0706 * , 066 ( 2007 ) ; c. panagiotakopoulos , k. tamvakis , phys . lett .
b * 446 * , 224 ( 1999 ) ; phys .
b * 469 * , 145 ( 1999 ) ; c. panagiotakopoulos , a. pilaftsis phys .
d * 63 * , 055003 ( 2001 ) ; a. dedes , et .
al . , phys .
d * 63 * , 055009 ( 2001 ) ; k. s. jeong , y. shoji and m. yamaguchi , jhep * 1204 * , 022 ( 2012 ) ; jhep * 1209 * , 007 ( 2012 ) .
g. blanger , f. boudjema , and a. pukhov , _ micromegas : a package for calculation of dark matter properties in generic model of particle interaction _ , ch .
12 , pp .
world scientific , 2012 . | in light of recent remarkable progress in higgs search at the lhc , we study the rare decay process @xmath0 and show its correlation with the decay @xmath1 in low energy susy models such as cmssm , mssm , nmssm and nmssm . under various experimental constraints , we scan the parameter space of each model , and present in the allowed parameter space the susy predictions on the @xmath2 and @xmath3 signal rates in the higgs production at the lhc and future @xmath4 linear colliders .
we have following observations : ( i ) compared with the sm prediction , the @xmath5 and @xmath6 signal rates in the cmssm are both slightly suppressed ; ( ii ) in the mssm , both the @xmath5 and @xmath3 rates can be either enhanced or suppressed , and in optimal case , the enhancement factors at the lhc can reach 1.1 and 2 respectively ; ( iii ) in the nmssm , the @xmath5 and @xmath3 signal rates normalized by their sm predictions are strongly correlated , and at the lhc the rates vary from 0.2 to 2 ; ( iv ) in the nmssm , the @xmath5 and @xmath3 rates are both greatly reduced . since the correlation behavior between the @xmath7 signal and the @xmath3 signal is so model - dependent , it may be used to distinguish the models in future experiments . |
let @xmath3 be a smooth complex projective curve of genus @xmath4 , and let @xmath5 be a complex reductive group .
the @xmath5-character variety of @xmath3 is defined as the moduli space of semisimple representations of @xmath6 into @xmath5 , that is , @xmath7=\operatorname{id}\}/ / g.\ ] ] for complex linear groups @xmath8 , the representations of @xmath9 into @xmath5 can be understood as @xmath5-local systems @xmath10 , hence defining a flat bundle @xmath11 which has @xmath12 .
a natural generalization consists of allowing bundles @xmath11 of non - zero degree @xmath13 .
the @xmath5-local systems on @xmath3 correspond to representations @xmath14 , where @xmath15 is a fixed point , and @xmath16 , @xmath17 a loop around @xmath18 , giving rise to the moduli space of twisted representations @xmath19=e^{2\pi { \sqrt{-1}}d / n } \operatorname{id}\}/ / g.\ ] ] the space @xmath20 is known in the literature as the betti moduli space .
this space is closely related to two other spaces : the de rham and dolbeault moduli spaces .
the de rham moduli space @xmath21 is the moduli space parameterizing flat bundles , i.e. , @xmath22 , where @xmath10 is an algebraic bundle of degree @xmath13 and rank @xmath23 ( and fixed determinant in the case @xmath24 ) , and @xmath25 is an algebraic connection on @xmath26 with a logarithmic pole at @xmath18 with residue @xmath27 .
the riemann - hilbert correspondence @xcite gives an analytic isomorphism ( but not an algebraic isomorphism ) @xmath28 .
the dolbeault moduli space @xmath29 is the moduli space of @xmath5-higgs bundles @xmath30 , consisting of an algebraic bundle @xmath10 of degree @xmath13 and rank @xmath23 ( also with fixed determinant in the case @xmath24 ) , and a homomorphism @xmath31 , known as the higgs field ( in the case @xmath24 , the higgs field has trace @xmath32 ) . in this situation
the theory of harmonic bundles @xcite gives a homeomorphism @xmath33 .
when @xmath34 these moduli spaces are smooth and the underlying differentiable manifold is hyperkhler , but the complex structures do not correspond under the previous isomorphisms .
the cohomology of these moduli spaces has been computed in several particular cases , mostly from the point of view of the dolbeault moduli space @xmath35 .
poincar polynomials for @xmath36 were computed by hitchin in his seminal paper on higgs bundles @xcite and for @xmath37 by gothen in @xcite .
more recently the case of @xmath38 has been solved in @xcite , and a recursive formula for the motive of the moduli space of higgs bundles of arbitrary rank and degree coprime to the rank has been given in @xcite .
in particular , this gives the betti numbers of the character variety for arbitrary coprime rank and degree .
hausel and rodriguez - villegas started the computation of the e - polynomials of @xmath5-character varieties focusing on @xmath8 and @xmath39 using arithmetic methods inspired by the weil conjectures .
in @xcite they obtained the e - polynomials of @xmath40 for @xmath41 in terms of a simple generating function . following these methods ,
mereb @xcite studied the case of @xmath42 giving an explicit formula for the e - polynomial in the case @xmath36 , while for @xmath42 these polynomials are given in terms of a generating function .
he proved also that the e - polynomials for @xmath42 are palindromic and monic .
logares , newstead and the second author @xcite introduced a geometric technique to compute the e - polynomial of character varieties by using stratifications and also handling fibrations which are locally trivial in the analytic topology but not in the zariski topology .
the main results of @xcite are explicit formulae for the e - polynomials for character varieties for @xmath36 and @xmath2 .
actually , the geometric technique allows for dealing with `` twisted '' character varieties in which the holonomy around the puncture is not diagonizable ( in this case there is no correspondence with a de rham or dolbeault moduli space as mentioned above ) .
the purpose in this paper is to give another step in the geometric technique of @xcite to the case of @xmath0 .
we prove the following [ thm : main ] let @xmath3 be a complex curve of genus @xmath0 . then the e - polynomials of the character varieties are : @xmath43 where @xmath44 .
the new input consists of using the map @xmath45 given by @xmath46 , @xmath47)$ ] , to stratify conveniently the space of representations in @xmath48 .
this leads us to analyse a fibration over a two - dimensional basis ( the quadric @xmath49 in figure [ fig1 ] ) , for which we have to compute the e - polynomial .
therefore the arguments of @xcite about hodge monodromy fibrations over one - dimensional bases have to be extended to general bases ( see theorem [ thm : general - fibr ] ) . in the case of
@xmath50 we also have to deal with reducible representations .
the arguments used in ( * ? ? ?
* section 8.2.2 ) to compute the e - polynomial of the irreducible locus give an increasingly large number of strata as the genus increases .
so here we use an alternative route ( suggested by the ideas in @xcite ) , consisting of computing the e - polynomial of the irreducible locus by subtracting the e - polynomial of the subset of reducible representations . as a byproduct of our analysis
, we obtain the behaviour of the e - polynomial of the parabolic character variety ( @xmath36 ) @xmath51= \begin{pmatrix } \lambda & 0 \\ 0 & \lambda^{-1 } \end{pmatrix } \right\}/ / g.\ ] ] when @xmath52 varies in @xmath53 , for the case @xmath54 .
this is given by the following formula @xmath55 which means that the e - polynomial of the invariant part of the cohomology is the polynomial accompanying @xmath56 , and the e - polynomial of the non - invariant part is the polynomial accompanying @xmath57 .
this monodromy is of interest from the point of view of mirror symmetry @xcite and answers a question raised to the authors by tams hausel . after finishing and submitting a first version of this paper , the authors have managed to complete the computation of e - polynomials of character varieties for surfaces of general genus @xmath58 in @xcite .
the case provided here can not however be subsumed in @xcite , since it serves as the starting point of induction on the genus carried there .
moreover , this specific case of @xmath0 has special features and has to be treated by an _
ad - hoc _ method ( it is the only case that needs the analysis of hodge monodromy representations over a @xmath59-dimensional basis )
. * acknowledgements .
* we would like to thank peter newstead , tams hausel , martin mereb , nigel hitchin , marina logares and oscar garca - prada for useful conversations
. this work has been partially supported by micinn ( spain ) project mtm2010 - 17389 .
the first author was also supported by a fpu scholarship from the spanish ministerio de educacin .
our main goal is to compute the hodge - deligne polynomial of @xmath1-character varieties .
we will follow the methods in @xcite , so we collect some basic results from @xcite in this section .
we start by reviewing the definition of e - polynomials . a pure hodge structure of weight @xmath60 consists of a finite dimensional complex vector space @xmath61 with a real structure , and a decomposition @xmath62 such that @xmath63 , the bar meaning complex conjugation on @xmath61 .
a hodge structure of weight @xmath60 gives rise to the so - called hodge filtration , which is a descending filtration @xmath64 .
we define @xmath65 .
a mixed hodge structure consists of a finite dimensional complex vector space @xmath61 with a real structure , an ascending ( weight ) filtration @xmath66 ( defined over @xmath67 ) and a descending ( hodge ) filtration @xmath68 such that @xmath68 induces a pure hodge structure of weight @xmath60 on each @xmath69 .
we define @xmath70 and write @xmath71 for the _ hodge number _ @xmath72 .
let @xmath73 be any quasi - projective algebraic variety ( maybe non - smooth or non - compact ) .
the cohomology groups @xmath74 and the cohomology groups with compact support @xmath75 are endowed with mixed hodge structures @xcite .
we define the _ hodge numbers _ of @xmath73 by @xmath76 the hodge - deligne polynomial , or e - polynomial is defined as @xmath77 when @xmath78 for @xmath79 , the polynomial @xmath80 depends only on the product @xmath81 .
this will happen in all the cases that we shall investigate here .
in this situation , it is conventional to use the variable @xmath82 .
if this happens , we say that the variety is _ of balanced type_. for instance , @xmath83 .
the key property of hodge - deligne polynomials that permits their calculation is that they are additive for stratifications of @xmath73 . if @xmath73 is a complex algebraic variety and @xmath84 , where all @xmath85 are locally closed in @xmath73 , then @xmath86 .
there is another useful property that we shall use often : if there is an action of @xmath87 on @xmath3 , then we have polynomials @xmath88 , which are the e - polynomials of the invariant and anti - invariant parts of the cohomology of @xmath3 , respectively . more concretely , @xmath89 and @xmath90 .
then if @xmath87 acts on @xmath3 and on @xmath91 , we have the equality ( see ( * ? ? ?
* proposition 2.6 ) ) @xmath92 we have to deal with fibrations @xmath93 that are locally trivial in the analytic topology .
we want to compute the e - polynomial of the total space in terms of the polynomials of the base and the fiber .
the fibration defines a local system @xmath94 , whose fibers are the cohomology groups @xmath95 , where @xmath96 , @xmath97 .
the fibers possess mixed hodge structures , and the subspaces @xmath98 are preserved by the holonomy .
we will assume from now on that @xmath68 is of balanced type , so @xmath99 .
associated to the fibration , there is a monodromy representation : @xmath100 suppose that the monodromy group @xmath101 is an abelian and finite group .
then @xmath102 are modules over the representation ring @xmath103 .
so there is a well defined element , the _
hodge monodromy representation _ , @xmath104 \ , .\
] ] as the monodromy representation ( [ eqn : exxtra ] ) has finite image , there is a finite covering @xmath105 such that the pull - back fibration @xmath106 \ar[d]_{\pi ' } & z \ar[d]_\pi \\
b_\rho \ar[r ] & b}\ ] ] has trivial monodromy .
let @xmath107 be the irreducible representations of @xmath108 ( there are @xmath109 of them , and all of them are one - dimensional ) .
these are generators of @xmath103 as a free abelian group .
we write the hodge monodromy representation of ( [ eqn : hodge - mon - rep ] ) as @xmath110 [ thm : general - fibr ] suppose that @xmath111 is of balanced type
. then @xmath73 is of balanced type .
moreover , there are polynomials @xmath112 $ ] ( only dependent on @xmath113 and @xmath108 , but not on the fibration or the fiber ) such that @xmath114 for @xmath115 .
the leray spectral sequence of the fibration ( [ fibration ] ) has @xmath116-term @xmath117 and abuts to @xmath75 . by @xcite
, @xmath118 has a mixed hodge structure for @xmath119 , and the differentials @xmath120 are compatible with the mixed hodge structure . therefore @xmath121 .
the mixed hodge structure associated to @xmath122 in ( [ eqn : fibration - bis ] ) is the product mixed hodge structure @xmath123 . by our assumption ,
@xmath111 and @xmath68 are of balanced type , so @xmath124 is of balanced type .
there is a map @xmath125 , which preserves the mixed hodge structures .
this map is surjective , so @xmath73 is of balanced type . by definition , @xmath126 the local systems
@xmath127 are @xmath128-dimensional and have a mixed hodge structure .
when we pull - back @xmath127 to @xmath111 , we get trivial local systems .
hence @xmath129 and it has the induced mixed hodge structure .
so @xmath130 . we define @xmath131 to have the statement .
write @xmath132 , @xmath133 .
so there is a @xmath134$]-linear map @xmath135 \to { \mathbb{z}}[q]\ ] ] satisfying the property that @xmath136 .
a trivial application of theorem [ thm : general - fibr ] happens when @xmath137 is a fibre bundle with fibre @xmath68 such that the action of @xmath138 on @xmath139 is trivial .
then @xmath140 , where @xmath56 is the trivial local system and @xmath141 ( this result appears in ( * ? ? ?
* proposition 2.4 ) ) .
the hypothesis that the action of @xmath138 on @xmath139 is trivial holds in particular in the following cases : * @xmath142 is irreducible and @xmath143 is locally trivial in the zariski topology .
* @xmath143 is a principal @xmath5-bundle with @xmath5 a connected algebraic group .
* @xmath73 is a @xmath5-space with isotropy @xmath144 such that @xmath145 is a fiber bundle , and @xmath5 is a connected algebraic group .
we can also recover easily the result in ( * ? ? ?
* proposition 2.10 ) , which is the main tool used in @xcite .
[ cor : general - fibr ] let @xmath146 .
suppose that @xmath111 is a rational curve .
then @xmath147 , where @xmath148 denotes the e - polynomial of the invariant part of the cohomology of @xmath68 .
let @xmath108 be the monodromy , @xmath109 , and @xmath149 the irreducible representations , where @xmath56 denotes the trivial representation .
we only need to see that @xmath150 and @xmath151 for @xmath152 . given that , if @xmath153 , then @xmath154 , @xmath155 , and @xmath156 , as required .
for the trivial representation , it is clear that @xmath157 .
now let @xmath158 be a ( one - dimensional ) irreducible representation of @xmath108 .
let @xmath159 be the image of @xmath160 , and let @xmath161 .
take the @xmath162-cover asociated to this group , @xmath163 .
then @xmath159 acts on @xmath164 with quotient @xmath142 . clearly , @xmath164 is a rational curve ( the quotient @xmath165 produces a covering map @xmath166 , and @xmath111 is a rational curve by assumption ) .
then we have a fibration @xmath167 , where @xmath91 is a finite set of @xmath162 points . clearly , @xmath168 , for some @xmath169 , where @xmath170 .
the covering @xmath171 can be extended to a ramified covering @xmath172 .
hurwitz formula then says that @xmath173 , where @xmath174 is the degree of the ramification divisor .
then @xmath175 has @xmath176 points .
therefore @xmath177 has @xmath178 the formula in theorem [ thm : general - fibr ] says that @xmath179 . therefore @xmath180 .
this happens for all choices of coverings associated to all representations @xmath181 .
hence @xmath182 , for all @xmath183 .
another case that we are going to use later is the following .
[ cor : thm1 ] let ( [ fibration ] ) be a fibration over @xmath184 with finite monodromy , such that @xmath68 is of balanced type . then @xmath73 is also of balanced type and its e - polynomial is given by @xmath185 first note that @xmath186 , so the monodromy @xmath108 , being a quotient @xmath187 , must be abelian . moreover , there is some @xmath188 such that @xmath189 , and the covering associated to @xmath190 is of balanced type ( is @xmath191 again ) .
hence @xmath111 is of balanced type , since there are coverings @xmath192 .
now let @xmath158 be an irreducible representation of @xmath103 . if @xmath193 , the trivial representation , then @xmath194 .
if @xmath158 is a non - trivial representation , then take the covering @xmath171 associated to @xmath158 .
again there are coverings @xmath195 .
the hodge monodromy representation associated to @xmath164 is @xmath196 , where @xmath170 . in cohomology @xmath197
are surjections , but the composition is an isomorphism ( multiplication by @xmath23 ) .
therefore @xmath198 , and @xmath199 .
this happens for all choices of @xmath158 , so it must be @xmath200 , for any irreducible non - trivial @xmath158 . we need to recall some e - polynomials from @xcite which serve as building blocks for the e - polynomials of character varieties for any genus @xmath201 .
they are basically associated to the case of @xmath202 .
first , we have that @xmath203 and @xmath204 .
consider the following subsets of @xmath1 : * @xmath205 conjugacy class of @xmath206 .
it has @xmath207 .
* @xmath208 conjugacy class of @xmath209 .
it has @xmath210 .
* @xmath211 conjugacy class of @xmath212 .
it is @xmath213 , with @xmath214 .
it has @xmath215 .
* @xmath216 conjugacy class of @xmath217 .
it is @xmath218 and @xmath219 .
* @xmath220 conjugacy class of @xmath221 , where @xmath222 .
note that @xmath223 , since the matrices @xmath224 and @xmath225 are conjugated . we have @xmath226 , where @xmath227 .
so @xmath228 .
* we also need the set @xmath229 , which is the union of the conjugacy classes @xmath230 , @xmath231 .
this has @xmath232 .
now consider the map @xmath233=aba^{-1}b^{-1}.\end{aligned}\ ] ] note that @xmath234 is equivariant under the action of @xmath1 by conjugation on both spaces .
we stratify @xmath235 as follows : * @xmath236 , * @xmath237 , * @xmath238 , * @xmath239 , * @xmath240
. we also introduce the varieties @xmath241 for fixed @xmath242 and define accordingly * @xmath243 .
there is a fibration @xmath244 , so @xmath245 .
* @xmath246 .
there is a fibration @xmath247 , and @xmath248 .
* @xmath249 , for @xmath250 .
we define also @xmath251 .
there is a fibration @xmath252 , and @xmath253
. it will also be convenient to define * @xmath254=\left(\begin{array}{cc } \lambda & 0 \\ 0 & \lambda^{-1}\\ \end{array } \right ) , \text { for some } \lambda\ne0 , \pm1,\ , a,\,b \in \operatorname{sl}(2,{\mathbb{c}})\right\}$ ] .
there is an action of @xmath87 on @xmath255 given by interchanging @xmath256 , with @xmath257 .
the set @xmath258 coincides with the union of all @xmath259 .
the hodge - deligne polynomials computed in @xcite are as follows : @xmath260 there is a fibration @xmath261 let @xmath262 be loops around the points @xmath263 , respectively .
the monodromies around @xmath264 are trivial , and the monodromy around @xmath265 is of order @xmath59 .
so the monodromy group is @xmath266 and the hodge monodromy representation is computed in ( * ? ? ?
* theorem 6.1 ) , @xmath267 where @xmath56 is the trivial representation and @xmath57 the non - trivial one
. then corollary [ cor : general - fibr ] gives @xmath268 ( where we have written @xmath269 , @xmath270 , @xmath271 ) .
taking the quotient by @xmath87 , we have the fibration @xmath272 and the hodge monodromy representation is a representation of the group @xmath273 , generated by the loops @xmath274 around the points @xmath275 , respectively . the hodge monodromy representation is computed in ( * ? ? ?
* theorem 7.1 ) , @xmath276 where @xmath56 is the trivial representation , @xmath277 is the representation that is non - trivial around the loop @xmath278 and @xmath279 .
now writing @xmath280 , corollary [ cor : general - fibr ] says that @xmath281 note that if @xmath280 , then @xmath282 , which agrees with ( [ eqn : rx4 ] ) .
to recover @xmath283 , we note that @xmath284 by ( * ? ? ?
* proposition 3.2 ) , we have that @xmath285 and @xmath286 . using ( [ eqn:+- ] ) , @xmath287 now using ( [ formulaepolyx4z2 ] ) and ( [ eqn : e(x4-bar ) ] ) we get @xmath288
we start by computing the e - polynomial of the twisted character variety @xmath289 for a curve @xmath3 of genus @xmath290 .
this space can be described as the quotient @xmath291 where @xmath292[a_2,b_2][a_3,b_3]=-\text{id } \rbrace.\ ] ] notice that we only need to consider a geometric quotient , since all representations are irreducible : if there was a common eigenvector @xmath293 for @xmath294 , we would obtain that @xmath295(v)=[a_2,b_2](v)=[a_3,b_3](v)=v$ ] and therefore @xmath295[a_2,b_2][a_3,b_3](v)=v\neq -\operatorname{id}(v)=-v$ ] . to compute the e - polynomial of the moduli space , we will stratify the space in locally closed subvarieties and compute the e - polynomial of each stratum , since e - polynomials are additive with respect to stratifications .
we stratify the space @xmath296 according to the possible values of the traces of the commutators .
to simplify the notation , we write @xmath295=\xi_{1 } , [ a_2,b_2]=\xi_{2 } , [ a_3,b_3]=\xi_{3}$ ] .
consider the map @xmath297 we are interested in the following condition : if @xmath298 share an eigenvector @xmath293 with eigenvalue @xmath299 in each case , then in a suitable basis @xmath300=\begin{pmatrix } \lambda_{1 } & \ast \\ 0 & \lambda_{1}^{-1 } \end{pmatrix } , \quad [ a_2,b_2]=\begin{pmatrix } \lambda_{2 } & \ast \\ 0 & \lambda_{2}^{-1 } \end{pmatrix } , \quad [ a_3,b_3]= \begin{pmatrix } \lambda_{3 } & \ast \\ 0 & \lambda_{3}^{-1 } \end{pmatrix},\ ] ] and therefore , as @xmath301 , we obtain that @xmath302 .
other possibility is that @xmath293 has eigenvalue @xmath303 for @xmath304 , @xmath305 for @xmath306 and @xmath307 for @xmath308 , yielding @xmath309 , etc .
working out all possibilities , we have that @xmath310 can share an eigenvector when @xmath311 equivalently , in terms of the traces , when @xmath312 this defines a ( smooth ) quadric @xmath313 , depicted in figure [ fig1 ] . .,width=7 ] [ lem : c ] @xmath298 share an eigenvector if and only if @xmath314 .
we have already seen the if part . for the only if part , note that if @xmath315 satisfies ( [ eqntraces ] ) then the eigenvalues @xmath316 satisfy one of the four relations in ( [ eqntracesbis ] ) . changing some @xmath317 by @xmath318 if necessary
, we can suppose that @xmath319 .
suppose that some @xmath320 is diagonalizable .
without loss of generality , we suppose it is @xmath304 , and choose a basis with @xmath321 .
write @xmath322 .
then we have the equation @xmath323 we get the equations @xmath324 and @xmath325 . hence @xmath326 and @xmath327 . if @xmath328 , it must be @xmath329 ; thus @xmath330 and hence @xmath331 , which implies that the matrices share an eigenvector .
if @xmath332 , then @xmath333 , and @xmath334 , so the matrices share their eigenvectors .
now suppose that all @xmath320 are not diagonalizable , so they are of jordan type .
let @xmath335 be the only eigenvector ( up to scalar multiples ) of @xmath320 . if @xmath320 do not share an eigenvector , then @xmath336 is a basis , on which @xmath337 , @xmath338 , where @xmath339 .
then @xmath340 hence @xmath341 .
so @xmath331 , which is a contradiction .
we stratify the space @xmath296 according to the traces of @xmath310 .
consider the planes @xmath342 and the quadric @xmath49 above . then consider as well the intersections of these seven subvarieties .
this gives the required stratification .
we shall compute the e - polynomials of the chunk of @xmath296 lying above each of these strata , starting by the lower - dimensional ones ( points ) and going up in dimension .
we shall use some of the polynomials computed in sections 8 - 12 in @xcite which correspond to some of the strata of the spaces of representations for the case of a curve of genus @xmath54 .
we shall point out which stratum we use each time .
the intersections of three planes , or of the quadric and two planes , is the collection of eight points given by @xmath343 .
let @xmath344 be the subset of @xmath345 with traces given by these possibilities .
note that if @xmath346 then @xmath347 or @xmath320 is of jordan type @xmath348 ; analogously , if @xmath349 then @xmath350 or @xmath320 is of jordan type @xmath351 .
* @xmath352 . then @xmath353 are all of jordan type ( if @xmath354 then @xmath355 , so @xmath356 and @xmath357 ) . choosing an adequate basis , we can assume that @xmath295=j_{+}$ ] , so @xmath358[a_3,b_3]=-(j_{+})^{-1 } = j_{-}$ ] .
this set has been determined in ( * ? ? ?
* stratum @xmath359 , section 12 ) .
it has polynomial @xmath360 .
therefore @xmath361 * @xmath362 and the cyclic permutations ( accounting for three cases ) .
if @xmath363 , then @xmath358[a_3,b_3]=-\operatorname{id}$ ] , with @xmath364 , and @xmath365 . if @xmath366 then @xmath367 ; and if @xmath368 then @xmath369
this produces the contribution ( multiplying by @xmath290 because of the three cyclic permutations ) @xmath370 if @xmath371 , then choosing an adequate basis , we can assume that @xmath295=j_{+}$ ] , so @xmath358[a_3,b_3]=-(j_{+})^{-1 } = j_{-}$ ] .
this set has been determined in ( * ? ? ?
* stratum @xmath372 , section 12 ) to be @xmath373
. therefore @xmath374 finally @xmath375 * @xmath376 and cyclic permutations ( which account for three cases ) .
if @xmath377 then it must be @xmath378 , which is not the case .
therefore @xmath379 is of jordan type @xmath380 . choosing a suitable basis , we can write @xmath381 , so @xmath295[a_2,b_2]=-j_{-}^{-1}= j_+$ ] .
this set has been determined in ( * ? ? ?
* stratum @xmath359 , section 11 ) and it has polynomial @xmath382
. therefore @xmath383 * @xmath384 . if @xmath377 , then @xmath358[a_3,b_3]=\operatorname{id}$ ] , with @xmath385 , and @xmath365 . if @xmath386 then @xmath367 ; and if @xmath387 then @xmath369 .
this produces the contribution @xmath388 if @xmath389 , then choosing an adequate basis , we can assume that @xmath295=j_{-}$ ] , so @xmath358[a_3,b_3]=-(j_{-})^{-1 } = j_{+}$ ] .
this set has been determined in ( * ? ? ?
* stratum @xmath390 , section 11 ) to be @xmath391 .
therefore @xmath392 hence @xmath393 adding all up , we obtain @xmath394 the intersection of two of the planes are the lines @xmath395 , and the cyclic permutations of these .
the intersection of one of the planes and the quadric is given by the lines @xmath396 , @xmath397 and the cyclic permutations .
we denote by @xmath398 the portion of @xmath296 lying over these lines , and we stratify in the following sets : * @xmath399 given by @xmath400 , @xmath401 and @xmath402 .
note that if @xmath403 then @xmath404 , with @xmath405 and @xmath406 , a contradiction .
then @xmath304 is of jordan type .
choosing an adequate basis , we can assume that @xmath407 .
then @xmath358[a_3,b_3]=-(j_+)^{-1}=j_{-}$ ] , where @xmath408 and @xmath409 .
this set has been computed in ( * ? ? ?
* stratum @xmath410 , section 12 ) and it has e - polynomial @xmath411 .
so @xmath412 * @xmath413 given by @xmath414 , @xmath401 and @xmath402 .
again @xmath304 is of jordan type .
fixing a basis we have @xmath415 and @xmath416 , where @xmath405 and @xmath417 .
this set has been computed in ( * ? ? ?
* stratum @xmath410 , section 11 ) and it has e - polynomial @xmath411 .
so @xmath418 * @xmath419 given by @xmath420 , @xmath363 and @xmath421 .
now @xmath355 , @xmath422 .
this is computed in ( * ? ? ?
* stratum @xmath423 , section 9 ) , @xmath424 .
so @xmath425 * @xmath426 given by @xmath420 , @xmath371 . using an adequate basis , we have @xmath427 , and @xmath428 , and @xmath422 .
this is computed in ( * ? ? ?
* stratum @xmath429 , section 12 ) , @xmath430 .
so @xmath431 * @xmath432 given by @xmath433 , @xmath377
. then @xmath434 , with @xmath435 .
this is computed in ( * ? ? ?
* stratum @xmath436 , section 8) , in @xcite , @xmath437 . ]
so @xmath439 * @xmath440 given by @xmath433 , @xmath441 .
choosing a basis so that @xmath415 , @xmath442 , with @xmath443 .
this is computed in ( * ? ? ?
* stratum @xmath429 , section 11 ) , @xmath444 .
so @xmath445 considering the possible permutations , and adding the e - polynomials of the strata , we get @xmath446 now consider the planes given by the equations @xmath447 , from which we remove the previous strata .
we do the case @xmath448 and multiply by three the result to account for the three cases @xmath449 .
the planes are given by @xmath448 , where @xmath450 and @xmath451 .
note that it can not be @xmath452 , since this would imply @xmath453 .
we have the following cases : * @xmath454 given by @xmath371 .
this implies that , in a suitable basis , @xmath455 , together with the previous restrictions for the traces .
this is the set given in ( * ? ? ?
* stratum @xmath456 , section 12 ) , which has e - polynomial @xmath457 .
so @xmath458 * @xmath459 given by @xmath441 .
this implies that , in a suitable basis , @xmath460 , together with the restrictions @xmath461 , @xmath462 .
this is the set given in ( * ? ? ?
* stratum @xmath456 , section 11 ) , which has e - polynomial @xmath463 .
so @xmath464 the total contribution of the planes is @xmath465 we now deal with the part of @xmath296 lying over the quadric @xmath49 , with @xmath466 . as we saw in lemma
[ lem : c ] , this corresponds to the case that @xmath467 share ( at least ) one eigenvector .
we deal now with the case where @xmath468 and @xmath469 share just one eigenvector ( up to scalar multiples ) , which we denote by @xmath293 . using it as the first vector of a basis that diagonalizes @xmath379 , we can arrange that @xmath470 where @xmath471 , @xmath472 are the eigenvalues of @xmath468 associated to the eigenvector @xmath293 .
our choice of basis for the above expressions of @xmath298 gives us a slice for the @xmath473-action , since their stabilizer is trivial .
so we shall only need to multiply by @xmath474 after computing the e - polynomial of .
this set can be regarded as a fibration @xmath475 with fiber @xmath476 to compute its e - polynomial , we would like to extend the fibration to the six curves @xmath477 and apply corollary [ cor : thm1 ] .
we can not extend the fibration ; however we can extend the local system definining the hodge monodromy fibration . by ( [ eqn : rx4 ] ) ,
the hodge monodromy fibration @xmath478 is trivial over @xmath479 , and it is of order @xmath59 over @xmath480 . consider the projections : @xmath481 then @xmath482 the hodge monodromy fibration @xmath483 can be extended ( locally trivially ) over the lines @xmath484 , @xmath485 and @xmath486 , to a hodge monodromy fibration @xmath487 over @xmath488 .
moreover , the monodromy around @xmath480 and @xmath489 is of order two .
the corresponding group is @xmath490 and the representation ring is generated by representations @xmath491 , where @xmath56 is the trivial representation , @xmath492 is the representation with non - trivial monodromy around the origin of the first copy of @xmath493 , and @xmath494 is the representation with non - trivial monodromy around the origin of the second copy of @xmath493 . pulling back the hodge monodromy representation of @xmath495 given in ( [ eqn : rx4 ] ) , we have that @xmath496 , @xmath497 and @xmath498 , where @xmath270 , @xmath271 . so the hodge monodromy representation of @xmath499 is @xmath500 we extend @xmath483 to a hodge monodromy fibration @xmath487 over @xmath488 with the same formula , and compute its e - polynomial , applying corollary [ cor : thm1 ] , @xmath501 now we substract the contribution of @xmath487 over the lines @xmath479 , @xmath485 and @xmath486 . * consider the curve defined by @xmath502 , @xmath503 .
the fibration has the hodge monodromy of a fibration over @xmath504 with fiber @xmath505 this has hodge monodromy representation equal to @xmath506 , where @xmath507 .
this is equal to @xmath508 using corollary [ cor : general - fibr ] and ( [ eqn : e(x4-bar ) ] ) , we get that the contribution equals @xmath509 * the computation for @xmath510 , @xmath511 is analogous and gives the same quantity . * by symmetry , the contribution for @xmath485 , and for @xmath512 is the same as for @xmath479 .
so we have to multiply ( [ eqn:333 ] ) by @xmath513 .
* the contribution of the four points @xmath514 is @xmath515 .
therefore , the e - polynomial of the original fibration is : @xmath516 and @xmath517 suppose now that @xmath518 and @xmath298 share two eigenvectors .
then they can all be simultaneously diagonalized . with respect to a suitable basis , we have that : @xmath300= \begin{pmatrix } \lambda & 0 \\ 0 & \lambda^{-1 } \end{pmatrix } , \quad [ a_2,b_2]=\begin{pmatrix } \mu & 0 \\ 0 & \mu^{-1 } \end{pmatrix } , \quad [ a_3,b_3]=\begin{pmatrix } -\lambda^{-1 } \mu^{-1 } & 0 \\ 0 & -\lambda \mu \end{pmatrix}.\ ] ] this defines a fibration @xmath519 , with fiber @xmath520 the stabilizer of @xmath298 is @xmath521 , where @xmath522 are the diagonal matrices and the @xmath523-action is given by the simultaneous permutation of the eigenvalues , i.e , by conjugation by @xmath524 .
therefore the stratum we are analysing is @xmath525 the action on the basis of @xmath526 takes @xmath527 to @xmath528 , producing a fibration @xmath529 \ar[d ] & \overline{z}':=\overline{z}/\mathbb{z}_{2 } \ar[d ] \\ { b } \ar[r ] & { b}'= { b}/\mathbb{z}_{2}}\ ] ] if we write @xmath530 ( @xmath87 acts on @xmath493 by @xmath531 ) , we can obtain three pullback bundles @xmath532 \ar[d ] & \overline{x}_{4}/\mathbb{z}_{2 } \ar[d ] \\ { b } ' \ar[r]^>>>>>{\pi_{i } } & ( \mathbb{c}^ * - \lbrace \pm 1 \rbrace ) /\mathbb{z}_{2 } } \ ] ] @xmath449 , such that @xmath533 , with @xmath534 . as a consequence ,
if we write @xmath280 , as in ( [ eqn : rx4z2 ] ) , the hodge monodromy representation is @xmath535 where @xmath536 and @xmath537 . to obtain the e - polynomial of the total space , we need to substitute each representation by its associated e - polynomial , by theorem [ thm : general - fibr ] .
we have * @xmath538 , * @xmath539 , * @xmath540 , for @xmath541 .
recall that the basis is @xmath542 .
we compute the e - polynomial of each representation case by case : * @xmath543 . using ( [ eqn:+- ] ) , we compute @xmath544 , since @xmath545 and @xmath546 .
also @xmath547 .
so @xmath548 .
* @xmath549 . since @xmath550
, we need to compute the e - polynomial of the pullback bundle of the fibration @xmath551 that maps @xmath552 , and which has hodge monodromy representation equal to @xmath553 .
the pullback bundle is @xmath554^{p } \ar[r ] & \mathbb{c}- \lbrace 0 , \pm 1\rbrace \ar[d]^{g } \\ { b } ' \subset \bar b'=(\mathbb{c}- \lbrace 0 , \pm 1 \rbrace)\times ( \mathbb{c}- \lbrace 0,\pm 1 \rbrace)/\mathbb{z}_{2 } \ar[r]^-{\pi_{1 } } & \mathbb{c } - \lbrace \pm 2 \rbrace \cong ( \mathbb{c}- \lbrace 0,\pm 1 \rbrace)/\mathbb{z}_{2}}\ ] ] where @xmath555 and @xmath556 is the quotient map .
@xmath557 , and @xmath558
. then @xmath559 , @xmath560 .
so @xmath561 .
finally , @xmath562 + the diagram still commutes if we change @xmath563 by @xmath564 or @xmath565 , so we infer that @xmath566 as local systems .
this implies that certain reductions can be made : since @xmath567 , whenever @xmath568 or @xmath569 appear in a tensor product we can switch one by another in order to simplify the expression . for example : @xmath570 @xmath571 and so on .
in fact , we deduce from this that , for all representations @xmath572 that appear in ( [ eqnlongrepresentations ] ) , either @xmath573 , @xmath574 , @xmath575 or @xmath576 , as any representation where @xmath577 appears can be simplified to one of these expressions . *
@xmath578 . to obtain @xmath578 , we take the pullback under the map @xmath579 , @xmath580 , which ramifies at @xmath581 . the pullback fibration is @xmath582 \ar[d]^{p } & x = y+y^{-1 } \in \mathbb{c } - \lbrace \pm 2,0 \rbrace \ar[d]^{2:1 }
\\ b'\subset \bar{b}'=\lbrace [ ( \lambda = y^{2},\mu)]\rbrace \ar[r ] & x^{2}-2=y^{2}+y^{-2}=\lambda+\lambda^{-1 } \in \mathbb{c}- \lbrace \pm 2 \rbrace}\ ] ] where @xmath583 , @xmath584 . therefore
@xmath585 . substracting the contribution corresponding to the two hyperbolas @xmath586 ,
which has e - polynomial @xmath587 , we get @xmath588 , and @xmath589 the diagram for @xmath590 is similar : @xmath591 \ar[d]^{p } & x = y+y^{-1 } \in \mathbb{c } - \lbrace \pm 2,0 \rbrace \ar[d]^{2:1 } \\ b'\subset \bar{b}'=\lbrace [ ( \lambda =- y^{2},\mu)]\rbrace \ar[r ] & 2-x^2=y^{2}+y^{-2}=\lambda+\lambda^{-1 } \in \mathbb{c}- \lbrace \pm 2 \rbrace}\ ] ] but now @xmath592 . since @xmath593 , we obtain that @xmath594 and analogous computations yield that @xmath595 . *
@xmath596 . to compute the e - polynomial of this representation
, we can use the fibration with hodge monodromy representation @xmath597 , which corresponds to the fibered product of two copies of the pullback fibration @xmath598 , one with @xmath599 , the other with @xmath600 .
the total space is @xmath601\},\ ] ] where @xmath602 and @xmath603 .
the e - polynomial is @xmath604 .
recall that we need to substract the contribution of the ( now four ) hyperbolas given by @xmath605 and @xmath606 .
the action @xmath607 acts on each of the first two hyperbolas giving a contribution of @xmath587 , whereas it interchanges the last two , which gives @xmath608 .
we get @xmath609 and @xmath610 computing the different fibered products of @xmath611 and @xmath598 gives us @xmath612 changing the projection does not alter the computations , so @xmath613 .
* @xmath614 . to compute this e - polynomial
, we can compute the e - polynomial of the representation @xmath615 , which corresponds to the fibered product of three copies of @xmath598 .
the total space of this fibered product is given by @xmath616 where @xmath617 and @xmath618 , @xmath619 and @xmath607 .
taking into account the two possible signs for @xmath620 , and substracting the contribution from the hyperbolas in @xmath621 , we obtain that the e - polynomial is @xmath622 .
this implies that @xmath623 an analogous computation gives the same polynomial for @xmath624 and cyclic permutations of signs .
* @xmath625 . as we did in the previous case , to compute the e - polynomial it suffices to take the fibered product of two copies of @xmath598 and @xmath611 .
the total space is again parametrized by @xmath616 where now @xmath602 and @xmath626 .
in particular , this implies that @xmath627 and @xmath628 , which gives two components that get identified under the @xmath523-action given by @xmath629 .
therefore , the quotient is parametrized by @xmath630 , which produces @xmath631 . substracting the four hyperbolas
, we get that the e - polynomial of the fibered product is @xmath632 .
so @xmath633 substituting the values just obtained for the e - polynomial of every irreducible representation in ( [ eqnlongrepresentations ] ) , and the values @xmath634 , we obtain the e - polynomial of the total fibration @xmath635 using the formula ( [ eqn : rx4-rx4z2->ex4 ] ) , we get that @xmath636 let us deal finally with the case @xmath637 , @xmath638 , which corresponds to the open subset @xmath639 in @xmath640 defined by those representations whose @xmath310 are diagonalizable and do not share an eigenvector . choosing a basis that diagonalizes @xmath379 , note that ,
if we write @xmath641 , then @xmath642 so that @xmath643 and @xmath644 ( see lemma [ lem : c ] ) . conjugating by a diagonal matrix , we can assume that @xmath645 .
as @xmath646 and @xmath647 , we have that @xmath648 are determined by the values of @xmath649 ; and @xmath650 is determined by the equation @xmath651 .
we see that for fixed @xmath652 , @xmath653 are fully determined , and so are @xmath654 .
consider the @xmath655-cover given by @xmath656 over @xmath657 .
let @xmath658 be the pull - back fibration .
the fiber over @xmath659 is isomorphic to @xmath660 let @xmath661 .
then @xmath295=\xi_1 $ ] , @xmath358=\xi_2 $ ] , @xmath662=\xi_3 $ ] , where @xmath663 .
take @xmath664 , @xmath665 matrices such that @xmath666 and @xmath667
. then @xmath668 identifies @xmath658 with the subset of @xmath669 where @xmath670 .
the second and third copies of @xmath87 act as the standard @xmath87-action on the second and third copies of @xmath255 , respectively .
the action of the first copy of @xmath87 is more delicate : it acts as conjugation by @xmath671 , sending @xmath672 , @xmath673 . then under the isomorphism @xmath674 ,
it acts by conjugation by @xmath675 ( resp .
@xmath676 ) on the second ( resp .
third ) factor of @xmath255 .
this matrix is easily computed to be diagonal , so the action is ( homologically ) trivial .
the conclusion is that @xmath677 is isomorphic to the open set of @xmath678 , where @xmath670 .
now we need to compute the e - polynomial of @xmath679 . here
we can parametrize @xmath680 .
the fiber over @xmath681 is isomorphic to @xmath682 , hence this space is isomorphic to @xmath683 , studied in ( [ eqn : barz ] ) . using ( [ eqn : ez ] ) ,
we get @xmath684 if we add all the e - polynomials of the different strata , we get @xmath685 then @xmath686 this agrees with the result in @xcite , obtained by arithmetic methods .
we introduce the following sets associated to the representations of a genus @xmath59 complex curve , and give the e - polynomials computed in @xcite : * @xmath687[a_2,b_2]=\operatorname{id}\}$ ] .
then @xmath688 , by ( * ? ? ?
* section 8.1 ) .
* @xmath689[a_2,b_2]=-\operatorname{id}\}$ ] .
then @xmath690 , by ( * ? ? ?
* section 9 ) .
* @xmath691[a_2,b_2]=j_+=\small{\left ( \begin{array}{cc } 1 & 1\\ 0 & 1\\ \end{array } \right)}\right\}$ ] , @xmath692 , by ( * ? ? ?
* section 11 ) . * @xmath693[a_2,b_2]=j_-=\small{\left ( \begin{array}{cc } -1 & 1\\ 0 & -1\\ \end{array } \right)}\right\}$ ] .
then @xmath694 , by ( * ? ? ?
* section 12 ) .
* @xmath695[a_2,b_2]=\small{\left ( \begin{array}{cc } \lambda & 0\\ 0 & \lambda^{-1}\\ \end{array } \right ) } \right\ } $ ] , for @xmath696 .
then @xmath697 , by ( * ? ? ?
* section 10 ) .
let @xmath698[a_2,b_2]=\small{\left ( \begin{array}{cc } \lambda & 0\\ 0 & \lambda^{-1}\\ \end{array } \right)},\ , \lambda\neq 0,\pm 1\right\}$ ] .
we have a fibration @xmath699 if we take the quotient by the @xmath87-action there is another fibration @xmath700 we are interested in the hodge monodromy representations @xmath701 and @xmath702 .
[ prop : ry4 ] @xmath703 .
we follow the stratification @xmath704 given in , and study the behaviour of each stratum when @xmath52 varies in @xmath705 to obtain the hodge monodromy representation of @xmath706 .
let @xmath707 .
as in ( * ? ? ?
* section 10 ) , we write @xmath708= \begin{pmatrix } a & b \\ c & d \end{pmatrix } , \quad \delta=[a_{1},b_{1 } ] = \xi\nu = \begin{pmatrix } \lambda a & \lambda b \\ \lambda^{-1}c & \lambda^{-1}d \end{pmatrix},\ ] ] and @xmath709 , @xmath710 .
note that every @xmath711 determines @xmath648 by @xmath712 then @xmath713 .
we look at the strata : * @xmath714 , corresponding to @xmath715 . in this case , both @xmath716 are of jordan type . if we take the basis given by @xmath717 , where @xmath718 is an eigenvector for @xmath719 and @xmath720 an eigenvector for @xmath721 , then @xmath722 for certain @xmath723 .
now , since @xmath724 , we obtain that @xmath725 .
we can fix @xmath726 by rescaling the basis , so @xmath727 is fixed and there is no monodromy around the origin . therefore @xmath728 where @xmath56 is the trivial representation .
* @xmath729 , corresponding to @xmath730 and @xmath731 .
we focus on the first case .
in this situation , @xmath331 , so there are three possibilities : either @xmath732 ( in which case @xmath733 ) or @xmath734 or @xmath735 ( in either case there is a parameter in @xmath493 and @xmath719 is of jordan type ) . in every situation
, @xmath719 has trivial monodromy , whereas @xmath736 .
this contributes @xmath737 .
therefore @xmath738 * @xmath739 , given by @xmath740 and @xmath741 .
this is analogous to the previous case , so @xmath742 * @xmath410 , defined by @xmath743 and @xmath744 .
both cases are similar , so we do the first case . for each @xmath52 ,
@xmath745 move in a punctured line @xmath746 , where @xmath719 is of jordan form and @xmath721 is of diagonal type , with trace @xmath747 .
both families can be trivialized , giving a contribution of @xmath748 and @xmath749 .
the missing fiber @xmath259 over @xmath750 , which needs to be removed , has monodromy representation @xmath737 as @xmath52 varies .
therefore @xmath751 the factor @xmath752 corresponds to the fact that now @xmath644 , so there is the extra freedom given by @xmath493 .
* @xmath753 , defined by @xmath754 and @xmath755 .
similarly to @xmath410 , we obtain @xmath756 * @xmath456 .
this stratum corresponds to the set @xmath757 , which is a hyperbola @xmath758 for every @xmath52 .
since @xmath331 , we get a contribution of @xmath759 , arising from the disjoint cases @xmath732 ; @xmath734 ; and @xmath760 .
parametrizing @xmath758 by a parameter @xmath761 as in ( * ? ? ?
* section 10 ) , we obtain a fibration over @xmath762 whose fiber over @xmath763 is @xmath764 , for fixed @xmath52 .
when @xmath52 varies over @xmath765 , note that we can extend the local system trivially to the cases @xmath766
. this extension can be regarded as a local system over the set of @xmath767 @xmath768 where @xmath769 maps @xmath770 .
the hodge monodromy representation of @xmath771 belongs to @xmath772 $ ] ( with generators @xmath773 denoting the representation which is not trivial over the generator of the fundamental group of the first and second copies of @xmath493 respectively , and @xmath774 ) . since @xmath775
, we get @xmath776 we write this as @xmath777 . to obtain the hodge monodromy representation over @xmath778
, we use the projection @xmath779 , @xmath780 .
then @xmath781 , @xmath782 , @xmath783 , @xmath784 for the representations .
using that @xmath785 and @xmath786 and substracting the contribution from the sets @xmath787 , which yield @xmath788 , we get @xmath789 and @xmath790 * @xmath791 , corresponding to the open stratum given by the set of @xmath745 such that @xmath792 , @xmath793 and @xmath794 .
if we forget about the condition @xmath795 , @xmath791 is a fibration over @xmath745 with fiber isomorphic to @xmath796 , @xmath797 , @xmath793 .
its monodromy is trivial , as the local system is trivial when @xmath52 varies .
the contribution over @xmath758 , already computed in the previous stratum , is @xmath798 .
so we get @xmath799 adding all pieces , we get @xmath800 dividing by @xmath801 , we get the formula ( [ eqn : rcm ] ) .
we want to compute the hodge monodromy representation of @xmath802 .
we have the following .
the hodge monodromy representation @xmath803 is of the form @xmath804 , for some polynomials @xmath805 $ ] .
the hodge monodromy representation @xmath803 lies in the representation ring of the fundamental group of @xmath806 . under the double cover @xmath807
, it reduces to @xmath808 . by proposition [ prop : ry4 ] , @xmath808 is of order @xmath59 .
hence @xmath803 has only monodromy of order @xmath59 over the loops @xmath278 around the points @xmath809 .
this is the statement of the lemma . to compute @xmath805 $ ]
, we compute the e - polynomial of the twisted @xmath1-character variety ( [ eqn : twisted ] ) in another way .
stratify @xmath810[a_2,b_2]=-[a_3,b_3]\}\ ] ] as follows : * @xmath811[a_2,b_2]=-[b_3,a_3]=\operatorname{id}\rbrace$ ] .
then @xmath812 * @xmath813[a_2,b_2]=-[b_3,a_3]=-\operatorname{id}\rbrace$ ] .
then @xmath814 * @xmath815[a_2,b_2]=-[b_3,a_3]\sim j_{+ } \rbrace$ ] .
then @xmath816 * @xmath817[a_2,b_2]=-[b_3,a_3]\sim j_{- } \rbrace$ ]
. then @xmath818 * @xmath819[a_2,b_2]=-[b_3,a_3]\sim \begin{pmatrix } \lambda & 0 \\ 0 & \lambda^{-1 } \end{pmatrix } , \lambda \neq 0 , \pm 1 \rbrace$ ] .
* @xmath820[a_2,b_2]=-[b_3,a_3]=\begin{pmatrix } \lambda & 0 \\ 0 & \lambda^{-1 } \end{pmatrix } , \lambda \neq 0 , \pm 1 \rbrace$ ] . * @xmath821[a_2,b_2]=-[b_3,a_3]=\begin{pmatrix } \lambda & 0 \\ 0 & \lambda^{-1 } \end{pmatrix}\rbrace$ ] , where @xmath822 . using the formula in section [ subsec : finalresult ] ,
@xmath823 for the last stratum , note that : @xmath824 so @xmath825 , where @xmath826 , @xmath827 . then @xmath828 using ( [ eqn : rx4-rx4z2->ex4 ] )
, we get @xmath829 which gives us the equation @xmath830 we can obtain another equation if we recall that @xmath831[a_{2},b_{2}]\sim \begin{pmatrix } \lambda & 0 \\ 0 & \lambda^{-1 } \end{pmatrix } , \lambda \neq 0 , \pm 1\}.\ ] ] using that @xmath832 , we obtain that : @xmath833 but the e - polynomial of @xmath834 can again be obtained using the hodge monodromy representation @xmath803 , using ( [ eqn : rx4-rx4z2->ex4 ] ) , @xmath835 finally , two more equations arise from the e - polynomial of the fiber of @xmath836 , @xmath837 and from the hodge monodromy representation @xmath838 given in proposition [ prop : ry4 ] . since @xmath839 , we get the equation @xmath840 from equations , , and ( [ eqn : hmry4z2ec4 ] ) , we find @xmath841 we have proved : [ ry4z2 ] @xmath842 .
let @xmath843 be the character variety of a genus @xmath290 complex curve @xmath3 , i.e , the moduli space of semisimple representations of its fundamental group into @xmath844 .
it can be defined as the space @xmath845 where @xmath846[a_2,b_2][a_3,b_3]=\operatorname{id}\rbrace.\ ] ] we stratify @xmath847 as follows : * @xmath848[a_2,b_2]=[b_3,a_3]=\operatorname{id}\rbrace$ ] . then @xmath849 * @xmath850[a_2,b_2]=[b_3,a_3]=-\operatorname{id}\rbrace$ ] . then @xmath851 * @xmath852[a_2,b_2]=[b_3,a_3]\sim j_{+ } \rbrace$ ] . @xmath853 * @xmath854[a_2,b_2]=[b_3,a_3]\sim j_{- } \rbrace$ ] . @xmath855 * @xmath856[a_2,b_2]=[b_3,a_3]\sim \begin{pmatrix } \lambda & 0 \\ 0 & \lambda^{-1 } \end{pmatrix } , \lambda \neq 0 , \pm 1 \rbrace$ ] . for computing @xmath857 ,
we define @xmath858[a_2,b_2]=[b_3,a_3]= \begin{pmatrix } \lambda & 0 \\ 0 & \lambda^{-1 } \end{pmatrix } \right\},\ ] ] for @xmath859 .
there is a fibration @xmath860 with fiber @xmath861 .
note that @xmath862 , so @xmath863 which we write as @xmath864
. if we apply , @xmath865 from this @xmath866 in ( * ? ? ?
* section 8) , this analysis is done in the case of @xmath54 , by stratifying the set of irreducible orbits , and computing the e - polynomial of each stratum .
the number of strata increases rapidly with the genus .
therefore , for @xmath0 we are going to follow the method in @xcite which consists on computing the e - polynomial of the reducible locus ( which has fewer strata ) and substracting it from the total . a reducible representation given by @xmath294 is @xmath158-equivalent to @xmath867 under the equivalence relation @xmath868 given by the permutation of the eigenvectors . under the action @xmath869 we have that @xmath870 , @xmath871 , so we obtain @xmath872 a reducible representation happens when there is a common eigenvector .
so in a suitable basis , it is @xmath873 this is parametrized by @xmath874 .
the condition @xmath875[a_{2},b_{2}][a_{3},b_{3}]=\operatorname{id}$ ] is rewritten as @xmath876 there are four cases : * @xmath877 given by @xmath878 and @xmath879
. then we can conjugate the representation ( [ eqn : bb ] ) to the diagonal form ( [ eqn : aa ] ) . in this case
we can suppose all @xmath880 , and the stabilizer are the diagonal matrices @xmath881 .
there is an action of @xmath87 given by interchanging of the two basis vectors , and if we write @xmath882 , the stratum is @xmath883 .
note that @xmath884 , @xmath885 , and @xmath886 , @xmath887 .
so @xmath888 * @xmath889 given by @xmath890 and @xmath879 .
then ( [ eqn : cc ] ) determines a hyperplane @xmath891 , and the condition for @xmath892 defines a line @xmath893 . if @xmath894 denotes the upper triangular matrices , we have a surjective map @xmath895 and the fiber is isomorphic to @xmath896 .
so @xmath897 * @xmath898 , given by @xmath899 , @xmath900 .
this is the case where @xmath901 , @xmath449 , which gives @xmath902 points .
therefore @xmath903 * @xmath904 , given by @xmath905 , @xmath906 . in this case , there is at least a matrix of jordan type .
the diagonal matrices act projectivizing the set @xmath907 and the stabilizer is isomorphic to @xmath639
. so @xmath908 | we compute the e - polynomials of the moduli spaces of representations of the fundamental group of a complex curve of genus @xmath0 into @xmath1 , and also of the moduli space of twisted representations .
the case of genus @xmath2 has already been done in @xcite .
we follow the geometric technique introduced in @xcite , based on stratifying the space of representations , and on the analysis of the behaviour of the e - polynomial under fibrations . |
studies of the spatial distribution , metallicities , and kinematics of the globular clusters ( gcs ) in the milky way have provided us with fundamental insights into the formation and evolutionary history of our own galaxy ( e.g. , searle & zinn 1978 , zinn 1985 , armandroff & zinn 1988 , zinn 1993 , mackey & van den bergh 2005 ) .
this link exists more generally as well : studies of gcs in other giant galaxies can provide us with clues to the host galaxies origins and structure .
these luminous , old star clusters seem to mark the major star formation and merger episodes in galaxies ( e.g. , schweizer 1987 , whitmore et al .
1993 , whitmore & schweizer 1995 , brodie & strader 2006 and references therein ) as well as acting as kinematic and dynamical tracers of the outer regions of galaxies , which are difficult to probe via integrated light ( e.g. , zepf et al .
2000 , ct et al .
2001 , richtler et al .
2004 , bridges et al .
2007 , strader et al . 2011 ) .
steady progress has been made in our understanding of extragalactic gc systems in the @xmath920 years since harris 1991 comprehensive annual review of gc systems in galaxies beyond the local group .
many of the major advances in extragalactic gc system research were made possible by the hubble space telescope ( hst ) : for example , hst has enabled systematic surveys of the gc systems of hundreds of galaxies in the ( relatively nearby ) virgo and fornax clusters ( ct et al .
2004 , jordn et al . 2007 ) as well as detection of gcs in galaxy clusters @xmath7100 mpc away ( e.g. , west et al . 2011 ) .
nevertheless one of the deficiencies pointed out in both the book on globular cluster systems by ashman & zepf ( 1998 ) and the annual review by brodie & strader ( 2006 ) is the relative rarity of wide - field ccd studies that are able to measure accurate global properties of giant galaxy gc populations .
reliable measurements of quantities like the spatial distribution of the gc system , the total number of gcs , the gc specific frequency , and the global color distribution require wide - field ccd imaging in multiple filters .
a ccd camera with a field - of - view ( fov ) many arc minutes on a side is needed to cover the full radial extent of the gc systems of galaxies within @xmath920@xmath1030 mpc and accurate multi - color photometry is crucial for minimizing contamination from non - gcs ( e.g. , rhode & zepf 2001 , dirsch et al .
our group is engaged in an ongoing survey aimed at quantifying the global properties of the gc populations of giant galaxies .
we use multi - color imaging with mosaic ccd cameras and target giant spiral , s0 , and elliptical galaxies at distances between @xmath910@xmath1025 mpc .
to date we have observed @xmath730 giant galaxies with a range of masses , morphological types , and luminosities ; results for the galaxies we have finished analyzing are published in a series of papers ( rhode & zepf 2001 , 2003 , 2004 ; rhode et al .
2005 , 2007 , 2010 , hargis et al . 2011
; hereafter rz01 , rz03 , rz04 , r05 , r07 , r10 , h11 ) .
other groups who use multi - color wide - field imaging techniques have for the most part targeted massive elliptical galaxies with very populous , extended gc systems ( e.g. , dirsch et al . 2003 , imaging of ngc 1399 ; tamura et al .
2006 , imaging of m87 ; dirsch et al .
2005 , imaging of ngc 4636 ) .
we have recently observed a number of moderate - luminosity spiral and s0 galaxies in order to fill in specific gaps in our survey and in the general census of giant galaxies that have been imaged with wide - field ccd cameras .
furthermore , these galaxies are a good match , in terms of their observational requirements , for the fov and sensitivity of the wiyn 3.5-m telescope and minimosaic imager . in this paper we present results from wiyn minimosaic imaging of the gc populations of four moderate - luminosity spiral and lenticular galaxies : the sb0 galaxy ngc 1023 , the sbb galaxy ngc 1055 , and the isolated s0/sbc galaxy pair ngc 7332 and ngc 7339 . basic properties of the four galaxies morphological type , distance , absolute magnitude , stellar mass , and environment are listed in table [ table : galstats ] .
the gc systems of two of the galaxies , ngc 1023 and ngc 7332 , were studied by @xcite and @xcite , respectively , so we have combined their previous results with our wide - field data to produce the full gc system radial distributions for those two galaxies .
also included in our wiyn imaging of ngc 1023 is the magellanic dwarf galaxy ngc 1023a , so we identify a subsample of gc candidates that may be associated with ngc 1023a and estimate this galaxy s total gc population and specific frequency .
the paper is laid out as follows .
the next two sections describe the data acquisition and image reductions ( section [ section : reductions ] ) and the detection and selection of gc candidates around each galaxy ( section [ section : detection ] ) . section [ section : analysis2 ] addresses the issues of completeness and contamination in the gc candidate samples .
the results of the analysis including the radial profiles of each galaxy s gc system , the total number and specific frequency of gcs , and analysis of the gc color distributions are presented in section [ section : results ] .
the last section of the paper gives a brief summary of the main conclusions .
images of the target galaxies were obtained in september 2008 and september 2009 at kitt peak national observatory with the 3.5 m wiyn telescope and minimosaic camera .
the detector is comprised of two 2048 @xmath11 4096 ccds with 0.141@xmath12 pixels , and a fov of 9.6@xmath13 @xmath11 9.6@xmath13 when paired with the wiyn telescope .
the relatively large fov allowed us to image the isolated galaxy pair ngc 7332 and ngc 7339 simultaneously . to obtain sufficient radial coverage of these galaxies gc systems , ngc
7339 was positioned near the top center of the field , while ngc 7332 was placed in the lower right .
ngc 1023 and 1055 were both positioned in the center of one ccd ( away from a narrow gap between the two minimosaic ccds ) .
figures [ fig : n1023 finder]-[fig : n7332 finder ] show the wiyn fov for each target galaxy overlaid on a digitized sky survey ( dss ) image . to help us differentiate gcs from contaminating foreground stars and background galaxies ,
multiple exposures were taken in each of three broadband filters ( @xmath0 ) . for a list of exposure times see table [ table : observations ] .
observations of photometric standard star fields @xcite were made in photometric conditions throughout the first night of the 2008 observing run and the third night of the 2009 run , along with at least one exposure of each target galaxy in all three filters .
these observations were used to determine the color terms and zero point constants for the data taken on the photometric nights .
the calculated errors on the zero point constants were all less than or equal to 0.007 magnitude , which demonstrates that those nights were genuinely photometric .
standard reductions ( overscan and bias level subtraction , flat - field division ) were performed on all of the minimosaic images using tasks in the iraf package mscred .
the mscred tasks _ msczero _ , _ msccmatch _ , and _ mscimage _ were then used to convert the multi - extension fits images into single - extension images .
the images taken of a given target galaxy were aligned to each other .
a constant sky background level was measured for each image and subtracted .
images of a given galaxy taken in the same filter were then scaled to a common flux level and combined using the iraf task _ imcombine _ and the _ ccdclip _ pixel rejection algorithm .
the constant sky background level was then added back to the final combined image .
the result of this processing was a single deep , stacked image of each target galaxy in each filter .
the mean full - width at half - maximum of the point - spread function ( fwhm psf ) in the combined @xmath0 images ranged from @xmath14 for ngc 1023 , @xmath15 for ngc 7332 & ngc 7339 , and @xmath16 for ngc 1055 .
the method we used for calibrating the photometry in the final stacked images ( using the data taken on the photometric nights of the 2008 and 2009 runs ) is explained in section [ section : photometry ] .
to facilitate detection of gc candidates close to each target galaxy , we executed a series of steps to remove the diffuse galaxy light from the deep , stacked images .
we began by smoothing each of the images with a circular median filter with a diameter [email protected] times the mean fwhm psf of the image .
the size of the median filter was chosen after testing to find the optimal size for removing diffuse light from the galaxy while preserving the psf of the compact , star - like objects around the galaxy .
this smoothed image was then subtracted from the original version .
a constant background level was added to the resultant images to preserve the photometric integrity of later measurements .
the rest of the analysis steps were performed on these `` galaxy - subtracted '' images ( one such image for each filter and each galaxy field ) .
some regions of the galaxy - subtracted images such as small areas around saturated stars and high - background areas in the galaxy disks were masked out prior to the detection step in order to minimize spurious detections .
the iraf task _ daofind _ was then used to identify sources above a given signal - to - noise threshold in each image .
the source lists for the @xmath17 , @xmath18 , and @xmath19 filters were matched , and only sources that appeared in all three filters were retained in the final list of detected objects for a particular galaxy field . for ngc 7332 & ngc 7339
a total of 1126 objects appeared in all three filters , while ngc 1055 had 343 matched objects and ngc 1023 had 2141 matched objects .
the relatively small numbers of objects found in the ngc 1055 field compared to the other two fields had two causes .
first , ngc 1055 has a higher galactic latitude than the other galaxy fields its latitude is @xmath20 @xmath2 @xmath1051 degrees , compared to @xmath20 @xmath2 @xmath1019 degrees for ngc 1023 and @xmath20 @xmath2 @xmath1029 degrees for ngc 7332 so many fewer galactic foreground stars appear in the ngc 1055 field .
secondly , ngc 1055 is an sb spiral galaxy , whereas ngc 1023 and ngc 7332 are s0 galaxies ; we are finding in our gc system survey that spiral galaxies typically have many fewer gc candidates around them than early - type galaxies .
given the distances to the target galaxies ( @xmath910@xmath1020 mpc ; see table [ table : galstats ] ) , gcs should appear as unresolved point sources in our ground - based images .
the median half - light radius for milky way gcs is 3 pc @xcite , which subtends 0.06@xmath12 at 10 mpc , and our images have fwhm psf values of [email protected]@xmath12 .
thus any extended objects are not real gcs and need to be removed from the source lists . to accomplish this
, we plotted the fwhm of the detected objects against instrumental magnitude ( separately for each galaxy field and for each filter ) .
in such a diagram , point sources produce a tight grouping that flares at faint magnitudes .
sources that deviated markedly from this sequence in any of the three filters were eliminated from the source list .
our extended source cut for ngc 1023 is illustrated in figure [ fig:1023magvfwhm ] ; a similar cut was applied to all three target fields .
once the extended objects were removed , the source list for ngc 1023 contained 917 objects .
the source list for ngc 1055 included 257 objects , and the list for ngc 7332/ngc 7339 had 642 objects .
calibrated @xmath17 , @xmath18 , and @xmath19 magnitudes for the matched list of point sources in each galaxy field were derived using the following procedure .
aperture photometry was performed on the point sources in each field ( i.e. , the 917 sources in the ngc 1023 field , the 257 objects in the ngc 1055 field , and the 642 objects in the ngc 7332/ngc 7339 field ) using the iraf task phot .
the radii of the apertures used to do the photometry were set equal to the mean fwhm value of the image .
an aperture correction was derived for each image by calculating the mean of the difference between the total magnitude and the magnitude measured within a one - fwhm - radius aperture for a set of @xmath910@xmath1020 bright point sources .
in addition to the aperture correction , a bootstrap correction was determined for calibrating our stacked images , which are a combination of images taken on both photometric and non - photometric nights . to calculate the bootstrap correction ,
instrumental magnitudes of bright objects were compared between single - frame images taken on photometric nights and the final stacked images of a given target field .
the mean magnitude difference between the single photometric image and the final , combined image was calculated for each case .
we also determined foreground galactic extinction corrections ( in the direction of each target galaxy field ) from the maps of infrared dust emission presented in @xcite .
total instrumental magnitudes were calculated for the matched point sources by applying the appropriate aperture , bootstrap , and airmass corrections to the instrumental magnitudes measured with the one - fwhm - radius aperture .
these magnitudes were then calibrated using the appropriate photometric coefficients and galactic extinction corrections , to produce final @xmath21 , @xmath22 , and @xmath23 magnitudes and colors for the point sources . in the final gc candidate selection step , we establish an expected range of @xmath0 magnitudes and colors for gcs at the target galaxies distances and keep only the objects that are consistent with those values .
the gc luminosity function ( gclf ) of the milky way and other giant galaxies indicates that the most luminous gcs will typically have @xmath24 @xmath9 @xmath1011 @xcite . for ngc 1055
, we eliminated point sources with @xmath24 brighter than @xmath1011 ( assuming the distance modulus given in table [ table : galstats ] ) .
for ngc 7332 and ngc 7339 , there were three point sources that were close to the host galaxies , looked like gcs , and had @xmath24 values a few tenths of a magnitude brighter than @xmath1011 .
we relaxed the bright - end magnitude criterion slightly in order to include only these three specific objects . in the case of ngc 1023 , we adjusted our bright - end cut to include a luminous , spectroscopically - confirmed gc originally identified by @xcite .
relaxing the bright - magnitude criterion yielded an extra seven objects around ngc 1023 that have gc - like colors and @xmath24 magnitudes brighter than @xmath1011 .
the sequence numbers , positions , @xmath18 magnitudes , and @xmath25 and @xmath26 colors and corresponding errors for these seven objects are given in table [ table : n1023 bright objects ] .
given our assumed distance modulus for ngc 1023 ( @[email protected] ) , the @xmath24 values for these seven objects range from @xmath1011.4 to @xmath1011.0 .
( we note that two of the objects , 119 and 133 , are located several arc minutes away from the galaxy center and thus would seem less likely than the objects at small projected radial distances to be bona fide gcs . ) for the ngc 7332/7339 field we also implemented a faint source cut , eliminating objects with @xmath28 .
we did this to remove a number of faint objects that had large magnitude errors ( @xmath290.15 ) and were located many arc minutes from the target galaxy ; these are likely to be unresolved background galaxies . for similar reasons a faint cut of @xmath30
was implemented for both the ngc 1023 and ngc 1055 fields .
( note that any gcs we miss due to the magnitude incompleteness of our images and the cuts imposed on the samples at the faint end are accounted for in the gclf correction step described in section [ section : gclf ] . )
we also implemented the standard @xmath0 color selection that we use for the wide - field survey ( see @xcite for more details ) : we removed point sources that lie more than @xmath31 away from the mean @xmath26 vs. @xmath25 relation for galactic gcs with metallicities between [ fe / h ] of @xmath32 and @xmath33 .
when we test whether each source is consistent within @xmath31 of the milky way gc color - color relation , we take into account the object s photometric error .
for ngc 7332 we added a single object that fell just barely outside the color criteria and was located close ( @xmath34@xmath13 ) to the galaxy center .
gc candidates positioned very close to the host galaxy disk may be subject to additional reddening and thus might appear in the @xmath0 color - color diagram just redward of the gc candidate selection box .
point sources that appeared in this part of the color - color diagram were individually examined in the images .
for all four of the target galaxies , none of the objects just redward of the color selection box were positioned close to the host galaxy disk in the images , so we did not add any such candidates to our final list .
an additional complicating factor in the detection of gc candidates around ngc 1023 was the presence of the dwarf galaxy ngc 1023a .
the galaxy is classified as a magellanic irregular in rc3 @xcite and is located @xmath92.5@xmath13 ( a projected distance of @xmath98 kpc ) to the southeast of ngc 1023 in the image .
according to the nasa extragalactic database ( ned ) , ngc 1023a is a member of the ngc 1023 group .
we determined that the light from ngc 1023a extends over a circular region with a radius of 0.6@xmath13 in the @xmath18-band image .
we did this by plotting the pixel intensities along the radial direction between ngc 1023a and ngc 1023 to find the radial distance at which the light from ngc 1023a begins to dominate .
we measured this radial distance in all three of the combined images and averaged the results to come up with the radial size of ngc 1023a in the images .
we excluded this region from the rest of our analysis of ngc 1023 s gc system .
eight of the point sources with magnitudes and colors like gcs are located within that circular region .
we note that , without additional information , we can not definitively say with which galaxy these objects are associated .
we explore the properties of these eight objects in further depth in section [ section : dwarf ] , including estimating how many of them are likely to be gcs in the dwarf galaxy ( rather than part of ngc 1023 s gc system ) .
we correct for the area that was masked around ngc 1023a when we construct the radial profile for ngc 1023 s gc system ( see section [ section : radial profile ] ) .
the final gc candidate list for ngc 1023 contained 192 objects , while ngc 1055 had 60 gc candidates and ngc 7732 and ngc 7339 together had 105 gc candidates .
figures [ fig:1023 color - color ] [ fig:7332 color - color ] show the @xmath0 color - color diagrams for the point sources detected around the four galaxies .
open squares represent the detected point sources , and the filled circles show the objects that were selected for the final gc candidate list .
the gc color selection box and the mean @xmath26 vs. @xmath25 relation for milky way gcs ( the line marking the midpoint of the box ) are marked with solid lines .
the tracks indicate the expected colors for galaxies of different morphological types at redshifts between @xmath35 and @xmath36 , to illustrate possible sources of background galaxy contamination ; i.e. , to show what types of galaxies , and at what redshifts , are likely to lie in the same part of the @xmath0 color - color plane as gcs .
( see rz01 for a full discussion of how the color cut for the survey was developed and details about how these galaxy tracks were produced . )
in order to determine the limit of our ability to detect point sources of a given magnitude in the galaxy - subtracted images , we created a series of tests to quantify the completeness level as a function of object magnitude in each filter for each galaxy field .
we first measured the typical psf for each image .
we added 200 artificial point sources with the appropriate psf and a given magnitude to each of the combined images ( @xmath17 , @xmath18 , and @xmath19 ) .
we did this in 0.2-magnitude steps over a range of 6@xmath107 magnitudes , creating a separate image for each magnitude step ( and each filter ) .
the images with the artificial point sources were then processed in a series of steps identical to the detection steps we had executed to find gc candidates in the original images . for each magnitude step and each filter , we determined what fraction of the artificial sources were recovered .
this gave us an estimate of the completeness at that magnitude , i.e. , our ability to detect point sources of a given magnitude in a given image .
the result was a series of completeness curves , one for each filter , for each set of galaxy - subtracted combined images .
the completeness information was taken into account in our analysis of what fraction of the gc luminosity function we had covered for each galaxy ( see section [ section : gclf ] ) .
the 50% completeness limits are typically @xmath37-@xmath38 magnitudes ; table [ table : completeness ] lists the exact values for each image .
contamination of the gc candidate lists by foreground stars and background compact galaxies occurs because some of these objects may match all of the selection criteria for gc candidates . here
we present our methods for quantifying the contamination level and , where possible , correcting for the presence of contaminating objects .
we used a galactic stellar population synthesis model to estimate the level of stellar foreground contamination that might be present in the gc candidate lists .
hereafter we refer to the model as the besanon model ; a complete description of the model and the observational constraints used to develop it are given in @xcite . briefly
, the model divides the milky way galaxy into four distinct stellar populations : a thin disk , a thick disk , a spheroid , and a bulge .
each component is defined by a star formation rate , initial mass function , metallicity distribution , kinematics , and evolutionary tracks within the model . given a range of stellar types , magnitudes , and colors , the model uses monte carlo methods to produce galactic star counts for a specific region of the sky .
comparisons of the star count predictions of the besanon model to photometric surveys have shown reasonable agreement ( e.g. , @xcite and @xcite ) .
we used the besanon model to estimate the number of galactic foreground stars that would appear within the fields of view of our minimosaic pointings . in order to arrive at an estimate of the number of possible contaminants in the gc candidate samples , we set the inputs of the model to provide counts only for galactic stars with @xmath25 colors , @xmath26 colors , and @xmath18 magnitudes defined by the range of properties in the final selected gc candidate list for each galaxy . to account for the statistical nature of the model
, we ran it ten times per target galaxy field and then calculated the mean galactic star counts in each case .
we then used this number and the fov of the minimosaic images to compute the surface density of contaminating galactic stars for the galaxy field .
the contamination was estimated to be 0.15 arcmin@xmath39 for ngc 1055 .
ngc 1023 has an estimated foreground contamination of 0.56 arcmin@xmath39 , while the field for ngc 7332 and 7339 has an estimated foreground contamination of 0.54 arcmin@xmath39 . as noted in the discussion of source detection ( section [ section : detection ] ) , the relatively low level of foreground stellar contamination for ngc 1055 compared to the other target galaxies
is expected because of its relatively high galactic latitude .
another way to evaluate the level of contamination in the wiyn data is by examining archival hst images .
many compact background galaxies that appear as point sources in ground - based imaging may be resolved by hst , so determining whether some of the wiyn gc candidates are actually extended objects can help us assess how much background galaxy contamination is present in the wiyn gc candidate lists .
accordingly , we analyzed hst archival data for ngc 7332 & ngc 7339 to determine if any of the wiyn gc candidates were background galaxies .
we found one wide field planetary camera 2 ( wfpc2 ) image in the hubble legacy archive ( hla ) that was taken in an appropriate broadband filter ( f606w ) and had sufficient depth ( a combination of three exposures totaling 1500 seconds ) for detecting our gc candidates .
the wfpc2 pointing ( hst proposal i d 7566 , pi : green ) was @xmath92.5@xmath13 from the center of ngc 7339 .
the footprint of the wfpc2 pointing is marked with a solid line in figure [ fig : n7332 finder ] .
five of the wiyn gc candidates lie within the wfpc2 fov .
aperture photometry was performed on these sources in the hst images , using apertures with radii of 0.5 and 3.0 pixels .
the ratio of counts in the larger aperture to the smaller aperture is expected to be @xmath40 12 for compact objects on the pc chip and @xmath40 8 for compact objects on the wf chip @xcite .
all five of the wiyn gc candidates had count ratios in the wfpc2 images consistent with their being point sources .
it is important to note here that this is not necessarily indicative of zero background contamination from background galaxies ( since only five of the wiyn gc candidates appear in the wfpc2 pointing ) , but it may indicate at least that the wiyn candidate sample is unlikely to be dominated by background objects . for ngc 1023 , there were four archival images of sufficient depth ( each a combination of two exposures totaling 2400 seconds ) and taken in useful filters ( f555w and f814w ) in the hla that we could use to assess our background galaxy contamination level .
the hst images included two slightly overlapping wfpc2 pointings near the galaxy center from hst proposal i d 6554 ( pi : brodie ) .
the wfpc2 footprints are marked in figure [ fig : n1023 finder ] .
twenty of our wiyn gc candidates were identified in both filters , and of those , two had count ratios indicating they were extended objects .
this again seems to suggest that the wiyn gc candidate samples are not dominated by contamination from background galaxies .
we did not remove these objects from our candidate list at this stage , since the presence of contaminants was dealt with by the application of the overall contamination correction ( section [ section : asymptotic ] ) . manually removing only a select
few contaminating objects at this point in the analysis would effectively result in a double correction and an incorrect result for the derived gc system radial profile .
these two objects were , however , excluded during the color distribution ( section [ section : tblue ] ) analysis step , where the measured properties of individual objects plays a role .
there are currently no images of ngc 1055 in the hst archive .
thus we could not do this type of analysis for the ngc 1055 field and gc candidate list .
since the gc population is expected to fall off with increasing distance from the center of the galaxy , we assume that beyond some point the gc candidate sample will be dominated by contaminating objects ( compact objects with magnitudes and colors like gcs , which are actually stars or background galaxies ) . at large projected radii
the radial distribution of gc candidates should flatten to some constant value ; the surface density in the outermost portion of the radial profile thus provides us with an estimate of the surface density of contaminants .
we refer to this estimate as the `` asymptotic contamination correction '' .
to investigate the radial distribution of each galaxy s gc system as well as to determine the asymptotic contamination correction , we assigned the gc candidates to one of a series of circular annuli extending outward from the galaxy center .
the annuli sizes we used to create this initial radial profile for each galaxy varied between 0.5 wide to 0.8 wide , depending on how many gc candidates appeared around the galaxy and how the candidates were distributed . for each annulus
, we calculated a corresponding `` effective area '' , equal to the area of the annulus minus any area in which gc candidates could not be detected ( e.g. , masked or saturated regions , or parts of the annuli that extended off the edges of the images ) .
we then calculated the surface density of gc candidates ( number per effective area ) in each annulus , and plotted this quantity versus the mean projected radius of the corresponding annuli .
for the gc system of ngc 1055 , we found that the surface density of gc candidates converged to a fairly constant value and remained flat in the annuli with mean radii greater than 4.7@xmath13 . taking the weighted mean of the surface density in the annuli beyond this point
, we find a contamination value of @xmath41 objects arcmin@xmath39 .
for ngc 1023 , the profile converged to a relatively constant level at radii beyond 6.7@xmath13 ; the weighted mean surface density in these outer annuli is @xmath42 objects arcmin@xmath39 .
ngc 7332 & ngc 7339 appear on the sky near to each other ( @xmath40 6@xmath13 separation ) , although they are not thought to be interacting @xcite .
the proximity of these galaxies in the minimosaic pointing complicated our efforts to determine an asymptotic contamination correction . both galaxies showed a gc surface density that decreased with increasing radius in the first few annuli of the profile , but then increased again due to the presence of the neighboring galaxy s gc system ( i.e. , the gc surface density around ngc 7332 dropped with radius in the first few arc minutes around the galaxy , but then increased in the region around ngc 7339 , and vice versa ) . to remedy this , we noted the approximate radius at which the gc population around each galaxy fell to a constant level , and then placed a circular mask with that radius over the galaxy and its gc system . for both galaxies ,
a circular region with radius 1.8@xmath13 ( @xmath912 kpc ) was sufficient to mask out the gc population .
we used the image with both masks in place to determine the asymptotic contamination correction for the entire field .
because ngc 7332 and ngc 7339 appear in the same wiyn pointing and we used a single set of gc candidate selection techniques for the @xmath0 images of that field , the surface density of contaminants that we apply should be the same for the two galaxies .
combining their asymptotic contamination corrections yields a weighted average contamination value of @xmath43 objects arcmin@xmath39 .
the last step in this process was to take the appropriate asymptotic contamination correction for each galaxy and multiply it by the effective area of each annulus in the radial profiles .
this provides us with an estimate of the number of contaminating objects in each radial bin of each galaxies gc system profile .
the _ fraction _ of contaminating objects in each annulus is the number of contaminating objects divided by the number of gc candidates .
this provides us with a radially - dependent contamination correction that we can use in subsequent analysis steps .
we note that for all of the galaxies , the contamination level determined from the asymptotic behavior of the radial profile is larger than the level calculated from the galactic star count models .
this is a useful consistency check ; we expect that the contamination in our gc candidate samples will come from both compact background galaxies and foreground galactic stars , so the total number density of contaminants should be larger than the number density from the galactic stellar population models ( in the limit that the model results are accurate ) .
we used the list of gc candidates for each galaxy to create a luminosity function ( lf ) a histogram of the number of gc candidates in a series of @xmath18 magnitude intervals for that galaxy s gc system . the radially - dependent contamination correction described in section [ section : asymptotic ]
provides us with an estimated contamination fraction for each annulus in the radial profile , so we used it to correct the lf data .
if a gc candidate was located , for example , in an annulus that had a contamination fraction of 15% , then 0.85 was added to the number of objects in the appropriate @xmath18 magnitude bin of the lf histogram .
we then divided each magnitude bin by its completeness correction ( section [ section : completeness ] ) , yielding a final corrected gclf . because we require that the gc candidates must be detected in the @xmath17 , @xmath18 , and @xmath19 images , the completeness correction we apply takes into account the detection limits in all three filters ( specifically , by convolving the completeness level in each filter at each @xmath18 magnitude bin ; see rz01 for additional details ) .
we can not reliably determine the exact shape of the gclf for these galaxies based on our data , because either we have insufficient numbers of gc candidates in the samples or we do not follow the luminosity function far enough past the peak of the gclf , or both . therefore to fit the gclf and determine how many gcs we may have missed , we assume that the intrinsic gc luminosity function is a gaussian function similar to that of the milky way gc system , which has a peak magnitude at @xmath44 and a dispersion of 1.23 @xcite .
given the distances in table [ table : galstats ] , we calculated the peak apparent magnitude for the gclf of our galaxies , and fitted gaussians with that peak and dispersions of 1.2 , 1.3 , and 1.4 magnitude to the lf data .
we allowed the amplitude of this theoretical gclf to vary and calculated the fraction of the total area of the best - fitting theoretical gclf that was covered by our observed lf data .
the size and starting position of the lf magnitude bins were also varied to investigate the effect on the calculated gclf coverage ; the resulting differences were included in our error calculations in section [ section : spec freq ] . the observed gclf data and best - fitting gaussian functions four the four target galaxies are plotted in figure [ fig : gclf ] .
the shaded histogram is the observed , contamination - corrected luminosity function of the gc candidates and the solid - line historgram is those same data corrected for the final convolved @xmath0 completeness .
the best - fitting gaussian functions for three different dispersions are shown as indicated in the figure legend .
table [ table : gclf ] lists the mean calculated gclf coverage for each galaxy . for ngc 1055 , ngc 7332 , and ngc 7339 , we detect about one - third of the gc population given the magnitude depth and detection limits of our data ; for ngc 1023 , which has a distance that is significantly closer than the other three galaxies , we cover about 60% of the theoretical gclf .
as mentioned in the explanation of the asymptotic contamination correction ( section [ section : asymptotic ] ) , a radial profile for the gc system of each galaxy was constructed by first assigning each gc candidate to one of a series of concentric circular annuli starting from the center of the host galaxy and continuing out to the edges of the images .
after some experimentation with various radial profile bin sizes , we settled on annuli corresponding to approximately 3 kpc in physical size at the distance of each galaxy .
accordingly , the circular annuli used to create the radial profile for ngc 1023 were 0.8@xmath13 wide , the ngc 1055 annuli were 0.8@xmath13 wide , and for ngc 7332 0.5@xmath13-wide annuli were used . for ngc 7339
the gc system is extremely compact around the galaxy , extending to only @xmath91@xmath13 from the galaxy center .
this necessitated the use of a more granular radial bin size ( 0.3@xmath13 ) to obtain more measurements at the cost of an increase in the relative error .
each annulus was corrected for missing areal coverage due to parts of the annuli being masked or extending off the edges of the images .
the surface density of contaminants , as determined from the asymptotic behavior of the radial profile ( section [ section : asymptotic ] ) , was multiplied by the effective area of each concentric annulus to yield the number of contaminants for that annulus . after correcting the total number of gc candidates in each annulus for contamination
, we applied the appropriate gclf correction , then divided by the effective area of the annulus to yield the surface density of gcs in each annulus .
the error on the surface density includes poisson errors on the number of gcs and the number of contaminating objects .
we calculated the mean distance of all unmasked pixels in each annulus , and used this as the radial distance of each annulus in the radial profile .
the final , corrected radial profile data for the gc systems of the four galaxies are listed in tables [ table:1023radprof ] through [ table:7339radprof ] . for each projected radial distance ,
a surface density and error are listed . in the case of ngc 1023 and ngc 7332 , we supplemented our wiyn data with data from previously - published studies of the gc populations to generate additional points in the gc system radial distributions
. therefore the radial profile tables for these galaxies also list the source of the measurement in the last column .
( more details on the supplementary data are given in the next section . )
the final radial profiles are plotted in figures [ fig:1023 radial profile ] through [ fig:7339 radial profile ] .
the top plot in each figure shows the gc surface density versus projected radius , while the bottom plot shows the log of the gc surface density versus the log of the projected radius .
we fitted both a de vaucouleurs law ( of the form @xmath45 ) and a power law ( of the form @xmath46 ) to each radial profile .
here , @xmath47 is the surface density of gcs in number per square arc minute and @xmath48 is the projected radius in arc minutes . the slope , intercept , and reduced @xmath49 value corresponding to the best - fitting power law and de vaucouleurs law functions are given in table [ table : fitting ] .
both functions are shown in the bottom plots of the radial profile figures but the function with the lowest reduced @xmath49 value was the one used in subsequent analysis steps ( see section [ section : totaln ] ) .
the power law is the best - fitting function for the radial profiles of the gc systems of ngc 1055 and ngc 7339 .
this is also the case for ngc 7332 , although the data points in the profile ( fig.[fig:7332 radial profile ] ) fall below the power law at large radius .
ngc 1023 is best fit by a de vaucouleurs profile . in this case , the points in the inner radial profile fluctuate between low and high values ; the reasons for this are explored in the next section .
the radial distribution of the gc system in the final corrected profile falls to zero surface density ( and remains at zero in adjacent bins ) within the calculated errors for each of the target galaxies .
we define the center of this radial bin as the radial extent of the gc system for that galaxy . for ngc 1023 , this occurs in the radial bin centered at 6.3@xmath13 and for ngc 1055 at 5.5@xmath13 . the gc system radial profiles of ngc 7332 and ngc 7339 drop to zero surface density within the errors in the 2.0@xmath13 and 1.5@xmath13 annuli , respectively .
it is worthwhile to transform the radial extent of each galaxies gc system into physical units of distance .
we calculate this value by combining the radial extents determined above with the distance moduli listed in table [ table : galstats ] .
the errors on the extent values are calculated for each case by combining the uncertainty in the galaxy distance modulus with an error equal to one bin width in the gc system radial profile . for ngc 1023
we find that the gc system extends out to @xmath50 kpc and for ngc 1055 out to @xmath51 kpc .
ngc 7332 and ngc 7339 possess less extended gc systems , at @xmath52 kpc and @xmath53 kpc , respectively .
the measured effective radii ( @xmath54 ) of the light distributions of our target galaxies have been tabulated by the atlas3d survey ( cappellari et al .
2011 ) . in their data tables
, ngc 1023 is listed as having @xmath55 @xmath2 47.9 , ngc 1055 has @xmath54 @xmath2 67.6 , ngc 7332 has @xmath54 @xmath2 17.4 , and ngc 7339 has @xmath55 @xmath2 22.4 .
we find , therefore , that the gc systems of ngc 1023 , ngc 1055 , ngc 7332 , and ngc 7339 extend to approximately 7.8 , 4.9 , 6.8 , and 3.9 @xmath54 , respectively .
our measurements of the gc system radial extent for these four galaxies fall along the established relationship between the extent of a gc system and the stellar mass of the host galaxy shown in previous papers from our wide - field survey ( r07 , r10 ) . fitting a second - order polynomial to the data including the four new points from this
paper yields a function of the form @xmath56 where @xmath57 is log(@xmath58 ) and @xmath59 is the radial extent of the gc system in kiloparsecs . the data points and best - fitting function are plotted in figure [ fig : extent mass ] . for two of the target galaxies
, we were able to supplement and directly compare our study with results from previous work . in the case of ngc 1023 ,
s. larsen and j. brodie provided us with @xmath18 and @xmath60 photometry , positions , and size measurements for 1058 point sources from their hst wfpc2 study of this galaxy s gc system @xcite . to select gc candidates from the set of point sources , we started by applying the same color and magnitude cuts as larsen and brodie used in their analysis of ngc 1023 s gc population : we selected objects with @xmath61 and @xmath62 , duplicating their sample of 221 gc candidates .
we decided to implement a more stringent magnitude cut at the faint end in order to simplify the rest of the analysis .
we removed objects with @xmath63 ; at magnitudes brighter than this threshold , the wfpc2 data are @xmath790@xmath10100% complete and a completeness correction is not required .
the final list we used included 151 objects and we assume no magnitude incompleteness in this list .
larsen and brodie assessed the contamination level in their wfpc2 data by examining a nearby comparison field pointing .
they found five point sources in the comparison field with @xmath18 @xmath40 24 and @xmath64 colors like gcs . based on this value and the size of the wfpc2 fov
, we assume the surface density of contaminants in the wfpc2 gc candidate list is 0.9 @xmath65 0.4 objects arcmin@xmath39 . using the same methods as in section [ section : gclf ] we fitted a gclf to the 151-object wfpc2 sample and found the gclf coverage to be @xmath66 . after assigning the hst candidates to annular bins and correcting for missing area , gclf coverage , and contamination , we added these values to the wiyn radial profile data to produce the profile shown in figure [ fig:1023 radial profile ] and listed in table [ table:1023radprof ] .
the advantage of combining hst and wiyn data is apparent in the plot and table : the hst data allow us to probe deeper into the central regions of the host galaxy , while the wiyn data enable us to observe the full radial extent of the gc system .
the innermost three points in the wiyn radial profile ( between @xmath90.75@xmath13 and 2.3@xmath13 ) overlap the radial range of the hst data , which extends from @xmath90.4@xmath13 to 2.5@xmath13 .
the wiyn gc surface density values at 1.5@xmath13 and 2.3@xmath13 closely match the hst gc surface density values in that radial region , but the innermost wiyn point at 0.76@xmath13 is low ( [email protected] arcmin@xmath39 ) compared to the nearby hst point ( [email protected] arcmin@xmath39 ) at a projected radius of 0.83@xmath13 .
we thoroughly investigated the reasons for this discrepancy .
a total of 88 hst - identified gc candidates had sky positions that put them in the same region as our innermost wiyn radial profile bin .
we marked these positions on our wiyn images , then individually examined them , looking for matches in the source lists .
we found that 26 hst - identified gc candidates within the magnitude detection limits for the wiyn data were located in masked regions in the wiyn images .
an additional seven candidates were in crowded regions , where sources in the wiyn images were sometimes blended together ; this apparently affected the stellar profiles and the measured colors enough that these objects were not chosen as gc candidates .
based on the wiyn surface density of gc candidates in the unmasked area , we expected only @xmath95 objects in the masked regions .
the unfortunate over - density of gc candidates in the masked regions in this bin explain the low calculated value for the surface density . for ngc 7332
, we were able to compare the results of our radial profile to those published in @xcite .
forbes et al .
studied ngc 7332 s gc system with @xmath17 , @xmath18 , and @xmath60 imaging from the keck - ii 10-m telescope and low resolution imaging spectrometer ( lris ) .
we have taken surface density values directly from the final gc system radial profile plotted in their paper and over - plotted them on our wiyn radial profile in figure [ fig:7332 radial profile ] .
we also list their radial profile values along with ours in table [ table:7332radprof ] . the radial profile from the forbes et al .
study agrees very well with our final wiyn radial profile points , though the errors on the forbes et al .
data are large .
a primary goal of our wide - field gc system survey is to determine an accurate value for the total number of gcs , @xmath1 , for each galaxy . to calculate this number we start by integrating the best - fitting radial profile function from the innermost bin in the profile to the outer edge of the bin in which the gc surface density vanishes within the errors . in the case of ngc 1055 the best - fitting function ( as measured by the reduced @xmath49 value ;
see table [ table : fitting ] ) was a power law that we integrated from 0.34@xmath13 to 5.9@xmath13 .
the result was 185 gcs .
we can not detect gc candidates inside 0.34@xmath13 ( 1.6 kpc at the distance of ngc 1055 ) because of saturation of the ccd by light from the galaxy bulge as well as the high noise level in that region .
we considered three scenarios for the behavior of the radial profile inside 0.34@xmath13 : ( 1 ) that the power law continues to very small radius ( @xmath48 very close to 0 ) ; ( 2 ) that the profile remains flat ( constant surface density ) from the innermost populated bin to @xmath48@xmath20 ; and ( 3 ) that the inner profile is similar to that of the milky way gc system .
we found that integrating the power law to very small radii ( close to @xmath48@xmath20 ) yielded unrealistically large numbers of gcs in this inner region ; it added 85 gcs in a region of radius only 1.6 kpc , and increased the total number of gcs in the galaxy by 46% .
thus we rejected this scenario and averaged the result of the other two methods .
the second possibility was straightforward to calculate and predicted an additional 17 gcs inside 0.34@xmath13 , to yield a total of 202 . to estimate the number of gcs for the third possibility , we examined the 2010 edition of the mcmaster catalog of milky way gcs ( harris 1996 ) and identified 19 galactic gcs within 1.6 kpc of the milky way center . if one assumes that ngc 1055 has a similar surface density of gcs within its central 1.6 kpc around the galaxy center , this yields an extra 28 gcs in the central region and a final total of 213 gcs for the system .
the average result for the two options ( flat or milky - way - like surface density inside 0.34@xmath13 ) yields 207.5 gcs .
we computed an error on @xmath1 by adding in quadrature the errors due to fractional coverage of the gclf , poisson errors on the number of gcs and contaminating objects , and uncertainty in the number of gcs in the central region of the galaxy .
our final value ( rounded to significant digits ) for the total number of gcs in ngc 1055 is @xmath1 @xmath2 @xmath4 .
we executed the analogous steps for ngc 1023 and integrated the best - fitting de vaucouleurs law function from 0.27@xmath13 to 6.67@xmath13 to derive the number of gcs ( 470 ) in that region . because we had made use of the hst data from @xcite
, we were able to start the integration at a smaller radius ( 0.27@xmath13 , or @xmath90.90 kpc at the distance of ngc 1023 ) than if we had only had the wiyn imaging .
inside 0.27@xmath13 , where we had no observations , we averaged the value obtained from continuing the de vaucouleurs integration to the center ( an additional 26 gcs ) , and the value calculated from assuming a constant surface density to @xmath48@xmath20 ( 16 gcs ) .
it seemed appropriate and valid to average the result from these two methods because the de vaucouleurs law does not become unphysical at small radii , and the two options yielded similar results in any case .
the outer integration edge was set at the point where the radial profile drops to levels consistent with zero . for ngc 1023 , we calculate a final total @xmath1 of @xmath3 .
a similar analysis was performed on the radial profiles of ngc 7332 and ngc 7339 . in both cases
a power law profile yielded the best fit to the radial profile data ; these functions were integrated from 0.2@xmath13 to 2.2@xmath13 for ngc 7332 , and 0.1@xmath13 out to 1.6@xmath13 for ngc 7339 to calculate the number of gcs in those radial ranges .
the result was 155 for ngc 7332 and 71 for ngc 7339 . to determine how many gcs were likely to be inside the innermost radial bin , we integrated the power law functions to very small radii ( @xmath48 close to 0 ) and found that they gave reasonable results this time ( for example , integrating the de vaucouleurs law fit all the way to the galaxy center gave results very similar to the power law integration for both galaxies ) . for ngc 7332
, the flat profile yielded 11 additional gcs in the central region , while the power - law profile gave a value of 27 gcs over the same region .
for ngc 7339 , the flat profile predicted three more gcs in the inner region while the power law profile predicted five more gcs .
the final estimates of @xmath1 @xmath2 @xmath5 for ngc 7332 and @xmath67 for ngc 7339 are the result of taking the mean of the values obtained when assuming a flat profile inside the innermost bin and a continuous power law into the center of the target galaxy ( very near @xmath48 @xmath2 0 ) .
a useful value for characterizing the gc system of a galaxy is the specific frequency , @xmath68 , or the total number of gcs normalized by the host galaxy luminosity .
specific frequency was first defined by @xcite as a tool for investigating whether the ability of a galaxy to form gcs depends on the galaxy s luminosity or on one or more other factors .
it is defined by the equation @xmath69 @xcite , where @xmath1 is the total number of gcs and @xmath24 is the galaxy s absolute @xmath18 magnitude . combining the final determinations of @xmath1 with the @xmath24 values listed in table [ table : galstats ] yields @xmath68 @xmath2 @xmath70 for ngc 1023 , @xmath68 @xmath2 @xmath71 for ngc 1055 , @xmath68 @xmath2 @xmath72 for ngc 7332 , and @xmath68 @xmath2 @xmath73 for ngc 7339 .
the quantity @xmath68 has a stellar population dependence i.e. , variations in the @xmath18-band stellar mass - to - light ratios for galaxies of different morphological types will contribute to differences in @xmath68 .
this in turn complicates the comparison of specific frequency values among galaxies of various types . to address this issue , @xcite formulated the @xmath74 parameter , which is the number of gcs normalized by the stellar mass of the parent galaxy and is defined by @xmath75 @xcite , where @xmath1 is again the total number of gcs and @xmath76 is the stellar mass of the galaxy .
table [ table : galstats ] lists the calculated @xmath76 for each galaxy based on the given @xmath24 and an assumed @xmath18-band mass - to - light ratio value that varies with morphological type as defined in @xcite .
the final @xmath74 values for the four galaxy targets are @xmath77 , for ngc 1023 , @xmath78 for ngc 1055 , @xmath79 for ngc 7332 , and @xmath80 for ngc 7339 .
errors for @xmath68 and @xmath74 were computed by propagating the error on @xmath1 and including the errors on the total galaxy magnitudes .
the final results for @xmath1 , @xmath68 , and @xmath74 for these four galaxies are included in the first four entries in table [ table : results ] in columns 4 , 5 , and 6 , respectively .
the other galaxies included in table [ table : results ] are introduced and discussed in section [ section : comparison ] and the quantity in column ( 7 ) , @xmath81 , is discussed in section [ section : tblue ] .
only one of the four target galaxies , ngc 7332 , has estimates of gc specific frequency already published in the literature .
we searched the literature and found no other published articles on the gc populations of ngc 1055 or ngc 7339 .
the @xcite paper on the gc system of ngc 1023 did not include estimates of the total population or radial extent of the full gc system .
their study was constrained by the small field - of - view of hst / wfpc2 and the relative proximity of ngc 1023 , which meant that their spatial coverage was very limited . in their keck - ii imaging study
, @xcite estimated that the total number of gcs in ngc 7332 is @xmath82 ; this is consistent with our final derived value of @xmath5 .
@xcite calculated a specific frequency @xmath68 @xmath2 @xmath83 , whereas we calculate @xmath68 @xmath2 @xmath72 . since our numbers for @xmath1
are consistent , this difference in @xmath68 is due to forbes etal . using a much smaller distance to the galaxy .
their assumed distance is 15.3 mpc and is derived by combining the redshift and assuming @xmath84 @xmath2 75 kms@xmath85 @xmath86 with an additional virgocentric infall correction . combining the distance derived from surface brightness fluctuations ( 23 mpc ; tonry et al .
2001 ) with their @xmath1 number yields @xmath68 @xmath2 @xmath87 , which is consistent with our result .
our gc system survey is ongoing and we add new observations of giant galaxies and their gc populations each observing season .
subsequent papers are planned that will include results for several more early - type galaxies ( hargis & rhode 2012a , in preparation ) and present a thorough multivariate analysis of how gc system properties vary with host galaxy luminosity , mass , morphology , and environment ( hargis & rhode 2012b , in preparation ) . for
now , we can examine how the four spiral and s0 galaxies analyzed here compare to the other spiral and e / s0 galaxies for which secure global gc system properties have been measured .
in table [ table : results ] we have compiled the measured global properties of the gc systems of a sample of spiral , s0 , and elliptical galaxies , including morphological type , @xmath24 , @xmath1 , @xmath68 , @xmath74 , and @xmath81 ( see section [ section : tblue ] ) . the data for thirteen of the galaxies are taken from our ongoing gc system survey ( rz01 , rz03 , rz04 , r07 , r10 , h11 , hargis & rhode 2012a , in preparation ) .
we supplement the survey data with data from other wide - field , multi - color imaging studies of galaxy gc system properties in the literature ( ngc 1052 from @xcite ; ngc 4374 from @xcite ; ngc 5128 from @xcite , with the blue gc fraction adopted from @xcite ) .
we also include values for the gc systems of the milky way @xcite and m31 ( @xcite , @xcite , @xcite).650 ( revised bologna catalog , galleti et al .
on the other hand , fan , de grijs , & zhou ( 2010 ) select a sample of objects from the revised bologna catalog and find 445 confirmed gcs in m31 . in any case , increasing @xmath1 for m31 from 450@xmath65100 to 650@xmath65150 , and thereby increasing the gc specific frequencies for that galaxy , does not change the results of the comparison presented here and in section [ section : tblue ] . ]
this gives us a total of 18 galaxies ( hereafter referred to as the @xmath88@xmath218 sample , which is made up of half spiral galaxies and half e / s0 galaxies ) with well - measured @xmath68 and @xmath74 to compare to the values measured for the four galaxies in the current paper . the weighted mean @xmath68 and @xmath74 values for the nine spiral galaxies in our @xmath88@xmath218 comparison sample are @xmath89 and @xmath90 , respectively .
the @xmath74 values for ngc 1055 and ngc 7339 , and the @xmath68 for ngc 7339 are consistent with the spirals in the @xmath88@xmath218 sample , while the @xmath68 for ngc 1055 falls slightly above , but within @xmath91 of , the weighted mean . including the results for ngc 1055 and ngc 7339 ( for a total of 11 spiral galaxies ) in the weighted mean
gives us @xmath92 and @xmath93 .
the nine e / s0 galaxies in the @xmath88@xmath218 sample have weighted mean values of @xmath94 and @xmath95 . the @xmath68 and @xmath74 for ngc 1023
are typical of the larger sample , whereas ngc 7332 falls below the weighted mean in both categories , but within the large range of measured @xmath74 values seen in the comparison galaxies . for the e / s0 galaxies in the @xmath88@xmath218 sample , the gc specific frequencies are in the range @xmath68 of @xmath96 and @xmath74 of @xmath97 . adding our results for ngc 1023 and ngc 7332 to the weighted mean ( for a total of 11 e / s0 galaxies ) results in a new weighted mean @xmath94 and @xmath98 .
as we add more galaxies to our sample , we will be able to divide the galaxies more finely according to their properties ( e.g. , specific morphological type and environment ) and look for trends of gc system properties with host galaxy properties .
for now , given the above , we can say with confidence that in terms of @xmath68 and @xmath74 , ngc 1023 and ngc 7332 possess values consistent with other early - type galaxies with carefully - measured global gc specific frequencies .
ngc 1055 and ngc 7339 likewise fall within the normal range of gc specific frequencies observed for spiral galaxies . quantifying both the gc system color distribution and the blue gc specific frequency (
the number of blue , metal - poor gcs normalized by the galaxy mass ) of the target galaxies are two of the goals of the overall gc system survey . in old stellar populations
( with ages greater than @xmath92 gyr ) , differences in broadband optical color are primarily due to differences in metallicity ( e.g. , @xcite and many references therein ) .
metal - poor gcs would therefore have bluer optical colors than metal - rich gcs .
the gc systems of many giant galaxies , including the milky way @xcite , m31 ( @xcite , @xcite ) , and many ellipticals ( @xcite , @xcite , @xcite ) have been shown to include two ( or sometimes more ) populations of gcs with different metallicities , likely formed in different episodes of star formation .
various galaxy formation models ( e.g. , @xcite , @xcite , @xcite , @xcite , @xcite ) make predictions for the relative numbers of blue and red gcs and how those numbers might vary in different types of galaxies and at different radii within galaxies
. one interesting piece of analysis that we can do , therefore , is to examine the numbers of blue , metal - poor gcs in our target galaxies and compare them to the values for other galaxies . in general ,
galaxies with higher stellar mass tend to have higher specific frequencies of blue , metal - poor gcs ( e.g. , r05 , r07 , peng et al .
2008 , spitler et al .
moreover , the high blue gc specific frequencies of the most massive giant elliptical galaxies suggest that they are unlikely to have formed from the straightforward merger of typical spiral galaxies and their gc populations ( e.g. , @xcite , r05 , @xcite ) . for each of the four target galaxies
, we set out to examine the morphology of the gc color distribution , determine the fraction of blue and red gcs in the galaxy , and calculate @xmath81 , which is defined in the same way as @xmath74 was defined in section [ section : sn_and_t ] except that @xmath99 this time includes only the blue portion of the gc population . to ensure no bias ( i.e. , toward red or blue objects ) we begin by constructing a list that includes only those gc candidates that lie within the 90% complete magnitude ranges for all three filters ( see rz01 for more details )
; we refer to this as the 90% complete sample . for ngc 1023
, the 90% complete sample contains 119 objects .
we used this sample as input to the kmm mixture modeling code ( ashman etal .
1994 ) to test for bimodality in the @xmath8 color distribution .
based on the gc candidates @xmath8 color , the kmm code found that unimodal fit was rejected at greater than 95% confidence .
the resulting bimodal distribution puts 68 out of 119 objects ( 57% ) in the blue ( lower metallicity ) peak . taking the fraction of blue gcs in the total sample from this estimate and combining it with @xmath1 for ngc 1023 and the stellar mass value from table [ table : galstats ] , we find a mass - normalized number of metal - poor gcs of @xmath100 . in typical giant galaxies with double - peaked gc color distributions ,
the dividing line between the blue- and red - dominated regions occurs at @xmath101 ( r01 , r04 ) .
we find reasonable agreement with that expectation here , as @xmath102 marks the overlap of the two populations in ngc 1023 .
a similar analysis was applied to the gc population of ngc 7332 . in this case , the 90% complete sample contained 53 objects , just above the 50-object lower limit for the kmm algorithm .
applying kmm to this sample resulted in a weak rejection of a unimodal distribution at an 81.2% confidence level .
for the bimodal case , kmm places 33 out of 53 objects ( 62% ) in the blue portion of the distribution .
the line marking the overlap between the two regions occurs at @xmath103 .
applying the 62% fraction to our full sample gives us @xmath104 . if we instead simply split the 90% sample into blue and red gc candidates at the typical dividing line for massive elliptical galaxies of @xmath8 @xmath2 1.23 ( rz01 , rz04 ) , the proportions of blue and red gcs in ngc 7332 are 51% and 49% , respectively .
the @xmath81 value then becomes [email protected] .
figure [ fig : color_dist ] illustrates the color distributions for gc candidates in ngc 1023 ( top ) and ngc 7332 ( bottom ) .
the unshaded region marks our full sample , while the shaded region contains the 90% complete sample .
the solid curve marks the bluer , more metal - poor population , while the dashed curve marks the redder , more metal - rich population . for ngc 1055 , our 90% complete sample contained only 13 objects . because this sample is so small , we were unable to use the kmm algorithm to investigate whether two populations might be present .
instead , we can look at what happens when we divide the population along the previously referenced @xmath8 @xmath9 1.23 criterion .
six of the 13 ( 46% ) objects in the 90% complete color sample for ngc 1055 are bluer than @xmath8 @xmath2 1.23 .
if we assume that the results from this small sample are representative of the entire gc system , we find that @xmath81 for ngc 1055 is [email protected] . if we assume instead that ngc 1055 has a blue gc fraction more like that of the milky way or m31 ( @xmath970% ; @xcite , @xcite , @xcite ) , the @xmath81 value becomes [email protected] .
the average of these two values ( rounded to significant digits ) is [email protected] , which we will take as our final estimate of @xmath105 for ngc 1055 .
the 90% complete sample for ngc 7339 also contained too few objects to use with kmm .
we find that for ngc 7339 , 21 out of 39 objects ( 54% ) fall on the blue side of @xmath101 , giving a @xmath106 .
if we assume a 70% blue fraction like the milky way or m31 , @xmath81 becomes [email protected] .
the average of these two values is [email protected] . as in section [
section : comparison ] we can compare and integrate our results for @xmath81 with well - determined values for other galaxies taken from our ongoing survey and the literature ( see table [ table : results ] ) .
the weighted mean @xmath81 for the nine spiral galaxies in the @xmath88@xmath218 sample is @xmath107 , which is consistent with our measurements for the spiral galaxies ngc 1055 and ngc 7339 . including the @xmath81 measurements for
ngc 1055 and ngc 7339 does not change the weighted mean : it stays at @xmath81 @xmath108 .
the nine e / s0 galaxies in the @xmath88@xmath218 sample have a weighted mean @xmath81 @xmath109 . adding in our new @xmath81 estimates for ngc 1023 and ngc 7332 again
does not change the weighted mean within the errors : the mean @xmath81 for all eleven e / s0 galaxies is @xmath110 .
the _ range _ of @xmath81 for early - type galaxies in the sample is fairly large , [email protected] , with massive cluster ellipticals having @xmath81 greater than @xmath92 .
the general trend identified in r05 i.e. , a rough increase in the number of blue , metal - poor gcs per galaxy mass with increasing galaxy stellar mass still holds with the addition of the four galaxies analyzed here .
this type of trend is consistent with the biased , hierarchical scenario of gc and galaxy formation outlined in @xcite and r05 . in this type of picture , today
s massive galaxies located in dense environments have comparatively high numbers of blue gcs for their stellar mass .
this is because a larger proportion of their mass was in place at high redshift ( massive galaxies began the collapse and assembly process earlier than lower - mass galaxies in low - density environments ) and thus could participate in star cluster formation when the first generation of gcs was forming in the universe .
detailed discussions of @xmath81 and possible scenarios that would give rise to the differences in @xmath81 values that we see are given in r05 , r07 and r10 ; suffice it to say here that the spiral and s0 galaxies in the current paper appear to have fairly typical @xmath81 for their masses and morphological types .
lastly , we note that as part of the analysis of the gc color distributions , we investigated whether any color gradients are present in the galaxies gc systems .
if the red , metal - rich gc population is less spatially extended compared to the blue , metal - poor gc population ( in other words , the ratio of red to blue gcs decreases with increasing radial distance from the galaxy center ) , one might expect to see a measurable color gradient in the gc system as a whole .
the presence of a color gradient could imply that the gas from which the red gcs formed was subject to additional dissipational collapse compared to the gas from which the blue gcs formed ( e.g. , az92 , forbes et al . 1997 , beasley et al .
2002 ) and/or that extra blue gcs are present in the galaxies halo that were accreted from metal - poor dwarf galaxies ( e.g. , ct et al .
1998 ) . in any case , only ngc 1023 and ngc 7332 had enough objects in the 90% sample to do this type of analysis .
we analyzed the @xmath8 color of the gc candidates as a function of projected radius and found no significant evidence for a radial color gradient in either galaxy s gc system .
figure [ fig : color_gradient ] shows the gc candidate colors versus projected radius for the 90% sample of ngc 1023 ( top ) and ngc 7332 ( bottom ) out to the radial extent of the gc system as determined in section [ section : radial profile ] .
the best fit least - squares line for each sample is plotted , with the slope determined by this fit indicated at the bottom of each plot . as explained in section [ section : final selection ] , during the gc candidate selection process for ngc 1023 we masked out a circular region with radius 0.6@xmath13 around the nearby dwarf satellite galaxy ngc 1023a .
eight gc candidates ( point sources with magnitudes and colors like gcs ) were located within the masked area . to determine what fraction of the eight objects are likely
to be truly associated with the dwarf galaxy ngc 1023a we must first consider possible sources of contamination .
we estimated the surface density of contaminants ( i.e. , foreground stars and background galaxies ) in the wiyn image based on the asymptotic behavior of the profile of ngc 1023 s gc system and derived a contamination level of @xmath111 objects arcmin@xmath39 ( section [ section : asymptotic ] ) . combining this surface density with
the area of the circular masked region around the dwarf galaxy implies that [email protected] contaminating objects should lie within that region .
we must also consider the gc system of ngc 1023 as a source of `` contaminants '' within the area around ngc 1023a .
accordingly , we used the de vaucouleurs fit to the radial profile of the gc system of ngc 1023 ( table [ table : fitting ] ) to calculate a surface density of gcs for each pixel in the masked region of ngc 1023a . multiplying this value by the area of each pixel and summing the values of all the pixels in the masked region together yields the total contribution from the gc system of ngc 1023 in the masked region . because we are concerned about the number of _ detected _ objects , we undo the gclf completeness scaling factor that was used to construct the final radial profile of ngc 1023 ( section [ section : gclf ] ) , yielding a value of @xmath112 detectable objects from the gc system of ngc 1023 coincident with the spatial position of the dwarf galaxy ngc 1023a .
considering both the contamination from foreground and background objects and from the ngc 1023 gc system , we expect a total contamination value of @xmath113 objects .
one more issue needs to be addressed before we can estimate the number of gcs in ngc 1023a .
this galaxy is a dwarf irregular ( dirr ) , so it has ongoing star formation .
in fact , @xcite spectroscopically confirmed the presence of two young ( @xmath9125@xmath10500 myr ) blue clusters in the central region of ngc 1023a , leading them to speculate that a recent ( @xmath114 myr ) close encounter with ngc 1023 may have led to an enhanced period of star formation in the dwarf galaxy .
young clusters like these may represent another source of contamination in our sample of gc candidates in ngc 1023a young , blue open clusters might have similar magnitudes and colors to gcs . to investigate this possibility ,
we have plotted in figure [ fig : color_mag_n1023a ] the integrated @xmath24 vs. @xmath115 for milky way open clusters @xcite , milky way globular clusters @xcite , and the eight gc candidates for ngc 1023a .
@xmath24 for the eight gc candidates in our sample is calculated by assuming the same distance to ngc 1023a as to ngc 1023 ( table [ table : galstats ] ) and errors are propagated from the distance modulus and photometric errors .
the dashed line marks the faint - end cut in our sample at @xmath116 ( section [ section : final selection ] ) . below this line
there is an area where the lowest - luminosity globular clusters and the most luminous , reddest open clusters do overlap . within our selection region
however , there is a clear separation between the milky way open clusters and globular clusters .
thus we do not consider open clusters to be a source of contamination in our gc candidate sample for ngc 1023a .
the number of gc candidates in ngc 1023a corrected for contamination is therefore [email protected] .
lastly , to produce a final estimated number of gcs in ngc 1023a , we need to correct for the magnitude incompleteness ( gclf coverage ) of our data .
@xcite studied the gc systems of 30 dirr galaxies and fitted the combined gclf of the systems with a gaussian function with a peak at @xmath117 and a dispersion of @xmath118 .
our number of detected gc candidates is far too small to attempt to directly fit a gclf to our data .
however , we note that the gclf peak in dirr galaxies is brighter than the gclf peak for giant galaxies ( @xmath44 ; @xcite ) .
given that we found our coverage of the gclf in the gc system of ngc 1023 to be 57% ( section [ section : gclf ] ) , we assume at a _ minimum _ the same level of coverage of the gclf of ngc 1023a . in an alternate scenario , we assume we have detected every gc in ngc 1023a . the 57%- and 100%-complete scenarios mark the range of possible values for our coverage of the gclf .
we average the result from these two possibilities and calculate that ngc 1023a has @xmath119 .
the error on this value was determined by adding in quadrature the poisson errors on the number of detected objects , the errors on the contamination estimates , and the uncertainty in the assumed gclf coverage scenarios .
@xcite published two estimates of total @xmath18 magnitude of ngc 1023a based on photoelectric photometry within two different apertures : @xmath120 @xmath2 15.27 @xmath65 0.03 for a 28@xmath121-diameter aperture and @xmath120 @xmath2 14.22 @xmath650.05 for a 45@xmath121-diameter aperture . using the distance modulus for ngc 1023 ( table [ table : galstats ] ) and taking the difference in values as the error
, we adopt a value of @xmath122 . calculating the luminosity - weighted specific frequency ( section [ section : sn_and_t ] )
we find that @xmath123 .
the error on @xmath68 was determined by propagating the error on @xmath1 and including the the uncertainty in the galaxy magnitude @xmath124 .
@xcite found a very broad range of @xmath68 for 13 dirr galaxies , spanning values from 0.3 to 11 .
the position of ngc 1023a toward the low end of this @xmath68 range indicates that the integrated luminosity is likely dominated by the young stellar population found by @xcite .
as this population fades , the @xmath68 value for ngc 1023a will increase .
we have obtained three - filter imaging with the 4096 x 4096-pixel minimosaic ccd imager on the wiyn 3.5-m telescope in order to study the gc populations of four s0 and spiral galaxies out to large galactocentric radii ( @xmath910@xmath1030 kpc ) . when possible , we combined the wiyn imaging data with archival and published hst / wfpc2 and keck - ii / lris data to help quantify the global properties of the galaxies gc populations .
we use the wiyn imaging and supplementary data to carefully construct the final radial distribution ( surface density of gcs vs. projected radius ) of each galaxy s gc system .
we then integrate the distributions to derive the total number of gcs ( @xmath99 ) .
the derived total numbers and specific frequencies for the four target galaxies , as well as values for other galaxies from our ongoing survey and from other published sources , are given in table [ table : results ] .
the number of gcs ranges from 75@xmath6510 for the least - massive galaxy of the set , the sbc galaxy ngc 7339 , to 490@xmath6530 for the massive s0 galaxy ngc 1023 .
the gc specific frequencies for these spiral and s0 galaxies are comparable to the mean values of other well - studied spiral and lenticular galaxies from our survey and from the literature .
we find that the @xmath8 color distribution of gcs in ngc 1023 is significantly bimodal ( @xmath795% confidence ) , whereas the color distribution of ngc 7332 s gc population is bimodal at only the @xmath981% confidence level .
the blue , metal - poor gcs in these galaxies make up @xmath960% of of the total population and the blue gc specific frequencies are consistent with the range of values expected for the galaxies morphological types .
we look for the presence of color gradients due to the changing ratio of blue to red gcs in the two galaxies with sufficient numbers of gc candidates to do this type of analysis and find no significant gradients in the gc systems . we have identified a subsample of eight gc candidates that coincide with the location of the dirr galaxy ngc 1023a , which lies @xmath92.5@xmath13 ( @xmath98 kpc ) in projected distance from ngc 1023 .
after accounting for possible contamination from the gc population of ngc 1023 and foreground and background sources , we estimate the dwarf galaxy has a total population of gcs equaling @xmath119 and a gc specific frequency @xmath123 .
both data acquisition and analysis for the wide - field gc system survey continue and we add new galaxies to the data set each semester .
the next two papers from the survey will describe the results for several elliptical and lenticular galaxies located both in the field and in galaxy clusters ( hargis & rhode 2012a , in preparation ) and present a multi - variate analysis of the current survey sample as a whole ( hargis & rhode 2012b , in preparation ) .
we thank sren larsen and jean brodie for providing us with the source list from their 2000 hst wfpc2 study of ngc 1023 s gc system . the research described in this paper
was supported by an nsf faculty early career development ( career ) award ( ast-0847109 ) to klr and also by a graduate fellowship from the indiana space grant consortium to jld .
we are grateful to the staff at the wiyn observatory and kitt peak national observatory for their assistance during our observing runs .
we thank an anonymous referee for providing a detailed report with many useful suggestions for improvements to the paper .
this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
lcrcccc ngc 1023 & s0 & @xmath125 & 11.4 & @xmath1021.2 & 11.3 & group + ngc 1055 & sb & @xmath126 & 16.3 & @xmath1021.0 & 11.1 & group + ngc 7332 & s0pec & @xmath127 & 23.0 & @xmath1020.7 & 11.1 & isolated pair + ngc 7339 & sbc & @xmath128 & 22.4 & @xmath1020.4 & 10.7 & isolated pair + [ table : galstats ] lcccccccc ngc1023 - 119 & 2:40:35.8 & + 38:58:00.9 & 19.022 & 0.001 & 0.987 & 0.002 & 0.616 & 0.001 + ngc1023 - 133 & 2:40:09.7 & + 38:58:07.2 & 19.100 & 0.001 & 0.800 & 0.002 & 0.466 & 0.001 + ngc1023 - 450 & 2:40:37.2 & + 39:02:06.9 & 19.116 & 0.001 & 0.814 & 0.002 & 0.484 & 0.001 + ngc1023 - 525 & 2:40:20.5 & + 39:02:47.6 & 18.903 & 0.001 & 0.573 & 0.002 & 0.347 & 0.001 + ngc1023 - 631 & 2:40:15.8 & + 39:03:30.2 & 19.289 & 0.001 & 0.754 & 0.002 & 0.406 & 0.001 + ngc1023 - 813 & 2:40:27.8 & + 39:04:40.7 & 19.155 & 0.001 & 0.738 & 0.002 & 0.520 & 0.001 + ngc1023 - 870 & 2:40:28.2 & + 39:05:17.6 & 19.056 & 0.001 & 0.733 & 0.002 & 0.430 & 0.001 + [ table : n1023 bright objects ] lrc 0.38 & @xmath133 & hst + 0.59 & @xmath134 & hst + 0.76 & @xmath135 & wiyn + 0.83 & @xmath136 & hst + 1.08 & @xmath137 & hst + 1.33 & @xmath138 & hst + 1.50 & @xmath139 & wiyn + 1.56 & @xmath140 & hst + 1.82 & @xmath141 & hst + 2.07 & @xmath142 & hst + 2.26 & @xmath143 & wiyn + 2.32 & @xmath144 & hst + 2.51 & @xmath145 & hst + 3.09 & @xmath146 & wiyn + 3.88 & @xmath147 & wiyn + 4.65 & @xmath148 & wiyn + 5.47 & @xmath149 & wiyn + 6.27 & @xmath150 & wiyn + 7.04 & @xmath151 & wiyn + 7.79 & @xmath152 & wiyn + [ table:1023radprof ] lrc 0.38 & @xmath161 & keck + 0.53 & @xmath162 & keck + 0.53 & @xmath163 & wiyn + 0.75 & @xmath164 & keck + 0.97 & @xmath165 & wiyn + 1.02 & @xmath166 & keck + 1.46 & @xmath167 & wiyn + 1.47 & @xmath168 & keck + 1.68 & @xmath169 & keck + 1.75 & @xmath170 & keck + 1.85 & @xmath171 & keck + 1.95 & @xmath172 & wiyn + 2.46 & @xmath173 & wiyn + 2.94 & @xmath174 & wiyn + 3.44 & @xmath175 & wiyn + 3.95 & @xmath176 & wiyn + 4.45 & @xmath177 & wiyn + 4.95 & @xmath178 & wiyn + [ table:7332radprof ] lr 0.32 & @xmath179 + 0.57 & @xmath180 + 0.86 & @xmath181 + 1.15 & @xmath182 + 1.46 & @xmath183 + 1.76 & @xmath184 + 2.05 & @xmath185 + 2.35 & @xmath186 + 2.65 & @xmath187 + 2.95 & @xmath188 + 3.25 & @xmath189 + 3.55 & @xmath190 + 3.85 & @xmath191 + [ table:7339radprof ] lrrrrrr ngc 1023 & @xmath192 & @xmath193 & 1.59 & @xmath194 & @xmath195 & 1.62 + ngc 1055 & @xmath196 & @xmath197 & 1.27 & @xmath198 & @xmath199 & 1.16 + ngc 7332 & @xmath200 & @xmath201 & 1.89 & @xmath202 & @xmath203 & 1.63 + ngc 7339 & @xmath204 & @xmath205 & 0.98 & @xmath206 & @xmath207 & 0.85 + [ table : fitting ] lccrcrrc ngc 1023 & sb0 & @xmath1021.2 & @xmath3 & @xmath70 & @xmath77 & @xmath208 & 14 + ngc 7332 & s0pec & @xmath1020.7 & @xmath5 & @xmath72 & @xmath79 & @xmath209 & 14 + ngc 1055 & sbb & @xmath1021.0 & @xmath4 & @xmath71 & @xmath78 & @xmath72 & 14 + ngc 7339 & sbc & @xmath1020.4 & @xmath67 & @xmath73 & @xmath80 & @xmath72 & 14 + ngc 1052 & e4 & @xmath1021.0 & @xmath210 & @xmath211 & @xmath212 & @xmath72 & 3 + ngc 3379 & e1 & @xmath1020.9 & @xmath213 & @xmath214 & @xmath79 & @xmath215 & 11 + ngc 4374 & e1 & @xmath1022.1 & @xmath216 & @xmath217 & @xmath218 & @xmath219 & 4 + ngc 4406 & e3 & @xmath1022.3 & @xmath220 & @xmath221 & @xmath222 & @xmath223 & 11 + ngc 4472 & e2 & @xmath1023.1 & @xmath224 & @xmath225 & @xmath226 & @xmath227 & 9 + ngc 5128 & epec & @xmath1022.1 & @xmath228 & @xmath229 & @xmath230 & @xmath231 & 7 + ngc 3384 & sb0 & @xmath1020.5 & @xmath232 & @xmath233 & @xmath234 & @xmath235 & 6 + ngc 4594 & s0/sa & @xmath1022.4 & @xmath236 & @xmath237 & @xmath238 & @xmath83 & 11 + ngc 7457 & s0 & @xmath1019.5 & @xmath239 & @xmath218 & @xmath240 & @xmath241 & 5 + milky way & sbc & @xmath1021.3 & @xmath242 & @xmath89 & @xmath90 & @xmath107 & 1 + m31 & sb & @xmath1021.8 & @xmath243 & @xmath71 & @xmath244 & @xmath214 & 1,2,8 + ngc 891 & sb & @xmath1020.8 & @xmath245 & @xmath246 & @xmath247 & @xmath248 & 13 + ngc 2683 & sb & @xmath1020.5 & @xmath249 & @xmath250 & @xmath251 & @xmath252 & 12 + ngc 3556 & sc & @xmath1021.2 & @xmath253 & @xmath254 & @xmath255 & @xmath256 & 12 + ngc 4013 & sb & @xmath1020.4 & @xmath257 & @xmath215 & @xmath212 & @xmath231 & 13 + ngc 4157 & sb & @xmath1020.4 & @xmath258 & @xmath247 & @xmath259 & @xmath260 & 12 + ngc 7331 & sb & @xmath1021.7 & @xmath261 & @xmath262 & @xmath263 & @xmath264 & 12 + ngc 7814 & sab & @xmath1020.4 & @xmath265 & @xmath266 & @xmath267 & @xmath268 & 10 + [ table : results ] | [ section : abstract ] we present results from a study of the globular cluster ( gc ) systems of four spiral and s0 galaxies imaged as part of an ongoing wide - field survey of the gc systems of giant galaxies .
the target galaxies the sb0 galaxy ngc 1023 , the sbb galaxy ngc 1055 , and an isolated pair comprised of the sbc galaxy ngc 7339 and the s0 galaxy ngc 7332 were observed in @xmath0 filters with the wiyn 3.5-m telescope and minimosaic camera . for two of the galaxies , we combined the wiyn imaging with previously - published data from the hubble space telescope and the keck observatory to help characterize the gc distribution in the central few kiloparsecs .
we determine the radial distribution ( surface density of gcs versus projected radius ) of each galaxy s gc system and use it to calculate the total number of gcs ( @xmath1 ) .
we find @xmath1 @xmath2 @xmath3 , @xmath4 , @xmath5 , and @xmath6 for ngc 1023 , ngc 1055 , ngc 7332 , and ngc 7339 , respectively .
we also calculate the gc specific frequency ( @xmath1 normalized by host galaxy luminosity or mass ) and find values typical of those of the other spiral and e / s0 galaxies in the survey .
the two lenticular galaxies have sufficient numbers of gc candidates for us to perform statistical tests for bimodality in the gc color distributions .
we find evidence at a high confidence level ( @xmath795% ) for two populations in the @xmath8 distribution of the gc system of ngc 1023 .
we find weaker evidence for bimodality ( @xmath781% confidence ) in the gc color distribution of ngc 7332 .
finally , we identify eight gc candidates that may be associated with the magellanic dwarf galaxy ngc 1023a , a satellite of ngc 1023 . |
in the main text , we have studied the @xmath53-limit and @xmath64-limit of the quadrupole susceptibility @xmath175 , and found that the relationship @xmath176 is satisfied .
the basis of this relationship is that the intra - band pauli term is absent in both @xmath82 and @xmath177 in the absence of the elastic and inelastic scattering .
however , the relationship @xmath176 is not trivial when the scattering processes exist . here , we calculate both @xmath82 and @xmath102 in the presence of the local nonmagnetic impurities based on the @xmath72-matrix approximation in the five - orbital model . for the charge quadrupole susceptivility , the relationship @xmath176 is confirmed even in the presence of impurities .
we assume that the impurity potential @xmath178 is diagonal in the orbital basis .
( we write @xmath179 , @xmath25 , @xmath180 , @xmath181 , @xmath182 orbitals as @xmath183 , respectively . ) then , the @xmath72-matrix in the orbital basis is given as @xmath184 where @xmath185 and the green function is @xmath186 , and @xmath187 is the impurity self - energy when the impurity concentration is @xmath188 . the bethe - salpeter equation for the one - particle operator
@xmath189 is @xmath190 where @xmath191 and @xmath192 .
we will show the significant role of the vc given by the second term ; ( @xmath193 ) .
first , we study the impurity effect on the bare - bubble @xmath39 for the @xmath194 quadrupole .
the impurity effect is divided into the ( i ) self - energy correction ( [ eqn : sigma ] ) and ( ii ) vertex correction ( [ eqn : threev ] ) .
if only ( i ) is taken into account , the bare - bubble within the @xmath25-orbital is given as @xmath195 where @xmath196 includes the self - energy , and the suffix @xmath197 in @xmath196 represents the @xmath25-orbital . if both ( i ) and ( ii ) is taken into account , it is given as @xmath198 for @xmath199 in eq .
( [ eqn : threev ] ) , where @xmath200 represents the @xmath21-orbital .
@xmath201 gives the correct susceptibility for @xmath202 , whereas @xmath203 is incorrect .
, ( b ) @xmath204 ( c ) @xmath205 , and ( d ) @xmath206 as functions of @xmath207 .
the correct results are given in ( b ) and ( d ) . here ,
@xmath64-limit values are obtained by extrapolating the data at @xmath208 with @xmath209 to the real axis numerically . ] here , we discuss the susceptibilities in the @xmath53-limit and @xmath64-limit . using eq .
( [ eqn : chisig ] ) or ( [ eqn : chisiglam ] ) , the former is simply given as @xmath210 at @xmath211 and @xmath212 .
here , we derive the latter numerically by extrapolating the data at @xmath208 with @xmath209 to the real axis .
this procedure is successful at sufficiently low temperatures .
figure [ fig : ap ] ( a ) and ( b ) represent the numerically obtained @xmath213 and @xmath214 for @xmath215 , respectively .
we fix @xmath216 mev and @xmath217 . in ( a ) , @xmath218 quickly increases with @xmath207 , and it is almost equal to @xmath219 just for @xmath220 . in ( b ) , in contrast , @xmath221 does not reach the @xmath53-limit value even for @xmath222 . in both ( a ) and ( b ) , impurity effect on the @xmath53-limit value
is very small .
since @xmath204 gives the true susceptibility , we conclude that the relationship @xmath223 is satisfied even for @xmath202 . in fig .
[ fig : ap ] ( a ) , @xmath224 approaches to the @xmath53-limit value for @xmath202 , since the intra - band pauli term also contributes to the @xmath64-limit ( @xmath212 and @xmath225 ) due to the broadening of the quasiparticle spectrum caused by im@xmath226 .
however , the impurity three - point vertex @xmath227 takes large value for @xmath228 , and it suppresses the pauli term .
these effects exactly cancel for conserved quantities : for this reason , the charge and spin susceptibilities become zero in the @xmath64-limit even for @xmath202 .
although @xmath194 is not conserved , the vc in @xmath227 is nonzero in the present model , and therefore the relationship @xmath229 is satisfied .
next , to discuss the al - vc , we calculate @xmath230 at @xmath231 introduced in the main text ( eq .
( [ eqn : l - k ] ) ) in the presence of impurities ( @xmath215 ) , by which the al - vc is given as @xmath232 .
figure [ fig : ap ] ( a ) shows the numerically obtained @xmath233 , in which only @xmath234 is included .
we see that @xmath235 increases with @xmath207 , and coincides with the @xmath53-limit value just for @xmath220 .
we also calculate @xmath206 , in which both @xmath234 and @xmath236 are taken into account properly . in this case , @xmath237 does not reach the @xmath53-limit value even for @xmath238 thanks to the vc in @xmath236 .
since @xmath206 gives the true vertex function , the relation of the al - vc @xmath239 is confirmed even for @xmath202 . in summary
, we confirmed that the relationship @xmath176 is satisfied in the presence of impurities , by taking both @xmath240 and @xmath236 into account correctly .
in other words , although the relation @xmath241 is obtained by including @xmath240 only , it is an artifact due to the neglect of the vc in @xmath236 .
( since @xmath242 for the charge current @xmath243 , such discontinuity will be absent for the conductivity . ) in real compounds , the raman vertex @xmath244 is very complex and momentum dependent . in this paper , we take the momentum average of @xmath245 ( @xmath246 ) , and consider the constant raman vertex @xmath247 to simplify the discussion . in the present multiorbital model , @xmath236 does not vanish even if the @xmath248-dependence of @xmath244 is taken into account , so the relationship @xmath176 should be satisfied for @xmath202 . | the emergence of the nematic order and fluctuations has been discussed as a central issue in fe - based superconductors . to clarify the origin of the nematicity , we focus on the shear modulus @xmath0 and the raman quadrupole susceptibility @xmath1 . due to the aslamazov - larkin vertex correction ,
the nematic - type orbital fluctuations are induced , and they enhances both @xmath2 and @xmath1 strongly .
however , @xmath1 remains finite even at the structure transition temperature @xmath3 , because of the absence of the band jahn - teller effect and the pauli ( = intra - band ) contribution , as proved in terms of the linear response theory .
the present study clarifies that origin of the nematicity in fe - based superconductors is the nematic - orbital order / fluctuations . in fe - based superconductors ,
the nematic order and fluctuations attract great attention as one of the essential properties of the electronic states . a schematic phase diagram of bafe@xmath4as@xmath4 as a function of carrier doping @xmath5 is shown in fig .
[ fig : fs ] : for @xmath6 ( e - doping ) , the non - magnetic orthorhombic ( @xmath7 ) phase transition occurs at @xmath3 , and the antiferro ( af ) spin order is realized at @xmath8 in the @xmath7 phase . in ba(fe@xmath9co@xmath10)@xmath4as@xmath4 ( @xmath11 ) , both the structural and magnetic quantum critical points ( qcps ) are very close , and strong magnetic fluctuations are observed near the qcps by nmr @xcite .
in addition , strong nematic susceptibility that couples to the @xmath7 structure deformation had been observed via the softening of shear modulus @xmath0 @xcite and in - plane anisotropy of resistivity @xcite .
similar softening of @xmath0 is also observed in ( ba@xmath9k@xmath10)fe@xmath4as@xmath4 ( @xmath12 ; h - doping ) @xcite and fe(se , te ) ( @xmath13 ) @xcite .
interestingly , in ba(fe@xmath9ni@xmath10)@xmath4as@xmath4 ( @xmath14 ) , magnetic qcp and structural qcp are well separated , and quantum criticalities are realized at both points @xcite .
then , a natural question is what is the `` nematic order parameter '' that would be closely related to the pairing mechanism . up to now , both the spin - nematic mechanism @xcite and ferro - orbital order mechanism @xcite had been proposed , and the softening of @xmath0 can be fitted by both mechanisms @xcite .
the former predicts that the spin - nematic order @xmath15 occurs above @xmath16 when the magnetic order @xmath17 is suppressed by the @xmath18-@xmath19 frustration . as for the latter scenario
, it was shown that the orbital order @xmath20 is induced by spin fluctuations , due to strong spin - orbital mode - coupling given by the vertex correction ( vc ) @xcite .
the large @xmath21-orbital level splitting @xmath22 mev in the @xmath7 phase @xcite may be too large to be produced by spin nematic order via spin - lattice coupling .
recently observed large quadrupole susceptibility @xmath1 by electron raman spectroscopy @xcite presents a direct evidence of the strong orbital fluctuations .
although this result favors the orbital nematic scenario , the observed enhancement of @xmath1 is apparently smaller than the orbital susceptibility extracted from @xmath0 .
for example , @xmath1 remains finite at @xmath23 , although @xmath24 diverges at @xmath3 .
therefore , it should be verified whether both @xmath0 and @xmath1 can be explained based on the orbital nematic scenario or not . .
the weight of @xmath25 orbital is stressed by green circles .
( c ) relation @xmath26 .
( d ) particle - hole excitation continuum .
, title="fig : " ] .
the weight of @xmath25 orbital is stressed by green circles .
( c ) relation @xmath26 .
( d ) particle - hole excitation continuum .
, title="fig : " ] in this paper , we analyze both @xmath0 and @xmath1 , both of which are key experiments to uncover the nematic order parameter .
it is found that both @xmath0 and @xmath1 are enhanced by the orbital fluctuations due to aslamazov - larkin type vc ( al - vc ) .
however , @xmath1 is less singular since the band jahn - teller ( band - jt ) effect and the pauli ( = intra - band ) quadrupole susceptibilities does not contribute to @xmath1 .
since both @xmath0 and @xmath1 are explained satisfactorily , the orbital nematic scenario is essential for many fe - based superconductors . as for the pairing mechanism ,
at present , both the spin fluctuation mediated @xmath27 wave state @xcite and orbital fluctuation mediated @xmath28 wave state @xcite have been discussed .
when both fluctuations coexist , nodal @xmath29-wave state can be realized @xcite .
the @xmath28-wave state is consistent with the robustness of @xmath30 against impurities @xcite and broad hump structure in the inelastic neutron scattering @xcite .
the self - consistent vertex correction ( sc - vc ) method @xcite predicts the developments of ferro- and af - orbital fluctuations , and the freezing of the latter fluctuations would explain the nematic order at @xmath31k ( @xmath32 ) @xcite .
first , we discuss the susceptibility at @xmath33 with respect to the quadrupole order parameter @xmath34 in the hubbard model . for @xmath35 ,
it is approximately given as @xcite @xmath36 where @xmath37 , and @xmath38 is the intra - orbital ( within @xmath25 orbital ) irreducible susceptibility : @xmath39 is the non - interacting susceptibility and @xmath40 is the vc for the charge channel . the orbital nematic order @xmath20 occurs when the charge stoner factor @xmath41 reaches unity , which is realized near the magnetic qcp since the al - vc is proportional to the square of the magnetic correlation length @xcite .
next , we discuss the `` total '' quadrupole susceptibility in real systems , by including the realistic quadrupole interaction due to the acoustic phonon for the orthorhombic distortion . according to ref .
@xcite , it is given as @xmath42 , where @xmath43 is the quadrupole operator , and @xmath44 is the phonon propagator multiplied by the coupling constants .
@xmath45 is the phonon velocity .
since the migdal s theorem tells that the effect of @xmath46 on the irreducible susceptibility is negligible , the total susceptibility is @xmath47 now , we discuss the acoustic and optical responses based on the total susceptibility ( [ eqn : chi - tot ] ) , by taking notice that any susceptibilities in metals are _ discontinuous _ at @xmath48 .
since the elastic constant is measured under the static ( @xmath49 ) strain with long wavelength ( @xmath50 ) , @xmath0 is given as @xmath51 where @xmath52 is called the @xmath53-limit , and the relation @xmath54 for @xmath49 is taken into account .
the structure transition occurs when @xmath24 diverges .
when the al - vc is negligible , @xmath55 is as small as @xmath56 . even in this case
, @xmath24 can diverge when @xmath46 is very large , which is known as the band - jt effect .
however , the band - jt mechanism can not explain the strong enhancement of @xmath1 , as we will clarify later .
in fact , the fitting of experimental data in the present paper indicates that the softening of @xmath0 is mainly given by the al - vc : the relation @xmath57 is satisfied in fe - based superconductors .
next , we derive the optical response in the dc limit , measured by using the low - energy photon with @xmath58 and @xmath59 . considering that the photon velocity @xmath60 is much faster than the fermi velocity @xmath61 and @xmath45 , it is given as @xmath62 where @xmath63 is called the @xmath64-limit @xcite .
since @xmath65 is zero for @xmath66 , the band - jt effect does not contribute to the raman susceptibility .
the physical explanation is that the acoustic phonons can not be excited by photons because of the mismatch of the wavelengths @xmath67 for the same @xmath64 as shown in fig .
[ fig : fs ] ( c ) .
also , since @xmath68 , low - energy photon can not induce the intraband particle - hole excitation as understood from the location of the particle - hole continuum shown in fig .
[ fig : fs ] ( d ) .
this fact leads to the relationship `` @xmath69 is smaller than @xmath55 '' as we discuss mathematically later . for the charge quadrupole susceptibility ,
this relationship holds even if the quasiparticle lifetime is finite due to impurity scattering ; see the supplemental material @xcite .
therefore , @xmath1 remains finite at @xmath70 although @xmath24 diverges at @xmath3 , consistently with experiments @xcite . and
( b ) @xmath71 as functions of @xmath72 .
their @xmath72-dependences originates from @xmath73 since @xmath74 is fixed . ]
hereafter , we perform the numerical calculation of the quadrupole susceptibility in the five - orbital model
. the unit of energy is ev unless otherwise noted .
first , we discuss the @xmath53-limit and @xmath64-limit of the bare bubble made of two @xmath25-orbital green functions .
they are connected by the following relation : @xmath75 where @xmath76 is the weight of the @xmath25-orbital on band @xmath77 , and @xmath78 . in eq .
( [ eqn : chi0k ] ) , @xmath79 is given by only the inter - band ( @xmath80 ) contribution , which is called the van - vleck term .
therefore , @xmath81 is always larger than @xmath82 due to the intra - band contribution given by the second term in eq .
( [ eqn : chi0k ] ) , called the pauli term .
we obtain @xmath83 and @xmath84 in the present model .
next , we analyze al - vc in detail , since it is the main driving force of the orbital fluctuations .
the analytic expression of the al term is given in refs .
@xcite . to simplify the discussion
, we consider the intra - orbital ( within @xmath25-orbital ) al - term .
then , @xmath85 is approximately given as @xmath86 where @xmath87 and @xmath88 at @xmath89 : @xmath90 is the three - point vertex made of three green functions @xcite . also , @xmath91 , where @xmath92 is the spin susceptibility for @xmath25-orbital . here , we assume the following millis - monien - pines form of @xmath92 @xcite : @xmath93 where @xmath94
, @xmath95 is the square of the spin correlation length , and @xmath96 is the spin - fluctuation energy scale .
quantitatively speaking , @xmath85 given by eq .
( [ eqn : xkw ] ) is underestimated since non - zero matsubara terms are dropped . however , in the classical region @xmath97 , which is realized in optimally - doped ba(fe , co)@xmath4as@xmath4 @xcite , @xmath98 for @xmath99 is negligibly small . in this case , we can safely use eq .
( [ eqn : xkw ] ) . according to eqs .
( [ eqn : xkw ] ) and ( [ eqn : mmp ] ) , we obtain @xmath100 for two - dimensional systems . @xmath101 and @xmath102 are connected by the following relation : @xmath103 where @xmath102 is the inter - band van - vleck term @xcite . for @xmath104 , @xmath101 increases strongly at low @xmath72 , because of the intra - band pauli term in the second term of eq .
( [ eqn : l - k ] ) .
its main contribution is given by @xmath105 and @xmath106 in fig .
[ fig : fs ] ( b ) . both pauli and van - vleck terms are negative in the present model .
therefore , the relationship @xmath107 is satisfied .
figure [ fig : numerical ] ( a ) shows the temperature dependence of @xmath108 given by eq .
( [ eqn : xkw ] ) , by using the static rpa spin susceptibility @xmath109 obtained at @xmath110 . in this calculation , @xmath111 is fixed , where @xmath112 is the spin stoner factor .
thus , we obtain the relationship @xmath113 , in which the factor @xmath114 originates from the strong @xmath72-dependence of @xmath115 .
we also show the temperature dependence of @xmath71 in fig .
[ fig : numerical ] ( b ) : the relation @xmath116 is realized due to the @xmath72-dependence of @xmath117 @xcite . therefore , the relationship @xmath107 is confirmed by the present calculation .
normalized by the 33% co - ba122 data in ref .
@xcite , shown by broad red lines .
the dotted lines @xmath0 is the fitting result under the constraint @xmath118 at @xmath23 , and the broken lines @xmath119 is the fitting without constraint .
( b ) the weiss temperature @xmath120 given by the present fitting .
@xmath121 is the weiss temperature of @xmath122 @xcite , and @xmath123 is given by the raman spectroscopy @xcite .
( c ) @xmath124 and @xmath125 given by the fitting of @xmath0 . experimental data of @xmath126 are shown by red circles @xcite . ] here , we perform the fitting of experimental data . to reduce the number of fitting parameters
, we put @xmath127 by assuming @xmath128 , which would be justified since the relation @xmath129 is predicted by the first principle study @xcite . also , we put @xmath130 by assuming that @xmath131 .
then , eqs .
( [ eqn : lr - c66 ] ) and ( [ eqn : lr - raman ] ) are simplified as @xmath132 where @xmath133 and @xmath134 : according to fig.[fig : numerical ] , @xmath135 and @xmath136 for @xmath137 .
first , we fit the data of @xmath138 , which is normalized by the shear modulus due to phonon anharmonicity ( = 33% co - ba122 data ) given in ref .
@xcite .
we putting @xmath139 , and the remaining fitting parameters are @xmath140 and @xmath120 . figure [ fig : c66 ] ( a ) shows the fitting result for ba(fe@xmath9co@xmath10)@xmath4as@xmath4 : the `` dotted line @xmath0 '' is the fitting result of @xmath138 under the constraint @xmath118 at @xmath23 .
we fix @xmath141 for all @xmath142 , and change @xmath120 from @xmath143k to @xmath144k .
the `` broken line @xmath119 '' is the fitting for @xmath145 without the constraint , by using @xmath146 .
thus , both fitting methods can fit the @xmath72- and @xmath142-dependences of @xmath138 very well by choosing only @xmath147 with a fixed @xmath148 .
figure [ fig : c66 ] ( b ) shows the obtained @xmath147 by @xmath0-fitting ( @xmath149 ) and by @xmath119-fitting ( @xmath150 ) , as explained above .
the obtained @xmath147 is very close to @xmath121 given by the curie - weiss fitting of @xmath122 @xcite , which manifests the importance of the al - vc .
also , @xmath123 is given by the raman spectroscopy @xcite .
k@xmath10)fe@xmath4as@xmath4 .
experimental data @xmath138 are shown by broad red lines @xcite . ] in fig .
[ fig : c66 ] ( c ) , we show @xmath124 obtained by the fitting of @xmath138 for ba(fe@xmath9co@xmath10)@xmath4as@xmath4 at @xmath151 and @xmath152 , we also show @xmath153 according to the numerical result in fig . [ fig : numerical ] , by putting @xmath154k . in fig .
[ fig : c66 ] ( c ) , all the data are normalized as unity at 300k . then , the relation @xmath155 is well satisfied , as expected from eq .
( [ eqn : lr - raman-2 ] ) .
in addition , the relation @xmath156 holds for @xmath70 , consistently with the report in ba(fe@xmath9co@xmath10)@xmath4as@xmath4 @xcite .
figure [ fig : c66-k ] ( a ) and ( b ) shows the fitting results for ( ba@xmath9k@xmath10)fe@xmath4as@xmath4 for @xmath157 ( @xmath139 ; @xmath141 for @xmath0 and @xmath146 for @xmath119 ) and @xmath158 ( @xmath159 ; @xmath160 for @xmath119 ) , respectively . in the present theory , we can explain the existence of inflection points of @xmath0 in over - doped region ( without structure transition ) reported experimentally @xcite , shown by large blue circles .
the inflection point originates from the factor @xmath161 in @xmath162 .
the fitting of over - doped data could be improved by considering the deviation from the relation @xmath163 at low @xmath72 , as recognized in fig .
[ fig : numerical ] ( a ) .
in addition , for @xmath164 , experimental pseudo - gap behavior of @xmath165 below @xmath166k @xcite would also be related to the inflection point of @xmath0 . in the present theory
, we can fit @xmath138 very well for both over - doped and under - doped regions in ba@xmath9k@xmath10fe@xmath4as@xmath4 .
however , different set of parameters should be used in each region : this fact indicates that the orthorhombic phase and superconducting phase are separated by the first - order transition .
in fact , the @xmath167-like resistivity at the optimum doping @xmath168 indicates the absence of the orbital - nematic qcp in this compound .
we also note that the change in the topology of the electron - pockets , called the lifshitz transition , occurs in ba@xmath9k@xmath10fe@xmath4as@xmath4 near the optimal doping . in this paper
, we showed that raman susceptibility at @xmath49 is enlarged by the al - vc .
the present theory predicts that the @xmath64-dependence of the ac raman susceptibility follows @xmath169 , and @xmath170 is approximately @xmath171 .
however , @xmath170 could be modified by the @xmath64-dependence of @xmath172 . in summary
, we presented a unified explanation for the softening of @xmath0 and enhancement of @xmath1 based on the five - orbital model . both @xmath2 and @xmath1
are enhanced by the nematic - type orbital fluctuations induced by the al - vc .
however , @xmath1 remains finite even at the structure transition temperature @xmath3 , because of the absence of the band - jt effect and the pauli ( = intra - band ) contribution .
the present study clarified that the origin of the nematicity , which is a central issue in fe - based superconductors , is the nematic - orbital order / fluctuations .
we are grateful to a.e .
bhmer for offering us her experimental data published in ref .
@xcite .
we also thank y. gallais , a.v .
chubukov , j. schmalian , r. fernandes and s. onari for useful discussions .
this study has been supported by grants - in - aid for scientific research from mext of japan .
part of numerical calculations were performed on the yukawa institute computer facility .
99 f. l. ning , k. ahilan , t. imai , a. s. sefat , m. a. mcguire , b. c. sales , d. mandrus , p. cheng , b. shen , and h .-
h wen , phys .
rev .
lett .
* 104 * , 037001 ( 2010 ) r.m .
fernandes , l. h. vanbebber , s. bhattacharya , p. chandra , v. keppens , d. mandrus , m.a .
mcguire , b.c .
sales , a.s .
sefat , and j. schmalian , phys .
rev .
lett . * 105 * , 157003 ( 2010 ) .
m. yoshizawa , d. kimura , t. chiba , s. simayi , y. nakanishi , k. kihou , c .- h .
lee , a. iyo , h. eisaki , m. nakajima , and s. uchida , j. phys .
. jpn . * 81 * , 024604 ( 2012 ) .
s. simayi , k. sakano , h. takezawa , m. nakamura , y. nakanishi , k. kihou , m. nakajima , c .- h .
lee , a. iyo , h. eisaki , s. uchida , and m. yoshizawa , j. phys .
. jpn . * 82 * , 114604 ( 2013 ) .
a. e. bhmer , p. burger , f. hardy , t. wolf , p. schweiss , r. fromknecht , m. reinecker , w. schranz , and c. meingast , phys .
rev .
lett .
* 112 * , 047001 ( 2014 ) .
t. goto , r. kurihara , k. araki , k. mitsumoto , m. akatsu , y. nemoto , s. tatematsu , and m. sato , j. phys .
. jpn . * 80 * , 073702 ( 2011 ) .
h .- h .
kuo , j. g. analytis , j .- h .
chu , r. m. fernandes , j. schmalian , and i. r. fisher , phys .
rev .
b * 86 * , 134507 ( 2012 ) .
m. yoshizawa , private communication .
r. zhou , z. li , j. yang , d. l. sun , c. t. lin , and g. zheng , nat .
commun * 4 * , 2265 , ( 2013 ) .
f. krger , s. kumar , j. zaanen , j. van den brink , phys .
rev .
b * 79 * , 054504 ( 2009 ) .
w. lv , j. wu and p. phillips , phys .
rev .
b * 80 * , 224506 ( 2009 ) ; w. lv , f. kruger , and p. phillips , phys
. rev .
b * 82 * , 045125 ( 2010 ) .
c .- c .
lee , w .- g .
yin , and w. ku , phys .
rev .
lett . * 103 * , 267001 ( 2009 ) s. onari and h. kontani , phys .
rev .
lett . * 109 * , 137001 ( 2012 ) .
h. kontani , y. inoue , t. saito , y. yamakawa and s. onari , solid state communications , * 152 * , 718 ( 2012 ) .
r.m .
fernandes and a.j .
millis , phys .
rev .
lett . * 111 * , 127001 ( 2013 ) .
y. ohno , m. tsuchiizu , s. onari , and h. kontani , j. phys .
. jpn . * 82 * , 013707 ( 2013 ) .
m. tsuchiizu , y. ohno , s. onari and h. kontani , phys .
rev .
lett . *
111 * , 057003 ( 2013 ) .
m. yi , d. h. lu , j .- h .
chu , j. g. analytis , a. p. sorini , a. f. kemper , b. moritz , s .- k .
mo , r. g. moore , m. hashimoto , w .- s .
lee , z. hussain , t. p. devereaux , i. r. fisher , and z .- x .
shen , proc .
natl .
acad .
sci .
usa * 108 * , 6878 ( 2011 ) .
h. miao , l .- m .
wang , p. richard , s .- f .
wu , j. ma , t. qian , l .- y .
xing , x .- c .
wang , c .- q .
jin , c .- p .
chou , z. wang , w. ku , and h. ding , arxiv.1310.4601 y. gallais , r. m. fernandes , i. paul , l. chauviere , y .- x .
yang , m .- a .
measson , m. cazayous , a. sacuto , d. colson , and a. forget , phys .
rev .
lett . * 111 * , 267001 ( 2013 ) . y .-
x .
yang , y. gallais , r. m fernandes , i. paul , l. chauviere , m .- a .
measson , m. cazayous , a. sacuto , d. colson , and a. forget arxiv:1310.0934 k. kuroki , s. onari , r. arita , h. usui , y. tanaka , h. kontani , and h. aoki , phys .
rev .
lett .
* 101 * , 087004 ( 2008 ) .
p. j. hirschfeld , m. m. korshunov , i. i. mazin rep .
. phys . * 74 * , 124508 ( 2011 ) .
a. v. chubukov , d. v. efremov , and i. eremin , phys .
rev .
b * 78 * , 134512 ( 2008 ) . h. kontani and s. onari , phys
. rev .
lett . * 104 * , 157001 ( 2010 ) .
s. onari , y. yamakawa and h. kontani , phys .
rev .
lett . * 112 * , 187001 ( 2014 ) .
t. saito , s. onari and h. kontani , phys .
rev .
b * 88 * , 045115 ( 2013 ) .
s. onari and h. kontani , phys .
rev .
lett . * 103 * , 177001 ( 2009 ) .
y. yamakawa , s. onari , and h. kontani , phys .
rev .
b * 87 * , 195121 ( 2013 ) .
s. onari , h. kontani and m. sato , phys .
rev .
b * 81 * , 060504(r ) ( 2010 ) s. onari and h. kontani , phys . rev .
b * 84 * , 144518 ( 2011 ) s. kasahara , h. j. shi , k. hashimoto , s. tonegawa , y. mizukami , t. shibauchi , k. sugimoto , t. fukuda , t. terashima , a. h. nevidomskyy , and y. matsuda , nature * 486 * , 382 ( 2012 ) .
y. k. kim , w. s. jung , g. r. han , k .- y .
choi , c .- c .
chen , t. p. devereaux , a. chainani , j. miyawaki , y. takata , y. tanaka , m. oura , s. shin , a. p. singh , h. g. lee , j .- y .
kim , and c. kim , phys .
rev .
lett . * 111 * , 217001 ( 2013 ) .
h. kontani , t. saito and s. onari , phys .
rev .
b * 84 * , 024528 ( 2011 ) .
p. nozieres , _ theory of interacting fermi systems _
( benjamin , new york , 1964 ) ; a. a. abrikosov , l. p. gorkov and i. e. dzyaloshinski , _ methods of quantum field theory in statistical physics _
dover , new york , 1975 ) ; a. j. leggett , phys . rev . * 140 * , a1869 ( 1965 ) .
h. kontani , and k. yamada , j. phys .
. jpn . * 65 * , 172 ( 1996 ) ; h. kontani , and k. yamada , j. phys .
soc .
jpn .
* 66 * , 2232 ( 1997 ) .
h. kontani and y. yamakawa , supplemental material a .- j .
millis , h. monien and d. pines , phys .
rev .
b * 42 * , 167 ( 1990 ) ; p. monthoux and d. pines , phys . rev .
b * 47 * , 6069 ( 1993 ) .
p. steffens , c.h .
lee , n. qureshi , k. kihou , a. iyo , h. eisaki , and m. braden , phys .
rev .
lett . * 110 * , 137001 ( 2013 ) .
the analytic expression of @xmath102 is given as @xmath173 t. miyake , k. nakamura , r. arita and m. imada , j. phys .
. jpn . * 79 * , 044705 ( 2010 ) .
m. hirano , y. yamada , t. saito , r. nagashima , t. konishi , t. toriyama , y. ohta , h. fukazawa , y. kohori , y. furukawa , k. kihou , c .- h .
lee , a. iyo , and h. eisaki , j. phys .
soc .
jpn . , * 81 * , 054704 ( 2012 ) . |
we obtained high - resolution , high - contrast two - photon excited fluorescence images of ca@xmath0 transients in live neuronal cell bodies at frame rates of up to 100 hz .
figure [ fig1]a shows widefield single - photon and widefield two - photon excited fluorescence images of live cells with continuous irradiation .
the frame rates were 1 hz , 10 hz and 100 hz .
the excitation power was adjusted to give a similar subjective appearance in the first images of all the series .
this strategy was adopted to achieve similar signal - to - noise values at the start of all the series .
the fluorescence signal intensities for three regions of interest ( rois ) located over neuronal cell bodies in each culture dish were measured from time t=0 to 590 seconds at 5-second intervals using imagej @xcite .
an average value of fluorescence intensity from the three rois was obtained for each dataset , and was then normalised relative to the fluorescence signal intensity at time t=0 seconds .
the resultant plots of normalised average fluorescence signal intensity with time for single - photon excitation and two - photon excitation are presented as figs .
[ fig1]b and [ fig1]c . for both single - photon and two - photon excitation at a frame rate of 1 hz , the overall fluorescence signal intensity
increased over time but was much greater in the single - photon case , with an increase in signal in excess of 100 % compared to a 25 % increase using two - photon excitation over the same time period . however , at a frame rate of 10 hz , there was very rapid fading using single - photon excitation ( following a large ca@xmath0 spike at around 30 seconds ) , whereas the fading using two - photon excitation was minimal , with less than 5 % loss of fluorescence signal intensity after 590 seconds . at a frame rate of 100 hz , however , recordings were quite different under the two illumination regimes . using single - photon excitation , there was a rapid decrease in fluorescence signal intensity , with almost no fluorescence recorded after 590 seconds . for two - photon excitation ,
the fluorescence signal intensity increased slowly over the first 200 seconds and then began to decrease .
so even at this fast framing rate ( with increased irradiation to compensate for the shorter integration time ) , useful images were being obtained with widefield two - photon excitation long after the conventional single - photon recordings had faded totally .
we noted also that that the rate of fluorescence intensity increase under two - photon excitation at 100 hz was less rapid than that measured for single - photon excitation at the much slower speed of 1 hz .
supplementary video 1 shows time - lapse recordings of live neuronal cell bodies loaded with fluo-4 am at a frame rate of 100 hz using single - photon ( left ) and two - photon ( right ) widefield excitation .
the neuronal cultures were irradiated continuously for 590 seconds and the video shows a frame every 5 seconds , demonstrating very clearly the rapid decrease in fluorescence intensity observed using widefield single - photon excitation , while no significant reduction in fluorescence intensity was observed using widefield two - photon excitation .
the mechanisms responsible for such light - induced changes over time are at present not understood .
increased fluorescence has often been observed and it has been suggested that this results from the photo - induced release of ca@xmath0 @xcite or from direct photo - activation of the fluorescent dye @xcite or effects on membrane permeability of sample heating @xcite , all of which can be expected at the short wavelengths applied in single - photon microscopy . the commoner problem of photo - bleaching due to the destruction of the dye by free - radicals generated by illumination @xcite is a serious limitation in all single - photon live cell imaging .
our present results suggest that widefield two - photon microscopy may overcome the bleaching problem at frame rates between 10100 hz .
to confirm the two - photon nature of the excitation process , we obtained images with frame rates of 1 hz , 10 hz , 50 hz and 100 hz at five time - averaged incident powers .
log - log plots of fluorescence intensity from five rois with excitation power gave gradients of between 1.8 and 2.3 for all frame rates , confirming two - photon excitation .
an example time course with three rois is shown in fig .
[ fig2 ] , and supplementary video 2 shows slow ( @xmath2seconds ) ca@xmath0 transients at a camera frame rate of 100 hz over a duration of 10 seconds .
for display purposes , the video file has been reduced in size by 50 % . in this video ,
three rois were chosen in three adjacent cell bodies to not only demonstrate the rate of image capture but also to confirm that the measured change in fluorescence signal intensity with time was not global across the image , but is localised to individual cell bodies at different times .
this evidence suggested that the change in measured fluorescence signal intensity was not a consequence of fluctuations in laser power or camera instability , but arose from localised changes in intracellular ca@xmath0 concentration .
having established that wide - field two - photon excitation reduces photo - bleaching at high frame rates , we then examined whether ca@xmath0 events induced by synaptic activity could be observed in primary hippocampal cultures .
figure [ fig3]a and supplementary video 3 ( left panel ) , acquired at a frame rate of 10 hz , show spontaneous changes in fluorescence intensity in live neuronal cell bodies over a 60-second time period , which were abolished in the presence of the ampa and nmda receptor antagonists nbqx ( 20 @xmath1 m ) and dl - ap5 ( 100 @xmath1 m ) ( fig .
[ fig3]b and right panel of supplementary video 3 ) , indicating that these events are driven by glutamatergic excitatory synaptic activity .
the sensitivity to nbqx / dl - ap5 is similar to that observed when using the whole - cell patch clamp technique to monitor synaptically driven events including spontaneous action potential firing using identical cultures ( fig . [ fig3]c and [ fig3]d ) @xcite thus highlighting the functional capability of this method .
additionally , in separate experiments , we show the consequence of reducing the intrinsic inhibitory tone present in the cultures through the application of the gaba@xmath3 receptor antagonist bicuculline ( 20 @xmath1 m ) .
figure [ fig3]e and supplementary video 4 , acquired at a frame rate of 50 hz , show that spontaneous changes in fluorescence intensity initially became more frequent and erratic in nature but at later time points were not present given the large rise in fluorescence within the neuronal cell bodies due to uncontrolled glutamatergic activity .
since the cell preparations used here were a thin monolayer of cells , we did not perform optical sectioning .
however , some optical sectioning of thicker specimens was possible . figure [ fig4 ] shows optical sectioning of an auto - fluorescent fixed _ taraxacum _ pollen specimen mounted in histomount , obtained at an excitation wavelength of @xmath4=820 nm and using the same chromatic reflectors and emission filters shown in fig .
[ fig5 ] . here
a 60x/1.35 na oil immersion objective was used , and the montage was obtained by moving the specimen by 1 @xmath1 m increments axially over a range of 25 @xmath1 m .
the frame rate was set to 10 hz for each image .
the spikes at the top and bottom of the pollen grain are clearly independently resolved without deconvolution or other image processing methods .
in evaluating this widefield two - photon microscopy method , it is natural to ask what advantages it confers in relation to conventional two - photon scanning microscopy .
the most important advantage of the method is the low photo - bleaching rate , compared with both two - photon scanning and single - photon widefield methods , when applied to the imaging of fast events .
this would appear to present a great experimental advantage .
the reason for this advantage is almost certainly that our widefield regime requires an excitation intensity that is about three orders of magnitude less than that used in point - scanning two - photon microscopy .
lowering the excitation intensity may bring a disproportionate benefit , since it is known that photo - bleaching has a dependency on intensity even higher than two - photon fluorescence @xcite .
we suggest that our method may prove advantageous when using fluorophores or photoproteins with fast photo - bleaching rates , including ca@xmath0-sensitive dyes , cfp @xcite , ebfp @xcite and yfp @xcite .
this technique may also reduce unwanted photo - bleaching in frap , flap and other similar experimental methods , where photo - bleaching of the untargeted region can compromise results @xcite .
we have used a single camera and fluorophore here , but our method could be easily adapted for simultaneous multi - channel recording using an image splitter and a second scmos camera . the advantage in time - resolution is plainly shown by the results at high frame rates .
the penetration ability of two - photon , as compared with single - photon excitation , which has been well established @xcite , has not been measured here , but is expected in this method also , because it depends on the lower scattering of longer wavelengths and the low absorption of the ultra - short pulsed infra - red excitation wavelengths .
the widefield two - photon method can not be expected to perform as well as conventional scanned two - photon excitation in optical sectioning ability . by imaging fluorescent test specimens using the 60x/1.35 na oil immersion lens at @xmath4=820 nm , we measured an axial point spread function of around 5 @xmath1 m and a measured lateral resolution of 850 nm .
these resolution values are lower than for a point - scanning two - photon microscope system with similar excitation parameters , but they are an expected consequence of weak focusing of the excitation beam .
nevertheless , the results from our live cell experiments and imaging of thicker specimens suggest that this method could be adapted for in vivo imaging , with the advantage of reduced photo - bleaching .
although no temporal focusing @xcite was used here , our widefield method could be used with temporal focusing to obtain improved optical sectioning , without the low image contrast and much reduced resolution observed in multi - focal two - photon microscopy @xcite .
ultra - short pulsed near - infrared lasers of the type used here have been previously used to generate ca@xmath0 waves in differentiated cells .
@xcite used a single diffraction - limited beam focus of @xmath5=0.3 @xmath1 m , with an average power @xmath6 > 20 mw , @xmath4=775 nm , repetition frequency @xmath7=82 mhz and pulse duration @xmath8=140 fs , which gives a peak intensity @xmath9 > 6.16 @xmath10 10@xmath11 w / m@xmath12 . however , since the peak intensity of illumination used in our two - photon widefield microscope is around three orders of magnitude less than smith _ et al . _ used in their experiments , it is highly unlikely that we are observing light - induced ca@xmath0 transients .
indeed , the sensitivity of the observed ca@xmath0 events to blockade of glutamatergic synaptic activity indicates that the events are synaptically driven and are presumably triggered by action potentials arriving at the neuronal cell body , the consequence of which is the transient increase in intracellular ca@xmath0 level due to membrane potential depolarisation .
the close similarity between the optical transients observed here and those recorded electrically in individual cells by the whole - cell path clamp technique in current clamp mode are a clear demonstration of the sensitivity , time resolution and usefulness of this method .
we have concentrated here on the improved time resolution and remarkably low level of photo - bleaching , but , as already pointed out by hwang _
_ , there may be other advantages over single - photon imaging , such as improved discrimination against auto - fluorescence and clearer imaging of dense tissues @xcite .
* excitation and detection of two - photon fluorescence . * awareness of the high peak intensities required for two - photon excitation has probably deterred experimentation on widefield two - photon microscopy .
here we compare excitation and detection parameters in the widefield two - photon microscope with that in a standard point - scanning two - photon microscope .
+ for the laser source , we assume a wavelength of @xmath4=780 nm , repetition frequency @xmath7=80 mhz and pulse duration @xmath8=140 fs with a 60x/0.9 na water dipping lens , as used in our live cell imaging experiments .
we also assume that pulse stretching is negligible , and is similar in both the laser scanning and widefield microscopes .
+ in a point scanning two - photon microscope , we can assume a time - averaged power at the specimen plane of @xmath6=15 mw . with a pulse duration @xmath8 and pulse repetition frequency @xmath7 , from @xmath13 a peak power of @xmath14=1.34 kw
is obtained .
+ the beam waist radius ( @xmath5 ) of the excitation laser at the specimen plane is given by @xcite @xmath15 where @xmath16 is the half - angle of the beam divergence given by the numerical aperture @xmath17 for the 60x/0.9 na water dipping lens ( @xmath18=1.33 ) , @xmath16=0.74 rad and thus at a wavelength of @xmath4=780 nm , the beam waist radius is calculated to be @xmath5=334 nm .
+ the peak intensity in the excitation spot is given by @xmath19 and from the calculated values of @xmath14 and @xmath5 , a peak intensity of @xmath9=3.82 @xmath10 10@xmath11 w / m@xmath12 is obtained for point - scanning . + in the widefield two - photon microscope , we use a time - averaged power of @xmath6=360 mw at the specimen plane which , from eq .
[ eq1 ] when considering the same pulse duration and pulse repetition frequency , gives a peak power of @xmath14=32 kw .
+ instead of focusing to a diffraction limited spot using the same 60x/0.9 na lens by overfilling the back aperture , we instead consider focusing to a small spot near the back aperture to obtain a weakly focusing spot at the specimen plane with a radius of @xmath5=45 @xmath1 m . using these values of @xmath14 and @xmath5 in eq .
[ eq4 ] above , the peak intensity in the widefield two - photon microscope is therefore @xmath9=5.03 @xmath10 10@xmath20 w / m@xmath12 .
we note that this peak intensity is approximately three orders of magnitude lower than the peak intensity for a point - scanning two - photon microscope .
+ the substantially lower peak excitation intensity in the widefield two - photon microscope is compensated by negating the need to scan the beam . by continuously irradiating across the whole field , we can use a much longer integration time for collection of fluorescence .
+ in a point - scanning two - photon microscope , we can assume a point dwell time of t=5 @xmath1s and , since the fluorescence lifetime of the fluorophore is very short ( in the order of ns ) , we can assume that fluorescence is generated and collected over the same 5 @xmath1s duration .
for an image that is 640 @xmath10 560 pixels , which is the same image size we use in the widefield setup , this point dwell time yields a frame rate of 0.56 hz .
the excitation radiation applied to the specimen for each point is therefore @xmath9 @xmath10 t=1.91 @xmath10 10@xmath21 w@xmath22s / m@xmath12 .
+ in the widefield two - photon microscope , the specimen is continuously irradiated with the femtosecond - pulsed laser radiation , but in comparison with the point - scanning two - photon microscope we can consider the dwell time to be equivalent to the frame rate of the camera . for images taken at 100 hz ( i.e. a t=10 ms exposure time ) , the excitation radiation applied to the specimen during the image capture can be considered to be @xmath9 @xmath10 t=5.03 @xmath10 10@xmath21 w@xmath22s / m@xmath12 . the overall total excitation radiation applied in the widefield two - photon microscope is therefore similar to that used in point - scanning two - photon microscopy .
+ in summary , when using widefield excitation , the peak intensity is orders of magnitude lower than in point - scanning excitation and the total excitation radiation applied is similar . by collecting the light on all the pixels in parallel we achieve long per pixel integration times , and can collect enough photons for imaging at fast frame rates .
moreover , the quantum efficiency of an scmos camera ( 60 % for the zyla camera ) is likely to be higher than the efficiency of a photomultiplier tube used in a point - scanning two - photon microscope ( @xmath230 % ) , therefore offering the possibility of high detection sensitivity . +
* experimental setup . *
our simple widefield two - photon microscope was configured as shown in fig .
[ fig5 ] . a commercial upright epi - fluorescence microscope ( bx51wi , olympus )
was modified for two - photon excitation . a chromatic reflector ( ff670-sdi01 , semrock ) reflecting wavelengths longer than 670 nm and transmitting shorter wavelengths was used to direct excitation towards the specimen and to transmit fluorescence .
a cooled scmos camera ( zyla 5.5 , andor ) controlled by a pc running the freely available software winfluor @xcite was used to detect fluorescence from the specimen .
for blocking of the laser source from the camera and to ensure that only the fluorescence signal contributed to the image , we used 680 nm and 694 nm short - wave pass filters ( ff01 - 680/sp and ff01 - 694/sp , semrock ) in the collection path .
a commercially available wavelength - tunable femtosecond - pulsed ti : sapphire laser ( chameleon ultra ii , coherent ) was used as the excitation source .
this laser delivered a maximum time - averaged power of 2.3 w at a repetition rate of 80 mhz with a pulse duration of 140 fs . the weakly divergent output from the ti :
sapphire laser was steered using a pair of highly reflecting mirrors ( bb1-e03 , thorlabs ) and then attenuated using a variable neutral density filter ( ndc-25c-4 , thorlabs ) and expanded to a diameter of 22 mm using a plano - convex lens pair of focal lengths f=+35 mm and f=+100 mm ( la1027-b and la1509-b , thorlabs ) .
the beam height was changed using a two - mirror periscope ( bb1-e03 , thorlabs ) to permit coupling of the laser into the upright microscope . at height
, the beam was focused using a single plano - convex lens of focal length f=+75 mm ( la1608-b , thorlabs ) , with the beam waist close to the back aperture of the objective lens .
this produced a wide and weakly - focused beam in the specimen plane which did not contribute significantly to the optical sectioning power .
a beam diameter was found experimentally which allowed excitation of the dye but with minimal photo - bleaching at a time - averaged laser power of 360 mw at a wavelength of 780 nm .
the illuminated area then had a diameter of 90 @xmath1 m and a 1.6@xmath10 magnifier was used to fill the camera sensor .
all live cell imaging data presented here used a 60x/0.9 na water dipping lens ( olympus ) . + for single - photon widefield imaging experiments , the short - wave pass filters were removed and the long - wavelength chromatic reflector was replaced with a filter cube ( xf22 , omega optical , inc . ) that comprised an excitation filter with a peak wavelength of 485 @xmath23 22 nm , a chromatic reflector with a 505 nm long - wave pass filter and an emission filter with a peak wavelength of 530 @xmath23 30 nm . the 75 mm plano - convex tube lens was removed , and the mercury arc lamp excitation source ( u - uls100hg , olympus ) and coupling optics originally supplied with the microscope were reattached . at a wavelength of 485 nm ,
the maximum time - averaged power at the specimen plane exiting the 60x/0.9 na water dipping objective lens was 240 @xmath1w .
the same camera was used with the same pc and controlling software as with widefield two - photon excitation . + * imaging primary rat hippocampal cultures . * we performed widefield two - photon microscopy of primary rat hippocampal cultures to demonstrate the capability of rapid imaging .
primary rat hippocampal cultures were prepared as described in ref .
briefly , sprague - dawley rat pups 12 days old were killed by cervical dislocation and decapitation , following uk home office guidelines , and the brain removed .
the hippocampi were then dissected out , incubated in a papain solution ( 1.5 mg / ml , sigma - aldrich ) at 37 @xmath24c for 20 minutes .
the hippocampi were then washed in solution containing bovine serum albumin ( 10 mg / ml ) , dissociated by trituration and plated onto coverslips previously coated with poly - l - lysine ( 0.1 mg / ml ) at a final density of 3 @xmath10 10@xmath25 cells / ml .
cultures were incubated in neurobasal - a medium ( invitrogen ) supplemented with 2 % ( v / v ) b-27 ( invitrogen ) and 2 mm l - glutamine and maintained in a humidified atmosphere at 37 @xmath24c/5 % co@xmath26 .
after 5 days _ in vitro _
( div ) , cytosine - d - arabinofuranoside ( 10 @xmath1 m ) was added to inhibit glial cell proliferation .
cells were used experimentally from 11 - 14 div .
+ for imaging , hippocampal cultures were washed twice with a hepes - buffered saline ( hbs ) containing ( in mm ) : nacl 140 , kcl 5 , mgcl@xmath26 2 , hepes 10 ,
d - glucose 10 and cacl@xmath26 2 , ph 7.4 , and transferred to hbs containing fluo-4 am ( 10 @xmath1 m , 4560 min , room temperature ) . once loaded , cells were washed with hbs and images were obtained from the cell bodies of neurones with constant irradiation and with frame rates of up to 100 hz .
pixel binning was used to increase the signal to noise ratio of the images : four pixel binning was used at frame rates of 150 hz , and eight pixel binning was used for camera frame rates of 100 hz .
experiments were performed on cultures at room temperature .
cells were identied as the cell bodies of neurones based on their morphological characteristics .
data were calculated as changes in fluorescence ratio from the fluorescence signal intensity at time t=0 s. + we also compared the measured fluorescence signal intensity over time in the live cell neuronal preparations using both single - photon and two - photon widefield excitation to investigate photo - bleaching at different frame rates . for these measurements , we used time - averaged powers of @xmath250 @xmath1w for single - photon and @xmath2100 mw for two - photon excitation at 1 hz , @xmath2400 @xmath1w and @xmath2170 mw at 10 hz , and @xmath22.5 mw and 360 mw at 100 hz .
these time - averaged powers were chosen to facilitate the best comparison of single - photon and two - photon excitation .
similar initial fluorescence signal intensity counts were obtained by adjusting the excitation radiation in one region of the specimen then moving to an adjacent region for recording .
it is perhaps helpful to note here that this setup procedure was possible by eye , since the two - photon excitation of fluorescence could be observed using the binocular viewer of the microscope . to our knowledge , this is the first observation of this kind .
+ to determine whether the transient changes in fluorescence signal intensity over time were caused by synaptically - driven events , nbqx ( 20 @xmath1 m , abcam ) and dl - ap5 ( 100 @xmath1 m , abcam ) , ampa receptor and nmda receptor antagonists respectively , were applied in order to block excitatory synaptic activity .
additionally in separate experiments , the gaba@xmath3 receptor antagonist bicuculline methiodide ( 20 @xmath1 m , abcam ) , was applied to determine the consequence of reducing the intrinsic inhibitory tone present in the cultures .
data were recorded using widefield two - photon excitation ( @xmath4=780 nm , @xmath6=180 mw and a frame rate of 10 hz for nbqx / dl - ap5 , @xmath6=250 mw and a frame rate of 50 hz for bicuculline ) before either the nbqx / dl - ap5 or bicuculline were added to the bath .
the excitation source was then blocked to prevent irradiation of the specimen for 10 minutes , and then further data were recorded using the same excitation and detection parameters . + this research was supported in part by the medical research council ( grant no .
mr / k015583/1 ) and the engineering and physical sciences research council ( grant no .
ep / i006826/1 ) .
is supported by a scottish universities physics alliance inspire studentship .
, t.b . and g.m .
designed the study .
j.d . developed the image acquisition and analysis program winfluor .
performed calculations .
l.w . and n.z.a.r .
prepared primary rat hippocampal cultures . w.b.a . and g.m .
prepared the pollen specimen .
, g.r . and g.m .
performed the experiments .
, w.b.a . and g.m .
analyzed the data and wrote the manuscript .
the authors declare no competing financial interests .
10 denk , w. , strickler , j.h . & webb , w.w .
two - photon laser scanning fluorescence microscopy .
_ science _ * 248 , * 7376 ( 1990 ) .
patterson , g.h .
& piston , d.w .
photobleaching in two - photon excitation microscopy . _
j. _ * 78 , * 21592162 ( 2000 ) .
nguyen , q - t . , callamaras , n. , hsieh , c. & parker , i. construction of a two - photon microscope for video - rate ca@xmath0 imaging . _ cell calcium _ * 30 , * 383393 ( 2001 ) .
brakenhoff , g.j .
_ real - time two - photon confocal microscopy using a femtosecond , amplified ti : sapphire system .
_ j. microsc . _ * 181 , * 253259 ( 1996 ) .
fittinghoff , d. , wiseman , p. & squier , j. widefield multiphoton and temporally decorrelated multifocal multiphoton microscopy . _
express _ * 7 , * 273279 ( 2000 ) .
niesner , r. , andresen , v. , neumann , j. spiecker , h. & gunzer , m. the power of single and multibeam two - photon microscopy for high - resolution and high - speed deep tissue and intravital imaging . _
j. _ * 93 , * 25192529 ( 2007 ) .
deniset - besseau , a. , lvque - fort , s. , fontaine - aupart , m.p . ,
roger , g. & georges , p. three - dimensional time - resolved fluorescence imaging by multifocal multiphoton microscopy for a photosensitizer study in living cells .
opt . _ * 46 , * 80458051 ( 2007 ) .
truong , t.v . , supatto , w. , koos , d.s . ,
choi , j.m .
& fraser , s.e .
deep and fast live imaging with two - photon scanned light - sheet microscopy .
methods _ * 8 , * 757760 ( 2011 )
. hwang , j.y .
_ et al . _
multimodal wide - field two - photon excitation imaging : characterization of the technique for in vivo applications .
express _ * 2 , * 356364 ( 2011 ) .
schneider , c.a . ,
rasband , w.s . & eliceiri , k.w .
nih image to imagej : 25 years of image analysis .
. methods _ * 9 , * 671675 ( 2012 )
. knight , m.m . ,
roberts , s.r . ,
lee , d.a . & bader , d.l .
live cell imaging using confocal microscopy induces intracellular calcium transients and cell death .
j. physiol .
cell physiol . _
* 284 , * c1083c1089 ( 2003 ) .
mcdonald , a. , harris , j. , macmillan , d. , dempster , j. & mcconnell , g. light - induced ca@xmath0 transients observed in widefield epi - fluorescence microscopy of excitable cells .
express _ * 3 , * 12661273 ( 2012 ) .
betzig , e. _ et al . _ imaging intracellular fluorescent proteins at nanometer resolution .
_ science _ * 313 , * 16421645 ( 2006 ) .
mitchison , t.j . ,
sawin , k.e . ,
theriot , j.a . , gee , k. & mallavarapu a. caged fluorescent probes .
_ methods enzymol . _
* 291 , * 6378 ( 1998 ) .
schnle , a. & hell , s.w .
heating by absorption in the focus of an objective lens . _
* 23 , * 325327 ( 1998 ) .
tsien , r.y . ,
ernst , l. & waggoner , a. [ fluorophores for confocal microscopy : photophysics and photochemistry ] _ handbook of biological confocal microscopy _ [ pawley , j. ( ed . ) ] [ 338352 ] ( springer , new york , 2006 ) .
gan , j. , greenwood , s.m . ,
& bushell , t.j . indirect modulation of neuronal excitability and synaptic transmission in the hippocampus by activation of proteinase - activated receptor-2 . _
j. pharmacol . _
* 163 , * 984994 ( 2011 ) .
ledgerwood , c.j . ,
greenwood , s.m . ,
brett , r.r . ,
pratt , j.a .
& bushell , t.j .
cannabidiol inhibits synaptic transmission in rat hippocampal cultures and slices via multiple receptor pathways . _
j. pharmacol . _
* 162 , * 286294 ( 2011 ) .
patterson , g.h . ,
knobel , s.m . ,
sharif , w.d .
, kain , s.r . & piston , d.w .
use of the green fluorescent protein and its mutants in quantitative fluorescence microscopy .
_ biophys .
j. _ * 73 , * 27822790 ( 1997 ) .
malkani , n. & schmid , j.a .
some secrets of fluorescent proteins : distinct bleaching in various mounting fluids and photoactivation of cyan fluorescent proteins at yfp - excitation .
_ plos one _ * 6 , * e18586 ( 2011 ) .
zaccolo , m. _ et al . _ a genetically encoded , fluorescent indicator for cyclic amp in living cells .
cell biol . _
* 2 , * 2529 ( 2000 ) .
mcananey , t.b .
_ et al . _
protonation , photobleaching , and photoactivation of yellow fluorescent protein ( yfp 10c ) : a unifying mechanism .
_ biochemistry _ * 44 , * 55105524 ( 2005 )
. weiss , m. challenges and artifacts in quantitative photobleaching experiments . _ traffic _ * 5 , * 662671 ( 2004 ) .
dunn , g.a . ,
dobbie , i.m . , monypenny , j. , holt , m.r . &
zicha , d. fluorescence localization after photobleaching ( flap ) : a new method for studying protein dynamics in living cells . _
j. microsc . _
* 205 , * 109112 ( 2002 ) .
centonze , v.e . & white , j.g .
multiphoton excitation provides optical sections from deeper within scattering specimens than confocal imaging .
_ biophys .
j. _ * 75 , * 20152024 ( 1998 ) .
oron , d. , tal , e. & silberberg , y. scanningless depth - resolved microscopy .
express _ * 13 , * 14681476 ( 2005 ) .
bewersdorf , j. , pick , r. & hell , s.w .
multifocal multiphoton microscopy .
* 23 , * 655657 ( 1998 ) .
smith , n.i .
_ et al . _
photostimulation of two types of ca@xmath0 waves in rat pheochromocytoma pc12 cells by ultra - short pulsed near - infrared laser irradiation .
_ laser phys . _ * 3 , * 154161 ( 2006 ) .
kogelnik , h. & li , t. laser beams and resonators .
. opt . _ * 5 , * 15501567 ( 1966 ) .
dempster , j. winfluor : fluorescence imaging & electrophysiology analysis program ( 2014 ) .
available at www.strath.ac.uk .
date of access : july 2 , 2014 .
m. ] m increments axially over a range of 25 @xmath1 m .
no post - processing was performed on the images except for cropping to display only a single pollen grain within the image field .
the optical sectioning shown here is closely similar to that obtained in a widefield single - photon fluorescence microscope ( results not shown ) .
this confirms that the optical depth of field is due to the focusing of the emission only .
scale bar=15 @xmath1 m . ] | we demonstrate fluorescence imaging by two - photon excitation without scanning in biological specimens as previously described by hwang and co - workers , but with an increased field size and with framing rates of up to 100 hz . during recordings of synaptically - driven ca@xmath0 events in primary rat neurone cultures loaded with the fluorescent ca@xmath0 indicator
fluo-4 am , we have observed greatly reduced photo - bleaching in comparison with single - photon excitation .
this method , which requires no costly additions to the microscope , promises to be useful for work where high time - resolution is required . immediately after its introduction by denk , strickler and webb in 1990 @xcite , two - photon excitation of fluorescence
was adopted in many fields of biomedicine .
its chief advantages are its ability to penetrate more deeply into tissues than single - photon excitation , the creation of optical sections ( by the combination of an excitation proportional to the square of the intensity with the conical beam geometry ) and the more efficient utilization of scattered emission than is possible in a confocal microscope .
the chief drawbacks were the slow scanning speed , which restricted the original instruments to a rate of approximately one image per second , often unusable for electrophysiology , where millisecond time resolution may be needed , and the high rate of photo - bleaching if the laser intensity was increased @xcite . to overcome the slowness ,
faster scanning mirrors have been used @xcite , and parallelism has been achieved in two - photon imaging by the use of slit scanning @xcite or the scanning of multiple foci , preferably uncorrelated in time @xcite .
two - photon light - sheet microscopy has also been used to increase the rate of imaging @xcite , but because scattering and absorption in the specimen causes inhomogeneous illumination , this method is best suited to highly transparent tissue volumes . a simple and elegant approach of illumination with a large stationary spot of light from a mode - locked femtosecond - pulsed laser in a conventional microscope and imaging of the emission from structures within the spot , though not favoured by fittinghoff _
et al . _ who were apparently the first to try it @xcite , was shown by hwang _
et al .
_ to provide images of good quality @xcite .
we show how this unexpected result can be explained by the vastly - increased exposure time of the simpler method . the work of hwang _
et al .
_ on several medically - important specimens seems to have only a limited effect on normal practice , possibly because the illuminated spot was no greater than 60 @xmath1 m in diameter and the loss of optical sectioning has been regarded as serious .
we have repeated and extended this approach , increasing the field size to a diameter of 90 @xmath1 m and concentrating on fast transients in living neurones , a specimen where the potentially unlimited time resolution of the stationary beam can be evaluated . in this work ,
we have used modest average powers from an 80 mhz repetition rate femtosecond - pulsed ti : sapphire laser and have collected images using a sensitive scmos camera at frame rates up to 100 hz .
all components of our simple microscope setup are commercially available .
we have explored the image quality , phototoxicity and bleach rate at different framing rates , making extended video recordings of synaptically - driven ca@xmath0 events in live neurones loaded with fluo-4 am , which we have chosen because they show both slow and fast activity .
our chief new finding is that photo - bleaching is greatly reduced in the widefield two - photon regime relative to that seen with single - photon excitation .
we have shown that this unexpected advantage relative to normal practice is particularly marked at high framing rates . |
in @xcite , dijkstra presents the following mutual exclusion protocol for a ring of nodes @xmath0 where each node can read the state @xmath1 \in \set{0}{k-1}$ ] of its anti - clockwise neighbour , and where node @xmath2 runs a different program than the other nodes .
node @xmath2 , privileged when @xmath3=x[n]$ ] @xmath3=x[n]$ ] @xmath3 \assign ( x[0 ] + 1 ) \bmod k$ ] node @xmath4 , @xmath5 , privileged when @xmath6 \neq x[i-1]$ ]
@xmath6 \neq x[i-1]$ ] @xmath6 \assign x[i-1]$ ] dijkstra proves self - stabilization of this protocol to a configuration where only one node is privileged at a time , for @xmath7 under a central daemon and says @xcite : `` for smaller values of @xmath8 , counter examples kill the assumption of self - stabilization '' .
failing to find a counter example for @xmath9 , we instead found the following proof that the system also stabilizes when @xmath9 , provided that @xmath10 .
we first define the legitimate configurations as those configurations that satisfy @xmath6=a$ ] for all @xmath4 with @xmath11 and @xmath6=(a-1 ) \bmod k$ ] for all @xmath4 with @xmath12 for some choice of @xmath13 and @xmath14 .
hence the configuration where all nodes have the same state is legitimate .
dijkstra already showed ( independent of any restriction on @xmath8 ) closure of the legitimate states , that no run of the protocol ever terminates , and that in each of these runs the exceptional node will change state ( aka `` fire '' ) infinitely often .
let @xmath10 .
consider the case where node @xmath2 fires for the first time .
then just before that , @xmath3=x[n]=b$ ] for some @xmath15 and the new value of @xmath3 $ ] becomes @xmath16 .
now consider the case when node @xmath2 fires again .
then just before that , @xmath3=x[n]=b+1 $ ] . in order for node @xmath17 to change value from @xmath15 to @xmath16 ,
it must have copied @xmath16 from its anti - clockwise neighbour @xmath18 $ ] ( which exists if @xmath10 ) .
this moment must have occurred after node @xmath2 changed state to @xmath3=b+1 $ ] .
but then , just after node @xmath17 copies @xmath16 from node @xmath19 we actually have @xmath18=x[n]=x[0]=b+1 $ ] . in other words , if @xmath10 , three different nodes hold the same value @xmath16 .
then the remaining @xmath20 nodes can each take a different value from the remaining @xmath21 values ( unequal to @xmath16 ) , which means that if @xmath22 ( so in particular when @xmath9 ) at this point in time there is a value @xmath13 ( among these @xmath21 values ) not occurring as the state of any node on the ring . because node @xmath2 fires infinitely often , eventually @xmath3 $ ] becomes @xmath13 . because the other nodes merely copy values from their anti - clockwise neighbours , at this point no other node holds @xmath13 .
the next time node @xmath2 fires , @xmath23=x[0]=a$ ] .
the only way that node @xmath17 gets the value @xmath13 is if all intermediate nodes have copied @xmath13 from node @xmath2 .
we conclude that for nodes , @xmath6=a$ ] , which is a legitimate state . | we show that , contrary to common belief , dijkstra s self - stabilizing mutual exclusion algorithm on a ring @xcite also stabilizes when the number of states per node is one less than the number of nodes on the ring .
* keywords * : distributed computing , fault tolerance , self - stabilization . |
to begin with , let us consider the following cauchy problem @xmath6 associated with the fractional schrdinger equation @xmath7 where @xmath8 is a potential .
this equation has recently attracted interest from mathematical physics .
this is because fractional quantum mechanics introduced by laskin @xcite is governed by the equation where it is conjectured that physical realizations may be limited to the cases of @xmath1 .
of course , the case @xmath9 corresponds to the ordinary quantum mechanics . by duhamel
s principle , the solution of is given by @xmath10 where the propagator @xmath11 is given by means of the fourier transform , as follows : @xmath12 then the standard approach to the problem is to obtain the corresponding strichartz estimates which control space - time integrability of in view of that of the initial datum @xmath13 and the forcing term @xmath14 . in the classical case @xmath9 ,
the strichartz estimates originated by strichartz @xcite have been extensively studied by many authors ( @xcite ) . over the past several years , considerable attention has been paid to the fractional order where @xmath1 in the radial case ( see @xcite and references therein ) . from these works , when @xmath15 , the homogeneous strichartz estimate @xmath16 holds for radial functions @xmath17 if @xmath18 here the condition is optimal if @xmath19 .
but when @xmath20 , is unknown for the endpoint @xmath21 .
also , it is known that the estimate does not hold in general if @xmath13 does not have radial symmetry .
now , by duality and the christ - kiselev lemma ( @xcite ) , one may use with @xmath22 to get some inhomogeneous estimates @xmath23 for @xmath24 and @xmath25 which satisfy with @xmath22 and @xmath26 .
this means that @xmath24 and @xmath25 lie on the segment @xmath27 in figure [ figure ] .
however , these trivial estimates are not enough to imply the well - posedness for the equation with the initial data @xmath17 beyond the case @xmath22 . when @xmath28 we need to obtain on a wider range of @xmath24 and @xmath25 .
see section [ sec2 ] for details .
let us first mention the following necessary conditions for : @xmath29 and @xmath30 the first is just the scaling condition and the second will be shown in section [ sec4 ] .
our main result in this paper is the following theorem where we obtain on the open segment @xmath31 in figure [ figure ] .
[ prop ] let @xmath32 and @xmath15 .
assume that @xmath33 is a radial function with respect to the spatial variable @xmath34 .
then we have @xmath35 if @xmath36 [ rem ] it should be noted that the range is sharp .
namely , the second condition in is the scaling condition for ( see ) , and the first one is the necessary condition when @xmath37 and @xmath38 .
[ figure ] . here
the open segment @xmath39 is the range for , and @xmath40 $ ] is the range for .,title="fig:",height=226 ] from interpolation between and , we can directly obtain further estimates when @xmath24 and @xmath25 are contained in the open quadrangle with vertices @xmath41 .
precisely , we have the following corollary .
[ mixedprop ] let @xmath32 and @xmath15 .
assume that @xmath33 is a radial function with respect to the spatial variable @xmath34 , and that @xmath24 and @xmath25 satisfy the necessary conditions and .
then we have @xmath42 if the following conditions hold : * for @xmath24 , @xmath43 and @xmath44 * similarly for @xmath25 .
let us give more details about the conditions in the above corollary .
the line @xmath45 in figure [ fig1 ] is when the equality holds in the first inequality of .
similarly , the lines @xmath46 and @xmath47 correspond to the second inequality in and the inequality , respectively . finally , the line @xmath48 is sharp because it is determined from the necessary condition .
now we apply the above strichartz estimates to the well - posedness theory for the fractional schrdinger equation in the radial case : @xmath49 where we assume that @xmath50 , @xmath13 and @xmath51 are radial functions with respect to the spatial variable @xmath34 .
we obtain the following well - posedness for with @xmath2 initial data @xmath13 below @xmath3 and potentials @xmath4 under the scaling - critical range @xmath5 .
the cauchy problem was studied in @xcite particularly when @xmath9 and @xmath22 .
[ thm ] let @xmath52 and @xmath53 for @xmath32 .
assume that @xmath17 and @xmath54;l_x^w(\mathbb{r}^n))$ ] for some @xmath55 .
then there exists a unique solution @xmath56;\dot{h}^\gamma(\mathbb{r}^n))\cap l_t^q([0,t];l_x^p(\mathbb{r}^n))$ ] of if @xmath57 @xmath58 and @xmath59 the condition @xmath5 on the potential is critical in the sense of scaling .
indeed , @xmath60 takes into @xmath61 with @xmath62 . hence the norm
@xmath63 is independent of @xmath64 precisely when @xmath5 . in proposition 3.9 of @xcite ,
the inhomogeneous estimates were shown in certain region that lies below the segment @xmath65 in figure [ figure ] .
( note that @xmath66 if and only if @xmath67 and @xmath68 . also , the point @xmath69 is the same as @xmath70 when @xmath20 . ) by interpolation between these and our estimates ,
we can also obtain further estimates in the triangle with vertices @xmath71 .
we omit the details since it does not affect the range @xmath53 of @xmath72 in theorem [ thm ] ( see section [ sec2 ] ) .
the rest of the paper is organized as follows . in section [ sec2 ] ,
we prove theorem [ thm ] by making use of the strichartz estimates and . section [ sec3 ] is devoted to proving theorem [ prop ] , and in section [ sec4 ] we show the necessary condition . in the final section , section [ sec5 ]
, we show lemma [ bessel ] which gives some estimates for bessel functions and is used for the proof of theorem [ prop ] . throughout the paper
, we shall use the letter @xmath73 to denote positive constants which may be different at each occurrence .
we also use the symbol @xmath74 to denote the fourier transform of @xmath13 , and denote @xmath75 and @xmath76 to mean @xmath77 and @xmath78 , respectively , with unspecified constants @xmath79 .
in this section we prove theorem [ thm ] . the proof is quite standard but we need to observe that if @xmath24 and @xmath25 satisfy the inhomogeneous estimate , then the midpoint of them lies on the segment @xmath27 .
note that @xmath80 if and only if @xmath67 and @xmath81 .
hence , if @xmath82 for @xmath83 , then @xmath25 should satisfy @xmath84 to give . in
what follows , it will be convenient to keep in mind this key observation . by duhamel
s principle , the solution of is given by @xmath85 then the standard fixed - point argument is to choose the solution space on which @xmath86 is a contraction mapping .
the strichartz estimates play a central role in this step . indeed , by the estimates and , we see that @xmath87;l_x^p)}\leq c\|f\|_{\dot{h}^\gamma}+c\|vu\|_{l^{\widetilde{q}'}_t([0,t ] ; l^{\widetilde{p}'}_x)}\ ] ] if @xmath88 @xmath89 @xmath90 and @xmath91 here , the conditions and are given from that the line @xmath92 lies in the closed triangle with vertices @xmath93 except the closed segments @xmath94,[c , d]$ ] .
note that @xmath95 when this line passes through the point @xmath73 .
similarly , the condition is given from that the line @xmath96 lies in the closed triangle with vertices @xmath97 except the closed segments @xmath98,[b , d]$ ] . by hlder s inequality , we now get @xmath99;l_x^p ) } \leq c\|f\|_{\dot{h}^\gamma}+c\|v\|_{l^r_t([0,t];l^w_x ) } \|u\|_{l^q_t([0,t];l^p_x)}\ ] ] if @xmath5 and the condition holds . indeed , when applying hlder s inequality to the second term in the right - hand side of , the conditions @xmath5 and follow from and , respectively .
from the above argument and the linearity , it follows that @xmath100;l_x^p ) } & \leq c\|v\|_{l^r_t([0,t];l^w_x)}\|u - v\|_{l^q_t([0,t];l^p_x)}\\ & \leq\frac12\|u - v\|_{l^q_t([0,t];l^p_x)},\end{aligned}\ ] ] which says that @xmath86 is a contraction mapping , if @xmath101 is sufficiently small . but here , since the above process works also on time - translated small intervals if @xmath102 for all @xmath103 , the smallness assumption on @xmath101 can be removed by iterating the process a finite number of times . for this
we will show that @xmath104 from , we first see that @xmath105 since @xmath106 is an isometry in @xmath3 , the first term in the right - hand side is clearly bounded by @xmath107 . on the other hand , by the inhomogeneous estimate the second term
is bounded by @xmath108 , where @xmath109 and @xmath110 .
here we use the sobolev embedding @xmath111 where @xmath112 with @xmath113 and @xmath114 , and hlder s inequality to get @xmath115 the required conditions here are summarized as follows : @xmath116 @xmath117 @xmath118 but , the inequalities @xmath109 and @xmath114 are satisfied automatically from the conditions on @xmath119 in theorem [ thm ] . on the other hand ,
the inequality @xmath120 is redundant because @xmath121 .
the remaining four equalities is reduced to the following one equality @xmath122 which is clearly satisfied from the condition .
consequently , we get .
in this section we prove theorem [ prop ] .
let us first consider the multiplier operators @xmath123 for @xmath124 defined by @xmath125 where @xmath126 $ ] is a smooth cut - off function which is supported in @xmath127 and satisfies @xmath128 .
then we will obtain the following frequency localized estimates ( proposition [ prop2 ] ) which imply theorem [ prop ] .
[ prop2 ] let @xmath32 and @xmath15 .
assume that @xmath33 is a radial function with respect to the spatial variable @xmath34 .
then we have @xmath129 uniformly in @xmath124 if @xmath130 indeed , since @xmath131 from the first condition in , by the littlewood - paley theorem and minkowski integral inequality , one can see that @xmath132 now , by , the right - hand side in the above is bounded by @xmath133 since @xmath134 , using the minkowski integral inequality and littlewood - paley theorem , this is bounded by @xmath135 . by this boundedness and @xmath136 , one may now use the christ - kiselev lemma ( @xcite ) to obtain @xmath137 as desired .
now it remains to prove the above proposition .
since we are assuming the scaling condition in , by rescaling @xmath138 , we may show only for @xmath139 .
let us first consider @xmath140 , @xmath141 and @xmath142 for @xmath143 , where @xmath144 , @xmath145 and @xmath146 . then by using the fact
( see @xcite , p. 347 ) that @xmath147 where @xmath148 denotes the bessel function with order @xmath149 , it is easy to see that @xmath150 d\lambda ds .
\end{aligned}\ ] ] now we define the operators @xmath151 , @xmath152 , as @xmath153 and for @xmath154 @xmath155 where @xmath156 denotes the characteristic function of a set @xmath70 and @xmath157 .
then the adjoint operator @xmath158 of @xmath159 is given by @xmath160 and for @xmath161 @xmath162 hence , since @xmath163 is independent of @xmath164 , by setting @xmath165 , it follows from that @xmath166 now we are reduced to showing that @xmath167 where we denote by @xmath168 the space @xmath169 . for lemma [ local_prop ] . here the open segment @xmath170 is the range for proposition [ prop2 ] ( see ) .,height=226 ] from now on , we will show by making use of the following lemma which will be obtained in subsection [ subsec3.2 ] .
[ local_prop ] let @xmath32 and @xmath15 .
then we have for @xmath171 @xmath172 if @xmath173 ( see figure [ fig2 ] ) .
* the case @xmath174 .
* we first decompose the sum over @xmath175 into two parts , @xmath176 and @xmath177 : @xmath178 when @xmath176 , using lemma [ local_prop ] , we then have @xmath179 note here that the first condition in implies @xmath180 from this , it follows that @xmath181 on the other hand , the second condition in implies @xmath182 since @xmath183 . consequently , we get @xmath184 as desired .
the other part where @xmath177 follows clearly from the same argument . * the case @xmath185 . *
the previous argument is no longer available in this case , since the left - hand side in becomes zero .
but here we deduce from bilinear interpolation between bilinear form estimates which follow from lemma [ local_prop ] .
this enables us to gain some summability as before .
let us first define the bilinear operators @xmath186 by @xmath187 where @xmath188 denotes the usual inner product on the space @xmath189 . then it is enough to show the following bilinear form estimate @xmath190 in fact , from we get @xmath191 as desired . for
, we first decompose the sum over @xmath175 into two parts , @xmath192 and @xmath193 : @xmath194 then we will use the following estimate which follows from hlder s inequality and lemma [ local_prop ] : @xmath195 for @xmath173 . by using this and , the first part where @xmath192 is now bounded as follows : @xmath196 for @xmath197 . if one applies this bound directly for @xmath198 satisfying the conditions in proposition [ prop2 ] as in the previous case , then one can not sum over @xmath199 because @xmath200 when @xmath185 .
but here we will make use of the following bilinear interpolation lemma ( see @xcite , section 3.13 , exercise 5(b ) ) together with to give @xmath201 [ bilinear_lem ] for @xmath202 , let @xmath203 be banach spaces and let @xmath101 be a bilinear operator such that @xmath204 then one has , for @xmath205 and @xmath206 , @xmath207 here , @xmath208 and @xmath209 .
indeed , let us first consider the vector - valued bilinear operator @xmath101 defined by @xmath210 where @xmath211 .
then , is equivalent to @xmath212 where @xmath213 , @xmath214 , @xmath215 , denotes the weighted sequence space equipped with the norm @xmath216
now , by we see that @xmath217 where @xmath197 and @xmath218 also , for @xmath219 satisfying , we can take a sufficiently small @xmath220 such that the ball @xmath221 with center @xmath222 and radius @xmath223 is contained in the region of @xmath222 given by @xmath197 ( see figure [ fig2 ] ) .
now , we choose @xmath224 such that @xmath225 then it is easy to check that @xmath226 and we get from the following three bounds @xmath227 then , by applying lemma [ bilinear_lem ] with @xmath228 and @xmath229 , we get @xmath230 since @xmath231 , by applying the real interpolation space identities in the following lemma , one can easily deduce from the above boundedness .
let @xmath232 and @xmath233 .
then one has @xmath234 if @xmath235 and @xmath236 , and if @xmath237 @xmath238 it is clear from the same argument that the second part where @xmath193 is bounded as follows : @xmath239 consequently , we have the desired estimate .
it remains to prove lemma [ local_prop ] .
we have to prove the following estimate : for @xmath240 , @xmath241 if @xmath173 . the proof is divided into the case @xmath242 and the cases where @xmath243 or @xmath139 . * the case @xmath244 . *
first we claim that for @xmath173 , @xmath245 indeed , we note that for @xmath246 and @xmath247 , @xmath248 which follows immediately from applying proposition 3.1 in @xcite with @xmath249 , @xmath250 , and @xmath251 . then by the usual @xmath252 argument , it is not difficult to see that implies @xmath253 for @xmath242 and @xmath173 .
replacing @xmath254 with @xmath255 , this gives .
when @xmath256 , follows now directly from .
so we are reduced to showing when @xmath257 .
for this we will obtain @xmath258 when @xmath257 . by interpolation between the estimates and
, we then get when @xmath257 .
indeed , when @xmath259 , by interpolation between with @xmath260 and , it is easy to check that holds for @xmath261 and @xmath262 .
this range of @xmath198 is wider than what is given by @xmath173 .
when @xmath263 , one can similarly get by interpolation between with @xmath264 and . now it remains to show the estimate . from and ,
we first write @xmath265 where @xmath266 is given as @xmath267 then , would follow from the uniform bound @xmath268 to show this bound , we will divide @xmath269 into four parts based on the following estimates for bessel functions @xmath270 .
[ bessel ] for @xmath271 and @xmath272 , @xmath273 where @xmath274 and @xmath275 assuming this lemma which will be shown in section [ sec5 ] , we see that @xmath276 and @xmath277 where the letter @xmath278 stands for constants different at each occurrence and depending only on @xmath279 .
now we write @xmath280 where @xmath281 then , @xmath269 is divided as @xmath282 , where @xmath283 first , it follows easily from that @xmath284 next , we shall consider @xmath285 .
since the factors @xmath286 and @xmath287 in @xmath288 would give a better boundedness than @xmath289 and @xmath290 , respectively , we only need to show the bound for @xmath291 let us now decompose @xmath292 as @xmath293 where @xmath294 . when @xmath295 , by the van der corput lemma ( see @xcite , chap .
viii ) it follows that @xmath296 hence , we get @xmath297 for the second part where @xmath298 or @xmath299 , we first note that @xmath300_{\rho=1/2}^{\rho=2}\\ & \qquad\qquad-\int \frac{1}{i(\pm\lambda \pm r+ \alpha t\rho^{\alpha-1 } ) } e^{\pm i\lambda\rho \pm ir\rho + it\rho^\alpha } \frac{d}{d\rho } \big ( \rho^{-1}\varphi^2 ( \rho ) \big ) d\rho\end{aligned}\ ] ] by integration by parts .
since we are handling the case where @xmath257 , we see that @xmath301 when @xmath298 or @xmath299 .
hence , using this , we get @xmath302_{\rho=1/2}^{\rho=2 } \big|\\ & \quad + \int\big|\frac{1}{(\pm\lambda \pm r+\alpha t\rho^{\alpha-1})}\frac{d}{d\rho } \big(\rho^{-1}\varphi^2 ( \rho ) \big ) \big| d\rho\\ & \lesssim2^{-m(j , k)}. \end{aligned}\ ] ] this implies @xmath303 it remains to bound @xmath304 and @xmath305
. we shall show the bound only for @xmath304 because the same type of argument used for @xmath304 works clearly on @xmath305 .
since the factor @xmath286 in @xmath306 would give a better boundedness than @xmath289 , we only need to show the bound for @xmath307 let us now decompose @xmath308 as @xmath309 where @xmath294 .
when @xmath295 , by the van der corput lemma as before , it follows that @xmath310}\big\{e_{\frac{n-2}{2}}(r\rho)\rho^{-\frac12}\varphi^2(\rho ) , \frac{d}{d\rho}\big(e_{\frac{n-2}{2}}(r\rho)\rho^{-\frac12}\varphi^2(\rho)\big)\big\}.\end{aligned}\ ] ] by and in lemma [ bessel ] , we see @xmath311 for @xmath312 .
thus , we get @xmath313 for the second part where @xmath298 or @xmath299 , we will use the following trivial bound when @xmath314 and @xmath315 : @xmath316 which follows from . on the other hand , when @xmath317 and @xmath315 , we will also show @xmath318 indeed , by integration by parts we see that @xmath319_{\rho=1/2}^{\rho=2}\\ & \qquad\qquad-\int \frac{1}{i(\pm\lambda+ \alpha t\rho^{\alpha-1 } ) } e^{\pm i\lambda\rho+it\rho^\alpha } \frac{d}{d\rho } \big ( e_{\frac{n-2}{2 } } ( r\rho ) \rho^{-\frac12}\varphi^2 ( \rho ) \big ) d\rho.\end{aligned}\ ] ] since @xmath317 , one can easily check that @xmath320 when @xmath298 or @xmath299 .
hence , using this and , we get @xmath321_{\rho=1/2}^{\rho=2 } \big|\\ & + \int\big|\frac{1}{(\pm\lambda+\alpha t\rho^{\alpha-1})}\frac{d}{d\rho } \big(e_{\frac{n-2}{2 } } ( r\rho ) \rho^{-\frac12}\varphi^2 ( \rho ) \big ) \big| d\rho\\ & \lesssim2^{-m(j , k)}2^{-\frac{3}{2}j}\end{aligned}\ ] ] as desired . consequently , if @xmath315 @xmath322 when @xmath298 or @xmath299 .
this implies @xmath323 * the cases where @xmath243 or @xmath139 .
* now we consider the following cases where @xmath243 or @xmath139 in : @xmath324 @xmath325 and @xmath326 where @xmath242 and @xmath173 .
since the second estimate follows easily from the first one using the dual characterisation of @xmath327 spaces and a property of adjoint operators , we only show and repeating the previous argument .
but here we use the following estimates ( see @xcite , p. 426 ) for bessel functions instead of lemma [ bessel ] : for @xmath328 and @xmath272 , @xmath329 first we shall show . recall that @xmath330 then , by changing variables @xmath331 , we see that @xmath332 thus , using plancherel s theorem in @xmath333 and , we get @xmath334}(r ) r^{-\frac{n-2}2 } j_{\frac{n-2}2}(r\rho^{1/{\alpha } } ) \varphi(\rho^{1/{\alpha}})h(\rho^{1/{\alpha}})\rho^{1/{\alpha } - 1 } \big\|_{\mathcal{l}_r^2 l_\rho^2 } \\ \nonumber&= c\big ( \int_0 ^ 1 r^{-(n-2 ) } \int_{1/2}^2 |j_{\frac{n-2}2}(r\rho)|^2|h(\rho)|^2 \rho^{1 - { \alpha } } d\rho r^{n-1}dr \big)^{1/2}\\ \nonumber&\lesssim\big ( \int_{1/2}^2 |h(\rho)|^2 \rho^{1-{\alpha } } \int_0 ^ 1 c r(r\rho)^{n-2}dr d\rho \big)^{1/2}\\ \nonumber&\lesssim\big(\int_{1/2}^2 |h(\rho)|^2 d\rho\big)^{1/2}\\ & = \|h\|_{l^2}.\end{aligned}\ ] ] also , by hlder s inequality , @xmath335 by interpolation between this and , we obtain @xmath336 for @xmath337
. then , by the usual @xmath252 argument as before , this implies @xmath338 for @xmath339 .
now we turn to .
first , by using and the dual estimate of , we see that @xmath340 for @xmath337 and @xmath341
. then , would follow from interpolation between this and the following estimate as before ( see the paragraph below ):
@xmath342 now we are reduced to showing . from and ,
we first write @xmath343 where @xmath344 then , we only need to show that @xmath345 recall from that @xmath346 by and , the part of @xmath269 coming from @xmath347 in @xmath348 is bounded as follows : @xmath349 now we may consider only the part of @xmath269 coming from @xmath350 , because the factor @xmath286 in @xmath348 would give a better boundedness than @xmath289 .
namely , we have to show the bound for @xmath351 let us now decompose @xmath352 as @xmath353 when @xmath354 , by the van der corput lemma as before , it follows that @xmath355}\big\{j_{\frac{n-2}{2}}(r\rho)\rho^{-\frac12}\varphi^2(\rho ) , \frac{d}{d\rho}\big(j_{\frac{n-2}{2}}(r\rho)\rho^{-\frac12}\varphi^2(\rho)\big)\big\}.\end{aligned}\ ] ] by , we see @xmath356 for @xmath312 .
thus , we get @xmath357 on the other hand , by integration by parts we see that @xmath358_{\rho=1/2}^{\rho=2}\\ & \qquad\qquad-\int\frac{1}{i(\pm\lambda+\alpha t\rho^{\alpha-1})}e^{\pm i\lambda\rho+it\rho^\alpha } \frac{d}{d\rho}\big(j_{\frac{n-2}{2}}(r\rho)\rho^{-\frac12}\varphi^2(\rho)\big)d\rho.\end{aligned}\ ] ] one can also easily check that @xmath359 when @xmath360 or @xmath361 .
hence , using this and , we get @xmath362_{\rho=1/2}^{\rho=2 } \big|\\ & + \int\big|\frac{1}{(\pm\lambda+\alpha t\rho^{\alpha-1})}\frac{d}{d\rho } \big(j_{\frac{n-2}{2 } } ( r\rho ) \rho^{-\frac12}\varphi^2 ( \rho ) \big ) \big| d\rho\\ & \lesssim2^{-k}r^{\frac{n-2}{2}}\end{aligned}\ ] ] when @xmath360 or @xmath361 .
this implies @xmath363
in this section we discuss the sharpness of theorem [ prop ] .
we will show that is a necessary condition for ( see remark [ rem ] ) .
if is valid with a pair @xmath24 on the left and a pair @xmath25 on the right , then it must be also valid when one switches the roles of @xmath24 and @xmath25 . by this duality relation ,
we only need to show the first condition @xmath364 in .
let @xmath365 be a smooth cut - off function supported in the interval @xmath366 $ ] .
let us now define @xmath367 by @xmath368 then one can easily see that @xmath369 , and @xmath370 by taking integration with respect to @xmath371 as @xmath372 now we recall from @xcite ( see p. 344
there ) that @xmath373 where @xmath374 , @xmath375 has a nondegenerate critical point at some point @xmath376 ( i.e. , @xmath377 and the matrix @xmath378(\xi_0)$ ] is invertible ) , and @xmath379 is supported in a sufficiently small neighborhood of @xmath376 . applying this with @xmath380 and @xmath381 we get @xmath382 for sufficiently large @xmath333 .
thus , if @xmath383 is sufficiently large , @xmath384 consequently , if holds , @xmath385 but , this is not possible as @xmath386 unless @xmath387 which is equivalent to the condition @xmath364 .
here we shall provide a proof of lemma [ bessel ] for estimates of bessel functions @xmath270 .
it is based on easy but quite tedious calculations . first , we recall from @xcite ( see p. 430
there ) that for @xmath271 and @xmath272 , @xmath388 where @xmath389 is the gamma function given by @xmath390 then , using the following identities @xmath391 and @xmath392 together with @xmath393 and @xmath394 one can rewrite @xmath395 here , @xmath396 and @xmath397 are the remainder terms in taylor series and , respectively , which are given by @xmath398 and @xmath399 for some @xmath400 and @xmath401 with @xmath402 . hence , when @xmath410 , by using and , it follows that @xmath413 on the other hand , when @xmath414 , by using and , we see that @xmath415 consequently , we get @xmath416 and is similarly shown by differentiating @xmath417 and using and .
z. guo and y. wang , _ improved strichartz estimates for a class of dispersive eqations in the radial case and their applications to nonlinear schrdinger and wave equations _
, j. anal . math .
124 ( 2014 ) , 1 - 38 . | in this paper we obtain some new inhomogeneous strichartz estimates for the fractional schrdinger equation in the radial case .
then we apply them to the well - posedness theory for the equation @xmath0 , @xmath1 , with radial @xmath2 initial data below @xmath3 and radial potentials @xmath4 under the scaling - critical range @xmath5 . |
the effects of propagating waves on the statistical properties of systems out of equilibrium remain an important topic of discussion . in the context of a coupled spin model in one - dimension
@xmath3@xcite it has been shown that the presence of such waves leads to weak dynamic scaling in that model .
in contrast , in a coupled burgers - like model in @xmath4 propagating waves do not affect the scaling properties of the correlation functions at all @xcite .
so far such issues have been considered only within very simplified one - dimensional nonequilibrium models @xcite .
magnetohydrodynamic ( @xmath0mhd ) turbulence , which is a hydrodynamic description of the coupled evolution of the velocity fields @xmath5 and the magnetic fields @xmath6 in a quasi - neutral plasma , stands as a very good candidate for a natural system with propagating waves in three dimensions as most of its natural realizations have proagating alfvn waves arising due to the presence of mean magnetic fields .
examples of such physical situations include solar wind , neutral plasma in fusion confinement devices etc .
the presence of the propagating alfvn modes , in addition to the usual dissipative modes due to the fluid and magnetic viscosities makes mhd turbulence a good natural example to study the interplay between the propagating and the dissipative modes in a system and their combined effects on the scaling properties of the correlation and structure functions , which are important issues from the point of view of nonequilibrium statistical mechanics .
the scaling of magnetic- [ @xmath7 and kinetic- [ @xmath8 energy spectra in the inertial range ( i.e. , wavevector @xmath9 lies in the region @xmath10 , @xmath11 and @xmath12 being the integral scale given by the system size and the dissipation scale , respectively ) in @xmath0mhd in the presence of alfvn waves originating due to a mean magnetic field @xmath13 remains controversial till the date .
numerical simulations , due to the lack of sufficient resolutions failed to conclusively distinguish between the kolmogorov s and kraichnan s predictions ( see below ) . in this communication , within one - loop self - consistent mode - coupling ( scmc ) approximations , we obtain the following results for homogeneous but anisotropic ( due to the mean magnetic field ) @xmath0mhd : * the bare alfvn wave speed , proportional to @xmath14 , renormalizes to acquire a singular @xmath9-dependence to become @xmath15 where @xmath9 is a fourier wavevector belonging to the inertial ( scaling ) range . from this result we are able to conclude that _ even _ in the presence of a mean magnetic field the kinetic- and the magnetic- energy spectra scale as @xmath1 in the inertial range , identical with the situation without any mean magnetic field . *
the dimensionless kolmogorov s constants for the kinetic- and magnetic - energy spectra depends on a dimensionless parameter @xmath16 which we identify as the ratio of the renormalised alfvn wave speed and the renormalised viscosities ( see below ) . *
the intermittency exponents of the elssser fields @xmath17 which approximately and qualitatively characterise the multiscaling properties of the structure functions in a log - normal model depend on the parameter @xmath16 and decrease with increasing @xmath16 .
thus our results show that alfvn waves in @xmath0mhd do not affect the scaling properties of the two - point correlation functions .
however , the multiscaling properties of the structure functions are shown to be affected by the mean magnetic fields .
the famous kolmogorov s arguments @xcite for fluid turbulence can be easily extended to @xmath0mhd turbulence . in mhd , in the unit where mass density @xmath18 , the kinetic energy dissipation rate per unit mass ( @xmath19 ) and the magnetic energy dissipation rate per unit mass ( @xmath20 ) have same physical dimensions .
thus , as in fluid turbulence , by claiming that the structure functions of the velocity and magnetic field differences in the inertial range must be constructed out of the mean energy dissipation rate @xmath21 per unit mass and the local length scale @xmath22 ( belonging to the inertial range ) one obtains for the @xmath23-th order structure function @xcite @xmath24\hat r_i]^n\rangle \sim ( \epsilon r)^{n/3},\;\;\eta_d\ll r\ll l\ ] ] where @xmath25 and @xmath26 for the velocity and the magnetic fields , @xmath12 is the ( small ) dissipation scale and @xmath11 is the ( large ) system size .
this yields , for the energy spectra in the inertial range ( as a function of wavevector @xmath9 ) @xmath27 this is known as the k41 theory in the relevant literature . however , in the presence of a mean magnetic field @xmath13 , the alfvn waves are generated ( see below ) with the propagation speed @xmath28 .
thus in such a case in addition to the usual kolmogorov time scale @xmath29 @xcite , there exists another time scale constructed out of the mean magnetic field @xmath14 , known as the kraichnan time scale @xmath30 @xcite .
kraichnan argued that this time scale would determine the energy cascade process and hence would enter in the expression of the structure function yielding @xmath31 .
there has not been any satisfactory resolution of this issue till this date ; due to the particularly difficult vectorial nature of the @xmath0mhd equations ( see below ) it is rather difficult to achieve high reynolds number in direct numerical solutions ( dns ) of the @xmath0mhd equations .
numerical solutions of an mhd shell - model in the presence of a small mean magnetic field - like term did not find any dependence of the multiscaling of the structure functions on the mean magnetic field @xcite .
analytically , it has been shown within the context of a @xmath32 coupled burgers - like model @xcite that the energy spectra are independent of a mean magnetic field and in case of @xmath32 mhd turbulence it exhibits the kolmogorov scaling for the energy spectra .
similar conclusions followed from an analogous one - dimensional model @xcite . various phenomenological approaches , including weak turbulence theories and three - wave interaction models , yield , in general , mean magnetic field dependences of the energy spectra @xcite when the mean magnetic field is strong
however , these theories , despite their predictions are either not directly derivable from the underlying @xmath0mhd eqs . of motion or involve additional assumptions on the flow fields .
analyses starting from the @xmath0mhd eqs . in this regard are still lacking .
most observational results on astrophysical systems seem to favour the k41 results @xcite .
simulations of incompressible mhd @xcite find results close to the k41 result . simulations of compressible mhd @xcite , even though in some cases find energy spectra closer to k41 , in general yields a less clear picture .
recent numerical results of mller _ et al _ @xcite suggest that in presence of finite magnetic helicity structure functions parallel and perpendicular to the mean magnetic fields are affected differently by the mean magnetic field . in this paper
, we address some of these issues by starting from the @xmath0mhd eqs . without making any further assumptions on the velocity and the magnetic fields , except for the validity of the perturbative approaches .
in particular , we show by applying one - loop mode coupling methods on the @xmath0mhd equations with a mean magnetic field and in the absence of any magnetic helicity that the one - dimensional energy spectra in @xmath0mhd turbulence are independent of the mean magnetic field @xmath13 and scale as @xmath1 in the inertial range where @xmath2 is a fourier wavevector belonging to the inertial range .
we then proceed to calculate the kolmogorov s constants for the kinetic- and the magnetic - energy spectra and show that they depend on @xmath14 .
lastly , we calculate the intermittency exponents for the velocity and the magnetic fields and find that they decrease with increasing @xmath14
. we do not distinguish between the longitudinal and the transverse structure functions . even though for analytical convenience we assume a weak mean magnetic field , we are able to obtain new and interesting results concerning scaling and multiscaling in @xmath0mhd in the presence of a mean magnetic field of small magnitude . in this respect
our results can be considered as complementary to some of the existing results @xcite .
the rest of the paper is organised as follows : in sec.[model ] we discuss the @xmath0mhd equations for incompressible fluids . in sec.[incom ] we show that for incompressible fluids the bare alfvn wave speed @xmath28 renormalizes to pick a correction @xmath33 in the inertial range .
this , as we argue , implies that the energy spectra scale as @xmath1 in the inertial range _ even _ in the presence of a mean magnetic field . in sec.[kolcons ] we calculate the kolmogorov s constants for the kinetic and the magnetic energy spectra .
we introduce a parameter @xmath16 which is the dimensionless ratio of the renormalised mean magnetic field and renormalised viscosities and show that the kolmogorov s constants depend on @xmath16 . in sec.[inter ] we elucidate the multiscaling properties of the structure functions by calculating the intermittency exponents and show that they decrease with increasing @xmath16 , i.e. , with increasing @xmath14 .
this suggests that a mean magnetic field tends to reduce multiscaling corrections to the k41 results for the structure functions . in sec.[conclu ] we summarize our results .
we begin by writing down the @xmath0mhd equations for the velocity fields @xmath5 and the magnetic fields @xmath6 : the velocity field @xmath5 is governed by the navier - stokes equation modified by the inclusion of the lorentz force @xcite @xmath34 and the dynamics of the magnetic field @xmath6 is governed by the induction equation @xcite constructed out of the ohm s law for a moving frame and the ampere s law : @xmath35 here , @xmath36 is the mass density , @xmath37 is the pressure , @xmath38 is the fluid viscosity and @xmath39 is the magnetic viscosity ( inversely proportional to the electrical conductivity of the fluid medium concerned )
. functions @xmath40 and @xmath41 are external forces needed to maintain a statistical steady state . in the present approach
these are taken to be stochastic forces .
we assume them to be zero - mean and gaussian distributed with specified variances ( see below ) .
in addition to eqs .
( [ mhdvcom ] ) and ( [ mhdb ] ) we also have @xmath42 ( maxwell s equation ) and , for incompressible fluids @xmath43 . if the magnetic fields @xmath44 are such that @xmath45 ( a constant vector ) then replacing @xmath6 by @xmath46 where now @xmath47 , in eqs .
( [ mhdvcom ] ) and ( [ mhdb ] ) one obtains additional linear terms proportional to wavevector @xmath2 leading to wave - like excitations , known as alfvn waves .
the resulting equations are @xmath48 and @xmath49 with @xmath50 .
the parameters @xmath51 are kept for book keeping purposes and can be set to unity ( see sec.[symm ] ) .
note that on dropping the nonlinear and the dissipative terms from eqs .
( [ mhdfullv ] ) and ( [ mhdfullb ] ) the resulting linear coupled partial differential equations admit wave - like solutions with dispersion relation linear in wavevector @xmath9 .
these are known as the alfvn waves @xcite in the literature which propagate with speed proportional to @xmath14 . in the following sections we would calculate the kinetic and the magnetic energy spectra , the kolmogorov s constants and the intermittency exponents in the presence of the alfvn waves , i.e. , for @xmath52 .
in the statistical steady state the time dependent correlation functions of @xmath5 and @xmath6 exhibit scaling which are characterized by the roughness exponents @xmath53 and @xmath54 respectively , and the dynamic exponent @xmath55 . in terms of the scaling exponents @xmath56 and @xmath55 the velocity and the magnetic field correlators have the form ( as a function of wavevector @xmath2 and frequency @xmath57 ) @xmath58 where @xmath59 and @xmath60 are scaling functions , @xmath61 is the transverse projection operator : @xmath62 to account for the incompressibility of the fields .
the kinetic- and magnetic - energy are simply related to the correlation functions : they are just the equal time velocity and magnetic field auto correlation functions multiplied by appropriate phasefactors .
we now set out to find whether the alfvn waves are relevant perturbations on the system .
it should be noted that the correlators ( [ corrform ] ) are chosen as they would be in the fully isotropic case . in the presence of a mean magnetic field @xmath13
there would however be additional anisotropic terms in the expression for the correlators above . in the small @xmath13 limit the lowest order corrections to the expressions ( [ corrform ] ) are @xmath63 .
neglecting these correction terms does not affect our scaling analyses below , and in any case , we are interested to see whether anisotropic alfvn waves are relevant perturbations in the large scale , long time limit on the isotropic @xmath0mhd as characterised by expressions ( [ corrform ] ) .
therefore , it suffices for us to work with the expressions ( [ corrform ] ) for our mode - coupling analyses below .
a complete characterisation of the nonequilibrium steady state ( ness ) of mhd , however , requires informations about , not just the scaling of the energy - spectra , but also of the @xmath23-th order equal time structure functions of the velocity and the magnetic fields in the inertial range .
these are defined by @xmath64\hat r_i|^n\rangle , \;\;a = u , b,\,\eta_d\ll r\ll l , \label{defstru}\ ] ] which scale as @xmath65 where @xmath22 belongs to the inertial range ( @xmath66 ) . according to the kolmogorov theory ( k41 ) the multiscaling exponents @xmath67 ; i.e. , they are linear in @xmath23 .
subsequent numerical and observational results for homogeneous and isotropic @xmath0mhd , i.e. , without any mean magnetic fields , @xcite suggested deviations from the k41 results which are similar to those of pure fluid turbulence @xcite .
much less results for the multiscaling of the structure functions are available when there are mean magnetic fields .
recent numerical simulations of mller _ et al _ @xcite suggested that structure functions parallel and perpendicular to the mean magnetic fields are differently affected by it when there is a finite magnetic helicity .
below we investigate the issue of multiscaling in presence of a mean magnetic field @xmath13 within the context of a log - normal model by evaluating the relevant intermittency exponents which are found to depend explicitly on @xmath14 . in our analyses we ignore the distinction between the structure functions parallel and perpendicular to the direction of the mean magnetic field ; nevertheless , as we discuss below , our results are significantly new .
we begin by re - expressing @xmath0mhd eqs .
( [ mhdvcom ] ) and ( [ mhdb ] ) by writing the magnetic fields @xmath6 as a sum of a space - time dependent part and a constant vector : @xmath68 . in terms of these fields and parameters , the @xmath0mhd eqs .
become @xmath69 and @xmath70 it should be noted that in eqs .
( [ mhdfullv1 ] ) and ( [ mhdfullb1 ] ) @xmath71 is _ not _ the mean magnetic field in general ; it would be so only if @xmath72 is zero in eqs .
( [ mhdfullv1 ] ) and ( [ mhdfullb1 ] ) .
note that our above way of splitting the magnetic fields fields does not change the actual mean magnetic field in the system : it is still given by @xmath73 .
the eqs . of motion ( [ mhdfullv1 ] ) and ( [ mhdfullb1 ] ) are invariant under the following continuous transformations @xcite : * the galilean transformation ( ti ) : @xmath74 and @xmath75 @xcite with @xmath76 in eqs . ( [ mhdfullv ] ) and ( [ mhdfullb ] ) .
this implies non - renormalization of @xmath77 @xcite . * the transformation ( tii )
@xmath78 , @xmath79 , @xmath80 .
this allows one to work with the _ effective _ magnetic fields defined by @xmath81 such that the coefficient of the lorentz force vertex constructed out of the effective magnetic fields does not renormalize .
this , therefore , ensures @xmath82 can be set to unity @xcite by treating all magnetic fields as _ effective fields_. here the shift @xmath83 is a vector . a transformation similar to tii above exists in a problem of passive scalar turbulence @xcite .
the transformation tii essentially signifies the freedom to split the total magnetic fields as a sum of a constant part and a space time dependent part .
it should be noted that the transformation tii keeps the mean magnetic field unchanged .
in fact , nonrenormalization of @xmath82 can be shown in a simpler way : let us assume that under mode eliminations and rescaling @xmath84 .
this scale factor of @xmath85 can now be absorbed by redefining the units of the magnetic fields by @xmath86 .
therefore , @xmath82 can be set to unity if all magnetic fields are considered as _ effective _ or rescaled magnetic fields . since the induction eq .
( [ mhdb ] is linear in the magnetic fields @xmath87 , such rescaling leaves every conclusion unchanged .
of course , the external force @xmath41 is also scaled by a factor @xmath88 which does not affect out analysis here as the assignment of canonical dimensions to various fields and parameters is done after absorbing @xmath82 in the definition of @xmath6 .
more specifically under the rescaling @xmath89 the bare parameters scale as @xmath90 and @xmath91 .
since there are no fluctuation corrections to the nonlinearities @xmath51 , which are the consequences of the invariances under the transformations ti and tii , they can be kept invariant under rescaling of space and time as mentioned above leading to @xmath92 and @xmath93 .
thus , under naive rescaling the bare parameter @xmath94 .
furthermore , the invariance under the transformation tii and the resulting ward identity ensure that different ways of implementing one - loop rg by having different values for @xmath95 and @xmath96 in eqs .
( [ mhdfullv1 ] ) and ( [ mhdfullb1 ] ) subject to the same bare mean magnetic field @xmath97 .
since any renormalisation scheme must respect a freedom of choice as represented by the transformation tii , respective terms must scale the same way under the rescaling of space and time @xcite .
hence , due to fluctuation corrections @xmath98 .
furthermore , since both @xmath96 and @xmath6 have the same scaling dimensions given by the exponent @xmath99 , the physical mean magnetic field in the system @xmath100 , if it receives fluctuation corrections under mode eliminations , must scale as @xmath101 , or if there are no fluctuation corrections to it due to some special symmetries ( as in the one - dimensional burgers - like model for mhd in ref.@xcite ) will be irrelevant ( in an rg ) sense as it will flow to zero as @xmath102 ( since @xmath55 for fully developed turbulence is less than unity ) .
this clearly suggests the possibility of a renormalisation of the alfvn wave speed , akin to the renormalisation of the sound speed in compressible fluid turbulence @xcite . as we will see below , for @xmath0mhd there are infra - red singular fluctuation corrections to the bare mean magnetic field leading it to renormalise in a way consistent with our predictions from the ward identities above .
it should be noted that our analyses above is independent of the strength of the bare value of the mean magnetic field .
although , in the discussion above we have rescaled space isotropically ( i.e. , @xmath103 and @xmath55 coordinates are rescaled the same way ) and analyze the scale dependence of the resulting effective parameters and the fields , it should be noted that such a rescaling does not imply that the effective parameters have isotropic structures .
we begin by writing down the @xmath0mhd equations in the incompressible limit ( i.e. , @xmath43 ) .
we write down the equations of motion in @xmath2-space in terms of the elssser variables @xmath104 .
the equations are ( we take the mean magnetic field @xmath13 to be along the @xmath105 direction ) @xmath106 the stochastic forces @xmath107 are linear combinations of @xmath108 , the tensor @xmath109 is the transverse projection operator which appears due to the divergence - free conditions on @xmath5 and @xmath6 , and @xmath110 . in the absence of any crosscorrelations between the velocity and the magnetic fields , the simplest choice for the noise variances , consistent with the divergence - free conditions on the velocity and the magnetic fields , are @xmath111 in eqs.([noisecor ] ) we choose @xmath112 .
in particular , the choice of @xmath113 in @xmath114-space dimensions ensures that the energy flux in the inertial range is a constant ( up to a weak logarithmic dependence on wavevector @xmath9 ) which forms the basis of the k41 theory @xcite , and as a result , the parameters @xmath115 and @xmath116 pick up dimensions of energy dissipation rate per unit mass .
hence , we will use @xmath113 in our calculations below . in experimental realisations of mhd ,
external forces act on the large scales . in analytical approaches ,
such forces are replaced by stochastic forces with variances given by eqs .
( [ noisecor ] ) for calculational convenience .
it should however be noted that in numerical solutions of the navier - stokes equation driven by stochastic forces with variances similar to eqs .
( [ noisecor ] ) structure functions of the velocity fields are shown to exhibit multiscaling similar to the experimental results @xcite .
our preliminary results from the numerical solutions of the isotropic @xmath0mhd equations ( i.e. , no mean magnetic fields ) yield multiscaling similar to those obtained from the @xmath0mhd eqs . driven by large - scale forces @xcite .
although no such numerical results exist for the case with a mean magnetic field , it will presumably be true for such a situation also . in a stochastically driven langevin description ,
correlation functions are proportional to the noise correlations and hence it is necessary to force the induction eq .
( [ mhdfullb ] ) stochastically , as is common in the literature ( see , e.g. , @xcite and references therein ) . for such noises in the absence of any mean magnetic field ( i.e. , for the isotropic case )
it has been shown by using a one - loop mode coupling theory that the scaling exponents have values @xmath117 @xcite . the eqs .
of motion ( [ zpmeq ] ) are coupled even at the linear level leading to a bare propagator matrix for the eqs .
( [ zpmeq ] ) of the form ( a @xmath118 refers to bare quantities ) @xmath119 we use a one - loop mode coupling theory which is conveniently formulated in terms of the self - energy matrix @xmath120 and the correlation functions given by eqs .
( [ corrform ] ) .
the self - energy matrix @xmath121 is defined by @xmath122 where @xmath123 is the renormalised propagator matrix . in terms of the scaling exponents @xmath99 and @xmath55 , and the _ renormalized _ parameters the self - energy matrix @xmath120 is given by @xmath124 where @xmath125 is the _ renormalised _ or the _ effective _ alfvn wave speed .
if there are diagrammatic corrections to the imaginary parts of @xmath126 at frequency @xmath127 which are singular in the infra - red limit , then @xmath125 is different from @xmath14 , the bare alfvn wave speed , else they are the same .
there are two one - loop diagrams which contribute to the fluctuation corrections to @xmath128 .
these are shown in fig.([diag1 ] ) .
these have both real and imaginary parts at frequency @xmath127 .
thus there are fluctuation corrections to the bare alfvn wave speed which are singular in the small wavenumber limit .
we assume , in the spirit of mode - coupling methods , @xmath129 ( we assume an isotropic scale dependence for @xmath125 for simplicity which suffices our purpose of finding the scale dependence of the effective alfvn wave speed ) . clearly , if @xmath130 , the small wavenumber limit of the problem is dominated by underdamped waves , if @xmath131 the alfvn waves are damped out in the small wavenumber limit .
in contrast if @xmath132 then in the small wavenumber limit both the propagating and the dissipative modes are present . our analyses in sec.[symm ] clearly suggest that if there are fluctuation corrections to @xmath14 then the _ effective _ scale dependent alfvn wave speed @xmath133 yielding @xmath134 . furthermore ,
since @xmath93 we have @xmath132 leading to the co - existence of the underdamped alfvn waves and the dissipative modes in the large - scale , long - time limit .
note that this situation allows us to define a dimensionless parameter @xmath135 where @xmath136 and @xmath137 are the renormalised alfvn speed and the viscosity respectively . in the scmc approach vertex corrections
are neglected .
lack of vertex renormalisations in the zero wavevector limit in @xmath0mhd allows scmc to yield _ exact relations _ between the scaling exponents @xmath99 and @xmath55 as in the noisy burgers / kardar - parisi - zhang equation @xcite . in this model
the nonlinearities and the noise variances do not renormalise , thus leading to @xmath138 , satisfying the exponent identity @xmath93 . in the present case , due to the singular nature ( in the small wavevector limit ) of the bare noise correlations ( [ noisecor ] )
, they do not pick up any further singular corrections to their scaling ; however the amplitudes get modified ( this is similar to the results in ref.@xcite ) .
we denote the renormalized amplitudes by @xmath139 and @xmath140 , respectively .
furthermore , for the @xmath0mhd eqs .
( [ mhdfullv ] ) and ( [ mhdfullb ] ) there are infra - red singular fluctuation corrections to @xmath14 [ see the one - loop diagrams in fig .
( [ diag1 ] ) ] leading to @xmath141 consistent with our arguments above .
the scmc approach involves consistency in the scaling and the amplitudes of the mode coupling equations .
it should be noted that the exponent values @xmath142 satisfy the mode coupling integral equations regardless of the strength of the mean magnetic field . for amplitude consistency
the one - loop integrals are required to be evaluated . due their complicated structure
, we evaluated them by assuming that the strength of the mean magnetic field is small .
therefore , only our amplitude relations and not the scaling exponents are affected by the approximation of small mean magnetic fields . .
a single line refers to a propagator and a line with a filled circle refers to a correlator.,title="fig:",width=384 ] .
a single line refers to a propagator and a line with a filled circle refers to a correlator.,title="fig:",width=384 ] demanding amplitude consistency in the mode coupling equations we obtain , @xmath143+\frac{\tilde d}{d(d+2)\eta_+^2}-\frac{\beta^2 d}{\eta_+^2d}\left[1-\frac{2}{d}+ \frac{1}{d(d+2 ) } -\frac{\beta^2 \tilde d}{\eta_+^2d(d+2)(d+4)}\right],\nonumber \\ \beta&\equiv & \frac{b}{\eta^+},\;b = b_o\frac{d}{d(d+2)\eta_+^3 } , \nonumber \\
\frac{1-\gamma}{1+\gamma}&=&\frac{(1+\gamma^2 ) ( 1-\frac{2}{d}+\frac{2}{d(d+2 ) } ) -0.5\gamma ( 1-\frac{2}{d})+f_1(\beta)}{(1+\gamma^2 ) ( 1-\frac{2}{d}+\frac{2}{d(d+2)})+0.5\gamma ( 1-\frac{2}{d})+f_2(\beta ) } , % d&=&d_o[1+d_o\frac{1-\beta^2/d + \frac{2\beta^2}{d(d+2)}- % \frac{2\beta^2}{d(d+2)(d+4)}-6\gamma^2\beta^2/d % + 6\gamma^2\beta^2/d(d+2)- \frac{6\gamma^2\beta^2}{d(d+2)(d+4 ) } } ] % { \label{ampli1}\end{aligned}\ ] ] where @xmath144 are the @xmath16-dependent parts of the amplitude - ratio . while calculating the amplitude consistency relations we worked in the limit of small @xmath135 and the renormalised magnetic prandtl number ( @xmath145 ) close to unity .
away from these limits the amplitudes of the underlying one - loop integrals become much more complicated functions of @xmath16 and also of the renormalised magnetic prandtl number , but the qualitative picture remains unchanged . the main physical picture that emerges from the expressions ( [ ampli1 ] )
is that in the absence of a bare mean magnetic field @xmath14 the effective alfvn wave speed is also zero which is a restatement of the fact that if the original theory is isotropic , so will be the renormalised theory .
moreover the renormalised alfvn speed increases with increasing @xmath14 . with these results we are now in a position to calculate the energy spectra in the inertial range .
we use the form of the effective ( scale dependent ) viscosity and the alfvn wave speed to obtain the equal time correlation function of @xmath146 in the long wavelength limit .
we define ( in @xmath0 ) @xmath147[i(\omega+bk^{-1/3}k_z)+\eta_+k^{2/3}]}. \label{corrdef}\end{aligned}\ ] ] as discussed before , we have omitted anisotropic corrections of @xmath63 or of @xmath148 to the correlation functions as we are trying to find out the relevance ( in an rg sense ) of alfvn waves on the scaling of the isotropic correlation functions in the absence of any mean magnetic fields . therefore , the equal time correlation functions have the following form ( in @xmath149 ) : @xmath150 therefore , the equal time autocorrelation functions of @xmath151 are independent of any mean magnetic field .
this holds true regardless of the scaling of the noise variances .
the equal time cross correlation function @xmath152 requires more careful considerations .
on integrating @xmath153 over all frequency @xmath57 one obtains ( in @xmath0 ) @xmath154 a form which is valid in the inertial range .
it is clear from the expression ( [ correxp2 ] ) that both the real and the imaginary parts of @xmath155 scale as @xmath156 in the inertial range .
thus the one - dimensional kinetic- and the magnetic - energy spectra ( which are simply related to the correlators defined in eqs .
( [ corrdef ] ) ) scale as @xmath1 in the inertial range .
the emerging physical picture is as follows : we find from the expression for @xmath153 above that this , as a function of frequency @xmath57 , has maxima at @xmath157 and the width at half - maxima @xmath29 .
in contrast , the auto correlations of @xmath158 are maximum at @xmath127 and their widths scale as @xmath29 . thus in the long wavelength limit the width and the location of the maxima of @xmath153 scale in the same way leading to the presence of the underdamped alfvn waves in the hydrodynamic limit .
therefore , it immediately follows that the kinetic- and the magnetic - energy spectra , being linear combinations of the correlators discussed above times appropriate phase factors , scale as @xmath1 _ even _ in the presence of a non - zero mean magnetic field .
it should be noted that we have used a small @xmath16 approximation to arrive at our results for the self - consistent amplitude relations . for
a finite @xmath16 one would require to work with a fully anisotropic form of the correlation functions and obtain self - consistent relations for scaling and anisotropic amplitudes , a task , which is analytically challenging , remains to be done in the future .
however , based on our calculations above , the exponent identity @xmath93 and the ward identity suggesting that the renormalised alfvn wave speed should scale as @xmath33 with wavevector @xmath9 being in the inertial range , we argue that _ even _ for a finite @xmath16 , i.e. , a finite mean magnetic field the energy spectra will scale as @xmath1 in the inertial range , a result supported by many observational evidences @xcite .
the self - consistent amplitude equations will then be anisotropic reflecting the presence of underdamped alfvn waves in the inertial range
. the full correlation matrix will be anisotropic in the hydrodynamic limit ; its eigenvalues have different amplitudes , but all of them scale the same way . our confidence on our result that the scaling of the correlation functions along directions parallel and perpendicular to the direction of the mean magnetic field is same is derived from the fact that our exponent values @xmath142 satisfy the one - loop integral equations regardless of the strength of the mean magnetic field and are consistent with the ward identity discussed above .
according to the kolmogorov s hypothesis for fluid turbulence @xcite , in the inertial range energy spectrum @xmath159 , where @xmath160 , a _ universal _ constant , is the kolmogorov s constant and @xmath21 is the energy dissipation rate per unit mass .
various calculations , based on different techniques by different groups @xcite show that @xmath161 in three dimensions .
having noted that the energy spectra , even in the presence of a mean magnetic field scale as @xmath1 extensions of kolmogorov s hypothesis for @xmath0mhd allows one to define kolmogorov s constants for the elssser fields : @xmath162 . since @xmath104 , we have @xmath163 and @xmath164 in absence of any crosscorrelations between the velocity and the magnetic fields , we have @xmath165 .
the noise strength @xmath139 and the rate of energy dissipation per unit mass is connected by the novikov s theorem @xcite:@xmath166 .
noting that the energy spectra @xmath167 of @xmath168 in the inertial range is given by @xmath169 where @xmath139 is the _ effective _ or renormalised noise strength , we identify @xmath170^{2/3}. \label{kolm}\ ] ] the notable feature of the expression ( [ kolm ] ) is that the constant @xmath171 depends on the dimensionless parameter @xmath16 which we introduced before . for @xmath172 ,
i.e. , for no magnetic fields , we find @xmath173 for pure fluid turbulence which is well - within the accepted range of values @xcite . before closing this sec .
we would like to point out that the presence of multiscaling raises questions about kolmogorov s constant @xmath174 being universal : a small but finite intermittency correction ( i.e. , multiscaling ) over the simple k41 scaling implies the presence of an _ arbitrary _ scale which may spoil the universality of @xmath174 .
we however refrain ourselves from getting into this question and adopt a point of view that regardless of whether or not @xmath174 remains universal due to multiscaling , the numerical value of this constant is likely to get affected by the presence of alfvn waves in the system which is reflected by the expression ( [ kolm ] ) .
experiments and numerical sumulations @xcite find nonlinear multiscaling corrections to the k41 prediction of @xmath175 for the structure functions in the inertial range . until the date
, no controlled perturbative calculation for @xmath176 is available . to account for multiscaling in fluid turbulence , however , obukhov @xcite and kolmogorov @xcite assumed a log - normal distribution for dissipation @xmath21 to arrive at @xmath177 where @xmath178 is the mean value of @xmath21 and ( a bound for @xmath21 in fluid turbulence has been discussed in ref.@xcite ) : @xmath179 for small @xmath180 , @xmath181 . a standard calculation on the randomly stirred model yields _ intermittency exponent _
@xcite where @xmath183 , whereas the best possible estimate from experiments is 0.23 @xcite .
this model , despite having well - known limitations and difficulties @xcite , serves as a qualitative illustration of multiscaling . as in fluid turbulence , in mhd the dissipations @xmath184 of the elssser variables @xmath168
fluctuate in space and time , and as a result one may define two intermittency exponents @xmath185 for them . in the present problem @xmath186 .
below we calculate the exponent @xmath187 in a one - loop expression which will give us an estimate of the alfvn wave speed - dependent deviation of the scaling of the structure functions from their simple - scaling values as predicted by the k41 theory .
we closely follow ref.@xcite .
we work with the self - consistent forms for the self - energies and correlation functions given above along with the consistency relations for the amplitude - ratios @xmath188 .
following ref.@xcite , we find the dissipation correlation functions in @xmath0 to be @xmath189 with @xmath85 being defined by the relation @xmath190 . from the self - consistent amplitude - relations ( [ ampli1 ] ) we find @xmath191^{1/3}\left[1 - 0.5\gamma -4\beta^2(0.7 - 0.03\gamma)\right]^{1/3}. \label{alpha}\ ] ] thus , @xmath85 which is a universal coefficient in ordinary fluid turbulence ,
varies with the parameter @xmath16 , or with the alfvn wave speed in mhd . substituting the values of @xmath160 and @xmath85 we find @xmath192^{4/3}. \label{delta}\ ] ]
thus we find that a decrease in the value of @xmath83 with an increase in @xmath16 , i.e. , with increasing mean magnetic field . despite the limited applicabilities of log - normal models in characterising multiscaling
, we can conclude , from our expression ( [ delta ] ) for the intermittency exponent in mhd in the presence of a mean magnetic field , that the intermittency corrections to the simple k41 scaling is likely to get affected ( reduced in our calculations ) in the presence of alfvn waves .
further calculations and/or numerical simulations are needed to find the exact extent of the dependence of multiscaling on the mean magnetic field and the possibilities of anisotropic multiscaling for the structure functions parallel and perpendicular to the mean magnetic field as demonstrated recently in ref.@xcite .
it should be noted that our conclusions on the multiscaling properties of the structure functions depend on a log - normal model for @xmath0mhd . such a description , unfortunately
, is unable to distinguish between the structure functions parallel and perpendicular to the direction of the mean magnetic field in the system .
moreover , the intermittency exponents above [ expressions ( [ delta ] ) ] are evaluated in the lowest order in mean magnetic field .
therefore , even though from our results we are not able to make firm comments on the possibility of parallel and perpendicular structure functions exhibiting different multiscaling , the real importance of our results lie in their elucidation of the multiscaling properties depending on the mean magnetic field .
in this paper we have considered the effects of the alfvn waves on the statistical properties of the correlation functions of the velocity and the magnetic fields or the elssser fields .
we considered the case when the velocity fields are incompressible . in a one - loop approximation
we find that the _ effective _ or the renormalised alfvn wave speed scales as @xmath33 where the wavevector @xmath9 is in the inertial range .
this immediately yields that the energy spectra , even in the presence of a mean magnetic field , scale as @xmath1 in the inertial range .
we identify a dimensionless parameter @xmath16 which is the ratio of the effective alfvn wave speed and the renormalised viscosity .
we obtain self - consistent relations between the amplitude ratio @xmath193 of the correlation functions and @xmath16 .
these relations allow us to calculate the dimensionless kolmogorov s constant and we show that it depends explicitly on @xmath16 or on the mean magnetic field .
finally , we calculate the intermittency exponent which in a log - normal model gives a qualitative account of the multiscaling in terms of the deviation from the k41 scaling for the structure functions .
we would like to emphasize that although the one - dimensional burgers - like model of mhd of ref.@xcite and the @xmath0mhd eqs .
yield energy spectra independent of the alfvn waves , the long wavelength physical pictures are different . in the former case , due to the nonrenormalization of the bare alfvn wave speed
, alfvn waves are overdamped in the hydrodynamic limit ; the dominant process in that limit is viscous dissipation .
in contrast , in @xmath0mhd , the alfvn wave speed picks up a singular correction in the hydrodynamic limit leading to the k41 scaling of the energy spectra and the presence of underdamped alfvn waves in the hydrodynamic limit . despite similar mathematical structures these crucial differences
arise principally because the one - dimensional model decouples completely when written in terms of the elssser variables , allowing to comove with the waves of each of them separately . as a result correlators of each of them
are independent of the alfvn waves and the bare alfvn wave speed remains unrenormalised leading to overdamped alfvn waves in the hydrodynamic limit .
however , the @xmath0mhd equations do not decouple when written in terms of the elssser variables and hence oppositely propagating waves can not be made to vanish by comoving . despite the limitations of the one - loop methods @xcite and the small @xmath16 approximation to facilitate easier analytical manipulations for the amplitude relations , we obtain results which are significantly new and open the intriguing possibilities of mhd multiscaling universality classes being parametrised by the mean magnetic field . some of the quantitative details will change if one retains terms which are higher order in @xmath16 ; however , we believe that the qualitative picture will essentially remain the same .
as mhd turbulence forms a natural example of a driven nonequilibrium system with alfvn waves , our results are the first of its kind for a natural system . in our log - normal model approach
, we did not distinguish between the longitudinal and the transverse structure functions our results can not be compared directly with those of ref.@xcite , where the longitudinal and the transverse structure functions are found to scale differently with multiscaling exponents which depend on the magnitude of the mean magnetic field in the presence of a finite magnetic helicity . in contrast our results apply to the multiscaling of the usual structure functions in the absence of any magnetic helicity which are combinations of the transverse and the longitudinal structure functions as considered in ref.@xcite which would also then exhibit a mean magnetic field dependent scaling . in particular
we find that the usual structure functions multiscale less in presence of an increasingly strong mean magnetic field which is in qualitative agreement with those of ref.@xcite .
although here we have restricted ourselves to the study of alfvn waves as relevant perturbations on the amplitude and the scaling of the isotropic correlation functions in the limit of small @xmath16 , our results indicate the possibility of the multiscaling exponents varying with the amplitude of the mean magnetic field .
our analyses , when extended for finite values of @xmath16 and for full anisotropic structure of the correlation functions , are likely to provide further understanding of and resolve some of the discrepancies between the various phenomenological scenarios available for situations with large mean magnetic fields @xcite .
we leave this task for the future . from a broader point of view our results
demonstrate the critical role of wave - like excitations in determining the statistical properties of @xmath0mhd .
the presence of propagating waves is not confined to mhd only ; they are a generic feature in many other naturally occurring soft - matter systems where such waves can be present in a viscous environment , e.g. , active polar gels in cytoskeletal dynamics @xcite and in the dynamics of self - propelled particles @xcite etc .
we believe our results will lead to similar theoretical and experimental studies in relevant nonequilibrium soft - matter systems .
this work was started when one of the authors ( ab ) was an alexander von humboldt fellow at the hahn meitner institut , berlin .
ab wishes to thank the avh stiftung , germany for partial financial support . 99 e-mail:[email protected] d. das , a. basu , m. barma and s. ramaswamy , _ phys .
rev e _ , * 64 * , 021402 ( 2001 ) .
a. basu , j. k. bhattacharjee and s. ramaswamy , _ eur . phys .
j b _ , * 9 * , 425 ( 1999 ) . m. k. verma , _ phys . plasmas _ , * 6 * , 1455 ( 1999 )
. a. n. kolmogorov , c. r. acad .
30 * , 301 ( 1941 ) .
p. goldreich and s. sridhar , _ astrophys .
_ , * 438 * , 763 ( 1995 ) .
a. basu , phd thesis , indian institute of science , bangalore ( 2000 ) .
v. yakhot and s.a .
orzag , _ phys .
lett . _ * 57 * , 1722 ( 1986 ) .
a. basu , _ europhys .
_ , * 65 * , 505 ( 2004 ) . r. h. kraichnan , _ phys . fluids _ , * 8 * , 1385 ( 1965 ) . s. galtier _
et al _ , astro - ph/0008148 .
c. s. ng and a. bhattacharjee , _ astrophys . j _ , * 465 * , 845 ( 1996 ) . j. v. shebalin _
et al _ , _ j. plasma phys . _ ,
* 29 * , 525 ( 1983 ) .
d. montgomery and l. turner , _ phys .
fluids _ , * 24 * , 825 ( 1981 ) . b. j. rickett and a. g. lyne , _ mon . not .
r. astron .
_ , * 244 * , 68 ( 1990 ) ; s. r. spangler and c. r. gewinn , _ astrophys .
_ , * 353 * , l29 ( 1990 ) ; y. gupta , b. j. rickett and w. a. coles , _ astrophys .
_ , * 403 * , 183 ( 1993 ) ; p. n. wilkinson , r. narayan and r. e. spencer , _ mon . not
r. astron .
_ , * 269 * , 67 ( 1994 ) .
j .cho and e. t. vishniac , _ astrophys .
_ , * 539 * , 273 ( 2000 ) ; j. cho , a. lazarian and e. t. vishniac , _ astrophys .
, * 564 * , 291 ( 2002 ) .
j. cho and a.lazarian , _ mon . not .
r. astron .
_ , * 345 * , 325 ( 2003 ) .
w - c mller , d biskamp and r. grappin , _ phys . rev .
e _ , * 67 * , 066302 ( 2003 ) .
j d jackson , _ classical electrodynamics _ , 2nd edn .
, wiley eastern , new delhi , ( 1975 ) ; a. raichaudhuri , _ the physics of fluids and plasmas _ , cambridge university press , cambridge ( 1998 ) .
d. montgomery , _ lecture notes on turbulence _ , eds . j. r. herring and j.c.mcwilliam , world scientific , singapore , 1989 .
a. basu , a. sain , s. k. dhar and r. pandit , _ phys .
_ , * 81 * , 2687 ( 1998 ) .
d. biskamp and w. mller , _ phys .
plasmas _ , * 7 * , 4889 ( 2000 )
. k. r. sreenivasan and r. a. antonia , _ ann .
fluid mech .
_ , * 29 * , 435 ( 1997 ) ; and s. k. dhar , a. sain , a. pande , and r. pandit , _ pramana - j. phys .
_ , * 48 * , 325 ( 1997 )
. a. basu , unpublished .
d. forster _
et al _ , _ phys .
rev a _ , * 16 * , 732 ( 1977 ) ; j.k .
bhattacharjee , _ j. phys .
a _ , * 31 * , l93 ( 1998 ) .
e. frey and u. tuber , _ phys .
e _ , * 50 * , 1024 ( 1994 ) . n. v. antonov , _ physica d _ , * 144 * , 370 ( 2000 )
. i. s. taroselsky , v. yakhot , s. kida and s. a. orszag , _ phys .
_ , * 65 * , 171 ( 1990 ) ; j. k. bhattacharjee , _ mod .
_ , * 7 * , 881 ( 1993 ) .
a. sain , manu and r. pandit , _ phys
_ , * 81 * , 4377 ( 1998 ) .
a. basu and r. pandit , work in progress .
r. h. kraichnan , _ j. fluid mech .
_ , * 47 * , 525 ( 1971 ) . c. d. leslie , _ developments in the theory of turbulence _ , clarendon press , oxford , 1973 .
a. das and j. k. bhattacharjee , _ europhys .
_ , * 26 * , 527 ( 1994 )
. w. a. woyczynski , _ burgers - kpz turbulence - gttingen lectures _ , springer - verlag , berlin , heidelberg , 1998 ; w.d mccomb , _ the physics of fluid turbulence _ , clarendon press , oxford , 1990 . j. cho , a. lazarin and e. t. vishniac , _ apj _ , * 564 * , 291 ( 2002 ) ; d. biskamp and w. mller , _ phys . plasmas _ , * 7 * , 4889 ( 2000 ) . a. m. obukhov , _ j. fluid mech .
_ , * 13 * , 77 ( 1962 ) . a. m. kolmogorov , _ j. fluid mech .
_ , * 13 * , 82 ( 1962 ) . c. r. doering and c. foias ,
_ j. fluid mech .
_ , _ 467 _ , 289 ( 2002 ) . u. frisch , _ turbulence : the legacy of a.n . kolmogorov _ ( cambridge university press , cambridge , 1995 )
. k. kruse _
et al _ , e - print : physics/0406058 .
r. a. simha and s. ramaswamy , _ phys .
_ , * 89 * , 058101 ( 2002 ) . | we analyse the effects of the propagating alfvn waves , arising due to non - zero mean magnetic fields , on the nonequilibrium steady states of three - dimensional ( @xmath0 ) homogeneous magnetohydrodynamic ( mhd ) turbulence .
in particular , the effects of alfvn waves on the universal properties of @xmath0mhd turbulence are studied in a one - loop self - consistent mode - coupling approach .
we calculate the kinetic- and magnetic energy - spectra .
we find that _ even _ in the presence of a mean magnetic field the energy spectra are kolmogorov - like , i.e. , scale as @xmath1 in the inertial range where @xmath2 is a fourier wavevector belonging to the inertial range .
we also elucidate the multiscaling of the structure functions in a log - normal model by evaluating the relevant intermittency exponents , and our results suggest that the multiscaling deviations from the simple kolmogorov scaling of the structure functions decrease with increasing strength of the mean magnetic field .
our results compare favourably with many existing numerical and observational results . |
one of the most interesting results obtained at rhic in au
au collisions was the observation that the baryon ( anti - baryon ) production at intermediate transverse momenta becomes comparable to that of mesons @xcite .
the topological decay reconstruction of k@xmath0and @xmath1 provides a unique opportunity to extend the baryon and meson identification to much larger transverse momenta than would be possible using conventional particle identification methods .
measurement performed by the star collaboration ( see , for example , @xcite ) , showed that the baryon / meson ratio reaches its maximum at @xmath6 gev/@xmath4 and starts decreasing at higher momenta .
the maximum value of the @xmath1/k@xmath0ratio in central collisions was found to exceed unity .
the question of why , in nucleus - nucleus collisions , baryons at intermediate @xmath7 appear to be more easily produced than mesons is still open .
possible explanations involve interplays between soft and hard mechanisms of particle production ( also at the partonic level ) like those discussed , for example , in ref .
the evolution of the baryon / meson ratio with collision energy may yield additional information about this `` baryon anomaly '' .
in this article , we present the @xmath1/k@xmath0ratios measured by the alice experiment at the lhc in 1.1@xmath810@xmath9 minimum bias pb pb events at @xmath10 tev as a function of transverse momentum and for different collision centrality bins , as well as in pp collisions at @xmath3 and 7 tev .
the alice experiment is well suited for k@xmath0and @xmath1 reconstruction over a wide momentum range . for the results discussed here
, the momentum range @xmath11 gev/@xmath4 is defined by our current level of understanding the systematic uncertainties .
the k@xmath0and @xmath1 particles were reconstructed via their v0 decay topology @xcite .
the method was the same for both pp and pb pb collisions .
the typical reconstruction efficiencies ( extracted from monte carlo studies ) were about 40 % for k@xmath0and 30 % for @xmath1 at @xmath12 gev/@xmath4 ( close to the @xmath1/k@xmath0maximum ) . in the momentum range
@xmath13 gev/@xmath4 , the variation of the ratio of the reconstruction efficiencies for k@xmath0and @xmath1 was no larger than 12 % .
the @xmath7 spectra of k@xmath0 obtained in centrality bins in pb
pb collisions were compared with the spectra of charged kaons reconstructed by the alice time projection chamber and the time of flight detector . at @xmath14 gev/@xmath4 ,
the two sets of spectra agreed within 12 % @xcite .
the spectra of @xmath1 were corrected for the contribution of @xmath1 s coming from decays of @xmath15 and @xmath16 .
this was done by re - scaling the corresponding distributions extracted from monte carlo simulations with the @xmath7 spectrum of @xmath15 reconstructed in real data , assuming that the number @xmath16 is proportional to the reconstructed number of @xmath15 .
the obtained feed - down corrections turned out to be of the order of 20 % and changed within only a few per cent as a function of event centrality and transverse momentum .
the efficiency and feed - down corrections were checked with the life - time distributions for the v0 particles .
these distributions were corrected as functions of two variables , @xmath7 and decay length .
this was done for all the event centrality bins separately .
the statistical error of the reconstructed life times for k@xmath0and @xmath1 was of the order of 1 % .
however , the systematic deviation from the corresponding nominal values was @xmath174 % in the case of the most central events . altogether , we considered the following main sources of systematic uncertainties , listed here along with their contributions to the overall uncertainty : signal extraction ( 3 % ) , efficiency correction ( 7 % for @xmath18 gev/@xmath4 , 1 % for @xmath19 gev/@xmath4 ) , feed - down correction ( 5 % ) , admixture of @xmath1 s generated in the detector material ( 2 % ) .
/k@xmath0ratios as a function of @xmath7 for different centralities in pb
pb collisions at @xmath10 tev , and also for minimum bias pp collisions at @xmath20 and 7 tev .
right : selected @xmath1/k@xmath0ratios shown on the left compared with those measured in au au collisions at @xmath21 gev .
[ fig1 ] , title="fig:",scaledwidth=47.0%,scaledwidth=37.0% ] /k@xmath0ratios as a function of @xmath7 for different centralities in pb
pb collisions at @xmath10 tev , and also for minimum bias pp collisions at @xmath20 and 7 tev .
right : selected @xmath1/k@xmath0ratios shown on the left compared with those measured in au au collisions at @xmath21 gev .
[ fig1 ] , title="fig:",scaledwidth=52.0%,scaledwidth=35.0% ] -2 mm the @xmath1/k@xmath0ratios as a function of @xmath7 for different centralities in pb pb collisions at @xmath10 tev are shown in fig .
[ fig1 ] ( left ) .
the same ratios for minimum bias pp events at 0.9 and 7 tev are also given .
the baryon / meson ratio in pp interactions always stays below 1 and is quite similar to what is observed in peripheral pb
pb collisions .
as the collision centrality increases , the baryon / meson ratio develops a maximum at @xmath12 gev/@xmath4 reaching a value of @xmath22 for the 05 % most central events .
comparing these preliminary @xmath1/k@xmath0ratios with those measured by the star collaboration in au au collisions at @xmath21 gev , we notice that , in the case of most central events , the baryon / meson ratio at the lhc decreases less rapidly with @xmath7 than at rhic ( see fig .
[ fig1 ] ( right ) ) .
the preliminary star data points @xcite shown in this figure are multiplied by the @xmath23 factor calculated from the data reported in @xcite ( to account for the non - unity of the anti - baryon / baryon ratio at rhic ) and subtracted a 10 % feed - down correction quoted in @xcite . a comparison between the maximum values of ( anti-)baryon / meson ratios measured by alice in pb pb collisions at @xmath10 tev and those obtained by star in au
au events at @xmath5 and 200 gev is presented in fig .
[ fig2 ] ( left ) .
the au au points at 62.4 gev are plotted as they are published in @xcite . to compare the star measurements at 200 gev with the alice results , we multiply the star values from @xcite by the same @xmath24 factor and apply the same feed - down correction as mentioned above . as is evident in fig .
[ fig2 ] ( left ) , the maximum value of the @xmath1/k@xmath0ratio increases with the beam energy . the position in @xmath7 of the @xmath1/k@xmath0maximum measured at pb pb collisions at @xmath10 tev is slightly shifted towards higher transverse momenta with respect to that observed in au au events at @xmath21 gev , as shown in the fig .
[ fig2 ] ( right ) .
/k@xmath0ratio as a function of number of participants compared between different colliding systems and energies .
right : position in @xmath7 of the @xmath1/k@xmath0maximum as a function of number of participants compared between pb pb at @xmath10 tev , pp at @xmath25 tev and au au at @xmath26 gev collisions .
[ fig2 ] , title="fig:",scaledwidth=48.0%,scaledwidth=36.0% ] /k@xmath0ratio as a function of number of participants compared between different colliding systems and energies .
right : position in @xmath7 of the @xmath1/k@xmath0maximum as a function of number of participants compared between pb pb at @xmath10 tev , pp at @xmath25 tev and au au at @xmath26 gev collisions .
[ fig2 ] , title="fig:",scaledwidth=48.0%,scaledwidth=36.0% ]
we have presented the measurements of @xmath1/k@xmath0ratios in pb
pb collisions at @xmath10 tev performed with the alice experiment at the lhc .
the ratios are compared to those measured by alice in pp collisions at @xmath27 and 7 tev as well as with the star results in au
au events at @xmath5 and 200 gev . in pb
pb collisions , the @xmath1/k@xmath0ratio as a function of the transverse momentum shows a broad maximum around @xmath12 gev/@xmath4 . the maximum value of the ratio increases with the collision centrality reaching the value of @xmath22 for the 0 - 5 % most central events .
as the collision centrality decreases , the maximum of the @xmath1/k@xmath0ratio becomes less pronounced and diminishes to a value of @xmath28 .
the same behaviour of the ratio is observed in pp events at @xmath20 and 7 tev which bracket in energy the pb
pb results reported here . comparison with similar measurements performed by the star collaboration in au
au collisions at @xmath5 and 200 gev shows that the value at the @xmath1/k@xmath0maximum increases with the beam energy . at the same time , the position of the maximum in @xmath7 shifts towards higher transverse momenta .
the magnitude of this shift is smaller than it was predicted , for example , in ref .
@xcite . however , the baryon enhancement in central nucleus - nucleus collisions at the lhc decreases less rapidly with @xmath7 and , at @xmath29 gev/@xmath4 , is a factor @xmath30 higher compared with that at rhic .
10 j. adams _ et al . _ [ star collaboration and star - rich collaboration ] , [ nucl - ex/0601042 ] . k. adcox _ et al . _
[ phenix collaboration ] , phys .
c * 69 * ( 2004 ) 024904 .
[ nucl - ex/0307010 ] .
m. a. c. lamont [ star collaboration ] , eur .
j. c * 49 * ( 2007 ) 35 .
r. j. fries , b. muller , eur . phys .
j. c * 34 * ( 2004 ) s279-s285 .
[ nucl - th/0307043 ] . | we present the study of k@xmath0and @xmath1 production performed with the alice experiment at the lhc in pb
pb collisions at @xmath2 tev and pp collisions at @xmath3 and 7 tev .
the k@xmath0and @xmath1 particles are reconstructed via their v0 decay topology allowing their identification up to high transverse momenta .
the corresponding baryon / meson ratios as a function of transverse momentum are extracted for pb
pb collisions in centrality bins and in the transverse momentum range from 1 to 6 gev/@xmath4 .
they are also compared with those measured in pp events at the lhc energies of 0.9 and 7 tev as well as in au
au collisions at @xmath5 and 200 gev from rhic . |
inflation , as the paradigm of the early universe , has not only solved a lot of fine - tuning problems of the big bang theory , but also predicted the primordial scalar and tensor perturbations .
the primordial tensor perturbations , i.e. primordial gravitational waves ( gws ) @xcite@xcite@xcite , have arisen great attentions after the bicep2 collaboration s announcement of the detection of b - mode signal in the cmb ( around @xmath1 ) @xcite , which was interpreted by them as the imprint of the primordial gws , though this result is doubtful due to the foregrounds of polarized dust emissions @xcite@xcite , see also @xcite .
the detection of primordial gws would verify general relativity ( gr ) and strengthen our confidence in inflation and quantum gravity@xcite , and also put more constrains on inflation models and modified gravity at the same time . besides the cmb experiments which mainly aimed at detecting low frequencies ( @xmath2 hz ) gws , many experiments relate to higher frequencies based on other methods , such as pulsar timing array ( @xmath3 hz ) , laser interferometer detectors ( @xmath4 hz ) ,
will be carried out in the coming decades . however , since the amplitude of the gws would stay constant after they are stretched outside the horizon , and decrease with the expansion of the universe after they reenter the horizon , the primordial gws with longer wavelength provide the most of opportunities for the detection @xcite
. therefore , the cmb observations , especially the cmb b - mode detections , are still the most promising experiments to detect the primordial gws if they actually exist .
einstein s gr is the most accepted theory of gravity .
however , it might be required to modify when dealing with the inflation in the primordial universe and the accelerated expansion of the current universe . during the matter and radiation dominated era ,
modified gravity has several effects on the cmb spectra , such as the lensing contribution to b - modes @xcite and the variation of propagating speed @xmath0 of primordial gws @xcite@xcite , we are especially interested in the latter in this paper , see e.g.@xcite for the case with the scalar perturbation . in gr , graviton is massless and propagates along the null geodesics , so the propagating speed @xmath0 of gws is naturally set to be unity , i.e. the speed of light .
but in modified gravity , e.g. , the low - energy effective string theory with high - order corrections @xcite@xcite@xcite@xcite@xcite , and also modified gauss - bonnet gravity @xcite , and generalized galileon ( horndeski theory @xcite ) @xcite@xcite@xcite@xcite , and beyond horndeski theories @xcite,@xcite , and the effective theory of fluids at next - to - leading order in derivatives ( e.g.@xcite ) , @xmath0 might deviate from unity .
because the value of @xmath0 determines the time of horizon crossing of gws , during the recombination the change of @xmath0 can result in a shift of the peak position of the primordial b - modes , see @xcite@xcite , for example .
in this paper , we focus on the effect of the variation of @xmath0 during slow - roll inflation on the cmb b - mode polarization , and show how it offers a distinct way to test the modified gravity theories .
we find that the step - like variation of @xmath0 may result in an oscillating modulation to the b - mode polarization spectrum , which can hardly be imitated by adjusting other cosmological parameters .
the intensity of the modulation is determined by the ratio of @xmath0 before and after the variation , which depends on the dynamics of theoretical models .
this oscillating modulation is so rich in feature that it may easily be discriminated from the variation of other parameters or other features .
thus provided that the foreground contribution is under control , high - precision cmb polarization observations will be able to put tight constraint on the variation of @xmath0 , and so the corresponding theories .
we begin with the action for the gws mode @xmath5 , e.g.@xcite@xcite s_(2)=dd^3x a^2 q_t8 , [ action2 ] where @xmath5 obeys @xmath6 and @xmath7 , @xmath8 is regarded as effective planck scale @xmath9 , @xmath0 is the propagating speed of primordial gws , and the prime is the derivative with respect to the conformal time @xmath10 , @xmath11 . during slow - roll inflation , the slow - roll parameter @xmath12 , as well as _
q = q_th q_t1 , s = c_thc_t1are required , e.g. , see @xcite for a recent study . here
, we will mainly focus on the effect of varying @xmath0 , i.e. , the condition @xmath13 might be broke at some point , on primordial gws spectrum . noting that the effects of varying sound speed of primordial scalar perturbations on the scalar power spectrum have been investigated in @xcite@xcite .
we do not get entangled with the details of ( [ action2 ] ) and the evolution of background , and will assume that the background is the slow - roll inflation , which is not affected by the variation of @xmath0 , and set @xmath8 constant and @xmath14 .
we will discuss a possibility of such a case in appendix a. in addition , there may be a mass term @xcite@xcite in ( [ action2 ] ) , which might also be time - dependent @xcite , but we will not involve it .
we can expand @xmath5 into fourier series as @xmath15 , where _ ij(,)=_=+,^()_ij ( ) , [ hij]where the polarization tensors @xmath16 are defined by @xmath17 , @xmath18 , which satisfy @xmath19 , @xmath20 , the commutation relation for the annihilation and creation operators @xmath21 and @xmath22 is @xmath23=\delta_{\lambda\lambda^{\prime}}\delta^{(3)}(\mathbf{k}-\mathbf{k}^{\prime})$ ] .
we define @xmath24 and @xmath25 .
thus we get the equation of motion for @xmath26 as u_k^+(c_t^2 k^2- ) u_k=0.[eom1 ] to phenomenologically investigate the effect on primordial gws spectrum induced by varying @xmath0 , we assume that the variation of @xmath0 can be described by a step - like function c_t=\ { ll c_t1(_0 ) + . , where @xmath27 is the transition time .
we take the background evolution to be the slow - roll inflation .
of course , it is also provided that the sudden change of @xmath0 wo nt affect the background evolution .
then , we have @xmath28 , and the equation of motion ( [ eom1 ] ) becomes u_k^+(c_t^2 k^2- ) u_k=0,[eom2 ] where @xmath29 . the solution to eq.([eom2 ] ) is familiar , we can write it as u_k1&= & , _ 0 , where @xmath30 and @xmath31 are the first and second kind hankel functions of @xmath32-th order , respectively .
these coefficients @xmath33 are functions of the comoving wave number @xmath34 , but constants with respect to conformal time @xmath10 . @xmath35 and @xmath36 are determined by the initial condition . here , we set the initial condition as the standard bunch - davies(bd ) vacuum .
therefore , when @xmath37 , which corresponds to perturbations deep inside the horizon , u_k~e^-ic_t1k.[bd ] when @xmath38 , @xmath39 in eq.([solution ] ) should approximate to eq.([bd ] ) .
allow for the hankel function @xmath40 and @xmath41 when @xmath42 , we get c_1,1=,c_1,2=0 . the coefficients @xmath43 and @xmath44 are determined by requiring @xmath26 and @xmath45 to be continuous at @xmath46 , i.e. the matching condition
. then we obtain c_2,1= & , + c_2,2= & .
[ c21 ] the spectrum of primordial gws is defined by @xmath47 with @xmath48 , which is only a function of comoving wave number @xmath34 . after neglecting the slow - roll parameter , from eq.([solution ] ) with @xmath49
, we have |u_k2|=| c_2,1-c_2,2 | .
therefore , we obtain the power spectrum of primordial gws as p_t & = & _ = + , |h_(,k)|^2 = p^inf_t | c_2,1-c_2,2|^2,[spectrum ] where p^inf_t=2h_inf^2/^2 [ pgr]is that of standard slow - roll inflation without modified gravity , i.e. @xmath50 and @xmath51 , and @xmath52 is the hubble parameter during inflation , which sets the scale of inflation .
the effect of varying @xmath0 is encoded in @xmath43 and @xmath44 .
we set @xmath53 and defined a new function f(k , k_0,x)=| c_2,1-c_2,2|^2 , where @xmath54 .
then , the gws spectrum ( [ spectrum ] ) may be rewritten as p_t = p^inf_t , [ ptt]where @xmath55 is obtained as f(k , k_0,x ) = ^2()+^2 . [ fk ]
we plot @xmath55 with respect to @xmath56 in fig.[fig : v1 ] , in which we set @xmath57 on the left panel and @xmath58 on the right panel , respectively . here , the transition time @xmath59 sets a character scale @xmath60 . when @xmath61 , i.e. the perturbation mode has longer wavelength than @xmath60 , we have @xmath62 , and @xmath63 is scale invariant , which is the result of slow - roll inflation with almost constant @xmath64 and @xmath8 @xcite@xcite .
when @xmath65 , we have f(k , k_0,x ) , thus @xmath55 oscillates between @xmath66 and @xmath67 , and @xmath68 oscillates correspondingly , just as we can see from fig.[fig : v1 ] . , where @xmath69 .
we have set @xmath57 on the left panel and @xmath58 on the right panel , respectively.,scaledwidth=95.0% ] , where @xmath69 .
we have set @xmath57 on the left panel and @xmath58 on the right panel , respectively.,scaledwidth=95.0% ] in the case of varying sound speed @xmath70 of primordial scalar perturbations , the sudden change of @xmath70 may lead to the oscillating modulation to the primordial scalar spectrum , as well as the cmb tt - mode spectrum , just as found in @xcite .
in addition , the oscillation in the primordial scalar spectrum can also be attributed to some other effects , such as inflaton potential with a small oscillation @xcite@xcite , a sudden change in inflaton potential or its derivative , e.g. , @xcite@xcite@xcite@xcite .
thus the oscillation in the primordial scalar spectrum may be implemented without modified gravity , as has been mentioned .
however , the oscillation in the primordial gws spectrum can only be attributed to the modified gravity .
when the gravity is gr , @xmath68 equals to @xmath71 , which is given in eq.([pgr ] ) .
the oscillation of @xmath71 certainly requires @xmath52 is oscillating , which is impossible , unless the null energy condition is violated periodically . though the particle production may also modify the gws spectrum @xcite@xcite@xcite , it only leads to a bump - like contribution , which is entirely different from the behavior of oscillation .
) ( brown line in the left panel and red solid line in the right panel ) and the power - law gws spectrum for reference(blue dashed line in the left panel and green dashed line in the right panel ) .
the insets of the right panels are the tt - mode spectra for our oscillating gws spectrum ( the yellow solid lines ) and the power - law gws spectrum ( the blue dashed lines ) for reference .
we set @xmath72 and @xmath73 . ,
scaledwidth=110.0% ] ) ( brown line in the left panel and red solid line in the right panel ) and the power - law gws spectrum for reference(blue dashed line in the left panel and green dashed line in the right panel ) .
the insets of the right panels are the tt - mode spectra for our oscillating gws spectrum ( the yellow solid lines ) and the power - law gws spectrum ( the blue dashed lines ) for reference .
we set @xmath72 and @xmath73 . , scaledwidth=110.0% ] )
( brown line in the left panel and red solid line in the right panel ) and the power - law gws spectrum for reference(blue dashed line in the left panel and green dashed line in the right panel ) .
the insets of the right panels are the tt - mode spectra for our oscillating gws spectrum ( the yellow solid lines ) and the power - law gws spectrum ( the blue dashed lines ) for reference .
we set @xmath72 and @xmath73 . , scaledwidth=110.0% ] )
( brown line in the left panel and red solid line in the right panel ) and the power - law gws spectrum for reference(blue dashed line in the left panel and green dashed line in the right panel ) .
the insets of the right panels are the tt - mode spectra for our oscillating gws spectrum ( the yellow solid lines ) and the power - law gws spectrum ( the blue dashed lines ) for reference .
we set @xmath72 and @xmath73 .
, scaledwidth=110.0% ] the primordial gws is imprint in cmb as the b - mode polarization .
thus the oscillation in the primordial gws spectrum will inevitably affect the b - mode polarization spectrum . to see such effects
, we plot the cmb bb and tt - mode correlations in fig.[fig : v2 ] , in which @xmath74 in ( [ ptt ] ) is parameterized as ^inf_t = ra_r^inf(kk_*)^n_t^inf . here
, we assume that the scalar spectrum is hardly affected by the modified gravity , which will be clarified in appendix a. thus the scalar perturbation spectrum is set as @xmath75 , in which @xmath76 is the amplitude of the scalar perturbations .
in addition , we also assume that after inflation the propagating speed @xmath0 is unity , so that the spectrum of b - mode polarization is not affected by relevant evolution at late time , or see @xcite,@xcite . ) with different @xmath77(solid lines ) and the power - law gws spectrum with different @xmath78(dashed lines ) .
, scaledwidth=80.0% ] in the left panels of fig.[fig : v2 ] , we see some obvious enhancements or suppressions to the reionization bump in the bb - mode spectrum , which depend on the oscillating effect on corresponding scales . the height of the reionization bump can be estimated roughly as @xcite , c_t , l~2^bb&&1100(1-e^-_ri)^2 c_t , l~2^tt , [ tau]where @xmath79 stands for the tt - mode spectrum from the primordial gws without the reionization and @xmath78 is the optical depth to the beginning of reionization . the periodic enhancements and suppressions of the reionization bump are a reflection of the oscillations of primordial gws spectrum on large scales . in addition , we can also see some obvious oscillations around the recombination peak at @xmath1 . in the right panels of fig.[fig :
v2 ] , we see that the tt - mode spectrum is hardly affected by the oscillating primordial gws power spectrum , since the contribution of gws to tt - mode spectrum is negligible , compared with the scalar perturbations .
the case of ee - mode polarization spectrum is actually also similar .
thus the main effect of varying speed of primordial gws is on the b - mode polarization , which makes the b - mode polarization spectrum show its obvious enhancements or suppressions to the reionization bump and oscillations around the recombination peak . in ref.@xcite ,
the authors have pointed out that it is possible to set to one the propagating speed of gws by a proper redefinition of the metric .
they got the gravitational waves spectrum in their eq.(12 ) same " as the standard one in the form .
however , since they have made a redefinition of the time coordinate and the scale factor in eq.(9 ) , the variation of their @xmath80 with respect to @xmath81 after redefinition is not the same as the variation of @xmath82 with respect to @xmath83 .
therefore , the oscillating feature induced by a step - like @xmath0 is encoded in @xmath80 , the result in both frame should be the same
. it might also be a concern whether such a b - mode polarization spectrum can be imitated by adjusting other cosmological parameters or not . in eq.([tau ] ) , the optical depth @xmath78 is relevant with the height of the reionization bump .
we show the bb - mode spectrum with different @xmath78 in fig.[fig : v3 ] .
we see that the change of @xmath78 can only alter the overall amplitude of bb - mode spectrum at low multipoles , but hardly create the oscillation .
in addition , in inflationary models with pre - inflation era , the reionization bump could also be suppressed ( e.g.@xcite,@xcite,@xcite,@xcite ) or enhanced ( e.g.@xcite ) .
however , similarly , these models also only alter the overall amplitude of the bb - mode spectrum , without oscillation , at low multipoles .
these results indicate that although the bb - mode spectrum may be modified by other ways , the oscillating modulation leaded by varying the speed of primordial gws is difficult to be imitated .
thus the measure of b - mode polarization spectrum provides an appropriate way for testing the corresponding gravity physics in the primordial universe .
in low - energy effective string theory and modified gravity theories , the propagating speed @xmath0 of primordial gws may deviate from unity .
we calculated the spectrum of primordial gws , assuming that @xmath0 has a step - like variation and the background of slow - roll inflation is not affected by it .
we found the spectrum of primordial gws acquires an oscillating modulation , which makes the b - mode polarization spectrum show its obvious enhancements or suppressions to the reionization bump and oscillations around the recombination peak .
the intensity of the modulation is determined by @xmath84 .
the frequency of the modulation is determined by @xmath85 .
both depend on the dynamics of theoretical models .
the oscillating behavior of the b - mode polarization can only be attributed to the effect of modified gravity , since it can hardly be imitated by adjusting other cosmological parameters .
the oscillating behavior is so rich in feature that it may be easily discriminated from the variation of other parameters or other features , thus the upcoming cmb experiments , such as cmbpol , b - pol , will be able to put a tight constraint on the propagating speed of primordial gws , and so the corresponding theories , provided that the foreground contribution is under control . in a certain sense ,
our paper again highlights the significance of b - mode polarization measures in exploring the fundamental physics of primordial universe . here , we only postulate a simple step - like variation of @xmath0 , which , however , might be far complicated in some modified gravity models , as well as accompanied by the variation of @xmath8
. the effect could be non - trivial in more general cases , which is under study .
but in a certain sense , a smooth change of @xmath0 will induce oscillations too , see e.g.@xcite for the case of scalar perturbations , so we expect that the case of gws is similar .
when we focus on the b - mode polarization spectrum , the assumption we adopted is that after inflation the propagating speed @xmath0 is unity , which may also be relaxed .
moreover , it may well be possible to produce oscillatory features beyond the standard slow - roll background
. the varying @xmath0 and @xmath8 will also affect the non - gaussianities of primordial perturbations .
the relevant issues are open .
* acknowledgments * this work is supported by nsfc , no .
11222546 , and national basic research program of china , no .
we acknowledge the use of camb .
in this appendix , we will argue how to realize the change of @xmath0 and the changeless of @xmath8 in ( [ action2 ] ) in low - energy effective string theory and modified gravity theories . in low - energy effective action of string theory ,
the simplest extension of the lowest - order action is e.g.@xcite _ correction~- ( c_1 r_gb^2+c_2 g^ _ _ ) [ action3]where @xmath86 , and @xmath87 is the gauss - bonnet term , @xmath88 is the inverse string tension , @xmath89 is a parameter allowing for different species of string theories , @xmath90 and @xmath91 are coefficients .
we have neglected the terms with @xmath92 and @xmath93 , since both do not contribute to gws .
the introducing of ( [ action3 ] ) will affect not only the gws , but also the adiabatic scalar perturbations , of course , it is interesting to check its effect on the latter .
however , for our purpose , we will regard @xmath94 as the spectator field , so that the effects of ( [ action3 ] ) on the background and the scalar perturbations are negligible .
the action for gws is ( [ action2 ] ) , and @xcite q_t&=&1- ( 8c_1 h_inf - c_2 ^ 2 ) , + c_t^2&= & , see also ref.@xcite for that with @xmath95 .
when @xmath96 is imposed , @xmath50 .
then , we have c_t^2=1 - 4c_1 ^(h_inf+ ) .
the step - like variation of @xmath97 requires that @xmath98 has the step at @xmath46 . as an example
, we will give a numerical result of the variation of @xmath97 and @xmath94 with respect to cosmological time @xmath83 in the case of @xmath99 . according to refs.@xcite@xcite
, we adopt ( ) = -e^- , c_1=-1 , ^=1 , = - ( ) .
we take the expression of @xmath100 as ( t)=t- , where @xmath101 , and @xmath102 is the gauss error function , @xmath103 and @xmath104 are some constants .
we plot @xmath94 and @xmath105 with respect to cosmological time @xmath83 in fig.[fig : v4 ] .
of course , the variation of @xmath97 could be more complicated than a step - like evolution , which would be harder to deal with .
however , as it may , @xmath106 and @xmath107 should be required to avoid ghost and gradient instabilities .
l. amendola , g. ballesteros and v. pettorino , phys .
d * 90 * , 043009 ( 2014 ) [ arxiv:1405.7004 [ astro-ph.co ] ] .
m. raveri , c. baccigalupi , a. silvestri and s. y. zhou , arxiv:1405.7974 [ astro-ph.co ] .
y. s. piao , phys .
d * 75 * , 063517 ( 2007 ) [ gr - qc/0609071 ] ; phys .
d * 79 * , 067301 ( 2009 ) [ arxiv:0807.3226 [ gr - qc ] ] .
i. antoniadis , j. rizos and k. tamvakis , nucl .
b * 415 * , 497 ( 1994 ) [ hep - th/9305025 ] . c. cartier , e. j. copeland and r. madden , jhep * 0001 * , 035 ( 2000 ) [ hep - th/9910169 ] . c. cartier , j. c. hwang and e. j. copeland , phys .
d * 64 * , 103504 ( 2001 ) [ astro - ph/0106197 ] .
y. s. piao , s. tsujikawa and x. m. zhang , class .
grav . * 21 * , 4455 ( 2004 ) [ hep - th/0312139 ] . g. w. horndeski , int .
j. theor .
* 10 * , 363 ( 1974 ) .
l. amendola , phys .
b * 301 * , 175 ( 1993 ) [ gr - qc/9302010 ] . c. deffayet , x. gao , d. a. steer and g. zahariade , phys .
d * 84 * , 064039 ( 2011 ) [ arxiv:1103.3260 [ hep - th ] ] .
j. gleyzes , d. langlois , f. piazza and f. vernizzi , arxiv:1404.6495 [ hep - th ] .
x. gao , phys .
d * 90 * , 081501 ( 2014 ) [ arxiv:1406.0822 [ gr - qc ] ] ; x. gao , phys .
d * 90 * , 104033 ( 2014 ) [ arxiv:1409.6708 [ gr - qc ] ] .
a. de felice and s. tsujikawa , arxiv:1411.0736 [ hep - th ] .
m. nakashima , r. saito , y. i. takamizu and j. yokoyama , prog .
phys . * 125 * , 1035 ( 2011 ) [ arxiv:1009.4394 [ astro-ph.co ] ] .
h. firouzjahi and m. h. namjoo , phys .
d * 90 * , 063525 ( 2014 ) [ arxiv:1404.2589 [ astro-ph.co ] ] .
r. flauger , l. mcallister , e. pajer , a. westphal and g. xu , jcap * 1006 * , 009 ( 2010 ) [ arxiv:0907.2916 [ hep - th ] ] .
a. a. starobinsky , jetp lett . * 55 * , 489 ( 1992 ) [ pisma zh .
55 * , 477 ( 1992 ) ] .
j. a. adams , b. cresswell and r. easther , phys .
d * 64 * , 123514 ( 2001 ) [ astro - ph/0102236 ] .
m. joy , v. sahni and a. a. starobinsky , phys .
d * 77 * , 023514 ( 2008 ) [ arxiv:0711.1585 [ astro - ph ] ] ; m. joy , a. shafieloo , v. sahni and a. a. starobinsky , jcap * 0906 * , 028 ( 2009 ) [ arxiv:0807.3334 [ astro - ph ] ] .
j. liu and y. s. piao , phys .
b * 705 * , 1 ( 2011 ) [ arxiv:1106.5608 [ hep - th ] ] .
j. l. cook , and l. sorbo , phys .
d * 85 * , 023534 ( 2012 ) [ arxiv:1109.0022 [ astro-ph.co ] ] .
l. senatore , e. silverstein and m. zaldarriaga , jcap * 1408 * , 016 ( 2014 ) [ arxiv:1109.0542 [ hep - th ] ] .
s. mukohyama , r. namba , m. peloso and g. shiu , jcap * 1408 * , 036 ( 2014 ) [ arxiv:1405.0346 [ astro-ph.co ] ] .
s. saito , k. ichiki and a. taruya , jcap * 0709 * , 002 ( 2007 ) [ arxiv:0705.3701 [ astro - ph ] ] .
p. creminelli , j. gleyzes , j. nore?a and f. vernizzi , phys .
lett . * 113 * , no .
23 , 231301 ( 2014 ) [ arxiv:1407.8439 [ astro-ph.co ] ] .
a. achucarro , v. atal , b. hu , p. ortiz and j. torrado , phys .
d * 90 * , no . 2 , 023511 ( 2014 ) [ arxiv:1404.7522 [ astro-ph.co ] ] . k. feng , t. qiu and y. s. piao , phys .
b * 729 * , 99 ( 2014 ) [ arxiv:1307.7864 [ hep - th ] ] . | in low - energy effective string theory and modified gravity theories , the propagating speed @xmath0 of primordial gravitational waves may deviate from unity .
we find that the step - like variation of @xmath0 during slow - roll inflation may result in an oscillating modulation to the b - mode polarization spectrum , which can hardly be imitated by adjusting other cosmological parameters , and the intensity of the modulation is determined by the dynamics of @xmath0 .
thus provided that the foreground contribution is under control , high - precision cmb polarization observations will be able to put tight constraint on the variation of @xmath0 , and so the corresponding theories . |
in recent years there has been a surge in experimental studies of dipole excitations lying on the low - energy tail of the isovector giant dipole resonance , the so - called pygmy dipole resonance ( pdr ) .
the pdr has been interpreted as an exotic mode of excitation due to the motion of a weakly bound neutron excess against an almost inert proton - neutron core @xcite , although single particle - hole excitations are also considered @xcite .
one major reason for the renewed interest in the pdr is the possibility of carrying out high - resolution measurements on these low - lying dipole excitations using heavy ion @xcite , proton @xcite , and @xmath1 inelastic scattering experiments @xcite .
an experimental technique , combining particle and @xmath3-ray detection techniques , to study the response of dipole excitations to isoscalar probes was pioneered by poelhekken _
_ @xcite and applied in several studies since @xcite .
these experiments provide complementary information to those obtained from ( @xmath3 , @xmath4 ) experiments which investigate the isovector nature of the excitations @xcite .
one of the surprising results from recent experiments is the isospin splitting of the pdr @xcite .
this provides intimate knowledge about the isospin nature of these excitations which would not be possible to infer from ( @xmath3 , @xmath4 ) experiments alone .
these experimental discoveries were followed by intensive theoretical investigations @xcite . incidentally , scattering experiments with isocalar probes for the study of the pdr have so far been limited to only certain regions of the nuclear chart and carried out mainly on nuclei with large neutron - to - proton ratios @xcite .
information on how the results from scattering reactions compare to those of ( @xmath3 , @xmath4 ) experiments in nuclei closer to @xmath61 are also becoming available @xcite .
since most of the incident isoscalar probes are sensitive to the surface of the nucleus , the information gathered advances our understanding of the evolution of the pdr with changing @xmath7 .
this information is extrapolated for obtaining better estimates of the total strength exhausted by the pdr in nuclei of astrophysical importance , many of which are still inaccessible with the available experimental facilities and techniques .
the pdr has been suggested to have a significant impact on neutron capture rates and isotopic solar abundance distributions in r - process nucleosynthesis @xcite .
further , the pdr could possibly constrain the equation - of - state of hot and dense neutron matter as found in neutron star remnants @xcite . in this contribution , we present results on @xmath0ge where a high - resolution measurement was carried out using the @xmath1 inelastic scattering reaction . in its ground state ,
@xmath0ge is a moderately deformed prolate nucleus @xcite with @xmath8 1.32 . for comparison and to facilitate the discussion , information about the @xmath5 strength distribution is also available from ( @xmath3 , @xmath4 ) data in @xmath0ge @xcite .
the experiment was performed at the separated sector cyclotron facility at ithemba labs with the african omnipurpose detector for innovative techniques and experiments ( afrodite ) @xmath3-ray detector array @xcite in conjunction with two identical particle - telescopes , each of them consisting of two silicon detectors ( in @xmath9 configuration ) .
the @xmath10particles with a beam energy of 48 mev impinged on a 500 @xmath11g/@xmath12 thick @xmath0ge target to populate excited states in the inelastic scattering reaction .
the experiment was carried out over a period of five days with an average beam current of @xmath13 14 particle na .
the telescopes were placed at an angle of @xmath1445@xmath15 with respect to the beam axis .
the dimensions of the w1-type double sided silicon strip detectors @xcite were 5 cm @xmath16 5 cm and consisted of 16 parallel and perpendicular strips 3 mm wide .
the distance from target to the telescopes was 5 cm yielding an angular range of @xmath17 to @xmath18 in the laboratory frame of reference .
thicknesses of the @xmath19__e _ _ and _ e _ detectors were 284 and 1000 @xmath11 m , respectively , and to suppress @xmath20 electrons an aluminum foil of 4.1 mg/@xmath12 areal density was placed in front of the @xmath19__e _ _ detectors .
calibration of individual strips of the silicon detectors was performed using a @xmath21th @xmath1 source .
afrodite , at the time of the experiment , consisted of nine clover hpge detectors with four detectors at 135@xmath22 and five at 90@xmath22 at a distance of 19.6 cm from the target .
the detectors were calibrated using standard @xmath23eu and @xmath24co sources .
high @xmath3-ray energy efficiency parameters for the afrodite array were available from ref . @xcite .
xia digital electronics @xcite was used to acquire timestamped online data in singles mode .
from the timestamped data , events with single , double , and higher fold coincidences were constructed with an offline coincidence time window of 600 ns . from double fold events ,
the @xmath25 coincidences were extracted by placing a gate on the @xmath10particles in the particle identification spectrum .
a projection of @xmath25 coincidences onto the @xmath10particle axis is shown in fig .
[ fig : particle - proj ] .
the selection of correlated events was made with a coincidence time of less than 140 ns by placing appropriate gates around the prompt time peak .
uncorrelated event contributions were extracted and subtracted from the data by placing off - prompt time gates to the early and late sides of the prompt timing peak .
spectrum of @xmath1 particles detected in coincidence with @xmath26rays .
solid and dashed curves are representing data with and without the subtraction of uncorrelated events , respectively .
visible peaks ( solid curve ) are strongly populated discrete states in @xmath0ge .
s@xmath27 indicates the location of the neutron separation energy . ]
kinematic corrections due to the recoil energy of @xmath0ge and the energy losses of scattered @xmath10particles in the target and aluminum foils were applied to the @xmath10particles .
although the target contained some oxygen and carbon contaminants , the recoil corrections for the scattered @xmath10particles from @xmath0ge are quite different compared to those of light contaminant nuclei , thereby allowing a clean extraction of the events of interest .
for instance , the corrections from @xmath0ge versus @xmath28o differ by @xmath13 1 mev and @xmath13 10 mev at @xmath17 and @xmath18 detection angles , respectively .
the energy resolution of the @xmath9 telescopes , measured from the elastic peak , was @xmath29 250 kev . despite the low velocities of the @xmath0ge
recoils , corrections for doppler effects of the high - energy @xmath26rays were found to be necessary and useful .
transitions ( @xmath30 ) to the ground state were extracted with the condition @xmath31 130 kev imposed on the @xmath25 coincidence events , where @xmath32 refers to the excitation energy of the decaying state and is determined from the energy of the scattered @xmath10particles . placing this stringent energy requirement upon the data , together with the differences in kinematic properties
ensures that only transitions from @xmath0ge are extracted , eliminating contributions due to contaminants .
additionally , various combination of angles between the direction of the recoiling nuclei ( as defined by the @xmath1 particles detected in the particle telescope ) and the @xmath3-rays detected in the clover detectors were used for the determination of angular distributions .
the spectrum of direct @xmath3-ray transitions to the ground state is shown in fig . [
fig : gamma - spectrum ] , where in addition to many states for @xmath33 6 mev , a high concentration of states / strength is also observed for 6.5 @xmath34 8 mev .
although the overall sensitivity to high - energy transitions is relatively poor , many transitions observed in ( @xmath3 , @xmath4 ) experiments @xcite can also be clearly identified in the present data .
unresolved strength was separated from intensities of individual transitions by simultaneously fitting the peaks using the root analysis package @xcite in a 16 kev per channel compressed @xmath26ray spectrum .
the unresolved , underlying intensity for 6.5 @xmath34 8 mev amounts to @xmath35 .
comparisons with the recent ( @xmath3 , @xmath4 ) measurement @xcite reveal several states , which were not populated in the ( @xmath1 , @xmath2@xmath3 ) reaction , but are observed in the ( @xmath3 , @xmath4 ) measurement .
however , the states at @xmath36 6850 and 7060 kev are populated only through the ( @xmath1 , @xmath2@xmath3 ) reaction .
( color online ) spectrum of @xmath3-ray transitions decaying directly to the ground state from defined excitation energies .
blue and red spectra correspond to correlated and uncorrelated events , respectively .
inset : the lower energy part of the spectrum where the arrow indicates the position of the unobserved 3558-kev transition , known from ( @xmath3 , @xmath4 ) experiments @xcite . ] the multipole nature of the high - energy transitions was determined through angular distribution measurements , shown in fig .
[ fig : ang_dist ] .
because of the paucity of the data , the angular distribution was extracted simultaneously for the total ( resolved and unresolved ) @xmath3-ray strength in the interval 6.5 @xmath34 8 mev . for comparison , angular distributions of known dipole ( @xmath30 = 2690 kev and @xmath30 = 3648 kev ) and quadrupole transitions ( @xmath30 = 596 kev ) in @xmath0ge are also included in fig .
[ fig : ang_dist ] .
although the 6.5 @xmath34 8 mev strength does not exhibit a perfect agreement with the expected distribution of a dipole transition , the similarity to the two known dipole transitions strongly supports the overall dipole nature .
natural - parity states are preferentially populated in this reaction @xcite , leading to an assignment of spin - parity @xmath37 to the decaying states .
( color online ) angular distributions of ( a ) the first - excited state @xmath38 596-kev transition , and ( b ) @xmath39 transitions from the known 2690 and 3648 kev states together with the total strength of resolved and unresolved transitions for 6.5 @xmath34 8 mev in @xmath0ge . ] in fig .
[ fig : alpha_compare ] panel ( a ) , relative cross - sections of observed @xmath40 states are plotted and normalized to the 4007 kev state . for comparison , panel ( b ) of fig . [
fig : alpha_compare ] displays relative integrated scattering cross - sections ( @xmath41 ) from ( @xmath3 , @xmath4 ) data @xcite , where the 4007 kev state is taken as the reference once again .
all states for @xmath42 6 mev from the ( @xmath3 , @xmath4 ) data are assumed to have negative parity and are plotted in fig .
[ fig : alpha_compare ] , whereas in both panels only states have been included with known negative parity for @xmath33 6 mev , as deduced from the ( @xmath3 , @xmath4 ) data .
an exception are the states 2690 , 3033 , and 3648 kev with assigned @xmath43 , respectively @xcite . the @xmath44 assignment to the 3648 kev state
is based on a polarization measurement @xcite .
however , this state has also been observed in an earlier ( @xmath1 , @xmath2 ) work @xcite . since inelastic @xmath1-scattering populates preferentially natural - parity states ,
the observed strong cross section in the present experiment contradicts this assignment .
hence , the transition is assumed to be electric dipole in character .
similar considerations are applied to the 2690 and 3033 kev states .
the complete absence of the 3558 kev state in the present data ( see arrow in inset of fig . [
fig : gamma - spectrum ] ) is noteworthy , since this state has been observed in the ( @xmath3 , @xmath4 ) work and was assigned @xmath45 @xcite .
( color online ) in panel ( a ) relative cross - sections of @xmath5 transitions from the ( @xmath1 , @xmath2@xmath3 ) data are plotted , while in panel ( b ) the relative integrated scattering cross - sections @xmath41 obtained from ( @xmath3 , @xmath4 ) data @xcite are shown .
numbers next to some transitions indicate the total value of relative cross section . in panel
( a ) , the sensitivity limit is shown by the black solid curve and was determined using the procedure outlined in ref .
uncertainties on the cross - sections in panel ( a ) are @xmath46 for weakly populated states and decrease to @xmath47 for strongly populated states . ]
the comparison shows the presence of two different regions in the energy range of the investigated dipole excitations . in the lower part ( 3 @xmath48 6 mev ) the excitations due to ( @xmath1 , @xmath2@xmath3 )
are enhanced compared to the upper part ( 6 @xmath48 9 mev ) . for ( @xmath3 , @xmath4 )
excitations the trend is reversed indicating a dominant isovector nature of the higher - energy dipole excitations .
this reduction in relative cross - section in the ( @xmath1 , @xmath2@xmath3 ) data becomes even more pronounced if the intensity of the 3648-kev state is taken as a normalization reference .
the reduction of cross - sections in the ( @xmath1 , @xmath2@xmath3 ) data for states @xmath49 6 mev , compared to cross - sections for @xmath50 6 mev , is larger than observed in previous cases . indeed , with respect to ( @xmath1 , @xmath2@xmath3 ) studies on @xmath51ce , @xmath52ba and @xmath53sn @xcite , the isoscalar response at low energies ( @xmath54 6 mev ) is much stronger .
the current result shows that many of the dipole excitations in the 6 @xmath34 9 mev range in @xmath0ge are mixed with larger isovector components .
however , a few weakly populated pure isoscalar , as well as several pure isovector states are found for @xmath55 mev .
these results indicate that the dipole excitations in @xmath0ge for @xmath49 6 mev do show the common scenario of dipole excitations splitting in two distinct parts : one at lower energy , whose states have a strong isospin mixing , and one at higher energy with predominant isovector character .
we have performed calculations of the dipole transition densities in @xmath0ge within the relativistic quasiparticle time blocking approximation ( rqtba ) @xcite based on the covariant energy density functional theory ( cedft ) @xcite .
the rqtba has been developed to include spreading mechanisms , other than landau damping ( one particle - one hole ( 1p1h ) , or two - quasiparticle ( 2q ) , configurations ) into the microscopic description of nuclear excitation modes within the relativistic framework .
the existing versions of rqtba include 2q @xmath56 phonon @xcite or two phonon @xcite configurations in a fully self - consistent way .
parameters ( in the present version with the nl3@xmath57 @xcite interaction - 8 parameters ) of the cedft were fixed by fitting masses and radii of several characteristic nuclei throughout the nuclear chart @xcite and no adjustments were involved in the subsequent calculations .
the calculations were performed in the following three steps : ( i ) the single - particle spectrum was obtained from the self - consistent relativistic mean - field solution ; ( ii ) the phonon spectrum was computed by the self - consistent relativistic quasiparticle random phase approximation ( rqrpa ) and ( iii ) the bethe - salpeter equation for the nuclear dipole response was solved within the rqtba employing the rqrpa phonons to construct the induced energy - dependent residual interaction .
the low - energy region of the dipole spectrum is calculated with the rqtba .
it includes mixing of quasiparticles with phonons , in particular , with the lowest 2@xmath58 collective state obtained in rqrpa at @xmath590.6 mev and the lowest 3@xmath60 state at @xmath593.4 mev , while without mixing there is no dipole strength at the energies of interest .
the phonon spectra are consistent with experimental observations for the first - excited 2@xmath58 and 3@xmath60 states at 596 and 2536 kev @xcite .
reduced transition probabilities from rqtba calculations with 25 kev smearing ( bunching ) for isoscalar and isovector dipole operators are plotted in panels ( a ) and ( b ) of fig .
[ fig : calculations_compare ] .
although these calculations also suggest a suppression in the isoscalar @xmath5 strength at higher energies , they underestimate the experimentally observed suppression in @xmath0ge .
( color online ) reduced transition probabilities in @xmath0ge from rqtba calculations plotted for the isoscalar ( a ) and electromagnetic ( isovector ) ( b ) dipole operators . ]
( color online ) transition densities for two calculated rqtba states at @xmath32=4.55 ( panel ( a ) ) and 7.05 mev ( panel(b ) ) in @xmath0ge . ]
figure [ fig : td ] shows the proton , neutron , isoscalar and isovector transition densities for calculated states at @xmath36 4.55 and 7.05 mev .
the lower - lying state ( panel ( a ) ) exhibits the usual pattern for an almost pure isoscalar dipole state , with the proton and neutron transition densities in phase inside the nucleus and at the nuclear surface .
consequently , the isoscalar transition density has a pattern typical of the compressional mode with a node close to the nuclear surface .
in contrast , the higher - lying state ( panel ( b ) ) exhibits the typical behavior of a pygmy dipole state where the proton and neutron transition densities are in phase inside the nucleus , while at the surface region the contribution comes from the neutron density only .
consequently , at the surface the isoscalar and isovector transition densities have the same intensity giving rise to a strong isospin mixing . for the calculated dipole states
this behavior is supported by the present data , which manifests significant isospin mixing in the energy region under investigation .
( color online ) cross sections of inelastic scattering of @xmath10particles are plotted as a function of scattering angle in the center of mass frame for the 2@xmath58 state in @xmath0ge at 596 kev and dipole states at ( a ) 4.55 and ( b ) 7.05 mev .
the blue shaded areas represent the angular coverage of scattered @xmath10particles in the present measurement . ] an estimate of the inelastic cross section of states due to different reaction mechanisms is obtained using the talys 1.6 reaction code @xcite .
these calculations suggest that for a 48 mev @xmath1 beam , the compound reaction does not contribute at any excitation energy under consideration , while for @xmath61 6 mev the contribution from pre - equilibrium reactions is an order of magnitude less than that from direct reactions and gradually increases with @xmath32 .
therefore , a direct comparison with other experimental data should be taken with some degree of caution at the highest excitation energies . in principle
, the presence of the coulomb interaction between the target and projectile has the capability to substantially contribute to the observed cross - sections @xcite . to investigate the effect of the coulomb interaction on the observed inelastic cross sections , theoretical cross sections were calculated both with and without taking the coulomb interaction into account .
these theoretical cross sections were obtained for the dipole states at @xmath32 = 4.55 and 7.05 mev by performing distorted wave born approximation ( dwba ) calculations , carried out using the fresco code @xcite .
the radial nuclear form factors were constructed within a double folding procedure using the microscopic transition densities of fig .
[ fig : td ] , see ref .
@xcite for more details on the procedure . for the coulomb form factors we have used the analytic expression built inside the fresco code . for these calculations the double folding potential was used as the real part of the optical potential , while for the imaginary part the same geometry as for the real part but with half the intensity was chosen @xcite .
these results are shown in fig .
[ fig : coulomb_effect_multipole ] where a negligible difference between the calculations performed using only the nuclear interaction ( red curve ) and using both the nuclear and coulomb interaction ( blue curve ) is observed for the detection angles under study ( blue shaded areas ) . for these low - lying dipole states
it has been shown that the nuclear and coulomb contributions interfere constructively in the nuclear surface region @xcite .
this feature is expected not to be visible for this relatively low incident energy since the coulomb contribution becomes important as the beam energy increases towards 30 mev / u @xcite .
we are aware of the fact that , while the relation between the inelastic cross section and the @xmath62 is clear for the coulomb excitation ( they are proportional ) , the relation between the isoscalar response and the inelastic excitation cross section due to an isoscalar probe is not so evident .
in fact , the ratio between the @xmath63 of the two states at 4.55 and 7.05 mev is 2.2 while the ratio between the corresponding values of the cross sections is 6.4 at the first maximum .
if we eliminate the effect of the q - value , by placing the two states at the same energy , then the ratio decreases to 4.1 , still far from 2.2 .
however , in ref .
@xcite a calculation of the cross section was presented in the framework of a semiclassical model , that provides the missing link to directly compare the results from the microscopic rqtba calculations to experimental data measured via the ( @xmath1 , @xmath2@xmath3 ) reaction , confirming the structural splitting of the low - lying @xmath5 strength .
it is instructive to also have an estimate of the cross - section of states with higher multipolarities .
therefore , we also performed calculations for the first - excited 2@xmath58 state in @xmath0ge , using a collective macroscopic nuclear form factor .
the @xmath64 value of the 596 kev transition is taken to be 3050 @xmath65 from ref .
@xcite with a deformation length of 1.43 fm .
the results are shown in fig .
[ fig : coulomb_effect_multipole ] , where the cross - sections for the 2@xmath58 state ( black curve ) are significantly higher when compared to the dipole states .
this is not only the case for the detection angles of the present experiment , but also for very forward angles .
it is interesting to point to a recent measurement of the photon strength function below the neutron separation energy in @xmath0ge @xcite , using the so - called oslo method . despite the limited @xmath3-ray detection resolution , a broad structure
is observed in the 6 @xmath66 8 mev range .
it is highly probable that this feature is the same pygmy dipole resonance structure as observed in this work .
we provide new results , which indicate a suppression in relative cross section for the excitation of the pdr in @xmath0ge populated through inelastic @xmath10scattering , when compared to photon scattering data for @xmath67 mev .
the observed dipole response splits into two distinct parts : one at lower energy , with excitations that have strong isospin mixing , and one at higher energy with predominant isovector character .
the results are particularly important in improving our understanding of the emergence and persistence of the pdr for low @xmath7 nuclei . as such , measurements in other mass regions are undoubtedly necessary to fully understand the evolution of the pdr from near - isospin saturated systems towards nuclei with large @xmath7 ratios .
finally , the present work highlights the importance of using complementary probes to photon scattering , in order to reveal detailed information about the underlying nature of dipole excitations .
the authors would like to thank the operational staff at ithemba labs for providing excellent beam quality throughout the experiment and lawrence berkeley national laboratory for making available the @xmath0ge target .
this work was supported by the national research foundation of south africa under grant no .
92789 , and no . 93500
; by the research council of norway , project grants no .
205528 , no . 213442 , and no .
210007 ; by us - nsf grants no .
phy-1204486 and no .
phy-1404343 ; by the us department of energy under contracts no .
de - ac52 - 07na27344 , and no . de - ac02 - 05ch11231 ; and by erc - stg-2014 grant no .
637686 .
60 n. paar , d. vretenar , e. khan , and g. col , rep .
phys . * 70 * , 691 ( 2007 ) .
d. savran , t. aumann , and a. zilges , prog .
nucl . phys . * 70 * , 210 ( 2013 ) .
a. bracco , f. c. l. crespi and e. g. lanza , eur .
j. a * 51 * , 99 ( 2015 ) .
a. m. lane , ann .
phys . * 63 * , 171 ( 1971 ) .
p. -g .
reinhard and w. nazarewicz , phys .
c * 87 * , 014324 ( 2013 ) .
l. pellegri _ et al .
_ , phys . lett .
b * 738 * , 519 ( 2014 ) .
f. c. l. crespi _ et al .
c * 91 * , 024323 ( 2015 ) .
i. poltoratska _ et al .
c * 85 * , 041304(r ) ( 2012 ) . a. m. krumbholz _
et al . _ ,
b * 744 * , 7 ( 2015 ) .
d. savran _ et al .
lett . * 97 * , 172502 ( 2006 ) .
j. endres , _ et al .
c * 80 * , 034302 ( 2009 ) .
t. d. poelhekken _ et al .
_ , phys . lett .
b * 278 * , 423 ( 1992 ) .
j. endres , _ et al .
lett . * 105 * , 212503 ( 2010 ) .
v. derya _ et al .
_ , j. phy .
* 366 * , 012012 ( 2012 ) .
v. derya _ et al .
_ , nucl . phys . * a906 * , 94 ( 2013 ) .
v. derya _ et al .
_ , phys . lett .
b * 730 * , 288 ( 2014 ) .
f. c. l. crespi _ et al .
lett . * 113 * , 012501 ( 2014 ) .
s. volz _ et al .
a779 * , 1 ( 2006 ) .
k. govaert _ et al .
c * 57 * , 2229 ( 1998 ) . c. romig
_ et al . _ ,
c * 88 * , 044331 ( 2013 ) .
t. shizuma _ et al .
c * 78 * , 061303 ( 2008 ) .
r. schwengner _
et al . _ ,
c * 76 * , 034321 ( 2007 ) .
n. benouaret _ et al .
_ , phys . rev .
c * 79 * , 014303 ( 2009 ) .
a. makinaga _ et al .
c * 82 * , 024314 ( 2010 ) .
n. tsoneva , and h. lenske , phys .
c * 77 * , 024321 ( 2008 ) .
n. paar _ et al .
rev . lett . * 103 * , 032502 ( 2009 ) .
m. martini , s. p ' eru , and m. dupuis , phys .
c * 83 * , 034309 ( 2011 ) .
m. tohyama , and t. nakatsukasa , phys .
c * 85 * , 031302(r ) ( 2012 ) .
d. vretenar _ et al .
_ , phys . rev .
c * 85 * , 044317 ( 2012 ) .
h. nakada , t. inakura , and h. sawai , phys .
c * 87 * , 034302 ( 2013 ) . s. goriely , phys . lett .
b * 436 * , 10 ( 1998 ) .
s. goriely , e. khan , and m. samyn , nucl .
a739 * , 331 ( 2004 ) .
e. litvinova _ et al .
_ , nucl . phys . *
a823 * , 26 , ( 2009 ) .
i. daoutidis , and s. goriely , phys .
c * 86 * , 034328 ( 2012 ) .
j. piekarewicz , phys .
c * 73 * , 044325 ( 2006 ) .
j. piekarewicz , phys .
c * 83 * , 034319 ( 2011 ) .
l. rosier , and e. i. obiajunwa , nucl .
a * 500 * , 323 ( 1989 ) .
p. mller
_ et al . _ , at .
data nucl .
data tables 94 , 758 ( 2008 ) .
a. jung _ et al .
_ , nucl . phys . *
a584 * , 103 ( 1995 ) .
r. massarczyk _ et al .
c * 92 * , 044309 ( 2015 )
. m. lipoglavek _ et al .
. instr . meth .
a * 557 * , 523 ( 2007 ) .
micron semiconductor , product cataloque , http://www.micronsemiconductor.co.uk/pdf/w1.pdf http://www.xia.com https://root.cern.ch w. w. eidson and j. g. cramer , jr .
* 9 * , 497 ( 1962 ) .
b. sch " urmann _ et al .
_ , nucl . phys . *
a475 * , 361 ( 1987 ) .
e. litvinova , p. ring , and v. tselyaev , phys .
c * 78 * , 014312 ( 2008 ) .
p. ring , prog .
37 * , 193 ( 1996 ) .
d. vretenar _ et al .
_ , phys . rep .
* 409 * , 101 ( 2005 ) .
e. litvinova , p. ring , and v. tselyaev , phys .
lett . * 105 * , 022502 ( 2010 ) .
e. litvinova , p. ring , and v. tselyaev , phys .
c * 88 * , 044320 ( 2013 ) .
g. a. lalazissis _
et al . _ ,
b * 671 * , 36 ( 2009 ) .
http://www.nndc.bnl.gov ( as of april 2015 ) .
a. j. koning
_ , _ nuclear data for science and technology _ , edited by o. bersillon _
( edp sciences , nice , france , 2008 ) , p. 211
( see also http://www.talys.eu ) .
e. g. lanza _ et al .
_ , phys . rev .
c * 89 * , 041601(r ) ( 2014 ) .
i. j. thompson , comp .
phys . rep .
* 7 * , 167 ( 1988 ) ; http://www.fresco.org.uk e. g. lanza , a. vitturi , and m. v. andrs , phys .
c * 91 * , 054607 ( 2015 ) .
e. g. lanza _ et al .
c * 84 * , 064602 ( 2011 ) .
r. leconte _ et al .
c * 22 * , 2420 ( 1980 ) .
t. renstrm _ et al .
c * 93 * , 064302 ( 2016 ) . | isospin properties of dipole excitations in @xmath0ge are investigated using the ( @xmath1 , @xmath2@xmath3 ) reaction and compared to ( @xmath3 , @xmath4 ) data .
the results indicate that the dipole excitations in the energy region of 6 to 9 mev adhere to the scenario of the recently found splitting of the region of dipole excitations into two separated parts : one at low energy being populated by both isoscalar and isovector probes and the other at high energy , excited only by the electromagnetic probe .
rqtba calculations show a reduction in the isoscalar @xmath5 strength with an increase in excitation energy which is consistent with the measurement . |
the study of synchronization of coupled oscillators has attracted the attention of widely diverse research disciplines such as neuroscience @xcite , physics @xcite , mathematics @xcite and engineering @xcite . since the seminal works of winfree @xcite and kuramoto @xcite , the kuramoto model has served as a canonical model for synchronization that can capture a quite rich dynamic behavior including multiple equilibria , limit cycles , and even chaos .
there are two main properties that characterize its behavior .
the first one is the coupling function , which is a trigonometric @xmath0 function in the case of the kuramoto model .
however , a broader class of the coupling functions has also been studied @xcite .
the second property , and perhaps the most important one , is the interconnection topology .
the most popular assumption is that all oscillators are connected to each other , which corresponds to a fully connected graph @xcite .
although a much more general approach is to study the systems of oscillators with arbitrary underlying topology @xcite . due to its complex behavior , several assumptions are usually made to make the study tractable .
for example , one can make the number of oscillators go to infinity , and use statistical mechanics tools to characterize its convergence . or one can assume that all oscillators have equal intrinsic frequencies , and therefore form a ( gradient ) system of homogeneous oscillators @xcite that has globally convergent properties . alternatively , as we do in this article , one may let the frequencies take distinct values @xcite and characterize sufficient conditions for synchronization . in this paper , we consider a system of finite number of heterogeneous kuramoto oscillators with general underlying topology . we show that when certain conditions on the coupling strength and initial phases are satisfied , the trajectories are bounded , which for kuramoto oscillators also implies synchronization .
the most relevant previous work is @xcite , where the authors studied the same setup . here
, we build upon @xcite to obtain less restrictive conditions for synchronization .
in particular , by using a novel computational - analytical approach , we are able to further exploit the particular features of each problem instance and outperform existing results .
several examples are used to illustrate our findings and characterize the scaling behavior of our conditions .
we consider a system of oscillators which are described by a kuramoto model , i.e. the behavior of each oscillator is governed by the following equation : @xmath1 where @xmath2 is a set of oscillators connected to oscillator @xmath3 , i.e. the set of its neighbors , @xmath4 is the coupling strength , which assumed to be the same for all connections , and @xmath5 is the total number of oscillators in the system .
we also assume that the oscillators are heterogeneous in intrinsic frequencies .
this means that their intrinsic frequencies @xmath6 are not necessary equal . frequencies , however , do not change their values with time , so each @xmath6 is a constant . in this article
we study frequency synchronization of the system .
the oscillators achieve synchronization if @xmath7 as @xmath8 , where @xmath9 is a constant common phase velocity .
it is easy to show that this common phase velocity @xmath9 is an average sum of intrinsic frequencies of the oscillators .
that is , @xmath10 indeed , when @xmath11 , adding up all the equations of gives : @xmath12 , because each @xmath13 is added to @xmath14 and gives zero .
we denote the average natural frequency by @xmath15 , and define the deviations of the natural frequencies by + @xmath16 , where @xmath17 . from now on we will study the following system instead of system : @xmath18 each limit cycle of system is an equilibrium of .
therefore , we will focus on finding conditions when system converges to an equilibrium , i.e. when @xmath19 . due to the rotational invariance of ,
we can shift without loss of generality , all the initial phases @xmath20 by the same value @xmath21 so that their sum becomes equal to zero : @xmath22 .
furthermore , since the phase average remains the same ( for system : @xmath23 ) , condition @xmath24 will be satisfied @xmath25 , where @xmath26 are the trajectories of system . in the rest of this article
we will assume that the phase values sum up to zero at each time @xmath27 . in this article
we show synchronization of oscillators by providing a lyapunov function and using lasalle s invariance theorem @xcite .
when all the intrinsic frequencies are equal , i.e. deviations @xmath28 , the oscillators are called homogeneous , and the following lyapunov function can be employed : @xmath29 where @xmath30 and @xmath31 is the edge set of a given graph .
it is easy to check that @xmath32 thus , since the function @xmath33 is @xmath34-periodic on each element @xmath35 in @xmath36 , it is also well - defined on a n - dimensional torus ( @xmath37 ) , which is compact .
thus , applying the lasalle s invariance theorem ( on @xmath37 ) guarantees synchronization of the oscillators .
when the intrinsic frequencies are not equal , we have a system of heterogeneous oscillators , and we still can provide a potential function for this case : @xmath38 we can check again that the time derivative of this function is also non - positive and equal to zero only at an equilibrium , i.e. when the frequencies are synchronized .
the problem here is that function @xmath39 is not bounded from below and can not be defined on the @xmath5-dimensional torus @xmath40 , and therefore , we are not able to apply directly the lasalle s invariance theorem . however ,
if we show that the trajectories @xmath41 of are bounded , then the function @xmath39 is bounded as well , and hence synchronization follows . therefore , the main goal of this study is to find the conditions that guarantee that trajectories are bounded .
we will achieve this by finding a compact positively invariant set ( pis ) for the oscillators phases .
that is , a compact ( closed and bounded ) set such that if the system s initial conditions are within this set , the trajectories will remain in the set .
the next section shows that when some conditions are met , such pis exists , and therefore system will converge to the set of equilibria .
this section is organized as follows .
we first formulate in proposition 1 a general sufficient condition for boundedness of the trajectories that leads to synchronization of system .
we then provide two solutions that guarantee fulfillment of proposition 1 .
our first solution , described in subsection @xmath42 , contains explicit requirements on the coupling strength and initial oscillators phases .
this solution is further refined using computational tools in subsection @xmath43 .
we will denote maximum and minimum phase values at time @xmath44 by @xmath45 and @xmath46 , where @xmath47 is a phase of oscillator @xmath3 at time @xmath44 .
let @xmath48 be defined as a maximum phase difference between two oscillators at time @xmath44 @xmath49 , i.e. @xmath50 then @xmath51 ( @xmath52 ) .
in other words , each phase lies between the minimum and maximum phases @xmath53 and @xmath54 .
the maximum initial ( at time @xmath55 ) pairwise phase difference is denoted by @xmath56 : @xmath57 if we can show that the maximum phase difference is always bounded , i.e. if @xmath58 , where @xmath59 is a constant satisfying @xmath60 , then the trajectories will be also bounded since the phase average remains the same .
the pis therefore is defined through the maximum phase difference that is bounded by the value of @xmath59 , i.e. @xmath61 which is obviously a compact .
we can now formulate a general condition that is sufficient to guarantee that the maximum phase difference is always bounded by a constant @xmath59 and thus the trajectories are also bounded .
* proposition 1 * _ if @xmath59 is a constant satisfying @xmath60 , and for all times @xmath62 such that @xmath63 , the following condition is satisfied : @xmath64 for every two oscillators @xmath65 and @xmath66 such that @xmath67 and @xmath68 , then the maximum phase difference is bounded by @xmath59 , i.e. @xmath69 for all @xmath62 , trajectories of system are bounded , and system achieves frequency synchronization . _ condition says that when the maximum phase difference achieves value @xmath59 , it can not grow anymore and thus does not exceed @xmath59 .
that s why the maximum phase difference will be always bounded by @xmath59 if is satisfied , and because the phase average remains the same , it also ensures that the trajectories of system are bounded in @xmath70 .
since function @xmath71 is well - defined in @xmath70 , we can now apply lasalle s invariance theorem to guarantee that each solution of approaches the nonempty set @xmath72 , and system achieves frequency synchronization .
it is possible that when @xmath73 , several oscillators have phase values equal to @xmath54 or @xmath53 . in this case condition
should be satisfied for each pair of oscillators with a phase difference equal to @xmath59 .
condition is very general by itself and difficult to check . in the rest of this section
we derive two conditions analytic and optimization - based that guarantee condition .
these conditions contain requirements on the coupling strength and initial phases of oscillators that can be verified for each given system .
we now introduce some additional notation .
let @xmath74 i.e. @xmath75 is the squared euclidean norm of a vector of phases at time @xmath44 . for simplicity we will use symbol @xmath76 instead of @xmath75 . at initial time
@xmath55 the value of this function is denoted by @xmath77 .
the euclidean norm of a vector of the natural frequencies deviations is defined as @xmath78 : @xmath79 let the topology of a given system be defined by an undirected graph @xmath80 with a set of nodes @xmath81 such that @xmath82 , and with an edge set @xmath31 . by @xmath83 we denote the set @xmath84 @xmath85 where @xmath86 is the set of @xmath87 edges of a complete graph with @xmath5 nodes .
we use @xmath88 to denote the minimum nodal degree of a graph .
the following two lemmas are based on the results from @xcite and will be used in the next two subsections ; their proofs are provided in appendix .
* lemma 1 * _ _ if @xmath24 , then @xmath89 where @xmath90 .
distance @xmath91 between two nodes @xmath92 in a graph @xmath80 is the number of edges in a shortest path between these two nodes .
* lemma 2 * _ if @xmath93 $ ] , then function @xmath94 satisfies the following differential inequality on @xmath95 $ ] : @xmath96 and if in addition @xmath97 then @xmath76 will be upper bounded by @xmath77 : @xmath98 $ ] .
_ it is assumed that not all initial phases are equal to zero , so that @xmath99 in condition .
lemma 2 states that if the trajectories of system stay in a pis defined by @xmath100 , and the coupling strength @xmath4 satisfies condition , then function @xmath94 does not exceed its initial value .
the main result of this subsection is theorem 1 which contains requirements on the initial phases and coupling strength such that all conditions of proposition 1 are satisfied and thus system achieves frequency synchronization . , @xmath101 and @xmath102 at time @xmath44 when @xmath103 @xmath104 _ if @xmath59 is a constant satisfying @xmath105 ; @xmath106 ; @xmath4 satisfies and @xmath107 where @xmath108 , then there exists a pis defined by @xmath69 for all @xmath27 , and system achieves frequency synchronization . when @xmath109 , only condition for @xmath4 is required .
_ we first consider the case when @xmath110 .
assume that at time moment @xmath62 : @xmath111 , and that before this moment @xmath44 the maximum phase difference has never exceeded @xmath59 .
then if holds , function @xmath76 has not exceeded its initial value @xmath77 .
we define : @xmath112 suppose that @xmath65 and @xmath66 are two oscillators with @xmath67 and @xmath68 . by the definition of @xmath101 :
@xmath114 such that @xmath115 , @xmath116 . on fig .
1 we plotted phases @xmath53 , @xmath54 at time @xmath44 with @xmath63 , and values of @xmath117 . if at time @xmath44 @xmath111 , then because @xmath24 , one of the following possibilities takes place : @xmath118 , or @xmath119 , or @xmath120
. let @xmath121 then @xmath122 therefore , @xmath123 and @xmath124 we want to show that at time @xmath44 the maximum phase difference does not start to increase by showing that condition of proposition 1 is satisfied , i.e. @xmath125 for every two oscillators @xmath65 and @xmath66 with @xmath67 and @xmath68 . using equation , and the fact that @xmath126 we have @xmath127 where follows from and @xmath128 , and from @xmath129 being decreasing and the theorem assumption @xmath106 .
similarly , we have @xmath130 now , from condition of proposition 1 we have @xmath131 where inequality follows from - , , the fact that @xmath0 is an increasing function on @xmath132 , and because @xmath133 , @xmath134 and @xmath135
. we will now show that @xmath136 where the last inequality holds because @xmath137 for simplicity we will introduce the following notation : @xmath138 since due to , @xmath139 , from and - , we have : @xmath140 and @xmath141 consider a function @xmath142 of one scalar argument @xmath143 : @xmath144 where @xmath143 satisfies and @xmath145 is a function of @xmath143 defined by .
we will now show that @xmath146 .
when @xmath147 we get an equality : @xmath148 assume first that @xmath149 .
now it is enough to show that the derivative of this function is positive : @xmath150 for all @xmath151 .
it can be verified , that @xmath152 notice , that @xmath153 \in ( 0,\frac{\pi}{4}]$ ] , and @xmath154 , @xmath155 and @xmath156 since @xmath151 , @xmath157 and @xmath158 .
therefore , @xmath159 now , since @xmath160 , it is sufficient to show that @xmath161 .
indeed , since @xmath162 , @xmath163 , which means that @xmath164 on the other hand , @xmath165 for @xmath166 the proof is similar . therefore ,
@xmath146 , and holds because of and .
notice that even though it is possible that one of the @xmath167 functions @xmath168 or @xmath169 is negative , we demonstrated that their sum is always positive . finally , from using : @xmath170 thus , @xmath125 , if @xmath171 .
when @xmath109 and condition is satisfied , @xmath48 will be always less than @xmath59 .
indeed , if at time @xmath44 @xmath111 , then @xmath172 , @xmath173 and @xmath174 in contradiction to lemma 2 .
therefore , we do not need to have an additional bound on @xmath4 to guarantee that @xmath69 . in theorem @xmath175
we have two conditions and on the lower bound of the coupling strength @xmath4 , thus the theorem will hold when @xmath4 satisfies the largest of these two lower bounds . similarly to the analytic synchronization condition described in a previous subsection ,
the optimization - based condition to be introduced in this subsection also guarantees that requirement of proposition 1 is satisfied .
numerical techniques , however , allow us to improve the analytic synchronization condition .
there are two bounds on the coupling strength @xmath4 in theorem 1 , and we will improve bound using optimization approach .
our optimization approach utilizes additional information that has not been used in the analytic condition , for example , topology information has not been taken into account ( except for the minimum nodal degree ) . for each pair of vertices
we solve an optimization problem posed below and find the lower bound on @xmath4 .
then we choose the maximum bound among these obtained @xmath87 lower bounds on the coupling strength .
condition in proposition 1 is satisfied if @xmath176 we find the minimum possible value of the denominator and then obtain corresponding bound on @xmath4 by .
the phases of oscillators constitute the phase vector @xmath177^{t}$ ] and are the variables of the optimization problem .
the optimization problem is formulated as follows : @xmath178 the first constraint guarantees that the phase distance between oscillators @xmath66 and @xmath65 is exactly @xmath59 .
second constraint requires that the sum of all phases is equal to zero .
third condition is necessary because function @xmath179 does not exceed its initial value when is satisfied .
the last constraint asks for all the phases to be between @xmath180 and @xmath181 .
let @xmath182 denote the value of the coupling strength in found with optimization for oscillators @xmath65 and @xmath66 , and suppose that @xmath183 is the maximum found coupling value among all pairs of oscillators
. then if @xmath184 and condition is satisfied , system achieves frequency synchronization .
we summarize this result in the following theorem : * theorem 2 * _ if @xmath59 is a constant satisfying @xmath185 ; @xmath186 ; if @xmath184 , where @xmath187 is the bound on the coupling strength obtained with optimization , and if @xmath4 satisfies , then there exists a pis defined by @xmath69 for all @xmath62 , and system achieves frequency synchronization . _ conditions @xmath185 and are required by lemma 2 which guarantees that @xmath188 whenever @xmath111 .
condition @xmath184 in turn ensures that requirement is satisfied and @xmath48 does not exceed the value of @xmath59 , which implies that there exists a pis .
condition @xmath186 guarantees that the optimal value of the optimization problem is strictly positive and there exists a finite positive value of @xmath182 that satisfies for each pair of oscillators @xmath65 and @xmath66 .
in addition to an improvement of condition , the numerical method has weaker requirements on the initial phases .
optimization approach can be applied when @xmath185 ( @xmath189 in theorem 1 ) and when @xmath186 ( @xmath106 in theorem 1 ) .
[ cols="^,^,^ " , ] [ table_example ] we compared our results with two existing frequency synchronization conditions .
we do not consider here the type of bounds of @xcite , @xcite and some results of @xcite , since they only provide existence and local stability . to the best of our knowledge
the only two bounds that guarantee the existence of a positively invariant set for arbitrary topologies are theorem 4.6 from @xcite , and @xcite .
therefore , here we compare our condition based on optimization approach , condition from theorem 4.6 in @xcite and condition from @xcite .
each of the three synchronization conditions consists of a bound on the coupling strength and constraints on the initial phases of oscillators . in particular , all synchronization conditions require that the difference between any two initial phases is less than @xmath190 ( i.e. @xmath191 ) .
in addition , each synchronization condition has its own special constraint on the initial phases .
the bounds on the coupling strength and corresponding requirements on the initial phases are summarized in table 1 .
our bound is the maximum between bound and the optimization bound .
the latter in turn is defined as the maximum bound among all pairs of oscillators . in our optimization - based synchronization condition , @xmath59 is a constant whose value can be chosen from the interval @xmath192 . while larger values of @xmath59 make the constraint on the initial phases less restrictive , the bound on the coupling strength is less restrictive on average for smaller values of @xmath59 . in the simulations we assigned a value of @xmath56 to the constant @xmath59 . in
the bound from @xcite , @xmath193 is the algebraic connectivity of a given graph , @xmath194 is the incidence matrix of the complete graph , @xmath195 is a vector of frequencies , @xmath196 vector of initial phases and @xmath197 . in the bound from @xcite , @xmath59 is a constant whose value is defined as @xmath198 , and @xmath199 is defined as @xmath200 , where @xmath201 is the diameter of a graph @xmath202 and @xmath203 is the cardinality of its set @xmath204 . in our analysis
we compare the requirements on both , the initial phases , and the bounds on the coupling strength
. * experiment 1 ( comparison of the constraints on initial phases ) .
* the first part of our comparison analysis examines how restrictive the constraints on the initial phases are .
we created @xmath205 samples of initial phases by choosing the phases randomly from the @xmath206 interval and then subtracting from each sample its mean .
the sum of phases in each sample therefore is equal to zero and maximum pairwise phase difference is not greater than @xmath190 .
we did this for @xmath207 , where @xmath5 is the number of oscillators in a network or the number of entries in each sample .
we then checked for each sample of initial phases , if it satisfies the initial phase constraints .
the results are shown on fig .
[ experiment1 ] , where the x - axis corresponds to the number of oscillators @xmath5 and the y - axis corresponds to a fraction of samples that satisfy the initial phase constraints for each synchronization condition . from this figure
we can see that our constraint on the initial phases is the least restrictive , whereas the constraint corresponding to condition from @xcite is the most restrictive on average . from this result
we may conclude that our synchronization condition has a larger region of applicability when initial phases are randomly chosen from the @xmath206 interval .
* experiment 2 ( comparison of the bounds on coupling strength ) . * [ experiment2 ] 0.3 0.3 0.3 0.3 0.3 0.3 [ experiment2 ] [ experiment2 ] in the second part of our comparison we applied the bounds on the coupling strength only to the samples of the initial phases that satisfy all three initial phase constraints .
we performed the comparison for star - tree ( a tree with one node connected to all others ) , ring and chain topologies with the number of oscillators @xmath5 varying from 4 to 8 .
because all three bounds contain oscillators frequencies , we also had to create random samples of frequencies .
the simulation process was organized as follows : we first fixed the underlying topology , that is the topology type ( star - tree , ring or chain ) together with the number of oscillators @xmath5 .
after that , for each topology we generated @xmath208 pairs of random samples of frequencies and initial phases .
then , for each pair of samples of frequencies and initial phases we calculated values of three bounds on the coupling strength .
we then found average values of the three bounds over @xmath208 samples for each topology .
we also calculated for each topology the fraction of samples for which our bound outperforms bounds from @xcite and from @xcite .
while all three bounds can be applied only when the initial maximum phase difference @xmath56 is from the interval @xmath209 , there is no similar requirement on the frequencies . to figure out if an interval from which the frequencies are sampled influences relative performance of the three bounds , we performed experiment 2 for two sample intervals of the frequencies : @xmath210 and @xmath211 . because results for these two cases are very similar , we provide only the results corresponding to the interval @xmath210 .
the average values of bounds are plotted on fig .
3 ( a - c ) . the performances of our bound and bound from @xcite seem to be very similar , especially when plotted together with the performance of bound from @xcite .
however , on fig .
3 ( d - f ) we plotted for each topology the fractions of frequency - phases samples for which bounds from @xcite and from @xcite are more restrictive .
as we can see , on the majority of samples our bound outperforms the two other bounds .
we can now summarize the results of our comparison analysis of the three synchronization conditions .
the constraint on the initial phases for our condition seems to be the least restrictive compared to the similar constraints for bounds two other bounds .
in addition , on the majority of samples our coupling strength bound based on optimization approach is the least restrictive as can be seen on fig .
3 . we used matlab s
r2012a @xmath212 function from the global optimization toolbox with the following options : maxfunevals = 300000 , maxiter = 500000 , tolfun = @xmath213 , tolcon = @xmath213 , tolx = @xmath213 .
this paper studies synchronization of heterogeneous kuramoto oscillators with arbitrary underlying topology .
we provide novel sufficient conditions on the coupling strength that guarantee the existence of a positively invariant set ( pis ) and then use lasalle s invariance principle to show frequency synchronization . moreover , we provide an optimization framework that can further improve our bounds .
we illustrate these results with simulations performed for chain , ring and star - tree topologies .
our bounds consistently improve existing bounds on average for every investigated case .
* lemma 1 * * lemma 2 * _ if @xmath93 $ ] , then function @xmath94 satisfies the following differential inequality on @xmath95 $ ] : @xmath220 and if in addition @xmath221 then @xmath76 will be upper bounded by @xmath77 : @xmath98 $ ] .
_ multiplying @xmath222 equation of by @xmath223 and summing them together : @xmath224 it can be verified that : @xmath225.\ ] ] thus , using lemma 1 and the schwartz inequality : @xmath226 we now consider the following differential equation : @xmath227 this equation has one asymptotically stable equilibrium @xmath228 , and @xmath229 monotonically decreases to @xmath228 if condition is satisfied and therefore @xmath230 . by comparison principle , @xmath231,\ ] ] and thus @xmath232 $ ] .
peter achermann and hanspeter kunz .
modeling circadian rhythm generation in the suprachiasmatic nucleus with locally coupled self - sustained oscillators : phase shifts and phase response curves .
_ journal of biological rhythms _ , 14(6):460 - 468 , 1999 .
ll bonilla , cj perez vicente , and r spigler .
time - periodic phases in populations of nonlinearly coupled oscillators with bimodal frequency distributions .
_ physica d : nonlinear phenomena _ , 113(1):79 - 97 , 1998 .
seung - yeal ha , zhuchun li , xiaoping xue , formation of phase - locked states in a population of locally interacting kuramoto oscillators , journal of differential equations , volume 255 , issue 10 , 15 november 2013 , pages 3053 - 3070 , issn 0022 - 0396 , http://dx.doi.org/10.1016/j.jde.2013.07.013 .
y. kuramoto , self - entrainment of a population of coupled non - linear oscillators , _ in int .
symposium on mathematical problems in theoretical physics _ , h. araki , ed .
39 of lecture notes in physics , springer , 1975 , pp .
420 - 422 .
e. mallada and a. tang .
`` distributed clock synchronization : joint frequency and phase consensus . '' in _ proceeding of the 50th ieee conference on decision and control , and european control conference _ , pages 6742 - 6747 .
ieee , 2011 .
p. monzn and f. paganini .
global considerations on the kuramoto model of sinusoidally coupled oscillators . in _ proceedings of the 44th ieee conference on decision and control , and european control
, pages 3923 - 3928 .
ieee , 2005 .
strogatz , from kuramoto to crawford : exploring the onset of synchronization in populations of coupled oscillators , physica d : nonlinear phenomena , volume 143 , issues 14 , 1
september 2000 , pages 1 - 20 , issn 0167 - 2789 . | we study synchronization of coupled kuramoto oscillators with heterogeneous inherent frequencies and general underlying connectivity .
we provide conditions on the coupling strength and the initial phases which guarantee the existence of a positively invariant set ( pis ) and lead to synchronization .
unlike previous works that focus only on analytical bounds , here we introduce an optimization approach to provide a computational - analytical bound that can further exploit the particular features of each individual system such as topology and frequency distribution .
examples are provided to illustrate our results as well as the improvement over previous existing bounds . |
the identification of each collective dynamical state is done by measuring a set of order parameters and then using the decision tree in fig . [
fig : order_parameters ] to classify the synchronization state .
the same procedure is applied independently for the stimulated and the unstimulated groups of nodes . at each numbered decision point ,
threshold values ( table [ table : op_table ] ) for the order parameters are employed to answer the following questions : 1 . whether the magnitude of the inhibitory component of each oscillator changes over time . for a non - oscillating state
, this quantity should be below a threshold @xmath70 .
2 . whether the magnitude of the inhibitory component of each oscillator is close to zero for all oscillators .
if this quantity is below a threshold @xmath71 , it signifies amplitude death .
3 . whether the temporal mean of each oscillator is the same .
if the difference is below a threshold @xmath72 , it signifies oscillator death ; otherwise it is an inhomogeneous steady state .
4 . whether all the oscillators are in phase . if the difference between the phases of the oscillators is less than a threshold @xmath73 , it corresponds to exact synchronization .
5 . whether the temporal mean of each oscillator is the same .
if the difference is greater than a threshold @xmath74 , it signifies inhomogeneous in - phase synchronization . in the oscillating case ,
once phase synchronization has been established , this order parameter checks if all oscillators have the same trajectory .
whether the total amount of phase space covered by the trajectory is small . for the case of gs , this quantity is less than a threshold @xmath75 ; if the quantity is greater than the threshold , the state is identified as qp which , by definition , completely fills a bounded region in phase space . [ cols="^,^,^",options="header " , ]
in the following , we demonstrate that the results presented in the main text are robust with respect to small changes in the system parameters or system size .
first , we consider the case of coupling - induced oscillations , shown in fig . 2 ( a ) of the main text . as seen in fig .
[ fig : si_fig_onoff ] , steady state behavior is exhibited by a globally coupled system of @xmath2 ( @xmath76 ) nodes , with only @xmath46 of them receiving a weak input stimulus @xmath42 .
although the stimulation is too weak to initiate activity in an isolated node , upon coupling these nodes with sufficient strength @xmath41 , we find that oscillations emerge , as in the case @xmath77 considered in the main text .
+ next , we consider the symmetry switching transition discussed in the main text . in fig . 3 ( e ) of the main text
, we see that for the case @xmath63 , @xmath48 there exists a range of values within which the distinct collective dynamical states ( es , iis ) and ( iis , es ) coexist . as seen in fig . 3 ( f ) , the likelihood of obtaining the pattern pair ( iis , es ) is higher at lower values of coupling strength @xmath41 , and only (
es , iis ) is obtained above a critical value of @xmath41 .
these results hold even for much larger system sizes , as shown in fig .
[ fig : si_fig_bif ] , which displays a bifurcation diagram for the case @xmath78 , @xmath61 .
as we can see , the pattern pair ( iis , es ) is seen for lower @xmath41 , and ( es , iis ) is seen for higher @xmath41 .
+ finally , we show that our results are robust with respect to changes in both the internal coupling parameters as well as the input stimulus applied to the inhibitory components ( @xmath79 ) . in fig .
[ fig : si_fig4 ] ( a ) , we show a parameter space diagram indicating the synchronization states obtained upon changing @xmath41 as well as the ratio @xmath80 , for the case @xmath63 , @xmath48 . note that , for these simulations , the values of all the internal coupling strengths @xmath81 have been halved relative to those used for the simulations reported in the main text . we find that for low values of the ratio ( @xmath82 ) the patterns observed , and their phase boundaries , are almost unchanged from the case @xmath83 . at larger values of @xmath79 , the phase boundaries change , but we continue to observe the symmetry switching transition from ( iis , es ) to ( es , iis ) .
furthermore , as shown in fig .
[ fig : si_fig4 ] ( b ) , we continue to observe a hysteresis - like behavior in the range of coupling strengths around the symmetry - switching transition even at large system sizes ( @xmath84 , @xmath85 ) .
these observations suggest that the results reported in the main text are robust with respect to small variations in the parameter values and system size .
the equations corresponding to the system of @xmath2 globally coupled nodes are solved numerically using the variable order method for stiff differential equations as implemented in the ` ode15s ` routine in matlab^^. the time - series data used for analysis is recorded after allowing sufficient time for any transient effects to dissipate ( typically @xmath86 time units ) . | many natural systems including the brain comprise coupled non - uniformly stimulated elements . in this paper we show that heterogeneously driven networks of excitatory - inhibitory units exhibit striking collective phenomena , including spontaneous oscillations upon coupling . on varying the coupling strength a novel transition
is seen wherein the pattern symmetries of stimulated and unstimulated groups undergo mutual exchange .
the system exhibits coexisting chaotic and non - chaotic attractors - an intriguing result in view of earlier reports of varying degrees of chaoticity in the brain .
complex patterns are observed to spontaneously emerge across a wide range of spatial and temporal scales in nature @xcite .
uncovering the fundamental mechanisms driving such pattern formation will contribute significantly towards understanding self - organization in non - equilibrium systems @xcite .
perhaps the most influential paradigm for explaining this , particularly in the biological context , is that of reaction - diffusion systems @xcite .
the basic principle , involving the interplay between self - activation and lateral inhibition mediated by diffusion , has been suggested to underlie the emergence of patterns in a broad range of contexts @xcite .
however , not all phenomena involving activator - inhibitor interactions arise through diffusive coupling , one of the best - known examples being populations in neighboring ecological habitats coupled through intra - specific competition @xcite . indeed , reaction - diffusion processes can be seen as a subset of the wider class of systems involving nonlinear interactions between spatially distributed elements .
novel dynamical transitions may accompany the breaking of any of the symmetries that are intrinsic to these systems .
thus , uncovering the diverse range of collective phenomena associated with non - diffusively coupled systems of activator - inhibitor units can contribute towards understanding how patterns can arise in a more general setting .
neurobiological phenomena arising in synaptically coupled neuronal populations provide some of the most varied and complex instances of nonlinear interactions resulting in spatiotemporal patterns @xcite .
indeed , such coordinated collective activity is seen across several spatial scales in the brain : from the network of cortical areas where brain regions comprising @xmath0-@xmath1 neurons @xcite interact with each other through fiber tracts @xcite , to the olfactory bulb , where around @xmath0 glomerular clusters coordinate the information received from sensory neurons at the nasal epithelium @xcite .
each glomerulus , which comprises circuits of excitatory and inhibitory neurons , interact with other glomeruli through interneurons @xcite giving rise to lateral inhibition @xcite and excitation @xcite .
as these interneurons are subject to turnover @xcite , the strength of inter - glomerular coupling can vary .
thus , this complex biological process can be potentially understood in terms of the collective dynamics of a network of excitatory - inhibitory units coupled nonlinearly with tunable strength @xcite . as each glomerulus
is activated by a specific odorant receptor type @xcite and different smells evoke responses in different combinations of glomerular clusters @xcite , it is important in this context to understand the implications of non - uniform stimulation on the global behavior of such a system . in this paper , we investigate the collective dynamics resulting from non - uniformly driven networks of identical nodes , each comprising excitatory and inhibitory subpopulations .
the heterogeneous stimulation is implemented through external inputs being applied only to a subset of the nodes .
this conceptual framework applies beyond the context of neurobiology to phenomena as diverse as ecological interactions between prey and predator populations @xcite and interdependencies between institutions in economic systems @xcite . to describe the dynamics of the individual nodes , we consider a model for interacting excitatory and inhibitory subpopulations [ fig .
[ fig : fig1 ] ( a ) ] proposed by wilson and cowan @xcite obtained by temporal coarse - graining of neuronal population dynamics . individual units of this type are capable of exhibiting oscillatory activity for a range of external stimuli strength . in order to focus on the effect of heterogeneous stimulation
, we consider the simplest connection topology , viz . , coupling within and between the subpopulations of all nodes @xcite .
one of the novel features we observe is the occurrence of coupling - induced oscillations .
thus , stimuli that generate only steady - state behavior in isolated nodes can drive a network into oscillatory behavior , suggesting that lateral connections between glomerular clusters can allow the network to recognize weak signals incapable of initiating activity in an isolated cluster . increasing the strength of coupling between the nodes results in a variety of transitions in the collective dynamics of the network , the most striking of which involves an exchange of symmetry between the synchronization states of the groups of stimulated and unstimulated nodes .
in addition , we observe that the network can converge to qualitatively distinct attractors for identical system parameters , exhibiting chaotic or non - chaotic activity depending only on the initial state .
this result is intriguing in light of observations reporting chaotic activity in the olfactory bulb , which has been suggested to have a functional role in sensory information processing @xcite .
we consider a network of @xmath2 nodes , with each node @xmath3 comprising excitatory ( @xmath4 ) and inhibitory ( @xmath5 ) components , whose activity evolves as @xcite : @xmath6 where @xmath7 are time constants ( @xmath8 ) , @xmath9 and @xmath10 are the inputs received by the respective components , @xmath11^{-1}$ ] is a sigmoidal response function with maximum value @xmath12^{-1}\,,$ ] and @xmath13 , @xmath14 and @xmath15 are system parameters . as mentioned earlier ,
the network is globally coupled ( i.e. , every node has @xmath16 links ) with each link having the same weight @xmath17 .
this normalization allows our results to be system - size independent .
the total inputs to each component of node @xmath3 are @xmath18 and @xmath19 , where @xmath20 ( @xmath21 ) . to implement heterogeneous stimulation different external inputs ( @xmath22 , @xmath23 )
can be applied to different nodes . for the results shown here we have used the following set of parameter values : @xmath24=16 , @xmath25=15 , @xmath26=12 , @xmath27=3 , @xmath28=1.3 , @xmath29=2 , @xmath30=4 , @xmath31=3.7 , @xmath32=1 , @xmath33=8 and @xmath34
we have verified that our results are robust with respect to changes in these parameter values . along with the resulting limit cycle attractor .
( c ) bifurcation diagram for the inhibitory component @xmath5 of an isolated node shown as a function of the stimulus @xmath35 .
the broken lines indicate the unstable branch ( black ) , as well as the peaks ( @xmath36 , pink ) and troughs ( @xmath37 , violet ) in the oscillatory regime .
the solid black and thick red curves indicate the stable and saddle branches , respectively .
( d ) the @xmath5 time - series of an isolated node receiving stimuli @xmath38 ( green ) and @xmath39 ( red ) .
( e ) phase - plane portraits for a pair of identically stimulated coupled nodes ( i.e. , @xmath40 ) for different @xmath41 and @xmath42 showing [ l - r ] exact synchronization ( es ) , quasiperiodicity ( qp ) , anti - phase synchronization ( aps ) and inhomogeneous in - phase synchronization ( iis ) .
the positions of the oscillators are denoted by black filled circles . ] .
( a ) by switching `` on '' a stimulus that is too weak ( @xmath43 ) to generate activity in the uncoupled nodes ( top ) , oscillations can be observed by strongly coupling the stimulated ( red ) and unstimulated ( blue ) nodes ( bottom ) .
( b ) the range of @xmath42 for which oscillations emerge increases at higher @xmath41 .
the amplitude of the oscillations is larger for the stimulated node at low @xmath41 ( osc1 ) and for the unstimulated node at high @xmath41 ( osc2 ) .
the other dynamical regimes observed correspond to a non - zero steady state ( ss ) and a quiescent state characterizing amplitude death ( ad ) .
the regimes are determined via order parameters and identified as the pattern obtained from the majority ( @xmath44 ) of random initial states ( see supplementary information ) .
( c - h ) symmetry breaking in a system of @xmath45 globally coupled nodes .
the stimulated ( large , red ) and unstimulated ( small , blue ) nodes are indicated schematically in top - left corner of ( c - f ) . phase space projections of the trajectories ( colored as per the schematic ) are shown for ( c ) @xmath46 , displaying ( es , iis ) state ( note the broken symmetry in unstimulated nodes ) and ( e ) @xmath47 , exhibiting chaos . in the corresponding bifurcation diagrams for ( d ) @xmath46 and ( f ) @xmath48 ,
the peaks @xmath49 , pink ) and troughs ( @xmath37 , violet ) of the inhibitory component are shown as a function of @xmath41 .
( g ) magnified view of ( f ) showing the coexistence of qualitatively distinct dynamical attractors corresponding to ( iis , es ) [ orange ] and ( es , es ) [ green ] . the system can exhibit either chaotic or non - chaotic behavior depending on its initial state ,
as illustrated in the top and bottom panels of ( h ) for @xmath50 . ] on receiving a stimulus @xmath42 of sufficient magnitude , a single node is capable of exhibiting limit - cycle oscillations around an unstable fixed point @xcite [ fig .
[ fig : fig1 ] ( b ) ] .
this limit cycle emerges via the collision of stable and saddle branches [ fig .
[ fig : fig1 ] ( c ) ] , and the amplitude of oscillation depends on the value of @xmath42 [ fig .
[ fig : fig1 ] ( d ) ] . as
shown in fig .
[ fig : fig1 ] ( e ) and discussed in detail in ref .
@xcite , coupling identically stimulated nodes yields a rich variety of synchronization patterns including exact synchronization ( es ) , quasiperiodicity ( qp ) , anti - phase synchronization ( aps ) and inhomogeneous in - phase synchronization ( iis ) at different @xmath41 and @xmath42 . in this work ,
we consider heterogeneously driven networks wherein the number of nodes receiving external stimulus @xmath51 .
we henceforth denote the synchronization state in such systems through the notation ( @xmath52,@xmath53 ) , where the first and second terms in the pair correspond to the collective pattern observed in the stimulated and unstimulated nodes , respectively .
the simplest case of non - uniform stimulation is when a single node receiving input @xmath35 is coupled to an unstimulated node ( i.e. , @xmath54 , @xmath55 ) with strength @xmath41 . at sufficiently large @xmath41 , oscillations can be observed even for very low @xmath35 [ fig .
[ fig : fig2 ] ( a ) ] . indeed , as shown in the phase diagram in fig .
[ fig : fig2 ] ( b ) , at high @xmath41 periodic activity can be observed over a much larger range of @xmath42 than the one capable of inducing oscillations in an isolated node .
furthermore , for lower ( higher ) @xmath41 the unstimulated node has lower ( higher ) amplitude oscillations than the stimulated one , denoted as osc1 ( osc2 ) .
this extremely surprising _ coupling - induced periodic activity _ is seen even when @xmath56 ( see supplementary information ) and suggests that a node is capable of detecting weak , subthreshold inputs when it is coupled to one or more unstimulated nodes .
such an increase in the sensitivity to stimuli beyond the capability of a single element is an emergent collective property of the network and may be understood as an effective renormalization of the parameters governing the nodal dynamics .
the observation of oscillations at larger values of @xmath35 , where an isolated node exhibits a stable state , may also be connected to the appearance of periodic activity in bistable systems , e.g. , excitable elements subjected to a sufficiently strong stimulus , upon appropriate coupling @xcite . on increasing @xmath41 beyond a critical value that depends on the stimulus intensity @xmath35 , the activity of all nodes ceases , which corresponds to a state of amplitude death ( ad ) .
a new feature , indicative of _ spontaneous symmetry breaking _ , appears on minimally increasing the size of the network to @xmath45 keeping @xmath46 .
this is manifested as iis in the unstimulated nodes [ fig .
[ fig : fig2 ] ( c ) ] , neither of which directly receive any external stimuli but are activated only by coupling with a common stimulated node . nevertheless these identical nodes exhibit distinct oscillation patterns over a range of coupling strengths [ fig .
[ fig : fig2 ] ( d ) ] . on stimulating a second node (
i.e. , @xmath45 , @xmath47 ) , another intriguing phenomenon emerges , viz .
, the _ coexistence _ of chaotic and non - chaotic attractors .
the existence of chaotic behavior [ fig .
[ fig : fig2 ] ( e ) ] , which arises via a period doubling route [ fig .
[ fig : fig2 ] ( f ) ] is confirmed by verifying the existence of positive lyapunov exponents @xcite .
note that if we use a smaller @xmath35 , chaos can be seen in the even simpler stimulation configuration @xmath57 .
the novel feature for @xmath48 is that , depending on initial conditions , it is possible to see either of two possible collective dynamical states corresponding to ( es , es ) or ( iis , es ) , the latter being a chaotic attractor .
as both the ( iis , es ) and ( es , es ) states have non - zero basin sizes [ fig . [ fig : fig2 ] ( g ) ] we can see either chaotic or non - chaotic behavior for identical system parameters over a range of @xmath41 [ fig . [
fig : fig2 ] ( h ) ] .
this observation lends support to the hypothesis , based on experimental recordings from the rabbit olfactory system , that attractors with varying degrees of chaoticity can coexist in the brain @xcite .
globally coupled nodes .
( a ) different synchronization states obtained by varying the number of stimulated nodes , @xmath58 , and coupling strength , @xmath41 , with ( @xmath59 , @xmath60 ) referring to patterns in stimulated and unstimulated groups .
( b - c ) variation of the attraction basin size ( measured as fraction of initial states reaching the attractor ) with @xmath41 at @xmath61 for the different regimes in ( a ) , shown separately for the ( b ) stimulated and ( c ) unstimulated groups .
basin sizes have been estimated using @xmath62 initial conditions .
( d ) phase space projections of the trajectories ( red : stimulated , blue : unstimulated ) corresponding to the different synchronization states indicated in ( a ) .
( e - g ) symmetry exchange transition in a system of @xmath63 globally coupled oscillators with @xmath48 .
( e ) bifurcation diagram showing the peaks @xmath64 of the inhibitory components of all nodes as @xmath41 is varied .
distinct coexisting attractors corresponding to ( iis , es ) and ( es , iis ) are indicated by orange and light blue , respectively .
( f - g ) variation of the ( f ) fractional basin size of these attractors and ( g ) the order parameter @xmath65 with @xmath41 , shown separately for the ( top ) stimulated and ( bottom ) unstimulated groups .
the distinct trends seen in ( g ) on increasing ( red circle ) or decreasing ( blue dots ) @xmath41 indicate the occurrence of hysteresis - like behavior . ]
the existence of numerous synchronization regimes in the @xmath66 parameter space becomes apparent as we increase the complexity of the system by incorporating more nodes [ fig .
[ fig : fig3 ] ( a ) ] .
these regimes are typically demarcated by sharp changes in the sizes of the basin of attraction of individual patterns indicated in fig .
[ fig : fig3 ] ( b - c ) .
apart from the states corresponding to es , qp , iis and ad described earlier , new collective dynamical patterns for the stimulated and unstimulated groups emerge .
these include oscillator death ( od ) which is a homogeneous non - zero steady state , and gradient synchronization ( gs ) , a generalization of aps for systems with @xmath67 @xcite .
fig .
[ fig : fig3 ] ( d ) shows several of the possible collective patterns , including those corresponding to symmetry breaking ( iis ) in one or both groups of stimulated and unstimulated nodes .
a particularly surprising feature that we investigate in detail below is the existence of a novel transition in which the ( broken ) symmetry of the patterns in the stimulated and unstimulated groups undergo a mutual exchange on varying the coupling strength .
this symmetry exchange manifests as a transition from the ( iis , es ) to the ( es , iis ) state [ fig .
[ fig : fig3 ] ( a ) ] .
we examine the nature of this transition in detail by considering the simplest system in which it can be observed . to observe broken symmetry in each group , both
the stimulated and unstimulated groups should contain at least two nodes each .
we observe that in the minimal case , i.e. , for @xmath63 , @xmath47 , the states ( iis , es ) and ( es , iis ) co - exist over a range of @xmath41 [ fig . [
fig : fig3 ] ( e ) ] .
the transition between them occurs through a change in the relative basin sizes for the patterns in the two groups [ fig .
[ fig : fig3 ] ( f ) ] .
the mechanism of the ( broken ) symmetry exchange is further established by examining how the order parameter @xmath68 ( non - zero values of which indicate iis in this regime ) changes upon varying @xmath41 in either direction in an annealed manner . in this procedure , the system is allowed to evolve starting from a random initial state at low ( high ) @xmath41 , following which the value of @xmath41 is adiabatically increased ( decreased ) . as seen in fig .
[ fig : fig3 ] ( g ) , a hysteresis - like behavior can be observed in the transition region in both groups of nodes , consistent with the mechanism of shrinking basin sizes underlying the symmetry exchange .
our results provide a simple framework for understanding aspects of the complex patterns of spatiotemporal activity that the brain is observed to exhibit @xcite .
the mesoscopic approach used here , focusing on the dynamics of a network of neuronal clusters , can yield significant insights into the mechanisms by which such patterns emerge .
furthermore , the phenomena we report here occurs in a globally connected system , which may help elucidate the important role of long - range intracortical interactions in olfactory processing @xcite . among the variety of dynamical transitions observed upon varying the coupling strength ,
the most striking one involves an exchange of broken and restored symmetries between the groups of stimulated and unstimulated nodes .
this result suggests an experimentally testable hypothesis , namely that simulated and non - stimulated glomeruli may show distinct collective dynamics at different levels of arousal which are analogous to having different coupling strengths between nodes . to conclude
, we have shown that non - uniformly driven networks of identical nodes , each comprising excitatory and inhibitory subpopulations , are capable of exhibiting surprisingly rich collective phenomena .
we find that the system exhibits _ coexistence of qualitatively distinct attractors _ ( chaotic and non - chaotic ) for identical system parameters .
this is intriguing in light of experimental observations of chaotic dynamics of varying complexity , particularly in the olfactory system @xcite , reported over several decades and which are yet to be fully understood .
furthermore , nodes that are quiescent in isolation can spontaneously oscillate for sufficiently strong coupling , thus enabling the system to be activated even by stimuli that are incapable of generating dynamical activity in isolated nodes .
these _ coupling - induced oscillations _ may help in understanding how excitatory - inhibitory feedback can cause spontaneous oscillations in networks of naturally quiescent stochastic spiking neurons @xcite , whose mean - field limit can in fact be described through the wilson - cowan equations .
moreover , this can be understood as belonging to a larger class of phenomena characterized by the emergence of activity in quiescent systems upon coupling @xcite . in summary
, our results suggest that transitions between symmetry broken and restored symmetry states in heterogeneously driven system of coupled neural oscillators may underlie the sequence of complex activity patterns seen in the brain .
we would like to thank bhaskar saha , rajeev singh , sudeshna sinha , deepak dhar and k. a. chandrashekar for helpful discussions .
snm is supported by the imsc complex systems project ( @xmath69 plan ) .
vs is supported by the itra project .
we thank imsc for providing access to the supercomputing cluster `` satpura '' , which is partially funded by dst . 0 p. ball , _ the self - made tapestry : pattern formation in nature _ ( oxford university press , oxford , 1999 ) .
m. c. cross and p. c. hohenberg , rev .
mod .
phys .
* 65 * 851 ( 1993 ) .
j. d. murray , sci . am . *
258 * , 80 ( 1988 ) .
a. j. koch and h. meinhardt , rev .
mod .
phys . * 66 * , 1481 ( 1994 ) .
m. tlidi , p. mandel , and r. lefever , phys .
rev .
lett . * 73 * , 640 ( 1994 ) .
s. kondo and t. miura , science * 329 * , 1616 ( 2010 ) .
t. bnsgi , v. k. vanag , and i. r. epstein , science * 331 * , 1309 ( 2011 ) .
r. kapral , and k. showalter , _ chemical waves and patterns _ ( springer , new york , 2012 ) .
a. de wit , d. lima , g. dewel , and p. borckmans , phys .
rev .
e * 54 * , 261 ( 1996 ) .
v. k. vanag and i. r. epstein , phys .
rev .
lett . * 92 * , 128301 ( 2004 ) . h. nakao and a. s. mikhailov , nat .
phys .
* 6 * , 544 ( 2010 ) .
r. singh and s. sinha , phys .
rev .
e * 87 * , 012907 ( 2013 ) .
m. e. gilpin and f. j. ayala , proc .
natl .
acad .
sci .
usa * 70 * , 3590 ( 1973 ) . h. i. wu , p. j. sharpe , j. walker , and l. k. penridge , ecological modelling * 29 * , 215 ( 1985 ) .
k. j. friston , phil .
trans .
roy .
soc .
b * 355 * , 215 ( 2000 ) .
g. palm , hippocampus * 3 * , 219 ( 1993 ) c. johansson and a. lansner , neural netw . *
20 * , 48 ( 2007 ) .
d. s. modha and r. singh , proc .
natl .
acad .
sci .
usa * 107 * , 13485 ( 2010 ) .
p. mombaerts , f. wang , c. dulac , s. k. chao , a. nemes , m. mendelsohn , j. edmondson , and r. axel , cell * 87 * , 675 ( 1996 ) .
g. lowe , olfactory bulb : synaptic organization .
_ in : encyclopedia of life sciences
_ ( john wiley & sons , ltd , chichester , 2013 ) .
t. imai , semin .
cell dev
. biol . * 35 * , 180 ( 2014 ) .
j. m. christie and g. l. westbrook , j. neurosci .
* 26 * , 2269 ( 2006 ) p.
m. lledo , f. t. merkle , and a. alvarez - buylla , trends . neurosci . * 31 * , 392 ( 2008 ) .
m. migliore , f. cavarretta , a. marasco , e. tulumello , m. l. hines , and g. m. shepherd , proc .
nat .
acad .
sci .
usa * 112 * , 8499 ( 2015 ) .
p. mombaerts , annu .
rev .
cell dev
. biol .
* 22 * , 713 ( 2006 ) .
b. malnic , j. hirono , t. sato , and l. b. buck , cell * 96 * , 713 ( 1999 ) . r. m. may and a. r. mclean , _ theoretical ecology : principles and applications _
( oxford university press , oxford , 2007 ) .
r. m. may , s. a. levin , and g. sugihara , nature * 451 * , 893 ( 2008 ) .
f. schweitzer _
et al . _ ,
science * 325 * , 422 ( 2009 ) .
h. r. wilson and j. d. cowan , biophys .
j. * 12 * , 1 ( 1972 ) .
r. singh , s. n. menon , and s. sinha , sci .
rep .
* 6 * , 22074 ( 2016 ) .
w. j. freeman , sci .
am . *
264 * , 78 ( 1991 ) .
v. in , a. r. bulsara , a. palacios , p. longhini , a. kho , and j. d. neff , phys .
rev .
e * 68 * 045102 ( 2003 ) .
a. wolf , j. b. swift , h. swinney , and j. a. vastano , physica d * 16 * , 285 ( 1985 ) .
w. j. freeman , _ how brains make up their minds_. ( columbia university press , new york , 2000 ) .
r. i. wilson and z. f. mainen , annu .
rev .
neurosci .
* 29 * , 163 ( 2006 ) m. luo , neuron * 72 * , 1 ( 2011 ) .
h. korn and p. faure , c. r. biologies * 326 * , 787 ( 2003 ) .
e. wallace , m. benayoun , d. w. van , and j. d. cowan , plos one * 6 * , e14804 ( 2011 ) .
a. f. taylor , m. r. tinsley , f. wang , z. huang , and k. showalter , science * 323 * , 614 ( 2009 ) .
r. singh , j. xu , n. g. garnier , a. pumir , and s. sinha , phys .
rev .
lett . * 108 * , 068102 ( 2012 ) . j. xu , s. n. menon , r. singh , n. b. garnier , s. sinha , and a. pumir , plos one * 10 * , e0118443 ( 2015 ) .
* supplementary information * |
we consider a generalized eigenvalue problem ( eigenproblem ) for a linear pencil @xmath0 with symmetric ( hermitian in the complex case ) matrices @xmath1 and @xmath2 with positive definite @xmath1 .
the eigenvalues @xmath3 are enumerated in decreasing order @xmath4 and the @xmath5 denote the corresponding eigenvectors .
the largest value of the rayleigh quotient @xmath6 where @xmath7 denotes the standard scalar product , is the largest eigenvalue @xmath8 .
it can be approximated iteratively by maximizing the rayleigh quotient in the direction of its gradient , which is proportional to @xmath9 .
preconditioning is used to accelerate the convergence ; see , e.g. , @xcite and the references therein . here
we consider the simplest preconditioned eigenvalue solver ( eigensolver)the gradient iterative method with an explicit formula for the step size , cf .
@xcite , one step of which is described by @xmath10 the symmetric ( hermitian in the complex case ) positive definite matrix @xmath11 in ( [ e.preeig ] ) is called the _
preconditioner_. since @xmath1 and @xmath11 are both positive definite , we assume that @xmath12 the following result is proved in @xcite for symmetric matrices in the real space . [ t.1 ] if @xmath13 then @xmath14 and @xmath15 the convergence factor @xmath16 can not be improved with the chosen terms and assumptions .
compared to other known non - asymptotic convergence rate bounds for similar preconditioned eigensolvers , e.g. , @xcite , the advantages of are in its sharpness and elegance .
method is the easiest preconditioned eigensolver , but still remains the only known sharp bound in these terms for any of preconditioned eigensolvers .
while bound is short and simple , its proof in @xcite is quite the opposite .
it covers only the real case and is not self - contained in addition it requires most of the material from @xcite .
here we extend the bound to hermitian matrices and give a new much shorter and self - contained proof of theorem [ t.1 ] , which is a great qualitative improvement compared to that of @xcite .
the new proof is not yet as elementary as we would like it to be ; however , it is easy enough to hope that a similar approach might be applicable in future work on preconditioned eigensolvers .
our new proof is based on novel techniques combined with some old ideas of @xcite .
we demonstrate that , for a given initial eigenvector approximation @xmath17 , the next iterative approximation @xmath18 described by belongs to a cone if we apply any preconditioner satisfying . we analyze a corresponding continuation gradient method involving the gradient flow of the rayleigh quotient and
show that the smallest gradient norm ( evidently leading to the slowest convergence ) of the continuation method is reached when the initial vector belongs to a subspace spanned by two specific eigenvectors , namely @xmath5 and @xmath19 .
this is done by showing that temple s inequality , which provides a lower bound for the norm of the gradient @xmath20 , is sharp only in @xmath21 .
next , we extend by integration the result for the continuation gradient method to our actual fixed step gradient method to conclude that the point on the cone , which corresponds to the poorest convergence and thus gives the guaranteed convergence rate bound , belongs to the same two - dimensional invariant subspace @xmath21 .
this reduces the convergence analysis to a two - dimensional case for shifted inverse iterations , where the sharp convergence rate bound is established .
we start with several simplifications : [ t.simp ] we can assume that @xmath22 , @xmath23 , @xmath24 is diagonal , eigenvalues are simple , @xmath25 , and @xmath26 in theorem [ t.1 ] without loss of generality . first , we observe that method and bound are evidently both invariant with respect to a real shift @xmath27 if we replace the matrix @xmath2 with @xmath28 , so without loss of generality we need only consider the case @xmath29 which makes @xmath30 second , by changing the basis from coordinate vectors to the eigenvectors of @xmath31 we can make @xmath2 diagonal and @xmath23 .
third , having @xmath14 if @xmath32 or @xmath33 , or both , bound becomes trivial .
the assumption @xmath22 is a bit more delicate .
the vector @xmath18 depends continuously on the preconditioner @xmath11 , so we can assume that @xmath22 and extend the final bound to the case @xmath34 by continuity
. finally , we again use continuity to explain why we can assume that all eigenvalues ( in fact , we only need @xmath3 and @xmath35 ) are simple and make @xmath36 and thus @xmath24 without changing anything .
let us list all @xmath2-dependent terms , in addition to all participating eigenvalues , in method ( [ e.ep ] ) : @xmath37 and @xmath18 ; and in bound ( [ e.muest ] ) : @xmath37 and @xmath38 .
all these terms depend on @xmath2 continuously if @xmath2 is slightly perturbed into @xmath39 with some @xmath40 , so we increase arbitrarily small the diagonal entries of the matrix @xmath2 to make all eigenvalues of @xmath39 simple and @xmath36 .
if we prove bound for the matrix @xmath39 with simple positive eigenvalues , and show that the bound is sharp as @xmath41 with @xmath40 , we take the limit @xmath40 and by continuity extend the result to the limit matrix @xmath42 with @xmath29 and possibly multiple eigenvalues .
it is convenient to rewrite
equivalently by theorem [ t.simp ] as follows denotes the euclidean vector norm , i.e. , @xmath43 for a real or complex column - vector @xmath17 , as well as the corresponding induced matrix norm . ] @xmath44 and if @xmath45 and @xmath26 then @xmath14 and @xmath46 now we establish the validity and sharpness of bound assuming and .
[ t.cone ] let us define$ ] between vectors by @xmath47 .
] @xmath48 then @xmath49 and @xmath50 .
let @xmath51 be defined as the vector constrained by @xmath52 and with the smallest value @xmath53
. then @xmath54 .
orthogonality @xmath55 by the pythagorean theorem implies @xmath56 , so @xmath57 , since @xmath58 as @xmath24 , and @xmath59 where @xmath60 as @xmath24 .
a ball with the radius @xmath61 by ( [ e.ass ] ) centered at @xmath62 contains @xmath63 by , so @xmath64 .
the statement @xmath65 follows directly from the definition of @xmath66 .
now , @xmath67 as @xmath24 , so @xmath68 and @xmath69 we denote by @xmath70 the circular cone around @xmath62 with the opening angle @xmath71 . theorem [
t.cone ] replaces @xmath18 with the minimizer @xmath66 of the rayleigh quotient on the cone @xmath72 in the rest of the paper , except at the end of the proof of theorem [ t.2d ] , where we show that bounding below the value @xmath53 instead of @xmath38 still gives the sharp estimate . later on , in the proof of theorem [ t.wcurve ]
, we use an argument that holds easily only in the real space , so we need the following last simplification .
[ t.real ] without loss of generality we can consider only the real case .
the key observation is that for our positive diagonal matrix @xmath2 the rayleigh quotient depends evidently only on the absolute values of the vector components , i.e. , @xmath73 , where the absolute value operation is applied component - wise .
moreover , @xmath74 and @xmath75 , so @xmath76 .
the cone @xmath72 lives in the complex space , but we also need its substitute in the real space .
let us introduce the notation @xmath77 for the _ real _ circular cone with the opening angle @xmath78 centered at the _ real _ vector @xmath79 next we show that in the real space we have the inclusion @xmath80 .
for any complex nonzero vectors @xmath17 and @xmath81 , we have @xmath82 by the triangle inequality , thus @xmath83 .
if @xmath84 then @xmath85 , i.e. , indeed , @xmath86 , which means that @xmath80 as required . therefore , changing the given vector @xmath17 to take its absolute value @xmath87 and replacing the complex cone @xmath72 with the real cone @xmath77 lead to the relations @xmath88 , but does not affect the starting rayleigh quotient @xmath89 this proves the theorem with the exception of the issue of whether the sharpness in the real case implies the sharpness in the complex case ; see the end of the proof of theorem [ t.2d ] .
[ t.wcurve ] we have @xmath90 and @xmath91 such that @xmath92 .
the inclusion @xmath93 implies @xmath94 . _ _ ] assuming that @xmath66 is strictly inside the cone @xmath72 implies that @xmath66 is a point of a local minimum of the rayleigh quotient .
the rayleigh quotient has only one local ( and global ) minimum , @xmath95 , but the possibility @xmath96 is eliminated by theorem [ t.cone ] , so we obtain a contradiction , thus @xmath90 .
the necessary condition for a local minimum of a smooth real - valued function on a smooth surface in a real vector space is that the gradient of the function is orthogonal to the surface at the point of the minimum and directed inwards . in our case
, @xmath72 is a circular cone with the axis @xmath62 and the gradient @xmath97 is positively proportional to @xmath98 ; see figure [ f1 ] .
we first scale the vector @xmath66 such that @xmath99 so that the vector @xmath100 is an inward normal vector for @xmath101 at the point @xmath66 .
this inward normal vector must be positively proportional to the gradient , @xmath102 with @xmath103 , which gives @xmath104 , where @xmath105 . here
@xmath106 as otherwise @xmath66 would be an eigenvector , but @xmath107 by theorem [ t.cone ] , where by assumptions @xmath108 , while @xmath26 by theorem [ t.simp ] , which gives a contradiction .
as the scaling of the minimizer is irrelevant , we denote @xmath109 here by @xmath66 with a slight local notation abuse .
finally , since @xmath92 , inclusion @xmath93 gives either the required inclusion @xmath110 or @xmath111 with @xmath112 for some @xmath113 and @xmath114 we now show that the latter leads to a contradiction .
we have just proved that @xmath115 , thus @xmath116 .
let @xmath117 , where we notice that @xmath118 and @xmath119 since @xmath17 is not an eigenvector
. then we obtain @xmath120 where @xmath121 , therefore @xmath122 .
since all eigenvalues are simple , @xmath123 we observe that @xmath124 , i.e. , in the mapping of @xmath17 to @xmath66 the coefficient in front of @xmath5 changes by a smaller absolute value compared to the change in the coefficient in front of @xmath19 . thus , @xmath125 using the monotonicity of the rayleigh quotient in the absolute values of the coefficients of the eigenvector expansion of its argument , which contradicts @xmath126 proved in theorem [ t.cone ] . theorem [ t.wcurve ] characterizes the minimizer @xmath66 of the rayleigh quotient on the cone @xmath72 for a fixed @xmath17 .
the next goal is to vary @xmath17 , preserving its rayleigh quotient @xmath37 , and to determine conditions on @xmath17 leading to the smallest @xmath53 in such a setting .
intuition suggests ( and we give the exact formulation and the proof later in theorem [ t.2dred ] ) that the poorest convergence of a gradient method corresponds to the smallest norm of the gradient , so in the next theorem we analyze the behavior of the gradient @xmath127 of the rayleigh quotient and the cone opening angle @xmath128 [ t.2dgrad ] let @xmath129 be fixed and the level set of the rayleigh quotient be denoted by @xmath130 .
both @xmath131 and @xmath132 with @xmath133 attain their minima on @xmath134 in @xmath135 by definition of the gradient , @xmath136 for @xmath134 .
the temple inequality @xmath137 is equivalent to the operator inequality @xmath138 , which evidently holds .
the equality here is attained only for @xmath139 finally , we turn our attention to the angles . for @xmath140
the pythagorean theorem @xmath141 shows that @xmath142 is minimized together with @xmath143 .
but for a fixed @xmath144 the function @xmath145 is strictly increasing in @xmath146 which proves the proposition for @xmath147 now we are ready to show that the same subspace @xmath21 gives the smallest change in the rayleigh quotient @xmath148 . the proof is based on analyzing the negative normalized gradient flow of the rayleigh quotient
. [ t.2dred ] under the assumptions of theorems [ t.wcurve ] and [ t.2dgrad ] we denote @xmath149the set of minimizers of the rayleigh quotient . then @xmath150 ( see figure [ f2 ] ) . the initial value problem for a gradient flow of the rayleigh quotient , @xmath151 has the vector - valued solution @xmath152 which preserves the norm of the initial vector @xmath66 since @xmath153 as @xmath154 . without loss of generality
we assume @xmath155 the rayleigh quotient function @xmath156 is decreasing since @xmath157 as @xmath158 , the function @xmath156 is strictly decreasing at least until it reaches @xmath159 as there are no eigenvalues in the interval @xmath160\subset(\mu_{i+1},\mu_i),$ ] but only eigenvectors can be special points of ode .
the condition @xmath161 thus uniquely determines @xmath162 for a given initial value @xmath66 .
the absolute value of the decrease of the rayleigh quotient along the path @xmath163 is @xmath164 our continuation method using the _ normalized _ gradient flow is nonstandard , but its advantage is that it gives the following simple expression for the length of @xmath165 , @xmath166 since the initial value @xmath66 is determined by @xmath17 , we compare a generic @xmath17 with the special choice @xmath167 , using the superscript @xmath168 to denote all quantities corresponding to the choice @xmath169 . by theorem [ t.wcurve ] @xmath170 implies @xmath171 , so we have @xmath172 @xmath173 as @xmath21 is an invariant subspace for the gradient of the rayleigh quotient . at the end points , @xmath174 by their definition .
our goal is to bound the initial value @xmath175 by @xmath176 , so we compare the lengths of the corresponding paths @xmath177 and @xmath165 and the norms of the gradients along these paths .
we start with the lengths .
we obtain @xmath178 by theorem [ t.2dgrad ] . here
the angle @xmath132 is the smallest angle between any two vectors on the cones boundaries @xmath101 and @xmath179 .
thus , @xmath180 as our one vector @xmath181 by theorem [ t.wcurve ] , while the other vector @xmath182 can not be inside the cone @xmath72 since @xmath183 by theorem [ t.cone ] . as @xmath184 is a unit vector , @xmath185 as the angle is the length of the arc the shortest curve from @xmath186 to @xmath182 on the unit ball . for our special @xmath168-choice , inequalities from the previous paragraph turn into equalities , as @xmath187 is in the intersection of the unit ball and the subspace @xmath21 ,
so the path @xmath177 is the arc between @xmath188 to @xmath189 itself . combining everything together , @xmath190 by theorem [ t.2dgrad ] on the norms of the gradient , @xmath191 for each pair of independent variables @xmath192 and @xmath193 such that @xmath194 using theorem [ t : iif ]
, we conclude that @xmath195 as @xmath196 , i.e. , the subspace @xmath21 gives the smallest value @xmath197 by theorem [ t.2dred ] the poorest convergence is attained with @xmath93 and with the corresponding minimizer @xmath110 described in theorem [ t.wcurve ] , so finally our analysis is now reduced to the two - dimensional space @xmath21 .
[ t.2d ] bound holds and is sharp for @xmath93 . assuming @xmath198 and @xmath199 , we derive @xmath200 and similarly for @xmath110 where @xmath92 . since @xmath24
, we have @xmath201 . assuming @xmath202
, this identity implies @xmath203 which contradicts our assumption that @xmath17 is not an eigenvector . for @xmath204 and @xmath115 by theorem [ t.wcurve ] , the inverse @xmath205 exists .
next we prove that @xmath206 and that it is a strictly decreasing function of @xmath207 indeed , using @xmath208 and our cosine - based definition of the angles , we have @xmath209 where @xmath210 we substitute @xmath211 , which gives @xmath212 using , multiplication by @xmath213 leads to a simple quadratic equation , @xmath214 for @xmath215 as @xmath216 , @xmath217 , and @xmath218 , the discriminant is positive and the two solutions for @xmath219 , corresponding to the minimum and maximum of the rayleigh quotient on @xmath72 , have different signs . the proof of theorem [ t.wcurve ]
analyzes the direction of the gradient of the rayleigh quotient to conclude that @xmath103 and @xmath220 correspond to the minimum . repeating the same arguments with @xmath221
shows that @xmath222 corresponds to the maximum . but
@xmath223 since @xmath24 , hence the negative @xmath219 corresponds to the maximum and thus the positive @xmath219 corresponds to the minimum .
we observe that the coefficients @xmath216 and @xmath217 are evidently increasing functions of @xmath224 , while @xmath218 does not depend on @xmath225 .
thus @xmath206 is strictly decreasing in @xmath225 , and taking @xmath226 gives the smallest @xmath227 since @xmath92 where now @xmath206 , condition @xmath228 implies @xmath229 and @xmath230 implies @xmath231 so we introduce the convergence factor as @xmath232 where we use and again @xmath92 .
we notice that @xmath233 is a strictly decreasing function of @xmath206 and thus takes its largest value for @xmath234 giving @xmath235 i.e. , bound that we are seeking .
the convergence factor @xmath236 can not be improved without introducing extra terms or assumptions .
but @xmath236 deals with @xmath237 , not with the actual iterate @xmath18 .
we now show that for @xmath129 there exist a vector @xmath93 and a preconditioner @xmath11 satisfying such that @xmath238 and @xmath239 in both real and complex cases . in the complex case , let us choose @xmath17 such that @xmath240 and @xmath241 according to , then the real vector @xmath242 is a minimizer of the rayleigh quotient on @xmath72 , since @xmath243 and @xmath244 . finally , for a real @xmath17 with @xmath240 and a real properly scaled @xmath84
there is a real matrix @xmath11 satisfying such that @xmath245 , which leads to with @xmath246 indeed , for the chosen @xmath17 we scale @xmath84 such that @xmath247 so @xmath248 . as vectors @xmath249 and @xmath250 are real and have the same length there exists a _
real _ householder reflection @xmath251 such that @xmath252 .
setting @xmath253 we obtain the required identity .
any householder reflection is symmetric and has only two distinct eigenvalues @xmath254 , so we conclude that @xmath11 is real symmetric ( and thus hermitian in the complex case ) and satisfies .
the integration of inverse functions theorem follows . [
t : iif ] let @xmath255\to{\bf r}$ ] for @xmath217 be strictly monotone increasing smooth functions and suppose that for @xmath256 $ ] we have @xmath257 .
if for all @xmath258 $ ] with @xmath259 the derivatives satisfy @xmath260 then for any @xmath261 $ ] we have @xmath262 for any @xmath261 $ ] we have ( using @xmath257 ) @xmath263 if @xmath264 , then for the derivatives of the inverse functions it holds that @xmath265 since @xmath266 and @xmath267 are strictly monotone increasing functions the integrands are positive functions and @xmath268 as well as @xmath269 . comparing the lower limits of the integrals gives the statement of the theorem .
we present a new geometric approach to the convergence analysis of a preconditioned fixed - step gradient eigensolver which reduces the derivation of the convergence rate bound to a two - dimensional case .
the main novelty is in the use of a continuation method for the gradient flow of the rayleigh quotient to locate the two - dimensional subspace corresponding to the smallest change in the rayleigh quotient and thus to the slowest convergence of the gradient eigensolver .
an elegant and important result such as theorem [ t.1 ] should ideally have a textbook - level proof .
we have been trying , unsuccessfully , to find such a proof for several years , so its existence remains an open problem .
we thank m. zhou of university of rostock , germany for proofreading .
m. argentati of university of colorado denver , e. ovtchinnikov of university of westminster , and anonymous referees have made numerous great suggestions to improve the paper and for future work .
d e. g. dyakonov , _ optimization in solving elliptic problems _ , crc press , 1996 .
kny1986 a. v. knyazev , _ computation of eigenvalues and eigenvectors for mesh problems : algorithms and error estimates _ ,
( in russian ) , dept
moscow , 1986 .
k99 a. v. knyazev , _ preconditioned eigensolvers : practical algorithms _ , in z. bai , j. demmel , j. dongarra , a. ruhe , and h. van der vorst , editors , templates for the solution of algebraic eigenvalue problems : a practical guide , pp .
siam , philadelphia , 2000 .
k00 a. v. knyazev , _ toward the optimal preconditioned eigensolver : locally optimal block preconditioned conjugate gradient method _ , siam j. sci .
comput . , 23 ( 2001 ) , pp
. 517541 .
knn2003 a. v. knyazev and k. neymeyr , _ a geometric theory for preconditioned inverse iteration .
iii : a short and sharp convergence estimate for generalized eigenvalue problems _ , linear algebra appl .
, 358 ( 2003 ) , pp . 95114 .
ney2001a k. neymeyr , _ a geometric theory for preconditioned inverse iteration .
i : extrema of the rayleigh quotient _ , linear algebra appl .
, 322 ( 2001 ) , pp .
ney2001b k. neymeyr , _ a geometric theory for preconditioned inverse iteration .
ii : convergence estimates _ , linear algebra appl .
, 322 ( 2001 ) , pp . 87104 . | preconditioned eigenvalue solvers ( eigensolvers ) are gaining popularity , but their convergence theory remains sparse and complex .
we consider the simplest preconditioned eigensolver the gradient iterative method with a fixed step size for symmetric generalized eigenvalue problems , where we use the gradient of the rayleigh quotient as an optimization direction .
a sharp convergence rate bound for this method has been obtained in 20012003 .
it still remains the only known such bound for any of the methods in this class . while the bound is short and simple , its proof is not .
we extend the bound to hermitian matrices in the complex space and present a new self - contained and significantly shorter proof using novel geometric ideas .
iterative method ; continuation method ; preconditioning ; preconditioner ; eigenvalue ; eigenvector ; rayleigh quotient ; gradient iteration ; convergence theory ; spectral equivalence 49m37 65f15 65k10
65n25 ( place for digital object identifier , to get an idea of the final spacing . ) |
cellular membranes in living cells are a complex and compound system , including various kinds of proteins and different lipid compositions .
some of them are `` decorated '' by sugar chains called glycoproteins or glycolipids @xcite . from theoretical and experimental aspects , pure lipid membrane in tensionless fluidic state
has extensively received much attention for a long time @xcite .
pioneer studies on the polymer - membrane compound system , especially on membrane with anchored polymers , have been recently reported by several groups both theoretically and experimentally @xcite .
they found that the mechanical properties of lipid membrane , such as the bending rigidity @xmath0 , the saddle - splay modulus @xmath1 , and the spontaneous curvature @xmath2 , can be dramatically changed by the interaction between membrane and the surrounding solvent and macromolecules @xcite .
liposomes decorated by grafted polymers ( peg etc . )
are often used for carriers in drug delivery systems @xcite .
thus , it is important to understand the effects of grafted polymers for both fundamental research and medical applications . in this paper
, we explain our simulation model and then briefly describe the effects of grafted polymers on membrane properties .
we employed a meshless membrane model with grafted polymer chains , since it allows large scale simulations and the mechanical properties of pure membranes are easy to control @xcite . for an isolated anchored polymer chain , the membrane properties have been investigated in details by using scaling analysis and partition function method @xcite .
here we focus on membranes with a high polymer density , where the interactions between polymer chains are not negligible and have significant effects .
in this work , we employ a coarse - grained lipid membrane grafted with a coarse - grained flexible linear polymer chain to form a grafted biomembrane system ( see fig .
[ fig : cart ] ) .
one membrane particle represents a patch of bilayer membrane , and possesses only transnational degree of freedom .
the particles form a quasi-2d membrane by a curvature potential based on moving least - squares ( mls ) method @xcite .
( ii ) polymer particles are linked by a harmonic potential , and freely move like a bead - spring behavior with a soft - core repulsion .
an end of each polymer chain is anchored on one membrane particle with the harmonic potential .
we consider a patch of membrane composed of @xmath3 membrane particles . among them , @xmath4 membrane particles are anchored by a polymer chain .
each polymer chain consists of @xmath5 polymer segments with a connected membrane particle .
the membrane and polymer particles interact with each other via a potential @xmath6 all particles have a soft - core excluded volume term with a diameter of @xmath7 . @xmath8f_{\rm { cut}}(r_{ij}/\sigma ) \label{rep}\ ] ] which @xmath9 is the distance between membrane ( or polymer ) particles @xmath10 and @xmath11 .
the diameter @xmath7 is used as the length unit , @xmath12 , and @xmath13 is a @xmath14 cutoff function @xmath15\right\ } & ( s < s_{\rm { cut } } ) \\ 0 & ( s \geqslant s_{\rm { cut } } ) \end{array } \label{cutoff } \right.\ ] ] with @xmath16 . the factor @xmath17 in eq .
@xmath18 is determined so that @xmath19 , which implies @xmath20 . in eq .
@xmath21 , we use the parameters @xmath22 , and @xmath23 .
the membrane potential @xmath24 consists of attractive and curvature potentials , @xmath25 where the summation is taken only over the membrane particles .
the attractive multibody potential is employed to mimic the `` hydrophobic '' interaction .
@xmath26\right\}-c , \label{att}\ ] ] which is a function of the local density of membrane particles @xmath27 with @xmath16 , @xmath28 [ satisfied with @xmath29 , and @xmath30 .
the constant @xmath31 $ ] is chosen so that @xmath32 at @xmath33 . here
we set @xmath34 in order to simulate 2d fluidic membrane . for @xmath35 ,
@xmath36 acts like a pairwise potential with @xmath37 . for @xmath38 ,
this potential saturates to the constant @xmath39 .
thus , it is a pairwise potential with a cutoff at higher densities than @xmath40 .
the curvature potential is given by the shape parameter called `` aplanarity '' , which is defined by @xmath41 with the determinant @xmath42 , the trace @xmath43 , and the sum of principal minors @xmath44 .
the aplanarity @xmath45 scales the degree of deviation from the planar shape , and @xmath46 , @xmath47 , @xmath48 are three eigenvalues of the weighted gyration tensor @xmath49 where @xmath50 . without loss of generality , we suppose @xmath51 . if @xmath46 is the minimum eigenvalue representing a relative deviation from the local plane patch formed by the neighbor membrane particles around , its corresponding eigenvector is collinear with the normal vector @xmath52 of this plane patch . when the @xmath10-th membrane particle has two or less neighbor particles within the cutoff distance @xmath53 , they could be localized on a certain plane , therefore @xmath54 . the mass center of local plane patch @xmath55 , where a gaussian @xmath56 cutoff function is employed to calculate the weight of the gyration tensor @xmath57 & ( r_{ij}<r_{cc } ) \\ 0 & ( r_{ij}\geqslant r_{\rm { cc } } ) , \end{array } \label{cutoff } \right . \label{weight}\ ] ] which is smoothly cut off at @xmath58 .
here we use the parameters @xmath16 , @xmath59 , and @xmath60 . the bending rigidity and the line tension of membrane edges
are linearly dependent on @xmath61 and @xmath62 , respectively , for @xmath63 , so that they can be independently varied by changing @xmath61 and @xmath62 , respectively . as described in the previous work @xcite , the aplanarity @xmath64 for the smallest eigenvalue @xmath65,@xmath48 .
locally a membrane patch can be expressed for small fluctuations in the monge representation as @xmath66 where @xmath67 and @xmath68 coordinates are set to the principal axes . @xmath69 and @xmath70 are the principal curvatures of the membrane patch . in the case of cylindrical membrane , @xmath71 and @xmath72 . by averaging over the local neighborhood with a weight function @xmath73 where @xmath74 , we have @xmath75 since @xmath76 , we easily know @xmath77
thus , a linear relation between the bending energy in helfrich s macroscopic model and our mesoscopic parameter is obtained as @xmath78 , which is numerically confirmed in the previous work @xcite .
the membrane has zero spontaneous curvature without polymers .
the details of this membrane model are described in ref .
@xcite . in each polymer chain ,
its polymer particles are connected by a harmonic potential . @xmath79 where the summation is taken only between neighboring particles in each polymer chain and between the end polymer particles and anchored membrane particles .
the harmonic potential is given by @xmath80 which @xmath81 is a kuhn length of polymer chain .
we choose @xmath82 here such that a polymer chain stays in the force - free state for @xmath83 .
the @xmath84 ensemble ( constant number of particles @xmath85 , volume @xmath86 , and temperature @xmath87 ) is used with periodic boundary conditions in a simulation box of dimensions @xmath88 .
the dynamics of both membrane and anchored flexible polymers are calculated by using underdamped langevin dynamics .
the motions of membrane and polymer particles are governed by @xmath89 where @xmath90 is the mass of a particle ( membrane or polymer particle ) and @xmath91 is the friction constant .
@xmath92 is a gaussian white noise , which obeys the fluctuation - dissipation theorem : @xmath93 where @xmath94 and @xmath95 is the thermal energy .
we use @xmath96 , @xmath97 , @xmath98 , and @xmath99 through this work .
front and lateral snapshots of a cylindrical membrane with anchored flexible polymers outside used for simulations .
it contains @xmath100 membrane particles and @xmath101 polymer chains.,width=294 ] a cylindrical membrane , composed of @xmath102 , with @xmath4 anchored polymers outside ( see fig .
[ fig : snap ] ) is used to estimate the polymer - induced spontaneous curvature and bending rigidity . for a cylindrical membrane with a radius @xmath103 and a length @xmath104 ,
the curvature free energy is given by @xmath105 da \nonumber \\ \label{eq : f_tube } & = & 2\pi rl_z \left [ \frac{\kappa}{2}\left ( \frac{1}{r } -c_0 \right)^2 \right],\end{aligned}\ ] ] where @xmath69 and @xmath70 are the principal curvatures at each position on the membrane surface , and the membrane area @xmath106 .
the coefficients @xmath0 and @xmath1 are the bending rigidity and saddle - splay modulus , respectively .
biomembranes with lipids symmetrically distributing in both leaflets of the bilayer have zero spontaneous curvature @xmath107 .
the anchored polymer outside induces asymmetry to the membrane .
the bending energy of the membrane generates a shrinking force along the cylindrical axis , @xmath108 since a shorter cylinder membrane has a larger cylindrical radius , it has a lower bending energy .
previously , we clarified that this force can be used to estimate the spontaneous curvature @xmath109 and bending rigidity @xmath0 in simulations @xcite .
we calculated @xmath109 and @xmath0 for another meshless model , which has an orientational degree of freedom and can have a finite @xmath109 .
the obtained values of @xmath109 agree very well with those calculated from an alternative method using a curved membrane strip .
we define a polymer density @xmath110 to quantitatively describe the relation between the change of mechanical properties of membranes and the anchored polymer concentration . figures [ fig : fz ] shows the axial force @xmath111 calculated from the pressure tensor with different polymer densities .
the force @xmath111 increases linearly with @xmath112 .
thus , @xmath109 and @xmath0 of the grafted membranes can be estimated from the least squares fits to eq .
( [ eq : fz ] ) .
the obtained values of @xmath109 and @xmath0 are shown in figs .
[ fig : spcv ] and [ fig : bd ] , respectively . both quantities increase with increasing anchored polymer density .
this quantitatively agrees with the previous theory for a low polymer density .
when the membrane is curved in the opposite direction from the polymer grafting , the polymer chains have more space to move .
these entropic effects generate the spontaneous curvature of the membranes .
the details of our results and comparison with the theory will be described elsewhere @xcite .
we have studied the effects of grafted polymer on the membrane properties . anchored polymer chains
are added to a meshless membrane model .
the free energy of polymer chains is dominated by the entropic effects induced by the degree of freedom of polymer conformation .
we confirm that the induced spontaneous curvature and bending rigidity both increase as the polymer chain density increases .
anchored polymer chains also affect other properties of biomembranes ( it will be described in ref .
the properties and stability of membranes can be controlled by the decoration of membrane with grafted polymers .
we thank h. shiba for helpful discussions .
hw acknowledges the support by a mext scholarship ( japan ) . this study is partially supported by a grant - in - aid for scientific research on priority area `` molecular science of fluctuations toward biological functions '' from the ministry of education , culture , sports , science , and technology of japan .
e. sackmann , `` biological membranes architecture and function '' in _ structure and dynamics of membranes : i. from cells to vesicles _ , edited by r. lipowsky , and e. sackmann , north holland , amsterdam , 1995 , pp . 165 . | we have studied biomembranes with grafted polymer chains using a coarse - grained membrane simulation , where a meshless membrane model is combined with polymer chains .
we focus on the polymer - induced entropic effects on mechanical properties of membranes .
the spontaneous curvature and bending rigidity of the membranes increase with increasing polymer density .
our simulation results agree with the previous theoretical predictions . |
the first sources of ionizing radiation in our universe transformed the cold and neutral cosmic gas that was present shortly after the big bang into the hot and ionized plasma that we observe in the intergalactic medium ( igm ) today .
research into this transformation , which took place in the first billion years , during the epoch of reionization , is crucial for understanding the formation and evolution of galaxies , including our own galaxy , the milky way ( for reviews see , e.g. , @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
a number of new telescopes are currently being developed or are already underway to unravel the astrophysics of these early times , including the _ low frequency array _ ( lofar ; e.g. , @xcite ) , the _ murchison widefield array _ ( mwa ; e.g. , @xcite ) , the _ square kilometer array _
( ska ; e.g. , @xcite ) , the _ james webb space telescope _ ( jwst ; e.g. , @xcite ) , and many others ( e.g. , @xcite ; @xcite ; @xcite ) .
these investments at the observational frontier have spurred the construction of models of the early universe to guide the design of observational campaigns and to help interpret the observations once the data arrives .
most of the current models are built on the observationally supported notion that reionization is driven mainly by star - forming galaxies ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
a natural implication of these models is a strong coupling between reionization and galaxy formation caused by ionizing radiative feedback ( e.g. , @xcite ) .
photoionization heating by star - forming galaxies typically raises the temperature in the reionized gas to a few times @xmath3 .
the associated increase in gas pressure increases the jeans scale , i.e. , the scale below which pressure prevents the collapse of gas into gravitationally bound objects ( e.g. , @xcite ) .
this boils gas out of low - mass dark matter ( dm ) haloes and impedes the accretion of gas from the igm onto them ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
photoheating therefore leads to a reduction in the fuel from which stars form , which lowers the star formation rates ( sfrs ) and the ionizing emissivities of low - mass galaxies .
this makes it more difficult for galaxies to reionize the igm and , therefore , photoheating provides a negative feedback on reionization .
photoheating also provides a positive feedback on reionization .
the increase in the gas pressure and jeans mass smooths density fluctuations in the igm .
this decreases the gas clumping factor , @xmath4 , where the angular brackets indicate the average in the igm , and @xmath5 is the cosmic mean hydrogen density ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
because the rate at which the igm recombines is proportional to the gas clumping factor ( e.g. , @xcite ) , photoheating decreases the igm recombination rate .
fewer ionizing photons are then needed to keep the gas ionized , which facilitates reionization ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) .
this positive feedback from photoheating will be especially strong if x - ray sources preheat the gas before it is reionized ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
the feedback from photoheating may be accompanied by feedback from the explosion of stars as supernovae ( sne ; e.g. , @xcite ; @xcite ) .
sne are thought to suppress star formation in low - mass galaxies by expelling gas in winds , which makes it more difficult for galaxies to reionize the igm .
however , the ejection of gas may open additional low - density channels through which ionizing photons may escape the ism more easily , thus making it easier for galaxies to reionize the igm ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) .
the suppression of star formation by sne may be amplified in a nonlinear manner by the feedback from photoheating ( @xcite ; @xcite ; @xcite ) .
locally , sne may also increase the rates at which stars form as interstellar gas is compressed to star - forming densities in colliding sn shocks ( e.g. , @xcite ) .
the wide range of possibilities by which star formation may feed back on the formation of structure and the reionization of the igm makes modelling reionization a challenging task .
cosmological simulations are generally considered to be among the most powerful techniques to investigate the impact of stellar feedback on reionization and galaxy formation ( e.g. , @xcite ; @xcite ) , but are also computationally highly demanding .
a principal computational challenge is the required large dynamic range .
the characteristic size of individual ionized regions is of order 10 comoving mpc ( cmpc ; e.g. , @xcite ) .
the redshift at which these regions percolate to reionize the universe shows spatial fluctuations due to the modulation of galaxy formation by structure formation modes on still larger scales , of order @xmath6 ( e.g. , @xcite ; @xcite ) .
the concurrent need to resolve the jeans mass in the igm as well as the population of atomically cooling galaxies with masses above @xmath7 , the prime candidates for driving reionization , extends the spectrum of relevant scales down to a few kpc , and less ( e.g. , @xcite ; @xcite ) . a second computational challenge is the large expense of the radiative transfer ( rt ) of ionizing photons .
a straightforward and accurate treatment of the rt consists of tracing photons from each ionizing source along straight rays , one source at a time ( for an overview of rt methods see , e.g. , @xcite ; @xcite ; @xcite ) .
the computational cost of such approaches therefore increases linearly with the number of ionizing sources .
this renders simulations of cosmological volumes that contain large numbers of sources computationally demanding .
the issue becomes particular severe if the ionizing radiation from recombining ions needs to be treated accurately ( e.g. , @xcite ; @xcite ; @xcite ) .
the high cost of ray - based rt methods has triggered the development of moment - based rt methods that are computationally less expensive , but also less accurate ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) . to simplify the problem
, cosmological hydrodynamical simulations of galaxy formation often assume the gas to be in ionization equilibrium with a spatially uniform ionizing background ( e.g. , @xcite ; @xcite ; @xcite ) .
such simulations enable the investigation of the response of the gas in galaxies and the igm to heating by ionizing photons at relatively low computational cost , but must assume the redshift at which the ionizing background is instantaneously turned on .
the main drawbacks of such simulations are that they do not yield insights in the timing and morphology of reionization , and that fluctuations in the intensity due to local sources ( e.g. , @xcite ; @xcite ; @xcite ) and the ability of the gas to throw shadows and self - shield ( e.g. , @xcite ; @xcite ; @xcite ) are ignored .
sometimes , self - shielding is approximately accounted for by setting the intensity of the ionizing background to zero at densities above which the gas is expected to be neutral in the absence of local sources ( e.g. , @xcite ; @xcite ; @xcite ) .
other simulations focus on the accurate treatment of the rt of ionizing photons at the expense of a detailed treatment of the feedback from radiative heating ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) . in this type of simulations ,
the computational complexity is reduced by computing the rt and the evolution of structure separately , ignoring the impact of photoionization heating on the dynamics and distribution of the gas . often , to simplify the problem further , the dynamics of the gas is ignored altogether and assumed to trace that of the dm ( e.g. , @xcite ) .
a semi - analytical model then specifies the location of the galaxies and the rates at which these galaxies form stars .
sometimes , the semi - analytical model is calibrated to approximately account for the radiation - hydrodynamical response to photoheating ( e.g. , @xcite ; @xcite ) . the most direct but also computationally most expensive way to investigate the coupling of reionization and galaxy formation ,
is to carry out three - dimensional cosmological radiation - hydrodynamical simulations .
several of these simulations now exist , including the pioneering works by @xcite and @xcite .
however , even after more than a decade of intense research , the field is still in its infancy .
radiation - hydrodynamical simulations have only very recently begun to approach the cosmological scales and the high resolution needed to capture the relevant physics ( @xcite ; @xcite ; @xcite ) .
most earlier radiation - hydrodynamical simulations of reionization were restricted to relatively small volumes that are less representative of the cosmological mean ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) . only some simulations afforded ray - based rt methods that ensure the accurate propagation of ionization fronts in inhomogeneous density fields ( @xcite ) .
the majority of simulations employed the computationally less expensive but also numerically diffusive moment - based rt methods ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) . in some cases , the computational expense needed to be reduced further by lowering the spatial resolution at which the rt is carried out , thus requiring a sub - resolution model for the propagation and consumption of photons at hydrodynamically resolved scales ( @xcite ) . in this work
we present radiation - hydrodynamical simulations of reionization in cosmological volumes at high spatially adaptive resolution , implementing accurate rt using the traphic code ( @xcite ) .
traphic solves the time - dependent rt equation by tracing photon packets at the speed of light and in a photon - conserving manner through the simulation box .
traphic thus enables an accurate treatment of shadows and self - shielding .
the photon packets are transported directly on the spatially adaptive , unstructured grid defined by the sph particles , thus exploiting the full dynamic range of the hydrodynamic simulation .
the large number of ionizing sources typical of these simulations does not pose a problem because traphic employs a photon packet merging technique that renders the computational cost of the rt independent of the number of sources .
traphic is implemented in the galaxy formation code gadget ( last described in @xcite ) , and this enables us to account for the radiation - hydrodynamical feedback from star formation and reionization .
we focus our investigations on the roles of radiative heating ( which we also refer to as photoheating ) and sne in shaping the sfr and reionization history in the physically motivated and observationally supported models of the early universe that our simulations provide .
it is currently not feasible to provide an ab initio description of the internal structure of galaxies in cosmological volumes , even in substantially less expensive simulations in which the detailed transfer of ionizing photons is ignored ( e.g. , @xcite ; @xcite ; @xcite ) .
simulations like ours therefore rely on sub - resolution models of the interstellar medium ( ism ) .
this introduces free parameters that require careful calibration , e.g. , using zoomed simulations that achieve a higher resolution by focusing on a smaller computational volume , or observations .
the two main parameters relevant to simulations of galaxy formation during reionization are the fraction @xmath8 of the energy released in sne and other unresolved stellar evolution processes that is available to heat the ism , and the fraction @xmath9 of the ionizing photons that escapes the unresolved ism and are available to reionize the gas .
this paper is organized as follows . in section [ sec : sim ] we describe our numerical techniques and the set of simulations we have carried out . in section [ sec : results ] we describe our results .
we start in section [ sec : reionization ] with an overview of the simulations and a comparison with observational constraints .
this is continued in section [ sec : feedback ] and extended by an analysis of the impact of stellar feedback on galaxy formation and reionization . in section [ sec : discussion ]
we discuss the sensitivity of our results to variations in physical parameters as well as some of the main limitations inherent to our numerical approach . in section [ sec : summary ]
we conclude with a brief summary .
lengths are expressed in physical units , unless noted otherwise .
[ cols= " < , < , < , < , < , < , < , < " , ] we use a modified version of the n - body / treepm smoothed particle hydrodynamics ( sph ) code gadget ( @xcite ; our specific version is derived from that discussed in @xcite ) to perform a suite of cosmological radiation - hydrodynamical simulations of galaxy formation down to redshift @xmath10 ( see table [ tab : sims ] ) .
the simulations are initialized at redshift @xmath11 .
initial particle positions and velocities are obtained by applying the zeldovich approximation ( zeldovich 1970 ) to particles arranged along a uniform grid of glass - like structure ( @xcite ) .
we use a transfer function for matter perturbations generated with camb ( @xcite ) and apply it to describe perturbations in both the dark matter and the gaseous components .
we adopt the @xmath12cdm cosmological model with parameters @xmath13 , @xmath14 and @xmath15 , @xmath16 , @xmath17 , and @xmath18 ( @xcite ) .
these parameters are consistent with the most recent constraints from observations of the cosmic microwave background by the _ planck _ satellite ( @xcite ; @xcite ) .
our reference simulation utilizes @xmath19 dm and gas particles in a box of size @xmath20 .
this implies a dm particle mass of @xmath21 and a gas particle mass of @xmath22 .
atomically cooling haloes with mass ( eq . 3.12 in @xcite ) @xmath23 where @xmath24 is the mean atomic weight , are therefore resolved by at least @xmath25^{-3/2}$ ] dm particles .
we apply a comoving plummer - equivalent gravitational softening length , @xmath26 , equal to 1/25 of the mean dm particle separation to both dm and baryonic particles , i.e. , @xmath27 , where @xmath28 is the number of dm particles .
sph quantities are estimated by averaging inside a sphere containing @xmath29 neighboring gas particles and adopting the entropy conserving formulation of sph ( @xcite ) .
the sph kernel , i.e. , the radius of this sphere , is prevented from falling below @xmath30 . while our simulations account for the thermal feedback from the explosion of massive stars as core - collapse sne ( see section [ sec : sn ] ) , for simplicity , we do not follow the ejection and transport of metals .
we therefore follow the non - equilibrium chemistry and radiative cooling of the gas , assuming that the gas is of primordial composition .
molecule formation is assumed to be suppressed by the lyman - werner background expected to pervade the universe soon after the formation of the first stars ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) , a process not resolved in our simulations .
our simulations thus ignore that molecular hydrogen may form in self - shielded regions ( e.g. , @xcite ) .
we consider all relevant atomic radiative cooling processes , using the rate coefficients in @xcite : cooling by collisional ionization , collisional excitation of atomic lines , the emission of free - free and recombination radiation , and compton cooling by the cmb .
the species fractions and the gas temperature are advanced in time using the explicit subcycling time integration scheme presented in @xcite .
once stars form and emit radiation , the chemical and thermal evolution of the gas is also affected by photoionization and photoheating , and by the injection of energy by sne , as we describe below .
solving for the chemistry and cooling of the gas is by far the most expensive operation in our simulations . to speed up the calculations and unless otherwise noted , we set @xmath31 , i.e. , we neglect helium . in section
[ sec : parameters ] , we will describe a preliminary comparison with a simulation in which we follow the radiative heating and cooling by helium assuming @xmath32 .
our simulations do not have sufficient resolution and do not capture the physics required to describe the multiple phases of the dense gas in galaxies that are expected to develop above densities @xmath33 ( e.g. , @xcite ) as observed in the local universe .
we therefore employ a subgrid model to describe the thermodynamics of the dense gas , imposing a pressure floor @xmath34 on gas with densities @xmath35 , where @xmath36 , normalized to @xmath37 at the critical density @xmath38 .
we use @xmath39 for which both the jeans mass and the ratio of the jeans length and the sph kernel of gas at the pressure floor are independent of the density , thus preventing spurious fragmentation due to a lack of numerical resolution ( @xcite ; @xcite ) . note that , unlike in simulations that impose a polytropic equation of state , in our simulations , gas with densities above the critical density can have pressure above the pressure floor .
our implementation follows @xcite and @xcite , to which we refer the reader for further details . in our simulations adopting a hydrogen mass fraction @xmath40 , we use a critical density higher than @xmath38 by a factor @xmath41 to facilitate comparisons with simulations in which @xmath32 .
we employ the star formation implementation of @xcite , to which we refer the reader for details .
briefly , gas with densities @xmath42 and temperatures @xmath43 is allowed to form stars using a pressure - dependent rate that reproduces the kennicutt - schmidt relation observed in the local universe ( @xcite ) in idealized simulations of isolated disk galaxies , @xmath44 where @xmath45 is the sfr surface density and @xmath46 the gas surface density .
the last equation has already been renormalized by a factor of @xmath47 to account for the fact that it assumes the @xcite imf whereas we are using a chabrier imf .
this conversion factor between sfrs has been computed using the @xcite population synthesis code for model galaxies of age @xmath48 forming stars at a constant rate and is insensitive to metallicity .
we set @xmath49 and use @xmath50 , where @xmath51 is the density - dependent temperature implied by the pressure floor below which particles are prevented from cooling .
moreover , to avoid numerical artefacts at high redshifts when the cosmological mean gas density is high , we require a minimum overdensity of 57.7 for gas to be allowed to form stars .
the star formation law is interpreted stochastically , and the probability that a star - forming gas particle is turned into a star particle in a time interval @xmath52 is given by @xmath53 , where @xmath54 is the gas consumption time .
gas particles are converted to star particles assuming a conversion efficiency of 100% , i.e. , the masses of the star particles are identical to those of the gas particles from which they are formed . at higher resolution ,
star formation and the associated feedback may be more extended in time and in space .
this may affect the impact of star formation on the gas and therefore alter , e.g. , the resolved escape fraction of ionizing photons .
we interpret the star particles as simple stellar populations , i.e. , coeval stellar clusters that are characterized by an initial mass function ( imf ) , metallicity , and age .
we compute the time - dependent hydrogen ionizing luminosities of these star formation bursts using the population synthesis models of @xcite .
since @xcite does not tabulate ionizing luminosities for the chabrier imf , which is the imf used by our star formation recipe , we adopt the ionizing luminosities of the @xcite models with salpeter initial mass function in the range @xmath55 and metallicity @xmath56 .
this yields ionizing luminosities close to those of stellar populations drawn from a chabrier imf in the range @xmath57 with subsolar metallicity @xmath58 and is consistent with the imf used in our star formation recipe and appropriate for stellar populations at @xmath59 ( e.g. , @xcite ; @xcite ; @xcite ) .
we multiply the luminosities of each star particle by the sub - resolution escape fraction @xmath60 to model the removal of photons due to absorption in the unresolved ism .
some previous works that carried out rt simulations of reionization have calibrated the ism escape fraction to yield reionization histories in agreement with current observational constraints ( e.g. , @xcite ; @xcite ; @xcite ) . however , as we will discuss below , the interpretation of these constraints is subject to substantial systematic uncertainties and limited by our incomplete understanding of the physical processes at play .
we therefore prefer to set @xmath61 , and use our simulations to explore the dependence of the reionization history on box size and resolution . in section [ sec : parameters ] , we will briefly discuss the impact of variations away from @xmath61 on our results .
we use the rt code traphic @xcite to transport ionizing photons .
traphic solves the time - dependent rt equation in sph simulations by tracing photon packets emitted by source particles through the simulation box in a photon - conserving manner .
the photon packets are transported directly on the spatially adaptive set of sph particles and hence the rt exploits the full dynamic range of the hydrodynamical simulations .
a directed transport of the photon packets radially away from the sources is accomplished despite the irregular distribution of sph particles by guiding the photon packets inside cones .
a photon packet merging technique renders the computational cost of the rt independent of the number of radiation sources . in the following ,
we provide a brief overview of traphic in order to motivate the meaning of the numerical parameters of the rt specified below .
the reader is referred to the descriptions in @xcite , @xcite and @xcite for further discussion .
the transport of radiation starts with the emission of photon packets by star particles in @xmath62 tessellating emission cones .
the photons in each photon packet are shared among the @xmath63 neighboring sph particles residing in the cones . in cones containing zero neighbors , an additional
, so - called virtual particle is inserted to which the photon packet is assigned .
star particles emit photons using emission time steps @xmath64 , in between which the orientation of the cones is randomly changed to increase the sampling of the volume with photons .
the spectrum of the emitted radiation is discretized using @xmath65 frequency bins , and each photon packet carries photons from one of these bins .
therefore , the number of photon packets emitted per emission time step is @xmath66 .
the newly emitted photons are assigned a propagation direction parallel to the central axis of the associated emission cone , and , together with any other photon in the simulation box , are then propagated to the downstream neighbors of the sph particle at which they reside . a particle is a downstream neighbor if it is among the @xmath67 neighboring gas particles and resides in the regular transmission cone centred on the propagation direction and subtending a solid angle of @xmath68 .
the parameter @xmath69 defines the angular resolution of the rt .
if there is no downstream neighbor inside a transmission cone , a virtual particle is created to which the photons are then propagated .
the transmission cones confine the propagation of photons to the solid angle in which they were originally emitted .
the transport of photons occurs at the user - specified speed @xmath70 and is discretized using rt time steps @xmath71 .
a given sph particle may receive several photon packets within the same rt time step @xmath71 .
these photon packets are collected according to their propagation directions using a set of @xmath72 tessellating reception cones .
photon packets whose propagation directions fall in the same reception cone are merged and replaced by a single new photon packet .
each reception cone subtends a solid angle @xmath73 . hence
, the parameter @xmath72 determines the angular resolution of the merging . in the absence of virtual particles , the merging strictly limits the maximum number of photon packets in the simulation box to @xmath74 , where @xmath75 is the number of sph particles , and renders the computation time independent of the number of sources .
note that the angular resolution at which photon packets are merged may be chosen independently of the angular resolution at which photon packets are transferred .
photons are absorbed as they propagate through the gas from sph particles to their neighbors depending on the optical depth between the two neighboring particles , respecting photon conservation ( @xcite ; @xcite ) .
the absorption of photons within each frequency bin is treated in the grey approximation using photoionization cross sections . ]
@xmath76^{-1},\ ] ] where @xmath77 is the spectrum , and @xmath78 and @xmath79 are the low and high energy limits of frequency bin @xmath80 .
the number of absorbed photons determines the photoionization rate @xmath81 of hi in the given frequency bin @xmath80 defined by ( e.g. , * ? ? ?
* ) @xmath82 the photoionization rate implies a photoheating rate given by @xmath83 , where @xmath84 \nonumber \\ & \times & \left [ \int_{\nu_{\rm l}}^{\nu_{\rm h } } { d\nu \frac{4 \pi j_{\nu}(\nu)}{h_{\rm p } \nu } } \sigma_{\rm hi } ( \nu ) \right]^{-1 } \end{aligned}\ ] ] is the grey excess energy of frequency bin @xmath80 , and @xmath85 is the photoionization threshold energy of hi .
the photoionization and photoheating rates are then passed to a chemistry solver to update the hi fraction and the temperature of the gas . in the simulations presented in this work
, we set @xmath86 , and choose an angular resolution of the transport of @xmath87 and of the merging of @xmath88 .
sources emit photons into @xmath89 directions using emission time steps @xmath90 .
photons are transported at a speed @xmath91 , where @xmath92 is the speed of light , using time steps of size @xmath93 , where @xmath94 is the smallest gadget particle time step .
we use a single frequency bin with bounding energies located at @xmath95 { ~\mbox{ev}}$ ] .
we use the fits to the frequency - dependent photoionization cross sections by @xcite and adopt a black body spectral shape of temperature @xmath96 , consistent with the parameters of the adopted population synthesis models , to compute the grey photoionization and photoheating rate associated with that bin .
this gives @xmath97 and @xmath98 .
the rt assumes periodic boundary conditions , i.e. , photons leaving the box through any given face are reinserted at the periodically opposing face .
the dependence of the rt on the parameters has been discussed in @xcite and @xcite , to which we refer the reader .
additional tests can be found in the appendices of @xcite and @xcite . for simplicity ,
we treat recombination radiation in the on - the - spot approximation , i.e. , we compute recombination rates using case b ( e.g. , @xcite ; @xcite ) . the feedback from the explosion of stars as core - collapse sne is modeled as an injection of thermal energy into the ism surrounding the star particles . our numerical implementation , which has been designed to control spurious radiative energy losses due to the limited numerical resolution , is described in @xcite .
core - collapse sne inject thermal energy after a delay of @xmath99 after the birth of the star particle , approximately corresponding to the maximum lifetime of stars that end their lives as core - collapse sne . for each core - collapse sn that occurs ,
@xmath100 of thermal energy is injected .
the energy is distributed stochastically to an average subset of @xmath101 of the 48 neighboring sph particles , instantaneously increasing their gas temperature by @xmath102 and assuming a mean atomic weight @xmath103 appropriate for fully ionized gas .
we set @xmath104 , which results in a good agreement of the star formation history in our reference simulation and observational constraints at @xmath105 ( see figure [ fig4 ] ) .
we post - process our simulations to extract haloes using the friends - of - friends ( fof ) halo finder , with linking length equal to 1/5th of the mean dm inter - particle distance , built into the substructure finder subfind ( @xcite ; @xcite ) .
for each fof halo , we use subfind to identify the particle for which the gravitational potential is minimum , and let its location mark the halo center .
we obtain the virial radius as the radius of the sphere centered on the halo center within which the average matter density equals 200 times the redshift - dependent critical density of the universe .
the total mass inside this sphere is the virial mass of the halo .
the total sfr of the halo is the sum of the sfrs of the gas particles it contains .
we will also compute the baryon fraction of the halo , which is the ratio of the total mass in gas and stars inside the virial radius and the virial mass .
in section [ sec : reionization ] we provide an overview of the simulations listed in table [ tab : sims ] .
thereafter , in section [ sec : feedback ] , we discuss how stellar feedback from sne and photoheating impact the formation and evolution of galaxies and the reionization of the gas .
we will make use of additional simulations identical to those in table [ tab : sims ] , except that we have turned off sne and/or photoheating to help isolate the impact of the feedback these processes provide .
most observational works on uv galaxies assume a salpeter imf to infer sfrs , while our simulations assume a chabrier imf . where necessary
, we have divided observationally inferred sfrs by a factor of 1.7 to account for the difference between the two imfs ( see section [ sec : sf ] ) .
note that the high - resolution ( high - res ) simulation l12n512 has been stopped at @xmath106 due to its large computational expense .
+ + figure [ fig1 ] shows images of the gas densities , temperatures and neutral and ionized hydrogen fractions in our reference simulation l25n512 at @xmath107 . by this redshift ,
reionization has already proceeded significantly , and individual reionized regions are just about to percolate the simulation box .
most of the gas inside the ionized regions is photoheated to about @xmath108 , the characteristic temperature of stellar hii regions .
near galaxies , gas may be heated to much higher temperatures in gas - dynamical shocks that accompany the accretion of gas in massive haloes and the explosion of stars as sne .
some of the gas in the ionized regions is sufficiently dense to shield from the ionizing radiation and remains neutral .
the high resolution enabled by our spatially adaptive rt technique is key to tracking this gas phase .
for comparison , a rt simulation on a uniform grid with a number of grid cells equal to the number of sph particles employed here would have limited our ability to resolve the structure of the ionized gas to scales above @xmath109 comoving .
this is larger than the typical scale of self - shielded ( lyman - limit ) systems that are thought to be among the main consumers of ionizing photons ( e.g. , @xcite ; @xcite ; @xcite ) .
matching our spatially adaptive resolution using a uniform grid would require @xmath110 grid points per dimension , which is beyond current computational capabilities .
figure [ fig2 ] compares a set of basic observables extracted from our simulations with current observational constraints .
the sfr density of currently observationally accessible galaxies with sfrs @xmath111 ( assuming a chabrier imf , see figure [ fig4 ] ) , is insensitive to changes in box size and resolution and in good agreement with observations .
the comparison with the total sfr density suggests that current observations uncover only a fraction of stars forming at @xmath112 .
there is a significant dependence of the total sfr on resolution . because density fluctuations are increased at higher resolution , the reference simulation l25n512 yields initially a higher total sfr than the low - resolution ( low - res ) simulation l25n256 , and a lower total sfr than the high - res simulation l12n512 . on the other hand , near the final simulation redshift at @xmath10 ,
the total sfr is slightly smaller in the reference simulation than in the low - res simulation , and slightly larger than in the high - res simulation .
this inversion in the dependence on resolution is caused by the stronger suppression of star formation by stellar feedback at higher resolution .
the top right panel of figure [ fig2 ] shows that the reionization history , i.e. , the evolution of the neutral hydrogen fraction , is insensitive to the size of the simulation box @xmath113 .
this rapid convergence may partly result from our use of identical amplitudes and phases that characterize the initial gaussian random density fluctuations from which our simulations start .
indeed , a larger variance in the reionization histories on these scales is seen by comparing correspondingly sized sub - volumes of larger - scale simulations ( @xcite ; @xcite ; @xcite ) .
however , this variance is caused primarily by the large range of environments sampled by the subvolumes , each of which may have a different mean density ( @xcite ) .
the top right panel of figure [ fig2 ] shows further that there is a significant dependence of the reionization history on resolution . at higher resolution
, reionization ends at higher redshifts .
this mostly reflects the increased sfrs at increased resolution that our simulations show during the early and middle phases of reionization . in our simulations
, reionization completes at higher redshifts than suggested by constraints on the neutral fraction from observations of lyman-@xmath114 emitting galaxies and high - redshift quasars ( e.g. , figure 5 in @xcite ) .
however , interpretations of these observations are dominated by large systematic uncertainties ( for a discussion see , e.g. , @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) . at @xmath10 ,
the simulated mean neutral fractions are in reasonable agreement with observational constraints from the lyman-@xmath114 forest ( @xcite ) , which are thought to be relatively robust ( e.g. , @xcite ; but see , e.g. , @xcite ; @xcite ; @xcite ) . nevertheless , it is important to keep in mind that our simulations may underestimate the duration and overestimate the redshift of reionization because the limited size of the simulated volume may bias the reionization histories ( e.g. , @xcite ; @xcite ) , and because of our choice of a large sub - resolution escape fraction @xmath115 .
a useful check for consistency between the total sfr densities and the reionization histories is provided by comparison with the critical sfr density @xmath116 needed to keep the igm ionized ( @xcite ; updated as in @xcite ) , @xmath117 where @xmath118 is the fraction of ionizing photons that , on average , escape galaxies to ionize the igm ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) , and we have assumed a chabrier imf , which implies an ionizing emissivity per unit stellar mass larger , and therefore a critical sfr density smaller , by a factor of about @xmath119 compared to that implied by a salpeter imf ( see section [ sec : sf ] ) .
an igm clumping factor @xmath120 is a conservative estimate from our simulations ( see figure [ fig8 ] ) , and consistent with earlier works ( @xcite ; see the summary of results in @xcite ) . in this case , the simulated star - forming galaxies can keep the gas at @xmath121 ionized provided @xmath122 . whether such a relatively large escape fraction is physically plausible
is a matter of active research ( see , e.g. , the recent overview in @xcite ) .
computing the escape fraction from our simulations would require additional rt computations beyond the current work .
observations currently probe sfrs @xmath123 ( assuming a chabrier imf ; e.g. , @xcite , @xcite ; @xcite ; figure [ fig4 ] ) . under the above assumptions , at @xmath10 , the observed population of galaxies alone is capable of keeping the universe
ionized , lending support to scenarios in which reionization is driven by star - forming galaxies ( e.g. , @xcite ; @xcite ; @xcite ) .
the simulations predict a strong increase in the volume - averaged mean and median hydrogen photoionization rates as the simulation volumes are reionized ( bottom left panel of figure [ fig2 ] ) .
this signals a rapid build - up of an ionizing background , which sets the equilibrium neutral hydrogen abundance in the reionized igm ( e.g. , @xcite ; @xcite ) .
the median photoionization rate is in good agreement with constraints at @xmath10 from observations of , e.g. , the lyman-@xmath114 forest ( @xcite ) and quasar near - zones ( e.g. , @xcite ) .
however , the mean photoionization rate exceeds the median by more than an order of magnitude .
the excess in the mean photoionization rate with respect to the observational constraints is similar to that seen in previous simulations of reionization ( e.g. , @xcite ; @xcite ) .
the optical depth towards reionization , @xmath124 , where @xmath125 is the number density of free electrons and @xmath126 is the redshift at which reionization begins , is an important integral constraint on the reionization history ( e.g. , @xcite ; @xcite ; @xcite ) . for the computation of the optical depth , we assume a hydrogen mass fraction @xmath32 , and that the fraction of ionized hydrogen is equal to that in our simulations , which assume @xmath31 .
furthermore , we make the standard assumptions that hydrogen is fully ionized at @xmath127 , that helium is neutral ( singly ionized ) whenever hydrogen is neutral ( ionized ) at @xmath128 , and that helium is doubly - ionized at @xmath129 ( e.g. , @xcite ) .
the bottom right panel of figure [ fig2 ] shows that our simulations yield optical depths consistent with the most recent constraints derived from observations of the cosmic microwave background , @xmath130 ( @xcite ) .
our simulations lack the resolution and the physics to follow the formation of stars in minihaloes , i.e. , haloes with virial temperatures below @xmath3 .
ionization by these stars is expected to increase the optical depth by @xmath131 ( e.g. , @xcite ; @xcite ; @xcite ) .
if these stars evolve into accreting black holes , the optical depth may be further increased due to the pre - ionization of the igm by x - rays ( e.g. , @xcite ) . if the additional contribution to ionization from minihaloes and x - ray sources is indeed that high , and
if we were to add this contribution to our estimates of the optical depth , then this would yield values in slight excess of the planck observational estimates .
however , consistency with observations may still be achieved by lowering the ism escape fraction of ionizing photons , which is a free parameter in our simulations . in this case ,
reionization in our simulations would occur later and yield a lower optical depth ( see figure [ fig9 ] ) .
current cosmological simulations of reionization often have difficulties in simultaneously matching observational constraints on the neutral fraction , the photoionization rate and the optical depth ( e.g. , @xcite ; @xcite ) .
figure [ fig2 ] shows that our simulations are no exception .
many works have argued that these difficulties may signal a significant role by physical processes that are currently not understood or unknown , and that are currently only poorly modelled or lacking altogether in cosmological simulations .
such processes include , among others , an evolving igm escape fraction ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) and the absorption of photons by lyman limit systems ( e.g. , @xcite ; @xcite ; which our reference simulation begins to resolve ) .
some works have successfully matched the observational constraints by a careful calibration of parameters ( e.g. , @xcite ) . in section
[ sec : parameters ] we confirm that calibrating the ism escape fraction of ionizing photons helps to improve the match with observational constraints .
however , it is also important to acknowledge that current observational constraints on the neutral fraction at @xmath132 are highly model - dependent and too weak to single out reionization scenarios ( e.g. , @xcite ; @xcite ) . investigating these issues is beyond the aims of the current work . here
we will focus on how feedback from photoheating and sne impacts the formation of galaxies and the reionization of the igm in physically motivated scenarios of reionization . [ sec : feedback ] in this section we investigate how stellar ionizing radiation and sne impact the formation and evolution of galaxies and the reionization of the igm .
we will compare the simulations listed in table [ tab : sims ] , which include both photoheating and sne , with simulations in which photoheating and sne are disabled but which are otherwise identical .
we use the suffix _ nofb _ to distinguish the latter simulations from the former .
we further compare with simulations that include sne but not photoheating , distinguished by the suffix _ noph _ , and with simulations which include photoheating but not sne , distinguished by the suffix _
nosn_. we focus on our reference simulation l25n512 to illustrate the physics at play , and we assess the impact of box size and resolution on our conclusions . figure [ fig3 ] quantifies the impact of stellar feedback on the evolution of the total cosmic sfr density in comparisons of our reference simulation l25n512 with simulations that differ only in the inclusion of the stellar feedback processes .
the comparison with simulation l25n512-nofb , in which both photoionization heating and sne were turned off , reveals the strong suppression of star formation by stellar feedback .
the comparison with simulations l25n512-noph , in which only photoheating is turned off , and l25n512-nosn , in which only sne are turned off , shows that the suppression of the sfr due to sne is much stronger than the suppression of the sfr due to photoheating .
sne have a significant impact already at early times , @xmath133 , long before the igm in the simulation is substantially ionized .
on the other hand , the impact of photoheating becomes noticeable only at @xmath134 , when the average ionized fraction approaches unity .
the belated impact of photoheating on the sfr suggests a stronger role of suppression of star formation due to the illumination of galaxies by the ionizing background than due to the exposure to local ionizing radiation from stars inside the galaxies .
however , the impact of photoheating on the cosmic sfr remains comparatively small also after reionization .
figure [ fig3 ] also shows that photoheating suppresses star formation more strongly in the simulation with sne than in the one without sne ( compare the difference between the two solid curves with the difference between the two dashed curves ) . on the other hand , sne suppress star formation more strongly in the simulation with photoheating than in that without ( compare the difference between the two blue curves with the difference between the two red curves ) . photoheating and
sne thus amplify each other in suppressing the formation of stars .
these results are in good agreement with our earlier work in @xcite , in which we first reported this effect in simulations in which the gas was in photoionization equilibrium with a uniform ionizing background instantaneously turned on at @xmath135 , and in which sn feedback was implemented by kicking gas particles in winds ( see also @xcite ; @xcite ; @xcite ) .
another important observable of galaxy formation is the uv luminosity function , i.e. , the volume density of galaxies per unit logarithmic uv luminosity .
it quantifies the contributions to the cosmic sfr density from galaxies with different sfrs , and therefore helps us to understand the processes that shape the cosmic sfr density discussed above .
figure [ fig4 ] shows the uv luminosity function at @xmath136 in our reference simulation l25n512 .
we converted sfrs and uv luminosities using @xmath137 ( e.g. , @xcite ; @xcite ) , which yields the sfr function ( e.g. , @xcite ) .
the factor 0.6 converts the original relation , which is based on a salpeter imf , to a relation assuming the chabrier imf used here ( see section [ sec : sf ] ) .
luminosities are related to ab magnitudes using @xmath138 - 48.6 $ ] .
our reference simulation l25n512 yields a sfr function in good agreement with the observations above the limiting sfr of @xmath123 ( assuming a chabrier imf ) currently probed by observations .
small difference exist at @xmath139 , where observational estimates are characterized by large uncertainties , and at @xmath121 , especially at sfrs @xmath140 , where dust corrections may be significant .
in the left - hand panel of figure [ fig5 ] we investigate how feedback impacts the uv luminosity function , where we focus for simplicity on a single characteristic redshift @xmath141 .
the comparison of the reference simulation with simulation l25n512-nofb , in which both photoheating and sne are turned off , shows that suppression of star formation by stellar feedback is a critical element of simulations that attempt to match the observed sfr function . in our reference simulation , reionization heating and sne suppress star formation at nearly complementary scales .
the sfr function in simulation l25n512-nosn , in which only sne are turned off and which serves to demonstrate the isolated impact of photoheating , approaches that in simulation l25n512-nofb at sfrs @xmath142 . on the other hand , the sfr function in simulation l25n512-noph , in which only photoheating is turned off and which serves to demonstrate the isolated impact of sne , approaches that in simulation l25n512-nofb at sfrs @xmath143 .
the inefficiency of photoheating at high sfrs is expected if sfrs correlate positively with halo mass , as is the case in our simulations ( figure [ fig6 ] ) , since reionization affects mostly the gas in low - mass haloes ( e.g. , @xcite ; @xcite ; @xcite ) .
however , the decreased efficiency of sn feedback at low sfrs is likely a numerical artefact of the limited resolution .
indeed , we show in appendix [ app1 ] that in galaxies with low sfrs , star formation is more strongly suppressed by sne at higher resolution .
note that our simulations also likely underestimate the local impact of photoheating by stars , especially in the lowest - mass star - forming galaxies ( @xcite ) , thus likely raising the relative importance of photoionization by the cosmological background .
the right - hand panel of figure [ fig5 ] shows how our results depend on resolution and box size . while our reference simulation yields sfr functions in good agreement with the observations , simulations at lower and higher resolution significantly overpredict and underpredict the observed sfr functions in the range of sfrs in which sn feedback is efficient
this can be understood by noting that the rate at which the sn heated gas cools and loses energy depends on resolution ( eq . 17 in @xcite ) . at higher resolution
, the sn heated gas can maintain a high ratio of the local radiative cooling time and the sound - crossing time across a resolution element , a requirement for efficient feedback , out to higher densities , increasing the ability of the sn heated gas to prevent the formation of new stars . because at higher resolution , the gas fraction in low - mass haloes is more strongly reduced by stellar feedback ,
also the hierarchically assembling larger - mass haloes grow systematically more baryon - deficient as resolution is increased ( see also , e.g. , @xcite ) .
the differences between the simulated and observed sfr functions in the high - res and low - res simulations may be reduced by increasing or lowering the fraction @xmath144 of the energy each sn injects .
this would be reasonable since our simulations lack the resolution and the physics that is needed to accurately model the radiative losses in the ism .
this strategy would enable one to investigate numerical convergence in simulations that reproduce observational constraints on star formation independent of resolution ( e.g. , @xcite ; @xcite ) .
the current set of simulations , on the other hand , highlights the dependence on resolution of the sn energy fraction required to match observational constraints on star formation .
this dependence demonstrates clearly that current cosmological simulations have limited power in predicting the properties of star - forming galaxies from first principles .
note however , that the difference in the simulated and observed sfr functions is relatively small , of the order of the difference introduced by uncertainties in the conversion between uv luminosities and sfrs ( see discussions in , e.g. , @xcite ; @xcite ; @xcite ) .
therefore , the agreement between simulated and observed sfr functions is still very good .
finally , we investigate the dependence on box size .
the reference simulation overpredicts the abundance of the most intense star - forming galaxies in comparison with observations .
this is likely due to the finite size of the box , which biases estimates of the abundance of the most massive and therefore rarest galaxies .
the comparison with the small and large box simulations l12n256 and l50n512 supports box size as a limiting factor in accurate determinations of the sfr function at the highest sfrs .
additional processes that may impact models of the sfr function at high sfrs include feedback from massive black holes ( e.g. , @xcite ; @xcite ) , which we have ignored here , and obscuration by dust ( e.g. , @xcite ; @xcite ) . while we compare to observations of dust - corrected sfrs , there are large systematic uncertainties in the amount of dust as well as its properties , especially at the high redshifts of interest here ( e.g. , @xcite ; @xcite ) .
it is also important to keep in mind that observational estimates of the sfr function are subject to cosmic variance ( e.g. , @xcite ) .
the reduction of the sfrs of low - mass galaxies due to photoheating may affect the slope of the sfr function at low sfrs .
several ( semi- ) analytical works have proposed to search for this type of signature of reionization in the sfr function ( e.g. , @xcite ; @xcite ) .
because the mass scale below which reionization suppresses star formation depends on the timing of reionization , locating this signature may constrain the reionization history ( e.g. , @xcite ; @xcite ) . unfortunately
, the imprint of photoheating in the sfr function may be quite weak and therefore difficult to detect , because of intrinsic scatter that causes galaxies of the same mass to show different sfrs ( e.g. , @xcite ) , because of hierarchical assembly that leaves also relatively massive haloes deficient in baryons as they assemble from low - mass photoevaporated progenitors ( e.g. , @xcite ; @xcite ) , and because of the averaging in volumes in which reionization may proceed in a patchy manner ( e.g. , @xcite ) . in our simulation l25n512-nosn , which does not include sn feedback and therefore lets us isolate the impact of radiative heating , photoheating indeed causes a pronounced change in the slope of the sfr function at sfrs @xmath145 .
however , in simulation l25n512 that includes both photoheating and sn feedback , the change in the slope of the uv luminosity function due to photoheating is strongly masked by the suppression of star formation by sne at sfrs @xmath146 . in the presence of sne
, the change in the slope of the sfr function occurs at lower sfrs , near the scale at which the lack of low - temperature physics prevents gas cooling and star formation in our simulations .
this result is robust to changes in resolution explored here .
we thus conclude that if sn feedback is as efficient as suggested by our simulations , detecting the signature of reionization heating in the sfr function will be challenging . however , it is important to keep in mind that our simulations ignore that stars may also form in minihaloes cooling through molecular hydrogen transitions , which would be more susceptible to feedback from photoheating .
the steep rise of the sfr function in the presence of radiative heating and sne at low sfrs down to the sfrs that our simulations do not accurately resolve is consistent with that seen in previous works ( e.g. , @xcite ; @xcite ; @xcite ) , and with extrapolations of the observed sfr function ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) . because the mass scale below which photoheating affects the sfrs of galaxies is close to the mass scale below which efficient gas cooling requires the presence of molecular hydrogen or metals , quantifying the precise impact of reionization on the uv luminosity function will require more detailed , higher - resolution simulations of galaxy formation that include the relevant gas physics . while this is challenging because it extends the relevant dynamic range to still smaller scales , impressive first steps towards such simulations
have already been made ( e.g. , @xcite ; @xcite ) . + inspecting the properties of individual galaxies
helps gain insight into the physical origin of the shape of the sfr function discussed above .
figure [ fig6 ] shows the median sfrs ( top row ) and the median baryonic mass fractions @xmath147 ( bottom row ) of galaxies in our simulations .
the left - hand panel in each row shows the impact of feedback in the reference simulation at @xmath141 , and the right - hand panel shows the evolution with redshift . in the top right panel , we have divided the sfrs by the virial mass of the haloes hosting the galaxies to improve the clarity of the presentation .
our reference simulation predicts that the faintest galaxies accessible by current observations reside in haloes with masses @xmath148 ( intersection of solid and diagonal dotted curves in the top middle panel ) , consistent with observational estimates based on matching the shapes of the dm halo mass function and the observed uv luminosity function ( e.g. , @xcite ) .
the top left panel of figure [ fig6 ] shows that in our reference simulation , the median sfr at @xmath141 drops sharply below @xmath149 . the comparison with simulation l25n512-nosn in which sne were turned off ,
shows that this scale is set primarily by feedback from reionization . in the absence of photoheating
, star formation continues in haloes with lower masses .
however , even without feedback , galaxies in haloes with masses @xmath150 do not form stars since gas cooling and condensation is prevented by the finite resolution and the lack of low - temperature coolants , such as molecular hydrogen and metals .
sn feedback suppresses star formation strongly across nearly the entire range of galaxy masses .
it only becomes inefficient at the lowest masses at which our finite resolution impacts the conversion of gas into stars , and at the highest masses , at which our simulations are also subject to finite box size effects .
stellar feedback reduces star formation primarily because it reduces the amount of dense gas , both by expelling gas and by limiting the rates at which gas is accreted . in our reference simulation ,
reionization strongly reduces the gas fractions at halo masses @xmath151 .
reionization also reduces , though less strongly , the gas fractions in up to @xmath152 times more massive galaxies , which assemble in mergers of lower - mass baryon - deficient galaxies and accrete gas from the reionized igm ( e.g. , @xcite ; @xcite ; @xcite ) .
sn feedback most strongly impacts the gas fractions at masses @xmath153 in our reference simulation , again demonstrating that reionization and sn feedback act mostly complementary at the reference resolution .
we show in appendix [ app1 ] that in our high - res simulation , both reionization and sn feedback reduce the baryon fractions more strongly , and sn feedback extends its impact to haloes with masses as low as @xmath154 .
the scale at which photoheating becomes effective in suppressing star formation and reducing the baryon fractions in our reference simulation evolves significantly only after @xmath155 , which coincides with the completion of reionization .
a similar effect is seen in the evolution of the baryon fractions .
this late impact of photoheating suggests a larger role of illumination by the external ionizing background than by localized ionizing sources internal to the galaxies , in agreement with the discussion of the belated impact of radiative heating on the cosmic sfr density above .
however , we caution that in our simulations the impact of local ionizing sources may be underestimated due to the finite resolution . at least before reionization , radiative feedback from local sources is expected to dominate over feedback from external illumination .
this is indeed seen in simulations of the first stars and the start of reionization , which focus on smaller volumes and therefore can afford a higher resolution and also accurately capture the relevant low - temperature physics , primarily the cooling and chemistry of molecular hydrogen ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) . both sne and radiative heating may have a strong impact on reionization .
we have seen above that sn feedback reduces the sfrs and therefore the ionizing luminosities of galaxies , which makes it more difficult for the galaxies to reionize the universe .
however , sne may also open low - density channels in the ism , e.g. by expelling gas in winds , through which ionizing photons may escape more easily ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) .
this makes it easier for galaxies to reionize the universe .
the net impact of sne on reionization is the result of the interplay of these two processes , and is therefore difficult to predict .
we have also seen above that radiative heating may help sne to suppress star formation , which impedes reionization . on the other hand , radiative heating raises the jeans mass in the igm , and this reduces the igm clumping factor and therefore the rate at which the igm recombines . if recombinations consume a significant number of the ionizing photons that escape into the igm , then radiative heating will help keeping the ionized gas ionized and thereby facilitate reionization ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) .
figure [ fig7 ] shows that sn feedback has only a small impact on the timing of reionization . in the reference and low - res simulations ,
the reionization histories with and without sn feedback are nearly identical , and in the high - res simulation ,
sn feedback delays reionization by @xmath156 . that the net impact of sne on reionization is small suggests that the strong reduction in the sfrs due to sne - in the high - res simulation by a factor @xmath152 by the end of reionization ( figure [ fig3 ] ) - is partially compensated by an increase , also due to sne , in the escape fraction of ionizing radiation
. a plausible physical mechanism by which this might be achieved , is the strong reduction of the baryon fractions by sne seen in our simulations ( figure [ fig6 ] ) .
however , the lack of a substantial impact on the reionization history could also be explained if reionization is driven mostly by the lowest - mass galaxies in which sn feedback is less effective in suppressing star formation due to the limited resolution .
rt computations of the escape of ionizing photons into the igm would help to identify which of these mechanisms is dominant , but this is beyond the scope of the current work .
the left - hand panel of figure [ fig8 ] shows that the reduction in the igm recombination rate due to radiative heating is strong .
we have followed @xcite and have equated the igm clumping factor , @xmath157 , to @xmath158 , which parametrizes the average recombination rate of gas with overdensities @xmath159 .
this enables us to separate recombinations in the igm from recombinations inside galaxies ( see also @xcite ) .
the latter are typically parametrized by the ionizing escape fraction .
alternative definitions of the clumping factor are sometimes employed , e.g. , by applying additional selection criteria to identify the ionized igm ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) . however , for typical reionization scenarios , these definitions , which are all designed to achieve the same objective , i.e. , locating the ionized gas in the igm , generally yield very similar results ( see , e.g. , the discussion in @xcite ) .
our simulations likely underestimate the clumping factor before reionization , since the jeans mass in the unheated igm is unresolved ( e.g. , @xcite ) .
thus , the positive radiative feedback from photoheating on reionization is also underestimated .
note however , as shown in @xcite , that the clumping factor at @xmath160 is insensitive to the redshift at which reionization occurs for reionization at @xmath161 . as already found in @xcite , sne move gas from galaxies to the igm , and this leads to a slight increase in the clumping factor ( see also @xcite ) .
the right - hand panel of figure [ fig8 ] shows that the clumping factor derived from the high - res simulation l12n512 is consistent with that in simulation r9l6n256 of @xcite , which had similar resolution but assumed that the igm is heated by a uniform uv background turned on instantaneously at @xmath135 .
this agreement likely results because in our rt simulations , reionization occurs sufficiently rapidly such that differences in the time at which individual regions inside the simulated volume are reionized are small and similar to the time it takes for the igm to respond dynamically to the increase in the jeans mass in the simulation in which the gas is instantaneously exposed to a uniform ionizing background .
the slight increase in the clumping factor by @xmath162 with respect to the earlier work is mostly due to the inclusion of sn feedback , although , as shown in @xcite , the differences in cosmological parameters also contribute . for reionization occurring at redshifts @xmath163 , the igm clumping factor @xmath158 at @xmath160 is numerically converged and insensitive to a further increase in resolution and box size ( @xcite ) .
in this section we briefly discuss how changes in physical parameters affect the outcome of our simulations and also mention some of the main physical processes our simulations have ignored .
figure [ fig9 ] shows how variations in the physical parameters impact the reionization history and the ability of galaxies to form stars . towards this end
we have repeated the small - box simulation l12n256 , which has the same resolution as the reference simulation l25n512 but allows for a computationally more efficient exploration of the parameter space , with different values for some of the parameters .
one of the main parameters of reionization simulations is the escape fraction @xmath60 of the unresolved ism .
decreasing this fraction from @xmath164 to @xmath165 and @xmath166 causes reionization to occur at lower redshifts , as expected , at which it extends over longer times , in good agreement with similar parameter studies by @xcite and @xcite . since photoheating is delayed , low - mass haloes can continue to form stars more efficiently down to lower redshifts
however , the associated increase in the ionizing emissivity is smaller than the reduction due to the decrease in the sub - resolution escape fraction , and so the impact on reionization due to the change in the minimum mass of haloes is small ( e.g. , @xcite ) . in the current simulations , adopting an escape fraction @xmath167 yields excellent agreement with observations of the evolution of the ionized fraction and the hydrogen photoionization rate at @xmath10 ( but observational constraints are weak ; see the discussion in section [ sec : reionization ] ) .
the amount of energy injected in each photoionization depends on the spectrum of the radiation sources and requires multi - frequency rt simulations for an accurate computation including spectral hardening ( e.g. , @xcite ; @xcite ; @xcite ) .
here we treat this energy as a parameter ( e.g. , @xcite ) . increasing it implies a slight acceleration in reionization , and a slight decrease in the mean neutral fraction after reionization .
this is caused by the increase in the gas temperatures implied by the higher photoheating rates , which in turn decreases the rate at which hydrogen recombines ( e.g. , @xcite ; @xcite ; @xcite ) .
finally , the increased gas temperature increases the negative impact of photoheating on the efficiency of low - mass galaxies to form stars , raising the mass scale below which star formation is strongly suppressed ( see also , e.g. , @xcite ) .
finally , we have carried out a preliminary comparison with results from a new set of simulations of reionization similar to those presented here , which feature increased physical realism and span a wider range of box sizes and resolutions .
these simulations will be discussed elsewhere in more detail and include , among others , helium chemistry and cooling / heating , and feedback and metal enrichment from agb stars , core - collapse and type ia sne , and metallicity - dependent population synthesis .
the treatment of helium was made feasible by replacing the explicit chemistry solver described in sec .
[ sec : chemistry ] with the implicit solver described in @xcite , which is faster .
moreover , the simulations are designed , by calibrating the sub - resolution escape fraction and the sn energy fraction , to match the observed uv luminosity functions and to exhibit similar reionization histories independent of resolution
. a simulation employing the same resolution as our reference simulation here and adopting a sub - resolution escape fraction of @xmath168 , yields reionization and sfr histories similar to those in the current simulation l12n256 adopting an escape fraction @xmath169 .
our simulations ignored a number of potentially relevant physical processes .
most importantly , perhaps , our simulations ignored the chemistry of and radiative cooling by molecular hydrogen , effectively assuming a soft uv background that prevents the build - up of hydrogen molecules
. this approximation fails at the earliest stages of reionization , where it artificially prevents the formation of stars inside low - mass minihaloes ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) .
this early population of stars may preionize the igm and provide a significant feedback on subsequent star formation and reionization ( e.g. , @xcite ; @xcite ) .
our simulations have also ignored the enrichment with metals that accompanies the explosion of stars as sne .
metal - enrichment affects the rates at which gas cools and forms stars ( e.g. , @xcite ; @xcite ) , which may be especially important in low - mass minihaloes in which radiative cooling by atomic hydrogen is suppressed .
finally , we have only followed the radiative feedback from the relatively soft ionizing radiation emitted by metal - enriched stars .
other sources of radiation , such as zero - metallicity stars ( pop iii stars ) or x - ray emitting black holes , may be an important source of ionization and feedback during reionization ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
we have carried out a suite of cosmological radiation - hydrodynamical simulations of galaxy formation during reionization .
the reference simulation was run in a box of size @xmath170 and contained @xmath171 dark matter and @xmath171 baryonic particles , thus resolving atomically cooling haloes with at least @xmath172 dark matter particles .
simulations using both larger and smaller boxes and higher and lower resolution allowed us to investigate the numerical convergence of our results .
simulations in which either sne or photoionization heating or both are turned off , enabled us to isolate and investigate the impact of feedback from star formation .
ionizing photons were transported using accurate and spatially adaptive rt , tracking the growth of ionized regions and the build - up of an ionizing background at the native high resolution at which the hydrodynamics was carried out .
current cosmological simulations lack both the resolution and the physics to provide an ab initio description of the structure of the interstellar gas and the rate at which the gas cools .
this necessitates the use of physically motivated but resolution - dependent parameters to control the energy that each sn injects and the fraction of ionizing photons that escape into the igm .
sfrs and reionization histories are sensitive to these parameters and this impedes the use of cosmological simulations in predicting these quantities from first principles . on the other hand , one may exploit this sensitivity and choose parameters for which simulated sfrs and reionization histories are consistent with current observational constraints , and investigate the implications of such observationally supported models of galaxy formation .
here we have focused on the role of feedback from sne and photoheating , two processes that critically shape galaxy formation and reionization .
our reference simulation yields sfr densities and a uv luminosity function in excellent agreement with observational constraints and completes reionization by @xmath155 . increasing the resolution leads to a strong increase in the cosmic sfr at high redshifts as it facilitates the condensation of gas into low - mass galaxies .
it leads to a mild decrease in the cosmic sfr at late times , when star formation is strongly regulated by stellar feedback . as a consequence , near the end of reionization , our high - res simulation yields a slightly smaller normalization of the uv luminosity function and our low - res simulation yields a slightly larger normalization of the uv luminosity function than our reference simulation and observations . because of the initial increase in the cosmic sfr , increasing the resolution also increases the redshift at which the igm is reionized .
increasing the size of the simulation box above @xmath173 has only a minor impact on the sfr and reionization histories .
photoheating reduces the baryon fractions and suppresses star formation primarily in haloes with masses below @xmath174 .
sne , on the other hand , reduce the baryon fractions and suppress star formation primarily in haloes more massive than @xmath175 .
therefore , the currently observable cosmic sfr is more strongly suppressed by sne than by photoheating , and sn feedback alone is sufficient to match observational constraints on the uv luminosity function .
the inefficiency of sne in the lowest mass galaxies is primarily a consequence of the lack of low - temperature gas physics and the limited resolution . nevertheless , the feedback from sne is sufficiently strong to mask the impact of photoheating on the abundance of low - mass atomically cooling star - forming galaxies .
we thus do not find a noticeable signature imprinted by reionization heating on the uv luminosity function , although we note that the resolution of our simulations is insufficient to model star - forming minihaloes that would be more strongly affected by photoheating . despite the relatively small impact on the cosmic sfr
, photoheating is an important feedback process .
first , photoheating amplifies the ability of sne to suppress star formation .
this amplification is nonlinear and mutual , demonstrating the need to simultaneously account for both feedback processes .
second , photoheating smooths out gas density fluctuations in the igm and thereby strongly reduces the igm recombination rate .
this makes it easier to keep the gas ionized , which facilitates reionization .
in contrast , the net impact of sne on reionization is small .
sne strongly suppress sfrs and slightly increase the igm recombination rate as gas is moved from the galaxies to the igm .
however , this leads only to a small delay in the timing of reionization , possibly because sne create additional low - density channels in the ism through which ionizing photons can escape , which increases the escape fraction of ionizing radiation , or because reionization is driven by the lowest mass galaxies in which sn feedback is inefficient in our simulations . in summary
, our work demonstrates that both photoheating by the stellar radiation that reionizes the universe and the explosion of massive stars as sne may have had a strong impact on structure formation and reionization in the first billion years .
we are grateful to volker springel for letting us use gadget and the halo finder subfind .
we thank ali rahmati , milan raicevic , myoungwon jeon , craig booth , and volker bromm for useful discussions , and we thank benedetta ciardi for a careful reading of an early draft .
we further thank ali rahmati for help with executing some of the simulations used in figure 9 .
we thank rychard bouwens for providing us with his upwards corrections of uv luminosity densities used in figure [ fig2 ] , and we thank kenneth duncan for his tables of observed sfrs and uv luminosity functions provided to us in electronic form .
computer resources for this project have been provided by the gauss centre for supercomputing / leibniz supercomputing centre under grant : pr83le .
we further acknowledge prace for awarding us access to resource supermuc based in germany at lrz garching ( proposal number 2013091919 ) .
some of the simulations presented here were run on odin at the rechenzentrum garching ( rzg ) and the max - planck - institute for astrophysics ( mpa ) and on hydra at the rzg .
this work was sponsored with financial support from the netherlands organization for scientific research ( nwo ) , also through a vidi grant and an nwo open competition grant .
we also benefited from funding from nova , from the european research council under the european unions seventh framework programme ( fp7/2007 - 2013 ) / erc grant agreement 278594-gasaroundgalaxies and from the marie curie training network cosmocomp ( pitn - ga-2009 - 238356 ) .
ahp received funding from the european union s seventh framework programme ( fp7/2007 - 2013 ) under grant agreement number 301096-professor .
cdv benefitted from marie curie reintegration grant perg06-ga-2009 - 256573 . 99 abel t. , norman m. l. , madau p. , 1999
, apj , 523 , 66 abel t. , haehnelt m. g. , 1999 , apj , 520 , l13 abel t. , wise j. h. , bryan g. l. , 2007 , apj , 659 , l87 ahn k. , shapiro p. r. , iliev i. t. , mellema g. , pen u .-
l . , 2009 , apj , 695 , 1430 ahn k. , iliev i. t. , shapiro p. r. , mellema g. , koda j. , mao y. , 2012 , apj , 756 , l16 altay g. , theuns t. , schaye j. , booth c. m. , dalla vecchia c. , 2013 , mnras , 436 , 2689 alvarez m. a. , shapiro p. r. , ahn k. , iliev i. t. , 2006 , apj , 644 , l101 alvarez m. a. , wise j. h. , abel t. , 2009 , apj , 701 , l133 alvarez m. a. , finlator k. , trenti m. , 2012 , apj , 759 , l38 aubert d. ,
teyssier r. , 2008 , mnras , 387 , 295 aubert d. , teyssier r. , 2010 , apj , 724 , 244 baek s. , ferrara a. , semelin b. , 2012 , mnras , 423 , 774 barkana r. , loeb a. , 1999 , apj , 523 , 54 barkana r. , loeb a. , 2000 , apj , 539 , 20 barkana r. , loeb a. , 2001 , phr , 349 , 125 barkana r. , loeb a. , 2004 , apj , 609 , 474 barkana r. , loeb a. , 2006 , mnras , 371 , 395 bate m. r. , burkert a. , 1997 , mnras , 288 , 1060 becker g. d. , bolton j. s. , madau p. , pettini m. , ryan - weber e. v. , venemans b. p. , 2014 ,
arxiv , arxiv:1407.4850 benson a. , venkatesan a. , shull j. m. , 2013 , apj , 770 , 76 bolton j. s. , haehnelt m. g. , 2007 , mnras , 382 , 325 bolton j. s. , becker g. d. , 2009 , mnras , 398 , l26 bolton j. s. , haehnelt m. g. , 2013 , mnras , 429 , 1695 booth c. m. , schaye j. , 2009 , mnras , 398 , 53 bouwens r. j. , et al . , 2011 ,
apj , 737 , 90 bouwens r. j. , et al . , 2014 ,
arxiv , arxiv:1403.4295 bouwens r. j. , et al . , 2014 , apj , 795 , 126 boylan - kolchin m. , bullock j. s. , garrison - kimmel s. , 2014 , mnras , 443 , l44 bromm v. , yoshida n. , 2011 , ara&a , 49 , 373 bruzual g. , charlot s. , 2003 , mnras , 344 , 1000 cai z .- y .
, lapi a. , bressan a. , de zotti g. , negrello m. , danese l. , 2014 , apj , 785 , 65 cen r. , kimm t. , 2014 , apj , 782 , 32 choudhury t. r. , ferrara a. , 2007 , mnras , 380 , l6 ciardi b. , stoehr f. , white s. d. m. , 2003 , mnras , 343 , 1101 ciardi b. , ferrara a. , 2005 , ssrv , 116 , 625 ciardi b. , scannapieco e. , stoehr f. , ferrara a. , iliev i. t. , shapiro p. r. , 2006 , mnras , 366 , 689 ciardi b. , bolton j. s. , maselli a. , graziani l. , 2012 , mnras , 423 , 558 coe d. , et al . , 2013 , apj , 762 , 32 dalla vecchia c. , schaye j. , 2012 , mnras , 426 , 140 dalla vecchia c. , schaye j. , 2008 , mnras , 387 , 1431 dekel a. , silk j. , 1986 , apj , 303 , 39 dijkstra m. , haiman z. , rees m. j. , weinberg d. h. , 2004 , apj , 601 , 666 dijkstra m. , mesinger a. , wyithe j. s. b. , 2011 , mnras , 414 , 2139 dijkstra m. , 2014 , arxiv , arxiv:1406.7292 di matteo t. , springel v. , hernquist l. , 2005 , natur , 433 , 604 dolag k. , borgani s. , murante g. , springel v. , 2009 , mnras , 399 , 497 dove j. b. , shull j. m. , ferrara a. , 2000 , apj , 531 , 846 duffy a. r. , wyithe j. s. b. , mutch s. j. , poole g. b. , 2014 , arxiv , arxiv:1405.7459 duncan k. , et al . , 2014 , mnras , 444 , 2960 ellis r. s. , et al . , 2013 , apj , 763 , l7 emberson j. d. , thomas r. m. , alvarez m. a. , 2013 , apj , 763 , 146 fan x. , carilli c. l. , keating b. , 2006 , ara&a , 44 , 415 fan x. , et al . , 2006 , aj , 132 , 117 ferrara a. , loeb a. , 2013 , mnras , 431 , 2826 fontanot f. , cristiani s. , vanzella e. , 2012 , mnras , 425 , 1413 finkelstein s. l. , et al . , 2012 , apj , 758 , 93 finkelstein s. l. , et al . , 2014 , arxiv , arxiv:1410.5439 finlator k. , zel f. , dav r. , oppenheimer b. d. , 2009 , mnras , 400 , 1049 finlator k. , dav r. , zel f. , 2011 , apj , 743 , 169 finlator k. , 2012 , arxiv , arxiv:1203.4862 finlator k. , oh s. p. , zel f. , dav r. , 2012 , mnras , 427 , 2464 fujita a. , martin c. l. , mac low m .-
, abel t. , 2003 , apj , 599 , 50 furlanetto s. r. , oh s. p. , 2005
, mnras , 363 , 1031 furlanetto s. r. , mcquinn m. , hernquist l. , 2006 , mnras , 365 , 115 furlanetto s. r. , oh s. p. , briggs f. h. , 2006 , phr , 433 , 181 geen s. , slyz a. , devriendt j. , 2013 , mnras , 429 , 633 gnedin n. y. , ostriker j. p. , 1997 ,
apj , 486 , 581 gnedin n. y. , 2000 , apj , 535 , 530 gnedin n. y. , 2000b , apj , 542 , 535 gnedin n. y. , abel t. , 2001 , newa , 6 , 437 gnedin n. y. , fan x. , 2006 , apj , 648 , 1 gnedin n. y. , kravtsov a. v. , chen h .- w . , 2008 ,
apj , 672 , 765 gnedin n. y. , 2010 , apj , 721 , l79 gnedin n. y. , 2014 , apj , 793 , 29 gnedin n. y. , kaurov a. a. , 2014 , apj , 793 , 30 greif t. h. , bromm v. , 2006 , mnras , 373 , 128 greif t. h. , johnson j. l. , klessen r. s. , bromm v. , 2008 , mnras , 387 , 1021 haiman z. , rees m. j. , loeb a. , 1997 , apj , 476 , 458 haiman z. , 2011 , natur , 472 , 47 hambrick d. c. , ostriker j. p. , johansson p. h. , naab t. , 2011 , mnras , 413 , 2421 hasegawa k. , umemura m. , 2010 , mnras , 407 , 2632 hasegawa k. , semelin b. , 2013 , mnras , 428 , 154 hopkins p. f. , keres d. , onorbe j. , faucher - giguere c .- a .
, quataert e. , murray n. , bullock j. s. , 2013 , arxiv , arxiv:1311.2073 iliev i. t. , scannapieco e. , shapiro p. r. , 2005 , apj , 624 , 491 iliev i. t. , et al . , 2006 , mnras , 371 , 1057 iliev i. t. , mellema g. , pen u .-
, merz h. , shapiro p. r. , alvarez m. a. , 2006 , mnras , 369 , 1625 iliev i. t. , mellema g. , shapiro p. r. , pen u .-
l . , 2007 , mnras , 376 , 534 iliev i. t. , et al . , 2009 , mnras , 400 , 1283 iliev i. t. , mellema g. , ahn k. , shapiro p. r. , mao y. , pen u .-
l . , 2014 ,
mnras , 439 , 725 jaacks j. , choi j .- h . , nagamine k. , thompson r. , varghese s. , 2012 , mnras , 420 , 1606 jappsen a .-
, klessen r. s. , glover s. c. o. , mac low m .-
, 2009 , apj , 696 , 1065 jensen h. , laursen p. , mellema g. , iliev i. t. , sommer - larsen j. , shapiro p. r. , 2013 , mnras , 428 , 1366 jeon m. , pawlik a. h. , bromm v. , milosavljevi m. , 2014 , mnras , 440 , 3778 jeon m. , pawlik a. h. , bromm v. , milosavljevic m. , 2014 , 2014 , mnras , 444 , jeon m. , bromm v. , pawlik a. h. , milosavljevi m. , 2015 , submitted ( arxiv:1501.01002 ) johnson j. l. , greif t. h. , bromm v. , klessen r. s. , ippolito j. , 2009 , mnras , 399 , 37 johnson j. l. , khochfar s. , 2011 , apj , 743 , 126 kaurov a. a. , gnedin n. y. , 2014 , apj , 787 , 146 kennicutt r. c. , jr . , 1998 , apj , 498 , 541 kimm t. , cen r. , 2014 , apj , 788 , 121 knevitt g. , wynn g. a. , power c. , bolton j. s. , 2014 , mnras , 445 , 2034 kohler k. , gnedin n. y. , hamilton a. j. s. , 2007 , apj , 657 , 15 komatsu e. , et al . , 2011 , apjs , 192 , 18 kuhlen m. , faucher - gigure c .- a . , 2012 , mnras , 423 , 862 kuhlen m. , madau p. , krumholz m. r. , 2013 , apj , 776 , 34 lewis a. , bridle s. , 2002 , phrvd , 66 , 103511 lidz a. , zahn o. , mcquinn m. , zaldarriaga m. , hernquist l. , 2008 , apj , 680 , 962 loeb a. , 2009 , jcap , 3 , 22 loeb a. , 2010 , hdfs.book , madau p. , haardt f. , rees m. j. , 1999 , apj , 514 , 648 madau p. , rees m. j. , volonteri m. , haardt f. , oh s. p. , 2004
, apj , 604 , 484 maio u. , ciardi b. , dolag k. , tornatore l. , khochfar s. , 2010 , mnras , 407 , 1003 maselli a. , ferrara a. , ciardi b. , 2003 , mnras , 345 , 379 maselli a. , ciardi b. , kanekar a. , 2009 , mnras , 393 , 171 mashian n. , loeb a. , 2013 , jcap , 12 , 17 mcgreer i. d. , mesinger a. , fan x. , 2011 , mnras , 415 , 3237 mcleod d. j. , mclure r. j. , dunlop j. s. , robertson b. e. , ellis r. s. , targett t. t. , 2014 , arxiv , arxiv:1412.1472 mcquinn m. , lidz a. , zahn o. , dutta s. , hernquist l. , zaldarriaga m. , 2007 , mnras , 377 , 1043 mellema g. , iliev i. t. , alvarez m. a. , shapiro p. r. , 2006 , newa , 11 , 374 mellema g. , et al . , 2013 , exa , 36 , 235 mesinger a. , furlanetto s. , cen r. , 2011 , mnras , 411 , 955 mesinger a. , aykutalp a. , vanzella e. , pentericci l. , ferrara a. , dijkstra m. , 2014 , arxiv , arxiv:1406.6373 miralda - escud j. , haehnelt m. , rees m. j. , 2000 , apj , 530 , 1 mitra s. , ferrara a. , choudhury t. r. , 2013 , mnras , 428 , l1 morales m. f. , wyithe j. s. b. , 2010 , ara&a , 48 , 127 muoz j. a. , loeb a. , 2011 , apj , 729 , 99 muratov a. l. , gnedin o. y. , gnedin n. y. , zemp m. , 2013 , apj , 773 , 19 nagamine k. , choi j .- h . , yajima h. , 2010 , apj , 725 , l219 nakamoto t. , umemura m. , susa h. , 2001 , mnras , 321 , 593 natarajan a. , yoshida n. , 2014 , ptep , 2014 , 6112 noh y. , mcquinn m. , 2014 , arxiv , arxiv:1401.0737 norman m. l. , reynolds d. r. , so g. c. , harkness r. p. , 2013 ,
arxiv , arxiv:1306.0645 oesch p. a. , et al . , 2013 ,
apj , 773 , 75 oesch p. a. , et al . , 2014 , apj , 786 , 108 okamoto t. , gao l. , theuns t. , 2008 , mnras , 390 , 920 okamoto t. , yoshikawa k. , umemura m. , 2012 , mnras , 419 , 2855 osterbrock d. e. , ferland g. j. , 2006 , agna.book , ota k. , et al . , 2008 ,
apj , 677 , 12 ouchi m. , et al . , 2010 ,
apj , 723 , 869 paardekooper j .-
, pelupessy f. i. , altay g. , kruip c. j. h. , 2011 , a&a , 530 , a87 paardekooper j .-
p . , khochfar s. , dalla vecchia c. , 2013 , mnras , 429 , l94 pawlik a. h. , schaye j. , 2008 , mnras , 389 , 651 pawlik a. h. , schaye j. , 2009 , mnras , 396 , l46 pawlik a. h. , schaye j. , van scherpenzeel e. , 2009 , mnras , 394 , 1812 pawlik a. h. , schaye j. , 2011 , mnras , 412 , 1943 pawlik a. h. , milosavljevi m. , bromm v. , 2011 , apj , 731 , 54 pawlik a. h. , milosavljevi m. , bromm v. , 2013 , apj , 767 , 59 petkova m. , springel v. , 2011 , mnras , 412 , 935 planck collaboration , 2014 , a&a , 571 , aa16 planck collaboration , 2015 , arxiv , arxiv:1502.01589 pritchard j. r. , loeb a. , 2012 , rpph , 75 , 086901 rahmati a. , pawlik a. h. , raievi m. , schaye j. , 2013 , mnras , 430 , 2427 rahmati a. , schaye j. , pawlik a. h. , raievi m. , 2013 , mnras , 431 , 2261 raievi m. , theuns t. , lacey c. , 2011 , mnras , 410 , 775 raievi m. , pawlik a. h. , schaye j. , rahmati a. , 2014 , mnras , 437 , 2816 razoumov a. o. , norman m. l. , abel t. , scott d. , 2002 , apj , 572 , 695 razoumov a. o. , sommer - larsen j. , 2006 , apj , 651 , l89 rees m. j. , 1986 , mnras , 218 , 25p ricotti m. , gnedin n. y. , shull j. m. , 2001 , apj , 560 , 580 ricotti m. , ostriker j. p. , 2004b
, mnras , 350 , 539 ricotti m. , ostriker j. p. , 2004
, mnras , 352 , 547 ricotti m. , ostriker j. p. , gnedin n. y. , 2005 , mnras , 357 , 207 ricotti m. , gnedin n. y. , 2005 , apj , 629 , 259 ricotti m. , gnedin n. y. , shull j. m. , 2008 , apj , 685 , 21 ricotti m. , 2010 , adast , 2010 robertson b. e. , ellis r. s. , dunlop j. s. , mclure r. j. , stark d. p. , 2010 ,
natur , 468 , 49 robertson b. e. , et al . , 2013 , apj , 768 , 71 rosdahl j. , blaizot j. , aubert d. , stranex t. , teyssier r. , 2013 , mnras , 436 , 2188 salpeter e. e. , 1955 , apj , 121 , 161 sawala t. , et al . , 2014 , arxiv , arxiv:1404.3724 scannapieco c. , et al . , 2012 , mnras , 423 , 1726 schaye j. , 2001 , apj , 559 , 507 schaerer d. , 2003 , a&a , 397 , 527 schaye j. , 2004 , apj , 609 , 667 schaye j. , dalla vecchia c. , 2008 , mnras , 383 , 1210 schaye j. , et al . , 2010 , mnras , 402 , 1536 schaye j. , et al . , 2015 , mnras , 446 , 521 schroeder j. , mesinger a. , haiman z. , 2013 , mnras , 428 , 3058 shapiro p. r. , giroux m. l. , babul a. , 1994 , apj , 427 , 25 shull j. m. , venkatesan a. , 2008 , apj , 685 , 1 shull j. m. , harness a. , trenti m. , smith b. d. , 2012 , apj , 747 , 100 smit r. , bouwens r. j. , franx m. , illingworth g. d. , labb i. , oesch p. a. , van dokkum p. g. , 2012 , apj , 756 , 14 so g. c. , norman m. l. , reynolds d. r. , wise j. h. , 2014 , apj , 789 , 149 sobacchi e. , mesinger a. , 2014 , mnras , 440 , 1662 springel v. , white s. d. m. , tormen g. , kauffmann g. , 2001 , mnras , 328 , 726 springel v. , hernquist l. , 2002 , mnras , 333 , 649 springel v. , 2005 , mnras , 364 , 1105 stiavelli m. , fall s. m. , panagia n. , 2004 , apj , 610 , l1 tanaka s. , yoshikawa k. , okamoto t. , hasegawa k. , 2014 , arxiv , arxiv:1410.0763 taylor j. , lidz a. , 2014 , mnras , 437 , 2542 thoul a. a. , weinberg d. h. , 1996 , apj , 465 , 608 trac h. , cen r. , 2007 , apj , 671 , 1 trac h. y. , gnedin n. y. , 2011 , asl , 4 , 228 trenti m. , stiavelli m. , bouwens r. j. , oesch p. , shull j. m. , illingworth g. d. , bradley l. d. , carollo c. m. , 2010 , apj , 714 , l202 verner d. a. , ferland g. j. , korista k. t. , yakovlev d. g. , 1996 , apj , 465 , 487 vogelsberger m. , genel s. , sijacki d. , torrey p. , springel v. , hernquist l. , 2013 , mnras , 436 , 3031 volonteri m. , gnedin n. y. , 2009 , apj , 703 , 2113 white s. d. m. , 1996 , in schaeffer r. , silk j. , spiro m. , & zinn- justin j. ed . , cosmology and large scale structure formation and evolution of galaxies , p. 349
wiersma r. p. c. , schaye j. , smith b. d. , 2009 , mnras , 393 , 99 wise j. h. , abel t. , 2005 , apj , 629 , 615 wise j. h. , abel t. , 2008 , apj , 685 , 40 wise j. h. , turk m. j. , abel t. , 2008 , apj , 682 , 745 wise j. h. , cen r. , 2009 , apj , 693 , 984 wise j. h. , demchenko v. g. , halicek m. t. , norman m. l. , turk m. j. , abel t. , smith b. d. , 2014 , arxiv , arxiv:1403.6123 wolcott - green j. , haiman z. , bryan g. l. , 2011 , mnras , 418 , 838 wyithe j. s. b. , bolton j. s. , 2011 , mnras , 412 , 1926 wyithe j. s. b. , loeb a. , 2013 , mnras , 428 , 2741 xu h. , ahn k. , wise j. h. , norman m. l. , oshea b. w. , 2014 , apj , 791 , 110 yajima h. , umemura m. , mori m. , nakamoto t. , 2009 , mnras , 398 , 715 yajima h. , choi j .- h . , nagamine k. , 2011 , mnras , 412 , 411 yajima h. , choi j .- h . , nagamine k. , 2012 , mnras , 427 , 2889 zackrisson e. , rydberg c .- e . , schaerer d. , stlin g. , tuli m. , 2011 , apj , 740 , 13 zaroubi s. , et al . , 2012 , mnras , 425 , 2964 zaroubi s. , 2013 , assl , 396 , 45
here we extend the discussion of the galaxy properties in our reference simulation in section [ sec : galprop ] ( figure [ fig6 ] ) with a brief investigation of the dependence on resolution .
figure [ figa1 ] shows that in our high - res simulation , in the absence of sn feedback , reionization suppresses star formation only at masses below @xmath176 .
this is significantly less than at the reference resolution , even though the igm in the two simulations is reionized at similar times .
the decrease in the suppression scale is caused by the increase in the sfrs that , in the absence of feedback , accompanies an increase in resolution , and for which photoheating does not entirely compensate in our simulations .
however , the high - res and reference simulations agree closely on the characteristic scale @xmath177 below which the median sfr is strongly suppressed by the combined action of radiative and sn feedback . in the low - res simulation ,
star formation is suppressed in haloes as massive as @xmath178 , primarily by the limited resolution . | we present a suite of cosmological radiation - hydrodynamical simulations of the assembly of galaxies driving the reionization of the intergalactic medium ( igm ) at @xmath0 .
the simulations account for the hydrodynamical feedback from photoionization heating and the explosion of massive stars as supernovae ( sne ) .
our reference simulation , which was carried out in a box of size @xmath1 using @xmath2 particles , produces a reasonable reionization history and matches the observed uv luminosity function of galaxies .
simulations with different box sizes and resolutions are used to investigate numerical convergence , and simulations in which either sne or photoionization heating or both are turned off , are used to investigate the role of feedback from star formation . ionizing radiation is treated using accurate radiative transfer at the high spatially adaptive resolution at which the hydrodynamics is carried out .
sn feedback strongly reduces the star formation rates ( sfrs ) over nearly the full mass range of simulated galaxies and is required to yield sfrs in agreement with observations .
photoheating helps to suppress star formation in low - mass galaxies , but its impact on the cosmic sfr is small . because the effect of photoheating is masked by the strong sn feedback , it does not imprint a signature on the uv galaxy luminosity function , although we note that our resolution is insufficient to model star - forming minihaloes cooling through molecular hydrogen transitions .
photoheating does provide a strong positive feedback on reionization because it smooths density fluctuations in the igm , which lowers the igm recombination rate substantially .
our simulations demonstrate a tight non - linear coupling of galaxy formation and reionization , motivating the need for the accurate and simultaneous inclusion of photoheating and sn feedback in models of the early universe .
[ firstpage ] cosmology : reionization methods : numerical radiative transfer galaxies : high - redshift intergalactic medium hii regions |
synthetic biology is in need of theoretical tools to design circuits with a high probability of exhibiting specific , user - defined dynamical behaviors .
conventionally , efforts in this regard have focused on explicit kinetic formulations of the dynamics of a circuit . indeed , deterministic models of the toggle switch @xcite and the repressilator @xcite , two early examples of successful synthetic circuits , directly predicted the dynamical behavior of these constructs and likely influenced the final design of the circuits themselves .
perhaps the most prominent challenge in assembling these models was a fundamental lack of knowledge regarding kinetic details , including the mechanistic form of the rate laws , as well as the precise values of rate constants themselves .
the primary focus of this work is to develop a modeling framework for understanding the dynamics of a biological circuit for which we have little detailed knowledge of the true kinetics . in particular
, we will try to derive generic results dependent exclusively on the `` topology '' of the circuit , rather than its detailed kinetics . to do so
, we will focus our efforts on a particular system known as the metabolator , a synthetic circuit constructed in _ escherichia coli k12 _ which exhibits limit cycle oscillations . at a coarse level of description ,
the metabolator consists of two metabolites which are interconverted between each other by two enzymes .
one of the metabolites in the circuit acts as a regulator , repressing the expression of the enzyme catalyzing the metabolite s production and activating the expression of the enzyme catalyzing its degradation . in ref .
@xcite , the authors observed that in conditions of sufficiently high metabolic inflow to the metabolator , a gene encoding one of the regulated enzymes exhibited sustained oscillations .
the metabolator is particularly interesting because it remains ( to our knowledge ) the only example of a synthetic circuit containing a combination of both metabolic and genetic ( transcriptional / translational ) components . in all other cases ,
the building blocks of circuits have been entirely transcriptional and translational .
the `` hybrid '' nature of the metabolator is particularly interesting because metabolism is well - known to take place at a faster rate than transcription and translation .
the role such separation of time scales actually plays in the dynamics of the metabolator is unresolved , although it has been observed to be quite in important in other metabolic contexts @xcite .
one of our primary approaches to studying the metabolator is based on a non - dimensionalization technique for studying dynamical systems commonly referred to as generalized modeling ( gm ) @xcite . in gm ,
a change of variables is applied to a dynamical system so that many of the otherwise unconstrained parameters in the system ( _ e.g. _ , rate constants ) are replaced by `` elasticity '' parameters with well - defined ranges .
then , the dynamics of the system can be studied around an arbitrary and potentially unknown steady - state .
gm s conclusions are invariant to the particular choice of rate law , and depend only on the sensitivity of the rate law at the system s equilibrum .
this enables the study of a _ topological _ model of the metabolator , in which no assumptions are generally made on the explicit form of the rate law , nor the true values of the parameters in the system . in practice , many gm studies have focused on using large - scale numerical simulations to tease out statistical properties relating elasticities to stability properties of the system . in this work ,
we supplement such numerical work with complementary analytical studies , and highlight the additional insight that can be gained through explicit calculations .
we complement our use of gm with explicit dynamical modeling of the metabolator , and highlight the complementary conclusions that can be drawn from the two techniques ( e.g. local vs. global bifurcation phenomena when applying gm or explicit dynamical modeling , respectively ) .
importantly , we demonstrate how the results from gm can prompt further study with explicit modeling , and vice versa .
the main results of the work are as follows .
first , a simple , `` core '' model of the metabolator ( two genes , two metabolites ) originally illustrated in ref .
@xcite is investigated using a conventional dynamical model .
we find that the system lacks the capability to pass through a hopf bifurcation .
second , using gm , we analytically prove that _ any _ dynamical model with a topology analogous to the core metabolator is incapable of hopf bifurcations . concluding that additional topological elements are necessary for a hopf bifurcation , we use numerical simulations with gm to progressively restore elements of the metabolator model from ref .
@xcite which we omitted in our core model .
doing so , we reveal the crucial topological components of the metabolator which endow it with the ability to oscillate .
third , we investigate the possibility of oscillations in an explicit model of a simply `` re - wired '' version of the metabolator in which the regulatory connections are reversed ( the reverse metabolator , `` rm '' ) .
we show that the rm is capable of sustained oscillations by both a local hopf bifurcation as well as global fold of limit cycles .
fourth , prompted by findings in the explicit model and using the gm tools we developed earlier , we study the role that metabolic and genetic time scales play in generating instability and potential oscillations in the rm _ in vivo_. -[sec : skm ] .
metabolite @xmath0 represses the transcription of gene @xmath1 and activates the transcription of gene @xmath2 .
( b ) core metabolator with intermediary delay gene @xmath3 , and ( c ) core metabolator with un - lumped metabolites , both studied in section [ sec : alt ] .
( d ) reverse metabolator , studied in section [ sec : rm ] . in contrast to the core metabolator
, @xmath0 activates the transcription of @xmath1 and represses the transcription of @xmath2.,title="fig : " ] [ fig : metabolator ]
to begin to understand the fundamental components which generate sustained oscillations in the metabolator , we developed a four - dimensional dynamical model of the system .
our particular choice of design was motivated by the schematic provided in ref .
@xcite and re - illustrated in [ fig : metabolator ] a. two enzymes ( @xmath1 and @xmath2 ) interconvert two substrates ( @xmath4 and @xmath0 ) .
gene @xmath1 codes for the enzyme which converts @xmath4 into @xmath0 , and gene @xmath2 codes for the enzyme which converts @xmath0 into @xmath4 .
we explicitly assume that the rate of gene expression is substantially faster than protein translation , so that the abundances of the mrna transcripts encoding @xmath1 and @xmath2 may be assumed to be in quasi - steady - state .
metabolite @xmath0 represses the expression of gene @xmath1 and activates the expression of gene @xmath2 .
furthermore , we assume that the rate of inflow of @xmath4 is constant , and the rate of outflow of @xmath0 is in proportion to its concentration . because of the symmetry in the dynamics of @xmath4 and @xmath0 ( both are produced and degraded by the same metabolic reactions ) , we replace @xmath4 with the total amount of substrate in the system @xmath5 .
furthermore , we assume bilinear mass - action rate laws for the dynamics of the metabolic reactions .
we dub our simplified model the _ core metabolator_. the nondimensional equations describing the dynamics of the core metabolator are @xmath6 @xmath7 @xmath8 @xmath9 [ eq : met_linear ] where the parameters @xmath10 are assumed to be strictly positive .
see appendix a for a derivation of ( [ eq : met_linear ] ) .
we emphasize that the core metabolator serves as a jumping off - point : observing oscillations here would suggest the presence of oscillations in more detailed models capturing the true topological elements of the _ in vivo _ metabolator . on the other hand , not observing oscillations in this simplified model
implies that one needs to examine a model with more components in order to observe hopf bifurcations , see section 4 . the fixed point of ( [ eq : met_linear ] ) is @xmath11 and @xmath12 .
the jacobian @xmath13 , evaluated at the fixed point , becomes + @xmath14 \label{eq : j_coredim}\ ] ] the characteristic equation for @xmath13 is @xmath15 in order to identify potential candidates for a hopf bifurcation , we make use of a constraint relating the coefficients of ( [ eq : char_met1 ] ) which must be satisfied in order for ( [ eq : char_met1 ] ) to have a pair of purely imaginary roots .
first , we note that ( [ eq : char_met1 ] ) has a root at @xmath16 .
then , we can substitute @xmath17 into the cubic polynomial in ( [ eq : char_met1 ] ) we find that these conditions are @xmath18 , where @xmath19 is the coefficient of the @xmath20th term of the cubic polynomial in parentheses in ( [ eq : char_met1 ] ) . after substituting the appropriate coefficients into these conditions and some algebra ,
we find @xmath21 it is not clear if ( [ eq : condition_met1 ] ) can be satisfied with real , positive values of the system parameters . rewriting ( [ eq : condition_met1 ] ) as a quadratic polynomial in @xmath22 , we obtain @xmath23 now the outcome becomes plainly clear : because ( [ eq : quad_et1 ] ) only contains positive coefficients , descartes rule of signs immediately proves that it can not have positive real roots .
thus , there exists no combination of @xmath24 which will yield a pair of purely imaginary eigenvalues , and the system does not exhibit a hopf bifurcation .
can we say anything more general about the oscillatory capabilities of the core metabolator illustrated in figure 1a ? in the previous section , we showed that one particular realization of this schematic model did not exhibit a hopf bifurcation , but it remains unclear whether this was a quality of our choice of model .
one may naturally wonder if choosing different kinetic rate laws for the metabolic reactions , or different regulatory kinetics for the feedback regulation , would yield qualitatively different behavior like oscillations . in this section ,
we propose a method for addressing these concerns by applying a technique known as generalized modeling ( gm ) .
gm is a nondimensionalization procedure for reformulating the kinetics of a dynamical system , enabling the study of local dynamics near an equilibrium in an efficient way .
to do so , parameters in the system are rewritten in terms of normalized parameters known as _ elasticities_. the elasticities have a direct connection to the original kinetic parameters , but are much easier to work with .
most importantly , elasticities typically have well - defined and limited ranges ( e.g. [ 0,1 ] ) , and sampling them across this range effectively samples all possible values of the original kinetic parameters ( which may otherwise span several potentially unknown orders of magnitude ) . furthermore , these elasticities are not necessarily tied to any single choice of kinetic rate law ; instead , they simply represent the sensitivity of any rate law to small changes in the equilibrium value of a variable , normalized by the steady - state magnitude of that rate law . in the context of metabolism , elasticities are similar to logarithmic derivatives of the rate law with respect to a change in substrate concentration @xcite .
this section is organized as follows .
first , we introduce a motivating example of the gm nondimensionalization procedure .
next , we illustrate the role this non - dimensionalization plays in the formulation of generalized models
. then , we apply this procedure to generate a gm of the core metabolator .
finally , we analyze the stability properties of the gm , and analytically prove that a topological model of the core metabolator with fairly general rate laws is incapable of a hopf bifurcation .
first , we present the normalization procedure upon which gm is based , known as generalized modeling , see for example ref .
consider a system of ordinary differential equations , @xmath25 here , @xmath26 and @xmath27 are both functions of time .
the stability of ( [ eq : motivating_1 ] ) around its equilibrium is determined by the eigenvalues of the jacobian @xmath13 . here , we approach the problem of calculating @xmath13 after first executing a particular change of variables central to gm .
assume that there exists an equilibrium @xmath28 and normalize ( [ eq : motivating_1 ] ) by its steady state concentration ( assuming @xmath29 ) : @xmath30 letting @xmath31 so that the variables are nondimensionalized and the equilibrium is at @xmath32 , we can write @xmath33 to find the components of the jacobian for @xmath34 , we simply differentiate each term by each independent variable and evaluate at the equilibrium @xmath35 .
if we let @xmath36 and write the components of the jacobian as @xmath37 and @xmath38 , we find @xmath39 where @xmath40 and @xmath41 . by repeating the calculations outlined above for @xmath42
, we can calculate the remaining entries of the jacobian .
we argue that these @xmath43 s ( herein referred to as _ elasticities _ and similar to the elasticities from metabolic control analysis @xcite ) have well - defined and limited ranges which makes performing calculations with them much easier than with similar calculations with the original parameters . to demonstrate the utility of elasticities
, we will reformulate the dynamics of @xmath1 in the core metabolator .
to do so , we will begin with the first equation of the dimensionalized dynamics of the metabolator , reproduced from appendix a ( see ( [ eq : dim_g1 ] ) ) , @xmath44 we let @xmath45 and @xmath46 ( where @xmath47 and @xmath48 are the steady - state concentrations of @xmath1 and @xmath0 , respectively ) .
then , we can write @xmath49 we will calculate the elasticity of the first term in ( [ eq : elast1 ] ) corresponding to the production of @xmath50 .
to do so , we normalize this production term by its magnitude at steady state : @xmath51 to calculate the elasticity , we take the derivative of @xmath52 with respect to @xmath53 , evaluated at @xmath54 , @xmath55 upon inspection , this elasticity can only take values in @xmath56 $ ] .
when @xmath57 , the elasticity approaches negative two and the production of @xmath1 becomes very sensitive to small changes in the abundance @xmath48 .
in contrast , when @xmath58 , the elasticity approaches zero and the production of @xmath1 is insensitive to changes in @xmath48 .
the well - defined and limited range of elasticities becomes extremely powerful in studying the dynamics of a system near its equilibrium .
as we show in the ensuing section , these elasticities feature prominently in the jacobian of a dynamical system reformulated with gm .
we will demonstrate how elasticities can be used to prove the generic stability of a _ topological _ model of a dynamical system , with only mild assumptions on the explicit rate laws governing the dynamics .
we can use the aforementioned normalization procedure to develop a generalized model of the core metabolator . as we will show , the generalized model enables us to draw conclusions about a topological model of the core metabolator , while relaxing many of the assumptions we made in our explicit model of the core metabolator in section 2 .
we begin with the following `` generalized '' form of the dynamics : @xmath59 the terms @xmath60 and @xmath61 correspond to the reactions catalyzed by @xmath1 and @xmath2 , respectively , while @xmath62 and @xmath63 correspond to the inflow and outflow of mass from the metabolator .
the @xmath64 and @xmath65 terms correspond to the production and degradation of proteins .
importantly , the @xmath64 terms capture the feedback regulation of metabolite @xmath0 on the two enzymes .
we make a few assumptions regarding the dynamics of the core metabolator in our generalized model .
first , we assume @xmath66 and @xmath65 are generic , monotonically increasing functions of their arguments ( i.e. @xmath67 ) .
many frequently used kinetic rate laws , including hill , michaelis - menten , and mass - action kinetics satisfy this assumption , among many other possible forms of kinetics .
we also retain the earlier assumption that the rates of metabolic reactions are linear with respect to the amount of gene product .
the structure of ( [ eq : skm_core1 ] ) corresponds precisely to the topological structure of the core metabolator in figure 1a .
before we proceed to calculating the jacobian of ( [ eq : skm_core1 ] ) , we briefly describe a simple method for calculating the prefactors ( those parameters which were not elasticities , i.e. @xmath68 ) which appear in the jacobian and were described in the prior section . in prior work on gm @xcite
, it has been shown that for metabolic networks , these prefactors correspond precisely to the rates of metabolic reactions at equilibrium .
as we show below , these rates can be calculated by enforcing that at equilibrium the fluxes of reactions flowing into and out of each metabolite are balanced @xcite .
furthermore , for the non - metabolic components of the metabolator , we can apply a similar balancing procedure by noting that at equilibrium , the production and degradation rates for each protein must be equal . applying these balancing conditions to ( [ eq : skm_core1 ] )
, we find that @xmath69 here , @xmath70 and @xmath71 indicate the steady state concentrations of metabolites 1 and 2 , and gene products 1 and 2 , respectively . the first two constraints in ( [ eq : constraints ] ) force the rate of reaction @xmath72 ( corresponding to the conversion of @xmath4 to @xmath0 ) to be precisely balanced by the rate of reaction @xmath73 ( the conversion of @xmath0 to @xmath4 ) and @xmath62 , the rate of inflow of @xmath4 to the system .
similarly , @xmath73 must be precisely equal to the sum of @xmath72 and @xmath63 ( the outflow of @xmath0 from the system ) .
the final two conditions impose that , for each gene , the rates of production and degradation must be balanced .
because of the interdependency of all the prefactors in ( [ eq : constraints ] ) , it becomes useful to explicitly introduce the parameter @xmath74 .
we also introduce a free parameter @xmath75 by setting @xmath76 . from the equilibrium condition ( [ eq : constraints ] ) , it follows that @xmath77 .
finally , we let @xmath78
. we can now rescale ( [ eq : skm_core1 ] ) in the same manner as ( [ eq : motivating_4 ] ) : @xmath79 where @xmath80 .
the jacobian matrix @xmath13 of the core metabolator is then @xmath81 .
\label{eq : j_core}\ ] ] the parameters @xmath82 in ( [ eq : j_core ] ) correspond to the elasticities of metabolic reactions described earlier : @xmath83 here , @xmath84 follows directly from the earlier assumption that all @xmath85 are monotonically increasing functions of their arguments .
similarly , the parameters @xmath86 in ( [ eq : j_core ] ) , corresponding to elasticities of the regulatory reactions , can be calculated by : @xmath87 for the generalized model of the core metabolator , @xmath88 since @xmath0 inhibits @xmath1 and activates @xmath2 .
furthermore , @xmath89 since an increase in the abundance of a gene product leads to an increase in its rate of degradation .
the utility of reformulating the core metabolator ( [ eq : met_linear ] ) with the gm in ( [ eq : skm_core2 ] ) is now plainly clear .
many of the prior assumptions made on the rate laws in our explicit dynamical model ( [ eq : met_linear ] ) have been relaxed .
our sole assumption is that the rate laws are monotonic , so that it is always the case that the signs of the elasticities are known .
now , by analytically studying the eigenvalues of @xmath13 , we can understand the stability properties of the core metabolator across the entire spectrum of possible parameter values and rate laws . in appendix b
, we show that ( [ eq : j_core ] ) can not have eigenvalues with positive real part for any admissible value of parameters .
the proof in the appendix follows by an application of the routh - hurwitz criterion , a well - known tool from classical control theory @xcite .
since ( [ eq : j_core ] ) does not have any eigenvalues with positive real part , we have proven that all equilibria of the gm of the core metabolator are exponentially stable to small perturbations .
this directly implies that the gm of the core metabolator can not undergo a hopf bifurcation . the main result of this section bears some elaboration : the gm of the core metabolator encompasses all possible explicit realizations of the topological model shown in figure [ fig : metabolator]a , under the general hypotheses about the rate laws stated above . with elementary techniques , it was possible to show that none of these potential realizations could ever pass through a hopf bifurcation .
we note that there may still be global mechanisms , such as a saddle node of limit cycles , that introduce oscillations .
this naturally suggests that some additional component of the topology of the _ in vivo _ metabolator , which has been removed in our model of the core metabolator , must be crucial to the existence of a hopf bifurcation .
this idea is explored further in the following section .
the analysis in sections 2 and 3 is suggestive : which topological simplifications in the core metabolator and in the gm of the core metabolator eliminated the possibility of a hopf bifurcation ? to address this question , we used gm to inspect which components would lead to oscillations when reintroduced into the core metabolator . to do so , successively more complete versions of the _ in vivo _ metabolator
were created , as shown in figure 1b - c .
we developed two hypotheses for the crucial topological features of the metabolator which would be necessary for sustained oscillations .
first , we reasoned that a third gene , labeled @xmath3 and representing laci from ref .
@xcite , might act as an intermediary delay in the negative feedback from @xmath0 to @xmath1 and consequently induce oscillations ( see figure 1b ) .
such delays in feedback have been implicated in the literature as a major source for the onset of oscillations .
second , we separated the lumped metabolite @xmath0 into its appropriate constitutive molecules ( now labeled @xmath0 and @xmath90 ) in figure [ fig : metabolator]c ) . in the revised topology ,
the enzyme transcribed from gene @xmath2 reconverts @xmath90 to @xmath4 , but the regulation of @xmath1 and @xmath2 remains dependent on @xmath0 . for each of our new topological models of the metabolator , the corresponding gm
was assembled ( see appendix c ) and numerical simulations completed to sample the space of parameters in the model . in each instance of the simulation , a random sample from the entire possible space of @xmath43 parameters was drawn . in order to efficiently sample the numerical space , we sampled uniformly from the ranges @xmath91 , \theta_{r1}=[-2,0)$ ] and @xmath92 $ ] for all other elasticities .
since the parameters @xmath93 and @xmath94 remained unconstrained , we decided to sample these parameters across seven orders of magnitude ( from @xmath95 to @xmath96 1/sec , assuming @xmath97 ) .
for each instance of our simulation , the eigenvalues of @xmath13 were calculated .
the process was repeated 20,000 times for each choice of @xmath98 and @xmath94 . in total , over @xmath99 simulations were conducted .
putative hopf bifurcations were detected by examining whether a pair of complex conjugate eigenvalues existed which were sufficiently close to the imaginary axis ( in our case , the magnitude of their real component was required to be less than @xmath95 ) .
the size of the dynamical system ( greater than four state - space variables ) made confirming these results using the routh - hurwitz criterion analytically intractable .
the results of our numerical studies were unexpected .
we found that reintroducing @xmath3 had no bearing on the appearance of a hopf bifurcation .
in fact , the model remained entirely stable , failing to exhibit a single instance of an eigenvalue with positive real part . on the other hand , the refined three - metabolite model did exhibit instability and hopf bifurcations .
these results suggest that the precise nature of the regulation is crucial for the metabolator to exhibit oscillations . in order for a hopf bifurcation to be possible , the regulatory metabolite could not be repressing the expression of its own degrading enzyme .
this delayed positive feedback appears to be a crucial element endowing the metabolator with the potential to undergo a hopf bifurcation .
the difficulty in constructing synthetic circuits with poorly studied components often motivates synthetic biologists to construct several alternative designs . frequently , these alternative designs are simply composed of rewired versions of the same regulatory components . in light of this
, we investigated the dynamical behavior of one such re - wired version of the core metabolator , which we dub the reverse metabolator ( rm ) . in the rm ,
the regulatory connections found in the core metabolator between @xmath0 and the two genes are reversed .
thus , @xmath0 now activates @xmath1 , and hence its own production , and represses @xmath2 , and hence its own degradation ( see figure 1d ) .
we selected this design precisely because synthetic biologists frequently build a number of alternative constructs in the course of circuit assembly , in order to increase the odds of identifying a succesful design . in nondimensional form , the rm is identical to ( [ eq : met_linear ] ) but with the regulation of @xmath50 and @xmath100 reversed : @xmath101 @xmath102 @xmath103 @xmath104 [ eq : revmet ] for the rm , we repeated the explicit dynamical analysis outlined in section 2 .
the fixed points of ( [ eq : revmet ] ) are identical to those of the core metabolator ( @xmath11 and @xmath12 ) , but the jacobian becomes @xmath105.\ ] ] the characteristic equation of the rm is now @xmath106 next , we calculate the necessary condition for ( [ eq : revmet_char ] ) to have a pair of purely imaginary roots ( i.e. that @xmath18 , where @xmath19 is the coefficient of the @xmath20th term of the cubic polynomial in ( [ eq : revmet_char ] ) ) .
assuming that @xmath17 , we find two conditions : @xmath107 equating these two expressions for @xmath108 , we arrive at a quadratic polynomial in @xmath22 @xmath109 using descartes rule of signs , ( [ eq : cond_revmet ] ) must contain positive , real roots since @xmath110 and the @xmath111 term is negative .
therefore , there exists an open set of points @xmath112 at which the reverse metabolator undergoes a hopf bifurcation . using the numerical continuation software package auto @xcite ,
a number of hopf bifurcations spanning several orders of magnitude in @xmath113 , and @xmath114 can be identified . ) . for small values of @xmath114 ,
two limit cycles exist ( a large amplitude unstable cycle , and a small amplitude stable cycle ) , while for large values of @xmath114 , a single , unstable limit cycle exists .
( b ) for @xmath114 = 5.5 ( black ) , 5.6 ( red ) , 5.8 ( green ) , and 6.0 ( blue ) , the unstable limit cycle exhibits increasingly sudden changes in abundance of @xmath0 .
( c ) for @xmath115 , the two limit cycles ( stable , black ; unstable , red ) exhibit smooth changes in amplitude .
for all cases , @xmath116 .
, title="fig : " ] [ fig : timeseries ] we used auto to study the behavior of the rm s limit cycles in different regions of parameter space .
as shown in figure 2 , the limit cycles exhibit two notable features depending on the magnitude of inflow into the rm , characterized by the parameter @xmath114 .
first , for small @xmath114 , two distinct periodic orbits exist : a small amplitude , stable limit cycle and a large amplitude unstable limit cycle .
these two cycles annihilate at a fold of limit cycles . both limit cycles display relatively smooth changes in amplitude ( figure 2c ) .
second , for large values of @xmath114 , unstable periodic solutions showed characteristically sudden changes in amplitude , reminiscent of relaxation oscillations commonly found in neuroscience @xcite . ) .
as @xmath114 and @xmath22 are varied , a curve of hopf bifurcations ( red ) and a fold of limit cycles ( black ) collide at a bautin point .
insets ( b ) , ( c ) , and ( d ) correspond to the dynamics of the rm when @xmath117 , in between the two bautin points .
phase portraits correspond to the dynamics of metabolite @xmath0 ( vertical axis ) and total metabolite @xmath5 ( horizontal axis ) , with green dots indicating initial conditions and red dots indicating final positions .
all simulations were run for a time interval of 1000 .
black curves indicate trajectories approaching a fixed point , blue curves correspond to trajectories approaching a stable limit cycle , and red curves correspond to trajectories spiralling away from an unstable limit cycle.(b ) for small @xmath114 , the rm exhibits dampled , spiralling oscillations to a stable fixed point .
( c ) for @xmath118 , two distinct limit cycles exist .
( d ) for large values of @xmath114 , the stable limit cycle in ( c ) undergoes a hopf bifurcation and creates a stable fixed point , surrounded by a large amplitude unstable limit cycle .
, title="fig : " ] [ fig : bif ] we investigate the stability of the limit cycles generated by the hopf bifurcations in the rm .
limit cycles arising from supercritical hopf bifurcations are stable to small perturbations , while those generated by subcritical hopf bifurcations are not . to explore the criticality of the observed hopf bifurcations , a two - parameter bifurcation analysis of the rm
was conducted in auto , shown in figure 3a .
as the two parameters @xmath114 and @xmath22 were varied , the criticality of the hopf bifurcation was tracked using the magnitude of the leading floquet multiplier .
we observed that as the value of @xmath114 was decreased , the hopf bifurcation switched from being subcritical to being supercritical .
as the value of @xmath114 was decreased further , a second transition in criticality was observed , and the hopf bifurcations returned to being subcritical .
the point at which a hopf bifurcation switches criticality is called a bautin bifurcation .
this higher - order bifurcation is associated with the collision of a fold of limit cycles with a manifold of hopf bifurcations .
using auto , we were able to follow the fold of limit cycles as it branched off from one bautin bifurcation , tracked closely with the manifold of hopf bifurcations , and reintersected with the hopf curve at the second bautin point ( figure 3a ) .
the fold of limit cycles is the boundary between two regions of parameter space where distinct qualitative dynamical behaviors of occur .
on one side of the fold ( in this case , for region b in figure 3 ) , no sustained oscillations exist . on the other side of the fold ( region c in figure 3 )
, a pair of limit cycles appears . for the rm
, this pair contained a small amplitude stable limit cycle and a large amplitude unstable limit cycle .
the existence of a bautin bifurcation indicates there are two distinct routes to oscillation in the reverse metabolator .
first , by decreasing @xmath114 for a fixed @xmath119 ( moving from point d to point c in figure 3a ) , the reverse metabolator passes through a supercritical hopf bifurcation . in this case , oscillations appear at infinitesimal amplitude , and their amplitude grows as @xmath114 is decreased further .
this type of bifurcation is identical to the one observed in the _ in vivo _
metabolator @xcite .
a second , and potentially more interesting , route to oscillation appears if the reverse metabolator passes through a fold of limit cycles . as the value of @xmath114
is increased for fixed @xmath119 ( moving from point b to c in figure 3a ) , a finite - amplitude stable limit cycle appears , surrounded by a larger amplitude unstable limit cycle .
points perturbed from the unstable steady state will immediately be attracted to oscillations that are much larger in magnitude than those observed in the immediate vicinity of a hopf bifurcation .
perturbations outside of the basin of attraction of the stable limit cycle will lead to spiraling instability , and eventually to the extinction of one of the metabolites .
we examine the biological implications of alternative modes of oscillation in the rm in the discussion , section 7 .
the appearance of relaxation oscillations inspired us to investigate the role time scales play in the dynamics of the rm .
the existence of such divergent time scales is well - known in the regulation of metabolism .
the allosteric feedback of small metabolites and the post - translational modification of enzymes target the activity of an enzyme pool at time - scales on the order of microseconds @xcite .
in contrast , the transcriptional regulation of enzyme levels ( which occurs through the binding of protein transcription factors to dna sequences ) takes place at a substantially slower rate , on the order of minutes @xcite , and affects the actual size of the enzyme pool directly .
remarkably , the rate of an enzymatically catalyzed reaction itself , which the above processes ultimately regulate , is known to take place at characteristic rates spanning seven orders of magnitude , from microseconds to minutes @xcite . in order to study the inherent time scales of the rm
, we constructed a generalized model .
this model was identical in structure to ( [ eq : skm_core2 ] ) , except that the signs of the elements corresponding to the transcriptional regulation of @xmath0 were reversed ( @xmath120 ) .
our intent was to use the entries in the jacobian @xmath13 to identify time - scales which characterized the dynamics of metabolites and genes in the rm . to understand the connection between @xmath13 and time - scales , recall that each row in @xmath13 describes the linearized dynamics of a small perturbation away from the equilibrium of one state variable .
for example , the first row of @xmath13 corresponds to the dynamics of a perturbation in @xmath4 : @xmath121 where @xmath122 , and @xmath123 correspond to perturbations in @xmath124 , and @xmath100 , respectively . upon inspection of ( [ eq : deltam1 ] ) ,
all terms on the right hand side share a common divisor @xmath125 which has units of time .
we can treat this shared prefactor as a natural measure of the time scale of @xmath4 . repeating this process for all of the rows in ( [ eq : j_core ] ) , we identified the characteristic time scale for each state space variable .
entries in rows one and two of @xmath13 , corresponding to the dynamics of the two metabolites , shared common divisors @xmath125 and @xmath126 , respectively .
similarly , entries in the final two rows of @xmath13 , corresponding to the genetic elements of the metabolator , shared common divisors @xmath127 and @xmath128 ( again with units of time ) . , and @xmath129 , so that there existed only two time scales , one metabolic and one genetic .
additional numerical investigations indicate results are identical when this assumption is relaxed ( results not shown).,title="fig : " ] [ fig : timescales ] we studied how the stability of @xmath13 depends on the ratio of the metabolic and genetic time scales .
to do so , we generated a large number of random instances of @xmath13 and calculated the real component of the leading eigenvalue @xmath130 of @xmath13 ( a positive value indicated instability , while a negative value indicated stability ) . the sampling was completed by selecting a particular choice of the ratio of metabolic and genetic time scales , and then creating a 30x30 grid of @xmath131 and @xmath132 values . for each point on the grid , we sampled 1000 instances of all other @xmath43 s uniformly from the interval @xmath133 $ ] for each point in the grid .
the results of this simulation are shown in figure 4 .
two major trends regarding the stabiilty of the rm emerged from our simulations .
first , we found that as the dynamics of the genetic elements in the reverse metabolator grew slower , the stability of the circuit as a whole increased .
as the value of @xmath94 grew smaller and eventually became faster than the metabolic time scale , the stability of the circuit was observed to fall significantly .
second , our simulations also revealed that small regulatory elasticities contributed to the stability of the circuit , while large elasticities significantly reduced the probability of a stable circuit .
this observation qualitatively agrees with our first observation relating high values of @xmath134 to instability .
large magnitude regulatory elasticities increase the same terms in the jacobian as @xmath134 .
together , our observations suggest that exceptionally slow regulation in the metabolator stabilizes the circuit , a conclusion that parallels commonly held notions on the speed of regulation and the role it plays in stability @xcite .
contemporary biology is charged with understanding how fundamentally different cellular processes such as signalling , metabolism , and transcriptional regulation , interact and function as a coherent whole .
our work here explored the different types of conclusions which may be drawn from detailed kinetic models and more generic generalized models in the study of a hybrid metabolic - genetic oscillator .
we first showed that the core metabolator ( [ eq : met_linear ] ) has a unique fixed point which is stable for all parameter values ( recall the end of section 2 ) .
hence , the core metabolator can not undergo hopf bifurcation to oscillations .
then , we developed a gm model of the core metabolator ( [ eq : skm_core1 ] ) , and we showed that this system also can not undergo hopf bifurcations by applying the routh - hurwitz criterion to the jacobian ( [ eq : j_core ] ) ( see section 3.2 and appendix b ) .
these analytical results for the core metabolator ( [ eq : met_linear ] ) and its generalized model ( [ eq : skm_core1 ] ) suggested that the network topology and the number of components in the network play an important role in determining whether or not oscillations are possible . as a result ,
in section 4 , we explored two more detailed versions of the core metabolator .
the first of these has an intermediary delay gene @xmath3 ( figure 1b ) , but did not exhibit hopf bifurcations in any of the large number of numerical simulations with randomly chosen system parameters we carried out .
the second of these consists of an extra metabolite , which is also present in the synthetically constructed circuit @xcite , effectively `` un - lumping '' @xmath0 .
this enhanced network exhibited hopf bifurcations and instability for many of the parameter values we simulated .
the delayed positive feedback created by the presence of this third metabolite is responsible for the induced oscillations .
moreover , this result also suggests that the self - repression of @xmath0 induced by the network topologies for the core metabolator , generalized model of the core metabolator , and the first `` enhanced '' network ( figures 1a and 1b ) , is responsible for the absence of oscillations in those systems . theory and simulation play important roles in synthetic biology .
as we have shown , they can highlight generic designs which have certain desirable behaviors , _
e.g. _ oscillations , and they can help exclude those that do not . in this respect , the analysis of the enhanced metabolator model with a third metabolite ( figure 1c ) suggested that we explore some other four - component models with related designs that do exhibit oscillations .
in particular , we studied system ( [ eq : revmet ] ) , the reverse metabolator , in which the wiring from @xmath0 to the two genes is reversed compared to the core metabolator , as shown in figure 1d .
this fundamentally altered the system dynamics so that @xmath0 now activates , rather than represses , the gene @xmath1 and hence @xmath0 is now self - activating .
the analysis of section 5 of the rm reveals that it exhibits both local and global routes to oscillations , via hopf bifurcations and saddle - node bifurcations of limit cycles , respectively , as shown in figure 3 .
most of the experimentally observed oscillations in synthetic circuits have arisen through local hopf bifurcations @xcite .
this is very likely to be due to the comparable ease of predicting local bifurcations .
indeed , the gm methods which are used throughout our work are only capable of identifying global bifurcations in very special and exceedingly rare cases ( _ i.e. _ when they are known to occur near the intersection of two local bifurcations , like a double hopf bifurcation ) .
in contrast to the hopf ( where oscillations appear at infinitesimal amplitude ) , oscillations arising from a fold of limit cycles ( as seen in the rm ) immediately appear at a finite , non - zero amplitude .
this feature may make such a `` global '' route to oscillation desirable in synthetic biology , which is still tackling the problem of designing robust cellular oscillators .
the similarity between the rm and the original metabolator ( ultimately , simply just a switch of promoters in the architecture of the circuit ) suggests that constructing the rm and investigating its dynamics _ in vivo _ may be relatively easy , and may lead to the experimental observation of a global bifurcation . finally , we propose that the construction of large generalized models of biological circuits may enable the resolution of several open problems regarding the dynamics of large , complex systems .
we investigated one such question here ( recall section 6 ) : the role that time scales play in determining the stability of the system .
while we found that stability generally increased as the time scale of regulation became longer ( figure 4 ) , it remains unclear how general this conclusion is among biological networks .
a closely related question is whether strongly stabilizing components exist within biological networks .
such components have been argued for in ecological networks @xcite , as well as in prior work on metabolism using gm @xcite .
however , the extent to which components of distinct types of biological networks ( _ e.g. _ signalling , transcriptional regulation , post - translational modifications ) act to stabilize or destabilize metabolism is unknown .
the investigation of such questions , aided by generalized modeling , seems potentially tractable numerically ; computational toolboxes for the automated assembly and inspection of generalized models already exist @xcite , and their extension to gm with regulation appears straightforward .
er and ds were supported by grants from the office of science ( ber ) , u.s .
department of energy ( de - sc0004962 ) and national science foundation nsf dms-0602204 emsw21-rtg biodynamics at boston university .
tjk was supported in part by nsf grant dms-1109587 .
the funders had no role in study design , data collection and analysis , decision to publish , or preparation of the manuscript .
er would like to acknowledge the kind support of the santa fe institute while studying the metabolator at the 2011 summer school .
10 uri alon . .
chapman and hall / crc , 2006 .
karl johan astrom and richard m. murray . .
princeton university press , 2008 .
arren bar - even , elad noor , yonatan savir , wolfram liebermeister , dan davidi , dan s tawfik , and ron milo .
, 50(21):440210 , 2011 . ken a dill , kingshuk ghosh , and jeremy d schmit . , 108(44):1787682 , 2011 .
eusebius j doedel , alan r champneys , thomas fairgrieve , yuri kuznetsov , bart oldeman , randy paffenroth , bjorn sandstede , xianjun wang , and chenghai zhang .
auto-07p : continuation and bifurcation software for ordinary differential equations , 2007 .
michael b. elowitz and stanislas leibler . , 403(6767):3358 , 2000 . d a fell . , 286 ( pt 2):31330 , september 1992 .
eileen fung , wilson w wong , jason k suen , thomas bulter , sun - gu lee , and james c liao . , 435(7038):11822 , 2005 .
tim s. gardner , charles r. cantor , and james j. collins .
, 403(6767):33942 , 2000 .
eva gehrmann , christine gl er , yaochu jin , bernhard sendhoff , barbara drossel , and kay hamacher . . , 84(2):021913 , august 2011
dorothee girbig , joachim selbig , and sergio grimbs . , 2012 .
sergio grimbs , joachim selbig , sascha bulik , herrmann - georg holzhutter , and ralf steuer .
the stability and robustness of metabolic states : identifying stabilizing sites in metabolic networks . , 3:146 , 2007 .
thilo gross and ulrike feudel .
generalized models as a universal approach to the analysis of nonlinear dynamical systems . , 73:016205016214 , 2006 .
thilo gross , lars rudolf , simon a levin , and ulf dieckmann .
, 325(5941):74750 , 2009 .
matthias heinemann and uwe sauer .
, 13(3):33743 , 2010 .
eugene m. izhikevich . .
mit press , 2006 .
neema jamshidi and bernhard palsson .
, 4:171 , 2008 .
edda klipp .
, 583(24):40138 , 2009 .
jeffrey d orth , ines thiele , and bernhard palsson .
, 28(3):2458 , 2010 .
jason a papin , tony hunter , bernhard o palsson , and shankar subramaniam .
, 6(2):99111 , 2005 .
flora picard , clmentine dressaire , laurence girbal , and muriel cocaign - bousquet .
, 332(11):95873 , 2009 .
oliver purcell , nigel j savery , claire s grierson , and mario di bernardo .
, 7(52):150324 , 2010 .
ed reznik and daniel segr . .
, 266(4):536549 , 2010 .
bjrn schwanhusser , dorothea busse , na li , gunnar dittmar , johannes schuchhardt , jana wolf , wei chen , and matthias selbach . .
, 473(7347):337342 , 2011 .
ralf steuer , thilo gross , joachim selbig , and bernd blasius .
structural kinetic modeling of metabolic networks .
, 103:1186811873 , august 2006 .
ralf steuer , adriano nunes nesi , alisdair r. fernie , thilo gross , bernd blasius4 , and joachim selbig .
from structure to dynamics of metabolic pathways : applications to the plant mitochondrial tca cycle . ,
23:13781385 , 2007 .
we begin by writing out the equations for the metabolator . in dimensionalized form , they are @xmath135 @xmath136 @xmath137 @xmath138 [ eq : mdim ] let s begin nondimensionalizing @xmath139 .
let @xmath140 .
we can rewrite ( [ eq : dim_g1 ] ) as @xmath141 where @xmath142 and @xmath143 .
then , the nondimensionalized equation for @xmath2 immediately follows as @xmath144 where @xmath145 and @xmath146 . for ( [ eq : dim_m ] ) , we rewrite it by substituting those nondimensionalized variables and parameters which we have already defined and choose to scale @xmath4 so that @xmath147 .
doing so , we get @xmath148 where @xmath149 and @xmath150 .
finally , the nondimensionalized equation for @xmath0 follows simply and is @xmath151 where @xmath152 .
we also note that a simpler way to define the system is in terms of `` total '' metabolite in the system @xmath5
. then , @xmath153 . we can non - dimensionalize this system as well to find that @xmath154 where @xmath155 . for the purpose of simplifying the analysis in the article
, we let @xmath156 and @xmath157 .
then , the system formed by ( [ eq : g1_nd]),([eq : g2_nd]),([eq : c_nd ] ) , and ( [ eq : n_nd ] ) is the nondimensionalized core metabolator we study in this article .
we work with the jacobian of the generalized model of the core metabolator ( [ eq : j_core ] ) : @xmath158 , \label{eq : j_corenew}\ ] ] where we have let @xmath159 and @xmath160 .
this means that @xmath161 and removes two otherwise free parameters from the analysis . adding multiples of one row to another
does not change the value of the determinant .
thus , to simplify matters , we multiply row one of ( [ eq : j_corenew ] ) by @xmath162 and add it to row 2 of ( [ eq : j_corenew ] ) to find @xmath163 .
\label{eq : j_corenew2}\ ] ] now expanding the determinant of ( [ eq : j_corenew2 ] ) along the first column , we find @xmath164 expanding again , we find @xmath165 expanding this expression a final time and grouping terms , we find the coefficients of the characteristic equation for the jacobian @xmath13 of the core metabolator are ( where @xmath166 corresponds to the coefficient of the @xmath167th term ) : @xmath168 the routh hurwitz criterion states that all of the eigenvalues of a 4x4 matrix have negative real parts when the following conditions on its characteristic polynomial are satisfied @xmath169 first , note that the condition @xmath170 is satisfied .
next , we calculate the quantity @xmath171 : @xmath172 now , @xmath173 = @xmath174 because @xmath175 . explicitly calculating @xmath176 : @xmath177 all the terms are positive ,
so the condition @xmath178 is satisfied .
now , we rewrite the final condition as @xmath179 . calculating @xmath180 ,
we find : @xmath181 also , rewriting @xmath174 and highlighting the two key components @xmath182 now , note that the first line of ( [ eq : a2a3minus ] ) multipled by term 2 of ( [ eq : a1 ] ) gives the first term of ( [ eq : a0a3 ] ) .
similarly , we can isolate a term like @xmath183 from the third line of ( [ eq : a2a3minus ] ) and multiply it by term 1 of ( [ eq : a1 ] ) to obtain the second term of ( [ eq : a0a3 ] ) , so that all the remaining terms of the product of the two lines are positive .
finally , we can multiply the second line of ( [ eq : a2a3minus ] ) by term 1 of ( [ eq : a1 ] ) to give the third term of ( [ eq : a0a3 ] ) and have all leftover terms be positive .
the remaining terms are all positive , proving that the routh - hurwitz criteria are satisfied and all of the eigenvalues of the jacobian have negative real part .
for the metabolator in figure 1c , we used the gm | the synthetic construction of intracellular circuits is frequently hindered by a poor knowledge of appropriate kinetics and precise rate parameters .
here , we use generalized modeling ( gm ) to study the dynamical behavior of topological models of a family of hybrid metabolic - genetic circuits known as `` metabolators . '' under mild assumptions on the kinetics , we use gm to analytically prove that all explicit kinetic models which are topologically analogous to one such circuit , the `` core metabolator , '' can not undergo hopf bifurcations .
then , we examine more detailed models of the metabolator . inspired by the experimental observation of a hopf bifurcation in a synthetically constructed circuit related to the core metabolator , we apply gm to identify the critical components of the synthetically constructed metabolator which must be reintroduced in order to recover the hopf bifurcation .
next , we study the dynamics of a re - wired version of the core metabolator , dubbed the `` reverse '' metabolator , and show that it exhibits a substantially richer set of dynamical behaviors , including both local and global oscillations . prompted by the observation of relaxation oscillations in the reverse metabolator , we study the role that a separation of genetic and metabolic time scales may play in its dynamics , and find that widely separated time scales promote stability in the circuit .
our results illustrate a generic pipeline for vetting the potential success of a potential circuit design , simply by studying the dynamics of the corresponding generalized model . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the engineering of biological circuits , synthetically constructed in the laboratory and designed to exhibit novel behaviors , has matured rapidly as a discipline over the past decade .
a major challenge for synthetic biology is understanding how different cellular subsystems , composed of distinct components and operating at widely different speeds , interface and work together to generate coherent cellular behaviors .
in particular , metabolic pathways can rapidly harvest energy and nutrients for reproduction , while genetic regulatory networks slowly control these pathways in response to environmental changes . here
, we study a generalized version of the metabolator , a unique synthetic circuit composed of both metabolic and regulatory components , previously constructed experimentally and shown to exhibit oscillations .
in addition to exploring multiple alternative network topologies , our generalized approach overcomes the inherent uncertainty associated with precise kinetic details of biological circuits . through this analysis
we identify circuit components that are essential for producing sustained oscillations .
furthermore , we find that certain dynamical regimes of a rewired metabolator display unexpected oscillatory properties , such as the capacity to suddenly transition to high amplitude oscillations .
this novel behavior could be tested experimentally , and lead to novel synthetic biology modules and applications . _
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ |
it was found in @xcite that a curve in the virasoro - bott group is a geodesic for the right invariant @xmath0-metric if and only if its right logarithmic derivative is a solution of the korteweg - de vries equation , see ! 2.3 .
vanishing geodesic distance for weak riemannian metrics on infinite dimensional manifolds was first noticed on shape space @xmath1 for the @xmath0-metric in @xcite . in @xcite
this result was shown to hold for the general shape space @xmath2 for any compact manifold @xmath3 and riemannian manifold @xmath4 , and also for the right invariant @xmath0-metric on each full diffeomorphism group with compact support @xmath5 .
in particular , burgers equation is related to the geodesic equation of the right invariant @xmath0-metric on @xmath6 or @xmath7 and it thus also has vanishing geodesic distance .
we even have @xcite the weak riemannian @xmath0-metric on each connected component of the total space @xmath8 for a compact manifold @xmath3 and a riemannian manifold @xmath9 has vanishing geodesic distance .
this result is not spelled out in @xcite but it follows from there : given two immersions @xmath10 in the same connected component , we first connect their shapes @xmath11 and @xmath12 by a curve of length @xmath13 in the shape space @xmath2 and take the horizontal lift to get a curve of length @xmath13 from @xmath14 to an immersion @xmath15 in the connected component of the orbit through @xmath16 .
now we use the induced metric @xmath17 on @xmath3 and the right invariant @xmath0-metric induced on @xmath18 to get a curve in @xmath19 of length @xmath13 connecting @xmath20 with @xmath21 . evaluating at @xmath16 we get curve in @xmath8 of length @xmath13 connecting @xmath22 with @xmath16 . in this article
we show that the right invariant @xmath0-metric on the virasoro - bott groups ( see ! 2.1 ) has vanishing geodesic distance .
this might be related to the fact that the riemannian exponential mapping is not a diffeomorphism near @xmath23 , see @xcite for @xmath6 and @xcite for the virasoro group over @xmath24 .
see @xcite for information on conjugate points along geodesics .
let @xmath25 be the group of diffeomorphisms of @xmath26 which rapidly fall to the identity .
this is a regular lie group , see @xcite .
the mapping @xmath27 satisfies @xmath28 , @xmath29 , @xmath30 and is a smooth group cocycle , called the bott cocycle : @xmath31 the corresponding central extension group @xmath32 , called the virasoro - bott group , is a trivial @xmath26-bundle @xmath33 that becomes a regular lie group relative to the operations @xmath34 other versions of the virasoro - bott group are the following : @xmath35 where @xmath7 is the group of all diffeomorphisms with compact support , or the periodic case @xmath36 .
one can also apply the homomorphism @xmath37 to the center and replace it by @xmath24 . to be specific
, we shall treat the most difficult case @xmath25 in this article .
all other cases require only obvious minor changes in the proofs .
the lie algebra of the virasoro - bott group @xmath38 is @xmath39 ( where @xmath40 ) with the lie bracket @xmath41 = \binom{-[x , y]}{\om(x , y ) } = \binom{x'y - xy'}{\om(x , y)}\ ] ] where @xmath42 is the @xmath43 , which is a bounded skew - symmetric bilinear mapping satisfying the cocycle condition @xmath44,z)+\om([y , z],x)+\om([z , x],y)=0.\ ] ] it is a generator of the 1-dimensional bounded chevalley cohomology @xmath45 for any of the lie algebras @xmath46 , @xmath47 , or @xmath48 .
the lie algebra of the virasoro - bott lie group is thus the central extension @xmath49 induced by this cocycle .
we have @xmath50 for each finite dimensional manifold of dimension @xmath51 ( see @xcite ) , which blocks the way to find a higher dimensional analog of the korteweg - de vries equation in a way similar to that sketched below . to complete the description , we add the adjoint action : @xmath52 where the @xmath53 is given by @xmath54 which measures the deviation of @xmath20 from being a mbius transformation : @xmath55 the schwartzian derivative of a composition and an inverse follow from the action property : @xmath56 we shall use the @xmath0-inner product on @xmath57 : @xmath58 we use the induced right invariant weak riemannian metric on the virasoro group .
according to @xcite , see @xcite for a proof in the notation and setup used here , a curve @xmath59 in the virasoro - bott group is a geodesic if and only if @xmath60 since we have @xmath61 thus @xmath62 is a constant in time and the geodesic equation is hence the _ korteweg - de vries equation _ @xmath63 with its natural companions @xmath64 to be complete , we add the invariant momentum mapping @xmath65 with values in the virasoro algebra ( via the weak riemannian metric ) .
we need the transpose of the adjoint action : @xmath66 thus , the invariant momentum mapping is given by @xmath67 along a geodesic @xmath68 , the momentum @xmath69 is constant in @xmath70 .
we consider the extension @xmath71 then @xmath72 is a riemannian submersion for the right invariant @xmath0-metric on @xmath73 , i.e. , @xmath74 is an isometry on the orthogonal complements of the fibers .
these complements are not integrable ; in fact , the curvature of the corresponding principal connection is given by the gelfand - fuks cocycle .
for any curve @xmath75 in @xmath73 its horizontal lift is given by @xmath76 since the right translation to @xmath77 of its velocity should have zero vertical component , see !
the horizontal lift has the same length and energy as @xmath20 .
.3.1 on all virasoro - bott groups mentioned in ! 2.1 geodesic distance for the right invariant @xmath0-metric vanishes .
in @xcite for @xmath7 it was first shown that there exists one non - trivial diffeomorphism which can be connected to @xmath79 with arbitrarily small length .
then , it was shown that the diffeomorphisms with this property form a normal subgroup . since @xmath7 is a simple group this concluded the proof .
but @xmath73 is not a simple group since @xmath7 is a normal subgroup .
so , we have to elaborate on the proof of @xcite as follows .
proof we show that any rapidly decreasing diffeomorphism can be connected to the identity by an arbitrarily short path .
we will write this diffeomorphism as @xmath80 , where @xmath81 is a rapidly decreasing function with @xmath82 .
for @xmath83 we define @xmath84 this is a ( non - smooth ) path defined for @xmath85 connecting the identity in @xmath73 with the diffeomorphism @xmath86 .
we define @xmath87 , where @xmath88 is a smoothing kernel with @xmath89 and @xmath90 .
thus @xmath91 is a smooth path defined on the finite interval @xmath92 connecting the identity in @xmath93 with a diffeomorphism arbitrarily close to @xmath86 for @xmath94 small .
( compare figure [ fig ] for an illustration . ) * claim . * the mappings @xmath101 and @xmath102 are continuous into each @xmath103 with @xmath72 even .
( the proofs are simpler when @xmath72 is even because there are no absolute values to be taken care of . ) to prove the claim , we calculate @xmath104 where @xmath105 is a polynomial without constant term in @xmath106 with coefficients also powers of @xmath107 .
integrals of the form @xmath108 with @xmath109 are finite and continuous in @xmath110 since @xmath111 is rapidly decreasing .
this shows that @xmath112 depends continuously on @xmath94 .
furthermore the sequence @xmath113 converges almost everywhere for @xmath114 , thus it also converges in measure . by the theorem of vitali , this implies convergence in @xmath103 ,
see for example ( * ? ? ?
* theorem 16.6 ) .
convolution with @xmath115 acts as approximate unit in each @xmath103 , which proves the claim for @xmath116 .
for @xmath117 it follows similarly .
the above claim implies that @xmath118 viewed as a mapping on @xmath119 ( first summand ) and on @xmath120 ( second summand ) is continuous in @xmath94 .
it also vanishes at @xmath121 since then @xmath122 and @xmath116 have disjoint support .
the cauchy - schwarz inequality @xmath123 implies that @xmath124 goes to zero as well .
ultimately , @xmath125 is arbitrarily close to @xmath80 .
proof the aim of the following argument is to construct a family of paths in the diffeomorphism group , parametrized by @xmath94 , with the following properties : all paths in the family start and end at the identity and their length in the diffeomorphism group with respect to the @xmath0 metric tends to @xmath23 as @xmath130 . by letting @xmath94 be time - dependent , we are able to control the endpoint @xmath131 of the horizontal lift for certain diffeomorphisms .
[ cc][cc]@xmath80[tl][tl]@xmath79[cr][cr]@xmath132[cr][cr]@xmath133[cr][cr]@xmath134[cc][cc]@xmath135[cc][cc]@xmath136[bc][bc]@xmath137[cl][cl]@xmath137 defined in !
3.3 connecting @xmath79 to @xmath80 , plotted at @xmath138 . between the dashed lines
, @xmath139 is constant .
, title="fig:",scaledwidth=50.0% ] we consider the function @xmath140 where @xmath141 is a function with @xmath142 $ ] and @xmath143 .
furthermore , let @xmath144 $ ] be a function with compact support contained in @xmath145 and @xmath82 , so that @xmath146 is a diffeomorphism . for @xmath147 and @xmath148 $ ] let @xmath149 be the path going away from the identity ( since @xmath150 , see also figure [ fig ] ) .
for given @xmath151 , let @xmath152 the path leading back again .
the only difference to @xcite is that the parameter @xmath94 may vary along the path .
we shall need some derivatives of @xmath20 and @xmath153 : @xmath154 * claim 1 . * the path @xmath20 followed by @xmath91 still has arbitrarily small length for the @xmath0-metric .
we are working with a fixed time interval @xmath155 $ ] .
thus arbitrarily small length is equivalent to arbitrarily small energy .
the energy is given by @xmath156 looking at the formula for @xmath157 we see that @xmath158 is bounded on a domain with bounded @xmath62 .
thus @xmath159 can be achieved by choosing @xmath94 , such that @xmath160 .
we will see later that this is possible . inspecting @xmath161 and looking at the formulas for @xmath162 and @xmath153 we see that for @xmath163 and for @xmath164 we have @xmath165 .
thus the domain of integration is contained in the compact set @xmath166\times[-\frac{t + \|g\|_{\infty}+2\|\ep\|_{\infty}}{\lambda},\frac{t + \|\ep\|_{\infty}}{\lambda}].\ ] ] therefore , it is enough to show that the @xmath167-norm of the integrand in goes to zero as @xmath168 goes to zero .
for all terms involving @xmath169 this is true by the above assumption since @xmath170 and @xmath158 are bounded . for the remaining parts @xmath171 and @xmath172 we follow the argumentation of @xcite . for @xmath70
fixed and @xmath173 close to @xmath174 , the function @xmath175 , when restricted to the support of @xmath162 , is bigger than @xmath176 only on an interval of length @xmath177 .
hence we have @xmath178 for the last part , we note that the support of @xmath179 is contained in the set @xmath180 .
now we define @xmath181 by @xmath182 and @xmath183 .
then @xmath184 the estimate for @xmath91 is similar and easier .
this proves claim 1 . * claim 2 . * for every @xmath185 and @xmath186 we may choose @xmath176 with @xmath187 such that @xmath188 we will subject @xmath94 and @xmath111 to several assumptions .
first , we partition the interval @xmath189 $ ] equidistantly into @xmath190 and the @xmath191-domain into two parts , namely @xmath192 \cup [ t_e , t ] ) \times \mathbb r$ ] and @xmath193 \times \mathbb r$ ] .
we want @xmath194 on a neighborhood of the interval @xmath195 $ ] .
we choose @xmath176 to be constant @xmath196 on @xmath197 \cup [ t_e , t]$ ] and to be symmetric in the sense , that @xmath198 .
in addition , we want @xmath176 small enough , such that @xmath194 on @xmath199.$ ] let @xmath193 \times \mathbb r$ ] be the region , where @xmath176 is not constant . in the interior , where @xmath206 we have by assumption @xmath207 .
therefore , one has in this region : @xmath208 we divide the integral over @xmath209 into two symmetric parts @xmath210 and apply the following variable substitution to the second integral @xmath211 thus @xmath212 .
together with @xmath213 this implies @xmath214 since @xmath215 changes sign , the term containing @xmath216 cancels out and leaves only @xmath217 a simple calculation shows that the integration limits transform @xmath218 to those of the first integral .
therefore , the sum of the integrals gives @xmath219 from formula we see : @xmath220 we can use this to rewrite the above integral : @xmath221 looking at the formula for @xmath222 @xmath223 we see that @xmath224 is non - zero only on the intervals @xmath225 and @xmath226 . for small @xmath94 , these are two disjoint regions .
therefore , the above integral equals @xmath227 for @xmath228 bounded and sufficiently small @xmath176 , the functions under the integral do not depend on @xmath176 any more as can be seen from the definitions of @xmath162 and @xmath222 .
thus @xmath229 is independent of @xmath70 and we have @xmath230 the same calculations can be repeated for the return path @xmath231 , where @xmath232 is constant in time : @xmath233 note that the sign is positive now , which comes from the @xmath70-derivative . putting everything together gives us @xmath234 let @xmath235 where @xmath236 is a bump function with height @xmath174 and @xmath237 is a small constant .
note that @xmath176 satisfies @xmath238 .
choosing @xmath239 and @xmath237 small independently we may produce any @xmath185 .
proof of theorem !
3.1 let @xmath240 . by proposition !
3.2 we get a smooth family @xmath241 for @xmath186 and @xmath242 $ ] such that @xmath243 , @xmath244 , @xmath245 , and such that the length of @xmath246 is @xmath247 . using !
2.4 consider the horizontal lift @xmath248 of this family which connects @xmath127 with @xmath249 for each @xmath186 and has length @xmath247 .
but one can see from the proof of lemma ! 3.3 that @xmath250 becomes unbounded for @xmath251 . using lemma !
3.3 we can find a horizontal path @xmath252 for @xmath253 $ ] in the virasoro group of length @xmath247 connecting @xmath127 with @xmath254 .
then the curve @xmath255 connects @xmath256 with @xmath257 and it has length @xmath247 .
p. w. michor .
some geometric evolution equations arising as geodesic equations on groups of diffeomorphism , including the hamiltonian approach . in : phase space analysis of partial differential equations .
bove , antonio ; colombini , ferruccio ; santo , daniele del ( eds . ) .
progress in non linear differential equations and their applications , vol .
birkhuser boston , 2006 . | the virasoro - bott group endowed with the right - invariant @xmath0-metric ( which is a weak riemannian metric ) has the kdv - equation as geodesic equation .
we prove that this metric space has vanishing geodesic distance . |
a resonant structure in the invariant mass of @xmath6 at 4.43 gev with a width of 45 mev was recently reported by belle in @xmath7 @xcite .
this result immediately provoked much interest from experiment and theory due to the possibility of it being a tetraquark candidate @xcite with flavor @xmath8 or @xmath9 , i.e. a charmonium with isospin @xmath10 .
it has also been proposed that this structure may be due to subthreshold effects in @xmath11 rescattering @xcite .
similar to the @xmath12 which is proposed to be a possible bound state of @xmath13 due to long - range @xmath14 exchange forces @xcite , such a mechanism will favor @xmath15 or @xmath16 for @xmath0 with @xmath10 @xcite .
also , it has been suggested by maiani _
et al . _
@xcite that the strong coupling of the @xmath0 to @xmath1 instead of @xmath17 could be a signal for a radially excited @xmath18 tetraquark .
long ago close and lipkin @xcite proposed that @xmath19 is a channel that could identify tetraquark states such as @xmath18 in s - wave or , most interesting , @xmath20 in p - wave , hence we shall also consider that @xmath21 occurs via @xmath22-wave transition if it has @xmath23 .
in particular , a large coupling to @xmath1 will have special implications for its photoproduction rate , and information about its quantum numbers may be extracted .
seeing this state in an independent production process will be essential for establishing its existence and determining its nature .
we propose here to look for the @xmath0 in photoproduction .
if @xmath0 is a genuine signal for a tetraquark state which couples to @xmath1 , then it can necessarily be photoproduced via @xmath14 exchange through the @xmath24 intermediate coupling .
it suggests that @xmath25 could be a dominant production mechanism for @xmath0 .
therefore , by looking for the signal of @xmath0 in @xmath26 , one may establish this state as a genuine resonance .
we also estimate a possible destructive interference from the intermediate @xmath27 coupling .
although belle does not see the decay of @xmath28 , which implies a weaker @xmath29 coupling , the relatively larger leptonic width for @xmath30 than @xmath31 may still lead to sizeable contributions from intermediate @xmath32 in photoproduction .
we will investigate such effects in a vector meson dominance ( vmd ) model .
in addition , we consider background contributions to this channel which involve pomeron exchanges and nucleon born terms .
interestingly , this shows that these backgrounds have rather small interferences with the @xmath33 production .
this enhances the possibility of identifying or refuting this state in photoproduction .
experimental data from desy may be able to clarify this directly . in the following we first provide a detailed description of the theoretical model by considering @xmath0 to have either @xmath3 , @xmath4 or @xmath5 .
the model predictions for the differential and total cross sections will be presented .
we then make a quantitative analysis of the possible background contributions to @xmath34 final state .
dalitz plot analysis will be provided .
the production of @xmath35 in @xmath36 can be described by the processes in fig .
[ fig:1 ] .
its production possesses different features at different kinematics .
we outline those features below as they are essential for constructing the theoretical model : \i ) in the @xmath37-channel , as the electromagnetic ( em ) interaction does not conserve isospin , the @xmath0 can be produced by either isoscalar or isovector exchange . taking into account the flavor content , i.e. @xmath35
has a hidden @xmath38 which does not carry isospin , the isoscalar exchange in the @xmath37-channel will be forbidden for the charged @xmath0 while suppressed for the neutral partner .
we will come back to this point later .
furthermore , as the @xmath0 has only been seen in the @xmath1 channel , we shall concentrate on the @xmath37-channel pion exchange as a leading contribution .
we shall argue that this is likely to give a lower limit for the photoproduction cross section .
we note in advance that the inclusion of e.g. @xmath39 exchange will increase the cross sections since there is no interferences between the pion and @xmath39 exchange amplitudes .
we will present some numerical estimates for the @xmath40 exchange later .
heavier @xmath37-channel meson ( regge ) exchanges contribute to the cross section at off - forward angles and could become important in the higher energy region . as we do not have information about their couplings to @xmath41 , we shall focus on the near - threshold region where the dominant pion exchange can be well - constrained and reliably estimated .
\ii ) for @xmath3 and @xmath5 , @xmath0 will couple to @xmath42 in @xmath22-wave at leading order , while for @xmath43 , it is via @xmath44-wave .
we apply different effective lagrangians to describe the @xmath45 vertex .
\iii ) there is no available information about the coupling for a @xmath46 to nucleons .
this will appear in the @xmath47 or @xmath48-channel where @xmath0 can directly couple to @xmath49 and @xmath50 .
the @xmath47-channel process generally becomes important near threshold , and at middle and backward scattering angles . with increasing energies ,
these contributions die out quickly due to the suppression of the @xmath47 and @xmath48-channel propagators and the effects of the nucleon s form factors .
in contrast , the @xmath37-channel pion exchange may remain dominant at forward angles and give the major contribution to the cross section .
based on the above rather general considerations , we identify processes that will drive the @xmath0 production . as the @xmath0
is observed to couple strongly to @xmath51 , we propose to look for signals in @xmath52 , and @xmath53 .
this gives rise to some background contributions as shown in fig .
[ fig:2 ] , in particular , in the neutral @xmath0 production .
\iv ) as the @xmath0 is observed to couple to @xmath1 , we derive the @xmath54 coupling by vector meson dominance ( vmd ) , and assume that the @xmath54 coupling is due to a sum of intermediate vector mesons which connect the photon with the @xmath0 and the exchanged @xmath14 .
we initially assume that @xmath55 is the dominant intermediate vector meson in the @xmath56 coupling , and then consider the possible effect of a coupling to @xmath57 . for @xmath3
, the following lagrangian is applied for the @xmath58 coupling , e.g. for the charged @xmath0 , @xmath59 where @xmath60 is levi - civita tensor , and the coupling constant @xmath61 will be determined by the width @xmath62
@xcite , i.e. @xmath63 where @xmath64 \nonumber \\ & = & \left(\frac{g_{z\psi^\prime\pi}}{m_z}\right)^2 \left [ \frac{(m_z^2+m_{\psi^\prime}^2-m_\pi^2)^2}{2 } - 2m_z^2 m_{\psi^\prime}^2 \right ] \ , \end{aligned}\ ] ] and @xmath65 is the three - vector momentum of the pion in the @xmath46 meson rest frame . with @xmath66 mev , and @xmath67 mev , we have @xmath68 . any coupling to @xmath57 will be analogously related to @xmath69 , which in turn is driven by the ( as yet unobserved ) decay @xmath70 . for the meson coupling vertices
, we apply a form factor as follows : @xmath71 with @xmath72 , and the cutoff is set as the mass of the intermediate vector meson @xcite , i.e. @xmath73 . the general expression for an intermediate - vector - meson @xmath74 coupling to @xmath75 at leading order is @xmath76 the coupling constant @xmath77 can be determined by @xmath78 , @xmath79 is the em fine - structure constant . with @xmath80 kev @xcite
, it gives @xmath81 .
the @xmath75-@xmath32 coupling analogously is @xmath82 . as this is larger than for the @xmath55 it is possible that any coupling @xmath83 could cause an additional contribution to the photoproduction amplitude , which could destructively interfere with the former and hence give a much reduced cross section .
if this were the case , it would also imply that a branching ratio @xmath84 should be visible .
we shall discuss the dependence of the photoproduction cross section on this . at the meson nucleon coupling vertices we adopt the commonly used lagrangian @xmath85 where a standard value , @xmath86 , is adopted @xcite . in addition ,
a form factor is applied for the @xmath87 vertex , @xmath88 with @xmath89 gev .
we note that this form factor successfully accounts for the photoproduction of @xmath90 and @xmath91 meson near threshold @xcite .
some works have argued for a larger value @xcite .
such values would increase the cross sections relative to what we compute here , and so we use the cited value in the spirit of seeking a lower limit for the cross - section .
the effective lagrangian leads to the amplitude for producing a charged @xmath0 via charged pion exchange in @xmath92 : @xmath93 where @xmath94 ( @xmath95 ) and @xmath96 ( @xmath97 ) are four - vector momenta of initial nucleon ( photon ) , and final nucleon ( @xmath46 meson ) , respectively ; @xmath98 and @xmath99 are the polarization vectors of @xmath75 and @xmath46 , respectively . by defining @xmath100 , @xmath101 ,
the differential cross section is @xmath102 where the invariant matrix element squared gives @xmath103 after integrating over the range of @xmath104 , i.e. within @xmath105 and @xmath106 , @xmath107 the total cross section can be obtained . in the above equation , @xmath108 is the photon energy in the overall c.m .
system , where @xmath109 is the photon energy in the rest - frame of the initial proton . in the appendix
, we provide the analogous expressions for @xmath37-channel @xmath39 exchange .
note that the pion and @xmath39 exchanges contribute to the real and imaginary part of the transition amplitude respectively .
thus , there are no interferences between these two channels , and the inclusion of the @xmath39 exchange will enhance the @xmath0 cross section . as we lack knowledge on the @xmath110 and @xmath111 coupling strengths , which are in any event expected to be much smaller than the @xmath14 exchange contributions , we will focus on the pion exchange here . for neutral @xmath46 production contributions
may also be allowed from pomeron exchange .
however , we argue that this scenario is suppressed due to the following reasons : i ) in the case where the photon ( isoscalar component ) couples to @xmath38 , isospin has to be transferred to the pair of light @xmath112 in order to form @xmath113 .
ii ) the photon can couple to isovector vector mesons made of light @xmath112 , which then pull out a pair of @xmath38 after exchanging a pomeron .
although this process conserves isospin , it is suppressed by the soft creation of a large mass @xmath38 , and also as it requires the @xmath46 meson to be formed with the @xmath112 and @xmath38 pairs at relatively large momentum transfers .
consequently suppression from the final state wavefunction is likely .
precise estimates will be model dependent but for the purpose of estimating a lower limit for the production cross section , we neglect this contribution here .
in contrast , pomeron contributions can generate background . as shown in fig .
[ fig:2](a ) and ( b ) , the direct production of @xmath55 allows pomeron exchanges as an important production mechanism .
an intermediate nucleon will then decay into a pion and nucleon . in principle , all the intermediate isospin 1/2 nucleon resonances can contribute to this process .
again , as a simple estimate of the background , we only consider the nucleon pole contributions . for the production of @xmath0 with @xmath43 ,
the difference from @xmath114 is at the @xmath45 vertex for which the following effective lagrangian is adopted @xcite : @xmath115 where the notations are the same as eq .
( [ lagrangian - v ] ) .
schematically , the @xmath14 exchange transition can be illustrated by fig .
[ fig:1 ] , and the definition of kinematic parameters is the same as the previous subsection .
the transition matrix elements are as follows : @xmath116 \nonumber \\ & \times & \frac{1}{q^2-m_\pi^2 } f_{\pi nn}(q^2 ) f_{z\psi^\prime\pi}(q^2 ) \ , \end{aligned}\ ] ] which gives @xmath117 one notices that the above equations have the same form as eq .
( [ trans - v - total ] ) .
this is because the longitudinal terms in @xmath56 coupling vanish in the real photon limit for the chiral partner @xmath118 and @xmath4 .
the difference between these two parities leads to different values for the decay constant @xmath61 . with the partial width : @xmath119 \ , \end{aligned}\ ] ] we have @xmath120 which is smaller than that for @xmath118 .
if we take @xmath0 as a pseudoscalar meson with @xmath122 , the following effective lagrangian for @xmath58 coupling is adopted : @xmath123 similarly , the @xmath14 exchange transition can be illustrated in fig .
[ fig:1](a ) .
the transition matrix element is as follows : @xmath124 which gives @xmath125 again , the coupling coupling constant @xmath61 is also determined by @xmath126 : @xmath127,\ ] ] and we have @xmath128 interestingly , the neutral @xmath129 has @xmath130 , which is an exotic quantum number . in figs .
[ fig : vec ] , [ fig : axi ] and [ fig : pse ] , the differential and total cross sections are plotted for @xmath3 , @xmath4 , and @xmath131 , respectively . the photon energy for the differential cross sections is @xmath132 gev ( @xmath133 gev ) , which corresponds to the peaking energy region in the total cross sections . in all three cases ,
the forward peaking turns to be a prominent feature in the differential cross sections due to the pion exchange .
the results of figs .
[ fig : vec ] , [ fig : axi ] and [ fig : pse ] show compatible production cross sections for those three spin - parity assignments . for spin - parity @xmath134 and @xmath135 ,
their production cross sections have the same expression in the real photon limit as noted before .
the difference between figs .
[ fig : vec ] and [ fig : axi ] arises from the different coupling strengths for @xmath61 extracted from @xmath136 .
the cross section for @xmath131 production is found larger than the other two assignments .
this is understandable by comparing eqs .
( [ trans - v - total ] ) , ( [ trans - v - axi ] ) , and ( [ tran - p ] ) with each other .
it also shows that in order to determine the quantum numbers , measurement of the angular distributions of @xmath137 is necessary . for spin - parity @xmath135
, the decay will be via relative @xmath44-wave between @xmath1 , while it is via relative @xmath22-wave for @xmath134 and @xmath131 .
therefore , an observation of @xmath22-wave decay will require further studies such as polarization observables to determine the quantum numbers of @xmath0 . for @xmath131 ,
@xmath39 exchange is forbidden and the pion exchange is the exclusive leading contribution . for other @xmath138 @xmath39 could contribute but in practice need not concern us here . even were its coupling as large as that of the pion , @xmath139 , its contribution to the cross - section is negligible ( in figs .
[ fig : vec ] and [ fig : axi ] , the results due to @xmath37-channel @xmath39 exchanges are presented by the dotted curves ) . this in practise allows us to focus on the pion exchange mechanism in the photoproduction reaction as a reliable estimate of the lower bound of the production cross sections . due to the pion exchange
the energy dependence of the total cross sections exhibits a strong threshold enhancement in @xmath140 .
it is similar to the charged @xmath91 meson production near threshold where the charged pion exchange ( unnatural parity exchange ) also plays a dominant role @xcite .
however , such a threshold enhancement seems to be absent in charmonium production , such as @xmath32 and @xmath55 @xcite .
this may be due to the dominance of the diffractive process which submerges the contributions from @xmath37-channel unnatural parity exchanges .
although @xmath32 has a relatively large coupling to @xmath141 , whereby @xmath37-channel pion exchange turns to be important in @xmath142 , the coupling is not as large as that for @xmath45 . consequently , the predicted threshold enhancement in @xmath46 meson photoproduction is strongly driven by the large @xmath45 coupling .
our main concern therefore is the possibility of some further contribution that is large and destructive .
the most important question here seems to concern the role played by the intermediate @xmath32 in the vmd model , which in a worst case could destructively interfere with the @xmath55 . by defining @xmath143
we include @xmath32 contributions to the @xmath14 exchange diagrams ; i.e. the real photon couples to @xmath32 and @xmath55 , which then couple to @xmath144 .
we then have @xmath145 where we have already set the form factors @xmath146 for simplicity , and @xmath147 is the relative phase between the two transition matrix element .
couplings @xmath148 and @xmath149 can be determined by eq .
( [ gamma - z ] ) and eq .
( [ vmd ] ) , respectively .
the cross section becomes @xmath150 the worst destructive situation is at @xmath151 .
but the destructive effects will depend on the value of @xmath152 .
note that belle has not seen @xmath153 decay , which allows an estimate of @xmath154 , and with which the total cross section would be still sizeable .
if the @xmath46 meson is a genuine resonance , the threshold enhancement will be a signature for its existence .
this can be further clarified by the study of background effects in @xmath155 in the next section .
experimentally , the final state particles identified are @xmath55 , @xmath14 and nucleon . as a result
, background contributions to the @xmath156 channel can interfere with the transitions we are interested in .
as discussed earlier , one of the major contributions is due to the diffractive production of @xmath55 ( fig .
[ fig:2 ] ) . in this section ,
we combine @xmath157 and @xmath158 together to make an estimate of the signal and background contributions
. this could be useful for experimental search for the @xmath35 in this channel .
since the production rate for all these three spin - parities are similar , we only consider @xmath159 in the following analysis .
the schematic transition diagrams are illustrated in fig .
[ fig:2 ] .
we first extend the formulae of the previous section to @xmath157 , and then include the diffractive contributions from the pomeron exchanges . by including the vertex coupling for @xmath160 , the transition amplitude for pion exchange
can be explicitly written down @xmath161 \bar{u}(p_3 ) \gamma_5 u(p_1 ) \epsilon_\mu(k)\epsilon_{\psi^\prime}^{\nu}(q ) \varepsilon_{\mu_1\nu_1\alpha_1\beta_1}k^{\mu_1}q_2^{\alpha_1 } \varepsilon_{\mu_2\nu\alpha_2\beta_2}q^{\mu_2}q_2^{\alpha_2}\nonumber \\ & \times & \frac{g^{\mu\nu_1}g^{\beta_1\beta_2}}{(q_1 ^ 2-m_\pi^2)(q_2 ^ 2-m_z^2+im_z\gamma ) } f_{z\psi^\prime\pi}(q_1 ^ 2)f_{\pi nn}(q_1 ^ 2)f_{z\psi^\prime\pi}(q_2 ^ 2 ) \ , \end{aligned}\ ] ] where the vertex functions are @xmath162 and the form factor is @xmath163 considering a moderate modification of the cutoff will not change the results significantly , we take @xmath164 as an input value .
the kinematic variables are denoted in fig .
[ fig:2 ] .
pomeron as a phenomenology accounting for the diffractive transitions has been widely studied in the literature @xcite . in this approach
, the pomeron mediates the long range interaction between a confined quark and a nucleon , and has been shown to behave just like a @xmath165 isoscalar photon .
the pomeron - nucleon coupling is determined by the vertex : @xmath166 where @xmath37 is the pomeron momentum squared , @xmath167 represents the coupling constants between a single pomeron and a light constituent quark , and @xmath168 is the isoscalar nucleon electromagnetic form factor : @xmath169 for the @xmath170 vertex ( @xmath74 represents the corresponding vector meson , here is @xmath55 ) , we take the on - shell approximation for restoring gauge invariance as @xcite .
therefore the equivalent vertex for the @xmath171 is reduced to a loop tensor as follows : @xmath172 where @xmath173 and @xmath174 is the effective coupling constant between pomeron and a charm quark within @xmath55 ; @xmath175 is a cutoff of the form factor related to the pomeron . the amplitudes for fig .
[ fig:2](a ) and ( b ) can then be expressed respectively as @xmath176 function @xmath177 is related to the pomeron trajectory @xmath178 via @xmath179 we also consider a form factor for the vertex @xmath180 since the intermediate nucleon is off - shell , i.e. @xmath181 where @xmath182 , @xmath183 . considering the kinematic region of the three - particles final state ,
we take a cutoff @xmath184 in our following calculation to avoid the singularity . combining the signal terms and background amplitudes ,
the invariant transition amplitude becomes @xmath185 + 2re[ac^ * ] + 2re[ad^*]+2re[bc^*]+2re[bd^*]+2re[cd^ * ] \ .\end{aligned}\ ] ] where @xmath186=0 $ ] , and @xmath187 $ ] , @xmath188 $ ] , @xmath189 $ ] , @xmath190 $ ] are found to be much smaller than the other terms .
the following values are applied in the pomeron exchange amplitudes : @xmath191 where @xmath174 is determined by considering the ratio @xmath192 through the pomeron exchange .
we take the following value h1 @xcite as a constraint : @xmath193 with the coupling @xmath194 fixed by the @xmath32 photoproduction @xcite , the coupling @xmath174 for @xmath55 is determined .
we note that these parameters lead to a reasonable description of both @xmath32 and @xmath55 photoproduction in comparison with the experiment @xcite .
as shown by the total production cross section for @xmath195 , the largest cross section is at @xmath196 .
we thus take this energy as an input to simulate the invariant mass spectrum for @xmath197 .
this could be the kinematic region which favors the separation of the @xmath35 signals . in fig .
[ fig : xsbg ] , we calculate the total cross sections for @xmath198 including both signal and background contributions . as shown by the solid curve ,
the overall production cross section has a quick increase just above threshold and is dominated by the @xmath35 production as shown by the dotted curve .
in contrast , the background contributions from the pomeron exchange are relatively small and do not vary drastically with the increasing energies at @xmath199 gev .
this is consistent with the familiar diffractive behavior in vector meson photoproduction .
the steep rise of the cross section near threshold turns out to be very different from other known heavy vector photoproductions , such as @xmath200 , @xmath32 and @xmath55 . in those cases ,
the threshold cross sections are rather smooth and no threshold enhancement has been observed . in the production of @xmath33 , its strong coupling
to @xmath1 gives rise to the predominant contributions from pion exchange .
note that @xmath201 is also strong .
these lead to an unusual prediction for its cross section near threshold .
it is more constructive to inspect the dalitz plot for the three - body final state .
we take a sample of the events at @xmath202 and @xmath203 gev , respectively . as shown by fig .
[ fig : dalitz1 ] , a clear band appears in the invariant mass spectrum of @xmath1 due to the @xmath46 resonance . although we do not include the @xmath47 and @xmath48 channel nucleon resonances , their contributions seem not to demolish the @xmath0 band since their effects will add to the plot horizontally and in the low @xmath204 region , e.g. @xmath205 gev@xmath206 . with the increasing center - mass energy ,
the background from pomeron - exchange becomes important , and the @xmath0 signals are submerged as shown by fig .
[ fig : dalitz2 ] . for the neutral @xmath113 production , the leading background contribution is the same as the charged @xmath46 production . in the @xmath37-channel @xmath207 production ,
an additional transition is the diffractive production of @xmath207 via pomeron exchange for @xmath118 or odderon exchange for @xmath135 .
it gives rise to two possibilities as shown in fig .
[ fig - z0 ] . in fig .
[ fig - z0](a ) , the initial photon couples to the isoscalar component of a vector meson ( @xmath38 ) , which then couples to @xmath207 by projecting the @xmath38 component to the @xmath10 tetraquark wavefunction .
note that the exchange of pomeron or odderon does not compensate the isospin .
this process will be suppressed due to isospin violation .
the other diffractive transition is that the initial photon couples to the isovector component of the intermediate meson as shown by fig . [ fig - z0](b ) . for the tetraquark scenario ,
the isovector component must be from the light @xmath112 instead of @xmath38 , hence requiring creation of a pair of @xmath38 from the vacuum which then couples to the @xmath10 tetraquark wavefunction .
note that the @xmath38 pair is much heavier than the light @xmath112 . to form a tetraquark of @xmath208
, it needs a sizeable overlap between the final state @xmath207 and the intermediate isovector @xmath112 .
since the @xmath38 is much heavier than the light @xmath112 , the @xmath207 wavefunction overlap with the @xmath38 and light @xmath112 should be small .
this allows us to neglect the pomeron or odderon exchange contributions to the direct @xmath207 production via diffractive processes . due to the above argument
, the neutral @xmath207 photoproduction turns out to be similar to the charged @xmath46 production studied earlier .
the neutral @xmath207 of @xmath131 assignment leads to production of exotic quantum number @xmath209 .
if indeed the @xmath210 has a partial width of @xmath211 mev , its photoproduction cross section near threshold will also be significant . by measuring the angular distribution of @xmath212 ,
its quantum numbers can also be determined .
in this work we have made a detailed analysis of the photoproduction reaction @xmath213 in an effective lagrangian approach for possible assignments of @xmath3 , @xmath4 and @xmath131 to the @xmath113 . as suggested by the observation of belle ,
we show that the @xmath37-channel pion exchange is an important production mechanism for creating @xmath33 at the rate of several tens of nano - barn . in this process
the @xmath87 coupling can be well constrained and the @xmath56 coupling can be extracted from the @xmath45 data via vmd model . we have also checked the possible effect of @xmath37-channel @xmath39 exchange .
we find its contributions are relatively small even if , as an upper limit , we fix the @xmath110 coupling to be the same as @xmath61 . since there is no interference between the @xmath39 and pion exchange amplitudes , we can neglect the @xmath39 exchange in the calculation and only adopt the relatively well - defined pion exchange for estimating the lower bound of the @xmath0 photoproduction cross sections .
we find that due to pion exchange , a clear threshold enhancement is predicted for @xmath0 photoproduction , which is different from photoproduction of @xmath32 and @xmath55 near threshold , and could be a signature for the @xmath0 existence .
this is also the kinematic region that uncertainties arising from other unknown processes can be overlooked as a reasonable approximate .
we also consider possible background contributions to the signals in @xmath214 , where the diffractive production of @xmath55 turns to be the leading source .
interestingly , it shows that the interferences from the background terms are rather small . in the dalitz plot at @xmath202 gev ( @xmath215 gev ) , a clear band of the @xmath46 meson signals appear in the invariant mass spectrum of @xmath51 .
similar results are obtained for the neutral production channel . for other background contributions from diffractive transitions
, we argue that they may be suppressed by isospin conservation and momentum mismatching .
although nucleon resonance contributions to the background in fig . [ fig:2](a ) and ( b ) have not been included in this calculation , their effects are unlikely to submerge the @xmath0 band in the dalitz plot .
the reason is that the nucleon resonance cross sections will accumulate horizontally in a form of uneven bands , and be located at @xmath216 gev@xmath206 . in the kinematics of higher @xmath204 interferences from the nucleon resonances
are expected to be negligible . as a consequence
, the @xmath0 production signals can be rather cleanly isolated from the background , and the measurement of the angular distributions between @xmath55 and @xmath14 will tell whether they are in a relative @xmath44 or @xmath22 wave .
the quantum numbers of the @xmath0 can thus be determined . in summary , in order to establish the @xmath0 as a concrete candidate of tetraquark state and
determine its quantum numbers , further search for its signals in other production processes is necessary .
supposing it is a genuine resonance with sizeable partial decay width to @xmath1 @xcite , we have argued that photoproduction should be useful for providing a clear and consistent evidence for its existence .
experimental data which have been collected at h1 and zeus may already contain such information .
useful discussion with zhiqing zhang about experiments at hera is acknowledged . this work is supported , in part , by the u.k .
epsrc ( grant no .
gr / s99433/01 ) , national natural science foundation of china ( grant no.10675131 ) , and chinese academy of sciences ( kjcx3-syw - n2 ) .
f.e.c . acknowledges support from the u.k .
science and technology facilities council ( stfc ) and the eu - rtn programme contract no .
mrtn - ct-2006 - 035 482 : flavianet " .
we give below the details for @xmath39 exchange .
the lagrangian for the @xmath110 is : @xmath217 where @xmath218 is the coupling constant . at this moment
, no information about this coupling is available .
a conservative estimate is to set it the same as @xmath61 as an upper limit . for the production of @xmath0 with @xmath43 ,
its difference from @xmath3 is at the @xmath45 and @xmath226 vertices for which the following effective lagrangians are adopted @xcite : @xmath227 where the notations are the same as eq .
( [ lagrangian - v ] ) . .
@xmath229 is the c.m .
the dashed line denotes the meson ( @xmath14 and @xmath39 ) exchange contribution , the dotted line is the pomeron - exchange contribution , and the solid line is the full contribution . ] | we propose a search for the newly discovered tetraquark candidate @xmath0 in photoproduction .
based on the belle results we show that if @xmath0 is a genuine resonance , its significantly large coupling to @xmath1 will cause it to stand out above the background in @xmath2 .
we consider the dependence of the cross section for the quantum numbers ( @xmath3 , @xmath4 or @xmath5 ) . |
kernel methods play an important role in solving various machine learning problems , such as non - linear regression and classification tasks @xcite .
the kernel methods rely on the so called `` kernel trick '' , which transforms a given non - linear problem into a linear one by using a kernel function @xmath0 , which is a similarity function defined over pairs of input data points @xmath1 .
this way , the input data @xmath2 is mapped into a high dimensional ( or even infinite - dimensional ) feature space @xmath3 , where the inner product @xmath4 can be calculated with a positive definite kernel function ( that is satisfying mercer s condition)@xcite : @xmath5 therefore , the mapping into the high dimensional feature space is done implicitly , without the need to explicitly map the data points @xmath3 . also , assuming that @xmath6 is the training data , then using the representer theorem any non - linear function @xmath7 can be expressed as a linear combination of kernel products evaluated on the training data points @xcite : @xmath8 the main methods for the kernel classification problems are the support vector machines ( svm)@xcite and the least - squares support vector machines ( ls - svm)@xcite . in this paper
we focus on the ls - svm classifier , where the main difficulty is the @xmath9 training complexity , where @xmath10 is the size of the training data set . because of this high complexity
, the ls - svm method is not a suitable candidate for applications with large data sets . here , we discuss several approximation methods using randomized block kernel matrices , that significantly reduce the complexity of the problem .
the proposed methods are based on the nystrm@xcite , kaczmarz@xcite and matching pursuit@xcite algorithms , and we show that they provide good accuracy and reliable scaling to relatively large multi - class classification problems .
in the binary classification setting the kernel svm method constructs an optimal separating hyperplane ( with the maximal margin ) between the two classes in the feature space @xmath3 .
the training problem is represented as a ( convex ) quadratic programming problem involving inequality constraints , which has a global and unique solution @xcite .
the kernel ls - svm method simplifies the optimization problem by considering equality constraints only , such that the solution is obtained by solving a system of linear equations @xcite . with these modifications ,
the problem is equivalent to a ridge regression formulation with binary targets @xmath11 .
also , the kernel ls - svm allows us to treat the multi - class classification task in a much simple and compact way .
more exactly , we assume that @xmath12 classes are encoded using the standard basis in the @xmath13 space . therefore ,
if @xmath14 is an input in class @xmath15 , then the output @xmath16 is encoded by a binary row vector with 1 in the @xmath17th position and 0 in all other positions : @xmath18 thus , for the input data @xmath19 and the feature mapping function @xmath3 we consider the following optimization problem:@xcite @xmath20 such that : @xmath21 where @xmath22 is the bias coefficient , @xmath23 is the vector of weights corresponding to the class @xmath24 , and @xmath25 is the approximation error .
the objective function @xmath26 is a sum of the least - squares error and a regularization term , and therefore it corresponds to a multi - dimensional ridge regression problem with the regularization parameter @xmath27 . in the primal weight space
the multi - class classifier takes the form : @xmath28 where @xmath29 is the nonlinear softmax function : @xmath30 because @xmath3 may be infinite dimensional we seek a solution in the dual space @xmath3 by introducing the lagrangian : @xmath31,\ ] ] with the lagrange multipliers @xmath32 .
the optimality conditions for the minimization problem are : @xmath33 by eliminating @xmath23 and @xmath25 we obtain : @xmath34 a_{nj } + b_j = y_{ij } , \quad i=1,\dots , n,\ ; j=1,\dots , k,\ ] ] where we applied the condition @xmath35 , and @xmath36 is kronecker s delta : @xmath37 if @xmath38 , and @xmath39 otherwise . therefore , in the dual space the multi - class classifier takes the form : @xmath40 one can see that the above system of equations ( 13 ) is equivalent to @xmath12 independent systems of equations with binary targets @xmath41 .
each system can be written in the following equivalent form : @xmath42 where @xmath43 is the @xmath44 identity matrix , @xmath45^t$ ] is an @xmath10 dimensional vector with all the components equal to 1 , @xmath46^t$ ] and @xmath47^t$ ] are the weight and respectively the target column vectors for the class @xmath48 .
each system has @xmath49 linear equations with @xmath49 unknowns , and requires the inversion of the same matrix : @xmath50 the above @xmath12 systems can be written in a compact matrix form as following : @xmath51 where @xmath52 and @xmath53 are @xmath54 matrices with the columns : @xmath55 despite these nice mathematical properties , the complexity of the problem is @xmath56 , and it is implied by the matrix inversion .
our goal is to reduce the complexity of the kernel ls - svm classification method by using iterative approximations based on randomized block kernel matrices .
more exactly we use the well known nystrm method,@xcite and we introduce new methods based on the kaczmarz@xcite and the matching pursuit@xcite algorithms . the nystrm method is probably the most popular low rank approximation of the matrix @xmath57.@xcite since the kernel matrix is a positive definite symmetric matrix we can find its eigenvalue decomposition : @xmath58 where @xmath59 is the diagonal matrix of the eigenvalues , and @xmath60 is the corresponding matrix of the eigenvectors . using the sherman - morrison formula@xcite one can show that the solution of the linear system of equations can be written as : @xmath61\ ] ] to reduce the computation one can consider only a small selection @xmath62 , @xmath63 , based on a random sample of the training data .
the eigenvalue decomposition of @xmath64 is therefore : @xmath65 one can show that the following relations exist between the eigenvalues and eigenvectors of the two matrices @xmath57 and respectively @xmath66 : @xmath67 such that : @xmath68 where @xmath69 and @xmath70 are the @xmath71 and respectively the @xmath72 block submatrices taken from @xmath57 , corresponding to the randomly selected @xmath73 columns . therefore , in this case only a matrix of size @xmath74 needs to be inverted , which simplifies the computation .
while this method seems to scale as @xmath75 it has the difficulty of optimally choosing the @xmath73 support vectors from the training data set .
it has been shown that an optimal method for choosing the @xmath73 support vectors is based on the maximization of the renyi entropy of the the matrix @xmath66.@xcite therefore , one can use a heuristic greedy algorithm to search for the best set of @xmath73 support vectors , that maximize the renyi entropy of @xmath66 , however this approach becomes difficult in the case of large data sets .
here we consider a much simpler approach , based on a committee machine made of @xmath76 classifiers .
let us assume that the classifiers are characterized by the following randomly selected matrices : @xmath77 , @xmath78 , @xmath79 , @xmath80 .
the random sampling can be done without replacement , such that after @xmath81 selections all the training data is used .
an efficient randomization strategy would be to use a random shuffling function which generates a random permutation @xmath82 of the set @xmath83 , and then to select contiguous subsets with size @xmath73 from @xmath84 .
after the algorithm consumes all the @xmath81 subsets one can re - shuffle @xmath84 in order to create another randomized list . for each classifier we solve the system : @xmath85 the solution is given by : @xmath86 where @xmath87 and @xmath88 are the moore - penrose pseudo - inverse matrices of the random block matrices @xmath89 and @xmath90 , which can be calculated at a lower cost of @xmath91 . in the end we aggregate the weights of the classifiers as following : @xmath92 and the classification process is performed using the equations ( 14 ) and ( 15 ) .
the kaczmarz method@xcite is a popular solver for overdetermined linear systems of equations of the form : @xmath93 with @xmath94 .
the method has numerous applications in computer tomography and image reconstruction from projections .
assuming that @xmath95 is an initial estimation of the solution , the algorithm proceeds cyclically and at each step it projects the estimate @xmath96 onto a subspace normal to the row @xmath97 of @xmath98 , such that : @xmath99 because of its inherent cyclical definition , the convergence of the method depends on the order of the rows . a simple method to overcome
this problem is to select the rows of @xmath98 in a random order : @xmath100 where @xmath101 is a random function .
it has been shown that if the rows are selected with a probability : @xmath102 where @xmath103 if the frobenius norm , then the algorithm converges in expectation exponentially , with a rate independent on the number of equations : @xmath104 where @xmath105 is the condition number of @xmath98 , with @xmath106 and @xmath107 the maximal and minimal singular values of @xmath98 respectively.@xcite this remarkable results shows that the estimate @xmath96 converges exponentially fast to the solution , in just @xmath108 iterations . also ,
since each iteration requires @xmath108 time steps , the overall complexity of the method is @xmath109 , which is much smaller than @xmath110 required by the standard approach .
another big advantage is the low memory requirement , since the method only needs the randomly chosen row at a given time , and it does nt operate with the whole matrix @xmath98 . this makes the method appealing for very large systems of equations .
it is interesting to note that the matrix @xmath98 can be `` standardized '' , such that each of its rows has unit euclidean norm : @xmath111 where @xmath112 therefore , by using this simple standardization method , the random row selection can be done with uniform probability , since all the rows have an identical norm @xmath113 , @xmath114 . in this case
the iterative equation takes the form : @xmath115 \hat{a}_{r(i)},\ ] ] and @xmath101 is a uniformly distributed random function . in the case of our multi - class classification problem , we prefer a randomized block version of the kaczmarz method , in order to use more efficiently the computing resources .
also we assume that the system of equations is standardized using the above described procedure . in this case at each iteration we randomly select a subset @xmath116 of the row indexes of @xmath98 , with a size @xmath117 , and we project the current estimate @xmath96 onto a subspace normal to these rows : @xmath118,\ ] ] where : @xmath119 ^{-1}\ ] ] is the moore - penrose pseudo - inverse of the matrix @xmath120 , and @xmath121 is the subvector of @xmath122 with the components from @xmath116 .
the above equation is equivalent to : @xmath123,\ ] ] where @xmath2 is the exact solution of the system .
therefore we have : @xmath124.\ ] ] this can be rewritten as : @xmath125.\ ] ] where @xmath126 . since @xmath127 is an orthogonal projection we have : @xmath128 \vert \leq \vert
x(t ) - x \vert^2,\ ] ] and the algorithm is guaranteed to converge .
in fact , for @xmath129 , the randomized block algorithm is equivalent to the simple ( one row at a time ) randomized algorithm , which guarantees exponential convergence.@xcite in our multi - class classification setting we solve simultaneously @xmath12 systems of equations , corresponding to the @xmath12 different classes : @xmath130 therefore , the iterative equations are : @xmath131 , \quad j=1,\dots , k,\ ] ] where @xmath132 and @xmath133 are the standardized versions of the randomly selected block matrix @xmath134 and the corresponding target subvector @xmath135 .
these equations can be written compactly in a matrix form as following : @xmath136.\ ] ] the iteration stops when no significant improvement for @xmath52 is made , or the number of iterations reach a maximum accepted value .
once the matrix @xmath52 containing the weights @xmath137 and the bias @xmath138 is calculated , one can use the multi - class classifier defined by the equations ( 14)-(15 ) , in order to classify any new @xmath139 data sample .
the matching pursuit ( mp ) method is frequently used to decompose a given signal into a linear expansion of functions selected from a redundant dictionary of functions.@xcite thus , given a signal @xmath140 , we seek a linear expansion approximation : @xmath141 where @xmath142 are the columns of the redundant dictionary @xmath143 , with @xmath144 . here , @xmath145 is a selection function which returns the index of the column from the dictionary .
thus , we solve the minimization problem : @xmath146 starting with an initial approximation @xmath147 , at each new step the algorithm iteratively selects a new column @xmath148 from the dictionary , in order to reduce the future residual @xmath149 .
therefore , from @xmath150 one can build a new approximation : @xmath151 by finding @xmath152 and @xmath148 that minimizes the residual error : @xmath153 the minimization condition : @xmath154 gives : @xmath155 and therefore we have : @xmath156 ^ 2 \leq \vert r_{m-1 } \vert^2.\ ] ] the index of the best column that minimizes the residual is given by : @xmath157 ^ 2.\ ] ] the mp algorithm can be easily extended for kernel functions.@xcite however , for large data sets this approach is not quite efficient due to the increasing time required by the search for the best column .
we notice that according to ( 54 ) the algorithm converges even if the selection of the column is randomly done .
therefore , here we prefer this weaker form , where the functions from the redundant dictionary are simply selected randomly . in our case
, we consider that the dictionary corresponds to the matrix @xmath57 , and as before we consider a block matrix formulation of the weak matching pursuit algorithm .
initially the matrix @xmath52 is empty , @xmath158 , and the residual is set to @xmath159 . at each iteration step
we select randomly ( without replacement ) @xmath73 columns from @xmath57 , which we denote as the @xmath71 matrix @xmath134 , where @xmath116 is the random list of columns .
thus , at each iteration step we solve the following system : @xmath160 from here we find the solution at time @xmath161 : @xmath162 the weights are updated as : @xmath163 here , @xmath164 is the block of @xmath52 containing only the rows from the random list @xmath116 .
also , the new residual is updated as following : @xmath165 due to the orthogonality of these quantities we have : @xmath166 and the algorithm is guaranteed to converge .
once the matrix @xmath52 is calculated , one can use the multi - class classifier defined by the equations ( 14)-(15 ) , in order to classify any new @xmath139 data sample .
interestingly , one can easily combine the kaczmarz and the mp methods into a hybrid kaczmarz - mp kernel method . in both methods we have to select a random list @xmath116 of @xmath73 indexes ,
therefore we can use the same list to perform a kaczmarz step followed by an mp step ( or vice versa ) .
the cost of this approach is similar to the randomized nystrm method , since it requires two matrix pseudo - inverses per iteration step .
in this section we provide several new numerical results obtained for the mnist@xcite and the cifar-10@xcite data sets , that illustrate the practical applicability and the performance of the proposed methods .
the mnist data set of hand - written digits is a a widely - used benchmark for testing multi - class classification algorithms .
the data consists of @xmath167 training samples and 10,000 test samples , corresponding to @xmath168 different classes .
all the images @xmath169 , @xmath114 , are of size @xmath170 pixels , with @xmath171 gray levels , and the intra - class variability consists in local translations and deformations of the hand written digits . some typical examples of images from the mnist data set are shown in figure 1 . while the mnist data set is not very large , it still imposes a challenge for the standard kernel classification approach based on ls - svm , because the kernel matrix has @xmath172 elements , which in double precision ( 64bit floating point numbers ) requires about 28.8 gb . in order to simulate even more drastic conditions we used a pc with only 8 gb of ram , and 4 cpu cores .
also , all the simulation programs were written using the julia language.@xcite obviously , with such a limited machine , the attempt to solve the problem directly ( for example using the gaussian elimination or the cholesky decomposition ) is not feasible .
however , we can easily use the proposed randomized approximation methods , and in order to do so we perform all the computations in single precision ( 32 bit floating point numbers ) , which provides a double storage capability , comparing to the case of double precision .
thus , we trade a potential precision loss for extra memory storage space , in order to adapt to the data size . in our numerical experiments
we have used only the raw data , without any augmentation or distortion .
it is well known that better results can be obtained by using more complex unsupervised learning of image features , data augmentation and image distortion methods at the pre - processing level .
however , our goal here is only to evaluate the performance of the discussed methods , and therefore we prefer to limit our investigation to the raw data . to our knowledge ,
the best reported results in the literature for the kernel svm classification of the mnist raw data have a classification error of @xmath173,@xcite and respectively @xmath174,@xcite and they have been obtained by combining ten kernel svm classifiers .
we will use these values for comparison with our results .
in our approach we used a simple pre - processing method , consisting in a two step normalization of each image , as following : @xmath175 where @xmath176 denotes the average .
also , the images are `` vectorized '' by concatenating the columns , such that : @xmath177 . therefore , after pre - processing all the images are unit norm vectors : @xmath178 , @xmath114 .
this normalization is useful because it makes all the inner products equal to the cosine of the angle between the vectors , which is a good similarity measure : @xmath179 , \quad \forall i , j=1,\dots , n.\ ] ] we have experimented with several kernel types ( gaussian , polynomial ) , and the best results have been obtained with a polynomial kernel of degree 4 : @xmath180 therefore all the results reported here are for this particular kernel function .
also , the regularization parameter was always set to @xmath181 , and the classification error @xmath182 was simply measured as the percentage of the test images which have been incorrectly classified . the full kernel matrix : @xmath183 would still require 14.4 gb , which is not feasible with the imposed constraints , and
therefore we must calculate the kernel elements on the fly at each step . here
, the exponent @xmath184 means that the power is calculated element - wise .
also we assume that the vectorized images correspond to the rows of the matrix @xmath185 .
the memory left is still enough to hold the matrix : @xmath186 which is required for the classification of the testing data .
this matrix is not really necessary to store in memory since the classification can be done separately for each test image , once the weights and the biases have been computed .
this matrix requires about 2.4 gb and it is convenient to store it just to be able to perform a fast measurement of the classification error after each iteration . for the iterative randomized kernel methods ( mnist data set ) : ( a ) nystrm method ; ( b ) kaczmarz method ; ( c ) mp method ;
( d ) kaczmarz - mp method .
here @xmath161 is the iteration time step and @xmath73 is the size of the random block submatrices . , width=566 ] for the randomized mp method with the fourier pre - processing ( mnist data set ) . here
@xmath161 is the iteration time step and @xmath73 is the size of the random block submatrices.,width=377 ] in figure 2(a ) we give the results obtained for the randomized nystrm method .
here we have represented the classification error as a function of the random block size @xmath73 and the number of aggregated classifiers @xmath161 .
one can see that the method converges very fast , and only few classifiers are needed to reach a plateau for @xmath182 .
unfortunately , the error @xmath182 is dependent on @xmath73 , and this result suggests that the method performs better for larger values of @xmath73 .
this is the fastest method considered here , since it requires the aggregation of only a few classifiers , such that for @xmath187 the total number of necessary classifiers is only @xmath188 .
the randomized kaczmarz method shows a different behavior , figure 2(b ) . in this case
we ran the algorithm for @xmath189 iterations using random data blocks of size @xmath190 .
this method shows a much slower convergence , and the whole computation process needed almost 5 hours to complete ( @xmath191 seconds per iteration time step ) , including the time for the error evaluation at each step .
after the first iteration step the classification error drops abruptly below @xmath192 , then the classification error decreases slowly , and fluctuates closer to @xmath193 , which is in very good agreement with the previously reported values .
better results for the kaczmarz method can be obtained by simply averaging the weights and biases of several separately ( parallel ) trained classifiers , with different random seeds for the random number generator , and/or different block sizes @xmath73 .
this form of aggregation is possible because the kernel matrix is the same for all classifiers , and the systems are linear in the dual space . assuming that we have @xmath26 classifiers , @xmath194 , each of them having the output : @xmath195 then the output of the averaged classifiers is : @xmath196 therefore , in the end one can store only the average values of the weights and the biases .
the advantage of averaging is the decrease of the amplitude of the fluctuations in the classification error , which gives better and more confident results , and also increases the generalization capabilities . in figure 2(c ) we give the results for the randomized mp method .
this method shows similar results to the randomized kaczmarz method , but it s convergence is faster and the fluctuations have a lower amplitude . also , the method is about twice faster , with @xmath197 seconds per iteration time step , for a random block of size of @xmath190 .
again , the obtained result @xmath198 is in very good agreement with the previously reported results .
the results for the hybrid kaczmarz - mp method are given in figure 2(d ) , and not surprisingly it shows inbetween convergence speed , and a classification error of @xmath199 , which confirms the previous results . in our last experiment
we have decided to modify the data pre - processing , in order to see if better results can be obtained .
we only did a simple modification by concatenating the images with the square root of the absolute value of their fast fourier transform ( @xmath200 ) .
since the fft is symmetrical , only the first half of the values were used , each image becoming a vector of 1176 elements .
the new image pre - processing consists in the following steps : @xmath201^t , \quad i=1,\dots , n.\end{aligned}\ ] ] with this new pre - processing we used the randomized mp method and the results are shown in figure 3 .
the classification error drops to @xmath202 , which means an improvement of @xmath203 over the previously reported results .
the 85 images incorrectly recognized are shown in figure 4 .
the cifar-10 dataset consists of 60,000 images .
each image has @xmath204 pixels and 3 colour ( rgb ) channels with 256 levels .
the data set contains 10 classes of equal size ( 6,000 ) , corresponding to objects ( airplane , automobile , ship , truck ) and animals ( bird , cat , deer , dog , frog , horse ) .
the training set contains 50,000 images , while the test set contains the rest of 10,000 images . in figure 5
we show the first 10 images from each class .
obviously this data set is much more challenging than the mnist data set , showing a very high variation in each class .
again , we only use the raw data without any augmentation and distortion , and without employing any other technique for advanced feature extraction and learning . in figure 6
we give the results of the numerical experiments for the randomized mp method , with the random data block size @xmath205 . in the first experiment ( @xmath206 ) we used the simple two - step normalization described in ( 61)-(62 ) .
thus , after the normalization steps each image becomes a unit vector with @xmath207 pixels , obtained by concatenating the columns from the three colour channels .
we ran the randomized mp algorithm for @xmath208 steps , such that the classification error settled at @xmath209 .
this is a good improvement over the random guessing `` method '' , which has an error @xmath210 . in the second experiment ( @xmath211 ) , before the two - step normalization we have applied a gaussian filter on each channel .
the filter is centred in the middle of the image and it is given by : @xmath212 \right ) , \
; i , j=1,\dots , l,\ ] ] where @xmath213 is the size of the image .
the filter is applied element - wise to the pixels and it is used to enhance the center of the image , where supposedly the important information resides , and to attenuate the information at the periphery of the image .
the best results have been obtained for a constant @xmath214 , and the classification error dropped to @xmath215 . in the third experiment ( @xmath216 ) we have used the gaussian filtering ( 76 ) and the fft normalization described by the equations ( 69)-(75 ) .
this pre - processing method improved the results again and we have obtained an error of only @xmath217 . the resulted confusion matrix is given in table 1 .
not surprisingly , one can see that main `` culprits '' are the cats and dogs .
for @xmath218 the cats are mistaken for dogs , while for @xmath219 the dogs are mistaken for cats . for the randomized mp method ( cifar-10 data set ) . here
@xmath161 is the iteration time step and @xmath73 is the size of the random block submatrices .
the following pre - processing steps have been used : @xmath206 - two - step data normalization ( 61)-(62 ) ; @xmath211 - gaussian filtering ( 76 ) and two - step data normalization ( 61)-(62 ) ; @xmath216 - gaussian filtering ( 76 ) and fft normalization ( 69)-(75 ) .
, width=340 ]
in this paper we have discussed several approximation methods for the kernel ls - svm multi - class classification problem .
these methods use randomized block kernel matrices , and their main advantages are the low memory requirements and the relatively simple iterative implementation , significantly reducing the complexity of the problem .
another advantage is that these methods do not require complex parameter tuning mechanisms .
the only parameters needed are the regularization parameter and the size of the random block matrices . despite of their simplicity
, these methods provide very good accuracy and reliable scaling to relatively large multi - class classification problems . also , we have reported new numerical results for the mnist and cifar-10 data sets , and we have shown that better results can be obtained by using a simple fourier pre - processing step of the raw data .
the results reported here for the mnist raw data set are in very good agreement with the previously reported results using the kernel svm approach , while the results for the cifar-10 data are surprisingly good for the small amount of pre - processing used .
t. hofmann , b. schlkopf , a. j. smola , _ kernel methods in machine learning , _ the annals of statistics 36(3 ) , 1171 ( 2008 ) .
j. mercer , _ functions of positive and negative type and their connection with the theory of integral equations , _ philos .
ser . a math .
a 209 , 415 ( 1909 ) . c. cortes , v. vapnik , _ support vector networks _ , machine learning 20 , 273 ( 1995 ) .
suykens , j. vandewalle , _ least squares support vector machine classifiers _ , neural processing letters 9(3 ) , 293 ( 1999 ) .
e. j. nystrm , _ ber die praktische auflsung von integralgleichungen mit anwendungen auf randwertaufgaben " , _ acta mathematica 54(1 ) , 185 ( 1930 ) .
s. kaczmarz , _ angenherte auflsung von systemen linearer gleichungen , _ bulletin international de lacadmie polonaise des sciences et des lettres .
classe des sciences mathmatiques et naturelles .
srie a , sciences mathmatiques , 35 , 355 ( 1937 ) . s. g. mallat , z. zhang , _ matching pursuits with time - frequency dictionaries , _ ieee transactions on signal processing 41(12 ) , 3397 ( 1993 ) .
c.k.i williams , m. seeger , _ using the nystrm method to speed up kernel machines , _
proceedings neural information processing systems 13 , 682 ( 2001 ) .
j. sherman , w.j .
morrison , _ adjustment of an inverse matrix corresponding to a change in one element of a given matrix , _ annals of mathematical statistics 21 , 124 ( 1950 ) .
m. girolami , _
orthogonal series density estimation and the kernel eigenvalue problem , _ neural computation 14 , 669 ( 2002 ) .
t. strohmer , r. vershynin , _ a randomized kaczmarz algorithm for linear systems with exponential convergence _ ,
journal of fourier analysis and applications 15 , 262 ( 2009 ) .
p. vincent , y. bengio , _ kernel matching pursuit , _ machine learning 48 , 169 ( 2002 ) .
y. lecun , l. bottou , y. bengio , p. haffner , _ gradient - based learning applied to document recognition , _ proceedings of the ieee 86(11 ) , 2278 ( 1998 ) .
a. krizhevsky , g. hinton .
_ learning multiple layers of features from tiny images , _ computer science department , university of toronto , tech . rep . , 2009 .
j. bezanson , s. karpinski , v.b .
shah , a. edelman , _ julia : a fast dynamic language for technical computing _ , arxiv:1209.5145 ( 2012 ) .
burges , b. schlkopf , _ improving the accuracy and speed of support vector machines _ , advances in neural information processing systems 9 , 375 ( 1997 ) . | the least - squares support vector machine is a frequently used kernel method for non - linear regression and classification tasks .
here we discuss several approximation algorithms for the least - squares support vector machine classifier .
the proposed methods are based on randomized block kernel matrices , and we show that they provide good accuracy and reliable scaling for multi - class classification problems with relatively large data sets . also , we present several numerical experiments that illustrate the practical applicability of the proposed methods . keywords : kernel methods ; multiclass classification .
pacs : 07.05.mh , 02.10.yn ; 02.30.mv |
macroscopic properties of granular materials such as soils depend on particle interactions . in dry granular materials , interparticle forces
are related to the applied external loads as different studies have shown @xcite . in unsaturated soils subjected to capillary effects , new features must be accounted for in order to understand properly their behaviour .
the presence of water leads to the formation of water menisci between neighboring grains , introducing new interparticle forces .
the effects of these forces depend on the degree of saturation of the medium . for low water content level corresponding to disconnected liquid bridges between grains ,
capillary theory allows the force induced by those bridges to be linked to the local geometry of the grains and to the matric suction or capillary pressure inside the medium @xcite .
since the disconnected menisci assumption is not valid for high water content levels due to water percolation , we consider here only the unsaturated state where the discontinuity of the water phase can be assumed , the so - called pendular regime .
+ there has been a wide debate on the various interpretations for the mechanical behaviour of unsaturated soils . at early stages of soil mechanics ,
terzaghi @xcite first introduced the concept of effective stress for the particular case of saturated soils enabling the conversion of a multiphase porous medium into a mechanically equivalent single - phase continuum . in unsaturated soils ,
water induced stresses are still debated .
the common practice @xcite is to use the suction or a modified version as a second stress variable within a complete hydro - mechanical framework .
an alternative method is introduced to develop homogenisation techniques in order to derive stress - strain relationships from forces and displacements at the particle level as proposed in @xcite for dry granular materials .
the basic idea is to consider the material as represented by a set of micro - systems , postulating that the behaviour of a material volume element depends on the intergranular interactions belonging to this volume .
we propose here to extend this micro - mechanical approach to unsaturated granular materials as proposed by li @xcite , jiang et al .
@xcite or lu and likos @xcite .
+ along these lines we present two micromechanical models which take into account capillary forces .
the first one is a three dimensional numerical model based on the discrete element method ( hereafter designed as the dem model ) pioneered by cundall and strack @xcite , and the second one is an analytical model ( hereafter designed as the microstructural model ) recently proposed by hicher and chang @xcite .
the microstructural model is a stress - strain relation which considers interparticle forces and displacements .
thanks to analytical homogenisation / localisation techniques , the macroscopic relation between stress and strain can be derived . in the dem model , a granular medium
is modelled by a set of grains interacting according to elementary laws .
direct simulations are carried out on grain assemblies , computing the response of the material along a given loading path . by studying their effects under triaxial loading
, we investigate capillary forces implications at the macroscopic level , and offer an insight into the unsaturated soil stress framework by introducing a capillary stress tensor as a result of homogenization techniques .
macroscopic interpretations of the mechanical behaviour of unsaturated soils have been mainly developed in the framework of elastoplasticity @xcite .
most of these models consider that the strain tensor is governed by the net stress tensor @xmath0 ( @xmath1 being the pore air pressure ) and the matric suction or capillary pressure @xmath2 ( @xmath3 being the pore water pressure ) inside the medium @xcite . in particular , they consider a new yield surface , called loading collapse ( lc ) surface in the plane ( ( @xmath0),(@xmath2 ) ) which controls the volume changes due to the evolution of the degree of saturation for a given loading path . as a matter of fact , all these formulations can be considered as extensions of the relationship initially proposed by bishop and blight @xcite for unsaturated soils : @xmath4 where @xmath5 is called the effective stress parameter or bishop s parameter , and is a function of the degree of saturation @xmath6 of the medium ( @xmath7 for a dry material , @xmath8 for a fully saturated material ) .
+ obviously , since the effective stress principle is by definition a macroscopic concept , several authors ( lu and likos @xcite or li @xcite ) have proposed to use a micromechanical approach for the effective stress principle . in order to further study this micromechanical approach to study unsaturated soil
stresses , we propose here a micromechanical analysis of the problem , examining the local water induced effects through a set of simulated laboratory experiments .
let us consider a representative volume element ( rve ) of a wet granular material , subjected to an assigned external loading .
when the water content decreases inside a saturated granular sample , the air breaks through at a given state . the capillary pressure ( @xmath2 ) corresponding to that point
is called the air - entry pressure , and strongly depends on the pore sizes .
thereafter , the sample becomes unsaturated and capillary forces start to grow due to interface actions between air and water .
since the the gaseous phase is discontinue , this is the capillary regime . from this state , a constant decrease in the degree of saturation corresponds to a gentle increase in pore water pressure .
the pendular regime starts when the water phase is no longer continuous . in this state ,
fluids equilibrium is obtained by the vapor pressure .
analytical and experimental results @xcite demonstrate that capillary effects at particle contacts produces a kind of bond between particles as a result of menisci ( fig.[fig1 ] ) .
liquid bridges may form between some pairs of adjoining particles not necessarily in contact , generating an attractive capillary force between the bonded particles .
if the drying process continues , these water bridges begin to fail , starting from the non - contacting grains , until the complete disappearance of capillary forces inside the assembly . + as the pendular regime is considered throughout this paper , water is considered to be solely composed of capillary menisci : each liquid bridge is assumed to connect only two particles .
therefore , two types of forces coexist within the granular medium . for dry contacts
, a contact force develops between contacting granules .
this repulsive force , that is a function of the relative motion between the contacting grains , is usually well described by an elastoplastic contact model . for water bonded particles , a specific attractive force exists .
this water - induced attractive interaction can be described by a resulting capillary force , rather than by a stress distribution as mentionned by haines @xcite or fisher @xcite .
this capillary force is a function of the bridge volume , of the size of particles , and of the fluid nature ( see section 3.1.1 for the details ) .
the objective of this section is to derive , in a simple manner , an expression relating the overall stress tensor within the rve to this internal force distribution .
+ for this purpose , the love @xcite static homogenisation relation is used .
this relation expresses the mean stress tensor @xmath9 within a granular volume @xmath10 as a function of the external forces @xmath11 applied to the particles @xmath12 belonging to the boundary @xmath13 of the volume : @xmath14 where @xmath15 are the coordinates of the particle @xmath12 with respect to a suitable frame .
it is worth noting that this relation is valid whatever the nature of the interactions between grains .
+ taking into account the mechanical balance of each particle of the volume @xmath10 ( including the boundary @xmath13 ) , eq.(2 ) can be written as : @xmath16 where @xmath17 is the number of particles within the volume , @xmath18 is the interaction force exerted by the particle @xmath19 onto the particle @xmath12 , and @xmath20 is the branch vector pointing from particle @xmath19 to particle @xmath12 ( @xmath21 ) .
+ as we consider partially saturated granular media , two independent kinds of interparticle forces can be distinguished : 1 . if particles @xmath12 and @xmath19 are in contact , a contact force @xmath22 exists
2 . if particles @xmath12 and @xmath19 are bonded by a liquid bridge , a capillary force @xmath23 exists . actually ,
depending on the local geometry , a liquid bond can exist between two grains in contact . in that case ,
solid contacts are surrounded by the continuous liquid phase providing the simultaneity of contact and capillary forces .
the two contributions have therefore to be accounted for by summation .
+ finally , in all cases and for any couple @xmath24 ^ 2 $ ] , it can be written that : @xmath25 thus , by combining eqs.(2 ) and ( 4 ) , it follows that : @xmath26 as a consequence , eq.(5 ) indicates that the stress tensor is split into two components : @xmath27 1 . a first component @xmath28 accounting for the contact forces transmitted along the contact network .
2 . a second component @xmath29 representing the capillary forces existing within the assembly . it is to be noted that @xmath30 is a stress quantity standing for intergranular contact forces in the same way as in saturated or dry conditions . considering the concept as initially introduced by terzaghi
, @xmath30 plays the role of the so - called effective stress by governing soil deformation and failure . besides , @xmath31 is the tensorial attribute to capillary water effects or suction , by extension .
by analogy with eq.(1 ) , we can therefore define a microstructural effective stress @xmath32 where @xmath33 could be affiliated to net stress , representing the apparent stress state in the material .
compared with eq.(1 ) , where the effect of water is intrinsically isotropic , @xmath31 implies a tensorial attribute to the water effects .
+ in fact , in both terms @xmath30 and @xmath31 , a fabric tensor can emerge from the summation @xcite .
the fabric tensor is useful to characterize the contact anisotropy of the assembly , which is known as a basic feature of granular assemblies . in dry granular materials ,
an induced anisotropy can develop when a deviatoric stress loading is applied . in partially saturated assembly , due to the possibility of interactions without contact ,
the conclusion is not so trivial . as an illustration ,
if we restrict our analysis to spherical particles @xcite , it can be inferred that : @xmath34 this relation points out that , contrary to the contact term where the fabric tensor @xmath35 is directly linked to the induced anisotropy ( a contact is associated with a force ) , two causes can be invoked for the capillary term .
first , the distribution of the liquid bonds can be anisotropic .
secondly , the geometry of the bonds being obviously dependent on the local geometry , it is possible that the distribution of both terms @xmath36 and @xmath37 is also anisotropic .
this is significant because the anisotropic attribute yields shear effects associated with pore fluid , which could mostly influence material behaviour . + in order to enrich our discussion , we present numerical investigations of these features in the following sections using both dem and micromechanical simulations .
we present here a numerical analysis of the stress variables using a micromechanical model based upon the discrete element method initially introduced by cundall and strack @xcite .
this technique starts with basic constitutive laws between interacting particles and can provide a macroscopic response of a particle assembly due to loading changes at the boundaries .
each particle of the material is a rigid sphere identified by its own mass , @xmath38 , radius , @xmath39 and moment of inertia , @xmath40 . for every time step of the computation , interaction forces between particles , and consequently resulting forces acting on each of them ,
are deduced from sphere positions through the interaction law .
newton s second law is then integrated through an explicit second order finite difference scheme to compute the new sphere positions .
a 3d software called yade ( yet another dynamic engine ) , kozicki and donz @xcite , has been enhanced in order to properly simulate partially saturated granular material features .
the contact interaction is described by an elastic - plastic relation between the force @xmath41 and the relative displacement @xmath42 of two interacting particles .
a normal stiffness @xmath43 is therefore defined to relate the normal force @xmath44 to the intergranular normal relative displacement @xmath45 : @xmath46 and a tangential stiffness @xmath47 allows us to deduce the shear force @xmath48 induced by the incremental tangential relative displacement @xmath49 ; this tangential behaviour obeying the coulomb friction law : @xmath50 where @xmath51 is the coulomb friction coefficient defined as @xmath52 , with @xmath53 the intergranular friction angle . + in the work presented here ,
@xmath43 and @xmath47 are dependent functions of the interacting particle sizes and of a characteristic modulus of the material denoted as @xmath54 : @xmath55 this definition results in a constant ratio between @xmath54 and the effective bulk modulus of the packing , whatever the size of the particles . + for simplicity , we assume that the water inside the sample is solely composed of capillary water as defined in the pendular state , with a discontinuous liquid phase . much attention has been given to these pendular liquid bridges , @xcite .
their exact shape between spherical bodies is defined by the laplace equation , relating the pressure difference @xmath56 across the liquid - gas interface to the mean curvature of the bridge and the surface tension of the liquid phase @xmath57 : @xmath58 in the cartesian coordinates of fig.[fig1](b ) the two curvature radii @xmath59 and @xmath60 ( fig.[fig1](a ) ) , are given by : @xmath61 and @xmath62 where @xmath63 defines the profile of the liquid - gas interface curve .
the @xmath64 axis coincides with the axis of symmetry of the bridge , passing through the centers of the connected spheres ( fig.[fig1](b ) ) .
according to the laplace equation , the profile of the liquid bridge is thus related to the capillary pressure @xmath65 through the following differential equation : @xmath66 the corresponding liquid bridge volume @xmath10 and intergranular distance @xmath67 can be obtained by considering the @xmath64-coordinates ( @xmath68 and @xmath69 ) of the three - phases contact lines defining the solid - liquid - gas interface , as defined by souli et al.,@xcite : @xmath70 and @xmath71 the capillary force due to the liquid bridge can be calculated at the profile apex @xmath72 according to the ` gorge method ' @xcite and consists of a contribution of the capillary pressure @xmath65 as well as the surface tension @xmath57 : @xmath73 the relation between the capillary pressure and the configuration of the capillary doublet is thus described by a system of non - linear coupled equations ( 15 , 16 , 17 , 18 ) where the local geometry ( @xmath67 ) and water volume arise as a result of the solved system @xcite .
so , to account for capillarity in the code , an interpolation scheme on a set of discrete solutions of the laplace equation has been developed in order to link directly the capillary pressure to the capillary force and water volume of the liquid bridge for a given grain - pair configuration .
this results in a suction - controlled model where , at every time - step during the simulation , capillary forces and water volumes ( @xmath74 ) are computed based upon the microstructure geometry @xmath67 and the imposed suction level @xmath65
. @xmath75 a schematic diagram of the implemented capillary law is shown in fig.[fig2 ] for a given value of the suction . with the intergranular distance @xmath67 for a given suction value : a meniscus can form for @xmath76 and breaks off for @xmath77.,width=283 ] in this paper , the choice was made to define the appearance of a bridge when particles come strictly in contact ( @xmath78 ) , neglecting the possibility of adsorbed water effects .
the capillary force is considered constant for the range of the elastic deformation ( @xmath79 ) , assuming local displacements to be very small compared to particle radii .
let us note that the formulation intrinsically defines the distance from which the meniscus breaks off as depending on the given capillary pressure and on the local geometry .
this maximum distance @xmath80 corresponds to the minimum @xmath67 value from which the laplace equation has no solution .
since this study covers the macroscopic and microscopic aspects of unsaturated granular media , stress tensors are calculated by both macro and micro - methods .
the macro - method is the conventional way used in laboratory to measure stresses in experiments , that is to say : @xmath81 where @xmath82 is the normal force acting on the boundary , and @xmath83 the surface of the boundary oriented by the normal direction @xmath84 .
this stress is equivalent to net stress @xmath85 used in unsaturated soils mechanics , with @xmath1 used as the reference pressure because the pore air pressure is effectively zero in many practical problems ( as well as in this study ) . +
as seen in section 2.2 , two other stress tensors can be considered through homogenization techniques : the intergranular stress tensor @xmath86 computed from intergranular forces , and the capillary stress tensor @xmath31 computed from capillary forces .
the studied particle assembly is a 1 mm length cubic sample composed of 10000 spheres , with a grain size distribution ranging from 0.025 mm to 0.08 mm , as shown in fig.[fig3 ] , and a porosity of 0.385 .
the input parameters are listed in table 1 , referring to equation ( 10 ) .
.dem model parameters [ cols="^,^,^,^ " , ] + fig.[fig11 ] presents the numerical simulations of three triaxial tests performed at three different confining pressures of @xmath87 , @xmath88 and @xmath89 for the simulated dry assembly having an initial void ratio equal to @xmath90 .
one can see that the results obtained with the microstructural model compared well to dem ones .
dem simulations were performed on unsaturated assemblies with an initial saturation degree of about @xmath91 , corresponding to a suction value equal to @xmath92 . in order to determine the corresponding capillary forces ,
eq.(25 ) includes two material parameters @xmath93 and @xmath94 which control the evolution of the water induced forces with the distance between two particles . according to experimental results presented in different studies @xcite
, we decided to take a value of @xmath95 and of @xmath96 .
these values give a standard evolution for the capillary forces , function of the distance between particles , as well as an initial isotropic distribution of these forces if the material structure is isotropic .
the evolution of the capillary forces with the degree of saturation requires two more parameters : @xmath97 and @xmath98 . according to previous studies
@xcite , we selected a value of @xmath99 and determined the value of @xmath100 . in order to obtain an initial value of the capillary stress in accordance with the one computed by dem .
we then performed numerical simulations of triaxial tests on wet samples for different confining pressures .
contrary to dem suction - controlled simulations , the microstructural model tests are water content controlled . in order to compare those two approaches
, samples were initially wetted at a common degree of saturation of about @xmath91 .
one should notice that , even if the test conditions are not strictly identical , the changes in the degree of saturation during loading obtained for both tests are sufficiently similar for us ( fig.[fig12 ] ) to compare the results obtained by the two approaches .
+ as presented in fig.[fig13 ] , the two models give quite similar results .
one can see that a material strength increase is obtained for unsaturated samples compared to dry ones at the same confining pressure .
the volume changes during triaxial loading create a small change in the degree of saturation ( fig.[fig12 ] ) . as a consequence ,
the capillary forces evolve , according to eq.(5 ) .
fig.[fig14 ] shows the evolution of the principal components of the capillary stress tensor during constant water content triaxial tests .
the initial state corresponds to an isotropic capillary stress tensor with a mean stress equal to @xmath101 as obtained in dem . during loading
, a structural anisotropy is created due to the evolution of the fabric tensor defined in eq.(38 ) .
therefore , the principal components of the capillary stress tensor evolve differently .
in the studied cases , this difference remains small and corresponds at the end of the test to a relative difference less than @xmath102 .
this small difference can be explained by two causes .
the first one is the small amount of induced anisotropy obtained by the evolution of the fabric tensor .
this evolution is due to the change in the branch length @xmath103 for each @xmath104 direction which , in this version of the model , is only due to elastic deformations of the grains in contact .
since all our numerical testing were performed at small confining stresses , the amount of elastic deformation remained quite small .
the second reason is linked to the fact that the capillary bridges can exist for non - touching neighboring grains .
this has been taken into account in calculating the mean capillary force and also in determining the capillary stress tensor ( eqs.(27 ) and ( 41 ) ) .
this result is in agreement with the distribution of the contacts and menisci distribution computed by dem ( figs.[fig8 ] and [ fig9 ] ) .
+ regarding the constitutive behaviour at contacts between solid particles , the results provided by the two methods are in rather good agreement concerning the influence of capillary forces at a macroscopic level .
the increase in the shear strength classically observed for partially saturated materials is clearly encountered starting from microscopical considerations , and the slight induced anisotropy of the capillary stress tensor is observed , confirming that suction effects in unsaturated materials can not be precisely accounted for by an equivalent pore pressure assumption .
starting from local capillary forces , a stress variable , denoted as the capillary stress tensor and intrinsically associated to water effects , has been defined through homogenisation techniques .
triaxial compression test simulations from two fundamentally different micromechanical models were performed on a granular assembly under several confining pressures for dry and partially saturated conditions .
both models reproduce in quite good agreement the main features of unsaturated granular materials , in particular the increase of the shear strength due to capillary effects .
+ the results also suggest that , in partially saturated materials within the pendular regime , the effects of pore fluid are adequately represented by a discrete distribution of forces rather than by an averaged pressure in the fluid phase .
effectively , as a representative quantity of the pore fluid distribution inside unsaturated materials , this suction associated stress tensor reveals that pore fluid has its own fabric which is inherently anisotropic and strongly dependent on the combined loading and hydric history of the material .
even if the induced anisotropy of the capillary stress tensor appears slight in this study , it is obvious that this tensorial nature of water in unsaturated material implies suction to produce shear effects on the solid phase .
this suction induced shear effect consequently makes it difficult to associate an isotropic quantity to water as expressed in the bishop s effective stress .
pore pressure is no longer an isotropic stress in unsaturated soil , and therefore , can not be considered as an equivalent continuum medium .
the analysis finally confirms that suction is a pore - scale concept , and that stress definitions for unsaturated soils should also include microscopic interparticle stresses as the ones resulting from capillary forces .
+ the multi - scale approach presented here appears to be a pertinent complementary tool for the study of unsaturated soil mechanics .
more precisely , discrete methods should convey new insights into the discussion about the controversial concept of generalized effective stress by relating basic physical aspects to classical phenomenological views .
nuth m , laloui l. effective stress concept in unsaturated soils : clarification and validation of a unified framework .
journal for numerical and analytical methods in geomechanics _ 2008 ; * 32*:771 - 801 .
rothenburg l , selvadurai aps .
micromechanical definition of the cauchy stress tensor for particulate media .
_ mechanics of structured media _ 1981 ; selvadurai a.p.s . editor .
amsterdam , the netherlands : elsevier .
469 - 486 souli f , cherblanc f , el youssoufi ms , saix c. influence of liquid bridges on the mechanical behaviour of polydisperse granular materials .
journal for numerical and analytical methods in geomechanics _ 2006 ; * 30*:213 - 228 .
de buhan p and dormieux l. on the validity of the effective stress concept for assessing the strength of saturated porous materials : a homogenization approach . _ journal of the mech . and phys . of solids
_ 1996 ; * 44*(10):1649 - 1667 .
dangla p , coussy o , eymard r. non - linear poroelasticity for unsaturated porous materials : an energy approach .
_ poromechanics , a tribute to m.a .
biot_. proceedings of the biot conference on poromechanics , balkema , 1998 ; 59 - 64 .
fleureau j - m , hadiwardoyo s , gomes correia a. generalised efective stress analysis of strength and small strains behaviour of a silty sand , from dry to saturated state .
_ soils and foundations _ 2003 ; * 43*(4):21 - 33 . | this paper presents a micromechanical study of unsaturated granular media in the pendular regime , based upon numerical experiments using the discrete element method , compared to a microstructural elastoplastic model .
water effects are taken into account by adding capillary menisci at contacts and their consequences in terms of force and water volume are studied .
simulations of triaxial compression tests are used to investigate both macro and micro - effects of a partial saturation .
the results provided by the two methods appear to be in good agreement , reproducing the major trends of a partially saturated granular assembly , such as the increase in the shear strength and the hardening with suction .
moreover , a capillary stress tensor is exhibited from capillary forces by using homogenisation techniques .
both macroscopic and microscopic considerations emphasize an induced anisotropy of the capillary stress tensor in relation with the pore fluid distribution inside the material . in so far as the tensorial nature of this fluid fabric implies shear effects on the solid phase associated with suction , a comparison has been made with the standard equivalent pore pressure assumption .
it is shown that water effects induce microstrural phenomena that can not be considered at the macro level , particularly when dealing with material history .
thus , the study points out that unsaturated soil stress definitions should include , besides the macroscopic stresses such as the total stress , the microscopic interparticle stresses such as the ones resulting from capillary forces , in order to interpret more precisely the implications of the pore fluid on the mechanical behaviour of granular materials .
[ published , doi : 10.1002/nag.767 ] , , , , . micromechanics ; granular materials ; unsaturated ; dem ; capillary forces ; microstructure |
after the discovery of 125-gev higgs boson with many properties consistent with the standard model ( sm ) expectations @xcite , we are left with two pieces of empirical evidence that call for new physics explanations .
one is the phenomenon of neutrino oscillations that leads to the question about how neutrino mass is generated .
the other is astronomical observation of gravitational effects caused by dark matter ( dm ) , whose abundance is bound to be about 5 times that of ordinary matter . in the sm ,
masses of quarks and massive gauge bosons are generated through yukawa and gauge couplings with the condensate of the higgs doublet field through the brout - englert - higgs ( beh ) mechanism @xcite .
one minimal extension for giving tiny mass to neutrinos is done by introducing a complex higgs triplet through the type - ii seesaw mechanism , the so - called higgs triplet model ( htm ) @xcite . in this case , the vacuum expectation value ( vev ) of the higgs triplet field is induced by electroweak symmetry breaking ( ewsb ) and controls whether the charged higgs bosons derived from the triplet field decay dominantly to leptons or weak gauge bosons .
direct searches of the doubly charged higgs boson predicted in this model at the large hadron collider ( lhc ) generally put a lower mass bound at about 400 gev @xcite . in the past few years
, many experiments have reported indirect evidence of dm , such as the excess of positron fraction observed by pamela @xcite , fermi - lat @xcite and ams-02 @xcite , the excess of positron+electron flux observed by atic @xcite , hess @xcite , fermi - lat @xcite , pamela @xcite and ams-02 @xcite , the excess of gamma - ray spectrum at the galactic center @xcite , and so on .
the ams collaboration also confirms with an unprecedented precision the excess of positron fraction in the energy range of @xmath2 $ ] gev @xcite , the positron+electron flux from 0.5 gev to 1 tev @xcite , and a deviation of the antiproton fraction from secondary astrophysical sources of cosmic ray collisions for the antiproton kinetic energy between @xmath3 gev @xcite .
although still uncertain whether the observed antiproton spectrum is still consistent with the background of secondary antiprotons @xcite , several studies @xcite have attempted to explain the possible excess of antiproton using dm annihilations and/or decays .
it is well - known that the excess of both electron and position fluxes require additional contributions to the thermally averaged dm annihilation cross section @xcite than is required by the antiproton fraction spectrum . in this work ,
we propose a weakly - interacting massive particle ( wimp ) dm model that readily accommodates the two sets of data . in this work , we extend the htm by having an additional hidden @xmath0 symmetry that is broken by a scalar quadruplet down to a residual @xmath1 symmetry _ a la _ krauss - wilczek mechanism @xcite , making the associated gauge bosons our dm candidates .
the quadruplet also has interactions with both the sm doublet and the higgs triplet and plays the role of mediator that connects the hidden sector and the visible sector of sm particles and exotic higgs bosons .
a model also invoking a scalar triplet field but using a singlet scalar as a cold dm candidate was proposed in ref . @xcite . as a result
, the dm can annihilate into a pair of exotic higgs bosons that in turn decay into leptons or weak gauge bosons as alluded to before .
the leptonic channel will lead to production of electrons and positrons , and the gauge channel to protons and antiprotons after hadronization . with an appropriate choice of the triplet vev
, the model can explain simultaneously the observed spectra of positron and antiproton fluxes .
it should be emphasized that the choice of @xmath0 gauge group is just to provide an explicit model with stable dm candidates .
other dm models from a hidden sector that couples mainly to the triplet higgs fields should catch the same features as well .
in this model , we extend the sm gauge group by an additional @xmath0 symmetry with the associated gauge field denoted by @xmath4 .
this symmetry is broken by an @xmath0 quadruplet field @xmath5 that does not carry sm gauge charges , where the subscript stands for the eigenvalue of @xmath6 , the third @xmath0 generator , of the field and @xmath7 .
finally , we also introduce a complex higgs field @xmath8 that is a triplet under the sm @xmath9 and carries hypercharge @xmath10 . here
we adopt the convention that the electric charge @xmath11 .
more explicitly , the complete lagrangian invariant under the @xmath12 gauge group is @xmath13 \non \\ & + \left ( d_\mu \phi_4\right)^\dagger d^\mu \phi_4 - v(\phi,\delta , \phi_4 ) - \frac{1}{4 } x^a_{\mu\nu } x^{a \mu \nu } ~ , \label{eq : lang}\end{aligned}\ ] ] where @xmath14 is the lagrangian of sm , @xmath15 denotes the left - handed lepton doublet field , @xmath16 is the charge conjugation , @xmath17 is the yukawa coupling matrix , @xmath18 is a higgs doublet , and the field strength tensor @xmath19 .
the most general higgs potential @xmath20 where , as in the sm , @xmath21 is required for the ewsb , @xmath22 is the sm higgs doublet field , and the scalar triplet of @xmath9 is written in an @xmath23 covariant form as @xmath24 where the triplet vev @xmath25 is induced by the doublet vev through the @xmath26 term and is constrained by the electroweak rho parameter to be less than about 1 gev . in fact , to produce the right positron and antiproton spectra given by the ams-02 experiment , @xmath25 is required to be @xmath27(10 - 100 ) kev . with @xmath28 assumed to be of @xmath27(tev ) , @xmath26 is also about the same scale as @xmath25 . to break the @xmath0 symmetry while preserving a discrete symmetry to stabilise dm candidates , we require @xmath29 so that the @xmath30 field spontaneously develops a vev in the @xmath31 component : @xcite @xmath32 where @xmath33 is assumed to be @xmath34 tev . as a result , the @xmath0 gauge bosons @xmath35 and
@xmath36 acquire their masses , @xmath37 and @xmath38 , at the tev scale after absorbing the @xmath39 and @xmath40 as their longitudinal components .
moreover , there is a residual @xmath1 symmetry in the model , under which @xmath41 and @xmath42 carry nonzero charges and serve as good candidates of dm . finally , the physical states in the hidden sector and the quadruplet are @xmath41 , @xmath42 , @xmath36 and @xmath43 , where @xmath43 plays the role of messenger between the hidden sector and the visible sector .
we now work out the relevant couplings for our analysis . from the kinetic term of @xmath30 and the breaking pattern of eq .
( [ eq : vev ] ) , the gauge interaction between @xmath43 and @xmath41 is given by @xcite @xmath44 where , as mentioned above , @xmath45(tev ) .
in general , the neutral components in @xmath46 , @xmath8 , and @xmath30 can mix under the mass matrix @xmath47 in the basis of @xmath48 , where @xmath49 , @xmath50 , and @xmath51 . as mentioned above ,
@xmath26 and @xmath25 are much smaller than the other mass parameters in eq .
( [ eq : mass ] ) . therefore , only @xmath52 and @xmath43 mix to render the mass eigenstates @xmath53 where @xmath54 denotes the sm - like higgs boson of mass 125 gev , @xmath46 is the other physical higgs boson with a mass at the tev scale , and the mixing angle @xmath55 given by @xmath56 is taken to be small in view of the hierarchy between @xmath57 and @xmath33 and by assuming small @xmath58 .
in addition , we require @xmath59 and @xmath60 to be sufficiently small so that the mass of @xmath61 is smaller than @xmath62 . finally , we obtain the following approximate formulas for the masses of physical scalar bosons : @xmath63 in the limit of vanishing @xmath58 , @xmath64 does not couple with sm particles directly . now the only important parameter that controls the phenomenology of dm interactions with the visible sector is @xmath65 . explicitly , the @xmath65 interaction term gives @xmath66 ~. \label{eq : hdd}\end{aligned}\ ] ] in this work , we consider the scenario where the dm particles annihilate through the @xmath67-@xmath68-@xmath46 and @xmath46-@xmath8-@xmath69 interactions given in eqs .
( [ eq : igg ] ) and ( [ eq : hdd ] ) , respectively , to a pair of higgs triplet bosons . to facilitate this pair annihilation process
, we employ the breit - wigner enhancement through the relation @xmath70 , which affects both relic density @xcite and positoron / antiproton fluxes @xcite .
subsequently , we have @xmath71 and @xmath72 dominantly and some small branching fractions of @xmath73 for @xmath74 , @xmath75 and @xmath76 , with details depending on the values of @xmath25 and the yukawa couplings @xmath77 . as a result
, the model can have simultaneous productions of @xmath78 and @xmath79 , with some of the latter hadronizing into antiprotons , at significantly different rates .
now the major free parameters in our analysis are @xmath80 , @xmath65 , @xmath62 , @xmath81 , @xmath82 , @xmath25 , and @xmath77 . in the dm annihilation amplitude ,
the gauge coupling @xmath80 and the quadruplet vev @xmath33 always appear in the combination of product and can be rewritten in terms of @xmath62 . for the mass of @xmath46 ,
we write @xmath83 and use @xmath84 as one positive free parameter of @xmath85 or smaller . a breit - wigner enhancement
is then obtained in the current dm annihilation rate because in the non - relativistic limit the annihilation cross section is written by @xmath86^{-1 } \sqrt{1 - \frac{m_\delta^2}{m_\chi^2 } } ~ , \label{eq : sigmav}\end{aligned}\ ] ] where @xmath87 denotes the total width of @xmath46 and is found to be much smaller than @xmath62 .
the average speed of dm @xmath57 in units of the speed of light is typically @xmath88 at the current universe and @xmath89 at the freeze - out . therefore , the dominant parameters in eq .
( [ eq : sigmav ] ) are @xmath65 , @xmath62 and @xmath84 .
the yukawa couplings can be fixed by fitting to the neutrino mass measurements @xcite and assuming the normal hierarchy as it is preferred by the positron flux excess @xcite .
moreover , we have taken the cp - violating phase , the majorana phases , and the mass of the lightest neutrino to be zero .
the conclusion will not change much if , for example , we take a tiny nonzero mass for the lightest neutrino .
both lepton flavour - conserving and -violating channels have been included in the decays of the triplet higgs bosons .
our choice of @xmath82 at the tev scale exempts us from the constraint of the lower bound of about 400 gev on the doubly - charged higgs boson mass through the searches of like - sign dilepton channels @xcite .
otherwise , the results of our numerical analysis are not sensitive to @xmath82 as long as it is smaller than @xmath62 . in the following analysis
, we take @xmath90 gev as a reference value . in the end , we are left with four independent free parameters : @xmath65 , @xmath62 , @xmath84 , and @xmath25 .
we start by investigating the dm relic abundance @xmath91 given by the planck collaboration at 90% confidence level ( cl ) @xcite . in our setup , the dm annihilation cross section @xmath92 is given by eq .
( [ eq : sigmav ] ) .
it is noted that the dm can also annihilate into a pair of sm particles ( @xmath93 , @xmath94 , @xmath95 , etc ) through the @xmath54-@xmath46 mixing , which is suppressed as we assume a small mixing angle @xmath55 .
our numerical analysis is done by utilizing the micromegas 4.1.5 package @xcite implemented with the model to solve the boltzmann equation for the dm relic density . in the lower horizontal axis , as required by the observed relic density .
the upper horizontal axis gives the corresponding value of @xmath65.,width=340 ] fig .
[ fig : xs ] shows the thermally averaged cross section of dm annihilation to all triplet higgs bosons , @xmath96 , at the present time for producing the right relic density .
it is shown as a function of @xmath84 defined in eq .
( [ eq : eps ] ) .
the corresponding values of @xmath65 are indicated in the upper horizontal axis as well .
although this curve does not have a sensitive dependence on the dm mass , we will fix @xmath97 tev because of better fits to the energy spectra of positron flux and antiproton flux ratio from the cosmic rays . as seen in the plot
, the cross section can reach @xmath98 @xmath99/s for @xmath100 because of the breit - wigner enhancement . in this case , the coupling @xmath101 . as shown below ,
such a cross section is desirable for rendering the correct excess of the positron flux .
again , using the micromegas 4.1.5 package @xcite , we compute the positron and antiproton fluxes resulting from the decays of the charged higgs bosons in the dm annihilation final states . for the dark matter density profile ,
we take the nfw model with a local halo density of @xmath102 gev/@xmath99 , a core radius of 20 kpc , and the distance from our solar system to the galactic center as @xmath103 kpc @xcite . for charged particle propagation through the space
, we consider the three schemes min , med , and max defined in ref .
@xcite to have the minimum , medium , and maximum charged particle flux , respectively .
the background for positron flux is provided by a fitting function given in refs .
the background for antiproton flux is estimated by combining ams proton flux data @xcite and the flux ratio @xmath104 estimated in ref .
@xcite . in the left plot and @xmath25 in the right plot .
the background in the left plot is given by the purple dotted curve , and those in the right plot are given by the thin curves.,title="fig:",width=302 ] in the left plot and @xmath25 in the right plot .
the background in the left plot is given by the purple dotted curve , and those in the right plot are given by the thin curves.,title="fig:",width=302 ] in fig .
[ fig : fluxes ] , the left plot shows the positron flux as a function of the positron energy .
the right plot shows the ratio of antiproton flux to proton flux as a function of the antiproton kinetic energy . in both plots ,
we fix @xmath97 tev . qualitatively ,
changing the dm mass will shift the peaks of signals horizontally .
the magnitudes of signals are mainly controlled by @xmath84 for the positron flux and by @xmath25 for the antiproton flux .
we take @xmath105 and @xmath106 kev , corresponding to @xmath107 , for the min scheme ( red solid curves ) ; @xmath108 and @xmath109 kev , corresponding to @xmath110 , for the med scheme ( green dashed curves ) ; and @xmath111 and @xmath112 kev , corresponding to @xmath113 , for the max scheme ( blue dash - dotted curves ) . here
we make some comments regarding the constraint on dm annihilation cross section from the @xmath114-ray flux provided from inclusive photon spectrum measurement @xcite and @xmath114-ray observation of dwarf galaxies @xcite by fermi - lat .
although dm annihilation to four - body final states , as in our case , are generally less constrained than two - body final states , our analysis assuming the normal hierarchy of neutrino masses has @xmath115 decaying to energetic @xmath116 and @xmath117 at a branching fraction of around 30% for both modes .
such @xmath118 leptons in the final states may produce @xmath114-rays subject to the fermi - lat constraint .
a detailed analysis with this taken into account will be presented in another work @xcite . finally , we note in passing that the scattering cross section between the dm and the nucleons is small because the interactions between the dm and the sm particles is suppressed by the mixing angle @xmath55
. therefore , the model can readily evade the constraints from current direct searches .
in the current work , we have constructed a model based on the higgs triplet model and an additional @xmath0 gauge symmetry that links neutrino mass and dark matter physics . with an @xmath0 quadruplet that develops a vacuum expectation value in its @xmath31 component ,
the @xmath0 symmetry is broken down to a discrete @xmath1 symmetry , thereby stabilising the @xmath119 components of the gauge field as the dark matter candidates .
we have then analyzed the dark matter phenomenology of the model particularly in view of the recent ams-02 data . through the physical scalar boson of the quadruplet as the messenger
, the dark matter particles can annihilate primarily into a pair of triplet higgs bosons , provided that the mixing between the quadruplet scalar particle and the sm - like higgs boson is sufficiently small .
the correct dark matter relic abundance can be obtained when the mass of messenger is about twice that of the dark matter .
the charged triplet higgs bosons in turn decay dominantly into leptonic final states to produce an excess of positrons .
however , there is a small branching fraction for the charged triplet higgs bosons to decay into weak gauge bosons , part of which eventually hadronize into antiprotons . the decay pattern of the charged triplet higgs bosons is largely fixed by assuming the normal hierarchy and mass measurements of neutrinos .
we have studied three charged particle propagation schemes and found the corresponding parameters of the model that fit well with the positron flux and antiproton flux ratio observed by the ams collaboration .
finally , we remark that the gauged @xmath0 for stable dark matter candidates in this study is only one choice .
any dm model from a hidden sector that couples mainly to the triplet higgs fields should have the same features as well .
the authors would like to thank yuan - hann chang for useful discussions about the ams-02 data .
this work was supported in part by the ministry of science and technology of taiwan under grant nos .
most-103 - 2112-m-006 - 004-my3 ( chc ) , most-100 - 2628-m-008 - 003-my4 ( cwc ) and most-103 - 2811-m-006 - 030 ( tn ) , as well as the world premier international research center initiative , ministry of education , culture , sports , science and technology , japan .
we also thank the national center for theoretical sciences for the support of useful facilities .
g. aad _ et al . _ [ atlas collaboration ] , phys .
b * 716 * , 1 ( 2012 ) [ arxiv:1207.7214 [ hep - ex ] ]
. s. chatrchyan _ et al . _
[ cms collaboration ] , phys .
b * 716 * , 30 ( 2012 ) [ arxiv:1207.7235 [ hep - ex ] ] . f. englert and r. brout , phys .
* 13 * , 321 ( 1964 ) .
p. w. higgs , phys .
* 13 * ( 1964 ) 508 .
m. magg and c. wetterich , phys .
b * 94 * , 61 ( 1980 ) ; g. lazarides , q. shafi and c. wetterich , nucl .
b * 181 * , 287 ( 1981 ) ; r. n. mohapatra and g. senjanovic , phys .
d * 23 * , 165 ( 1981 ) ; e. ma and u. sarkar , phys . rev .
lett . * 80 * , 5716 ( 1998 ) [ hep - ph/9802445 ] .
w. konetschny and w. kummer , phys .
b * 70 * , 433 ( 1977 ) ; j. schechter and j. w. f. valle , phys . rev .
d * 22 * , 2227 ( 1980 ) ; t. p. cheng and l. -f .
li , phys .
d * 22 * , 2860 ( 1980 ) ; s. m. bilenky , j. hosek and s. t. petcov , phys .
b * 94 * , 495 ( 1980 ) .
a. g. akeroyd , c. w. chiang and n. gaur , jhep * 1011 * , 005 ( 2010 ) [ arxiv:1009.2780 [ hep - ph ] ] . c. w. chiang , t. nomura and k. tsumura , phys .
d * 85 * , 095023 ( 2012 ) [ arxiv:1202.2014 [ hep - ph ] ] .
s. chatrchyan _ et al . _
[ cms collaboration ] , eur .
j. c * 72 * , 2189 ( 2012 ) [ arxiv:1207.2666 [ hep - ex ] ] .
g. aad _ et al . _ [ atlas collaboration ] , eur .
j. c * 72 * , 2244 ( 2012 ) [ arxiv:1210.5070 [ hep - ex ] ] .
o. adriani _ et al . _
[ pamela collaboration ] , nature * 458 * , 607 ( 2009 ) [ arxiv:0810.4995 [ astro - ph ] ] .
m. ackermann _ et al .
_ [ fermi lat collaboration ] , phys . rev .
lett . * 108 * , 011103 ( 2012 ) [ arxiv:1109.0521 [ astro-ph.he ] ] .
m. aguilar _ et al .
_ [ ams collaboration ] , phys .
* 110 * , no .
14 , 141102 ( 2013 ) .
j. chang , j. h. adams , h. s. ahn , g. l. bashindzhagyan , m. christl , o. ganel , t. g. guzik and j. isbert _ et al .
_ , nature * 456 * , 362 ( 2008 )
. f. aharonian _ et al .
_ [ h.e.s.s .
collaboration ] , phys .
* 101 * , 261104 ( 2008 ) [ arxiv:0811.3894 [ astro - ph ] ] .
f. aharonian _ et al .
_ [ h.e.s.s .
collaboration ] , astron .
astrophys . * 508 * , 561 ( 2009 ) [ arxiv:0905.0105 [ astro-ph.he ] ] . m. ackermann _ et al .
_ [ fermi lat collaboration ] , phys . rev .
d * 82 * , 092004 ( 2010 ) [ arxiv:1008.3999 [ astro-ph.he ] ] .
o. adriani _ et al . _
[ pamela collaboration ] , phys .
lett . * 106 * , 201101 ( 2011 ) [ arxiv:1103.2880 [ astro-ph.he ] ] .
t. daylan , d. p. finkbeiner , d. hooper , t. linden , s. k. n. portillo , n. l. rodd and t. r. slatyer , arxiv:1402.6703 [ astro-ph.he ] .
b. zhou , y. f. liang , x. huang , x. li , y. z. fan , l. feng and j. chang , arxiv:1406.6948 [ astro-ph.he ] .
f. calore , i. cholis and c. weniger , arxiv:1409.0042 [ astro-ph.co ] . k. n. abazajian , n. canac , s. horiuchi , m. kaplinghat and a. kwa , arxiv:1410.6168 [ astro-ph.he ] . f. calore , i. cholis , c. mccabe and c. weniger , arxiv:1411.4647 [ hep - ph ] . l. accardo _ et al . _ [ ams collaboration ] , phys .
lett . * 113 * , 121101 ( 2014 ) . m.
aguilar _ et al .
_ [ ams collaboration ] , phys .
* 113 * , 221102 ( 2014 ) .
ams-02 collaboration , talks given at the _ ams days at cern _ , april 15 - 17 , 2015 .
g. giesen , m. boudaud , y. genolini , v. poulin , m. cirelli , p. salati , p. d. serpico and j. feng _ et al .
arxiv:1504.04276 [ astro-ph.he ] . h. b. jin , y. l. wu and y. f. zhou , arxiv:1504.04604 [ hep - ph ] .
m. ibe , s. matsumoto , s. shirai and t. t. yanagida , arxiv:1504.05554 [ hep - ph ] .
k. hamaguchi , t. moroi and k. nakayama , arxiv:1504.05937 [ hep - ph ] .
s. j. lin , x. j. bi , p. f. yin and z. h. yu , arxiv:1504.07230 [ hep - ph ] . c. h. chen and t. nomura , jhep * 1409 * , 120 ( 2014 ) [ arxiv:1404.2996 [ hep - ph ] ] . k. belotsky , m. khlopov , c. kouvaris and m. laletin , adv . high energy phys .
* 2014 * , 214258 ( 2014 ) [ arxiv:1403.1212 [ astro-ph.co ] ] .
t. hambye , jhep * 0901 * , 028 ( 2009 ) [ arxiv:0811.0172 [ hep - ph ] ] . c. w. chiang , t. nomura and j. tandean , jhep * 1401 * , 183 ( 2014 ) [ arxiv:1306.0882 [ hep - ph ] ] .
v. v. khoze and g. ro , jhep * 1410 * , 61 ( 2014 ) [ arxiv:1406.2291 [ hep - ph ] ] .
s. baek , p. ko and w. i. park , jcap * 1410 * , 067 ( 2014 ) [ arxiv:1311.1035 [ hep - ph ] ] .
c. h. chen and t. nomura , arxiv:1501.07413 [ hep - ph ] .
s. b. dev , d. k. ghosh , n. okada and i. saha , phys .
d * 89 * , 095001 ( 2014 ) [ arxiv:1307.6204 [ hep - ph ] ] .
p. gondolo and g. gelmini , nucl .
b * 360 * , 145 ( 1991 ) .
d. feldman , z. liu and p. nath , phys .
d * 79 * , 063509 ( 2009 ) [ arxiv:0810.5762 [ hep - ph ] ] .
m. ibe , h. murayama and t. t. yanagida , phys .
d * 79 * , 095009 ( 2009 ) [ arxiv:0812.0072 [ hep - ph ] ] .
k. a. olive _ et al .
_ [ particle data group collaboration ] , chin . phys .
c * 38 * , 090001 ( 2014 ) .
p. a. r. ade _ et al .
_ [ planck collaboration ] , arxiv:1502.01589 [ astro-ph.co ] . g. belanger , f. boudjema , a. pukhov and a. semenov , arxiv:1407.6129 [ hep - ph ]
. j. f. navarro , c. s. frenk and s. d. m. white , astrophys . j. * 490 * , 493 ( 1997 ) [ astro - ph/9611107 ] . f. donato , n. fornengo , d. maurin and p. salati , phys . rev .
d * 69 * , 063501 ( 2004 ) [ astro - ph/0306207 ] .
e. a. baltz and j. edsjo , phys .
d * 59 * , 023511 ( 1998 ) [ astro - ph/9808243 ] .
e. a. baltz , j. edsjo , k. freese and p. gondolo , phys .
d * 65 * , 063511 ( 2002 ) [ astro - ph/0109318 ] .
m. aguilar , phys .
* 114 * , no .
17 , 171103 ( 2015 ) .
m. ackermann _ et al .
_ [ fermi - lat collaboration ] , phys .
d * 86 * , 022002 ( 2012 ) [ arxiv:1205.2739 [ astro-ph.he ] ] .
m. ackermann _ et al .
_ [ fermi - lat collaboration ] , phys . rev .
d * 89 * , 042001 ( 2014 ) [ arxiv:1310.0828 [ astro-ph.he ] ] . c. h. chen ,
c. w. chiang and t. nomura , work in progress . | we propose a dark matter explanation to simultaneously account for the excess of antiproton - to - proton and positron power spectra observed in the ams-02 experiment while having the right dark matter relic abundance and satisfying the current direct search bounds .
we extend the higgs triplet model with a hidden gauge symmetry of @xmath0 that is broken to @xmath1 by a quadruplet scalar field , rendering the associated gauge bosons stable weakly - interacting massive particle dark matter candidates . by coupling the complex higgs triplet and the @xmath0 quadruplet , the dark matter candidates can annihilate into triplet higgs bosons each of which in turn decays into lepton or gauge boson final states . such a mechanism gives rise to correct excess of positrons and antiprotons with an appropriate choice of the triplet vacuum expectation value .
besides , the model provides a link between neutrino mass and dark matter phenomenology .
ipmu15 - 0060 |
let @xmath0 be a real - valued function on a closed bounded rectangular domain @xmath4 .
fix some @xmath5 .
let @xmath6 be a probability measure on @xmath7 and absolutely continuous with respect to the lebesgue measure .
for simplicity we will always assume that @xmath8 and @xmath9 . in this paper
we address the following question : what is the necessary and sufficient condition on @xmath2 , real numbers @xmath10 and a measure @xmath6 such that the following inequality holds @xmath11 for all borel measurable @xmath12 .
we assume that @xmath13 and @xmath14 are strictly positive numbers .
essential supremum in ( [ ragaca ] ) is taken with respect to the lebesgue measure .
our main result is the following theorem .
[ mth1 ] suppose that @xmath15 and @xmath16 never vanish in @xmath17 . for inequality ( [ ragaca ] ) to hold it is necessary that @xmath18 moreover , if @xmath19 is a gaussian measure then the above conditions are also sufficient . by gaussian measure we mean a probability measure of the form @xmath20 the symbols @xmath21 and @xmath22 denote partial derivatives .
@xmath23 denotes transpose of the row vector @xmath24 .
constraint @xmath25 on numbers @xmath26 in ( [ numbers ] ) can be rewritten as @xmath27 and @xmath28 . moreover , if @xmath29 then it is necessary that @xmath30 is the zero vector . in the applications usually @xmath31 .
therefore the most important condition the reader needs to keep in mind is the partial differential inequality ( pdi ) in ( [ numbers ] ) .
we should also notice an _ independence from the dimension _
, i.e. , the necessity conditions follow from the one dimensional case @xmath32 , and for the gaussian measures ( [ numbers ] ) is sufficient for ( [ ragaca ] ) to hold for all @xmath33 . in section [ proof1 ]
we present the proof of theorem [ mth1 ] .
we should mention that ( [ numbers ] ) already appeared in @xcite as a sufficient condition for inequality ( [ ragaca ] ) to hold in case of the gaussian measure with supremum in ( [ ragaca ] ) and smooth compactly supported functions @xmath34 .
namely , it was proved in @xcite that if @xmath35 are nonvanishing and @xmath2 satisfies ( [ numbers ] ) then the following inequality holds @xmath36 for all smooth compactly supported functions @xmath34 and the gaussian measure @xmath6 .
in fact , in @xcite inequality ( [ weaker ] ) was obtained in more general setting where one can allow @xmath2 to have arbitrary number of variables and supremum in ( [ weaker ] ) is taken over an affine subspace of @xmath7 . by slightly extending the class of test functions in ( [ weaker ] ) ( which can be done by using the methods in @xcite ) one can derive ( [ ragaca ] ) from ( [ weaker ] ) by using some subtle approximation arguments , the fact that @xmath2 is monotone in each variable and elaborating an argument used in appendix of @xcite . in this paper
we still prefer to directly work with ( [ ragaca ] ) .
our first immediate extension is that we have ( [ ragaca ] ) with essential supremum and borel measurable functions .
our second extension is that we obtain if and only if characterization , moreover we obtain the necessity part for almost arbitrary measures .
our approach to ( [ ragaca ] ) sheds light to a question about _ optimizers _ , and it provides us with some quantitative version of ( [ ragaca ] ) ( see lemma [ lem-5 ] and lemma [ lem-7 ] which are interesting in itself ) , and it shows a link between two different type of pdes considered in @xcite ( see pde ( 1.3 ) and ( 1.5 ) in @xcite ) .
our proof is completely different from the methods developed in @xcite . in particular , we obtain the new proof of the ehrhard inequality .
the main idea of the current paper is based on two arguments : one of brascamp
lieb @xcite rewriting essential supremum as the limit of @xmath37 norms , and the other one is a subtle inequality which was studied and developed only recently in @xcite . in section [ degenerate ]
we give answers to the questions : what happens in theorem [ mth1 ] when @xmath15 and @xmath16 vanish in @xmath17 .
it turns out that in this degenerate case the problem depends very much on the class of test functions @xmath38 , namely , it matters whether @xmath38 are continuous or discontinuous functions . in section [ app ]
we illustrate various applications of the theorem .
the first immediate application is the ehrhard inequality .
this corresponds to the function @xmath39 which does satisfy ( [ numbers ] ) where @xmath40 is the gaussian distribution function . in theorem [ theorem2 ] and
corollary [ last3 ] we find necessary conditions on the probability measures @xmath6 which satisfy the functional version of the ehrhard inequality in dimension one where the gaussian distribution @xmath40 is replaced by the distribution of @xmath6 , i.e. , @xmath41)$ ] . as a corollary ( see corollary [ even - measure ] ) we show that in the class of even probability measures the gaussian measure is the only one which satisfies the functional form of the ehrhard inequality on the real line .
another immediate application is the prkopa leindler inequality .
this corresponds to the function @xmath42 with @xmath31 which also does satisfy ( [ numbers ] ) . in subsection [ lebeg - hom ]
we describe @xmath43-homogeneous functions @xmath2 , i.e. , @xmath44 for all @xmath45 , and measures @xmath6 which satisfy ( [ ragaca ] ) .
it turns out that in this particular class of functions the description is simple : @xmath2 is a convex function , or @xmath2 is one of the counterparts of the prkopa leindler functions : @xmath46 with @xmath31 , @xmath47 with @xmath48 , and @xmath49 with @xmath50 .
the last two functions provide us with a natural extension of the prkopa leindler inequality for any real power @xmath51 ( see corollary [ character ] ) . in the case @xmath46 as it is well known the only possible measures which satisfy ( [ ragaca ] ) must be log - concave measures . on the other hand appearance of the convex functions
@xmath2 makes it clear that some of the inequalities of the form ( [ ragaca ] ) are trivial consequences of the jensen s inequality ( see proposition [ jensen2 ] and remark [ trivial ] ) . in subsection [ universality - prekopa ]
we characterize all smooth concave functions @xmath2 which satisfy ( [ ragaca ] ) for a probability measure @xmath6 .
it turns out that @xmath2 must satisfy homogeneous monge ampre equation , and as a consequence of this , it generically must be the prkopa
leindler function @xmath52 with @xmath31 , or one of its counterpart ( see proposition [ conc - sol ] ) .
thus we observe some universality of the prkopa leindler inequality , as a unique reverse to jensen s inequality . section [ other ] illustrates other examples of @xmath2 such that ( [ ragaca ] ) holds for the gaussian measure ( see corollary [ ee-1 ] , [ ee-2 ] and [ ee-3 ] ) . in section
[ come - back ] we explain how theorem [ mth1 ] allows to restore the ehrhard inequality by solving pde problem over an obstacle .
i am grateful to christos saroglou who initiated this project and with whom i had many discussions .
he should be considered as co - author ( despite his insistence to the contrary ) .
i am extremely thankful to the kent state analysis group especially fedor nazarov who gave me some valuable suggestions in obtaining the necessity part , and artem zvavitch for providing c. borell s lecture notes .
the talk given by grigoris paouris on the informal analysis seminar at kent state university served as a guide and inspiration for the present article .
first we notice that if ( [ ragaca ] ) holds for some @xmath33 then it holds for @xmath32
. indeed we can test ( [ ragaca ] ) on the functions @xmath53 , @xmath54 for some measurable @xmath55 . in what follows
we will assume that @xmath32 .
finiteness of the fifth moment together with the lebesgue dominated convergence theorem implies that @xmath56 we need several technical lemmas .
we fix a number @xmath57 close to @xmath58 which will be determined later .
[ decay ] @xmath59 as @xmath60 .
we have @xmath61 similarly for the second integral .
let @xmath62 be the point in the interior of @xmath17 .
let @xmath63 and @xmath64 .
[ first - form ] if @xmath2 satisfies ( [ ragaca ] ) then @xmath65 for all real numbers @xmath66 and @xmath67 such that @xmath68 .
let @xmath69 be a number which will be determined later .
we consider the following test functions @xmath38 : @xmath70 we notice that @xmath71 for some fixed constant @xmath72 . since @xmath73 it is clear that if @xmath74 is sufficiently small then @xmath75 for all @xmath76 where @xmath77 is a small fixed number .
ideally we want to choose @xmath78 for all @xmath79 but then the image of @xmath38 will escape from the rectangle @xmath17 . on the other hand ,
unfortunately , we can not just make it compactly supported ( i.e. , multiply by some cutoff function ) as we will see from the computations given below .
let @xmath80 and @xmath81 .
choose @xmath82 so that @xmath83 .
notice that for each fixed @xmath69 by ( [ tends - to - zero ] ) and lemma [ decay ] we have @xmath84 by taylor s formula we obtain @xmath85 on the other hand we have @xmath86 since @xmath87 we have @xmath88 .
first we should compare small order terms in order to get a restriction on @xmath89 .
since @xmath34 are continuous clearly the essential supremum of the integrand in ( [ ragaca ] ) becomes @xmath90 .
thus , introducing new variables @xmath91 and using the fact that supremum of the sum is at most the sum of the supremums , we obtain that ( [ ragaca ] ) implies the inequality @xmath92 notice that @xmath93 for @xmath94 and it is bounded as @xmath95 otherwise .
therefore ( [ miss1 ] ) implies that @xmath96 . on the other hand we can considered the new test functions @xmath97 and
we can obtain the opposite inequality @xmath98 .
this implies that if @xmath29 then @xmath99 .
thus we have obtained the constraint on @xmath100 .
further we should assume that @xmath99 because in the remaining case @xmath31 the first order terms still cancel and the proof proceeds in the same way .
next we should compare the higher order terms .
( [ ragaca ] ) implies the inequality @xmath101d\mu(t ) \geq \nonumber \\ & \varepsilon^{2 } \beta ( pa^{2 } + qb^{2 } ) + o(\varepsilon^{2 } ) \label{final - inequality}\end{aligned}\ ] ] since @xmath102 , ( [ tends - to - zero ] ) and lemma [ decay ] we obtain @xmath103 set @xmath104.\end{aligned}\ ] ] we remind that @xmath66 and @xmath67 are chosen in such a way that @xmath105 .
we need the following lemma .
[ second - form ] let @xmath69 be such that for all @xmath106 , @xmath107 we have @xmath108 then there exist sufficiently small positive constants @xmath109 and @xmath110 such that @xmath111 for all real @xmath112 and all @xmath113 .
before we proceed to the proof of the lemma , let us mention that lemma [ first - form ] follows from lemma [ second - form ] .
indeed , first we choose @xmath69 such that ( [ about - delta ] ) holds .
such choice is possible because of the continuity and the assumption on the numbers @xmath66 and @xmath67 .
lemma [ second - form ] , ( [ final - inequality ] ) , ( [ tends - to - zero ] ) , lemma [ decay ] and the fact that @xmath114 on the complement of the interval @xmath115 $ ] imply ( [ main - form ] ) . thus it remains to prove lemm [ second - form ] .
set @xmath116 we should describe behavior of @xmath117 on the red line @xmath118 ( see figure [ fig : dom ] ) .
if @xmath119 then the line @xmath118 will cross the sides @xmath120 and @xmath121 of the rectangle @xmath122 as it is shown on figure [ fig : dom ] . ]
we have @xmath123 & x \leq -\delta \varepsilon^{-\alpha};\\ \left ( \frac{h_{uu}}{h_{u}^{2 } } + p\right)x^{2}+2\frac{h_{uv}}{h_{u}h_{v}}x \varepsilon^{-\alpha } + \left(\frac{h_{vv}}{h_{v}^{2}}+q \right ) \varepsilon^{-2\alpha } & t - \varepsilon^{-\alpha } \geq x \geq -\delta \varepsilon^{-\alpha};\\ \left ( \frac{h_{uu}}{h_{u}^{2 } } + p\right)x^{2}+2\frac{h_{uv}}{h_{u}h_{v}}x ( t - x ) + \left(\frac{h_{vv}}{h_{v}^{2}}+q \right)(t - x)^{2 } & t - \varepsilon^{-\alpha } \leq x \leq \varepsilon^{-\alpha};\\ \left ( \frac{h_{uu}}{h_{u}^{2 } } + p\right)\varepsilon^{-2\alpha}+2\frac{h_{uv}}{h_{u}h_{v } } \varepsilon^{-\alpha}(t - x ) + \left(\frac{h_{vv}}{h_{v}^{2}}+q
\right ) ( t - x)^{2 } & \varepsilon^{-\alpha } \leq x \leq t+\delta \varepsilon^{-\alpha};\\ \varepsilon^{-2\alpha } \left [ \left ( \frac{h_{uu}}{h_{u}^{2 } } + p\right)-2\frac{h_{uv}}{h_{u}h_{v}}\delta + \left(\frac{h_{vv}}{h_{v}^{2}}+q \right ) \delta^{2}\right ] & x \geq
t+\delta \varepsilon^{-\alpha}. \end{cases}\end{aligned}\ ] ] notice that because of the assumption ( [ about - delta ] ) the values of @xmath124 approach to negative infinity as @xmath125 if @xmath126 or @xmath127 where @xmath128 is some constant .
if @xmath129 then let us reparametrize the function @xmath124 as follows @xmath130 where @xmath131
. then @xmath132.\end{aligned}\ ] ] clearly if @xmath133 then @xmath134 and by ( [ about - delta ] ) the maximal value of @xmath124 behaves as @xmath135 for some @xmath128 on the interval @xmath136
. behavior of @xmath124 on the interval @xmath137 is completely symmetric to the previous case .
one can check that the maximum of the function @xmath124 on the remaining interval @xmath138 $ ] is attained at the point @xmath139 it is clear that if @xmath112 for some sufficiently small number @xmath109 then @xmath140 $ ] .
finally we have @xmath141 } w(x , t - x)= \frac{p\frac{h_{vv}}{h_{v}^{2}}+pq + q \frac{h_{uu}}{h_{u}^{2}}+\frac{h_{uu}h_{vv}-h_{uv}^{2}}{h_{u}^{2}h_{v}^{2 } } } { \frac{h_{uu}}{h_{u}^{2 } } + p-2\frac{h_{uv}}{h_{u}h_{v } } + \frac{h_{vv}}{h_{v}^{2}}+q } \cdot t^{2},\end{aligned}\ ] ] for all @xmath112 and all @xmath113 where @xmath142 and @xmath109 are some sufficiently small numbers .
thus we obtain ( [ convergence ] ) .
inequality ( [ main - form ] ) holds for all real @xmath66 and @xmath67 with @xmath143 if and only if @xmath144 let us rewrite ( [ main - form ] ) as follows @xmath145+\\ & q\left[b^{2}\left ( \frac{h_{vv}}{h_{v}^{2}}-2\frac{h_{uv}}{h_{u } h_{v}}\right)+(b^{2}-1)\frac{h_{uu}}{h_{u}^{2 } } \right]-\frac{h_{uu}h_{vv}-h_{uv}^{2}}{h_{u}^{2}h_{v}^{2}}\geq 0\end{aligned}\ ] ] on the half plane where @xmath143 and @xmath146 in order for the quadric form @xmath147 to be nonnegative on the half plane it is necessary that @xmath148 and this gives the constraints on the numbers @xmath13 and @xmath14 .
further , without loss of generality we can assume that @xmath149 is invertible , i.e. , @xmath150 .
indeed , otherwise we can slightly perturb the numbers @xmath13 and @xmath14 .
notice that @xmath147 is nonnegative on the boundary of the half plane , i.e. , @xmath151 let @xmath152 be the vertex of the paraboloid @xmath153 , i.e. , @xmath154 .
it is enough to make sure that if @xmath155 then the point @xmath156 lies outside of the half plane .
indeed , the direct computations show that @xmath157 therefore @xmath158 this finishes the proof of the lemma and thereby the necessity conditions are obtained .
our main ingredient will be a subtle theorem 3 from @xcite .
let us precisely formulate it in the way we will use it . let @xmath159 and @xmath160 be some positive integers .
let @xmath161 be @xmath162 size matrices of full rank for @xmath163 and @xmath164 .
set @xmath165 to be @xmath166 size .
let @xmath167 be in @xmath168 where @xmath17 is a closed bounded rectangular domain , i.e. , parallelepiped with sides parallel to the standard orthonormal system in @xmath169 .
let @xmath149 be a positive definite @xmath170 matrix .
set @xmath171 by @xmath172 we denote the transpose of the matrix @xmath173 .
let @xmath174 be a row vector , i.e. , @xmath175 . let @xmath176 denotes @xmath177 matrix @xmath178 , i.e. , @xmath179 is constructed by the blocks @xmath180 .
[ ivo-1 ] @xmath181 on @xmath17 if and only if @xmath182 for all borel measurable @xmath183 .
the theorem was formulated for smooth @xmath184 and compactly supported functions @xmath185 , @xmath186 .
we should mention that the theorem still remains true for @xmath187 and for smooth bounded @xmath185 .
the proof proceeds absolutely in the same way as in @xcite .
the only property we need is that the heat extension of @xmath185 and @xmath188 must be well defined , which is definitely true under these assumptions as well . in order to obtain ( [ ivo-2 ] ) for all borel measurable functions @xmath189 we approximate pointwise almost everywhere by smooth bounded functions @xmath190 such that @xmath191 .
finally , the lebesgue dominated convergence theorem justifies the result ( notice that all functions @xmath192 and @xmath184 are uniformly bounded ) . we should also mention that inequality ( [ ivo-2 ] ) for the function @xmath193 recovers the reverse young s inequality for convolutions with sharp constants , and the latter was used in @xcite in obtaining the prkopa leindler inequality . in our case
the situation is slightly different
. we will be using ( [ ivo-2 ] ) for some sequence of functions @xmath184 , matrices @xmath149 , @xmath173 , and gaussian measures . further in obtaining the sufficiency condition , without loss of generality
we can assume that @xmath194 is the standard @xmath195-dimensional gaussian measure @xmath196 .
the case of the arbitrary gaussian measure ( [ gauss - measure ] ) follows by testing ( [ ragaca ] ) on the shifts and dilates of @xmath34 and change of variables in ( [ ragaca ] ) .
next we consider two different cases , when @xmath197 and when @xmath198 .
to the first case we refer as _ parabolic case _ and the second case we call _ elliptic case_. these names originate from studying the solutions of the partial differential inequality in ( [ numbers ] ) . in this subsection
we consider the case when @xmath199 ( [ heat ] ) holds if and only if @xmath200 or @xmath201 .
without loss of generality we will assume that @xmath202 for some @xmath203 .
moreover , we can choose @xmath203 so that @xmath204 .
set @xmath205 the choice of the numbers @xmath206 will be specified later .
so far we assume that @xmath207 .
notice that @xmath208 is a probability measure on @xmath7 .
we need the following lemma .
[ lem-5 ] for any @xmath209 with @xmath210 there exists @xmath211 such that for all @xmath212 we have @xmath213 where @xmath214 , @xmath215 if @xmath200 and @xmath216 if @xmath217 . before we proceed to the proof of the lemma let us explain that the lemma implies the desired result ( [ ragaca ] ) .
it is clear that if @xmath218 then the right hand side of ( [ mq1 ] ) tends to @xmath219 .
we claim that the left hand side of ( [ mq1 ] ) tends to @xmath220 . indeed ,
let @xmath221 we claim that @xmath222 a.e .
as @xmath223 .
notice that @xmath224 { } \operatorname*{ess\,sup}_{y } h ( f((x - y)/a ) , g(y / a)).\end{aligned}\ ] ] on the other hand let @xmath225 .
consider @xmath226 .
let @xmath227 be a sufficiently large number such that @xmath228 . here
@xmath229 denotes the ball centered at the origin with radius @xmath227 .
then @xmath230 { } \operatorname*{ess\,sup}_{y } h ( f((x - y)/a ) , g(y / a ) ) - \varepsilon.\end{aligned}\ ] ] the last passage follows from the fact that the power in the exponent ( [ measure ] ) is of order @xmath231 where @xmath232 . since @xmath74 is arbitrary we obtain the pointwise convergence for @xmath233 .
finally , since @xmath233 are uniformly bounded the lebesgue dominated convergence theorem implies @xmath234 { } \int_{\mathbb{r}^{n } } \operatorname*{ess\,sup}_{y } \;h\left ( f\left ( \frac{x - y}{a}\right ) , g\left(\frac{y}{b}\right ) \right)d\gamma_{n}(x).\end{aligned}\ ] ] it remains to prove the lemma .
take an arbitrary borel measurable @xmath235 such that @xmath236 for some @xmath237 .
let @xmath238 .
first , we show that @xmath239 we should apply theorem a. in order to do that notice that @xmath240 where @xmath241 and @xmath242 where @xmath243 is @xmath244 identity matrix .
clearly @xmath72 if and only if @xmath207 .
set @xmath245 let @xmath246 be @xmath247 matrix .
notice that @xmath248 where @xmath249 is @xmath250 matrix . by theorem a the desired inequality ( [ pirveli ] ) holds , namely , @xmath251 if and only if @xmath252 where @xmath253 is given on @xmath254 .
notice that in this case @xmath255 therefore @xmath256 \otimes i_{n\times n } ; \\ & a_{1}^{*}ca_{1 } = \langle \tilde{c}a_{1 } , a_{1 } \rangle = \frac{1}{a^{2}}\left[1+\frac{1}{p}+2\cdot \frac{q}{p}+\frac{q^{2}}{p^{2 } } \right ] \otimes i_{n\times n}.\end{aligned}\ ] ] ideally we want to have @xmath257 for all @xmath163 and @xmath164 .
unfortunately , it is not possible , i.e. , there are no solutions @xmath66 and @xmath67 which satisfy ( [ heat ] ) .
we choose @xmath258 , and @xmath259 if @xmath200 and @xmath260 if @xmath201 .
we will set @xmath261 but so far lets think of it as a sufficiently small number .
then @xmath262 and @xmath263 .
first we consider the case when @xmath31 .
the remaining cases are similar .
we obtain @xmath264 for @xmath253 we have @xmath265 we have @xmath266 where @xmath267 is a diagonal matrix with entries @xmath43 , @xmath43 and @xmath268 on the diagonal , and @xmath269 thus @xmath181 if and only if @xmath270
. if we set @xmath271 and @xmath272 then we have @xmath273 by choosing @xmath274 we will make diagonal entries positive .
notice that such choice is possible because @xmath275 and @xmath276 .
let us investigate the sign of @xmath277 leading minor .
we have @xmath278 where @xmath279 where @xmath280 is some absolute constant which depends only on @xmath281 and @xmath282 .
thus we see that choosing @xmath283 for some absolute @xmath284 the determinant of the minor will be positive . for the next @xmath285
minor we have @xmath286 notice that @xmath287 and since @xmath288 we have @xmath289 and the last expression is nonnegative if @xmath290 for some absolute @xmath291 . in the similar way we obtain that all @xmath285 minors are nonnegative provided that @xmath292 is sufficiently large .
so it remains to check the sign of @xmath293 .
we have @xmath294+(n - m)\frac{n \varepsilon}{2 } h_{x}^{2 } h_{y}^{2}(a+b)^{2}+\nonumber \\ & h^{2}(a^{2}b^{2}+\varepsilon^{2})(h_{xx}h_{yy}-h_{xy}^{2})+\varepsilon h^{2}(h_{xx}h_{yy}(a^{2}+b^{2})+2ab h_{xy}^{2 } ) .
\nonumber\end{aligned}\ ] ] notice that the first term , i.e. , @xmath295 is nonnegative by the condition of the theorem . for the rest of the terms
we notice that if we set @xmath296 and @xmath297 for any @xmath209 , @xmath210 , then we obtain that for sufficiently large @xmath298 we have @xmath299 , @xmath300 therefore the second term @xmath301 will dominate the rest of the terms , and in the last terms we notice that @xmath302 will dominate all the bounded terms . for the remaining cases when @xmath217 and @xmath303 we have @xmath304 @xmath305 in both of the cases
there is a diagonal matrix @xmath306 having entries @xmath307 on the diagonal such that @xmath308 notice that ( [ diagonal ] ) is the same as @xmath309 in ( [ matrixa ] ) except @xmath14 has switched the sign
. formulas ( [ first - minor ] ) , ( [ second - minor ] ) and ( [ third - minor ] ) are still valid if we switch the sign of @xmath14 to @xmath310 .
the rest of the discussions proceed without any changes . finally in order to obtain ( [ mq1 ] )
we take infimum of the left hand side of ( [ pirveli ] ) over all positive @xmath235 such that @xmath311 . indeed , for the convenience of the reader let us mention the following classical result .
[ classical ] let @xmath312 and @xmath313 be such that @xmath314 then for any positive bounded @xmath315 we have @xmath316 the equality holds if @xmath317 for a constant @xmath318 . indeed ,
notice that @xmath319 is a 1-homogeneous convex function for @xmath320 .
therefore ( [ jensen ] ) follows from the jensen s inequality . in case of ( [ pirveli ] )
we take @xmath321 and @xmath322 . taking infimum over all
positive and bounded @xmath235 with @xmath311 and finally rising the obtained inequality to the power @xmath323 we obtain ( [ mq1 ] ) .
in fact the infimum is attained on the following function @xmath324 where the constant @xmath325 is chosen so that @xmath326
. clearly such optimizer satisfies @xmath327 for some nonzero @xmath237 because @xmath2 is bounded and @xmath328 .
this finishes the proof in the case @xmath197 .
[ opt - structure ] inequality ( [ ivo-2 ] ) was obtained in @xcite as a result of an identity , involving ornstein uhlenbeck semigroups and quadratic form @xmath176 . since in ( [ optimizer-1 ] ) we have explicit optimizer we conclude that ( [ mq1 ] ) and thereby inequality ( [ ragaca ] ) can be rewritten as an equality in terms of limits , ornstein uhlenbeck semigroups and quadratic form @xmath179 .
one can use this information to obtain some structure of the optimizers @xmath38 in ( [ ragaca ] ) . in this subsection
we consider the following case @xmath329 [ lem-7 ] there exist positive constants @xmath330 and @xmath331 such that for any @xmath212 we have @xmath332 where @xmath208 is defined as in ( [ measure ] ) and @xmath333 as before using lemma [ classical ] it is enough to prove ( [ chveni ] ) for all bounded , positive and uniformly separated from zero @xmath235 where @xmath334 , @xmath109 will be determined later .
notice that ( [ choiceofp ] ) implies that @xmath335=1 ; \quad \langle \tilde{c}\tilde{a}_{1 } , \tilde{a}_{1 } \rangle = \frac{1}{a^{2}}\left[1+\frac{1}{p}+2\cdot \frac{q}{p}+\frac{q^{2}}{p^{2 } } \right]=1.\end{aligned}\ ] ] we have @xmath336 notice also that @xmath337 is positive - definite if and only if @xmath338 and @xmath29 .
let @xmath339 be the same matrix as before . by theorem
a , the inequality @xmath340 holds if and only if @xmath181 where again @xmath253 .
we have @xmath341 notice that as before it is enough to check positive definiteness of the following matrix @xmath342 where @xmath271 and @xmath272 .
if @xmath298 is sufficiently large then all diagonal entries are positive .
one can notice that all principal @xmath285 minors have positive determinant provided that @xmath298 is sufficiently large and @xmath343 .
this follows from the fact that @xmath344 and @xmath345 .
so it remains to check the sign of @xmath293 .
we have @xmath346+\\ & ( n - m)\left[\frac{1}{2}n h_{x}^{2}h_{y}^{2}(1-(a - b)^{2})((a+b)^{2}-1)+ h(4a^{2}b^{2}(h_{xx}h_{y}^{2}+h_{yy}h_{x}^{2})-2h_{xy}h_{x}h_{y}(1-a^{2}-b^{2})^{2})\right]\\ & \frac{1}{2}(n - m)n h_{x}^{2}h_{y}^{2}(1-(a - b)^{2})((a+b)^{2}-1)+4a^{2}b^{2}h^{2}\left(h_{xx}h_{yy}-h_{xy}^{2}\left ( \frac{1-a^{2}-b^{2}}{2ab}\right)^{2 } \right)=\\ & = m i_{1}+(n - m)i_{2}+i_{3}+i_{4}.\end{aligned}\ ] ] notice that the first term @xmath347 by ( [ numbers ] ) . the second term @xmath348 contains a factor of the form @xmath349 which will dominate the remaining subterms as @xmath350 .
finally the sum of the last two terms @xmath351 will be positive provided that @xmath352 the last inequality holds provided that @xmath109 is sufficiently large number . indeed ,
notice that if @xmath212 for some large @xmath353 then @xmath354 .
on the other hand the right hand side of ( [ final ] ) is bounded .
this finishes the proof of the lemma .
in this section we give the answers to the question what happens when @xmath15 and @xmath16 vanish in @xmath17 .
unfortunately our technique developed in section [ suf-1 ] directly does not apply to the cases when partial derivatives of @xmath2 vanish .
this obstacle could be overcome by using theorem 2 in @xcite .
let us illustrate its application .
let @xmath275 for some closed bounded rectangular domain @xmath4 .
let @xmath13 and @xmath14 be strictly positive numbers .
if @xmath355 , @xmath28 and the following conditions hold @xmath356 then @xmath357 holds for all smooth compactly supported @xmath358 , and any gaussian measure @xmath6 .
if @xmath29 then we require that @xmath359 .
the theorem was proved in @xcite corollary 5.2 under the extra assumptions that @xmath15 and @xmath16 never vanish in @xmath17 .
let us explain that the general case follows from theorem 2 given in @xcite .
we should mention that the assumption of smoothness ( actually continuity ) of @xmath38 in ( [ axali ] ) is crucial , and this makes the statement ( [ axali ] ) completely different than what we have in ( [ ragaca ] ) .
we will see that there are examples of discontinuous functions @xmath34 when ( [ axali ] ) does not hold at all , although @xmath2 satisfies ( [ odin ] ) and ( [ dva ] ) ( see the discussions after the proof of theorem [ ll - th ] ) . as before it is enough to obtain ( [ axali ] ) for the standard gaussian measure @xmath360 .
first let us mention that if ( [ axali ] ) holds for @xmath32 then it holds for all @xmath361 .
indeed , assume ( [ ragaca ] ) is true for @xmath32 .
let @xmath362 .
then @xmath363 for all @xmath364 .
let @xmath365 , @xmath366 and @xmath367 .
we apply one dimensional ( [ axali ] ) to the following functions @xmath368 we obtain @xmath369 then we consider new functions @xmath370 and we again apply ( [ ragaca ] ) .
after iterating this process we will obtain that @xmath371 thus we obtain the desired result .
let @xmath372 .
set @xmath373 by theorem 2 in @xcite the inequality @xmath374 for all @xmath375 implies the inequality @xmath376 if @xmath377 on the subspace @xmath378 . notice that if we take smooth bounded @xmath379 such that @xmath380 for all @xmath381 then ( [ before ] ) is the same as ( [ pi-1 ] ) for @xmath32 because @xmath382 for all @xmath163 and @xmath164 . on the other hand
the matrix @xmath383 is negative semidefinite on the subspace @xmath378 where @xmath384 if and only if @xmath385 these are precisely conditions ( [ odin ] ) and ( [ dva ] ) .
let @xmath6 be defined as in the introduction .
let us consider the following degenerate case @xmath386 where @xmath387 .
let @xmath388 be a closed bounded set @xmath389 . notice that ( [ axali ] ) for all borel measurable @xmath390 takes the following form @xmath391 where @xmath392 denotes the convex hull of the set @xmath393 .
notice that if @xmath394 is continuous then @xmath395 , therefore ( [ deg ] ) becomes equality and it is always true without any restrictions on @xmath184 .
however , the sufficiency condition , namely ( [ dva ] ) , requires that if @xmath396 then @xmath397 .
thus sufficiency condition ( [ dva ] ) is unnecessary if we consider inequality ( [ axali ] ) on the space of smooth compactly supported functions .
however , if we slightly enlarge the class of test functions , for example , if we allow to include discontinuous functions in ( [ axali ] ) then the sufficiency conditions ( [ odin ] ) and ( [ dva ] ) become necessary . in the next theorem
we assume that there exists a point @xmath398 such that @xmath399 .
[ ll - th ] let @xmath8 for some closed bounded rectangular domain @xmath4 .
let @xmath6 be defined as in introduction .
then for the inequality @xmath400 to hold for all continuous functions @xmath358 having at most one point of discontinuity it is necessary that @xmath401 , ( [ odin ] ) and ( [ dva ] ) holds . if @xmath29 then it is necessary that @xmath359 .
again we restrict the measure @xmath6 to the case @xmath32 , i.e. , @xmath6 is a probability measure on the real line with finite absolute fifth moment . if @xmath402 then condition ( [ odin ] ) together with constraints on numbers @xmath26 , i.e. , @xmath403 was already obtained in section [ nec-1 ] .
if @xmath404 but @xmath405 then we consider the following test functions @xmath406 where @xmath235 is a bounded function .
@xmath407 on the other hand @xmath408 therefore ( [ fun1 ] ) implies that @xmath409 assume the contrary that @xmath410 .
then we obtain that @xmath411 .
there exists a point @xmath412 such that @xmath413 .
take @xmath414 if @xmath415 and @xmath416 otherwise .
since @xmath417 we obtain that @xmath418 .
thus @xmath419 .
the similar reasoning works in the case when @xmath420 but @xmath421 . finally , if @xmath422 at point @xmath62 , then in the similar way we can show that @xmath423 and @xmath424 . by considering the following test functions @xmath425 and @xmath426 where @xmath235 and @xmath427 are bounded continuous functions , and using taylor s formula we obtain @xmath428 for all bounded continuous @xmath235 and @xmath427 . if @xmath429 then there is nothing to prove because we have that @xmath430 and @xmath431 are nonnegative .
further assume that @xmath432 .
consider the following function @xmath433 in a rectangular domain @xmath434 such that @xmath435 and @xmath436 never vanish .
@xmath153 does satisfy ( [ ragaca ] ) with bounded continuous test functions @xmath34 .
notice that the test functions ( [ test-1 ] ) , ( [ test-2 ] ) and ( [ test-3 ] ) which we used in derivation of necessity conditions for ( [ ragaca ] ) are continuous functions . therefore we can apply theorem [ mth1 ] , and we have @xmath437 .
this implies that @xmath438 for all @xmath439 such that @xmath440 and @xmath441 .
we should assume that @xmath442 otherwise there is nothing to prove .
therefore ( [ last1 ] ) can be rewritten as follows @xmath443 for all nonzero @xmath444 and @xmath445 .
the last inequality implies ( [ dva ] ) , and this finishes the proof of the theorem .
let us make the following remark .
notice that for inequality ( [ fun1 ] ) to hold for the standard gaussian measure and the class of functions @xmath34 having at most one point of discontinuity it is necessary that conditions ( [ odin ] ) and ( [ dva ] ) must be satisfied .
these conditions are also sufficient for inequality ( [ fun1 ] ) to hold for a class of smooth functions but not discontinuous .
for example , consider @xmath446 .
this function does satisfy ( [ odin ] ) , ( [ dva ] ) . on the other hand , inequality ( [ fun1 ] ) fails for the function @xmath447 for @xmath448 and @xmath449 otherwise .
thus we see that in the degenerate case , i.e. , when we allow @xmath15 and @xmath16 to switch the signs in @xmath17 , it is crucial whether the space of test functions consists by continuous or discontinuous functions .
in the applications @xmath2 will not belong to the class @xmath450 .
instead we will only have @xmath451 and @xmath2 is lower - semicontinuous on @xmath17 .
thus we can not directly apply theorem [ mth1 ] . in order to avoid this obstacle
we will slightly modify the functions @xmath2 and then pass to the limit in ( [ ragaca ] ) .
for example , if @xmath452^{2}$ ] we will consider auxiliary functions @xmath453 for @xmath454 , and we apply ( [ ragaca ] ) to these functions .
finally we just send @xmath455 and @xmath456 in the appropriate order .
let @xmath457 .
the ehrhard inequality states that if @xmath458 and @xmath459 then @xmath460 for all borel measurable @xmath461 such that the minkowski sum @xmath462 is measurable .
the equality is attained in ( [ ehrhard ] ) for the half spaces with one containing the other one .
inequality ( [ ehrhard ] ) was proved by ehrhard @xcite when @xmath173 and @xmath184 are convex sets under the assumptions that @xmath31 .
ehrhard , by developing gaussian symmetrization method showed that ( [ ehrhard ] ) is enough to prove in the case @xmath32 .
it was an open problem whether ( [ ehrhard ] ) holds for the borel measurable sets @xmath173 and @xmath184 ( see @xcite ) .
latala @xcite showed that the inequality is true if at least one of the sets is convex ( again under the constraints @xmath31 ) .
it was also noticed that the inequality is equivalent to its functional version @xmath463 for all smooth @xmath464^{2}$ ] for some @xmath465 .
finally , borell in his series of papers @xcite using a maximum principle obtained ( [ borel - result12 ] ) .
we refer the reader to @xcite for further developments of the subject . among many applications of the ehrhard ineuqlaiy
, it also allows to find _ isoperimetric profile _ for the gaussian measure .
let @xmath6 be a probability measure on @xmath7 .
let @xmath466 be an epsilon neighborhood of the set @xmath173 .
set @xmath467 the function @xmath468 is called isoperimetric profile of the measure @xmath469 .
@xmath468 measures minimal _
perimeter _ of the set @xmath173 under the constraint that @xmath470 is fixed .
one can obtain from ( [ ehrhard ] ) that @xmath471 which is regarded as an infinitesimal version of @xmath472 .
a subtle result of bobkov @xcite asserts that for any even , log - concave measure @xmath6 on the real line we have @xmath473 where @xmath474 . we should also mention that if @xmath6 is a probability measure with positive distribution function @xmath40 on the real line then there is a trivial upper bound @xmath475 the inequality is exhausted by half - lines .
if @xmath6 is a log - concave measure then we also have a trivial lower bound @xmath476 for @xmath31 via the prkopa leindler inequality .
these considerations motivate to the following question : which measures @xmath6 satisfy ( [ ehrhard ] ) or ( [ borel - result12 ] ) with @xmath477 replaced by @xmath478 , i.e. , with a distribution function of @xmath6 .
further by @xmath479 we denote product measure , i.e. , @xmath480 .
[ theorem2 ] let @xmath6 be a probability measure with density function @xmath481 and finite absolute fifth moment .
let @xmath482 .
let @xmath13 and @xmath14 be some fixed positive numbers .
let @xmath483 on @xmath484^{2 } \setminus \{(0,1 ) \cup ( 1,0)\}$ ] and set @xmath485 . for the inequality @xmath486 to hold for @xmath487 and all borel measurable @xmath488^{2}$ ] it is necessary that @xmath489 moreover , if @xmath490 is the gaussian measure then the above conditions are also sufficient for the inequality to hold for any @xmath361 .
further inequality ( [ borel - result ] ) we call the ehrhard inequality .
first we prove the necessity part .
we consider @xmath0 on the domain @xmath491^{2}$ ] for some @xmath492 .
clearly @xmath493 and in particular ( [ borel - result ] ) holds for borel measurable @xmath494 .
notice that @xmath15 and @xmath16 never vanish in @xmath495 .
it remains to use the theorem [ mth1 ] .
direct computations show that @xmath496 .
\end{aligned}\ ] ] therefore if we introduce new variables @xmath497 and @xmath498 we see that by theorem [ mth1 ] the necessary condition for ( [ borel - result ] ) is ( [ log - prime ] ) . for the sufficiency condition we should introduce an auxiliary function @xmath499 notice that @xmath500^{2})$ ] and it satisfies ( [ numbers ] ) . by theorem [ mth1 ] we have ( [ ragaca ] ) for @xmath501 .
we consider @xmath502 since @xmath2 and hence @xmath501 is increasing in each variable we have @xmath503 almost everywhere .
as @xmath60 the sequence of functions @xmath504 is decreasing in @xmath74 .
these two conditions imply that @xmath505 almost everywhere for some borel measurable @xmath506 .
clearly @xmath507 .
we claim that @xmath508 a.e .. indeed , otherwise @xmath509 for some @xmath412 on some set of positive measure . since continuous function on a compact set is uniformly continuous we can choose @xmath142 so that @xmath510^{2 } ) } \leq \tau/2 $ ] for all @xmath511 .
this would contradict to the fact that @xmath509 .
thus @xmath512 almost everywhere . since @xmath513 the lebesgue dominated convergence theorem implies that @xmath514 in @xmath515 . using the fact that @xmath2 is increasing in each variable we obtain @xmath516 .
this gives the left hand side of ( [ borel - result ] ) .
for the right hand side notice that if the point @xmath517 coincides with @xmath518 or @xmath519 then there is nothing to prove because @xmath520 . in the remaining case when @xmath517 is the point of continuity of @xmath2 in @xmath484^{2}$ ]
we obtain the right hand side of ( [ borel - result ] ) by taking the limit .
the next corollary says that in the class of even probability measures on the real line , the only measures which satisfy the ehrhard inequality are the gaussian measures .
[ even - measure ] an even probability measure @xmath6 with finite absolute fifth moment and the density function
@xmath521 satisfies ehrhad inequality ( [ borel - result ] ) with @xmath32 and some @xmath522 if and only if it is the gaussian measure . by ( [ log - prime ] ) and
the fact that @xmath523 is an odd function we obtain @xmath524 therefore @xmath525 for all @xmath381 .
if we take derivative with respect to @xmath444 we obtain @xmath526 .
choose @xmath445 so that @xmath527 then we obtain that @xmath528 for all @xmath529 .
thus @xmath530 if @xmath29 and @xmath531 if @xmath31 .
further we will just write @xmath532 instead of considering previous cases separately . in both cases @xmath128 because @xmath533 .
on the other hand testing the ehrhard inequality ( [ borel - result ] ) with @xmath534 and test functions @xmath535 , @xmath536 we see that after the change of variables the inequality holds for the probability measures @xmath537 with @xmath538 .
it turns out that the measures @xmath539 which satisfy @xmath540 for all @xmath381 and all @xmath541 with @xmath27 and @xmath542 have a simple geometrical description .
[ last3 ] let @xmath6 be a probability measure with the density function @xmath543 and finite absolute fifth moment . assume that the ehrhard inequality ( [ borel - result ] ) holds for all @xmath522 with @xmath27 and @xmath28 .
then @xmath544 and @xmath523 is a convex function .
moreover , there exist constants @xmath545 with @xmath546 such that @xmath547 .
by theorem [ theorem2 ] we have @xmath548 for all real @xmath439 and for all positive numbers @xmath10 with @xmath27 and @xmath459 .
inequality ( [ geometry ] ) has the following geometrical meaning .
let @xmath549 be the epigraph of @xmath523 .
condition ( [ geometry ] ) means that @xmath550 for all @xmath27 and @xmath28 .
it follows that the infinity parallelogram @xmath551 ( see figure [ fig : pizza ] ) with sides @xmath552;\\ & ( x , v'(x))+s \cdot ( ( x , v'(x))+(y , v'(y ) ) ) , \quad s \in [ 0,\infty)\\ & ( y , v'(y))+s \cdot ( ( x , v'(x))+(y , v'(y ) ) ) , \quad s \in [ 0,\infty)\\ \end{aligned}\ ] ] belongs to @xmath553 . since this is true for all @xmath381 it follows that this can happen if and only if @xmath523 is convex and @xmath553 contains all lines @xmath554 of the form @xmath555 for all @xmath529 .
then it follows that there exist real numbers @xmath556 , @xmath557 such that @xmath558 . since @xmath559 it follows that there exists sufficiently large @xmath560 such that @xmath561 and @xmath562 .
this implies that @xmath545 .
, parallelogram @xmath551 and the line @xmath554 ] one can observe that @xmath564 if and only if @xmath6 is the gaussian measure .
it would be interesting to see whether the converse of corollary [ last3 ] is also true at least for @xmath31 , i.e. , a probability measure with density @xmath539 and the function @xmath565 described in corollary [ last3 ] satisfies the ehrhard inequality ( [ borel - result ] ) with @xmath32 .
if this is the case then for such measures we obtain @xmath566 .
next we investigate 1-homogeneous functions which satisfy ( [ ragaca ] ) .
the class of 1-homogeneous functions was studied in a remarkable paper of borell @xcite .
one should compare our results of subsection [ lebeg - hom ] with the results of borell .
for the convenience of the reader we have included borell s theorem in appendix ( see theorem b ) . in this section
we describe all 1-homogeneous functions @xmath2 which satisfy ( [ ragaca ] ) .
it turns out that they are either convex functions , or the prkopa leindler type functions ( [ p-1 ] ) , ( [ p-2 ] ) and ( [ p-3 ] ) .
further we will always assume that the numbers @xmath541 satisfy the constraint in ( [ numbers ] ) , i.e. , @xmath27 and @xmath28 . [ concavity - gaussian ] let @xmath567 be 1-homogeneous function with @xmath276 .
partial differential inequality ( [ numbers ] ) holds on @xmath568 if and only if one of the following holds : @xmath569 since @xmath2 is 1-homogeneous we have @xmath570 for some @xmath379 .
conditions @xmath571 and @xmath572 imply that @xmath573 and @xmath574 .
notice that ( [ numbers ] ) takes the following form @xmath575 where @xmath576 and @xmath577 .
notice that @xmath578 .
we have @xmath579 the last expression becomes zero if and only if @xmath580 and @xmath581 .
but if @xmath580 then @xmath582 ( since @xmath582 ) .
we consider several cases .
suppose @xmath583 .
then @xmath584 for some nonzero @xmath149 . then @xmath585 and
we obtain that @xmath586 .
the last equality implies that either @xmath31 or @xmath50 .
thus we obtain that @xmath587 for some nonzero @xmath149 with @xmath31 or @xmath50 .
suppose @xmath588 .
then @xmath589 for some nonzero @xmath149 . then @xmath590 and
we obtain that @xmath591 .
the last equality can happen if and only if @xmath48 .
thus @xmath592 with some nonzero @xmath149 and @xmath48 .
thus if ( [ diffur1 ] ) becomes zero on some interval then the several cases might happen : 1 ) @xmath31 and @xmath593 ; 2 ) @xmath50 and @xmath593 ; 3 ) @xmath48 and @xmath589 .
the obtained functions are either convex functions or concave functions .
notice that non of these concave functions can be glued smoothly with a convex function @xmath594 .
it follows that ( [ pde1 ] ) is nonnegative if and only if either @xmath379 is a convex function , or @xmath379 is a concave function of the form @xmath595 since @xmath596 and @xmath2 is 1-homogeneous , then convexity / concavity of @xmath379 is equivalent to convexity / concavity of @xmath2 . in the first case
when @xmath594 , @xmath597 and @xmath598 then @xmath0 becomes 1-homogeneous convex function such that @xmath15 and @xmath16 never vanish . in the remaining cases
the descriptions are given in ( [ ff-1 ] ) , ( [ ff-2 ] ) and ( [ ff-3 ] ) .
thus in case of smooth 1-homogeneous functions there are two instances : @xmath2 is convex , or @xmath2 coincides with one of the functions ( [ p-1 ] ) , ( [ p-2 ] ) and ( [ p-3 ] ) .
next we describe measures @xmath6 which satisfy ( [ ragaca ] ) for 1-homogeneous functions @xmath2 .
first we consider the case when @xmath2 is a function of the form ( [ ff-1 ] ) , ( [ ff-2 ] ) and ( [ ff-3 ] ) . the case of convex functions we consider separately in the next subsection .
functions found in ( [ f-1 ] ) , ( [ f-2 ] ) and ( [ f-3 ] ) provide us with the following inequalities .
[ character ] let @xmath6 be the gaussian measure ( or the lebesgue measure ) .
we have @xmath599 for all nonnegative borel measurable @xmath600 . moreover , if @xmath6 is even then @xmath601 for all bounded compactly supported nonnegative borel measurable @xmath34 with positive @xmath602 and @xmath603 .
inequalities in the corollary follow from the application of theorem [ mth1 ] to the functions ( [ p-1 ] ) , ( [ p-2 ] ) and ( [ p-3 ] ) .
the only obstacle to directly apply theorem [ mth1 ] is that @xmath604 . to avoid this obstacle one needs to consider an auxiliary function @xmath605 for @xmath606 and then send @xmath607 ( see the similar discussions in ( [ auxiliary ] ) ) .
case of the lebesgue measure follow from the gaussain measure and the fact that @xmath2 is 1-homogeneous . indeed , we can test inequalities in the corollary for the following test functions @xmath608 and @xmath609 . by making change of variables and using 1-homogeneity of @xmath2
we can send @xmath610 and obtain the desired result .
inequality ( [ prekopa-1 ] ) is the classical prkopa leindler inequality @xcite . among its many applications we should mention a remarkable paper @xcite .
stability of ( [ prekopa-1 ] ) was studied in @xcite .
the inequality implies that the marginals of log - concave measures are log - concave .
for a local version of the latter fact we refer the reader to @xcite ( see also @xcite for the complex setting ) .
an extension of the inequality was obtained in @xcite .
inequality ( [ prekopa-2 ] ) can be understood as an extension of the classical prkopa leindler inequality for @xmath611 $ ] .
in fact , one can show that ( [ prekopa-1 ] ) and ( [ prekopa-2 ] ) are equivalent if instead of essential infimum in ( [ prekopa-2 ] ) we would have only infimum . since the measure is even in ( [ prekopa-2 ] )
, it is easy to show that ( [ prekopa-2 ] ) for @xmath612 is the same as ( [ prekopa-2 ] ) for @xmath613 .
notice that if we take @xmath614 and @xmath615 in ( [ prekopa-2 ] ) for compact borel measurable @xmath616 with @xmath617 then we obtain @xmath618 since the measure is even then if we interchange @xmath619 , @xmath620 and we denote @xmath621 then we obtain from ( [ difference ] ) that @xmath622 \setminus c | = 0\ } |^{\lambda } \cdot |b|^{1-\lambda}. \end{aligned}\ ] ] inequality ( [ pre-2 ] ) does not follow immediately from the general brunn minkowski inequality @xmath623 . however , it is easy to see that weaker version of ( [ pre-2 ] ) , i.e. , @xmath624 \setminus c = \emptyset \ } |^{\lambda } \cdot |b|^{1-\lambda } \end{aligned}\ ] ] is equivalent to the brunn minkowski inequality .
it is the remarkable result of borell @xcite that ( [ prekopa-1 ] ) holds if and only if @xmath6 is a log - concave measure .
one can show that the weaker version of ( [ prekopa-2 ] ) , i.e. , when essential infimum is replace by infimum , also holds for even log - concave measures .
finally we would like to mention that even though among @xmath625 smooth 1-homogenous functions @xmath2 with nonvanishing @xmath15 and @xmath16 there are only two instances either @xmath2 is convex or @xmath2 is of the form ( [ p-1 ] ) , ( [ p-2 ] ) and ( [ p-3 ] ) , it is not the case in general if we drop the assumption of smoothness .
we can always take the maximum of any two functions which satisfy ( [ cl - max ] ) .
indeed , next proposition says that ( [ ragaca ] ) is closed under taking maximum .
[ cl - max ] if @xmath626 and @xmath627 satisfy ( [ ragaca ] ) then @xmath628 also satisfies ( [ ragaca ] ) .
indeed , suppose @xmath629 then since @xmath630 we have @xmath631 next we describe measures @xmath6 and convex functions @xmath2 which satisfy ( [ ragaca ] ) .
if @xmath2 is a convex function ( not necessarily 1-homogeneous ) one should be careful in this case especially when @xmath31 .
this case is a trivial application of the fact that @xmath632 and the jensen s inequality , i.e. , ( [ ragaca ] ) holds for any probability measure @xmath6 .
if in addition @xmath2 is 1-homogeneous it easily follows that ( [ ragaca ] ) holds for any @xmath633-finite measure @xmath634 .
let us formulate the last statement precisely .
[ jensen2 ] let @xmath635 be a nonnegative 1-homogeneous convex function .
then for any @xmath633-finite measure @xmath634 absolutely continuous with respect to the lebesgue measure we have @xmath636 there exist measurable sets @xmath637 such that @xmath638 and @xmath639 for all @xmath640 .
we apply jensen s inequality to the probability measures @xmath641 . by using the fact that @xmath2 is 1-homogeneous we obtain @xmath642 finally we just send @xmath643 and we use the continuity of @xmath2 .
[ trivial ] fix some positive @xmath26 such that @xmath31 .
let @xmath644 for @xmath645 .
then @xmath646 is 1-homogeneous convex function on @xmath647 for @xmath648 .
thus ( [ jensen3 ] ) holds for @xmath649 with @xmath648 .
one may think that remark [ trivial ] improves the borell brascamp lieb inequality for @xmath648 because @xmath650 pointwise .
we caution the reader that this is not the case because in the borell brascamp lieb inequality @xmath646 is set to be zero if @xmath651 ( see @xcite ) . under these restrictions
the new function @xmath652 fails to be convex and the inequality becomes nontrivial ( see also @xcite for a riemannian version of the borell brascamp lieb inequality ) .
there are nontrivial functions @xmath34 for which inequality ( [ jensen3 ] ) for @xmath649 is equality .
indeed , set @xmath653 and @xmath654 for some positive function @xmath655 such that @xmath656 is convex
. then @xmath657 .
thus the integrand in the left hand side of ( [ jensen3 ] ) becomes @xmath658 and since @xmath659 we obtain equality in ( [ jensen3 ] ) .
this is the example of function @xmath2 for which partial differential inequality ( pdi ) in ( [ numbers ] ) holds with not being identically equality , however , ( [ ragaca ] ) is sharp .
the last remark also sheds light about connection of existence of nontrivial optimizers and equality cases of pdi in ( [ numbers ] )
. in this case inequality ( [ ragaca ] ) for the gaussian measure seems to be a nontrivial statement , i.e. , it does not follow from the application of jensen s inequality as before .
let @xmath660 for some closed bounded rectangular domain @xmath4 .
assume that @xmath2 is a convex function and @xmath35 never vanish in @xmath661 .
let @xmath26 be some positive numbers such that @xmath29 .
then ( [ ragaca ] ) holds for the gaussian measure @xmath6 if and only if @xmath28 and @xmath359 .
the necessity condition follows from theorem [ mth1 ] . for the sufficiency
it is enough to check that convex functions @xmath2 satisfy ( [ numbers ] ) . indeed , @xmath662 finally we need to invoke the auxiliary function ( [ auxiliary ] ) to relax the requirement @xmath8 to @xmath660 thus we finish characterization of ( [ ragaca ] ) in case of 1-homogeneous functions , and in case of convex functions . in the next section
we describe all concave functions @xmath2 which satisfy ( [ ragaca ] ) .
we noticed in the previous section that if @xmath2 is a convex function then ( [ numbers ] ) holds .
actually if we assume even less than convexity of @xmath2 , for example , the following matrix is positive semidefinite @xmath663 then @xmath2 still satisfies ( [ numbers ] ) .
indeed , the proof proceeds in the same way as in ( [ conv1 ] ) .
notice that since @xmath664 condition ( [ con1 ] ) is weaker than convexity of @xmath2 : any convex function @xmath2 satisfies ( [ con1 ] ) but not any function @xmath2 which satisfies ( [ con1 ] ) is a convex function .
there are many functions which are not convex and satisfy ( [ con1 ] ) .
such functions appeared in the proof of hypercontractivity of the ornstein uhlenbeck semigroup and in obtaining the gaussian noise stability .
for example , let @xmath665 and @xmath666 be two standard real gaussian random variables but they are not independent @xmath667 , @xmath668
. consider @xmath669 where @xmath670 .
then @xmath2 is not convex but it satisfies ( [ con1 ] ) with @xmath671 .
we refer the reader to @xcite for more details . on the other hand
it is an interesting question whether there are concave functions @xmath2 which satisfy ( [ numbers ] ) .
for example , we already found some of them : ( [ p-1 ] ) , ( [ p-2 ] ) and ( [ p-3 ] ) .
these functions correspond to the prkopa
leinlder s inequality which sometimes is regarded as reverse to hlder s inequality because ( [ ragaca ] ) can be treated as a reverse to jensen s inequality .
are there other nontrivial concave functions @xmath2 which satisfy ( [ numbers ] ) ?
one trivial example is @xmath672 .
it turns out that all nontrivial concave functions which satisfy ( [ ragaca ] ) are of the form ( [ p-1 ] ) , ( [ p-2 ] ) and ( [ p-3 ] ) .
[ conc - sol ] let @xmath673 be a concave function such @xmath276 are nonvanishing .
assume that @xmath2 is nowhere linear .
@xmath2 satisfies ( [ numbers ] ) if and only if @xmath2 is generically equivalent to one of the following functions @xmath674 by _ generically equivalent _ we mean up to shifts , translations and dilations . for example , @xmath675 with @xmath72 and some constants @xmath676 generically is equivalent to @xmath46 .
it is clear that the functions of the form ( [ f-1 ] ) , ( [ f-2 ] ) and ( [ f-3 ] ) are concave functions and they satisfy ( [ numbers ] ) .
let us show that there are no other concave functions which satisfy ( [ numbers ] ) . without loss of generality we can assume that @xmath677 . indeed , otherwise if @xmath678 but @xmath679 then we can consider @xmath680 in the reflected domain @xmath681 .
notice that ( [ numbers ] ) does not change by this transformation .
@xmath682 therefore we must have @xmath683 it follows that if one of the second partial derivative becomes zero then all of them become zero .
consider a point @xmath62 in the interior of @xmath17 where @xmath684 .
then @xmath685 .
since @xmath686 we have that @xmath2 satisfies homogeneous monge ampre equation .
locally at point @xmath62 there exists a foliation , i.e. , some family of lines such that @xmath687 is constant on each line .
this is the result of pogorelov @xcite , and we also refer the reader for the terminology to @xcite , see lemma 3 .
the lines provide us with directions @xmath688 for some @xmath689 smooth @xmath690 and @xmath691 for some interval @xmath393 .
each line can be identified with some point @xmath691 , therefore we can think that @xmath687 depends only one parameter @xmath691 .
we have @xmath692 for all @xmath691 where @xmath693 are smooth functions of @xmath106 .
the last equality corresponds to the fact that @xmath2 has constant gradient on each line .
we can treat @xmath694 as a @xmath689 function in a neighborhood of @xmath62 such that @xmath695 .
since @xmath696 we obtain @xmath697 , i.e. , @xmath698 .
equation ( [ monge2 ] ) implies @xmath699 thus we obtain that @xmath700 .
the last equality implies that @xmath701 .
thus we obtain that @xmath702 with some @xmath72 .
equality @xmath703 implies two cases @xmath31 or @xmath704 .
we consider these cases separately .
* case a+b = 1 . * in this case since @xmath705 we obtain from ( [ monge2 ] ) that @xmath706 . then @xmath702 is the same as @xmath707 .
the last equality implies @xmath708 .
since @xmath706 we obtain that the right choice of the sign should be negative . thus @xmath709 . by using methods of characteristics
we obtain @xmath710 where @xmath711 , @xmath712 , @xmath106 belongs to some open interval and @xmath713 and @xmath714 are some constants .
if we set @xmath715 and @xmath716 then we obtain @xmath717 , and the result follows . * case |a - b|=1*. in this case we have @xmath718 and therefore @xmath719
. then @xmath720 .
we consider the case @xmath217 ( the remaining case is similar ) . again using the method of characteristics we obtain
obtain @xmath2 is generically equivalent to ( [ f-2 ] ) .
the case @xmath303 is similar .
thus we obtain that locally functions are generically equivalent to one of the functions in ( [ f-1 ] ) , ( [ f-2 ] ) and ( [ f-3 ] ) . since @xmath2 is a smooth function the only possibility to glue these local pieces is when @xmath2 just coincides with one of these functions everywhere on @xmath17 .
this finishes the proof of the proposition .
further we will always assume that @xmath541 , @xmath355 , @xmath28 and @xmath6 is the gaussian measure ( [ gauss - measure ] ) .
whenever @xmath29 we will always assume that @xmath721 .
for any real @xmath66 we define @xmath722 we set that @xmath723 almost everywhere if and only if @xmath724 . if @xmath725 then we set @xmath726 .
[ ee-1 ] let @xmath66 and @xmath67 be the arbitrary real numbers .
the following inequality holds @xmath727 for all borel measurable @xmath728 and @xmath729 if and only if @xmath730 .
we notice that the case @xmath731 with @xmath31 recovers the prkopa leindler inequality . without loss of generality
we assume that @xmath732 .
first let us obtain ( [ vse-1 ] ) for bounded @xmath394 and @xmath733 .
we start with the case when @xmath734 .
set @xmath735 on some bounded closed rectangular domain @xmath736
. then ( [ ragaca ] ) holds if and only if @xmath730 . indeed ,
notice that ( [ numbers ] ) takes the form @xmath737 finally , in ( [ ragaca ] ) we should take the following test functions @xmath738 and @xmath739 .
thus we obtain ( [ vse-1 ] ) for bounded functions @xmath34 and uniformly separated from zero , i.e. , @xmath740 for some @xmath741 .
the general case of bounded @xmath34 follows by considering @xmath742 and @xmath743 . by sending @xmath60 and using the dominated convergence theorem
we obtain ( [ vse-1 ] ) for @xmath34 .
if @xmath206 become zero then we consider the new exponents @xmath744 for @xmath225 sufficiently small so that @xmath745 for all @xmath511 .
since the map @xmath746 is increasing on @xmath747 we have that if ( [ vse-1 ] ) holds for @xmath206 then it holds for @xmath748 .
thus @xmath749 . sending @xmath60
we obtain that @xmath730 . for the converse
we notice that inequality @xmath730 implies the inequality @xmath750 . since @xmath751 we obtain ( [ vse-1 ] ) with @xmath752 and @xmath753 .
finally , notice that @xmath754 and @xmath755 . for arbitrary @xmath756 and @xmath757 we can approximate by bounded @xmath758 and @xmath759 with @xmath760 in @xmath3 and @xmath761 in @xmath762 . since @xmath763 almost everywhere we obtain the desired result .
[ ee-2 ] let @xmath764 then @xmath765 for all nonnegative @xmath756 and @xmath757 if and only if @xmath766 and @xmath767 .
indeed , consider @xmath768 .
let @xmath769 . then notice that ( [ numbers ] ) takes the form @xmath770 for the quantity in ( [ ex21 ] ) to be nonnegative it is necessary that @xmath771 .
we can assume that @xmath772 otherwise the conclusion follows easily .
finally ( [ ex22 ] ) implies that ( [ ragaca ] ) holds if and only if @xmath773 . now we consider ( [ ragaca ] ) with the test functions @xmath774 , @xmath775 and we obtain ( [ minkowski ] ) .
it is interesting to mention that if @xmath776 then we can take @xmath777 in ( [ minkowski ] ) , since @xmath778 , we obtain the following corollary [ ee-3 ] let @xmath779 be such that @xmath776 .
then @xmath780 for all nonnegative @xmath756 and @xmath757 .
let @xmath388 and @xmath393 be bounded closed subsets of the real line @xmath747 . by symbol @xmath781
we denote all pair of borel measurable functions @xmath38 such that @xmath782 and @xmath783 . in other words
@xmath784 let @xmath6 be a probability measure on @xmath7 such that @xmath785 for some positive borel measurable function @xmath235 .
let @xmath786 be an arbitrary borel measurable function .
fix some @xmath787 .
we consider the following extremal problem @xmath788 we will omit variables @xmath789 and we will just write @xmath790 instead of @xmath791 . since @xmath6 is a probability measure it is easy to notice that @xmath790 is defined on a bounded rectangular domain @xmath792 , i.e. , convex hull of @xmath793 .
the goal of the paper is to find a good lower bound for @xmath790 under the assumptions that @xmath794 , @xmath28 and @xmath355 .
[ fi ] let @xmath795 be such that @xmath796 on @xmath793 , @xmath2 satisfies ( [ numbers ] ) and @xmath35 never vanish in @xmath797 .
then @xmath798 on @xmath797 .
indeed , pick any @xmath799 such that @xmath800 and @xmath801 .
then @xmath802 finally we take infimum over all such test functions @xmath34 and we obtain @xmath798 . since ( [ ragaca ] )
is closed under taking maximum it is reasonable to seek for the maximal function @xmath2 such that @xmath796 on @xmath793 and @xmath2 satisfies ( [ numbers ] ) .
this procedure together with proposition [ fi ] allows to find good lower bounds for @xmath790 .
next we consider a particular case when when @xmath803 , and @xmath804 and @xmath805 .
then we have @xmath806 where @xmath807 is understood as the essential sum of the sets @xmath173 and @xmath184 ( see @xcite ) . then finding the function @xmath184 becomes the following boundary value problem : find the maximal function @xmath2 on @xmath808^{2}$ ] which satisfies ( [ numbers ] ) and has the obstacle condition : @xmath809 and @xmath810
. one should seek such functions among those when partial differential inequality in ( [ numbers ] ) becomes equality .
solution of the pde ( [ numbers ] ) reduces to laplacian eigenvalue problem , and sometimes to backwards heat equation ( see @xcite ) .
it turns out that the ehrhard function is the best possible @xmath811 .
thus this algorithm indicates how to find the ehrhard s function by solving pde problem with the obstacle condition @xmath153 over the square @xmath484^{2}$ ] .
the following remarkable result belongs to borell @xcite .
[ borel - th ] let @xmath812 be a continuously differentiable function such that @xmath813 let @xmath814 and @xmath815 .
further suppose @xmath816 $ ] is a continuous 1-homogeneous function and increasing in each variable .
then the inequality @xmath817 holds for all nonempty @xmath818 if and only if there are sets @xmath819 @xmath820 such that @xmath821 for all @xmath822 , @xmath823 , @xmath824 and for every @xmath825 .
moreover if ( [ bo - je ] ) is holds then @xmath826 and @xmath827 will do .
let us consider a particular case when @xmath32 . since in the current paper we are interested when the inequalities of the form @xmath363 imply its integral version @xmath828 then in order to apply borell s result we should take @xmath829 for @xmath645 .
then ( [ borell - integral ] ) takes the following form @xmath830 . the form ( [ bo - je ] )
reduces to the form ( [ ragaca ] ) if @xmath831 and @xmath379 is supported on @xmath832 .
this may happen if and only if @xmath833 .
since @xmath173 and @xmath184 is arbitrary we obtain that @xmath834 .
therefore the last condition takes the form @xmath835 thus if ( [ borell - integral2 ] ) holds for all @xmath836 and all nonnegative @xmath837 then we obtain the integral inequality @xmath838 since @xmath839 may not be measurable we should understand the integral in the left hand side of ( [ integral - version ] ) as an upper integral .
let @xmath840 be 1-homogeneous , nonnegative and increasing in each variable .
assume @xmath833 .
if @xmath40 satisfies ( [ borell - integral2 ] ) for all positive @xmath841 , and with some positive @xmath26 such that @xmath31 then @xmath842 .
@xmath40 is one homogeneous therefore @xmath843 for some nonnegative increasing function @xmath325 .
( [ borell - integral2 ] ) simplifies to @xmath844 for all @xmath845 .
let @xmath846 and @xmath847 then we obtain @xmath848 .
set @xmath849 .
then by taylor s expansion we obtain that for sufficiently small @xmath74 we have @xmath850 .
since @xmath74 can be negative as well we obtain @xmath851 and hence @xmath852 for some @xmath72 .
therefore @xmath853 .
thus the corollary shows that in the particular case @xmath854 the functions which satisfy the assumption of borell s theorem ( [ borell - integral2 ] ) and hence would give us integral inequality ( [ integral - version ] ) are of the form @xmath855 .
the reader can recognize that this is the instance of the prkopa leindler inequality .
notice that this also confirms ours results : in subsection [ lebeg - hom ] we have found that @xmath40 has to convex function or @xmath40 has to be function of the form ( [ p-1 ] ) , ( [ p-2 ] ) and ( [ p-2 ] ) . since in application of borell
s theorem we require that @xmath856 , and @xmath857 then the only possibility is @xmath858 . indeed , 1-homogeneous convex nonnegative function @xmath859 on @xmath647 with values zero at the points @xmath860 , @xmath518 and @xmath519 must be identically zero . h. j. brascamp , e. h. lieb , _ on extensions of the brunn minkowski and prkopa leindler theorems , including inequalities for log concave functions , and with an application to the diffusion equation _
, journal of functional analysis , * 22 * , issue 4 , 366389 ( 1976 ) | in this paper we study a natural extension of the prkopa leindler inequality involving the essential supremum , almost arbitrary probability measures , and arbitrary function @xmath0 of two variables instead of the classical one @xmath1 .
we obtain the necessary conditions on @xmath2 in terms of partial differential inequalities , and in case of gaussian measures we show that these conditions are also sufficient .
we completely describe convex , concave and 1-homogeneous functions @xmath2 which satisfy our necessary conditions , moreover , we illustrate some other nontrivial examples . as an immediate application we obtain the new proof of the ehrhard inequality . in particular , we show that in the class of even probability measures the gaussian measure is the only one which satisfies the functional form of the ehrhard inequality on the real line with their own distribution functions .
our approach is based on two arguments : one of brascamp lieb rewriting the essential supremum as a limit of @xmath3 norms , and the other one is a subtle inequality which is an extension of the classical brascamp
lieb inequality recently developed for the gaussian measures .
namedefsubjclassname@2010 mathematics subject classification |
in this paper we will be concerned with the numerical calculation of the casimir interaction energy in geometries with translational invariance along one direction , i.e. very long cylinders of arbitrary section . for the sake of simplicity
, we will first discuss the case of a massless quantum scalar field that satisfies dirichlet or neumann boundary conditions on the surfaces of the cylinders .
as we will see , in some particular situations the generalization to the electromagnetic field and/or more general boundary conditions will be straightforward .
the casimir energy is formally given by @xmath0 where @xmath1 are the eigenfrequencies of the scalar field satisfying the appropriate boundary conditions on the surfaces of the shells , and @xmath2 are those corresponding to a situation in which the distances between the shells is very large .
the subindex @xmath3 denotes the set of quantum numbers associated to each eigenfrequency .
we have introduced an exponential cutoff for high frequency modes . for the particular geometry considered here ,
@xmath4 and the eigenfrequencies are of the form @xmath5 , where @xmath6 is a continuous variable associated to the translational invariance along the z - direction and @xmath7 are the eigenvalues of the laplacian on the two - dimensional transversal section @xmath8 contained in the plane ( x , y ) : @xmath9 the eigenfunctions @xmath10 satisfy dirichlet or neumann boundary conditions on @xmath11 , the boundary of @xmath8 .
the helmholtz equation ( [ helmholtz ] ) arises in many branches of physics , from the vibration of membranes to quantum billiards , and there are a plethora of methods to compute numerically its eigenfunctions and eigenvalues @xcite . among them , the `` boundary methods '' are based on the following strategy : the solution @xmath12 is written as a ( finite ) linear combination of basis functions that satisfy helmholtz equation inside @xmath8 .
the coefficients of the linear combination are chosen in such a way that the boundary conditions are satisfied at a finite number of points on @xmath11 .
the linear system of equations that determine the coefficients has a non trivial solution only for some particular values of @xmath13 , the eigenvalues of the system .
for example , in the point matching method ( pmm)@xcite , one expands the eigenfunction @xmath12 in terms of a basis of solutions of the helmholtz equation in free space @xmath14 @xmath15 in the numerical calculation this expansion is truncated at given @xmath16 , and the boundary conditions are imposed on @xmath17 points on @xmath11 .
these boundary conditions become a set of homogeneous , linear equations for the unknown coefficients @xmath18 ( @xmath19 , with @xmath20 a @xmath13-dependent @xmath21 matrix ) which has nontrivial solutions only when @xmath22 .
the last equation can be used to determine numerically the eigenvalues @xmath23 . in a similar approach , known as the method of fundamental solutions ( mfs ) @xcite ,
the eigenfunction @xmath12 is expanded in terms of solutions of the helmholtz equation with a point source at an arbitrary location @xmath24 , that we denote by @xmath25 @xmath26 if the sources are located _ outside _ @xmath8 , this is a solution of the homogeneous helmholtz equation _ inside _ @xmath8 .
once again , the sum is truncated at @xmath16 , and the coefficients @xmath27 are determined by solving the linear system that results when imposing the boundary conditions on a finite number of points on @xmath11 .
the roots of the determinant of the associated matrix are the eigenfrequencies of the problem .
this is the simplest version of the mfs , in which the locations of the sources are fixed .
one can find in the literature discussions about spurious solutions , improvements and alternative methods to find the eigenvalues .
we refer the reader to refs .
@xcite for more details .
a crucial point is that , at a practical level , the knowledge of the spectrum of the helmholtz equation is not enough to compute the casimir energy .
the reason is that the numerical evaluation of the sum over modes in eq.([modes ] ) is extremely unstable @xcite , and one has to subtract very large numbers to compute the finite interaction energy .
the calculation is complicated even for the simplest case of casimir effect in @xmath28 dimensions . instead of performing explicitly the summation
, it is far more efficient to combine the methods mentioned previously with the cauchy s theorem @xmath29 where @xmath30 is an analytic function in the complex @xmath31 plane within the closed contour @xmath32 , with simple zeros at @xmath33 within @xmath32 .
we use this result to replace the sum over the eigenvalues of the helmholtz equation in the casimir energy eq.([modes ] ) by a contour integral . in this way , it is not necessary to solve numerically the equation @xmath22 for the eigenvalues , but to take @xmath34 in the cauchy s theorem . in other words ,
if in the numerical method to solve the helmholtz equation the eigenvalues are the roots of a given function , one can integrate this function in the complex plane in order to get the casimir energy .
the combination of the use of numerical methods to compute the eigenvalues with the cauchy s theorem is the main idea we want to put forward in this paper . in the next section
we will describe the simplest version of the pmm to a situation in which the surfaces separate regions of different permittivities , generalizing our previous results @xcite for perfect conductors . in sections 3 and 4 we will review some numerical evaluations of the casimir interaction energy for perfect conductors . in section 5
we include our final remarks .
a media - separated waveguide presents an interesting setup for the application of the pmm .
this technique has been widely used to solve eigenvalue problems in many areas of engineering science @xcite .
the boundary conditions are imposed at a finite number of points around the periphery of both media . for the sake of concreteness , we will bear in mind the situation in which one perfect conductor encloses two dielectric media , as shown in fig.[fig1 ] , although the method could be applied to more general cases . the general solution of the helmholtz equation in region i ( inside the inner cylinder ) is @xmath35 while in region ii ( annular region ) @xmath36 e^{i m \theta},\label{c2}\ ] ] where @xmath37 are polar coordinates , and @xmath38 and @xmath39 are the @xmath40-th order bessel functions .
the constants @xmath41 and @xmath42 are determined by the boundary conditions . in both equations
we have defined @xmath43 .
we assume the outer surface to be a `` perfect conductor '' , and impose dirichlet boundary conditions on a finite number of points @xmath44 of @xmath45 : @xmath46 e^{i m \theta_q } , \label{perfectcond}\ ] ] ( alternatively , for the te modes of the electromagnetic field , one should impose neumann boundary conditions ) .
the surface @xmath47 as a dielectric interphase separating media @xmath48 and @xmath49 , and therefore we impose continuity of the field and its derivative : @xmath50 e^{i m \theta_p } \\
\sum_{m = -s}^{s } a_m j_m'(\lambda^{(i ) } r_p ) e^{i m \theta_p } & = & \frac{\lambda^{(ii ) } } { \lambda^{(i ) } } \sum_{m = - s}^s [ b_m j_m'(\lambda^{(ii ) } r_p ) + c_m h'^{(1)}_m(\lambda^{(ii ) } r_p ) ] e^{i m \theta_p},\nonumber \label{continuity}\end{aligned}\ ] ] where @xmath51 are points on the curve @xmath52 .
the boundary conditions can be written , in matrix form , as @xmath53 eliminating the coefficients @xmath54 we end with @xmath55 where @xmath56 and @xmath57 can be written as @xmath58 it is worthy to note that as @xmath59 is proportional to @xmath60 , then @xmath61 when @xmath62 .
thus , the matrices @xmath63 and @xmath64 , re - obtaining in this way , the usual perfect conductor wave - guide case studied in @xcite . for the system of eq.([syst ] ) to have non trivial solutions , the determinant must be zero , i.e. @xmath65 = det p_2 \cdot det n_1 \cdot det ( 1 - n_2 p_2^{-1 } p_1n_1^{-1})=0 ~.\ ] ] this equation determines the eigenfrequencies associated to the geometry .
however , as already mentioned , in order to compute the casimir energy it is not necessary to find each eigenvalue but to integrate the determinant @xmath66 in the complex plane .
we have developed a numerical fortran routine in order to evaluate the casimir interaction energy in the case in which the field satisfies dirichlet or neumann boundary conditions on both curves @xmath47 and @xmath67 . in this case
one should consider the fields only in region ii with @xmath68 .
after some straightforward steps one can re - write the casimir energy as a single integral in the imaginary axis @xmath69 .
for dirichlet boundary conditions the result is @xmath70 while for neumann boundary conditions one can derive a similar expression with a different function @xmath71 .
it is worth to stress that these casimir energies correspond to those of tm and te modes of the electromagnetic field in the presence of perfect conductors .
when two concentric cylinders have corrugations , the vacuum energy produces a torque that could , in principle , make one cylinder rotate with respect to the other . this `` cylindrical rack and pinion '' has been proposed in ref .
@xcite , where the torque has been computed using the proximity force approximation .
it was further analyzed in @xcite , where the authors obtained perturbative results for dirichlet boundary conditions in the limit of small amplitude corrugations . in this section
, we numerically evaluate the casimir interaction energy for two concentric corrugated , perfect conductor cylinders .
the cylinders have radii @xmath72 and @xmath73 , and we will denote by @xmath74 the mean distance between them and by @xmath75 the sum of the radii .
we will use the notation @xmath76 .
the points in the mesh , that give us the corrugated cylinder boundaries , are described by the following functions : @xmath77 where @xmath78 is the corrugation amplitude and @xmath79 is the frequency associated with these corrugations .
the casimir torque can be calculated by taking the derivative of the interaction energy with respect to the shifted angle @xmath80 . in fig.[fig10
] we show the numerical evaluation of the tm casimir interaction energy for this geometry .
the plot shows the results obtained using the pmm with @xmath81 and corrugation frequency @xmath82 , for different values of the amplitude of the corrugation @xmath78 . as expected the amplitude of the oscillations grows with @xmath78 . for each value of @xmath78
we have performed a numerical fit of the data in order to compare with the analytical prediction . with dotted lines
we have plotted the fit @xmath83 for each curve in fig.[fig10 ] .
the agreement between dots and dotted lines is extremely good .
similar results can be obtained for the neumann ( te ) modes ( see @xcite for details ) . for @xmath84 and different values of the perturbation @xmath78 .
the different shaped dots are the numerical data obtained with our program while the lines represent the numerical fit of each curve .
energies are measured in units of @xmath85 , and distances in units of @xmath72.,width=325 ] for @xmath81 , @xmath86 and @xmath87 .
the different shaped dots are the numerical data obtained by our program while the line represents the numerical fit of each curve . in this case
, the plot shows that the exact result can not be fitted by a function @xmath88 .
energies are measured in units of @xmath85.,width=325 ] it is worth to remark that , when the amplitude of the corrugation is not very small , the exact results can not be reproduced with a simple fit of the form @xmath89 .
this is illustrated in fig.[fig12 ] , where we see that , for the biggest corrugated amplitude @xmath90 , the exact result differs from the cosine function @xcite .
some time ago , there was a conjeture @xcite based on a geometric optics approximation , about the possibility of focusing vacuum fluctuations in parabolic mirrors .
it was argued that a parabolic mirror is capable of focusing the vacuum modes of the quantized electromagnetic field , therefore creating large physical effects near the mirror s focus . with this motivation , in this section we shall evaluate the casimir interaction energy for configurations in which the outer conducting shell has a cross section that contains focal points .
we will consider one small inner cylinder and an outer ellipse .
we will denote by @xmath72 the radius of the inner cylinder , by @xmath91 and @xmath92 the minor and major semiaxes of the ellipse , respectively , and by @xmath93 the distance between the foci and the center of the ellipse .
the coordinates of the center of the cylinder with respect to the center of the ellipse will be @xmath94 .
we will use an additional tilde to denote adimensional quantities , i.e distances in units of @xmath72 : @xmath95 , etc . for this configuration ,
we use a mesh where with an inner cylinder , and an outer ellipse with semiaxes @xmath96 and @xmath97 .
the ellipse has two focal points at @xmath98 .
we present the results for the casimir energy in fig.[fig15 ] .
.,width=325 ] from fig.[fig15 ] it is possible to see that there is an unstable equilibrium position at the origin under displacements of the inner cylinder along the ( vertical ) @xmath99 direction . as expected , it is also possible to check that the energy grows as well as the cylinder gets closer to the surface of the outer ellipse .
fig.[fig15 ] also shows a monotonic behaviour of the energy as a function of the position , even when passing through the focus .
so we do not see a focusing of vacuum fluctuations near the focus of the ellipse . however , in order to confirm this result one should consider much smaller inner cylinders , in order to explore shorter wavelengths .
this will require much more computational effort .
finally , we have also checked @xcite that there is an unstable equilibrium position at the origin when moving the inner cylinder in the ( horizontal ) @xmath100 direction .
we have presented new numerical methods to compute the vacuum energy for arbitrary geometries with translational invariance .
the approach is based on the use of traditional boundary methods to compute eigenvalues of the two dimensional helmholtz equation , combined with cauchy s theorem . as a particular example
, we have described a straightforward version of the point - matching method to compute the casimir interaction energy for a waveguide with different permittivities , and reviewed some numerical calculations for perfect conductors . in all examples , for the numerical calculations we have chosen pair of points with the same angular coordinate with respect to the inner cylinder . for less symmetric configurations , and
when the surfaces of both conductors are closer to each other , it will be necessary to consider grids with a larger number of points , and to optimize their positions .
as in the applications to acoustic or classical electromagnetism , special care must be taken for surfaces with pronounced edges , clefts or `` handles '' , where the point - matching technique may not be accurate to determine the eigenfrequencies . in these cases , more sophisticated approaches
@xcite could be necessary to optimize the numerical evaluation and to avoid spurious solutions .
kuttler and v.g .
sigillito , siam review 26 , 163 ( 1984 ) .
r bates , ieee trans . on microwave theory and techniques , vol .
mtt-17 , 297 ( 1969 ) ; h. y. yee and n.f.audeh , ieee trans . on microwave theory and techniques , vol .
mtt-13 , 847 ( 1965 ) ; ibidem vol .
mtt-14 , 487 ( 1966 ) ; j.r .
kuttler and v.g .
sigillito , siam review bf 26 , 163 ( 1984 ) and references therein . for a generalization and applications to scattering problems see f.m .
kahnert , j. quant .
transfer * 79 - 80 * , 775 ( 2003 ) and references therein .
a. karageorghis , appl .
14 , 837 ( 2001 ) ; c.c .
tsai et al , proc .
a462 , 1442 ( 2006 ) .
villadsen and e. stewart , chem .
22 , 1483 ( 1967 ) ; t. betcke and l.n .
trefethen , siam review 47 , 469 ( 2005)c.j.s . alves and p.r.s .
antunes , cmc 2 , 251 ( 2005 ) ; d. cohen , n. lepore and e.j .
heller , j. phys . a : math .
37 , 2139 ( 2004 ) ; p. amore , arxiv 0910.4798v1 [ quant - ph ] .
a. rodriguez , m. ibanescu , d. iannuzzi , f. capasso , j. d. joannopoulos , and s.g .
johnson , phys .
lett . * 99 * , 080401 ( 2007 ) .
lombardo , f.d .
mazzitelli , p.i .
villar , and m. vzquez , phys .
rev . d*80 * , 0605018 ( 2009 ) .
f. d. mazzitelli , f. c. lombardo and p. i. villar , j. phys .
161 , 012015 ( 2009 ) .
i. cavero - pelez , k.a .
milton , p. parashar and k.v .
shajesh , phys .
d * 78 * , 065019 ( 2008 ) .
ford and n.f .
svaiter , phys .
a * 62 * , 062105 ( 2000 ) . | we discuss new approaches to compute numerically the casimir interaction energy for waveguides of arbitrary section , based on the boundary methods traditionally used to compute eigenvalues of the 2d helmholtz equation .
these methods are combined with the cauchy s theorem in order to perform the sum over modes . as an illustration
, we describe a point - matching technique to compute the vacuum energy for waveguides containing media with different permittivities .
we present explicit numerical evaluations for perfect conducting surfaces in the case of concentric corrugated cylinders and a circular cylinder inside an elliptic one . |
the bcs theory is one of the most successful theories in physics .
it has not only revealed the mechanism of phonoic superconductivity but also developed our understanding of spontaneous breaking of gauge symmetry .
the bcs theory as a field theory has been applied to various stages of physics , elementary particles , condensed matter and cosmology .
the discovery of high-@xmath3 cuprate superconductors has renewed our interests of superconductivity .
low dimensionality on layered compounds , carrier doping effect , electron correlation effect , antiferromagnetic ordering and mott transition play the key roles to understand the high-@xmath3 physics . since the high-@xmath3 superconductivity emerges in close to antiferromagnetic instability , the significant role of magnetism is associated with the high-@xmath3 mechanism . since coulomb repulsion between electrons can produce spin and charge fluctuations , a close relation between magnetism and superconductivity
is suggested .
the superconducting properties except the high @xmath3 and vortex matter physics are conventional .
the electronic and magnetic properties in the normal conducting state are unconventional .
the high-@xmath3 cuprate superconductors have still attracted great interests .
we may anticipate new development in understanding the solid state physics .
mercury - based cuprates hgba@xmath0cuo@xmath1 with 0 @xmath2 0.2 ( hg1201 ) are the superconductors with a single cuo@xmath0 layer in unit cell and the optimally oxidized one has the highest @xmath3 = 98 k among the ever reported single - cuo@xmath0-layer superconductors @xcite .
mercury - based cuprates hgba@xmath0cacu@xmath0o@xmath4 with 0.05 @xmath2 0.35 ( hg1212 ) are the superconductors with double cuo@xmath0 layers in unit cell and the optimally oxidized one has the highest @xmath3 = 127 k among the ever reported double - cuo@xmath0-layer superconductors @xcite . hg1201 and hg1212 are the layered compounds and contain one and two cuo@xmath0 planes in unit cell , respectively @xcite .
triple - cuo@xmath0-layer hgba@xmath0ca@xmath0cu@xmath9o@xmath10 ( hg1223 ) is the highest @xmath3 = 134 k at ambient pressure @xcite and @xmath3 = 164 k at high pressure @xcite .
hg1234 is also known to be synthesized @xcite .
but it has lower @xmath3 than hg1223 @xcite . in this article
, we present the hg - based superconducting cuprates and the microscopic magnetic properties studied through nuclear magnetic resonance ( nmr ) and nuclear quadrupole resonance ( nqr ) techniques .
especially , we present the cu nmr evidence of the existence of a pseudo spin - gap in the norma - state magnetic excitation spectrum of the single - layer cuprate superconductor .
the discovery of a big pseudo spin - gap of the underdoped single - layer cuprate superconductor has turn our attention to the unconventional electronic state above @xmath3 .
conventional metallic states are well described by the landau - fermi liquid theory .
uniform magnetic susceptibility @xmath11 exhibits pauli paramagnetism . since the density of electron states is finite at the fermi level , @xmath11 is finite at @xmath12 = 0 k and nearly independent of temperature .
nuclear spin - lattice relaxation time @xmath13 satisfies korringa relation with the spin knight shift . since the quasi - particle scattering through the finite density of states induces the nuclear spin relaxation , 1/@xmath14 is finite at @xmath12 = 0 k. magnetic itinerant compounds involving the transition metal elements often exhibit curie - weiss magnetism , in spite of the absence of localized moments .
itinerant magnetism breaks down the korringa relation . in antiferromagnetic compounds ,
the nuclear spin - lattice relaxation rate divided by temperature 1/@xmath15 is enhanced in a curie - weiss law , more than the uniform spin susceptibility @xmath11 .
the normal - state pseudo spin - gap was first confirmed by nmr experiments @xcite .
the decrease of a static uniform spin susceptibility @xmath16 with cooling down was observed in the nmr knight shift measurements for the widely - studied double - layer yba@xmath0cu@xmath9o@xmath17 .
the decrease of the nuclear spin - lattice relaxation rates divided by temperature 1/@xmath15 was also observed for the underdoped cuprates in the normal state .
thus , the existence of a pseudogap in the magnetic excitation spectrum was observed in the underdoped double - layer superconductors in the normal states .
however , the existence of such a pseudogap in the magnetic excitation spectrum was not obvious for the single - layer superconductor la@xmath18sr@xmath19cuo@xmath20 .
then , there had been an issue whether the pseudogap in the magnetic excitation spectrum is intrinsic in a cuo@xmath0 plane or it is secondary due to spin singlet formation between adjacent cuo@xmath0 planes in the double layer .
the nmr studies of hg1201 gave us an obvious evidence of the pseudo spin - gap in the single layer .
the review articles of the intensive nmr studies of the high-@xmath3 yba@xmath0cu@xmath9o@xmath17 can be seen in @xcite .
hence , we focus on the nmr studies of hg1201 and hg1212 .
cuo@xmath1 and double - layer hgba@xmath0cacu@xmath0o@xmath4 .
strong o - hg - o bonds like dumbbell " along the @xmath21-axis and dilute oxygen concentration in the hgo layers are characteristics of these crystal structures . ]
phase diagrams of hg1201 and hg1212 reproduced from @xcite .
@xmath3 is plotted against excess oxygen concentration @xmath22 @xcite(a ) , the in - plane @xmath23-axis lattice constant @xcite(b ) , and hole concentration @xmath24 defined by the ionic formal valence of cu@xmath25 ( @xmath24 = 2@xmath22 for hg1201 and @xmath24 = @xmath22 for hg1212 ) @xcite(c ) . ]
figure [ cs ] shows the crystal structures of the hg - based superconductors hg1201 and hg1212 @xcite .
strong covalent o - hg - o bonds along the @xmath21-axis , which look like dumbbell " shape , and dilute oxygen concentration in the hgo@xmath26 layers are characteristics of the crystal structures .
the change in oxygen concentration of the hgo@xmath26 layers can yield a wide carrier doping region .
figure [ pd ] ( a ) shows the oxygen concentration dependences of @xmath3 of hg1201 , hg1212 , and hg1223 @xcite .
the typical bell "- shaped dependences of @xmath3 are seen for hg1201 and hg1212 . with changing only the oxygen concentration
, the electronic states of hg1201 and hg1212 develop from the underdoped to the overdoped regions although no magnetic ordering states nor insulating states were confirmed for the deeply underdoped samples of hg1201 and hg1212 , the existence of hg - based cuprate insulators was observed for hg@xmath0ba@xmath0ycu@xmath0o@xmath27 ( hg2212 ) which contains the double hg layers @xcite .
figure [ pd ] ( b ) shows @xmath3 plotted against the in - plane @xmath23-axis lattice parameter for hg1201 , hg1212 , and hg1223 @xcite .
the in - plane cu - cu distances of hg1212 and hg1223 are shorter than that of hg1201 .
the shrunk cuo@xmath0 planes characterize the higher @xmath3 cuprates . figure [ pd ] ( c ) shows @xmath3 plotted against the hole concentration @xmath24 defined by the ionic formal valence of cu@xmath25 @xcite .
we estimated @xmath24 = 2@xmath22 for hg1201 and @xmath24 = @xmath22 for hg1212 .
the optimal hole concentration is nearly the same as the typical values of 0.18 - 0.20 .
then , the mechanism of the record high @xmath3 is not only due to the substantial hole concentration .
figure [ cuo2 ] illustrates a cuo@xmath0 plane and two key structure parameters .
the two key parameters are the distance from the plane - site cu ion to the apical oxygen and the bond angle between the plane - site cu ions via the plane - site oxygen .
both parameters are associated with the flatness of the cuo@xmath0 planes .
the large separation between cu and the apical oxygen and the nearly 180 degree of the in - plane cu - o - cu bond angle are observed in hg1201 and hg1212 @xcite .
the optimal @xmath3 = 38 k la@xmath18sr@xmath19cuo@xmath28 with the single cuo@xmath0 layers and the optimal @xmath3 = 93 k yba@xmath0cu@xmath9o@xmath29 with the double cuo@xmath0 layers have the buckling structures in the cuo@xmath0 planes .
the cuo@xmath0 planes of hg1201 and hg1212 are the flattest among the reported superconducting cuprates @xcite .
many researchers believe that the nearly perfect flat cuo@xmath0 plane is a significant factor to realize the higher @xmath3 . actually , the double - hg - layer ( hg , tl)2212 has the relatively lower @xmath3 and involves the bucking in the cuo@xmath0 planes @xcite . in the ( hg , tl)2212 , the in - plane cu - o - cu bond is bended and the bond angle is about 170 degree .
the flatness of the cuo@xmath0 planes is associated with a higher @xmath3 mechanism or at least an inevitable background of crystal structure .
plane ( left ) and two key structure parameters ( right ) are illustrated .
the two key structure parameters ( right ) are the distance from the plane - site cu to the apical oxygen and the bond angle between the cu ions via the oxygen ion in the cuo@xmath0 plane .
the flatness of the cuo@xmath0 plane is believed to be a significant factor to the higher @xmath3 . ]
anisotropy of electrical resistivity has been measured for hg1201 single crystals @xcite and hg1212 epitaxial thin films on vicinal substrates @xcite .
figure [ rhot ] shows the in - plane resistivity @xmath30 and the out - of - plane resistivity @xmath31 for hg1201 single crystal of @xmath3 = 97 k reproduced from @xcite ( a ) and those of ( hg , re)1212 of @xmath3 = 117 k reproduced from @xcite ( b ) .
dash lines are @xmath12 linear functions for the in - plane resistivity @xmath30 .
the @xmath12-linear resistivity is associated with the conduction electron scattering due to two dimensional antiferromagnetic spin fluctuations @xcite .
the deviation from the @xmath12-linear behavior is associated with the scattering suppression due to the opening of a pseudo spin - gap in the magnetic excitation spectrum .
the onset temperatures of deviation from the @xmath12 linear functions are denoted by @xmath32 . at @xmath32 ,
the out - of - plane resistivity @xmath31 also takes the minimum value . the out - of - plane resistivity @xmath31 is two or three order higher than the in - plane resistivity @xmath30 and the temperature dependence is different from each other .
the metallic in
plane resistivity @xmath31 and the semi - conducting out - of - resisitivity @xmath31 above @xmath32 is an evidence of two dimensional electrical conduction in hg1201 and hg1212 .
this has been known for the other superconducting cuprates .
the in - plane and the out - of - plane electrical conduction is correlated with each other with respect to the @xmath32 .
the two dimensional electrical resistivity is understood in terms of a hot spot and a cold spot on the two dimensional fermi surface @xcite . and @xmath31 of hg1201 ( upper panel ) @xcite and hg1212 ( lower panel ) @xcite .
dash lines for the in - plane resistivity are @xmath12 linear functions .
the onset temperatures of deviation from the @xmath12 linear functions are denoted by @xmath32 , which are also the minimum temperatures of the out - of - plane resistivity . ]
nuclear magnetic resonance ( nmr ) and nuclear quadrupole resonance ( nqr ) are powerful techniques to characterize microscopically magnetic insulators , metals , superconductors , alloys and compounds @xcite .
microscopic studies using the nmr and nqr techniques have provided us with rich information inside unit cell through the nuclear sites in a site - selective way .
using the nmr and nqr techniques , one can obtain static and dynamic information of the electronic systems .
cu nqr frequency spectra for single - cuo@xmath0-layer superconducting cuprates ( left panels ) ( a ) and double - cuo2-layer superconducting cuprates ( right panels ) ( b ) at @xmath12 = 4.2 k. ] @xmath5cu nuclei have spin @xmath33 = 3/2 and then quadrupole moments @xmath34 . in a non - cubic crystalline symmetry , they interact with the electric field gradient @xmath35 ( @xmath36 = @xmath37 ) of electric static crystalline potential @xmath38 @xcite . nuclear quadrupole hamiltonian is given by @xmath39 \label{nq}\ ] ] where @xmath40 is the maximum component of the electric field gradient @xmath41 and @xmath42 is called the asymmetry parameter @xmath43 @xmath5cu nuclear quadrupole resonance frequency @xmath44 is given by @xmath45 where @xmath46 = @xmath47/2 .
the ratio of the natural abundance of @xmath7cu and @xmath48cu atoms is about 69.1 to 30.9 .
the ratio of the nuclear quadrupole moments @xmath34 is about 0.211 to 0.195 .
for one crysallographic cu site , a pair of @xmath7cu and @xmath48cu nqr lines is observed .
figure [ cunqr ] shows actual zero - field plane - site @xmath5cu nqr frequency spectra for various optimally doped single - cuo@xmath0-layer superconducting cuprates ( left panels ) ( a ) and double - cuo2-layer superconducting cuprates ( right panels ) ( b ) at @xmath12 = 4.2 k. the cu nqr spectra were measured for hg1201 in @xcite , tl@xmath0ba@xmath0cuo@xmath4 @xcite , la@xmath49sr@xmath50cuo@xmath28 in @xcite , ( ca , na)@xmath0cuo@xmath0cl@xmath0 in @xcite , hg1212 in @xcite , bi@xmath51sr@xmath52ca@xmath53cu@xmath0o@xmath10 in @xcite , yba@xmath0cu@xmath9o@xmath54 in @xcite and yba@xmath0cu@xmath28o@xmath55 in @xcite .
solid curves are simulations using multiple gaussian functions .
more than two pairs of @xmath7cu and @xmath48cu nqr lines were needed to reproduce the broad nqr spectra except yba@xmath0cu@xmath9o@xmath56 and yba@xmath0cu@xmath28o@xmath55 .
non - stoichiometry in compounds gives rise to inhomogeneous broadening in nmr and nqr spectra , because magnetic shift and quadrupole frequency are distributed .
yba@xmath0cu@xmath28o@xmath55 with double cuo@xmath57 planes and double cuo chains in unit cell has @xmath3 = 82 k. in fig .
[ cunqr ] , this stoichiometric and naturally underdoped compound shows sharp cu nqr spectra . yba@xmath0cu@xmath9o@xmath54 with double cuo@xmath57 planes and a single cuo chain in unit cell has @xmath3 = 93 k. in fig .
[ cunqr ] , this stoichiometric and slightly overdoped compound shows sharp cu nqr spectra .
although the cuo@xmath0 planes of these compounds have the buckling structures , the cu nqr spectra are rather sharp .
the superconducting hg1201 and hg1212 at any doping level , however , show the inhomogeneously broad cu nqr spectra @xcite . in general ,
crystalline imperfection is the origin of the inhomogeneous broadening of an nqr spectrum . in fig .
[ cunqr ] , bi@xmath51sr@xmath52ca@xmath53cu@xmath0o@xmath10 ( bi2212 ) also shows a broad cu nqr spectrum .
this is attributed to a wide range structural modulation of the bio layers . in spite of the dilute oxygen concentration in the hgo layers of hg1201 and hg1212 , most of the cu nuclei feel inhomogeneous electric field gradients .
thus , the effect of dilute oxygen ions on the crystalline potential may be long ranged .
long - range friedel oscillations from the excess oxygen ions may yield such a broad nqr spectrum .
the inhomogeneously broad cu nqr spectrum is observed in the nearly perfect flat cuo@xmath0 plane .
this is instructive for us to understand the relation between nqr and the crystal structure .
the broad cu nqr spectrum results from the nonstoichiometry but not from the buckling of the cuo@xmath0 planes . in fig .
[ cunqr ] , one should note that the value of cu @xmath44 of hg1201 is similar to that of tl@xmath0ba@xmath0cuo@xmath4 and about a half of those of yba@xmath0cu@xmath9o@xmath56 and yba@xmath0cu@xmath28o@xmath55 .
the local density approximation calculations of the nqr frequencies account for these experimental similarity and difference @xcite .
a static magnetic field @xmath58 is applied to a material along the @xmath59-axis and then a magnetization @xmath60 is measured by a magnetometer .
the bulk magnetization @xmath60 is the sum of respective electron spins @xmath61 @xmath62 where @xmath63 is a thermal average of the electron spin along the @xmath64-axis .
then the bulk magnetic susceptibility @xmath65 is defined by @xmath66 the paramagnetism of 3@xmath6 transition metal oxides results from the unpaired @xmath6 electron spins and the orbital momentums .
the bulk magnetic susceptibility is expressed by the sum of the spin susceptibility @xmath67 , the van vleck orbital susceptibility @xmath68 , and the diamagnetic susceptibility @xmath69 of inner core electrons @xmath70 of the optimally doped powdered sample of hg1201 ( @xmath3 = 98 k ) in @xcite at a magnetic field of 1 tesla @xcite . ]
fgiure [ xthg1201 ] shows the uniform magnetic susceptibility @xmath65 of the optimally doped powdered sample of hg1201 ( @xmath3 = 98 k ) in @xcite at a magnetic field of 1 tesla @xcite .
the pseudo spin - gap behavior is seen in the bulk magnetic susceptibility .
the magnitude of @xmath65 above 150 k is the same order of the other high-@xmath3 superconducting cuprates
. nuclear spins of constituent ions in a crystal interact with electron spins through an electron - nuclear hyperfine coupling .
when a static uniform magnetic field @xmath71 is applied to the system along the @xmath59-axis , the electron medium is polarized along the @xmath59-axis .
the magnetic polarization produces an additional magnetic field at the nuclear sites through the hyperfine coupling and then leads to a shift of the resonance field ( frequency ) of the nuclear spins .
this is knight shift @xmath72 .
the static nuclear spin hamiltonian of a zeeman coupling and the electron - nuclear hyperfine coupling is given by @xmath73 where @xmath74 is the nuclear gyromagnetic ratio and @xmath75 is a hyperfine coupling constant along the @xmath64-axis with the @xmath76-site electron spin .
the experimental knight shift is defined by a shift of the observed resonance frequency @xmath77 from the reference frequency @xmath78 @xmath79 in the 3@xmath6 transition metal oxides , the knight shift @xmath72 is decomposed into the spin shift @xmath80 due to unpaired electron spins and the orbital shift @xmath81 due to the van vleck orbital susceptibility @xmath82 the spin knight shift @xmath83 is given by @xmath84 where @xmath85 is the uniform fourier component of the hyperfine coupling constant and @xmath86 is the static ( @xmath87 = 0 ) uniform ( @xmath40 = 0 ) electron spin susceptibility .
@xmath88 is the bohr magneton .
@xmath89 is the avogadro number .
@xmath90 is the measured bulk susceptibility in emu / mole - atom .
the hyperfine coupling constant reflects the characters of the wave functions of the electron orbitals .
measurement of the knight shift at each nuclear enables us to obtain the site - specific information .
the temperature dependence of the knight shift reveals that of the intrinsic uniform spin susceptibility @xmath67 at each site . in covalent bonded compounds ,
the transferred and supertransferred hyperfine coupling constants play important roles . in the high-@xmath3 cuprate superconductors ,
the fourier transformed hyperfine coupling constants at the plane - site cu and the oxygen are expressed by @xmath91 and @xmath92 where @xmath93 ( @xmath64 = @xmath94 and @xmath95 ) is an on - site anisotropic hyperfine coupling constant , @xmath96(@xmath97 0 ) is a supertransferred isotropic hyperfine coupling constant from a cu to a cu through an oxygen , and @xmath98(@xmath97 0 ) is a transferred hyperfine coupling constant from a cu to an oxygen @xcite . for the optimally doped hg1201
, we estimate @xmath99 = 145 koe / mole - cu-@xmath88 from the @xmath100 plot , where the knight shift @xmath101 @xcite is plotted against the bulk magnetic susceptibility @xmath65 in fig .
[ xthg1201 ] with temperature as an implicit parameter .
nuclear spins are coupled by the fluctuating hyperfine fields of electron spins through a time - dependent hyperfine coupling hamiltonian @xmath102 in a spin - echo recovery technique , the nuclear moments excited by an inversion @xmath103-pulse are in a thermal non - equilibrium state .
the energy dissipation from the nuclear moments to a lattice takes place through the fluctuating hyperfine fields .
the recovery time of the nuclear moments to a thermal equilibrium state is the nuclear spin - lattice relaxation time @xmath104 .
moriya derived a general expression of @xmath104 @xcite , @xmath105 @xmath106 is the dimension of space on which the electronic system lies .
1/@xmath14 is the square of the hyperfine coupling constant @xmath107 times the wave vector averaged low frequency dynamical spin susceptibility @xmath108 @xcite .
thus , from measurements of the knight shift and the nuclear spin - lattice relaxation time , one can infer the @xmath40 dependence of the dynamical spin susceptibility @xmath109 .
the modified korringa ratio can give us a criterion which the electronic system is , ferromagnetic or antiferromagnetic and where @xmath108 is enhanced in the @xmath40 space .
in general , the @xmath110 dependence of the hyperfine coupling constant @xmath111 is slower than that of @xmath112 .
but , the form factor of the plane - site oxygen @xmath113 of eq .
( [ oaq ] ) acts as a filter for a finite @xmath110 correlation .
the antiferromagnetic correlation of @xmath114 = [ @xmath115 , @xmath115 ] and @xmath117 = [ @xmath115(1@xmath118 ) , @xmath115(1@xmath118 ) ] is cancelled at the plane - site oxygen through the form factor of eq .
( [ oaq ] ) .
the plane - site cu nuclei can probe the antiferromagnetic correlation through eq .
( [ cuaq ] ) .
strong indirect cu nuclear spin - spin interaction was first found in yba@xmath0cu@xmath9o@xmath56 @xcite .
gaussian decay in the transverse relaxation of the plane - site cu nuclear moments is predominantly induced by the indirect nuclear spin - spin interaction through the in - plane antiferromagnetic electron spin susceptibility .
strong antiferromagnetic fluctuations persist in all the high-@xmath3 cuprate superconductors .
thus , the gaussian decay rate of the plane - site cu nuclear spin - spin relaxation provides us fruitful information on the antiferromagnetic electron spin susceptibility @xcite .
the indirect nuclear spin - spin interaction is given by @xmath119 where a range function @xmath120(r@xmath121 ) is given by @xmath122 the gaussian decay rate 1/@xmath123 of the nuclear spin - spin relaxation is given by @xmath124 the kramers - kronig relation is @xmath125 the gaussian decay rate 1/@xmath123 reflects the full range frequency integration of the dynamical spin susceptibility @xmath126(@xmath127 , @xmath87 ) .
the cross section of inelastic neutron scattering is expressed by @xmath128 where @xmath40 is a momentum transfer of the neutron , @xmath87 the energy transfer , and @xmath129 = 1/@xmath130 @xcite . in the high-@xmath3 cuprate superconductors ,
the enhancement of @xmath126(@xmath127 , @xmath87 ) was observed over a finite range of @xmath87 and at around @xmath131 = @xmath114 or @xmath132 @xcite . in general , it is hard to see @xmath126(@xmath127 , @xmath87 ) over the full range . from measurements of the nuclear spin - lattice relaxation and the gaussian decay rate of nuclear spin - spin relaxation
, one can infer the dynamical spin susceptibility at low and high frequency regions .
the electronic state of a cuo@xmath0 plane of the high-@xmath3 cuprate superconductor is described by a single band picture .
the anisotropy of knight shift and nuclear spin - lattice relaxation and the site differentiations are explained by anisotropic hyperfine coupling and different @xmath110 dependence of the coupling constant due to transferred and supertransferred hyperfine couplings .
that is mila - rice - shastry hyperfine coupling hamiltonian @xcite .
two dimensional antiferromagnetic spin fluctuations through the mila - rice - shastry hyperfine coupling yield the anisotropy and the site difference @xcite .
the two dimensional spin fluctuation models were successfully applied to account for nmr , neutron scattering and conductivity results @xcite .
we employ the two dimensional nearly antiferromagnetic spin fluctuation model in @xcite .
the dynamical spin susceptibility is expressed by a relaxation mode @xmath133 the leading terms of @xmath134 and @xmath135 in a random phase approximation ( rpa ) are expressed using a long wave length expansion around a specific mode @xmath40 = @xmath136 by @xmath137 and @xmath138 @xmath139 is the inverse of a magnetic correlation length @xmath140 defined around @xmath40 = @xmath136 , @xmath141 @xmath142 is the spin fluctuation amplitude at @xmath40 = @xmath136 and @xmath143 is the characteristic spin fluctuation energy .
the staggered spin susceptibility @xmath144 is then @xmath145 we obtain the plane - site cu nuclear spin - lattice relaxation rate expressed by the leading term of @xmath140 @xmath146 where @xmath147 is the spherical radius of the same volume as the first brillouin zone and @xmath148 1 .
we also obtain the gaussian cu nuclear spin - spin relaxation rate expressed by the leading term of @xmath140 @xmath149 thus , for @xmath106 = 2 we have @xmath150 and @xmath151 in the self - consistent renormalization ( scr ) theory for two dimensional antiferromagnetic spin fluctuations , the curie - weiss behavior of the sqaure of the antiferromagnetic correlation length @xmath152 ( @xmath153 the staggered spin susceptibility @xmath154 ) is reproduced as a function of the distance from the quantum critical point and the spin fluctuation energy @xmath143 @xcite . using @xmath142 = 1/2@xmath155 ( @xmath64 = @xmath156 ) , @xmath143 = 2@xmath157 and @xmath158 = @xmath159
, we obtain @xcite @xmath160 which represents the inverse of the antiferromagnetic correlation length with a unique parameter @xmath161 , and @xmath162 which represents the integrated spin fluctuation weight .
using eqs .
( [ scr1 ] ) and ( [ scr2 ] ) , the nmr relaxation data for tl@xmath0ba@xmath0cuo@xmath4 , yba@xmath0cu@xmath9o@xmath56 and yba@xmath0cu@xmath28o@xmath55 were analyzed @xcite .
cu nmr results of single - layer superconductors hg1201 .
temperature dependences of the plane - site @xmath7cu knight shifts @xmath163 ( top panels ) , nuclear spin - lattice relaxation rates ( 1/@xmath164 ( middle panels ) , and gaussian decay rates ( 1/@xmath165 of nuclear spin - spin relaxation ( bottom panels ) are shown for the underdoped ( left ) , the optimally doped ( center ) and the overdoped samples ( right ) .
dash lines denote the respective @xmath3 s .
the data are reproduced from @xcite . ] the left panels of fig .
[ fig : hg1201f1 ] show @xmath7cu nmr results of the underdoped hg1201 of @xmath3 = 50 k @xcite .
the nmr experiments were performed for the powdered polycrystalline sample .
all the powder samples were magnetically aligned along the @xmath21 axis . in general the mercury compounds
can not be easily aligned by a magnetic field .
then , the nmr experiments were performed for partially oriented powder samples .
this does not indicate that the nmr data are the partially powder - averaged ones .
the sharply aligned nmr lines can be separated in the powder pattern , so that the selected signals surely come from the aligned grains . the plane - site @xmath7cu knight shift @xmath163 , nuclear spin - lattice relaxation rate ( 1/@xmath164 , and gaussian nuclear spin - echo decay rate ( 1/@xmath165 are shown as functions of temperature .
the subscripts of @xmath95 and @xmath94 denote the data in the external magnetic field @xmath166 8 tesla along the @xmath21 and @xmath94 axis , respectively .
no appreciable field dependence was observed within @xmath71 = 4 @xmath167 8 tesla .
the drastic decreases of the cu knight shift @xmath72 and nuclear spin - lattice relaxation rate 1/@xmath15 with cooling down are clearly seen from room temperature .
this is an obvious evidence of the existence of the pseudo spin - gap in the single layer cuprate .
the decrease of the cu knight shift is independently found in @xcite .
the pseudo spin - gap behavior of the uniform spin susceptibility is also found by the in - plnae @xmath168o nmr experiments @xcite . in passing , for the underdoped triple - layer hg1223 ,
the pseudo spin - gap behavior has been observed by cu nmr experiments @xcite .
the absence of the hebel - slichter peak of 1/@xmath15 just below @xmath3 excludes the weak coupling @xmath169-wave pairing symmetry .
the slow decrease of the cu 1/@xmath170 is also seen below 200 k. it suggests that the large pseudo spin - gap leads to the loss of the total weight of the frequency integrated @xmath171 and then that the static staggered spin susceptibility @xmath172 decreases .
the drastic decrease of 1/@xmath15 but the moderate decrease of 1/@xmath170 are theoretically reproduced by the numerical calculations involving the self - energy correction due to the enhanced @xmath173-wave superconducting fluctuations @xcite .
in this theory , the pseudogap is a consequence from the resonance scattering in the two dimensional strong coupling superconductivity .
the finite 1/@xmath170 below @xmath3 is an evidence for the @xmath173 wave pairing symmetry @xcite .
figure [ fig : hg1201f1 ] shows the hole doping dependence of @xmath7cu nmr results of hg1201 @xcite .
temperature dependences of the plane - site @xmath7cu knight shifts @xmath163 ( top panels ) , nuclear spin - lattice relaxation rates ( 1/@xmath164 ( middle panels ) , and gaussian decay rates ( 1/@xmath165 of nuclear spin - spin relaxation ( bottom panels ) are shown for the underdoped ( left ) , the optimally doped ( center ) and the overdoped samples ( right ) .
the pseudo spin - gap temperature @xmath174 is defined by the maximum temperature or the onset of the decrease of ( 1/@xmath164 .
@xmath174 decreases from the underdoped to the overdoped samples .
the temperature region of the curie - weiss behavior of ( 1/@xmath164 above @xmath174 is broadened with the hole doping .
cu nmr results of double - layer superconductors hg1212 .
temperature dependences of the plane - site @xmath7cu knight shifts @xmath163 ( top panels ) , nuclear spin - lattice relaxation rates ( 1/@xmath164 ( middle panels ) , and gaussian decay rates ( 1/@xmath165 of nuclear spin - spin relaxation ( bottom panels ) are shown for the underdoped ( left ) , the optimally doped ( center ) and the overdoped samples ( right ) .
dash lines denote the respective @xmath3 s .
the data are reproduced from @xcite . ]
the middle panels of fig .
[ fig : hg1212f1 ] show @xmath7cu nmr results of the optimally doped hg1212 of @xmath3 = 127 k @xcite , which is the maximum @xmath3 among the ever reported single and double layers .
the plane - site @xmath7cu knight shift @xmath163 , nuclear spin - lattice relaxation rate ( 1/@xmath164 , and gaussian nuclear spin - echo decay rate 1/@xmath170 are shown as functions of temperature .
the decrease of the @xmath7cu knight shift @xmath163 with cooling down is obvious even at the optimal hg1212 . above @xmath174 = 200
k , 1/@xmath15 shows a curie - weiss behavior , because of the development of the antiferromagnetic correlation length .
the decrease of 1/@xmath15 starts below about @xmath174 = 200 k. this @xmath174 = 200 k of hg1212 is larger than @xmath174 = 140 k of the optimally doped bi2212 with @xmath3 = 86 k in @xcite .
the record optimal @xmath3 of bi2212 is about 96 k @xcite . for the double layer systems of hg1212 and bi2212 ,
the ratio ( @xmath167 1.4 ) of the optimal @xmath174 s is nearly the same as that of @xmath3 .
the slight decrease of 1/@xmath170 is also seen , because of the large pseudo spin - gap effect on the frequency integrated @xmath171 .
the finite 1/@xmath170 below @xmath3 indicates the @xmath173 wave pairing symmetry @xcite .
figure [ fig : hg1212f1 ] shows the hole doping dependence of @xmath7cu nmr results of hg1212 @xcite .
temperature dependences of the plane - site @xmath7cu knight shifts @xmath163 ( top panels ) , nuclear spin - lattice relaxation rates ( 1/@xmath164 ( middle panels ) , and gaussian decay rates ( 1/@xmath165 of nuclear spin - spin relaxation ( bottom panels ) are shown for the underdoped ( left ) , the optimally doped ( center ) and the overdoped samples ( right ) .
the pseudo spin - gap temperature @xmath174 decreases from the underdoped to the overdoped samples .
the values of @xmath174 of hg1212 are higher than those of bi2212
the temperature region of the curie - weiss behavior in ( 1/@xmath164 above @xmath174 is broadened with the hole doping .
it should be noted that the overdoped sample shows the curie - weiss behavior in ( 1/@xmath164 but not korringa behavior above @xmath174 @xcite .
cu and @xmath8hg nuclear spin - lattice relaxation rates 1/@xmath15 for optimally doped hg1201 ( a ) and for overdoped hg1201 ( b ) @xcite .
temperature dependence of @xmath175 is nearly the same as that of @xmath176 for both samples . ]
the hg site is located just halfway between the cuo@xmath0 planes and between the respective cu ions .
the hg nuclei can serve as a probe of the interplane coupling .
if the interplane coupling is antiferromagnetic , the in - plane antiferromagnetic correlation is also cancelled out at the hg site .
if the inteplane coupling is uniform and ferromagnetic , the in - plane antiferromagnetic correlation is also seen at the hg site .
figure [ cuvshgt1 ] shows @xmath7cu and @xmath8hg nuclear spin - lattice relaxation rates 1/@xmath15 for the optimally doped hg1201 ( a ) and for the overdoped hg1201 ( b ) @xcite .
the temperature dependence of @xmath175 is nearly the same as that of @xmath177 for both samples .
thus , the inteplane coupling is not antiferromagnetic but uniform .
this uniform interplane coupling is also observed by @xmath7cu and @xmath8hg nmr 1/@xmath15 for hg1212 @xcite .
in contrast to the reports @xcite that @xmath7cu and @xmath8hg nuclear spin - lattice relaxation rates 1/@xmath15 show the different behaviors with each other in hg1201 , these results in fig .
[ cuvshgt1 ] indicate that the interplane coupling is uniform in hg1201 irrespective of the doping level . and superconducting transition temperatures
@xmath3 are plotted against the hole concentration @xmath24 per plane - site cu .
the value @xmath24 of cu@xmath25 is estimated from the excess oxygen concentration @xmath22 and the charge neutrality condition .
solid and dash curves are guides to the eyes . ] figure [ mphase ] shows the magnetic phase diagrams of hg1201(a ) and hg1212 ( b ) @xcite .
pseudo spin - gap temperatures @xmath174 and superconducting transition temperatures @xmath3 are plotted against the hole concentration @xmath24 .
the hatched region of the overdoped hg1201 indicates a pseudo korringa behavior below @xmath174 .
it should be noted that the pseudo spin - gap persists at the optimally doped regions for hg1201 and h1212 .
the doping dependence of @xmath174 of hg1212 is different from that of hg1201 even in the underdoped region .
the @xmath174 as a function of @xmath24 does not seem to be universal .
as siglet - rvb formation temperatures @xmath178 in a two dimensional @xmath158-@xmath179 model with two type fermi surfaces reproduced from @xcite .
@xmath158 is a cu - to - cu transfer integral and @xmath179 is a superexchange interaction in charge - transfer type compounds .
the inset figures are the two type fermi surfaces .
( b ) theoretical hole doping dependences of the mean - field @xmath180 and the true @xmath3 suppressed by @xmath173-wave superconducting fluctuations reproduced from @xcite . ]
we present two theoretical explanations for the doping dependence of @xmath174 . in the two dimensional @xmath158-@xmath179 model with spinon - holon decomposition technique ,
the pseudo spin - gap temperature @xmath174 is regarded as the onset temperature @xmath178 of a spinon singlet rvb ( resonating valence bond ) state @xcite .
the real transition @xmath3 is given by a bose - einstein condensation temperature @xmath181 of holons , leading to the underdoped regime . with an existing approximation , @xmath178 is a second order phase transition temperature .
figure [ theorysg ] ( a ) shows a numerical @xmath178 as a function of doped hole concentration @xcite .
the doping dependence of @xmath178 depends on the contour of a basal fermi surface . for each high-@xmath3 family with different fermi surface , @xmath178 exhibits the different doping dependence . in the two dimensional superconducting fluctuation theory with the strong coupling , the pseudo spin - gap temperature @xmath174
is regarded as the onset of enhancement of @xmath173-wave superconducting fluctuations and a mean - field @xmath3 .
the actual @xmath3 is reduced by the strong superconducting fluctuations so that the underdoped regime appears .
thus , the mean - field @xmath3 is a crossover temperature .
figure [ theorysg ] ( b ) shows the mean - field @xmath182 and the suppressed @xmath3 @xcite .
the doping dependence of @xmath182 depends on the shape of the fermi surface . from the above two theories
, one may conclude that the difference in the hole doping dependence of @xmath174 of hg1201 from hg1212 can be attributed to the different shape in the basal fermi surface of hg1201 and hg1212 .
the band structure calculations have been performed by the full potential linear muffin - tin orbital method for hg1201 , hg1212 and hg1223 @xcite .
the band calculations indicate the different electronic structures and fermi surfaces between hg1201 and the others @xcite .
the fermi surface of a single crystal hg1201 was observed by angle - resolved photoemission measurement @xcite .
( a ) and @xmath183 ( b ) for hg1201 .
the data are reproduced from @xcite . ]
( a ) and @xmath183 ( b ) for hg1212 .
the data are reproduced from @xcite .
solid lines are eye guides for @xmath184 = constant ( a ) and @xmath183 = constant ( b ) . ] the ratios of @xmath15 and @xmath123 tell us the spin fluctuation parameters through eqs .
( [ z1 ] ) , ( [ z2 ] ) , ( [ scr1 ] ) and ( [ scr2 ] ) . in figs . [ z1z2hg1201 ] and [ z1z2hg1212 ] , @xmath184 (
a ) and @xmath183 ( b ) for hg1201 and hg1212 are plotted against temperatures from the underdoped to the overdoped samples @xcite .
the pseudo scaling of @xmath184 = constant is observed in the limited temperature region for hg1212 .
the pseudo scaling temperature region decreases and shifts at lower temperatures by the hole doping . at high temperatures , @xmath183 = constant holds .
the value of @xmath183 indicates the product of the spin fluctuation amplitude and the spin fluctuation energy .
thus , the product of @xmath185 decreases from the underdoped to the overdoped regimes of hg1201 and hg1212 .
equation ( [ z1 ] ) indicates that @xmath184 is proportional to the inverse of the antiferromagnetic correlation length @xmath140 and has a unique scale parameter @xmath161 . in figs .
[ z1z2hg1201 ] and [ z1z2hg1212 ] , @xmath184 decreases from the underdoped to the overdoped samples of hg1201 and hg1212 .
this is inconsistent with the theoretical doping dependence of the magnetic correlation length near the two dimensional quantum critical point .
the @xmath186 in the scr theory increases from the weakly antiferromagnetic to the nearly antiferromagnetic regimes and away from the quantum critical point @xcite .
the decrease of @xmath184 suggests the decrease of the spin fluctuation energy @xmath161 from the underdoped to the overdoped samples . in the overdoped regime ,
the decrease of @xmath3 can be associated with the decrease of the spin fluctuation energy @xmath161 . in the rpa for the two dimensional @xmath158-@xmath179 model , both @xmath184 ( a ) and @xmath183 ( b ) depend on temperature more or less @xcite . the value of @xmath184 increases with the hole doping @xcite , that is inconsistent with the experimental doping dependence .
but the decreases of @xmath183 with doping is reproduced within the rpa calculations . in the numerical calculations for the small size @xmath158-@xmath179 model using the lanczos diagonalization method , the doping dependence of @xmath184 below @xmath12 = 1390
k agrees with the experimental tendency . above @xmath12
= 2320 k , the doping dependence of @xmath184 above @xmath12 = 2320 k reproduces those of the scr and the rpa calculations @xcite . are plotted against the spin fluctuation parameters of hg1201 , hg1212 and tl1212 ( tlsr@xmath0cacu@xmath0o@xmath17 ) .
pseudo spin - gap temperatures @xmath174 are also plotted .
the cu nmr results of hg1201 and hg1212 are reproduced from @xcite .
the cu nmr results of tl1212 are reproduced from @xcite . ] within the framework of eliashberg - nambu strong coupling superconductivity theory , the actual @xmath3 is determined by competition between the pairing effect and the depairing effect @xcite . in the spin - fluctuation - mediated superconductors , the depairing effect due to low frequency spin fluctuations competes the paring effect due to high frequency ones . for the @xmath173 superconductivity on a square lattice ,
the numerical calculation and theoretical consideration tell us that @xmath3 is proportional to the characteristic energy scale of antiferromagnetic spin fluctuations @xcite , @xmath187 the spin - fluctuation - induced superconductivity theories are studied for the two and three dimensional hubbard models @xcite . as a general tendency , two dimensional systems have higher @xmath3 than three dimensional ones @xcite .
then , the mystery of the high @xmath3 of the layered compounds is traced back to the two dimensionality and the large scale of the antiferromagnetic spin fluctuation energy .
figure [ tcvssf ] shows @xmath3 plotted against @xmath183 of hg1201 , hg1212 , and tl2212 reproduced from @xcite . if one assumes @xmath188 of eq .
( [ z2 ] ) , the linear relation between @xmath3 and the spin fluctuation energy does not seem to hold .
the spin fluctuation product @xmath185 decreases monotonically with the hole doping as in figs .
[ z1z2hg1201 ] and [ z1z2hg1212 ] .
this has been recognized in the other systems @xcite .
thus , @xmath183 is a monotonic function of the doped hole concentration and then the indicator . at some threshold value of @xmath183 in the overdoped regime
, @xmath3 starts to increase toward the optimally doping level . beyond the optimally doping level , @xmath3 decreases and @xmath174 increases as the hole concentration is reduced .
the antiferromagnetic spin fluctuation spectrum is different between the underdoped and the overdoped samples .
thus , the overdoped part of the @xmath3-vs-@xmath183 curves indicates some correlation between @xmath3 and the spin fluctuation parameters , but the underdoped part indicates the suppression of @xmath3 due to the grown of the large pseudo spin - gap . , @xmath189 , and @xmath190 for hg1201 ( black symbols ) and hg1212 ( red symbols ) from figs .
[ fig : hg1201f1 ] and [ fig : hg1212f1 ] @xcite . ] from the fact that the uniform spin susceptibility @xmath11 is suppressed at low temperatures in yba@xmath0cu@xmath9o@xmath191 and yba@xmath0cu@xmath20o@xmath192 more than la@xmath18sr@xmath193cuo@xmath28 , the bilayer ( double - layer ) coupling effect had been proposed to be the primary origin of the pseudo spin - gap @xcite .
the double - layer exchange scattering effect was propsed to account for the different behavior of 1/@xmath15 and 1/@xmath123 below @xmath174 @xcite .
now the single cuo@xmath0 layer hg1201 is found to possess the pseudo spin - gap in the low - lying excitation spectrum . however , the existence of the double - layer coupling was actually confirmed by the neutron scattering @xcite and nmr @xcite experiments for low doped yba@xmath0cu@xmath9o@xmath4 .
spin - echo double resonance techniques have been applied to estimate the double - layer coupling constant in y@xmath0ba@xmath28cu@xmath56o@xmath194 @xcite and in bi@xmath0sr@xmath0ca@xmath0cu@xmath9o@xmath195 @xcite .
the angle dependence of the gaussian decay rate 1/@xmath123 was measured to estimate the like - spin interlayer coupling for hg1223 @xcite .
figure [ hg1201vshg1212 ] shows @xmath196 , @xmath189 , and @xmath190 for hg1201 ( black symbols ) and hg1212 ( red symbols ) from figs .
[ fig : hg1201f1 ] and [ fig : hg1212f1 ] to compare two systems @xcite .
the nmr results of hg1212 is quantitatively different from those of hg1201 . as to the underdoped hg1201 and hg1212 samples
, @xmath189 of hg1212 is nearly the same as that of hg1201 but the @xmath3 is about twice higher than hg1201 .
the different point is the underdoped @xmath190 of hg1212 higher than that of hg1201 .
thus , not the low frequency spin fluctutions but the high frequency ones contribute the higher @xmath3 . as to the optimally doped hg1201 and hg1212 samples ,
the spin part of the cu knight shift of hg1212 and @xmath189 are smaller than those of hg1201 , whereas the @xmath190 is higher than that of hg1201 .
the electron spin - spin correlation function @xmath197 due to the double - layer coupling is competitive in the in - plane correlation function @xmath198 of @xmath196 and @xmath189 but is additive in @xmath190 @xcite .
although the double - layer coupling is not a primary origin of the pseudo spin - gap , it surely affects the microscopic magnetic properties .
the flat cuo@xmath0 plane and the large pseudo spin - gap are the characteristics of the mercury - based high-@xmath3 superconducting cuprates hg1201 and hg1212 .
the role of the pseudo spin - gap in the higher @xmath3 is still unclear .
is the pseudo spin - gap a consequence of the enhanced superconducting fluctuations to suppress the mean field @xmath199 ?
then , we should explore the layered compounds with higher energy spin fluctuations and any method to suppress the superconducting fluctuations to get higher @xmath3 .
we would like to thank s. adachi , a. yamamoto , a. fukuoka , k. tanabe , n. koshizuka , k. yoshimura , and h. yasuoka for the fruitful collaboration , and y. ohashi , j. kishine , and y. yanase for the valuable discussions on theoretical studies .
takigawa m. ; reyes a. p. ; hammel p. c. ; thompson j. d. ; heffner r. h. ; fisk z. ; ott k. c. _ phys .
_ 1991 , _ b43 _ , 247 - 257 .
pennington c. h. ; slichter c. p. in _ physical properties of high temperature superconductors _ ; ginsberg , d. m. ; ed . ;
world scientific publishing co. : new jersey , 1998 ; vol .
2 , pp 269 - 367 .
itoh y. cond - mat/0711.1688 .
fukuoka a. ; tokiwa - yamamoto a. ; itoh m. ; usami r. ; adachi s. ; tanabe k. _ phys .
_ 1997 , _ b 55 _ , 6612 - 6620 .
radaelli p. g. ; marezio m. ; perroux m. ; de brion s. ; tholence j. l. ; huang q. ; santoro a. _ science _ 1994 , _ 265 _ , 380 - 383 .
radaelli p. g. ; marezio m. ; tholence j. l. ; de brion s. ; santoro a. ; huang q. ; capponi j. j. ; chaillout c. ; krekels t. ; van tendeloo g. _ j. phys .
chem . solids _ 1995 , _ 56 _ , 1471 - 1478 .
tokiwa - yamamoto a. ; tatsuki t. ; adachi s. ; tanabe k. _ physica _ 1996 , c 268 , 191 - 196 .
yamamoto a. ; hu w .- z . ; tajima s. _ phys . rev .
_ 63 _ , 024504 - 1 - 6 .
wagner j. l. ; radaelli p. g. ; hinks d. g. ; jorgensen j. d. ; mitchell j. f. ; dabrowski b. ; knapp g. ; beno m. _ physica _ 1993 , _ c 210 _ , 447 - 454 .
loureiro s. m. ; antipov e. v. ; tholence j. l. ; capponi j. j. ; chmaissem o. ; huang q. ; marezio m. _ physica _ 1993 , _ c 217 _ , 253 - 264 .
antipov e. v. ; capponi j. j. ; chaillout c. ; chmaissem o. ; loureiro s. m. ; marezio m. ; putilin s. n. ; santoro a. ; tholence j. l. _ physica _ 1993 , _ c 218 _ , 348 - 355 .
huang q. ; lynn j. w. ; meng r. l. ; chu c. w. _ physica _ 1993 , _ c 218 _ , 356 - 364 .
wagner j. l. ; hunter b. a. ; hinks d. g. ; jorgensen j. d. _ phys .
_ 1995 , _ 51 _ , 15407 - 15414 .
radaelli p. g. ; wagner j. l. ; hunter b. a. ; beno m. a. ; knapp g. s. ; jorgensen j. d. ; hinks d. g. _ physica _ 1993 , _ 216 _ , 29 - 35 .
ohta t. ; izumi f. ; tokiwa - yamamoto a. ; tanabe k. ; hewat a. w. _ physica _ 1997 , _ c 281 _ , 228 - 236 .
daignere a. ; wahl a. ; hardy v. ; maignan a. _ physica _ 2001 , _ c 349 _ , 189 - 195 .
ogawa a. ; sugano t. ; wakana h. ; kamitani a. ; adachi s. ; tarutani y. ; tanabe k. _ jpn . j. appl .
_ 2004 , _ 43 _ , no . 1a / b , l 40 - 43 .
ogawa a. ; sugano t. ; wakana h. ; kamitani a. ; adachi s. ; tarutani y. ; tanabe k. j. appl .
phys . 2005 , _ 97 _ , 013903 - 1 - 9 .
moriya t. ; takahashi y. ; ueda k. _
_ 1990 , _ 59 _ , 2905 - 2915 . itoh y. ; machi t. ; kasai c. ; adachi s. ; koshizuka n. ; murakami m. _ phys .
_ b 67 _ , 064516 - 1 - 8 .
gippius a. a. ; antipov e. v. ; hoffmann w. ; luders k. _ physica _ 1997 , _ c 276 _ , 57 - 64 .
s. ohsugi , t. tsuchiya , t. koyama , k. fueki , _ j. of low temp .
_ 1996 , _ 105 _ , nos .
3/4 , 419 - 423 .
julien m .- h . ; horvati m. ; carretta p. ; berthier c. ; berthier y. ; sgransan p. ; loureiro s. m. ; capponi j .- j . _
physica _ 1996 , _ c 268 _ , 197 - 204 .
singh d. j. _ physica _ 1993 , _ c 212 _ , 228 - 232 .
itoh y. ; machi t. ; adachi s. ; tokiwa - yamamoto a. ; tanabe k. ; koshizuka n. in _ advances in superconductivity xi _ ; proceedings of the 11th international symposium on superconductivity ( iss 98 ) ; springer - verlag : tokyo , 1999 , pp 61 - 64 . | mercury - based cuprates hgba@xmath0cuo@xmath1 with 0 @xmath2 0.2 ( hg1201 ) are the superconductors with a single cuo@xmath0 layer in unit cell and the optimally oxidized one has the highest @xmath3 = 98 k among the ever reported single - cuo@xmath0-layer superconductors .
double cuo@xmath0 layered cuprates hgba@xmath0cacu@xmath0o@xmath4 with 0.05 @xmath2 0.35 ( hg1212 ) have the highest @xmath3 = 127 k at the optimal oxygen concentration .
this is the highest @xmath3 among the ever reported double - cuo@xmath0-layer superconductors .
the hg1201 has the nearly perfect fat cuo@xmath0 plane .
the hg1212 has the flattest cuo@xmath0 plane among the other lower @xmath3 double - layer cuprates , which is associated with the mystery of the highest @xmath3 .
both systems have a pseudo spin - gap in the magnetic excitation spectrum of the normal states . in this article , we present the microscopic studies of magnetic and electric properties of the hg - based superconducting cuprates using nuclear magnetic resonance ( nmr ) and nuclear quadrupole resonance ( nqr ) techniques .
nmr and nqr are powerful to detect local information through the nuclear sites in materials and have supplied us with information on low frequency magnetic response of electronic systems .
although the structure analysis indicates the flat cuo@xmath0 planes , zero field @xmath5cu nqr spectra , which are sensitive to local electric charge distribution , show inhomogeneous broadening .
the local electrostatic states are rather inhomogeneous .
although the @xmath6-wave superconductivity must be fragile to imperfection and non - magnetic impurities , the pure hg - based superconducting cuprates show impure @xmath5cu nqr spectra but rather robust pseudo spin - gap in the @xmath7cu nmr knight shift and nuclear spin - lattice relaxation rate over the wide doping regions .
there had been an issue whether the pseudo spin - gap results from a double - layer coupling or a single layer anomaly .
the nmr results for hg1201 served as the evidence for the existence of the single - layer pseudo spin - gap .
the pseudo spin - gap is explained by a precursory phenomena of superconducting pairing fluctuations or spin singlet correlation .
the different doping dependence of the pseudo spin - gap of hg1201 and hg1212 is associated with the different fermi surface contour .
the similar temperature dependence of the @xmath8hg and the @xmath7cu nuclear spin - lattice relaxation times indicates uniform interlayer coupling .
the double - layer coupling effect is revisited through the comparison of hg1201 and hg1212 . |
this paper is a companion to @xcite , which established the algebraic generating function for @xmath0-avoiding permutations conjectured by vaclav kotesovec @xcite . in similar vein
, vladimir kruchinin @xcite has conjectured the @xmath1 for @xmath2-avoiding permutations , where @xmath3 denotes the for the catalan numbers .
we will show that do indeed have this .
first , we use the combinatorial interpretation of the invert transform to reduce the problem to counting _ indecomposable _ .
then we exhibit a bijective mapping from the set of indecomposable of length @xmath4 to the union of cartesian products @xmath5 , where @xmath6 is the set of indecomposable 321-avoiding permutations of length @xmath7 and @xmath8 is the set of integer sequences @xmath9 satisfying @xmath10 and @xmath11 for @xmath12 .
the counting sequences for the sets @xmath13 and @xmath8 are known , and the result follows readily .
section 2 recalls the notion of indecomposability and the application of the invert transform to indecomposable permutations .
section 3 reviews nonnegative lattice paths and integer sequences whose successive entries increase by at most 1 , that is , elements of @xmath8 .
section 4 defines some notions relevant for our bijection .
section 5 presents the main bijection and section 6 explains why it works .
section 7 ties everything together .
a _ standard _ permutation is one on an initial segment of the positive integers and to _ standardize _ a permutation on a set of positive integers means to replace its smallest entry by 1 , next smallest by 2 and so on , thereby obtaining a standard permutation . in the context of pattern avoidance , we consider standard permutations written in one - line form ( that is , as lists ) . when a standard permutation is written in two - line form
, it may be possible to insert some vertical bars to obtain subpermutations , not necessarily standard , as in @xmath14 .
after inserting the largest possible number ( 0 or more ) of such bars , as in @xmath15 , we obtain the _ components _ of the permutation , here 312 , 654 , 7 .
a permutation is _ indecomposable _ if it has exactly one component .
( thus the permutation 1 is indecomposable but the empty permutation is not . ) let @xmath16 denote the for and @xmath17 the for indecomposable .
clearly , a permutation avoids @xmath2 if and only if each of its components does so .
hence , the combinatorial interpretation of the invert transform ( see @xcite for example ) implies that @xmath18 and our problem is reduced to showing that @xmath19 .
it is well known that the `` ballot number '' @xmath20 @xmath21 ( with @xmath22 if @xmath23 and @xmath24 if @xmath25 ) counts nonnegative paths of @xmath26 upsteps @xmath27 and @xmath4 downsteps @xmath28 , where nonnegative means the path never dips below , the horizontal line through its initial vertex ( see , e.g. , @xcite ) .
a nonnegative path of @xmath4 upsteps and @xmath4 downsteps is a _ dyck _ path and its _ size _ is @xmath4 .
a nonempty dyck path is _ indecomposable _ if its only return to is at the end . the returns to split a dyck path into its ( indecomposable ) components .
the number of indecomposable dyck paths of size @xmath4 is @xmath29 ( delete the first and last steps to obtain a one - size - smaller dyck path ) .
@xmath30 [ count1 ] the number of indecomposable 321-avoiding permutations on @xmath31 $ ] is @xmath29 .
one method is to observe that krattenthaler s bijection @xcite from 321-avoiding permutations on @xmath31 $ ] to dyck paths of size @xmath4 preserves components in the obvious sense and so sends indecomposable permutations to indecomposable paths . given a nonnegative path , successively delete the first peak ( @xmath32 ) recording its height above until no peaks remain , as in figure 1 ( heights prepended to existing list ) .
= 0.4 cm @xmath33 the path on the left produces the list of heights 3442 and this is a map from nonnegative paths of @xmath26 @xmath34s and @xmath4 @xmath35s to @xmath36 . to reverse the map ,
suppose given @xmath37 .
start with a path of @xmath38s . then successively insert a peak at height @xmath39 into the initial ascent of the current path so that its top vertex is at height @xmath40 .
this produces a path whose first peak is at height @xmath39 and the growth condition @xmath41 is just what is needed to enable the next step .
so the map is a bijection and we have @xmath30 [ count2 ] @xmath42 .
the preceding construction will be mirrored in section [ whyworks ] below when we insert a `` peak '' entry into a permutation so that if it has `` height '' @xmath43 , then there are @xmath44 possibilities for the next insertion .
by ( slight ) abuse of language , to delete an entry @xmath45 $ ] from a permutation @xmath46 on @xmath31 $ ] means to erase @xmath47 and then subtract 1 from each entry @xmath48 ; @xmath49 denotes the resulting permutation .
the non-@xmath47 entries of @xmath46 correspond in an obvious way to the entries of @xmath49 .
conversely , to insert @xmath47 in position @xmath50 means to increment by 1 each entry @xmath51 and then place @xmath47 in position @xmath50 ; we use @xmath52 to denote the result .
thus , for @xmath53 .
again , the entries of @xmath46 correspond naturally to the non-@xmath47 entries of @xmath54 . the adjective / noun is short for left - to - right maximum in a permutation .
* _ henceforth , for brevity , we use the unadorned term `` avoider '' to mean an indecomposable @xmath55-avoider . _ * a _
key-2 _ entry in an avoider is an entry that serves as the `` 2 '' in either a 321 pattern or a 4312 pattern .
for example , the key-2 entries in 6174235 are 4 and 3 .
clearly , an avoider with no key-2 entries is a 321-avoider .
the term key-2 is mnemonic but somewhat ungainly and to add a little color , * we will refer to a key-2 entry as a _
entry*. the _ peak _ entry in an avoider that contains ( one or more ) 321 patterns is the larger of the last `` 1 '' of a 321 and its ( immediate ) predecessor . the terminology is justified because the peak is indeed a : suppose @xmath56 is the last `` 1 '' of a 321 in an avoider , say the last `` 1 '' of @xmath57 , and @xmath47 is the predecessor of @xmath56 .
if @xmath58 , then @xmath56 is the `` 2 '' of the 4312 pattern @xmath59 . if @xmath60 , then @xmath61 or @xmath62 . in the former case , @xmath47 is the `` 2 '' of the 321 pattern @xmath57 ; in the latter case , @xmath63 ( else @xmath59 is a proscribed 3241 ) and so @xmath47 is the `` 2 '' of the 321 pattern @xmath64 .
see figure 2 for some examples of avoiders with blue entries so colored and peak blue entry highlighted .
we let @xmath65 denote the set of avoiders ( indecomposable @xmath66-avoiding permutations ) of length @xmath4 , and similarly @xmath67 is the set of indecomposable @xmath68-avoiding permutations . set @xmath69 .
for @xmath70 , there is a bijection @xmath71 here is its description .
suppose given @xmath72 .
if @xmath73 , then @xmath46 is already 321-avoiding and @xmath74 is the image pair , where @xmath75 denotes the empty list . if @xmath76 , the idea is to successively delete the ( current ) peak entry recording , in the same right to left fashion as in figure 1 , its `` height '' , appropriately defined , until a 321-avoider @xmath77 is obtained .
then the image pair is @xmath78 , where @xmath79 is the list of heights .
the trick is to find the correct definition of height , and it s a doozy . to this end ,
associate to each 321-containing permutation @xmath46 a triple @xmath80 , all integers except that @xmath81 may be infinite : @xmath56 is the last `` 1 '' of a 321 in @xmath46 , @xmath82 is the rightmost entry to the left of @xmath56 that exceeds @xmath56 , and @xmath81 is the first non - lrmax entry after @xmath56 ( with @xmath83 if there is no such lrmax ) .
thus , for @xmath84 , we have @xmath85 and for @xmath86 , we have @xmath87 . [ w ] if @xmath46 is a @xmath68-containing avoider with associated triple @xmath88 , then there is an entry @xmath89 in @xmath46 such that @xmath90 is a @xmath68 pattern in @xmath46 . proof .
since @xmath56 is the `` 1 '' of a 321 , there is a sublist @xmath91 in @xmath46 with @xmath92 . by definition of @xmath82
, @xmath93 must lie weakly to the left of @xmath82 and so @xmath94 . if @xmath95 , take @xmath96 .
otherwise , @xmath97 and @xmath93 must both be @xmath98 and @xmath99 is a forbidden 3241 pattern .
if @xmath46 is a @xmath68-containing avoider with associated triple @xmath88 and @xmath81 is finite , then @xmath100 . proof . if not , @xmath101 would be a 321 , violating the definition of @xmath56 as the last `` 1 '' of a 321 . for reasons to become clear , we call the disjoint union @xmath102 \cup [ c+1,n]$ ] the _ peak - insertion set _ for @xmath46 where @xmath103=\emptyset$ ] if @xmath104 .
furthermore , for a 321-avoiding permutation on @xmath31 $ ] , set @xmath105 and define its peak - insertion set to be @xmath103=[2,n]$ ] ( the @xmath56 and @xmath82 evaporate in this case ) .
next , we arrange the peak - insertion set of an avoider @xmath46 into a suitably ordered list , called the _ peak - insertion list _ of @xmath46 . taken left to right , the entries @xmath106 of @xmath46 form a list @xmath107 and the non - entries @xmath108 form a list @xmath109 .
thus @xmath110 $ ] .
obviously , @xmath107 is an increasing list , and so is @xmath109 for otherwise , in the 321-containing case , @xmath56 would not be the last `` 1 '' of a 321 , and in the 321-avoiding case , a 321 would actually be present . split @xmath107 into maximal runs of consecutive integers @xmath111 .
likewise , split @xmath109 into maximal runs of consecutive integers but this time written as @xmath112 , where @xmath113 is the first entry of the @xmath50-th run and @xmath114 may be empty .
there is the same number of runs in @xmath107 as in @xmath109 because ( i ) the smallest run contains @xmath115 and comes from @xmath109 since @xmath81 is not a , ( ii ) thereafter the runs alternate between @xmath107 and @xmath109 , and ( iii ) the largest run contains @xmath4 , a , and so comes from @xmath107 . the _ peak - insertion list _ of @xmath46 is now defined to be the peak - insertion set of @xmath46 listed in the following order ( @xmath116 denotes the reversal of the list @xmath79 ) : @xmath117 where the terminal segment starting at @xmath82 is omitted if @xmath46 is 321-avoiding .
note that @xmath115 is missing and the list consists of @xmath102 \cup [ c+1,n]$ ] ( or @xmath103 $ ] in the 321-avoiding case ) , as it should .
for example , for the avoider shown in matrix form in figure 3 , we have @xmath118 and runs in @xmath107 and @xmath109 as follows .
@xmath119 a_{i } & = & 13 \ \,14 \ \ , 15 & \
\,17\ \ , 18 & \ \,22 \
\,23 \ \ , 24 \
\ , 25 \\
b_{i}|b_{i } & = & 10 \ | \ 11 \ \,12 & \ \,16 \ |\ \eps & \ \,19 \ | \ 20 \ \,21 \end{array}\ ] ] here , @xmath120 and the ordering in the peak - insertion list is @xmath121 we can now define the _ height _ of the peak entry @xmath47 in a 321-containing avoider @xmath46 : it is the position of @xmath47 in the peak - insertion list of @xmath49 .
( we will see later that @xmath47 must be in the peak - insertion set of @xmath49 . )
there is a graphical way to visualize the ordering in the peak - insertion list . as illustrated in figure 3 , for each @xmath47 in the peak - insertion set @xmath102 \cup [ c+1,n]$ ] of @xmath46 , insert a bullet at vertex @xmath122 in the matrix diagram of @xmath46 where the abscissa @xmath123 is determined as follows . for @xmath124 , insert the bullet as far right as possible so that the region below and to the right of the bullet is nonempty . for @xmath125 $ ] ,
let @xmath50 denote the position of @xmath56 in @xmath46 . then , for @xmath126,\ x = i+1 $ ] , and for @xmath127 .
if the bullets are arranged in order of distance from the vertical line @xmath128 and , for bullets at the same distance from @xmath128 , in order of distance from the horizontal line @xmath129 or @xmath130 if @xmath104 ( heavy line in figure 3 ) , then their @xmath47 s form the peak - insertion list . the mapping is illustrated in figure 4 , which shows that @xmath131 .
we need to establish several facts to show the map does all it claims to and is invertible .
suppose @xmath47 is the peak entry of a @xmath68-containing avoider @xmath46 .
then @xmath132i@xmath133 @xmath49 is again an avoider , @xmath132ii@xmath133 the blue entries of @xmath46 other than @xmath47 become the entries of @xmath49 , and @xmath132iii@xmath133 @xmath47 is in the peak - insertion set of @xmath49 . proof .
( i ) @xmath49 inherits the @xmath2-avoiding property from @xmath46 . if @xmath49 was decomposable then the entries other than @xmath47 in the 321 or 4312 pattern containing @xmath47 in @xmath46 would correspond to entries in the same component of @xmath49 .
but then @xmath46 would also be decomposable , obviously in the 321 case , and because the `` 1 '' and `` 2 '' can be chosen adjacent in the 4312 case .
( ii ) no new blue entry can be introduced and no non - peak blue entry is lost because if the deleted entry @xmath47 is the `` 1 '' of a 321 , then the `` 2 '' , a blue entry in @xmath46 , is still the `` 2 '' of a 321 in @xmath49 since the predecessor of @xmath47 in @xmath46 is @xmath134 and so serves as a `` 1 '' in place of @xmath47 .
also , the peak blue entry can not possibly be the `` 4 '' , `` 3 '' , or `` 1 '' of a 4312 , so no blue entry in @xmath46 that is the `` 2 '' of a 4312 loses its blue status in @xmath49 .
( iii ) this will be proved in contrapositive form in proposition [ bigone ] below .
[ misc ] suppose @xmath46 is a @xmath68-containing avoider with associated triple @xmath88 .
@xmath132i@xmath135 if @xmath81 is finite , then all entries after @xmath56 in @xmath46 are @xmath108 .
@xmath132ii@xmath135 if @xmath104 or @xmath81 is finite and @xmath136 , then @xmath137 lies to the left of @xmath82 in @xmath46 .
@xmath132iii@xmath135 if @xmath104 , then @xmath56 is the last entry of @xmath46 .
@xmath132iv@xmath135 suppose @xmath138 is an entry of @xmath46 .
then @xmath139 is a in @xmath46 provided @xmath139 lies to the left of @xmath81 in @xmath46 in case @xmath81 is finite . proof .
( i ) all entries after @xmath81 are @xmath140 else @xmath56 would not be the last `` 1 '' of a 321 . if the assertion fails , take @xmath47 to be the rightmost offending entry in @xmath46 .
clearly , @xmath47 lies between @xmath56 and @xmath81 in @xmath46 and @xmath141 and region @xmath142 in the schematic of figure 5 is empty because @xmath47 is the rightmost offender .
also , @xmath143 is empty since @xmath47 is a ( by definition of @xmath81 ) , @xmath144 is empty else @xmath56 is the `` 2 '' of a 321 , and @xmath145 is empty else @xmath81 is the `` 2 '' of a 321 .
these empty regions force @xmath47 to be a fixed point and @xmath46 to be decomposable .
\(ii ) first , @xmath137 can not lie between @xmath82 and @xmath56 in @xmath46 by definition of @xmath82 . if @xmath104 , we are done by part ( i ) .
so suppose @xmath81 is finite .
the entry @xmath146 whose existence is guaranteed by prop.[w ] implies that @xmath137 is not a , violating the definition of @xmath81 if @xmath137 lies between @xmath56 and @xmath81 . if @xmath137 lies to the right of @xmath81 , then @xmath147 is the `` 21 '' of a 321 ( since @xmath81 is not a ) , contradicting the assumption that @xmath56 is the last `` 1 '' of a 321 .
\(iii ) if not , then all entries after @xmath56 would be entries , and the last entry would be @xmath4 , violating indecomposability .
\(iv ) suppose @xmath138 is an offender .
if @xmath104 , @xmath139 lies to the left of @xmath82 by part ( iii ) and the definition of @xmath82 . if @xmath81 is finite , @xmath139 lies to the left of @xmath56 by definition of @xmath81 , and so lies to the left of @xmath82 by definition of @xmath82 . in either case , @xmath139 is a non - lying to the left of @xmath82 .
then @xmath148 is the `` 321 '' of a forbidden 4321 . for each @xmath47 in the peak - insertion set of an avoider @xmath46 on @xmath31 $ ]
, there is exactly one position @xmath50 such that @xmath149 @xmath132insertion of @xmath47 at position @xmath50@xmath133 satisfies @xmath132i@xmath133 @xmath77 is an avoider , @xmath132ii@xmath133 the peak entry of @xmath77 is @xmath47 , and @xmath132iii@xmath133 @xmath77 has just one more blue entry than @xmath46 .
also , for @xmath47 not in the peak - insertion set of @xmath46 , there is no such @xmath50 . proof .
first , suppose @xmath150 $ ] .
let @xmath139 be the rightmost entry of @xmath46 that is @xmath151 .
insert @xmath47 immediately to the left of @xmath139 .
suppose @xmath46 has the matrix form depicted schematically in figure 6 where the bullet represents the inserted entry and @xmath139 its successor .
then @xmath145 is empty ( contains no entries ) by definition of @xmath47 .
if @xmath143 were also empty , @xmath46 would be decomposable .
thus @xmath47 is the `` 2 '' of a 321 , making @xmath47 blue in @xmath152 and , clearly , it is the peak blue entry . on the other hand ,
if @xmath47 is inserted to the right of @xmath139 it will not be blue , and if inserted to the left of @xmath139 but not adjacent to @xmath81 , it may be blue but will not be the peak blue . now suppose @xmath125 $ ] .
if @xmath153 , insert @xmath47 just before @xmath56 ( @xmath47 will be the `` 2 '' of a 321 ) , and if @xmath126 $ ] , insert @xmath47 just after @xmath56 ( @xmath47 will be the `` 2 '' of a 4312 ) .
similar considerations show that , for this insertion point , @xmath47 will be the peak blue entry in @xmath152 and the only new blue entry .
also no other insertion point will do .
as for the last assertion , if @xmath46 is 321-avoiding , the peak - insertion set is @xmath154 $ ] and 1 can not be a blue entry in @xmath155 because , by definition , blue entries exceed 1 .
now suppose @xmath46 is 321-containing with associated @xmath156 .
if @xmath157 is inserted to the left of @xmath56 , it can not be the larger of the last `` 1 '' of a 321 and its predecessor in @xmath152 ; if inserted to the right of @xmath56 , a descending quadruple is present in @xmath152 .
next , suppose @xmath158 $ ] ( @xmath159 in case @xmath104 ) .
if @xmath47 is inserted to the left of @xmath56 , then it is not the larger of @xmath56 and the predecessor of @xmath56 unless it actually is the predecessor of @xmath56 , but in that case @xmath160 is a forbidden 3241 by lemma [ misc](ii ) ; if @xmath47 is inserted between @xmath56 and @xmath81 in case @xmath81 is finite or after @xmath56 in case @xmath104 , @xmath47 can not be the `` 1 '' of a 321 in @xmath152 by lemma [ misc](iv ) and so @xmath47 is certainly not the larger of the last `` 1 '' of a 321 and its predecessor in @xmath152 ; if @xmath47 is inserted after @xmath81 , then it is the last `` 1 '' of a 321 in @xmath152 , but is not larger than its predecessor and so is not peak .
[ bigone ] for an avoider @xmath46 , as @xmath47 ranges from left to right over the peak - insertion list of @xmath46 , the length of the peak - insertion list of @xmath161 ranges from left to right over the interval @xmath162 , where @xmath7 denotes the length of the peak - insertion list of @xmath46 and @xmath163 is the @xmath50 of the preceding proposition . proof .
recall that every avoider @xmath46 is associated with an @xmath88 triple if it is 321-containing and with a singleton @xmath105 otherwise , and the peak - insertion set for @xmath46 is @xmath102 \cup [ c+1,n]$ ] with the first interval absent if @xmath46 is 321-avoiding and the second interval absent if @xmath104 .
we need to determine the triples , denoted @xmath164 , for each @xmath161 with @xmath47 in the peak - insertion list of @xmath46 . in the peak - insertion list of @xmath46 ,
the entries @xmath140 all occur before the entries @xmath165 . split the entries @xmath140 into segments consisting of increasing runs as illustrated in figure 7 for an avoider on @xmath31=[25]$ ] with @xmath118 and peak - insertion set @xmath102 \cup [ c+1,n ] = [ 5,8 ] \cup [ 11,25]$ ] .
then for @xmath124 in the peak - insertion list of @xmath46 , the @xmath166 and @xmath167 of @xmath161 are as follows : ( i ) @xmath168 is one less than the smallest entry in the segment containing @xmath47 , ( ii ) @xmath169 , and ( iii ) @xmath167 is one more than the largest entry in the segment containing @xmath47 ( or @xmath170 if this largest entry is @xmath4 ) . for @xmath63 in the peak - insertion list of @xmath46 ,
the @xmath171 is @xmath137 , the @xmath167 is @xmath172 , while @xmath173 for @xmath174 $ ] and @xmath175 for @xmath61 .
we leave the reader to verify the truth of these assertions with the visual aid that each pair @xmath176 is shown as a bullet at vertex @xmath177 in figure 7 .
so , `` expanding '' the @xmath176 bullet into a cell containing the entry @xmath47 gives the matrix diagram of @xmath178 .
it is now clear that the size of the peak - insertion set ( sps in figure 7 ) starts at 2 and increases by 1 thereafter as @xmath47 ranges across the peak - insertion list of @xmath46 .
from propositions [ count1 ] and [ count2 ] and the preceding bijection , we find that the number @xmath179 of indecomposable @xmath2-avoiding permutations of length @xmath4 is given by @xmath180 and , for @xmath181 , @xmath182 this is sequence in oeis @xcite with @xmath183 where @xmath184 is the for the catalan numbers , and by section [ indecinvert ] , the claimed for @xmath2-avoiding permutations follows .
the counting sequence for @xmath185-avoiders can be succinctly described as the invert transform of the catalan transform of the catalan numbers . | we show that permutations avoiding both of the ( classical ) patterns 4321 and 3241 have the algebraic conjectured by vladimir kruchinin .
[ theorem]definition [ theorem]lemma [ theorem]proposition [ theorem]corollary permutations avoiding 4321 and 3241 have an algebraic generating function david callan + department of statistics , university of wisconsin - madison + 1300 university avenue , madison , wi 53706 - 1532 + * [email protected] * + june 12 , 2013 |
in qcd , the chiral condensate is important as it is the order parameter of chiral symmetry breaking . as such , its behavior is key in understanding qcd in extreme conditions .
finite temperature and pressure effects have been studied extensively . on the other hand ,
comparatively few efforts have focused on the effects of finite electromagnetic fields .
prior studies have been done either using chiral perturbation theory ( @xmath5pt)@xcite , the effective theory for low - energy qcd , or models compatible with the large-@xmath6 expansion@xcite ( the njl model , in particular@xcite ) .
the method used in this paper is @xmath5pt .
this approach has the advantage of being model - independent and systematic , but has the disadvantage of containing a number of undetermined parameters ( low - energy constants or lecs ) .
we extend prior work done at the @xmath7 limit @xcite and at one loop @xcite to two loops for @xmath8 .
other calculations using @xmath5pt have focused on @xmath3 , and while this may be of interest theoretically , it has at best a narrow window of validity@xcite . in addition , it is unlikely to find in nature a real electric or magnetic field with @xmath9 or @xmath10 , which is required for the approximation to be reasonable . incidentally , the opposite limit , @xmath11 , might be of more interest in the sense that it can be produced in the laboratory , but in such a regime the shift in the condensate is miniscule . continuing a calculation to higher orders in an expansion is always of interest at least in the trivial sense of finding a more precise result . in this case
, large-@xmath6 qcd@xcite provides another possible motivation for why the @xmath0 result might be of interest .
low - energy constants ( lecs ) of the same chiral order will have different orders of @xmath6 , depending upon the number of flavor traces in the term they multiply .
this can be understood as follows : a flavor trace corresponds to a quark loop in the analogous qcd calculation , and large-@xmath6 counting rules indicate quark loops are down by a power of @xmath6 , and the lec is the only parameter available to absorb this difference .
this large-@xmath6 dependence of the lecs can provide hints to the convergence of the chiral expansion .
there are several processes which have been calculated at two - loop order , most of which show a close match between the @xmath0 results and the experimental results @xcite .
the @xmath0 correction required to achieve this agreement varies , but in particular , the @xmath0 calculation of the process @xmath12 is strikingly more accurate than the @xmath13 result@xcite .
this is also the process which provides a ( very rough ) estimate of the lec we use here , so it is conceivable that the @xmath0 correction will be important in our case , as well .
we will first proceed with a brief overview of @xmath1 @xmath5pt .
we then follow with an analytical calculation of the shift in the chiral condensate due to a magnetic field at @xmath0 , and finally , a numerical analysis of the shift for general electromagnetic fields .
this section is a brief summary of the chiral perturbation theory notation that will be used in this paper . for a detailed description of the theory , see the original papers by gasser and leutwyler@xcite or a number of reviews(refs .
@xcite , for example ) .
treatment of the chiral lagrangian to @xmath0 can be found in @xcite .
the building blocks that are used to construct the chiral lagrangian include @xmath14 , containing the dynamical pion fields , and external fields @xmath15 , @xmath16 , @xmath17 , and @xmath18 .
we will consider here @xmath1 flavor symmetry with @xmath19 , with external fields corresponding to only a constant electromagnetic field and quark masses .
we then have @xmath20 we also need to define the covariant derivative as @xmath21 from these , we define the following operators that will contribute to the terms relevant in this paper ( in the general @xmath22 notation ) : @xmath23 where are defined @xmath24 \\f_l^{\mu\nu } = \partial^\mu l^\nu - \partial^\nu l^\mu - i[l^\mu , l^\nu ] . \end{split}\ ] ] note that in the case we are considering , @xmath25 .
@xmath26 can be paramaterized in several ways , but we will use the weinberg parameterization @xmath27 here , @xmath28 are still the dynamical fields , and @xmath29 is represented as an expansion in terms of @xmath28 from the second equation . using these definitions ,
@xmath30 and @xmath31 can be written as follows ( in the general @xmath32 form ) , where @xmath33 denotes the trace of @xmath34 @xcite : @xmath35\rangle - \frac{l_7}{16}\langle \chi_- \rangle^2 \\&+ \textrm{contact terms } \end{split}\ ] ] the calculation here is up to @xmath0 , which means that we will be using the @xmath30 lagrangian up to two loops and the @xmath31 lagrangian up to one loop ; the @xmath4 lagrangian will also contribute at tree level .
this lagrangian has been calculated in ref .
@xcite , and has more terms than we will list ( 112 for @xmath22 and 53 for @xmath1 ) .
fortunately , only one of these ( in @xmath1 ) will be relevant for our calculation , as we will see later , and it can be expressed as : @xmath36 renormalization of the theory to @xmath13 was calculated in ref .
it has also more recently been calculated for the @xmath0 lagrangian in @xcite .
using the @xmath1 lecs , the renormalized couplings can be written : @xmath37 when @xmath38 , these can be expressed in terms of scale - independent parameters as : @xmath39 the renormalization of the @xmath4 term that we will be using later can be expressed similarly in terms of the renormalized lecs from @xmath31 as @xmath40
the term in the qcd lagrangian which is relevant to @xmath41 is @xmath42 . since we know that @xmath43 , in the isospin limit of @xmath44 we can calculate the condensate from the vacuum energy as @xmath45 to first order in @xmath46 , the gell - mann - oakes - renner relation @xmath47@xcite applies , and can be used to calculate the shift in the condensate .
unfortunately , we will need the next order result , which will introduce some ambiguity as follows .
neither the chiral condensate or the quark mass can be defined independently ; only their product has a physical meaning . in the language of @xmath5pt@xcite , @xmath48 the ambiguity here is codified in the unphysical lec @xmath49 , which will vary according to the renormalization convention .
this is an ambiguity in the definition of @xmath41 . in order to avoid this difficulty
, we choose to normalize our results according to the quantity @xmath50 , which we define by @xmath51 with @xmath52 and @xmath53 at their physical values .
we will thus express our results in terms of @xmath54 , where @xmath55 .
as we will see , this ratio is unambiguous at the order to which we work .
we will also need the relationships between @xmath56 and @xmath53 as well as @xmath57 and @xmath52 ( @xmath58 ) .
the difference between the lowest - order @xmath13 result and the @xmath59 result is @xmath60 ; we therefore only need one order of corrections@xcite : @xmath61 \\f_\pi = & f\left[1 + \frac{m^2}{16\pi^2f^2}\bar{l}_4 + { \cal o}(m^4)\right ] .
\end{split}\ ] ] this will only be applicable in the first term of the expansion , which is two powers of momentum less than the maximum order for the calculation . our object is to calculate the condensate for the case of a constant electromagnetic field at @xmath0 in @xmath5pt .
as described above , we can accomplish this by calculating the vacuum energy to the same order . as noted by ref .
@xcite , this calculation will involve two - loop diagrams with @xmath30 vertices , one - loop diagrams with an @xmath31 vertex , and tree - level diagrams with an @xmath4 vertex .
diagrams which contribute to the vacuum energy will contain only external photon lines coming from the constant em field ( which is the `` vacuum '' in this case ) .
insertions of the electromagnetic field at the @xmath30 level are calculated as part of the propagator of the @xmath62 .
thus , the @xmath13 calculation corresponds roughly to a closed single propagator and was calculated in ref .
contributing diagrams must be dependent on the electromagnetic field , either through a direct insertion or through the propagator of the @xmath62 . with these criteria , we find that the diagrams contributing to our calculation are as pictured in fig . [
fig : vacdiagrams ] .
returning , then , to the @xmath5pt lagrangian , we find that only the terms proportional to @xmath63 , @xmath64 , @xmath65 , and @xmath66 can contribute from the @xmath31 lagrangian , and only the term proportional to @xmath67 ( as we anticipated above ) can contribute to the vacuum energy from @xmath4 .
+ we simplify the chiral lagrangians for @xmath1 and @xmath68 up to the relevant terms in @xmath4@xcite , including only terms which will contribute to the diagrams in figure [ fig : vacdiagrams ] : @xmath69 ^ 2 \\&-
\frac{m\pi^2}{8 f^2}\left[2\pi^+\pi^- + ( \pi^0)^2\right]^2 \\{\cal l}_4 = & -\frac{2 l_5}{f^2}(e f_{\mu\nu})^2\pi^+\pi^- \\&-\frac{2il_6}{f^2}e f_{\mu\nu}\left[\partial^\mu\pi^-\partial^\nu\pi^++iea^\mu\partial^\nu(\pi^+\pi^-)\right ] \\&-2l_3\frac{m^4}{f^2}\pi^+\pi^- \\{\cal l}_6 = & 4 c_{34 } m^2 ( ef_{\mu\nu})^2 \end{split}\ ] ] here , terms proportional to @xmath63 and @xmath64 have been added to the lagrangian from ref .
@xcite , which are down by an order of @xmath70 but have the same overall chiral order . as a first case , we will work with the case of pure magnetic fields , where @xmath71 .
this simplifies the calculations and allows us to obtain an analytic result .
we will later generalize to an arbitrary combination of @xmath72 and @xmath73 fields for numerical analysis .
the propagator for a scalar particle in a constant @xmath73 field was first calculated in ref .
@xcite , and here we use the convenient form also used in ref .
@xcite : @xmath74 where @xmath75 , @xmath76 and @xmath77 .
we will also need the scalar propagator @xmath78 \\\lambda = & \frac{1}{16\pi^2(d-4)}. \end{split}\ ] ] @xmath79 and @xmath80 are both divergent quantities , whereas @xmath81 is finite : @xmath82 with @xmath83 .
this is the same integral as was calculated in ref .
@xcite , and can be expressed analytically as : @xmath84 with these , the diagrams in fig .
[ fig : vacdiagrams ] can be calculated fairly straightforwardly to be @xcite @xmath85 the only one of these diagrams which is not expressed solely in terms of @xmath86 and @xmath79 is @xmath87 , which vanishes generally as well as in the @xmath3 case . in order to make the divergences and scale - dependence explicit , we first make the substitution @xmath88 . any term which is dependent on neither @xmath73 nor @xmath86 can then be re - absorbed into the vacuum energy ; we are only looking for the shift due to @xmath73 . after making this substitution , we see that @xmath89 is divergent and cancelled by a counterterm generated by the combination @xmath90 .
@xmath91 has both a finite piece , which will contribute to the calculation , and a divergent piece , which is cancelled by a counterterm in @xmath67 .
@xmath92 and @xmath93 are finite , aside from the aforementioned counterterms .
combining , then , all of these terms , we find the vacuum energy to be @xmath94 \\&+ \beta \bar{d}(m^2)\bigg\ } , \end{split}\ ] ] where we have defined the scale - independent quantity @xmath95 we then substitute eq .
( [ eqn : constcorrect ] ) to find @xmath96 from @xmath70 in the first -order term .
we find that the @xmath97 term cancels , and that the correction to @xmath57 does not play a role ( as it only appears at the second order ) .
then , taking a derivative , and applying the gell - mann - oakes - renner relation as above , with the first - order corrections for the @xmath13 term , we find ( @xmath98 ) : @xmath99 + \bar{d}(eh ) \bigg\ } , \end{split}\ ] ] with @xmath100 taking @xmath101 , @xmath102 , and the shift we find agrees with the expression found in ref .
@xcite for the case @xmath3 . for the case of an @xmath72 field
, we can make the substitution @xmath103 and get a similar analytic expression .
the @xmath104 case is somewhat more complicated , but we will write an integral expression which we can later evaluate numerically . for a convenient paramaterization of the general case , we introduce the variables @xmath105 and @xmath106 such that with @xmath107 and @xmath108 @xcite , @xmath109 expressed in terms of these variables , the shift in the condensate due to an arbitrary combination of fields will become ( @xmath110 ) : @xmath111 \end{split}\ ] ] this is the same integral as in ref .
@xcite , and as before we have had to avoid some potential ambiguity .
the poles in the integrand indicate an instability in the system , which is interpreted as due to pair creation in an electric field@xcite . we have chosen to regulate the divergence in a manner which has an imaginary part corresponding to this pair creation .
the magnitude of the imaginary part indicates the importance of this instability , though some caution is warranted in interpreting it quantitatively .
this issue was discussed in more detail in ref .
@xcite .
equations ( [ eqn : hshift ] ) and ( [ eqn : ehshift ] ) are the principal results of this work .
because the behavior of the theory is encoded in the lecs , a real - world interpretation of the low - energy behavior requires the use of measured lecs . for the @xmath31 lecs ,
this is straightforward , as these are individually determined with relatively small error . on the other hand , @xmath4 lecs , such as @xmath67 , are more problematic , as there are in general many more lecs than easily measurable processes to determine them .
these lecs are often estimated ( at a particular scale ) by resonance exchange .
unfortunately , @xmath67 in particular is difficult to extract , as the resonance processes to which it contributes involve only scalar exchange , and because it appears squared in these processes , its sign is undetermined .
this resonance exchange occurs at a scale @xmath112@xcite , and it is the ( scale - dependent ) value determined by experiment that has an undetermined sign .
the scale - independent @xmath113 is positive in both cases .
the values we use for these constants are @xcite : @xmath114 with these experimental values , we can plot realistic values of the shift in the condensate . in fig .
[ fig : shush ] , for the case of a pure magnetic field , we compare the value calculated for a finite @xmath53 to that for @xmath7 .
it is clear that these values are significantly different , as in the @xmath13 case@xcite . in this and the figures following , we have chosen to extend our results up to @xmath115 ( the expansion parameter is @xmath116 ) . a comparison of the shift due to a pure magnetic field in the @xmath7 case to the @xmath117 case .
shaded regions indicate uncertainty due to the @xmath4 constant @xmath118 . ] here , we use the same numerical trick as in ref .
@xcite to extract the principal value of the integral numerically .
we remove the singularities due to the poles located at @xmath119 with residue @xmath120 by subtracting the expression @xmath121 the principal value of the integral of this expression is zero , but it has a singularity at @xmath119 which exactly cancels the singular behavior of the integrand in @xmath122 . using this method , we plot the total shift in the condensate up to @xmath0 from a general @xmath72 and @xmath73 field in fig . [ fig : totval ] , and in fig . [ fig : fracval ] , we plot the ratio of the added correction at @xmath0 to the total shift . in these plots
, we have included a shaded region to indicate the possible values for the shift based on a range for @xmath118 of @xmath123 .
the asymptotic expression for the shift as @xmath124 ( for an @xmath73 field , which also provides some qualitative insight to other cases ) is @xmath125 which is , of course , zero for @xmath126 ( @xmath127 ) .
this expression encodes low - energy behavior for a more realistic regime , ; namely , that of the actual pion mass and a very small magnetic field .
we see from fig .
[ fig : fracval ] and in ( [ eqn : asymshift ] ) that the ( unknown ) sign of @xmath118 has a profound impact on the importance of the @xmath0 calculation .
the shift in the condensate for a positive @xmath118 is significant enough even as @xmath128 , whereas the shift for a negative one is negligible up to large values of @xmath129 . .
another notable feature is that the contribution to the imaginary part is larger at @xmath0 order as a pure @xmath73 field is approached ( while the total imaginary part is going to zero ) . except in this case
where the total imaginary part is negligible , the fraction of the imaginary part at two loops will be much less significant than its real counterpart in regimes where the chiral expansion would be expected to converge ( @xmath130 or below ) . the calculation as a whole will only be valid when the real part is significantly larger than the imaginary part .
when the imaginary part dominates , the system will break down due to the instability from pair creation .
we have studied the shift in the chiral condensate due to an electromagnetic field using chiral perturbation theory , which is a powerful tool for analyzing the low - energy behavior of qcd .
our analysis was done at @xmath0 with @xmath131 .
it is obvious that the inclusion of a nonzero pion mass greatly affects the result .
the importance of the @xmath0 correction is less clear .
large-@xmath6 reasoning coupled with the results of model - based calculations give circumstantial evidence that it could play an important role in the final result .
however , because the sign of the relevant lec at @xmath4 is undetermined by experiment , its effect at @xmath0 could be significant or virtually irrelevant .
i would like to thank tom cohen for many useful discussions .
this research is funded by the u. s. department of energy under grant number de - fg02 - 93er-40762 .
99 j. gasser and h. leutwyler , annals phys .
* 158 * ( 1984 ) 142 ; j. gasser and h. leutwyler , nucl . phys .
* b250 * ( 1985 ) 465 ; for reviews , see g. ecker , prog .
. phys . * 35 * ( 1995 ) 1 ; a. pich , rept .
* 58 * ( 1995 ) 563 i. a. shushpanov and a. v. smilga , phys
. lett . *
b402 * ( 1997 ) 351 n. o. agasian and i. a. shushpanov , phys . lett .
* b472 * ( 2000 ) 143 a. yu .
babansky , e. v. gorbar , g. v. shchepanyuk , phys . lett . * b419 * ( 1998 ) 272 a. goyal and m. dahiya , phys . rev .
* d62 * ( 2000 ) 025022 s. p. klevansky and r. h. lemmer , phys .
* d39 * ( 1989 ) 3478 y. nambu and g. jona - lasinio , phys . rev .
* 122 * ( 1961 ) 45 ; * 124 * ( 1961 ) 246 ; for a review , see s. p. klevansky , rev .
* 64 * ( 1992 ) 649 h. sugamura and t. tatsumi , ann
* 208 * ( 1991 ) 470 ; b. muller , s. schramm , and a. j. schramm , mod .
. lett . *
a7 * ( 1992 ) 973 ; k. g. klimenko , z. phys . * c54 * ( 1992 ) 323 ; k. g. klimenko , hep - ph/9809218 ; d. ebert , k. g. klimenko , m. a. vdovichenko , and a. s. vshivtsev , phys . rev . * d61 * ( 2000 ) 025005 , hep - ph/9905253 v. p. gusynin , v. a. miransky , i. a. shovkovy , phys . lett .
* b349 * ( 1995 ) 477 t. cohen , d. mcgady and e. werbos , phys . rev
. * c76 * ( 2007 ) 055201 g. t hooft , nucl . phys . *
b72 * ( 1974 ) 461 ; g. t hooft , nucl . phys .
* b75 * ( 1974 ) 461 e. witten , nucl . phys .
* b160 * ( 1979 ) 57 j. bijnens , prog .
nucl . phys .
* 58 * ( 2007 ) 521 , hep - ph/0604043 s. bellucci , j. gasser and m. e. sainio , nucl
. phys . *
b423 * ( 1994 ) 80 a. pich , rept .
* 58 * ( 1995 ) 563 g. ecker , prog .
* 35 * ( 1995 ) 1 j. bijnens , g. colangelo and g. ecker , annals phys .
* 280 * ( 2000 ) 100 j. bijnens , g. colangelo , g. ecker , jhep * 9902 * ( 1999 ) 020 m. gell - mann , r. j. oakes and b. renner , phys
* 175 * ( 1968 ) 2195 m. jamin , phys .
* b538 * ( 2002 ) 71 , hep - ph/0201174 j. r. pelez , phys . rev .
* d66 * ( 2002 ) 096007 , hep - ph/0202265 weinberg , s. physica * 96a * ( 1979 ) 327 j. schwinger , phys . rev .
* 82 * ( 1951 ) 664 j. bijnens and p. talavera , nucl
. phys . *
b489 * ( 1997 ) 387 , hep - ph:9610269 ; j. bijnens , arxiv/0708.1377 | we examine the shift in the chiral condensate due to a constant electromagnetic field at @xmath0 using @xmath1 chiral perturbation theory and a realistic @xmath2 .
we find that this value differs significantly from the value calculated using @xmath3 , while the magnitude of the two - loop correction is unclear due to the uncertainty in the experimentally determined value of the relevant @xmath4 lec . |
urban crowds are characterised by the presence of a large number of social groups .
the ratio between individual pedestrians and pedestrians moving in groups may change considerably between different environments and at different times of the day @xcite , but it is in general never negligible , with groups representing up to 85% of the walking population @xcite . despite this empirical evidence about the importance of groups , the standard approach in microscopic ( agent - based )
pedestrian modelling has been for long time to assume that the crowd is composed of individuals , moving without any preferential ties to other pedestrians .
this is an extremely strong simplification of the system , although it was obviously understandable as a first approach to the problem .
nevertheless , it is intuitive that groups behave in a specific way ( they move together and close ) and their presence should clearly influence the dynamics of the crowd .
not taking in account the group component of crowds may have an impact on the planning of buildings and emergency evacuation plans .
for example , @xcite reports that around 48% of people that evacuated by foot the city of sendai ( miyagi , japan ) during the 2011 tohoku great earthquake and tsunami did it by moving in groups , with a probable effect on evacuation times ( for the smaller city of kamaishi , in iwate prefecture , the figure was as high as 71% ) .
furthermore an understanding of pedestrian behaviour is essential for robots and automatic navigation vehicles ( such as wheelchairs or delivery carts ) that will become arguably common in future pedestrian areas @xcite .
indeed , in recent years , a few studies concerning empirical observations and mathematical modelling of the groups characteristic configuration and velocity have been introduced @xcite . in a recent series of papers @xcite
, we focused on the development of a mathematical model to describe group interaction , and in specific the group spatial structure and velocity .
the model proposed in @xcite introduced a non newtonian @xcite potential for group interaction on the basis of few and intuitive ideas about social interaction in pedestrian groups , and its predictions for group size , structure and velocity are in agreement with the observed natural behaviour of pedestrians . in @xcite
we studied a large data set of pedestrian trajectories to see how an _ extrinsic _ , i.e. , environmental , property such as crowd density influences the dynamics of groups , and in @xcite we introduced a mathematical model to explain such a crowd density effect on groups .
nevertheless , we may expect that a social behaviour such as walking in groups depends also on _ intrinsic _ properties of the groups .
it is known by studies with subjects that age , gender and height affect walking speed @xcite , but here we are interested on how group behaviour is affected by the nature of the group itself : not only by the characteristics of the individuals that compose it , but also by the relation between them , which is expected to have a strong impact on group dynamics ( see @xcite and our preliminary study @xcite ) . to study natural human behaviour
we use a large _ ecological _
( i.e. obtained by observing unconstrained pedestrians in their natural environment , see @xcite ) data set to describe how the group spatial structure , size and velocity of _ dyads _ ( two people groups ) change based on the following intrinsic properties of groups : 1 .
* purpose * of movement , 2 .
* relation * between the members , 3 . * gender * of the members , 4 .
* age * of the members , 5 .
* height * of the members .
being the data set based on unconstrained trajectories of unknown pedestrians , such features are necessarily ( with the exclusion of pedestrian height , obtained automatically through our tracking system @xcite ) _ apparent _ , i.e. based on the judgement of human coders , and thus an analysis of their reliability is performed .
furthermore , being social behaviour cultural dependent , the results are probably influenced by the place in which data were collected ( a shopping mall in osaka , western japan ) . nevertheless , they provide a useful insight on how these intrinsic features affect in a quantitative way the behaviour of dyads .
the pedestrian group data base used for this work is based on the freely available set @xcite , introduced by @xcite .
this set is again based on a very large pedestrian trajectory set @xcite , collected in a @xmath0 m@xmath1 area of the asia and pacific trade center ( atc ) , a multi - purpose building located in the osaka ( japan ) port area . for the purpose of this work , in order to avoid taking in consideration the effect of architectural features of the environment @xcite , such as its width , we use data only from the corridor area as defined in @xcite .
+ the trajectories have been automatically tracked using 3d range sensors and the algorithm introduced in @xcite , which provides , along with the pedestrian position on the plane , the height of their head , for more than 800 hours during a one year time span .
at the same time , we video recorded the tracking area using 16 different cameras . a subset of the video recordings were used by a human coder to identify the pedestrian social groups reported in data set @xcite .
the data set concerns the natural behaviour of pedestrians , i.e. the pedestrians were behaving in an unconstrained way , and observed in their natural environment .
collecting data in the pedestrians natural environment obviously presents some technical problems and an overall lower quality in tracking data ( higher tracking noise ) , but it is an approach with growing popularity @xcite , that allows for removing possible influence on pedestrian behaviour due to performing experiments in laboratories , i.e. artificial environments , using selected subjects following the experimenters instructions .
this is extremely important for this study , since we may hardly believe that social pedestrian group behaviour could be observed in such controlled laboratory experiments @xcite .
the pedestrians in this data set are all _ socially interacting _
, i.e. they were , on the basis of conversation and gaze clues , coded as not only moving together , but also performing some kind of social interaction @xcite . in order to obtain the `` ground truth '' for the inter - group composition and social relation , we proceeded similarly to our previous works @xcite , and asked three different human coders to observe the video recordings corresponding to the data set @xcite and analyse the group composition , and in detail to code , when possible , 1 .
the apparent purpose of the group s visit to the area ( * work * or * leisure * ) , 2 .
the apparent gender of their members , 3 . their apparent relation ( * colleagues * , * friends * , * family * or * couples * ) , 4 . their apparent age ( in decades , such as 0 - 9 , 10 - 19 , etc . )
while one coder examined data from five different days ( three working days and two holidays ) , corresponding to 1168 different socially interacting dyads , the other two examined only one day ( 283 dyads ) .
the coders are not specialised in pedestrian studies , are not aware of our mathematical models of pedestrian behaviour , and did not have access to our quantitative measurements of position and velocity .
they thus relied only on visual features such as clothing , gestures , behaviour and gazing @xcite to identify the groups social roles and composition .
the coding process is obviously strongly dependent on the subjective evaluation of the coder .
nevertheless , the 283 dyads examined by all coders may be used to examine their agreement , and provide thus some information about the reliability of their coding .
to this end , we use in appendix [ coderel ] two different approaches . on one hand
, we use the standard approach used in social sciences of directly comparing the results of the coding , through statistical indicators such as cohen s and fleiss s kappa , or krippendorf s alpha ( appendix [ coderag ] ) . on the other hand
, we also use an approach more rooted in the `` hard '' sciences , and treat the different codings as independent experiments , and quantitatively and quantitatively compare the findings ( appendix [ codercomp ] ) .
while our tracking system provides us with pedestrian positions and velocities at time intervals @xmath2 in the order of tens of milliseconds , we average pedestrian positions over time intervals @xmath3 s , to reduce the effect of measurement noise and the influence of pedestrian gait . as a result , we obtain pedestrian positions at discrete times @xmath4 , as @xmath5 where @xmath6 gives the height of the top of the pedestrian head , and define pedestrian velocities in 2d as @xmath7.\ ] ] following @xcite , only data points with both the average group velocity @xmath8 ( eq . [ ev ] ) and all individual velocities @xmath9 larger 0.5 m / s , and with all pedestrian positions falling inside a square with side 2.5 m centred on the group centre , were used .
pedestrian height measurement is obviously subject to oscillations ( see @xcite ) .
a major problem with height tracking is that there are situations in which the head is hidden or poorly tracked , and the pedestrian height is wrongly assigned as the height of the shoulders . to avoid this problem , for each pedestrian
we first compute the median height over the whole trajectory , and then define the pedestrian height as the average @xmath10 of all @xmath6 measurements above the median we use only data points for which the distance between pedestrians in the group was less than 4 meters .
this is a conservative threshold justified by our findings in @xcite suggesting that interacting groups have extremely low probability of having a distance larger than 2 meters . ] . by being smaller and
thus more easily occluded , the tracking of children is more difficult than the tracking of adults . since our statistical analysis identified a few interesting results related to children , we visually analysed the video recordings to verify that this problem did not affect significantly our findings .
as we report in @xcite , group velocity and spatial configuration depend on crowd density . in this paper
we follow again our main analysis of @xcite and compute pedestrian density with a good spatial resolution ( more than a good time resolution ) as time averages over 300 seconds in a @xmath11 meters square area .
more details , along with a discussion of possible density definitions , may be found in @xcite ( refer also to @xcite and @xcite for possible alternative definitions of density ) .
based on our analysis performed in works @xcite , we define the following quantitative observables for the dynamics of a pedestrian dyad ( fig [ f0 ] ) : 1 .
the _ group velocity _ @xmath8 , defined as @xmath12 @xmath13 being the velocities of the two pedestrians in an arbitrary reference frame co - moving with the environment ( i.e. in which the velocity of walls and other architectural features is zero ) .
2 . the _ pedestrian distance _ or _ group size _ @xmath14 , defined as @xmath15 @xmath16 being the positions of the two pedestrians in the above reference frame .
the _ group abreast distance _ or _
abreast extension _
@xmath17 , that may be defined as follows .
first we identify a unit vector in the direction of @xmath18 @xmath19 for each pedestrian we compute the clockwise angle @xmath20 between @xmath21 and @xmath16 , and define the projection of each @xmath16 orthogonal to the velocity @xmath22 if necessary , we rename the pedestrians so that @xmath23 and finally define @xmath24 4 .
we can also define the group extension in the direction of motion , @xmath25 this is a signed quantity , a property that was particularly useful in our previous works .
nevertheless , for the purpose of this paper , it resulted more useful to analyse the _ group depth _
@xmath26 as described in detail in appendix [ detstat ] , to which the reader should refer for technical details , for each observable @xmath27 and relation or composition category @xmath4 we provide the number of groups @xmath28 , the category average @xmath29 , standard deviation @xmath30 and standard error @xmath31 , all based on the analysis of groups that contributed with at least 10 usable data points , and reported in tables as @xmath32 furthermore we provide an analysis of the overall observable probability distribution function , and some parameters to estimate the differences between categories ( anova @xmath33 values , effect size @xmath34 , coefficient of determination @xmath35 ) .
the cross - analysis regarding the common effect of different `` factors '' ( i.e. , purpose , relation , gender , age and height ) may be found in appendix [ accounting ] .
the results related to the purpose dependence of all observables concerning the 1088 dyads whose purpose was coded ( and that provided enough data to be analysed ) are shown in table [ table1a ] ( refer to appendix [ detstat ] for an explanation of all terms ) .
@xmath14 & @xmath17 & @xmath36 + leisure & 716 & 1118 @xmath37 7.3 ( @xmath38=195 ) & 815 @xmath37 9.5 ( @xmath38=253 ) & 628 @xmath37 6.1 ( @xmath38=162 ) & 383 @xmath37 12 ( @xmath38=334 ) + work & 372 & 1271 @xmath37 8.2 ( @xmath38=158 ) & 845 @xmath37 12 ( @xmath38=228 ) & 713 @xmath37 8 ( @xmath38=154 ) & 332 @xmath37 15 ( @xmath38=289 ) + @xmath39 & & 169 & 3.75 & 69.4 & 6.25 + @xmath33 & & @xmath40 & 0.053 & @xmath40 & 0.0126 + @xmath35 & & 0.135 & 0.00344 & 0.0601 & 0.00572 + @xmath34 & & 0.832 & 0.124 & 0.533 & 0.16 + we have thus a very strong and significant evidence that pedestrians that visited the environment for working walk at an higher velocity and with a larger abreast distance , as shown by the comparison of averages and standard errors , and by the corresponding high @xmath41 and @xmath34 and low @xmath33 values ( see appendix [ detstat ] for definitions and meaning of these quantities ) .
we have also a difference in distance and `` group depth '' , although its significance is less strong .
we can get further insight about the differences in behaviour between workers and leisure oriented people by studying explicitly the probability distribution functions for the observables @xmath8 , @xmath14 , @xmath17 and @xmath36 , which are shown respectively in figures [ f1d ] , [ f1a ] , [ f1b ] and [ f1c ] , and whose statistical analysis is reported in appendix [ overpur ] ( refer again to appendix [ detstat ] for the difference between the analysis reported in the main text an the one of appendix [ overpur ] ) .
observable in leisure ( blue , centre of bin identified by circles ) and work ( red , squares ) oriented dyads .
all pdfs in this work are shown after having been smoothed with a moving average filter . ]
observable in leisure ( blue , circles ) and work ( red , squares ) oriented dyads . ] observable in leisure ( blue , circles ) and work ( red , squares ) oriented dyads . ] observable in leisure ( blue , circles ) and work ( red , squares ) oriented dyads . ]
these pdfs provide an easy interpretation of the data .
the @xmath17 and @xmath8 peaks and tails are displaced to higher values for workers .
the @xmath14 peak is also displaced to an higher value , but leisure distribution has a fatter tail . correspondingly , the @xmath36 distribution is slightly more spread in leisure oriented pedestrians .
furthermore , the @xmath17 distribution presents a considerably higher value for low @xmath17 values in the `` leisure '' case .
these latter consideration show that while `` workers '' walk strongly abreast , the `` leisure '' dyads are less ordered .
in appendix [ furtherpur ] we further analyse these results , to understand the effect on them of age , gender , density and height , while in appendix [ coderpur ] we verify that the major findings are confirmed by all coders .
we may see that in general , even when age , gender , density and height are kept fixed , the results exposed above are confirmed by this further analysis .
groups may also be analysed according to the relation between their members ( colleagues , couples , friends , family ) .
there is obviously a strong overlap between the `` colleagues '' category and the `` work '' one analysed above ( and obviously between `` leisure '' and the three categories couples , friends and families ) , but since they are are conceptually different ( colleagues could visit the shopping mall for lunch , or for shopping outside of working time ) , we will provide an independent analysis we usually drop the analysis of purpose and focus on relation . ] .
the results related to the relation dependence for all observables concerning the 1018 dyads whose purpose was coded ( and that provided enough data to be analysed ) are shown in table [ table2a ] .
@xmath14 & @xmath17 & @xmath36 + colleagues & 358 & 1274 @xmath37 8.3 ( @xmath38=157 ) & 851 @xmath37 12 ( @xmath38=231 ) & 718 @xmath37 8.3 ( @xmath38=157 ) & 334 @xmath37 15 ( @xmath38=292 ) + couples & 96 & 1099 @xmath37 17 ( @xmath38=169 ) & 714 @xmath37 22 ( @xmath38=219 ) & 600 @xmath37 15 ( @xmath38=150 ) & 291 @xmath37 24 ( @xmath38=231 ) + families & 246 & 1094 @xmath37 13 ( @xmath38=197 ) & 863 @xmath37 19 ( @xmath38=302 ) & 583 @xmath37 11 ( @xmath38=171 ) & 498 @xmath37 25 ( @xmath38=391 ) + friends & 318 & 1138 @xmath37 11 ( @xmath38=200 ) & 792 @xmath37 11 ( @xmath38=199 ) & 662 @xmath37 7.5 ( @xmath38=134 ) & 314 @xmath37 15 ( @xmath38=268 ) + @xmath42 & & 60.7 & 12.2 & 42.3 & 21.4 + @xmath33 & & @xmath40 & 7.39@xmath43 & @xmath40 & @xmath40 + @xmath35 & & 0.152 & 0.0349 & 0.111 & 0.0595 + @xmath34 & & 1.03 & 0.529 & 0.828 & 0.587 + we may see that , as expected from the previous analysis , there is a considerable difference between the velocity of colleagues and the velocity of the other groups .
friends appear to be faster than couples or families , although the difference is limited to 2 - 3 standard errors .
we may also see that couples walk at the closest distance , followed by friends , colleagues and then families . on the other hand ,
families walk at the shortest abreast distance , although at a value basically equivalent to that of couples .
the abreast distance of friends is significantly larger , and the one of colleagues assumes the greatest value .
the `` depth '' @xmath36 assumes the smallest value in couples , followed by friends and workers , and the , by a large margin , highest value in families .
these results may be completely understood only by analysing the probability distribution functions , which are shown in figures [ f2d ] , [ f2a ] , [ f2b ] and [ f2c ] for , respectively , @xmath8 , @xmath14 , @xmath17 and @xmath36 ( the statistical analysis of these distributions is reported in section [ overrel ] ) .
observable in dyads with colleague ( blue , squares ) , couple ( red , circles ) , family ( orange , triangles ) and friend ( green , stars ) relation . ] observable in dyads with colleague ( blue , squares ) , couple ( red , circles ) , family ( orange , triangles ) and friend ( green , stars ) relation . ] observable in dyads with colleague ( blue , squares ) , couple ( red , circles ) , family ( orange , triangles ) and friend ( green , stars ) relation . ] observable in dyads with colleague ( blue , squares ) , couple ( red , circles ) , family ( orange , triangles ) and friend ( green , stars ) relation . ] the pdfs provide again an easy interpretation of the data .
the @xmath8 distributions for friends , families and couples are quite similar , while the one for colleagues is clearly different ( displaced to higher values ) .
this suggests that `` relation '' influences velocity in a limited way , with respect to `` purpose '' .
the peaks of both @xmath14 and @xmath17 distributions assume the minimum value for couples , followed by families , friends and colleagues .
the distributions for families present the following peculiar properties : the @xmath14 distribution has a fat tail ( causing the high average value ) , the @xmath17 distribution assumes large values for small @xmath17 , and the @xmath36 distribution is more spread ( on the other hand , @xmath36 distributions are very similar in the other categories ) . we may thus conclude that `` relation '' has an influence on distance , with couples walking at the closest distance , followed by families , friends and colleagues . at the same time , families walk in a less ordered formation ( less abreast ) .
this behaviour is probably mainly due to children ( see also section [ ageef ] ) , and influences the results of the previous section ( `` leisure '' oriented dyads walking less abreast ) . in appendix [ furtherrel ]
we further analyse these results , to understand the effect on them of age , gender , density and height , while in appendix [ coderrel ] we verify if the major findings are confirmed by all coders .
the major trends exposed above are all confirmed by this further analysis . in particular , the tendency of families to have a wider @xmath36 distribution may be diminished but does not disappear when we keep fixed gender , age or height , showing that it is probably not only due to children .
the results related to the relation dependence for all observables concerning the 1089 dyads whose gender was coded ( and that provided enough data to be analysed ) are shown in table [ table3a ] .
@xmath14 & @xmath17 & @xmath36 + two females & 252 & 1102 @xmath37 12 ( @xmath38=193 ) & 790 @xmath37 14 ( @xmath38=227 ) & 647 @xmath37 7.8 ( @xmath38=123 ) & 321 @xmath37 20 ( @xmath38=311 ) + mixed & 371 & 1111 @xmath37 9.5 ( @xmath38=183 ) & 824 @xmath37 14 ( @xmath38=273 ) & 613 @xmath37 9 ( @xmath38=174 ) & 416 @xmath37 18 ( @xmath38=350 ) + two males & 466 & 1254 @xmath37 8.3 ( @xmath38=178 ) & 846 @xmath37 11 ( @xmath38=228 ) & 699 @xmath37 7.7 ( @xmath38=166 ) & 349 @xmath37 14 ( @xmath38=293 ) + @xmath44 & & 84.6 & 4.37 & 30.7 & 7.69 + @xmath33 & & @xmath40 & 0.0129 & @xmath40 & 0.000484 + @xmath35 & & 0.135 & 0.00798 & 0.0535 & 0.014 + @xmath34 & & 0.825 & 0.248 & 0.51 & 0.282 + although we may easily see that the differences between the distributions are statistically significant ( with stronger differences in the @xmath8 and @xmath17 distributions ) , it is again useful , in order to understand these results , to analyse the probability distribution functions , which are shown in figures [ f3d ] , [ f3a ] , [ f3b ] and [ f3c ] for , respectively , the @xmath8 , @xmath14 , @xmath17 and @xmath36 observables ( the statistical analysis of these distributions is reported in section [ overgen ] ) observable in dyads with two females ( red , circles ) , mixed ( orange , triangles ) and two males ( blue , squares ) . ]
observable in dyads with two females ( red , circles ) , mixed ( orange , triangles ) and two males ( blue , squares ) . ]
observable in dyads with two females ( red , circles ) , mixed ( orange , triangles ) and two males ( blue , squares ) . ]
observable in dyads with two females ( red , circles ) , mixed ( orange , triangles ) and two males ( blue , squares ) . ]
the difference between the females and males distributions is very clear , with both peaks and tails in the velocity and ( abreast or absolute ) distance distributions displaced to higher results . regarding the @xmath36 distribution
, we may see that the male distribution is more spread than the female one , and thus females have a stronger tendency to walk abreast .
the mixed dyads absolute and abreast distance distribution are characterised by low values for the peaks and fat tails , in particular for the absolute value distribution .
the @xmath17 distribution presents relatively high values at low @xmath17 , and correspondingly the @xmath36 distribution is very spread ( tendency not to walk abreast ) .
the mixed dyads velocity distribution is interestingly very similar to the female one
. the peculiarity of the mixed distributions may be better understood by taking in consideration the other effects , in particular those related to relation , as shown in appendix [ furthergen ] .
coder reliability is analysed in appendix [ codergen ] .
to study the dependence of the @xmath14 , @xmath17 , @xmath36 and @xmath8 observable on age , we used three different approaches , namely to study how these observable change depending on _ average _ , _ maximum _ and _ minimum _ group age .
the latter analysis appears to be the most interesting one , since it allows us to spot the presence of children , and we limit ourselves to it in the main text .
results corresponding to the dependence on average and maximum age are found in appendix [ maxavage ] .
table [ table4a ] and figures [ f4b ] , [ f4b2 ] show the minimum age dependence of all observables ( based on the analysis of 1089 dyads ) .
although differences between distributions are statistically significant , both velocity and distance observables are mostly constant for groups whose minimum age is in the 20 - 60 years range .
we nevertheless find that the group depth @xmath36 ( the observable characterising thus the tendency of pedestrians not to walk abreast ) assumes a very high value in groups with children , a minimum in the 20 - 29 years range , and then grows with age . on the other hand ,
abreast distance @xmath17 is relatively low for groups with children ( as we will see below , @xmath17 grows with body size distance in elderly people , due to the shorter height of elderly people in the japanese population @xcite . ] ) .
velocity is mostly constant below 60 years , but drops for elderly groups .
@xmath14 & @xmath17 & @xmath36 + 0 - 9 years & 31 & 1143 @xmath37 42 ( @xmath38=235 ) & 995 @xmath37 69 ( @xmath38=383 ) & 529 @xmath37 34 ( @xmath38=189 ) & 701 @xmath37 87 ( @xmath38=485 ) + 10 - 19 years & 63 & 1158 @xmath37 33 ( @xmath38=259 ) & 791 @xmath37 33 ( @xmath38=259 ) & 624 @xmath37 19 ( @xmath38=148 ) & 359 @xmath37 40 ( @xmath38=320 ) + 20 - 29 years & 364 & 1181 @xmath37 9.1 ( @xmath38=173 ) & 793 @xmath37 11 ( @xmath38=218 ) & 668 @xmath37 8.1 ( @xmath38=154 ) & 307 @xmath37 14 ( @xmath38=264 ) + 30 - 39 years & 292 & 1204 @xmath37 12 ( @xmath38=202 ) & 836 @xmath37 14 ( @xmath38=238 ) & 673 @xmath37 10 ( @xmath38=176 ) & 364 @xmath37 18 ( @xmath38=307 ) + 40 - 49 years & 149 & 1181 @xmath37 14 ( @xmath38=176 ) & 841 @xmath37 18 ( @xmath38=224 ) & 664 @xmath37 13 ( @xmath38=158 ) & 384 @xmath37 26 ( @xmath38=311 ) + 50 - 59 years & 111 & 1164 @xmath37 18 ( @xmath38=193 ) & 825 @xmath37 21 ( @xmath38=219 ) & 649 @xmath37 15 ( @xmath38=160 ) & 378 @xmath37 30 ( @xmath38=318 ) + 60 - 69 years & 67 & 1028 @xmath37 21 ( @xmath38=170 ) & 881 @xmath37 41 ( @xmath38=335 ) & 638 @xmath37 20 ( @xmath38=164 ) & 468 @xmath37 52 ( @xmath38=422 ) + @xmath45 70 years & 12 & 886 @xmath37 29 ( @xmath38=99.8 ) & 786 @xmath37 79 ( @xmath38=275 ) & 588 @xmath37 19 ( @xmath38=66.6 ) & 385 @xmath37 100 ( @xmath38=363 ) + @xmath46 & & 10.7 & 3.96 & 4.23 & 8.02 + @xmath33 & & @xmath40 & 0.000282 & 0.000128 & @xmath40 + @xmath35 & & 0.065 & 0.025 & 0.0267 & 0.0494 + @xmath34 & & 1.6 & 0.583 & 0.808 & 1.37 + dependence on minimum age .
dashed lines provide standard error confidence intervals .
the point at 75 years corresponds to the `` 70 years or more '' slot . ] , @xmath17 and @xmath36 dependence on minimum age .
black circles : @xmath14 ; red squares : @xmath17 ; blue triangles : @xmath36 .
dashed lines provide standard error confidence intervals .
the point at 75 years corresponds to the `` 70 years or more '' slot . ]
the probability functions for different observables in different age ranges are shown in figures [ f4d ] , [ f4e ] , [ f4f ] and [ f4 g ] respectively for observables @xmath8 , @xmath14 , @xmath17 and @xmath36 , and their statistical analysis is presented in section [ overage ] .
we may easily see from the large tail of the @xmath14 distribution , the high values for the @xmath17 distribution , the spread of the @xmath36 distribution , that the presence of a child causes the group not to walk very abreast .
the abreast distance peak is higher in `` working age people''with respect to young and elderly dyads .
elderly people have a very narrow peak in the @xmath36 distribution , but also a fat tail .
velocity in the 0 - 19 age range assumes lower peaks than in the 20 - 59 range , but has a large spread , while in elderly people it assumes clearly lower values , the tracking of short people ( and thus of children ) is more difficult , and thus the tracked position could be affected by higher sensor noise , although our time filter ( see again section [ traj ] ) should remove this problem .
we thus examined a portion of the videos corresponding to group with children , and noticed that children have indeed an erratic behaviour that leads them to sudden accelerations and non - abreast formations .
we thus believe that the large spread of observables for dyads with children is due to actual pedestrian behaviour . ] .
( minimum ) age in the 0 - 9 years range : green and circles ; in the 10 - 19 range : red and diamonds , in the 30 - 39 range : orange and squares ; in the 50 - 59 range : blue , and triangles ; in the over 70 range : black and stars . ] .
( minimum ) age in the 0 - 9 years range : green and circles ; in the 10 - 19 range : red and diamonds , in the 30 - 39 range : orange and squares ; in the 50 - 59 range : blue , and triangles ; in the over 70 range : black and stars . ] .
( minimum ) age in the 0 - 9 years range : green and circles ; in the 10 - 19 range : red and diamonds , in the 30 - 39 range : orange and squares ; in the 50 - 59 range : blue , and triangles ; in the over 70 range : black and stars . ] .
( minimum ) age in the 0 - 9 years range : green and circles ; in the 10 - 19 range : red and diamonds , in the 30 - 39 range : orange and squares ; in the 50 - 59 range : blue , and triangles ; in the over 70 range : black and stars . ] in appendix [ furtherage ] we analyse possible effects due to density , relation , gender and age .
interesting results reported in the appendix suggest a tendency of families _ not _ to walk abreast even when formed only by adults , and differences in groups with children based on gender ( probably affected by the gender of the parent ) .
coder reliability is analysed in appendix [ coderage ] .
a further interesting result is that , as shown in figure [ fastchild ] ( based on the analysis of appendix [ accounting ] ) , dyads with children _ walk faster at higher density _ , in contrast with the usual pedestrian behaviour . on minimum age at different densities .
* blue squares : @xmath47 ped / m@xmath1 range ; red circles : @xmath48 ped / m@xmath1 range .
the point at 75 years corresponds to the `` 70 years or more '' slot .
dashed lines show standard error confidence intervals . ]
height is the only pedestrian feature that is not the result of coding , since it is automatically tracked by our system @xcite
. we again considered ( see appendix [ heicomp ] ) average , minimum and maximum height .
the three indicators give similar results , and in the following we use minimum height to better identify the presence of children .
the dependence of all observables on minimum height ( based on 1089 dyads ) is shown in table [ table_h_min ] .
we have significant statistical difference for all observables , but the interpretation of the results is not straightforward , due to the peculiar behaviour of dyads including short people ( most probably children ) . as shown in figure [ f5d ] velocity grows ( as expected , see for example @xcite and @xcite ) , with height , but dyads with a very short individual represent an exception ( children move fast despite the short height ) . in figure [ f5d2 ]
we may see that distance is mostly independent of height above 150 cm , but assumes a very high value for dyads including short pedestrians .
figure [ f5b ] shows the height dependence of @xmath36 , which results to be a decreasing function , although a comparison with a linear fit shows that dyads including people under 140 cm walk with a particularly spread ( non abreast ) @xmath36 distribution , while above 150 cm the group depth is almost constant .
the @xmath17 observable , on the other hand , appears to grow mostly in a linear way ( figure [ f5b2 ] ) .
this could lead us to think that abreast distance depends only on body size .
nevertheless , while there is probably a strong dependence of abreast distance on height , this linear dependence is also due to the balance between the non - linear male and female behaviour , as shown in figure [ fastchild2 ] , based on the analysis of appendix [ accounting ] .
furthermore , as we will see below when studying the distribution probability functions of the @xmath17 observable , the growth in @xmath17 with height is a combination of a increase of peak position and decrease of people walking in non abreast formation ( figure [ f5f ] ) .
@xmath14 & @xmath17 & @xmath36 + @xmath49 140 cm & 39 & 1130 @xmath37 34 ( @xmath38=211 ) & 1004 @xmath37 65 ( @xmath38=404 ) & 573 @xmath37 34 ( @xmath38=210 ) & 672 @xmath37 80 ( @xmath38=501 ) + 140 - 150 cm & 39 & 1106 @xmath37 50 ( @xmath38=311 ) & 875 @xmath37 46 ( @xmath38=289 ) & 619 @xmath37 25 ( @xmath38=156 ) & 469 @xmath37 64 ( @xmath38=403 ) + 150 - 160 cm & 234 & 1104 @xmath37 13 ( @xmath38=197 ) & 797 @xmath37 16 ( @xmath38=246 ) & 631 @xmath37 8.9 ( @xmath38=136 ) & 360 @xmath37 21 ( @xmath38=328 ) + 160 - 170 cm & 498 & 1169 @xmath37 8.1 ( @xmath38=182 ) & 821 @xmath37 11 ( @xmath38=243 ) & 657 @xmath37 7.7 ( @xmath38=172 ) & 362 @xmath37 14 ( @xmath38=311 ) + 170 - 180 cm & 262 & 1242 @xmath37 11 ( @xmath38=173 ) & 827 @xmath37 12 ( @xmath38=197 ) & 699 @xmath37 9.6 ( @xmath38=155 ) & 321 @xmath37 16 ( @xmath38=251 ) + @xmath50 180 cm & 17 & 1232 @xmath37 51 ( @xmath38=211 ) & 793 @xmath37 48 ( @xmath38=198 ) & 689 @xmath37 33 ( @xmath38=135 ) & 270 @xmath37 53 ( @xmath38=217 ) + @xmath51 & & 14.5 & 5.25 & 7.45 & 9.69 + @xmath33 & & @xmath40 & 9.03@xmath52 & 6.9@xmath53 & @xmath40 + @xmath35 & & 0.0626 & 0.0237 & 0.0333 & 0.0428 + @xmath34 & & 0.744 & 0.591 & 0.773 & 0.922 + dependence on minimum height .
data points shown by red circles . continuous black line : linear best fit with @xmath54 , @xmath55=715 mm / s , @xmath56=2.81 s@xmath57 .
dashed lines provide standard error confidence intervals , the point at 135 cm corresponds to the `` less than 140 cm '' slot , the one at 185 cm to the `` more than 180 cm '' slot . ]
dependence on minimum height .
data points shown by red circles . continuous black line : linear best fit with @xmath58 , @xmath55=1390 mm , @xmath56=-3.35 . dashed lines provide standard error confidence intervals , the point at 135 cm corresponds to the `` less than 140 cm '' slot , the one at 185 cm to the `` more than 180 cm '' slot . ]
dependence on minimum height .
data points shown by red circles . continuous black line : linear best fit with @xmath59 , @xmath55=1530 mm , @xmath56=-7.01 .
dashed lines provide standard error confidence intervals , the point at 135 cm corresponds to the `` less than 140 cm '' slot , the one at 185 cm to the `` more than 180 cm '' slot . ]
dependence on minimum height .
data points shown by red circles . continuous black line : linear best fit with @xmath60 , @xmath55=258 mm , @xmath56=2.42 .
dashed lines provide standard error confidence intervals , the point at 135 cm corresponds to the `` less than 140 cm '' slot , the one at 185 cm to the `` more than 180 cm '' slot . ]
dependence on minimum height for different genders .
red circles : two females ; blue squares : two males ( continuous lines : linear fits @xmath60 , @xmath61=154 mm , @xmath62=2.98 ; @xmath63=211 mm , @xmath64=2.79 ) .
the points at 135 and 185 cm represent the `` less than 140 '' and `` more than 180 '' cm slots ) .
dashed lines show confidence intervals . ] the probability functions for different observables in different minimum height ranges are shown in figures [ f5h ] [ f5e ] , [ f5f ] and [ f5 g ] , respectively for the @xmath8 , @xmath14 , @xmath17 and @xmath36 observables , and their statistical analysis is shown in section [ overh ] .
we see that the abreast distance distributions are displaced to the right with growing height , with a corresponding decrease in the values assumed around zero ( particularly high in the 0 - 140 cm distribution , probably due to children behaviour ) . similarly , the @xmath36 distribution becomes narrower with growing height , and presents a very different behaviour in the shortest height slot .
the absolute distance distributions are displaced to the right with growing height , but the very fat tail for the 0 - 140 cm distribution causes the average value to have a more complex dependence on height .
the @xmath8 distribution shows a clear displacement to the right with growing height , both in peaks and tails , although the 0 - 140 cm distribution has again a peculiar behaviour due to its very pronounced width . .
( minimum ) height in the 0 - 140 cm range : green circles ; in the 140 - 150 range : red diamonds ; in the 150 - 160 range : orange squares ; in the 160 - 170 range : blue triangles ; in the 170 - 180 range : black stars . ] .
( minimum ) height in the 0 - 140 cm range : green circles ; in the 140 - 150 range : red diamonds ; in the 150 - 160 range : orange squares ; in the 160 - 170 range : blue triangles ; in the 170 - 180 range : black stars . ] .
( minimum ) height in the 0 - 140 cm range : green circles ; in the 140 - 150 range : red diamonds ; in the 150 - 160 range : orange squares ; in the 160 - 170 range : blue triangles ; in the 170 - 180 range : black stars . ] .
( minimum ) height in the 0 - 140 cm range : green circles ; in the 140 - 150 range : red diamonds ; in the 150 - 160 range : orange squares ; in the 160 - 170 range : blue triangles ; in the 170 - 180 range : black stars . ] in appendix [ furtherhei ] we analyse the validity of these results on height dependence when we consider other effects such as age , relation , gender and density , and verify that , although sometimes diminished , height related results are present also when analysing groups with fixed age , relation and gender .
by analysing how pedestrian dyad behaviour depends on the group s `` intrinsic properties '' , namely the characteristics of its members and the relation between them , we observed that females dyads are slower and walk closer than males , that workers walk faster , at a larger distance and more abreast than leisure oriented people , and that inter - group relation has a strong effect on group structure , with couples walking very close and abreast , colleagues walking at a larger distance , and friends walking more abreast than family members .
pedestrian height influences velocity and abreast distance , observables that grow with the average or minimum group height .
we also found that elderly people walk slowly , while active age adults walk at the maximum velocity .
dyads with children have a strong tendency to walk in a non abreast formation , with a large distance but a shorter abreast distance . in the supplementary materials appendices
, we analysed how these features affect each other , and we verified that the effects of the different features are present , even though sometimes diminished , even when the other features are kept fixed ( e.g. , when we compare colleagues of different gender , and the like ) .
the cross - analysis of the interplay between these features revealed also a richer structure .
interesting results are , for example , that the velocity of dyads with children appears to increase with density ( at least in the low - medium density range ) , and that children behaviour appears to be influenced by the gender of the parent . in this work we focused on `` group
features '' more than `` individual features '' , i.e. we did not explicitly address questions such as the age or height difference , and similar .
we may nevertheless infer from our results some information about how group members with different height , age and gender `` compromise '' on group dynamics .
we may see , for example , in appendix [ heicomp ] that average age gives , for the @xmath8 and @xmath17 observables that are growing function of height , a result in between those obtained for minimum and maximum height .
height is a physical and not social feature , and it appears that , after averaging over all social features , the chosen velocity and abreast distance are the averages of those preferred by the individuals .
gender and age appear , on the other hand , to have a deeper impact on social interactions . in mixed groups ,
males appear to adapt to female velocity when we average over relations , but when we analyse for secondary effects in appendix [ accounting ] , we see that this is true for couples and families , but it does not apply to friends or colleagues . similarly , while couples walk closer than male or female same sex dyads , mixed colleague groups walk farther than same sex dyads of both genders . in a similar way , due also to the peculiar behaviour of children , it is impossible to find a simple `` compromise '' rule for age related behaviour .
more information could be inferred by an analysis taking in explicit account age differences , that we reserve for a future work .
the exact figures found in this work may depend strongly on the environment in which they have been recorded , and vary not only with density , but also with other macroscopic crowd dynamics features ( uni - directional flow , bi - directional flow , multi - directional flow , presence or not of standing pedestrians , etc . ) as well as architectural features of the environment ( open space or large corridor or narrow corridor , etc . ) .
for this reason , attempts to verify our findings in different environments should be directed not at specific quantitative figures ( e.g. male dyads walk at 1.25 m / s and females at 1.1 meters per second ) but at qualitative patterns ( e.g. males walk faster than females in a statistically significant way , with a difference in velocity comparable to the standard deviation in distributions ) .
it would be in particular very interesting to compare our findings with different cultural settings , since it may be expected that social group behaviour is strongly dependent on culture , so that at least some of the patterns could change when similar data collection experiments are performed outside of ( western ) japan .
a possible extension of this work regard the analysis of three people group behaviour .
furthermore , as stated above , in this work we limited ourselves to group properties and not individual properties ( e.g. , we verified if a group was mixed , but we did not study the specific position of the male or female ) . after a revision of the coding procedure , we could analyse if , according to gender , age or height differences , roles such as `` leader '' or `` follower '' emerge .
finally , a mathematical modelling following @xcite and @xcite could be performed . besides the obvious applications to pedestrian simulations , with possible influence in building and events planning , disaster prevention , and even in entertainment industries such as movies and video games , we are particularly interested in applications in the field of robotics and more in general slow vehicles with automatic navigation capabilities deployed in pedestrian facilities , such as delivery vehicles or automatic wheelchairs and carts .
such vehicles will arguably become more common in the future , and in order to navigate safely inside human crowds , and to move together with other humans `` as in a group '' , they will need and understanding of pedestrian and group behaviour . more specifically , a `` companion '' robot or an automatic wheelchair will need 1 . to be able to recognise pedestrian groups , using an automatic recognition algorithm
@xcite 2 . to be able to predict their behaviour , both in order to be able to safely avoid them and to perform a socially acceptable behaviour @xcite 3 .
to be able to move together with other humans , and behave as a member of a group @xcite for all these applications it is extremely important to understand deeply how pedestrians actually behave and we plan to use these findings to improve our previous algorithms and systems as part of the development of a platform for autonomous personal mobility systems .
this research is partly supported by the ministry of internal affairs and communications ( mic ) , japan , _ research and development project on autonomous personal mobility including robots _ , by crest , jst and jsps kakenhi grand number 16j40223 .
in this work we are interested in describing how pedestrian group behaviour is influenced by some _ intrinsic features _ , such as purpose , relation , gender , age or height .
each feature ( or factor ) may be divided in @xmath4 categories ( e.g. , in the case of relation @xmath65 and the categories are colleagues , couples , family and friends ) .
each group is coded as belonging to a specific category , so that each category has @xmath66 groups . as described in section [ traj ] , for each group @xmath67 we can measure the value of observable @xmath68 every 500 ms .
we may call these measurements @xmath69 with @xmath70 ( i.e. we have @xmath71 measurements , or events , corresponding to group @xmath72 in category @xmath4 ) .
we believe that the largest amount of quantitative information regarding the dependence of group behaviour on intrinsic features is included in the overall probability distributions functions concerning all @xmath73 measurements of a given observable , as shown for example in figure [ f1d ] , since from the analysis of these figures we can understand what is the probability of having a given value for each observable in each category .
it is nevertheless useful to extract some quantitative information , such as average values and standard deviations , from these distributions .
furthermore , although the purpose of this paper is not to provide a `` @xmath33 value statistical independence label '' to each feature , to compare such average values it is customary and useful to compute , along with other statistical indicators such as effect size and determination coefficient , the standard error of each distribution and to perform the related analysis of variance ( anova ) .
the computation of these latter statistical quantities is nevertheless based on an assumption of statistical independence of the data , an assumption that clearly does not hold for all our @xmath74 observations we were following a single group ( @xmath75 ) for one hour ( @xmath76 ) .
we will have then , if we ignore measurement noise , a perfect information regarding the behaviour of that group in that hour and , under the strong assumption of time independence in the group behaviour , a good statistics about the behaviour _ of that particular group_. we still do not have any information about how group behaviour changes between groups in the category , since that information depends on the number of groups analysed , @xmath28 .
furthermore , since in general we track a given group only for the few seconds it needs to cross the corridor , the observations @xmath77 at fixed @xmath72 are also strongly time correlated . ] .
we thus proceed in the following way , justified by having a similar number of observation for each group . for each observable @xmath68
we compute the average over group @xmath72 @xmath78 and then provide its average value in the category @xmath4 as @xmath79 where @xmath80 and the standard error @xmath81 are given by @xmath82 @xmath83 and the standard deviation is @xmath84 this rule of thumb is obviously related to the anova analysis reported in the text .
the anova analysis proceeds as follows .
we define @xmath87 as the number of categories for a given feature , @xmath88 as the total number of groups , and the overall average of the observable as @xmath89 we then define the distance between @xmath80 and @xmath29 as @xmath90 and the degrees of freedom @xmath91 the @xmath41 factor is then defined as @xmath92 this result is reported in our tables as @xmath93 , along with the celebrated @xmath33 value , that provides the probability , under the hypothesis of independence of data , that the difference between the distributions is due to chance @xcite @xmath94 the @xmath95 distribution has to be computed numerically @xcite , but a value @xmath96 assures a small @xmath33 value .
let us see how this relates to the rule of thumb for standard errors .
let us assume we have two categories with the same number of groups for category @xmath97 we clearly have @xmath98 @xmath99 and @xmath100 using but the numbers shown in the tables use the approximate definition @xmath101 . for @xmath102 or more ,
as it is usually the case in this work , the difference is at most 5% . ]
@xmath103 we get the expression @xmath104 so that the rule of thumb eq .
[ thumb ] corresponds to have an high @xmath41 value and thus a low @xmath33 value .
eq [ effe ] says that the @xmath41 factor is high if the @xmath30 are smaller than the @xmath105 , i.e. if the variation inside the categories are smaller than outside the category , and if the total number of observation is high .
due to the large number of data points , the @xmath41 values in appendix [ overstat ] ( where we use all the observable measurement instead of group averages ) are always very high , and the corresponding @xmath33 values very low , but the hypothesis of statistical independence of data underlying the usual interpretation of @xmath33 is obviously not valid .
there are nevertheless some statistical estimators that do not depend dramatically on the number of observations , and that will thus have a similar value either if performed using all the data points or if performed using only group averages .
one such estimator is the coefficient of determination @xmath106 which can also be computed as from the @xmath41 factor as @xmath107 and provides an estimate of how much of the variance in the data is `` explained '' by the category averages . the @xmath35 coefficient may attain low values if two or more category distribution functions are very similar , as it usually the case in our work . to point out the presence of at least one distribution that is clearly different from the others we may use the following definition of the effect size @xmath34 .
we first define @xcite @xmath108 where @xmath109 , @xmath110 are the number of points used for computing the averages and standard deviations if we are using group averages , @xmath111 if we are using overall distributions . ] , and then we consider the maximum pairwise effect size @xmath112 while a @xmath33 value tells us about the significance of the statistical difference between two distributions , the difference may be often so small that if can be verified only if a large amount of data are collected .
but if we have also @xmath113 , then the two distributions are different enough to be distinguished also using a relatively reduced amount of data .
we refrain from applying the machinery of two way or @xmath114 way anova to our data , since our ecological data set is extremely unbalanced , and it is unbalanced for the very reason that our `` factors '' are not independent variables .
it is nevertheless useful to analyse the interplay between the different features , and we do that in section [ accounting ] by performing a statistical analysis similar to the one described above of a given feature @xmath115 while keeping fixed the value of another feature @xmath116 to a category @xmath117 . , we use also the external feature of pedestrian crowd density . since the same group may contribute to different densities , when operating at a fixed density we use for group averages all groups that contribute with at least 5 data points ( instead of the usual 10 ) to the observable distribution for that density value . ]
sometimes this analysis is performed on a reduced number of groups , and thus the corresponding @xmath33 value may be high .
this does not imply that the analysis is valueless , at least in our opinion , since it provides new information .
the @xmath41 and @xmath33 values are , in this situation , useful to compare different observables on the given condition . as an example ,
table [ table2c1 ] tells us that @xmath17 has a stronger variation between relation categories for fixed gender than @xmath14 , and so on .
furthermore , in these situations , an analysis of statistical indicators that do not depend critically on the number of observations , such as the effect size , is particularly valuable .
table [ overt1 ] provides a statistical analysis of the overall probability distributions for the purpose categories .
@xmath14 & @xmath17 & @xmath36 + leisure & 38501 & 1096 @xmath37 1.3 ( @xmath38=251 ) & 799 @xmath37 1.6 ( @xmath38=309 ) & 630 @xmath37 1.2 ( @xmath38=236 ) & 360 @xmath37 2 ( @xmath38=388 ) + work & 18936 & 1257 @xmath37 1.7 ( @xmath38=235 ) & 834 @xmath37 2.1 ( @xmath38=287 ) & 714 @xmath37 1.6 ( @xmath38=227 ) & 315 @xmath37 2.5 ( @xmath38=341 ) + @xmath118 & & 5400 & 169 & 1640 & 184 + @xmath33 & & @xmath40 & @xmath40 & @xmath40 & @xmath40 + @xmath35 & & 0.0859 & 0.00293 & 0.0278 & 0.00319 + @xmath34 & & 0.652 & 0.115 & 0.36 & 0.12 + table [ overt2 ] provides a statistical analysis of the overall probability distributions for the relation categories .
@xmath14 & @xmath17 & @xmath36 + colleagues & 18172 & 1262 @xmath37 1.7 ( @xmath38=234 ) & 840 @xmath37 2.2 ( @xmath38=290 ) & 720 @xmath37 1.7 ( @xmath38=229 ) & 317 @xmath37 2.6 ( @xmath38=344 ) + couples & 5273 & 1085 @xmath37 3.2 ( @xmath38=231 ) & 699 @xmath37 3.7 ( @xmath38=271 ) & 584 @xmath37 2.6 ( @xmath38=188 ) & 290 @xmath37 4.4 ( @xmath38=318 ) + families & 12596 & 1072 @xmath37 2.2 ( @xmath38=246 ) & 834 @xmath37 3.2 ( @xmath38=357 ) & 592 @xmath37 2.3 ( @xmath38=260 ) & 452 @xmath37 4 ( @xmath38=447 ) + friends & 17634 & 1113 @xmath37 2 ( @xmath38=260 ) & 788 @xmath37 2 ( @xmath38=265 ) & 659 @xmath37 1.6 ( @xmath38=214 ) & 312 @xmath37 2.5 ( @xmath38=338 ) + @xmath119 & & 1940 & 362 & 975 & 485 + @xmath33 & & @xmath40 & @xmath40 & @xmath40 & @xmath40 + @xmath35 & & 0.0978 & 0.0198 & 0.0517 & 0.0264 + @xmath34 & & 0.795 & 0.493 & 0.614 & 0.392 + table [ overt3 ] provides a statistical analysis of the overall probability distributions for the relation categories .
@xmath14 & @xmath17 & @xmath36 + two females & 14688 & 1075 @xmath37 2.1 ( @xmath38=251 ) & 773 @xmath37 2.2 ( @xmath38=268 ) & 647 @xmath37 1.7 ( @xmath38=202 ) & 302 @xmath37 2.9 ( @xmath38=346 ) + mixed & 19311 & 1098 @xmath37 1.7 ( @xmath38=239 ) & 803 @xmath37 2.4 ( @xmath38=334 ) & 614 @xmath37 1.8 ( @xmath38=248 ) & 388 @xmath37 3 ( @xmath38=411 ) + two males & 23516 & 1237 @xmath37 1.6 ( @xmath38=249 ) & 839 @xmath37 1.9 ( @xmath38=292 ) & 702 @xmath37 1.6 ( @xmath38=239 ) & 337 @xmath37 2.3 ( @xmath38=355 ) + @xmath120 & & 2570 & 225 & 791 & 232 + @xmath33 & & @xmath40 & @xmath40 & @xmath40 & @xmath40 + @xmath35 & & 0.0822 & 0.00778 & 0.0268 &
0.008 + @xmath34 & & 0.647 & 0.233 & 0.365 & 0.224 + table [ overt4 ] provides a statistical analysis of the overall probability distributions for the minimum age ranges .
@xmath14 & @xmath17 & @xmath36 + 0 - 9 years & 1041 & 1127 @xmath37 8.4 ( @xmath38=272 ) & 983 @xmath37 15 ( @xmath38=480 ) & 573 @xmath37 9.5 ( @xmath38=306 ) & 663 @xmath37 18 ( @xmath38=580 ) + 10 - 19 years & 3443 & 1110 @xmath37 5.2 ( @xmath38=303 ) & 767 @xmath37 5.1 ( @xmath38=298 ) & 626 @xmath37 3.8 ( @xmath38=222 ) & 322 @xmath37 6.2 ( @xmath38=364 ) + 20 - 29 years & 18679 & 1167 @xmath37 1.8 ( @xmath38=240 ) & 788 @xmath37 2.1 ( @xmath38=289 ) & 665 @xmath37 1.6 ( @xmath38=223 ) & 301 @xmath37 2.6 ( @xmath38=349 ) + 30 - 39 years & 15552 & 1179 @xmath37 2.1 ( @xmath38=264 ) & 816 @xmath37 2.4 ( @xmath38=294 ) & 667 @xmath37 2 ( @xmath38=248 ) & 343 @xmath37 2.9 ( @xmath38=357 ) + 40 - 49 years & 7974 & 1167 @xmath37 2.7 ( @xmath38=242 ) & 838 @xmath37 3.3 ( @xmath38=296 ) & 668 @xmath37 2.7 ( @xmath38=243 ) & 374 @xmath37 4.2 ( @xmath38=378 ) + 50 - 59 years & 6025 & 1153 @xmath37 3.3 ( @xmath38=253 ) & 812 @xmath37 3.7 ( @xmath38=284 ) & 653 @xmath37 2.9 ( @xmath38=223 ) & 358 @xmath37 4.7 ( @xmath38=367 ) + 60 - 69 years & 3969 & 1001 @xmath37 3.5 ( @xmath38=219 ) & 836 @xmath37 5.4 ( @xmath38=340 ) & 643 @xmath37 3.8 ( @xmath38=242 ) & 409 @xmath37 6.7 ( @xmath38=419 ) + @xmath45 70 years & 832 & 877 @xmath37 6 ( @xmath38=172 ) & 793 @xmath37 13 ( @xmath38=363 ) & 599 @xmath37 7.8 ( @xmath38=224 ) & 383 @xmath37 16 ( @xmath38=453 ) + @xmath121 & & 400 & 89.1 & 46.7 & 175 + @xmath33 & & @xmath40 & @xmath40 & @xmath40 & @xmath40 + @xmath35 & & 0.0464 & 0.0107 & 0.00566 & 0.0208 + @xmath34 & & 1.16 & 0.619 & 0.382 & 0.991 + table [ overt5 ] provides a statistical analysis of the overall probability distributions for the minimum height ranges .
@xmath14 & @xmath17 & @xmath36 + @xmath49 140 cm & 1579 & 1127 @xmath37 6.9 ( @xmath38=274 ) & 942 @xmath37 11 ( @xmath38=457 ) & 605 @xmath37 7.6 ( @xmath38=300 ) & 578 @xmath37 14 ( @xmath38=553 ) + 140 - 150 cm & 2206 & 1032 @xmath37 6.7 ( @xmath38=315 ) & 855 @xmath37 8 ( @xmath38=374 ) & 644 @xmath37 5.3 ( @xmath38=248 ) & 420 @xmath37 10 ( @xmath38=468 ) + 150 - 160 cm & 13064 & 1076 @xmath37 2.2 ( @xmath38=251 ) & 779 @xmath37 2.5 ( @xmath38=281 ) & 628 @xmath37 1.8 ( @xmath38=209 ) & 337 @xmath37 3.2 ( @xmath38=365 ) + 160 - 170 cm & 26345 & 1151 @xmath37 1.5 ( @xmath38=245 ) & 810 @xmath37 1.9 ( @xmath38=306 ) & 655 @xmath37 1.5 ( @xmath38=243 ) & 348 @xmath37 2.3 ( @xmath38=374 ) + 170 - 180 cm & 13497 & 1234 @xmath37 2.1 ( @xmath38=243 ) & 819 @xmath37 2.3 ( @xmath38=269 ) & 700 @xmath37 2 ( @xmath38=232 ) & 309 @xmath37 2.8 ( @xmath38=323 ) + @xmath50 180 cm & 824 & 1224 @xmath37 9.3 ( @xmath38=268 ) & 823 @xmath37 11 ( @xmath38=325 ) & 686 @xmath37 8.1 ( @xmath38=234 ) & 309 @xmath37 14 ( @xmath38=404 ) + @xmath122 & & 648 & 102 & 149 & 171 + @xmath33 & & @xmath40 & @xmath40 & @xmath40 & @xmath40 + @xmath35 & & 0.0533 & 0.00875 & 0.0128 & 0.0146 + @xmath34 & & 0.796 & 0.534 & 0.398 & 0.532 +
work - oriented dyads are more frequently found during working days , in which the environment presents a lower density ( and thus higher velocity and inter - group pedestrian distance , @xcite ) .
it is thus important to analyse the results of section [ purpose ] when they are divided for density ranges , for example by comparing results in the @xmath123 pedestrian per square meter range with those in the @xmath124 range .
the results are reported in table [ table1b1 ] and [ table1b2 ] , showing that the differences in @xmath8 and @xmath17 remain significant regardless of density .
the difference in @xmath36 becomes significant at high density , while at very low density is not significant ( while the opposite happens to @xmath14 ) .
@xmath14 & @xmath17 & @xmath36 + leisure & 426 & 1113 @xmath37 10 ( @xmath38=215 ) & 851 @xmath37 15 ( @xmath38=303 ) & 658 @xmath37 9 ( @xmath38=186 ) & 400 @xmath37 18 ( @xmath38=373 ) + work & 209 & 1274 @xmath37 12 ( @xmath38=169 ) & 924 @xmath37 20 ( @xmath38=296 ) & 741 @xmath37 14 ( @xmath38=203 ) & 409 @xmath37 26 ( @xmath38=373 ) + @xmath125 & & 89.6 & 8.17 & 25.7 & 0.0807 + @xmath33 & & @xmath40 & 0.0044 & 5.36@xmath53 & 0.776 + @xmath35 & & 0.124 & 0.0127 & 0.039 &
0.000128 + @xmath34 & & 0.8 & 0.242 & 0.428 & 0.024 + @xmath14 & @xmath17 & @xmath36 + leisure & 145 & 1084 @xmath37 14 ( @xmath38=170 ) & 764 @xmath37 17 ( @xmath38=209 ) & 560 @xmath37 13 ( @xmath38=158 ) & 390 @xmath37 26 ( @xmath38=308 ) + work & 22 & 1229 @xmath37 27 ( @xmath38=125 ) & 754 @xmath37 26 ( @xmath38=123 ) & 673 @xmath37 25 ( @xmath38=117 ) & 237 @xmath37 40 ( @xmath38=186 ) + @xmath126 & & 14.7 & 0.0513 & 10.2 & 5.05 + @xmath33 & & 0.000182 & 0.821 & 0.00167 & 0.026 + @xmath35 & & 0.0817 & 0.000311 & 0.0583 & 0.0297 + @xmath34 & & 0.881 & 0.0521 & 0.735 & 0.516 + in table [ table1b3 ] and [ table1b4 ] we report , respectively , @xmath33 and @xmath34 values for purpose corresponding to each observable and density range , showing that the @xmath8 , @xmath17 and @xmath36 distributions are different in a statistically significant way at different density ranges , although the effect on @xmath36 grows with density .
differences in @xmath14 are significant only at the lowest density range . & @xmath36 + 0 - 0.05 ped / m@xmath1 & @xmath40 & 0.0044 & 5.36@xmath53 & 0.776 + 0.05 - 0.1 ped / m@xmath1 & @xmath40 & 0.682 & @xmath40 & 0.000517 + 0.1 - 0.15 ped / m@xmath1 & @xmath40 & 0.221 & @xmath40 & 1.32@xmath127 + 0.15 - 0.2 ped / m@xmath1 & 0.000182 & 0.821 & 0.00167 & 0.026 + & @xmath36 + 0 - 0.05 ped / m@xmath1 & 0.8 & 0.242 & 0.428 & 0.024 + 0.05 - 0.1 ped / m@xmath1 & 0.914 & 0.0292 & 0.515 & 0.248 + 0.1 - 0.15 ped / m@xmath1 & 0.812 & 0.117 & 0.627 & 0.467 + 0.15 - 0.2
ped / m@xmath1 & 0.881 & 0.0521 & 0.735 & 0.516 + the work and leisure populations are strongly biased regarding gender . in tables
[ table1c1 ] , [ table1c2 ] and [ table1c3 ] we show the results for the work and leisure observables when limited to , respectively , female , mixed and male dyads . while velocity is still significantly different also when gender is fixed , absolute distance in men , and all distance observables in females are not significantly different .
we may thus conclude that differences between workers and leisure oriented people are present regardless of gender , but are magnified by the gender difference in the two populations .
@xmath14 & @xmath17 & @xmath36 + leisure & 222 & 1092 @xmath37 13 ( @xmath38=194 ) & 794 @xmath37 16 ( @xmath38=235 ) & 644 @xmath37 8.4 ( @xmath38=125 ) & 328 @xmath37 22 ( @xmath38=322 ) + work & 29 & 1184 @xmath37 30 ( @xmath38=162 ) & 755 @xmath37 28 ( @xmath38=150 ) & 663 @xmath37 19 ( @xmath38=101 ) & 274 @xmath37 38 ( @xmath38=204 ) + @xmath128 & & 5.97 & 0.733 & 0.615 & 0.777 + @xmath33 & & 0.0153 & 0.393 & 0.433 & 0.379 + @xmath35 & & 0.0234 & 0.00293 & 0.00247 & 0.00311 + @xmath34 & & 0.484 & 0.17 & 0.155 & 0.175 + @xmath14 & @xmath17 & @xmath36 + leisure & 330 & 1097 @xmath37 9.9 ( @xmath38=180 ) & 814 @xmath37 15 ( @xmath38=267 ) & 602 @xmath37 9.6 ( @xmath38=174 ) & 415 @xmath37 19 ( @xmath38=341 ) + work & 41 & 1226 @xmath37 26 ( @xmath38=167 ) & 902 @xmath37 48 ( @xmath38=308 ) & 698 @xmath37 24 ( @xmath38=152 ) & 420 @xmath37 65 ( @xmath38=419 ) + @xmath129 & & 18.9 & 3.79 & 11.3 & 0.00662 + @xmath33 & & 1.77@xmath52 & 0.0524 & 0.000849 & 0.935 + @xmath35 & & 0.0488 & 0.0102 & 0.0298 & 1.79@xmath52 + @xmath34 & & 0.722 & 0.323 & 0.558 & 0.0135 + @xmath14 & @xmath17 & @xmath36 + leisure & 164 & 1196 @xmath37 16 ( @xmath38=207 ) & 846 @xmath37 19 ( @xmath38=246 ) & 660 @xmath37 14 ( @xmath38=173 ) & 392 @xmath37 25 ( @xmath38=325 ) + work & 302 & 1285 @xmath37 8.7 ( @xmath38=152 ) & 846 @xmath37 13 ( @xmath38=218 ) & 720 @xmath37 9 ( @xmath38=157 ) & 325 @xmath37 16 ( @xmath38=271 ) + @xmath130 & & 28.1 & 0.000251 & 14.4 & 5.5 + @xmath33 & & 1.83@xmath53 & 0.987 & 0.000165 & 0.0195 + @xmath35 & & 0.057 & 5.42@xmath53 & 0.0301 & 0.0117 + @xmath34 & & 0.515 & 0.00154 & 0.369 & 0.228 + in tables ( [ table1d1 ] ) and ( [ table1d2 ] ) we show the results for the work and leisure observables when limited to groups of a given average age .
the results suggest that differences may be present at any age ( in particular concerning @xmath8 ) , but are definitely more strong for more mature walkers .
@xmath14 & @xmath17 & @xmath36 + leisure & 292 & 1164 @xmath37 10 ( @xmath38=177 ) & 798 @xmath37 14 ( @xmath38=242 ) & 656 @xmath37 9.7 ( @xmath38=165 ) & 326 @xmath37 17 ( @xmath38=290 ) + work & 78 & 1242 @xmath37 19 ( @xmath38=166 ) & 775 @xmath37 17 ( @xmath38=152 ) & 684 @xmath37 12 ( @xmath38=108 ) & 266 @xmath37 22 ( @xmath38=197 ) + @xmath131 & & 12.2 & 0.608 & 1.91 & 2.93 + @xmath33 & & 0.000536 & 0.436 & 0.168 & 0.088 + @xmath35 & & 0.0321 & 0.00165 & 0.00515 & 0.00789 + @xmath34 & & 0.446 & 0.0996 & 0.176 & 0.219 + @xmath14 & @xmath17 & @xmath36 + leisure & 61 & 1053 @xmath37 21 ( @xmath38=164 ) & 808 @xmath37 32 ( @xmath38=247 ) & 601 @xmath37 20 ( @xmath38=155 ) & 404 @xmath37 46 ( @xmath38=356 ) + work & 53 & 1276 @xmath37 21 ( @xmath38=153 ) & 845 @xmath37 24 ( @xmath38=173 ) & 706 @xmath37 20 ( @xmath38=144 ) & 345 @xmath37 36 ( @xmath38=261 ) + @xmath132 & & 54.8 & 0.808 & 13.7 & 0.966 + @xmath33 & & @xmath40 & 0.371 & 0.000328 & 0.328 + @xmath35 & & 0.329 & 0.00716 & 0.109 & 0.00855 + @xmath34 & & 1.4 & 0.17 & 0.702 & 0.186 + in table [ table1d3 ] and [ table1d4 ] we report , respectively , @xmath33 and @xmath34 values for purpose corresponding to each observable and average age range , showing again that differences have a tendency to grow with age .
@xmath17 & @xmath36 + 20 - 29 years & 0.000536 & 0.436 & 0.168 & 0.088 + 30 - 39 years & @xmath40 & 0.0689 & 6.34@xmath43 & 0.144 + 40 - 49 years & @xmath40 & 0.12 & 4.78@xmath127 & 0.264 + 50 - 59 years & @xmath40 & 0.371 & 0.000328 & 0.328 + 60 - 69 years & 0.0233 & 0.221 & 0.48 & 0.463 + @xmath17 & @xmath36 + 20 - 29 years & 0.446 & 0.0996 & 0.176 & 0.219 + 30 - 39 years & 0.994 & 0.224 & 0.682 & 0.18 + 40 - 49 years & 0.97 & 0.226 & 0.68 & 0.162 + 50 - 59 years & 1.4 & 0.17 & 0.702 & 0.186 + 60 - 69 years & 1.21 & 0.649 & 0.373 & 0.389 + in tables ( [ table1e1 ] ) and ( [ table1e2 ] ) we show the results for the work and leisure observables when limited to groups of a given average height , and in tables [ table1e3 ] and [ table1e4 ] we report , respectively , @xmath33 and @xmath34 values for purpose corresponding to each observable and average height range .
differences appear to grow with height , probably affected also by the gender distributions .
@xmath14 & @xmath17 & @xmath36 + leisure & 108 & 1107 @xmath37 25 ( @xmath38=260 ) & 821 @xmath37 26 ( @xmath38=268 ) & 629 @xmath37 15 ( @xmath38=153 ) & 389 @xmath37 34 ( @xmath38=352 ) + work & 10 & 1152 @xmath37 51 ( @xmath38=160 ) & 709 @xmath37 27 ( @xmath38=86 ) & 631 @xmath37 28 ( @xmath38=88.5 ) & 264 @xmath37 32 ( @xmath38=101 ) + @xmath133 & & 0.283 & 1.71 & 0.00249 & 1.24 + @xmath33 & & 0.596 & 0.194 & 0.96 & 0.268 + @xmath35 & & 0.00243 & 0.0145 & 2.15@xmath52 & 0.0106 + @xmath34 & & 0.177 & 0.434 & 0.0166 & 0.37 + @xmath14 & @xmath17 & @xmath36 + leisure & 188 & 1138 @xmath37 12 ( @xmath38=168 ) & 801 @xmath37 15 ( @xmath38=212 ) & 628 @xmath37 12 ( @xmath38=159 ) & 366 @xmath37 21 ( @xmath38=291 ) + work & 233 & 1291 @xmath37 10 ( @xmath38=157 ) & 850 @xmath37 15 ( @xmath38=230 ) & 730 @xmath37 10 ( @xmath38=153 ) & 322 @xmath37 18 ( @xmath38=273 ) + @xmath134 & & 92.2 & 5.22 & 44 & 2.53 + @xmath33 & & @xmath40 & 0.0228 & @xmath40 & 0.113 + @xmath35 & & 0.18 & 0.0123 & 0.0951 & 0.006 + @xmath34 & & 0.944 & 0.225 & 0.652 & 0.156 + @xmath17 & @xmath36 + 150 - 160 cm & 0.596 & 0.194 & 0.96 & 0.268 + 160 - 170 cm & @xmath40 & 0.0557 & 0.00126 & 0.953 + 170 - 180 cm & @xmath40 & 0.0228 & @xmath40 & 0.113 + @xmath50 180 cm & 0.773 & 0.959 & 0.289 & 0.522 + @xmath17 & @xmath36 + 150 - 160 cm & 0.177 & 0.434 & 0.0166 & 0.37 + 160 - 170 cm & 0.696 & 0.217 & 0.368 & 0.00667 + 170 - 180 cm & 0.944 & 0.225 & 0.652 & 0.156 + @xmath50 180 cm & 0.0998 & 0.0177 & 0.368 & 0.22 + as discussed above , work - oriented ( and thus colleagues ) dyads are more present during working days , in which the environment presents a lower density
. tables [ table2b1 ] and [ table2b2 ] show the observables dependence for fixed density ranges ( @xmath123 ped / m@xmath1 and @xmath124 ped / m@xmath1 , respectively ) .
the major trends exposed in the main text are present at any density , as confirmed also by tables [ table2b3 ] and [ table2b4 ] , reporting @xmath33 and @xmath34 values , respectively , for all density ranges .
@xmath14 & @xmath17 & @xmath36 + colleagues & 202 & 1276 @xmath37 12 ( @xmath38=169 ) & 934 @xmath37 21 ( @xmath38=298 ) & 751 @xmath37 15 ( @xmath38=207 ) & 409 @xmath37 26 ( @xmath38=376 ) + couples & 62 & 1103 @xmath37 25 ( @xmath38=193 ) & 760 @xmath37 38 ( @xmath38=297 ) & 600 @xmath37 22 ( @xmath38=177 ) & 359 @xmath37 40 ( @xmath38=314 ) + families & 125 & 1084 @xmath37 19 ( @xmath38=208 ) & 894 @xmath37 30 ( @xmath38=331 ) & 617 @xmath37 16 ( @xmath38=175 ) & 512 @xmath37 37 ( @xmath38=413 ) + friends & 193 & 1130 @xmath37 17 ( @xmath38=230 ) & 830 @xmath37 19 ( @xmath38=258 ) & 685 @xmath37 12 ( @xmath38=162 ) & 338 @xmath37 23 ( @xmath38=326 ) + @xmath135 & & 30.5 & 7.59 & 19 & 6.17 + @xmath33 & & @xmath40 & 5.52@xmath52 & @xmath40 & 0.000396 + @xmath35 & & 0.137 & 0.0379 & 0.0896 & 0.031 + @xmath34 & & 1.03 & 0.585 & 0.754 & 0.482 + @xmath14 & @xmath17 & @xmath36 + colleagues & 22 & 1229 @xmath37 27 ( @xmath38=125 ) & 754 @xmath37 26 ( @xmath38=123 ) & 673 @xmath37 25 ( @xmath38=117 ) & 237 @xmath37 40 ( @xmath38=186 ) + couples & 19 & 1064 @xmath37 28 ( @xmath38=124 ) & 663 @xmath37 31 ( @xmath38=135 ) & 542 @xmath37 22 ( @xmath38=97.1 ) & 290 @xmath37 36 ( @xmath38=159 ) + families & 68 & 1068 @xmath37 22 ( @xmath38=180 ) & 802 @xmath37 30 ( @xmath38=247 ) & 532 @xmath37 21 ( @xmath38=170 ) & 465 @xmath37 44 ( @xmath38=362 ) + friends & 57 & 1107 @xmath37 22 ( @xmath38=168 ) & 753 @xmath37 22 ( @xmath38=164 ) & 603 @xmath37 20 ( @xmath38=149 ) & 332 @xmath37 33 ( @xmath38=251 ) + @xmath136 & & 5.54 & 2.59 & 5.87 & 4.77 + @xmath33 & & 0.0012 & 0.055 & 0.000794 & 0.00327 + @xmath35 & & 0.0931 & 0.0457 & 0.098 & 0.0811 + @xmath34 & & 1.32 & 0.613 & 0.888 & 0.694 + & @xmath36 + 0 - 0.05 ped / m@xmath1 & @xmath40 & 5.52@xmath52 & @xmath40 & 0.000396 + 0.05 - 0.1 ped / m@xmath1 & @xmath40 & 0.000158 & @xmath40 & @xmath40 + 0.1 - 0.15 ped / m@xmath1 & @xmath40 & 2.22@xmath52 & @xmath40 & @xmath40 + 0.15 - 0.2 ped / m@xmath1 & 0.0012 & 0.055 & 0.000794 & 0.00327 + 0.2 - 0.25 ped / m@xmath1 & 0.378 & 0.327 & 0.144 & 0.664 + & @xmath36 + 0 - 0.05 ped / m@xmath1 & 1.03 & 0.585 & 0.754 & 0.482 + 0.05 - 0.1
ped / m@xmath1 & 1.07 & 0.476 & 0.732 & 0.538 + 0.1 - 0.15 ped / m@xmath1 & 1.16 & 0.666 & 0.857 & 0.761 + 0.15 - 0.2 ped / m@xmath1 & 1.32 & 0.613 & 0.888 & 0.694 + 0.2 - 0.25
ped / m@xmath1 & 0.675 & 0.798 & 0.983 & 0.974 + we now compare the results regarding relation for groups of given gender ( two females , mixed and two males ) in tables [ table2c1 ] , [ table2c2 ] and [ table2c3 ] .
differences in the distributions ( and the corresponding trends ) are still significant in fixed gender groups , with the exception of the @xmath14 female and male distributions , although , as shown by the relatively high @xmath34 values , this may be due due by the low amount of data in some categories . the patterns analysed in the main text are mostly respected ,
although we may notice some differences such as female friends walking at an higher distance than colleagues , and two male families walking at a very high speed .
@xmath14 & @xmath17 & @xmath36 + colleagues & 24 & 1167 @xmath37 30 ( @xmath38=145 ) & 735 @xmath37 26 ( @xmath38=128 ) & 664 @xmath37 20 ( @xmath38=95.9 ) & 238 @xmath37 34 ( @xmath38=168 ) + families & 28 & 1023 @xmath37 32 ( @xmath38=171 ) & 847 @xmath37 58 ( @xmath38=305 ) & 565 @xmath37 27 ( @xmath38=140 ) & 488 @xmath37 77 ( @xmath38=405 ) + friends & 184 & 1105 @xmath37 15 ( @xmath38=197 ) & 777 @xmath37 15 ( @xmath38=205 ) & 658 @xmath37 8.5 ( @xmath38=115 ) & 293 @xmath37 20 ( @xmath38=274 ) + @xmath137 & & 3.85 & 1.9 & 7.85 & 6.48 + @xmath33 & & 0.0227 & 0.153 & 0.000503 & 0.00182 + @xmath35 & & 0.032 & 0.016 & 0.0631 & 0.0527 + @xmath34 & & 0.902 & 0.465 & 0.817 & 0.783 + @xmath14 & @xmath17 & @xmath36 + colleagues & 35 & 1228 @xmath37 30 ( @xmath38=175 ) & 923 @xmath37 55 ( @xmath38=327 ) & 702 @xmath37 27 ( @xmath38=158 ) & 440 @xmath37 75 ( @xmath38=445 ) + couples & 96 & 1099 @xmath37 17 ( @xmath38=169 ) & 714 @xmath37 22 ( @xmath38=219 ) & 600 @xmath37 15 ( @xmath38=150 ) & 291 @xmath37 24 ( @xmath38=231 ) + families & 183 & 1078 @xmath37 13 ( @xmath38=182 ) & 860 @xmath37 21 ( @xmath38=285 ) & 588 @xmath37 13 ( @xmath38=173 ) & 493 @xmath37 28 ( @xmath38=372 ) + friends & 20 & 1153 @xmath37 41 ( @xmath38=183 ) & 820 @xmath37 43 ( @xmath38=192 ) & 616 @xmath37 43 ( @xmath38=192 ) & 391 @xmath37 70 ( @xmath38=311 ) + @xmath138 & & 7.47 & 7.99 & 4.63 & 7.26 + @xmath33 & & 7.49@xmath52 & 3.72@xmath52 & 0.00345 & 9.96@xmath52 + @xmath35 & & 0.0636 & 0.0677 & 0.0404 & 0.0619
+ @xmath34 & & 0.832 & 0.83 & 0.67 & 0.61 + @xmath14 & @xmath17 & @xmath36 + colleagues & 299 & 1287 @xmath37 8.8 ( @xmath38=152 ) & 852 @xmath37 13 ( @xmath38=220 ) & 724 @xmath37 9.3 ( @xmath38=160 ) & 329 @xmath37 16 ( @xmath38=273 ) + families & 35 & 1234 @xmath37 39 ( @xmath38=229 ) & 891 @xmath37 63 ( @xmath38=375 ) & 571 @xmath37 31 ( @xmath38=182 ) & 537 @xmath37 79 ( @xmath38=467 ) + friends & 114 & 1187 @xmath37 19 ( @xmath38=198 ) & 811 @xmath37 17 ( @xmath38=186 ) & 676 @xmath37 14 ( @xmath38=147 ) & 335 @xmath37 23 ( @xmath38=246 ) + @xmath139 & & 14.5 & 2.1 & 16.1 & 8.35 + @xmath33 & & 8.13@xmath53 & 0.124 & 1.76@xmath53 & 0.000276 + @xmath35 & & 0.0611 & 0.00933 & 0.0675 & 0.0362 + @xmath34 & & 0.607 & 0.329 & 0.94 & 0.698 + tables [ table2d1 ] and [ table2d2 ] show the observables dependence for fixed minimum age ranges ( 20 - 29 and 50 - 59 years , respectively ) .
the major trends exposed in the main text are present even when the age is kept fixed .
we may notice that @xmath36 assumes a very high value in families even when children are not present .
@xmath33 and @xmath34 values for the relation feature at different minimum age ranges are shown in , respectively , tables [ table2d3 ] and [ table2d4 ] .
@xmath14 & @xmath17 & @xmath36 + colleagues & 86 & 1255 @xmath37 18 ( @xmath38=165 ) & 813 @xmath37 20 ( @xmath38=185 ) & 706 @xmath37 16 ( @xmath38=144 ) & 291 @xmath37 24 ( @xmath38=219 ) + couples & 74 & 1115 @xmath37 19 ( @xmath38=165 ) & 711 @xmath37 27 ( @xmath38=229 ) & 600 @xmath37 18 ( @xmath38=154 ) & 281 @xmath37 28 ( @xmath38=243 ) + families & 23 & 1109 @xmath37 39 ( @xmath38=187 ) & 877 @xmath37 58 ( @xmath38=277 ) & 581 @xmath37 37 ( @xmath38=177 ) & 527 @xmath37 78 ( @xmath38=373 ) + friends & 164 & 1186 @xmath37 13 ( @xmath38=164 ) & 801 @xmath37 16 ( @xmath38=208 ) & 683 @xmath37 10 ( @xmath38=128 ) & 298 @xmath37 21 ( @xmath38=265 ) + @xmath140 & & 10.9 & 5.05 & 11.1 & 5.9 + @xmath33 & & 7.21@xmath53 & 0.00195 & 5.89@xmath53 & 0.00062 + @xmath35 & & 0.0872 & 0.0423 & 0.0883 & 0.0491 + @xmath34 & & 0.861 & 0.684 & 0.825 & 0.882 + @xmath14 & @xmath17 & @xmath36 + colleagues & 52 & 1274 @xmath37 22 ( @xmath38=159 ) & 844 @xmath37 24 ( @xmath38=172 ) & 700 @xmath37 20 ( @xmath38=142 ) & 350 @xmath37 36 ( @xmath38=263 ) + families & 28 & 1048 @xmath37 32 ( @xmath38=169 ) & 846 @xmath37 55 ( @xmath38=289 ) & 562 @xmath37 34 ( @xmath38=182 ) & 492 @xmath37 78 ( @xmath38=410 ) + friends & 22 & 1051 @xmath37 36 ( @xmath38=167 ) & 759 @xmath37 44 ( @xmath38=208 ) & 637 @xmath37 24 ( @xmath38=115 ) & 308 @xmath37 59 ( @xmath38=276 ) + @xmath141 & & 23.4 & 1.29 & 7.63 & 2.55 + @xmath33 & & @xmath40 & 0.28 & 0.000825 & 0.0835 + @xmath35 & & 0.321 & 0.0254 & 0.134 & 0.0489 + @xmath34 & & 1.39 & 0.339 & 0.878 & 0.514 + @xmath17 & @xmath36 + 10 - 19 years & 0.558 & 0.049 & 0.615 & 0.0313 + 20 - 29 years & 7.21@xmath53 & 0.00195 & 5.89@xmath53 & 0.00062 + 30 - 39 years & @xmath40 & 0.0128 & @xmath40 & 0.0513 + 40 - 49 years & 1.02@xmath52 & 0.39 & 0.000266 & 0.537 + 50 - 59 years & @xmath40 & 0.28 & 0.000825 & 0.0835 + 60 - 69 years & 0.0525 & 0.248 & 0.745 & 0.388 + @xmath45 70 years & 0.385 & 0.251 & 0.198 & 0.237 + @xmath17 & @xmath36 + 10 - 19 years & 0.888 & 3.23 & 0.609 & 2.49 + 20 - 29 years & 0.861 & 0.684 & 0.825 & 0.882 + 30 - 39 years & 1.48 & 0.77 & 1.05 & 0.636 + 40 - 49 years & 1.2 & 0.529 & 0.848 & 0.557 + 50 - 59 years & 1.39 & 0.339 & 0.878 & 0.514 + 60 - 69 years & 1.21 & 1.12 & 0.386 & 0.669 + @xmath45 70 years & 0.612 & 0.844 & 0.937 & 0.865 + tables [ table2e1 ] and [ table2e2 ] show the observables dependence for fixed average height ranges ( 160 - 170 and 170 - 180 cm , respectively ) , showing that the distributions are still different in a significant way , and that the major patterns exposed in the main text are confirmed . tables [ table2e3 ] and [ table2e4 ] show respectively the @xmath33 and @xmath34 values for all height ranges .
@xmath34 values are always high showing that some reduced @xmath33 values are due to the low number of groups in some ranges .
@xmath14 & @xmath17 & @xmath36 + colleagues & 89 & 1240 @xmath37 16 ( @xmath38=149 ) & 862 @xmath37 26 ( @xmath38=249 ) & 686 @xmath37 18 ( @xmath38=167 ) & 380 @xmath37 37 ( @xmath38=350 ) + couples & 47 & 1106 @xmath37 28 ( @xmath38=191 ) & 731 @xmath37 33 ( @xmath38=226 ) & 622 @xmath37 28 ( @xmath38=189 ) & 287 @xmath37 30 ( @xmath38=204 ) + families & 121 & 1090 @xmath37 18 ( @xmath38=196 ) & 854 @xmath37 27 ( @xmath38=295 ) & 593 @xmath37 15 ( @xmath38=169 ) & 487 @xmath37 34 ( @xmath38=371 ) + friends & 172 & 1135 @xmath37 13 ( @xmath38=169 ) & 798 @xmath37 17 ( @xmath38=221 ) & 659 @xmath37 10 ( @xmath38=131 ) & 321 @xmath37 23 ( @xmath38=302 ) + @xmath142 & & 13.3 & 3.96 & 7.08 & 7.42 + @xmath33 & & 2.47@xmath43 & 0.0084 & 0.000119 & 7.42@xmath52 + @xmath35 & & 0.086 & 0.0272 & 0.0476 & 0.0498 + @xmath34 & & 0.843 & 0.542 & 0.551 & 0.599 + @xmath14 & @xmath17 & @xmath36 + colleagues & 231 & 1293 @xmath37 10 ( @xmath38=157 ) & 859 @xmath37 15 ( @xmath38=232 ) & 738 @xmath37 10 ( @xmath38=157 ) & 325 @xmath37 18 ( @xmath38=274 ) + couples & 45 & 1089 @xmath37 22 ( @xmath38=145 ) & 700 @xmath37 33 ( @xmath38=219 ) & 576 @xmath37 14 ( @xmath38=95.4 ) & 300 @xmath37 39 ( @xmath38=264 ) + families & 56 & 1107 @xmath37 22 ( @xmath38=166 ) & 818 @xmath37 31 ( @xmath38=234 ) & 557 @xmath37 20 ( @xmath38=148 ) & 462 @xmath37 48 ( @xmath38=361 ) + friends & 71 & 1162 @xmath37 20 ( @xmath38=166 ) & 811 @xmath37 19 ( @xmath38=156 ) & 679 @xmath37 17 ( @xmath38=145 ) & 328 @xmath37 26 ( @xmath38=215 ) + @xmath143 & & 38.9 & 6.77 & 31.5 & 4.13 + @xmath33 & & @xmath40 & 0.000183 & @xmath40 & 0.00672 + @xmath35 & & 0.226 & 0.0485 & 0.192 &
0.0301 + @xmath34 & & 1.32 & 0.692 & 1.16 & 0.503 + @xmath17 & @xmath36 + @xmath49 140 cm & 0.362 & 0.108 & 0.61 & 0.0849 + 140 - 150 cm & 0.12 & 0.181 & 0.785 & 0.299 + 150 - 160 cm & 0.842 & 0.133 & 0.803 & 0.0402 + 160 - 170 cm & 2.47@xmath43 & 0.0084 & 0.000119 & 7.42@xmath52 + 170 - 180 cm & @xmath40 & 0.000183 & @xmath40 & 0.00672 + @xmath50 180 cm & 0.00432 & 0.551 & 0.126 &
0.951 + @xmath17 & @xmath36 + @xmath49 140 cm & 0.767 & 1.41 & 0.423 & 1.53 + 140 - 150 cm & 0.977 & 0.798 & 0.159 & 0.612 + 150 - 160 cm & 0.405 & 0.658 & 0.521 & 0.594 + 160 - 170 cm & 0.843 & 0.542 & 0.551 & 0.599 + 170 - 180
cm & 1.32 & 0.692 & 1.16 & 0.503 + @xmath50 180 cm & 2.78 & 0.972 & 1.41 & 0.331 + tables [ table3c1 ] and [ table3c2 ] show the dependence on gender of observables at fixed density ranges ( @xmath123 ped / m@xmath1 and @xmath124 ped / m@xmath1 , respectively ) .
we may see that only the @xmath14 observable loses the statistically significant gender dependence at high density ( but still shows it at lower density , when pedestrians may move more freely ; furthermore , the effect size is almost not affected by density ) , while the other observables preserve it at any density .
tables [ table3c3 ] and [ table3c4 ] show the dependence of , respectively , the gender @xmath33 and @xmath34 values at different density values .
@xmath14 & @xmath17 & @xmath36 + two females & 160 & 1095 @xmath37 17 ( @xmath38=219 ) & 818 @xmath37 21 ( @xmath38=267 ) & 669 @xmath37 11 ( @xmath38=138 ) & 337 @xmath37 27 ( @xmath38=346 ) + mixed & 217 & 1112 @xmath37 13 ( @xmath38=196 ) & 870 @xmath37 23 ( @xmath38=340 ) & 642 @xmath37 13 ( @xmath38=194 ) & 448 @xmath37 28 ( @xmath38=409 ) + two males & 259 & 1254 @xmath37 12 ( @xmath38=196 ) & 914 @xmath37 18 ( @xmath38=283 ) & 733 @xmath37 13 ( @xmath38=217 ) & 404 @xmath37 22 ( @xmath38=351 ) + @xmath144 & & 41.5 & 5.06 & 14.1 & 4.11 + @xmath33 & & @xmath40 & 0.00658 & 1.04@xmath127 & 0.0169 + @xmath35 & & 0.116 & 0.0157 & 0.0426 & 0.0128 + @xmath34 & & 0.771 & 0.346 & 0.441 & 0.289 + @xmath14 & @xmath17 & @xmath36 + two females & 35 & 1073 @xmath37 28 ( @xmath38=164 ) & 714 @xmath37 26 ( @xmath38=152 ) & 572 @xmath37 18 ( @xmath38=107 ) & 318 @xmath37 39 ( @xmath38=230 ) + mixed & 73 & 1062 @xmath37 18 ( @xmath38=152 ) & 782 @xmath37 29 ( @xmath38=247 ) & 521 @xmath37 20 ( @xmath38=172 ) & 448 @xmath37 42 ( @xmath38=361 ) + two males & 59 & 1171 @xmath37 23 ( @xmath38=178 ) & 767 @xmath37 19 ( @xmath38=147 ) & 644 @xmath37 18 ( @xmath38=136 ) & 304 @xmath37 28 ( @xmath38=218 ) + @xmath145 & & 7.81 & 1.4 & 11.2 & 4.6 + @xmath33 & & 0.000578 & 0.249 & 2.88@xmath52 & 0.0114 + @xmath35 & & 0.0869 & 0.0168 & 0.12 & 0.0531 + @xmath34 & & 0.665 & 0.308 & 0.786 & 0.471 + & @xmath36 + 0 - 0.05 ped / m@xmath1 & @xmath40 & 0.00658 & 1.04@xmath127 & 0.0169 + 0.05 - 0.1 ped / m@xmath1 & @xmath40 & 0.0448 & @xmath40 & 0.00164 + 0.1 - 0.15 ped / m@xmath1 & @xmath40 & 0.897 & 9.41@xmath43 & 0.00478 + 0.15 - 0.2 ped / m@xmath1 & 0.000578 & 0.249 & 2.88@xmath52 & 0.0114 + 0.2 - 0.25 ped / m@xmath1 & 0.0304 & 0.0628 & 0.31 & 0.43 + & @xmath36 + 0 - 0.05 ped / m@xmath1 & 0.771 & 0.346 & 0.441 & 0.289 + 0.05 - 0.1
ped / m@xmath1 & 0.771 & 0.235 & 0.537 & 0.291 + 0.1 - 0.15 ped / m@xmath1 & 0.737 & 0.0554 & 0.582 & 0.35 + 0.15 - 0.2 ped / m@xmath1 & 0.665 & 0.308 & 0.786 & 0.471 + 0.2 - 0.25 ped / m@xmath1 & 1.31 & 1.56 & 0.942 & 0.751 + tables [ table3b1 ] , [ table3b2 ] and [ table3b3 ] show the gender dependence of observables in , respectively , colleagues , families and friends ( couples are not shown being exclusively of mixed gender ) .
@xmath14 & @xmath17 & @xmath36 + two females & 24 & 1167 @xmath37 30 ( @xmath38=145 ) & 735 @xmath37 26 ( @xmath38=128 ) & 664 @xmath37 20 ( @xmath38=95.9 ) & 238 @xmath37 34 ( @xmath38=168 ) + mixed & 35 & 1228 @xmath37 30 ( @xmath38=175 ) & 923 @xmath37 55 ( @xmath38=327 ) & 702 @xmath37 27 ( @xmath38=158 ) & 440 @xmath37 75 ( @xmath38=445 ) + two males & 299 & 1287 @xmath37 8.8 ( @xmath38=152 ) & 852 @xmath37 13 ( @xmath38=220 ) & 724 @xmath37 9.3 ( @xmath38=160 ) & 329 @xmath37 16 ( @xmath38=273 ) + @xmath146 & & 8.49 & 4.82 & 1.78 & 3.7 + @xmath33 & & 0.00025 & 0.00862 & 0.17 & 0.0256 + @xmath35 & & 0.0457 & 0.0264 & 0.00995 & 0.0204 + @xmath34 & & 0.798 & 0.709 & 0.38 & 0.561 + @xmath14 & @xmath17 & @xmath36 + two females & 28 & 1023 @xmath37 32 ( @xmath38=171 ) & 847 @xmath37 58 ( @xmath38=305 ) & 565 @xmath37 27 ( @xmath38=140 ) & 488 @xmath37 77 ( @xmath38=405 ) + mixed & 183 & 1078 @xmath37 13 ( @xmath38=182 ) & 860 @xmath37 21 ( @xmath38=285 ) & 588 @xmath37 13 ( @xmath38=173 ) & 493 @xmath37 28 ( @xmath38=372 ) + two males & 35 & 1234 @xmath37 39 ( @xmath38=229 ) & 891 @xmath37 63 ( @xmath38=375 ) & 571 @xmath37 31 ( @xmath38=182 ) & 537 @xmath37 79 ( @xmath38=467 ) + @xmath147 & & 12.3 & 0.197 & 0.308 & 0.198 + @xmath33 & & 8.41@xmath127 & 0.821 & 0.735 & 0.821 + @xmath35 & & 0.0917 & 0.00162 & 0.00253 & 0.00163 + @xmath34 & & 1.03 & 0.128 & 0.135 & 0.112 + @xmath14 & @xmath17 & @xmath36 + two females & 184 & 1105 @xmath37 15 ( @xmath38=197 ) & 777 @xmath37 15 ( @xmath38=205 ) & 658 @xmath37 8.5 ( @xmath38=115 ) & 293 @xmath37 20 ( @xmath38=274 ) + mixed & 20 & 1153 @xmath37 41 ( @xmath38=183 ) & 820 @xmath37 43 ( @xmath38=192 ) & 616 @xmath37 43 ( @xmath38=192 ) & 391 @xmath37 70 ( @xmath38=311 ) + two males & 114 & 1187 @xmath37 19 ( @xmath38=198 ) & 811 @xmath37 17 ( @xmath38=186 ) & 676 @xmath37 14 ( @xmath38=147 ) & 335 @xmath37 23 ( @xmath38=246 ) + @xmath148 & & 6.02 & 1.27 & 1.91 & 1.76 + @xmath33 & & 0.00272 & 0.283 & 0.15 & 0.173 + @xmath35 & & 0.0368 & 0.00798 & 0.012 & 0.0111 + @xmath34 & & 0.412 & 0.213 & 0.388 & 0.354 + we may see that males dyads are farther and faster than female ones regardless of relation ( although the differences in @xmath14 , @xmath17 and @xmath36 are quite reduced in families and friends ) .
mixed dyad behaviour , on the other hand , depends strongly on relation .
mixed dyads are the only ones including couples , and this affects strongly their behaviour , and represent also the largest part of families .
they are very little represented in friends ( interestingly , mixed dyads of friends walk much closer , in abreast distance , than same sex dyads , although their absolute distance is higher than in females ) .
the `` colleagues '' category could represent a fair field for comparing the effect of gender , and in it the mixed behaviour is somehow in between the two sexes ( although the absolute distance @xmath14 and group depth @xmath36 are very large , suggesting not abreast formations ) but in our set colleagues are extremely biased towards males , and thus the analysis in hindered by low female and mixed dyads numbers .
finally we may notice that in families and friends , the effect of gender on distance ( @xmath14 , @xmath17 and @xmath36 ) is very reduced , but the one on velocity is persistent .
the velocity effect size in families is nevertheless more than two times the one for friends .
tables [ table3d1 ] and [ table3d2 ] show the dependence on gender of observables at fixed average age ranges ( 20 - 29 years and 50 - 59 years , respectively ) .
interestingly , the differences between young two females and two males dyads are reduced ( and almost absent regarding the distance observables @xmath14 , @xmath17 and @xmath36 ) , while they are very strong in elder groups .
young mixed dyad behaviour is strongly influenced by the presence of couples .
@xmath14 & @xmath17 & @xmath36 + two females & 111 & 1166 @xmath37 16 ( @xmath38=170 ) & 791 @xmath37 21 ( @xmath38=220 ) & 686 @xmath37 10 ( @xmath38=110 ) & 275 @xmath37 26 ( @xmath38=271 ) + mixed & 125 & 1122 @xmath37 16 ( @xmath38=175 ) & 784 @xmath37 23 ( @xmath38=255 ) & 612 @xmath37 16 ( @xmath38=182 ) & 360 @xmath37 27 ( @xmath38=307 ) + two males & 134 & 1247 @xmath37 14 ( @xmath38=164 ) & 803 @xmath37 17 ( @xmath38=201 ) & 689 @xmath37 13 ( @xmath38=148 ) & 301 @xmath37 20 ( @xmath38=235 ) + @xmath149 & & 18 & 0.235 & 10.5 & 3.03 + @xmath33 & & 3.37@xmath43 & 0.791 & 3.77@xmath52 & 0.0496 + @xmath35 & & 0.0895 & 0.00128 & 0.054 & 0.0162 + @xmath34 & & 0.739 & 0.0832 & 0.47 & 0.291 + @xmath14 & @xmath17 & @xmath36 + two females & 20 & 1010 @xmath37 30 ( @xmath38=136 ) & 708 @xmath37 21 ( @xmath38=95.5 ) & 613 @xmath37 27 ( @xmath38=121 ) & 254 @xmath37 34 ( @xmath38=151 ) + mixed & 34 & 1071 @xmath37 29 ( @xmath38=170 ) & 856 @xmath37 48 ( @xmath38=278 ) & 608 @xmath37 32 ( @xmath38=189 ) & 462 @xmath37 66 ( @xmath38=388 ) + two males & 60 & 1255 @xmath37 22 ( @xmath38=168 ) & 847 @xmath37 25 ( @xmath38=192 ) & 686 @xmath37 18 ( @xmath38=141 ) & 369 @xmath37 38 ( @xmath38=298 ) + @xmath150 & & 22.8 & 3.69 & 3.37 & 2.81 + @xmath33 & & @xmath40 & 0.0281 & 0.0379 & 0.0643 + @xmath35 & & 0.291 & 0.0623 & 0.0573 & 0.0482 + @xmath34 & & 1.52 & 0.646 & 0.486 & 0.646 + tables [ table3d3 ] and [ table3d4 ] show the @xmath33 and @xmath34 values for gender in different average age ranges .
minimum ages ranges are shown in tables [ table3d5 ] and [ table3d6 ] .
@xmath17 & @xmath36 + 10 - 19 years & 0.0301 & 0.685 & 0.573 & 0.903 + 20 - 29 years & 3.37@xmath43 & 0.791 & 3.77@xmath52 & 0.0496 + 30 - 39 years & @xmath40 & 0.0477 & 7.66@xmath43 & 0.0433 + 40 - 49 years & @xmath40 & 0.106 & 0.000167 & 0.856 + 50 - 59 years & @xmath40 & 0.0281 & 0.0379 & 0.0643 + 60 - 69 years & 0.00145 & 0.495 & 0.17 & 0.655 + @xmath45 70 years & 0.245 & 0.564 & 0.543 & 0.598 + @xmath17 & @xmath36 + 10 - 19 years & 0.769 & 0.241 & 0.309 & 0.14 + 20 - 29 years & 0.739 & 0.0832 & 0.47 & 0.291 + 30 - 39 years & 0.87 & 0.48 & 0.732 & 0.322 + 40 - 49 years & 1.3 & 0.329 & 0.619 & 0.0946 + 50 - 59 years & 1.52 & 0.646 & 0.486 & 0.646 + 60 - 69 years & 1.49 & 0.457 & 0.666 & 0.27 + @xmath45 70 years & 1.56 & 0.802 & 0.915 & 0.788 + @xmath17 & @xmath36 + 0 - 9 years & 0.0872 & 0.17 & 0.577 & 0.198 + 10 - 19 years & 0.00563 & 0.497 & 0.981 & 0.484 + 20 - 29 years & 1.67@xmath53 & 0.665 & 8.91@xmath52 & 0.0654 + 30 - 39 years & @xmath40 & 0.0904 & 1.99@xmath43 & 0.027 + 40 - 49 years & 3.02@xmath43 & 0.193 & 0.000602 & 0.778 + 50 - 59 years & 3.78@xmath43 & 0.0458 & 0.0555 & 0.0743 + 60 - 69 years & 0.00245 & 0.446 & 0.105 & 0.651 + @xmath45 70 years & 0.245 & 0.564 & 0.543 & 0.598 + @xmath17 & @xmath36 + 0 - 9 years & 1.11 & 1.21 & 0.449 & 1.17 + 10 - 19 years & 0.949 & 0.473 & 0.0487 & 0.421 + 20 - 29 years & 0.715 & 0.111 & 0.475 & 0.271 + 30 - 39 years & 0.907 & 0.396 & 0.749 & 0.306 + 40 - 49 years & 1.62 & 0.343 & 0.65 & 0.113 + 50 - 59 years & 1.47 & 0.616 & 0.487 & 0.659 + 60 - 69 years & 1.45 & 0.541 & 0.79 & 0.275 + @xmath45 70 years & 1.56 & 0.802 & 0.915 &
0.788 + tables [ table3e1 ] and [ table3e2 ] show the dependence on gender of observables at fixed average height ranges ( 150 - 160 and 170 - 180 cm , respectively ) . the results ( in particular for the shorter height range the effect size
, that helps in dealing with the reduced number of groups ) show that differences between the sexes are still present when we consider individuals of similar height .
@xmath14 & @xmath17 & @xmath36 + two females & 75 & 1094 @xmath37 27 ( @xmath38=232 ) & 791 @xmath37 26 ( @xmath38=225 ) & 627 @xmath37 15 ( @xmath38=131 ) & 352 @xmath37 36 ( @xmath38=316 ) + mixed & 25 & 1045 @xmath37 35 ( @xmath38=176 ) & 796 @xmath37 43 ( @xmath38=217 ) & 603 @xmath37 33 ( @xmath38=167 ) & 376 @xmath37 60 ( @xmath38=300 ) + two males & 18 & 1272 @xmath37 82 ( @xmath38=346 ) & 921 @xmath37 92 ( @xmath38=390 ) & 674 @xmath37 42 ( @xmath38=180 ) & 493 @xmath37 110 ( @xmath38=447 ) + @xmath151 & & 4.95 & 1.89 & 1.22 & 1.24 + @xmath33 & & 0.00869 & 0.156 & 0.3 & 0.294 + @xmath35 & & 0.0792 & 0.0318 & 0.0207 & 0.0211 + @xmath34 & & 0.873 & 0.493 & 0.415 & 0.408 + @xmath14 & @xmath17 & @xmath36 + two females & 16 & 1091 @xmath37 48 ( @xmath38=191 ) & 741 @xmath37 30 ( @xmath38=119 ) & 662 @xmath37 22 ( @xmath38=88.5 ) & 238 @xmath37 32 ( @xmath38=128 ) + mixed & 121 & 1127 @xmath37 16 ( @xmath38=171 ) & 797 @xmath37 23 ( @xmath38=250 ) & 598 @xmath37 14 ( @xmath38=149 ) & 396 @xmath37 31 ( @xmath38=338 ) + two males & 284 & 1270 @xmath37 9.6 ( @xmath38=161 ) & 846 @xmath37 13 ( @xmath38=213 ) & 723 @xmath37 9.4 ( @xmath38=159 ) & 324 @xmath37 15 ( @xmath38=258 ) + @xmath152 & & 36.7 & 3.4 & 27.8 & 3.89 + @xmath33 & & @xmath40 & 0.0344 & @xmath40 & 0.0212 + @xmath35 & & 0.149 & 0.016 & 0.117 & 0.0183 + @xmath34 & & 1.1 & 0.502 & 0.799 & 0.492 + tables [ table3e3 ] and [ table3e4 ] show , respectively , the gender @xmath33 and @xmath34 values for different average height ranges .
@xmath17 & @xmath36 + @xmath49 140 cm & 0.614 & 0.0596 & 0.958 & 0.148 + 140 - 150 cm & 0.000737 & 0.372 & 0.0226 & 0.306 + 150 - 160 cm & 0.00869 & 0.156 & 0.3 & 0.294 + 160 - 170 cm & 1.0 @xmath52 & 0.0653 & 0.0212 & 0.000455 + 170 - 180 cm & @xmath40 & 0.0344 & @xmath40 & 0.0212 + @xmath50 180 cm & 0.0241 & 0.191 & 0.137 &
0.647 + @xmath17 & @xmath36 + @xmath49 140 cm & 0.774 & 1.89 & 0.183 & 1.39 + 140 - 150 cm & 2.06 & 0.693 & 1.47 & 0.868 + 150 - 160 cm & 0.873 & 0.493 & 0.415 & 0.408 + 160 - 170 cm & 0.523 & 0.279 & 0.235 & 0.416 + 170 - 180 cm & 1.1 & 0.502 & 0.799 &
0.492 + @xmath50 180 cm & 1.28 & 0.708 & 0.811 & 0.246 + tables [ table4fur_dl ] and [ table4fur_dh ] show the age dependence of observables in , respectively , the @xmath47 ped / m@xmath1 and @xmath48 ped / m@xmath1 density ranges .
results mostly reflect those of the main text , with high or relatively high @xmath34 values suggesting that some not very good @xmath33 values may be due to the scarcity of data in the children and elderly categories ( i.e. the categories with the most different behaviour ) .
a remarkable feature , presented with the caveats related to sensor noise in the tracking of children , is that while in general velocity decreases with density , this is not true for dyads with children , as shown in figure [ fastchild ] in the main text .
@xmath14 & @xmath17 & @xmath36 + 0 - 9 years & 9 & 1075 @xmath37 68 ( @xmath38=205 ) & 1078 @xmath37 97 ( @xmath38=291 ) & 663 @xmath37 62 ( @xmath38=186 ) & 704 @xmath37 140 ( @xmath38=414 ) + 10 - 19 years & 44 & 1175 @xmath37 43 ( @xmath38=288 ) & 802 @xmath37 40 ( @xmath38=262 ) & 661 @xmath37 25 ( @xmath38=167 ) & 337 @xmath37 44 ( @xmath38=294 ) + 20 - 29 years & 184 & 1198 @xmath37 14 ( @xmath38=196 ) & 853 @xmath37 23 ( @xmath38=313 ) & 694 @xmath37 14 ( @xmath38=193 ) & 357 @xmath37 27 ( @xmath38=372 ) + 30 - 39 years & 185 & 1196 @xmath37 16 ( @xmath38=217 ) & 894 @xmath37 22 ( @xmath38=306 ) & 696 @xmath37 16 ( @xmath38=223 ) & 418 @xmath37 27 ( @xmath38=368 ) + 40 - 49 years & 87 & 1150 @xmath37 20 ( @xmath38=191 ) & 909 @xmath37 32 ( @xmath38=297 ) & 683 @xmath37 22 ( @xmath38=202 ) & 440 @xmath37 42 ( @xmath38=395 ) + 50 - 59 years & 71 & 1157 @xmath37 23 ( @xmath38=198 ) & 844 @xmath37 27 ( @xmath38=228 ) & 678 @xmath37 18 ( @xmath38=149 ) & 381 @xmath37 38 ( @xmath38=320 ) + 60 - 69 years & 47 & 1022 @xmath37 25 ( @xmath38=174 ) & 912 @xmath37 51 ( @xmath38=348 ) & 670 @xmath37 27 ( @xmath38=182 ) & 481 @xmath37 62 ( @xmath38=424 ) + @xmath45 70 years & 9 & 891 @xmath37 31 ( @xmath38=92.6 ) & 815 @xmath37 100 ( @xmath38=307 ) & 605 @xmath37 19 ( @xmath38=55.8 ) & 411 @xmath37 130 ( @xmath38=394 ) + @xmath153 & & 6.98 & 1.61 & 0.527 & 1.94 + @xmath33 & & 4.97@xmath43 & 0.129 & 0.815 & 0.0613 + @xmath35 & & 0.0722 & 0.0176 & 0.00584 & 0.0211 + @xmath34 & & 1.59 & 1.03 & 0.419 & 1.16 + @xmath14 & @xmath17 & @xmath36 + 0 - 9 years & 6 & 1284 @xmath37 110 ( @xmath38=258 ) & 693 @xmath37 51 ( @xmath38=126 ) & 485 @xmath37 47 ( @xmath38=116 ) & 401 @xmath37 92 ( @xmath38=225 ) + 10 - 19 years & 14 & 1146 @xmath37 47 ( @xmath38=176 ) & 806 @xmath37 65 ( @xmath38=244 ) & 571 @xmath37 39 ( @xmath38=147 ) & 426 @xmath37 91 ( @xmath38=341 ) + 20 - 29 years & 72 & 1099 @xmath37 16 ( @xmath38=133 ) & 745 @xmath37 20 ( @xmath38=167 ) & 598 @xmath37 17 ( @xmath38=145 ) & 322 @xmath37 30 ( @xmath38=255 ) + 30 - 39 years & 39 & 1102 @xmath37 31 ( @xmath38=192 ) & 766 @xmath37 33 ( @xmath38=208 ) & 575 @xmath37 24 ( @xmath38=149 ) & 372 @xmath37 50 ( @xmath38=313 ) + 40 - 49 years & 17 & 1121 @xmath37 32 ( @xmath38=131 ) & 763 @xmath37 37 ( @xmath38=152 ) & 547 @xmath37 42 ( @xmath38=172 ) & 403 @xmath37 68 ( @xmath38=279 ) + 50 - 59 years & 10 & 1057 @xmath37 53 ( @xmath38=167 ) & 739 @xmath37 60 ( @xmath38=190 ) & 485 @xmath37 58 ( @xmath38=184 ) & 416 @xmath37 110 ( @xmath38=343 ) + 60 - 69 years & 7 & 1021 @xmath37 75 ( @xmath38=199 ) & 967 @xmath37 130 ( @xmath38=354 ) & 616 @xmath37 84 ( @xmath38=222 ) & 618 @xmath37 160 ( @xmath38=423 ) + @xmath45 70 years & 2 & 760 @xmath37 18 ( @xmath38=25.4 ) & 644 @xmath37 22 ( @xmath38=31 ) & 585 @xmath37 8.9 ( @xmath38=12.5 ) & 185 @xmath37 52 ( @xmath38=73.5 ) + @xmath154 & & 2.76 & 1.46 & 1.11 & 1.21 + @xmath33 & & 0.00993 & 0.184 & 0.359 & 0.299 + @xmath35 & & 0.108 & 0.0606 & 0.0466 & 0.0506 + @xmath34 & & 2.22 & 0.987 & 0.652 & 1.1 + tables [ table4fur_d1 ] and [ table4fur_d2 ] show , respectively , the minimum age @xmath33 and @xmath34 values in different density ranges & @xmath36 + 0 - 0.05 ped / m@xmath1 & 4.97@xmath43 & 0.129 & 0.815 & 0.0613 + 0.05 - 0.1 ped / m@xmath1 & @xmath40 & 0.00286 & 0.0232 & 1.26@xmath127 + 0.1 - 0.15
ped / m@xmath1 & 8.51@xmath43 & 0.0207 & 0.00346 & 7.22@xmath127 + 0.15 - 0.2 ped / m@xmath1 & 0.00993 & 0.184 & 0.359 & 0.299 + 0.2 - 0.25 ped / m@xmath1 & 0.651 & 0.504 & 0.118 & 0.328 + @xmath17 & @xmath36 + 0 - 0.05 ped / m@xmath1 & 1.59 & 1.03 & 0.419 & 1.16 + 0.05 - 0.1 ped / m@xmath1 & 1.6 & 0.941 & 0.605 & 1.36 + 0.1 - 0.15 ped / m@xmath1 & 2.25 & 0.689 & 0.93 & 0.924 + 0.15 - 0.2 ped / m@xmath1 & 2.22 & 0.987 & 0.652 & 1.1 + 0.2 - 0.25 ped / m@xmath1 & 0.513 & 2.14 & 1.15 & 1.14 + tables [ table4fur_col ] , [ table4fur_coup ] , [ table4fur_fam ] and [ table4fur_fri ] show the age dependence of observables in , respectively , colleagues , couples , families and friends .
observable values almost have no age dependence in the 20 - 59 years age ( with the exclusion of friend velocity ) .
it is interesting to note that the @xmath36 distribution assumes a larger value for families even when only adults are involved .
another interesting , although expectable , result is that while dyads with teenagers are very abreast in the friends category , they are not abreast in the family one ( the @xmath36 values is almost doubled in families ) .
@xmath14 & @xmath17 & @xmath36 + 20 - 29 years & 86 & 1255 @xmath37 18 ( @xmath38=165 ) & 813 @xmath37 20 ( @xmath38=185 ) & 706 @xmath37 16 ( @xmath38=144 ) & 291 @xmath37 24 ( @xmath38=219 ) + 30 - 39 years & 145 & 1293 @xmath37 14 ( @xmath38=167 ) & 861 @xmath37 21 ( @xmath38=257 ) & 734 @xmath37 14 ( @xmath38=165 ) & 331 @xmath37 25 ( @xmath38=301 ) + 40 - 49 years & 71 & 1258 @xmath37 14 ( @xmath38=119 ) & 870 @xmath37 28 ( @xmath38=236 ) & 714 @xmath37 19 ( @xmath38=161 ) & 363 @xmath37 38 ( @xmath38=324 ) + 50 - 59 years & 52 & 1274 @xmath37 22 ( @xmath38=159 ) & 844 @xmath37 24 ( @xmath38=172 ) & 700 @xmath37 20 ( @xmath38=142 ) & 350 @xmath37 36 ( @xmath38=263 ) + 60 - 69 years & 4 & 1217 @xmath37 36 ( @xmath38=72 ) & 1075 @xmath37 220 ( @xmath38=433 ) & 692 @xmath37 100 ( @xmath38=208 ) & 617 @xmath37 320 ( @xmath38=632 ) + @xmath155 & & 1.17 & 1.72 & 0.702 & 1.62 + @xmath33 & & 0.326 & 0.146 & 0.591 & 0.168 + @xmath35 & & 0.013 & 0.0191 & 0.00789 & 0.018 + @xmath34 & & 0.461 & 1.32 & 0.252 & 1.33 + @xmath14 & @xmath17 & @xmath36 + 10 - 19 years & 2 & 958 @xmath37 180 ( @xmath38=253 ) & 919 @xmath37 53 ( @xmath38=74.7 ) & 725 @xmath37 58 ( @xmath38=81.7 ) & 480 @xmath37 180 ( @xmath38=257 ) + 20 - 29 years & 74 & 1115 @xmath37 19 ( @xmath38=165 ) & 711 @xmath37 27 ( @xmath38=229 ) & 600 @xmath37 18 ( @xmath38=154 ) & 281 @xmath37 28 ( @xmath38=243 ) + 30 - 39 years & 17 & 1049 @xmath37 37 ( @xmath38=151 ) & 670 @xmath37 35 ( @xmath38=143 ) & 572 @xmath37 32 ( @xmath38=130 ) & 274 @xmath37 30 ( @xmath38=124 ) + 40 - 49 years & 3 & 1091 @xmath37 110 ( @xmath38=187 ) & 897 @xmath37 110 ( @xmath38=198 ) & 684 @xmath37 66 ( @xmath38=115 ) & 501 @xmath37 130 ( @xmath38=223 ) + @xmath156 & & 1.2 & 1.53 & 0.966 & 1.36 + @xmath33 & & 0.315 & 0.211 & 0.412 & 0.261 + @xmath35 & & 0.0376 & 0.0476 & 0.0306 & 0.0424 + @xmath34 & & 0.946 & 1.78 & 1.19 & 1.65 + @xmath14 & @xmath17 & @xmath36 + 0 - 9 years & 31 & 1143 @xmath37 42 ( @xmath38=235 ) & 995 @xmath37 69 ( @xmath38=383 ) & 529 @xmath37 34 ( @xmath38=189 ) & 701 @xmath37 87 ( @xmath38=485 ) + 10 - 19 years & 36 & 1163 @xmath37 38 ( @xmath38=230 ) & 831 @xmath37 49 ( @xmath38=296 ) & 617 @xmath37 30 ( @xmath38=179 ) & 415 @xmath37 58 ( @xmath38=347 ) + 20 - 29 years & 23 & 1109 @xmath37 39 ( @xmath38=187 ) & 877 @xmath37 58 ( @xmath38=277 ) & 581 @xmath37 37 ( @xmath38=177 ) & 527 @xmath37 78 ( @xmath38=373 ) + 30 - 39 years & 46 & 1078 @xmath37 23 ( @xmath38=159 ) & 814 @xmath37 33 ( @xmath38=225 ) & 561 @xmath37 24 ( @xmath38=163 ) & 458 @xmath37 49 ( @xmath38=330 ) + 40 - 49 years & 41 & 1116 @xmath37 31 ( @xmath38=199 ) & 801 @xmath37 28 ( @xmath38=181 ) & 582 @xmath37 23 ( @xmath38=149 ) & 431 @xmath37 40 ( @xmath38=256 ) + 50 - 59 years & 28 & 1048 @xmath37 32 ( @xmath38=169 ) & 846 @xmath37 55 ( @xmath38=289 ) & 562 @xmath37 34 ( @xmath38=182 ) & 492 @xmath37 78 ( @xmath38=410 ) + 60 - 69 years & 37 & 1030 @xmath37 24 ( @xmath38=145 ) & 911 @xmath37 63 ( @xmath38=382 ) & 642 @xmath37 25 ( @xmath38=154 ) & 512 @xmath37 75 ( @xmath38=456 ) + @xmath45 70 years & 4 & 847 @xmath37 52 ( @xmath38=104 ) & 926 @xmath37 190 ( @xmath38=384 ) & 550 @xmath37 38 ( @xmath38=75.6 ) & 575 @xmath37 240 ( @xmath38=477 ) + @xmath157 & & 2.83 & 1.52 & 1.46 & 1.74 + @xmath33 & & 0.00758 & 0.162 & 0.182 & 0.101 + @xmath35 & & 0.0767 & 0.0427 & 0.0412 & 0.0486 + @xmath34 & & 1.42 & 0.679 & 0.659 & 0.686 + @xmath14 & @xmath17 & @xmath36 + 10 - 19 years & 23 & 1143 @xmath37 61 ( @xmath38=292 ) & 681 @xmath37 15 ( @xmath38=73.9 ) & 621 @xmath37 16 ( @xmath38=78.5 ) & 217 @xmath37 19 ( @xmath38=93.2 ) + 20 - 29 years & 164 & 1186 @xmath37 13 ( @xmath38=164 ) & 801 @xmath37 16 ( @xmath38=208 ) & 683 @xmath37 10 ( @xmath38=128 ) & 298 @xmath37 21 ( @xmath38=265 ) + 30 - 39 years & 56 & 1143 @xmath37 28 ( @xmath38=206 ) & 817 @xmath37 24 ( @xmath38=178 ) & 644 @xmath37 22 ( @xmath38=162 ) & 366 @xmath37 38 ( @xmath38=286 ) + 40 - 49 years & 19 & 1089 @xmath37 47 ( @xmath38=206 ) & 819 @xmath37 49 ( @xmath38=213 ) & 682 @xmath37 21 ( @xmath38=92.3 ) & 341 @xmath37 68 ( @xmath38=295 ) + 50 - 59 years & 22 & 1051 @xmath37 36 ( @xmath38=167 ) & 759 @xmath37 44 ( @xmath38=208 ) & 637 @xmath37 24 ( @xmath38=115 ) & 308 @xmath37 59 ( @xmath38=276 ) + 60 - 69 years & 26 & 996 @xmath37 38 ( @xmath38=192 ) & 808 @xmath37 40 ( @xmath38=202 ) & 625 @xmath37 33 ( @xmath38=169 ) & 383 @xmath37 59 ( @xmath38=299 ) + @xmath45 70 years & 8 & 906 @xmath37 32 ( @xmath38=91.4 ) & 716 @xmath37 56 ( @xmath38=159 ) & 606 @xmath37 18 ( @xmath38=52.2 ) & 290 @xmath37 85 ( @xmath38=239 ) + @xmath158 & & 7.17 & 1.82 & 1.98 & 1.29 + @xmath33 & & 3.58@xmath53 & 0.0947 & 0.0678 & 0.262 + @xmath35 & & 0.121 & 0.0339 & 0.0368 & 0.0242 + @xmath34 & & 1.73 & 0.904 & 0.608 & 0.731 + tables [ table4fur_0o ] , [ table4fur_1o ] and [ table4fur_2o ] show the age dependence of observables in , respectively , dyads with two females , mixed dyads and two males .
the results are similar to those shown in the main text .
although based on an extremely reduced number of data , it is interesting to note the large difference in velocity between two males and two females dyads with children ( mixed dyads show a value in between , probably due to the fact that they include male and female parents ) , and the very large @xmath36 ( non - abreast formation ) value assumed in two females dyads ( mixed dyads on the opposite are more abreast ) .
this values are based on very few groups , but differences are nevertheless larger than standard errors , and could reflect differences in the relation that children have with fathers and mothers ( at least in the observed cultural environment ) .
@xmath14 & @xmath17 & @xmath36 + 0 - 9 years & 6 & 985 @xmath37 88 ( @xmath38=215 ) & 1252 @xmath37 150 ( @xmath38=378 ) & 525 @xmath37 78 ( @xmath38=192 ) & 993 @xmath37 220 ( @xmath38=535 ) + 10 - 19 years & 20 & 1075 @xmath37 58 ( @xmath38=258 ) & 738 @xmath37 48 ( @xmath38=216 ) & 621 @xmath37 23 ( @xmath38=103 ) & 291 @xmath37 64 ( @xmath38=288 ) + 20 - 29 years & 110 & 1169 @xmath37 16 ( @xmath38=167 ) & 789 @xmath37 21 ( @xmath38=221 ) & 684 @xmath37 10 ( @xmath38=107 ) & 273 @xmath37 26 ( @xmath38=277 ) + 30 - 39 years & 55 & 1108 @xmath37 25 ( @xmath38=188 ) & 777 @xmath37 23 ( @xmath38=171 ) & 639 @xmath37 16 ( @xmath38=118 ) & 330 @xmath37 33 ( @xmath38=246 ) + 40 - 49 years & 24 & 1040 @xmath37 33 ( @xmath38=163 ) & 827 @xmath37 54 ( @xmath38=266 ) & 622 @xmath37 25 ( @xmath38=123 ) & 404 @xmath37 77 ( @xmath38=379 ) + 50 - 59 years & 17 & 1015 @xmath37 35 ( @xmath38=143 ) & 704 @xmath37 24 ( @xmath38=97.3 ) & 623 @xmath37 29 ( @xmath38=120 ) & 240 @xmath37 32 ( @xmath38=133 ) + 60 - 69 years & 17 & 923 @xmath37 31 ( @xmath38=130 ) & 791 @xmath37 52 ( @xmath38=213 ) & 580 @xmath37 36 ( @xmath38=149 ) & 390 @xmath37 85 ( @xmath38=349 ) + @xmath45 70 years & 3 & 958 @xmath37 14 ( @xmath38=23.6 ) & 629 @xmath37 31 ( @xmath38=53.5 ) & 587 @xmath37 40 ( @xmath38=69 ) & 186 @xmath37 22 ( @xmath38=37.4 ) + @xmath159 & & 6.27 & 4.83 & 3.74 & 5.64 + @xmath33 & & 8.87@xmath53 & 4.06@xmath52 & 0.000721 & 4.77@xmath127 + @xmath35 & & 0.153 & 0.122 & 0.0969 & 0.139 + @xmath34 & & 1.51 & 1.94 & 1.42 & 1.78 + @xmath14 & @xmath17 & @xmath36 + 0 - 9 years & 12 & 1119 @xmath37 56 ( @xmath38=193 ) & 888 @xmath37 75 ( @xmath38=259 ) & 573 @xmath37 61 ( @xmath38=212 ) & 547 @xmath37 83 ( @xmath38=287 ) + 10 - 19 years & 16 & 1060 @xmath37 45 ( @xmath38=181 ) & 840 @xmath37 54 ( @xmath38=214 ) & 620 @xmath37 36 ( @xmath38=145 ) & 417 @xmath37 78 ( @xmath38=313 ) + 20 - 29 years & 120 & 1123 @xmath37 16 ( @xmath38=175 ) & 782 @xmath37 23 ( @xmath38=251 ) & 619 @xmath37 17 ( @xmath38=182 ) & 352 @xmath37 27 ( @xmath38=301 ) + 30 - 39 years & 93 & 1143 @xmath37 19 ( @xmath38=180 ) & 834 @xmath37 30 ( @xmath38=286 ) & 601 @xmath37 19 ( @xmath38=179 ) & 435 @xmath37 40 ( @xmath38=386 ) + 40 - 49 years & 53 & 1141 @xmath37 26 ( @xmath38=191 ) & 802 @xmath37 25 ( @xmath38=178 ) & 614 @xmath37 21 ( @xmath38=150 ) & 400 @xmath37 34 ( @xmath38=245 ) + 50 - 59 years & 34 & 1078 @xmath37 29 ( @xmath38=168 ) & 848 @xmath37 47 ( @xmath38=277 ) & 604 @xmath37 32 ( @xmath38=188 ) & 455 @xmath37 66 ( @xmath38=387 ) + 60 - 69 years & 38 & 1042 @xmath37 26 ( @xmath38=160 ) & 905 @xmath37 61 ( @xmath38=378 ) & 642 @xmath37 25 ( @xmath38=152 ) & 506 @xmath37 73 ( @xmath38=451 ) + @xmath45 70 years & 5 & 831 @xmath37 44 ( @xmath38=99 ) & 868 @xmath37 160 ( @xmath38=363 ) & 563 @xmath37 33 ( @xmath38=72.8 ) & 484 @xmath37 210 ( @xmath38=463 ) + @xmath160 & & 3.64 & 1.11 & 0.387 & 1.32 + @xmath33 & & 0.000822 & 0.358 & 0.91 & 0.24 + @xmath35 & & 0.0656 & 0.0209 & 0.0074 & 0.0248 + @xmath34 & & 1.76 & 0.431 & 0.536 & 0.652 + @xmath14 & @xmath17 & @xmath36 + 0 - 9 years & 13 & 1237 @xmath37 65 ( @xmath38=233 ) & 975 @xmath37 120 ( @xmath38=425 ) & 491 @xmath37 42 ( @xmath38=153 ) & 709 @xmath37 150 ( @xmath38=540 ) + 10 - 19 years & 27 & 1277 @xmath37 48 ( @xmath38=252 ) & 803 @xmath37 58 ( @xmath38=303 ) & 628 @xmath37 34 ( @xmath38=175 ) & 376 @xmath37 65 ( @xmath38=337 ) + 20 - 29 years & 134 & 1241 @xmath37 14 ( @xmath38=156 ) & 806 @xmath37 16 ( @xmath38=180 ) & 697 @xmath37 13 ( @xmath38=148 ) & 296 @xmath37 18 ( @xmath38=208 ) + 30 - 39 years & 144 & 1280 @xmath37 16 ( @xmath38=190 ) & 860 @xmath37 18 ( @xmath38=222 ) & 732 @xmath37 14 ( @xmath38=173 ) & 331 @xmath37 22 ( @xmath38=259 ) + 40 - 49 years & 72 & 1257 @xmath37 14 ( @xmath38=122 ) & 875 @xmath37 27 ( @xmath38=233 ) & 715 @xmath37 19 ( @xmath38=159 ) & 365 @xmath37 39 ( @xmath38=328 ) + 50 - 59 years & 60 & 1254 @xmath37 22 ( @xmath38=168 ) & 846 @xmath37 25 ( @xmath38=193 ) & 683 @xmath37 19 ( @xmath38=144 ) & 373 @xmath37 38 ( @xmath38=297 ) + 60 - 69 years & 12 & 1134 @xmath37 48 ( @xmath38=167 ) & 930 @xmath37 89 ( @xmath38=307 ) & 711 @xmath37 54 ( @xmath38=189 ) & 462 @xmath37 120 ( @xmath38=406 ) + @xmath45 70 years & 4 & 902 @xmath37 48 ( @xmath38=95.7 ) & 802 @xmath37 92 ( @xmath38=183 ) & 619 @xmath37 19 ( @xmath38=37.6 ) & 410 @xmath37 140 ( @xmath38=289 ) + @xmath161 & & 3.77 & 1.82 & 5.09 & 4.07 + @xmath33 & & 0.000553 & 0.081 & 1.39@xmath52 & 0.000239 + @xmath35 & & 0.0544 & 0.0271 & 0.0722 &
0.0586 + @xmath34 & & 2.01 & 0.445 & 1.41 & 1.64 + tables [ table4fur_150 ] and [ table4fur_170 ] show the age dependence of observables in , respectively , the 150 - 160 cm and 170 - 180 cm minimum height ranges .
these data , which respect the patterns highlighted in the main text , present a sufficient number of groups in each category and are thus reliable .
in some situation , a noisy tracking of a child may cause to have a category with very poor and not reliable representation ( e.g. groups with children but with a tall minimum height ) causing some irregular behaviour in the @xmath33 and @xmath34 values of tables [ table4fur_h1 ] and [ table4fur_h2 ] .
@xmath14 & @xmath17 & @xmath36 + 10 - 19 years & 21 & 1124 @xmath37 54 ( @xmath38=246 ) & 783 @xmath37 47 ( @xmath38=215 ) & 606 @xmath37 31 ( @xmath38=143 ) & 352 @xmath37 73 ( @xmath38=333 ) + 20 - 29 years & 75 & 1157 @xmath37 21 ( @xmath38=184 ) & 800 @xmath37 27 ( @xmath38=232 ) & 668 @xmath37 14 ( @xmath38=122 ) & 311 @xmath37 35 ( @xmath38=307 ) + 30 - 39 years & 48 & 1109 @xmath37 24 ( @xmath38=168 ) & 804 @xmath37 32 ( @xmath38=223 ) & 624 @xmath37 20 ( @xmath38=139 ) & 392 @xmath37 44 ( @xmath38=305 ) + 40 - 49 years & 32 & 1108 @xmath37 39 ( @xmath38=218 ) & 828 @xmath37 37 ( @xmath38=211 ) & 589 @xmath37 27 ( @xmath38=151 ) & 452 @xmath37 58 ( @xmath38=328 ) + 50 - 59 years & 19 & 1067 @xmath37 39 ( @xmath38=171 ) & 757 @xmath37 34 ( @xmath38=146 ) & 611 @xmath37 36 ( @xmath38=156 ) & 334 @xmath37 52 ( @xmath38=228 ) + 60 - 69 years & 33 & 1008 @xmath37 28 ( @xmath38=163 ) & 808 @xmath37 65 ( @xmath38=375 ) & 641 @xmath37 20 ( @xmath38=112 ) & 365 @xmath37 75 ( @xmath38=429 ) + @xmath45 70 years & 5 & 883 @xmath37 52 ( @xmath38=117 ) & 666 @xmath37 48 ( @xmath38=108 ) & 538 @xmath37 30 ( @xmath38=67.2 ) & 272 @xmath37 88 ( @xmath38=197 ) + @xmath162 & & 3.65 & 0.427 & 2.13 & 0.849 + @xmath33 & & 0.00177 & 0.86 & 0.0506 & 0.533 + @xmath35 & & 0.0883 & 0.0112 & 0.0536 & 0.022 + @xmath34 & & 1.52 & 0.802 & 1.09 & 0.569 + @xmath14 & @xmath17 & @xmath36 + 20 - 29 years & 95 & 1209 @xmath37 16 ( @xmath38=152 ) & 808 @xmath37 18 ( @xmath38=173 ) & 688 @xmath37 16 ( @xmath38=156 ) & 309 @xmath37 22 ( @xmath38=219 ) + 30 - 39 years & 90 & 1269 @xmath37 21 ( @xmath38=203 ) & 838 @xmath37 24 ( @xmath38=225 ) & 729 @xmath37 16 ( @xmath38=155 ) & 300 @xmath37 26 ( @xmath38=246 ) + 40 - 49 years & 45 & 1265 @xmath37 18 ( @xmath38=120 ) & 820 @xmath37 20 ( @xmath38=137 ) & 720 @xmath37 20 ( @xmath38=131 ) & 298 @xmath37 26 ( @xmath38=176 ) + 50 - 59 years & 30 & 1241 @xmath37 33 ( @xmath38=182 ) & 862 @xmath37 44 ( @xmath38=238 ) & 635 @xmath37 27 ( @xmath38=148 ) & 436 @xmath37 68 ( @xmath38=371 ) + @xmath163 & & 2.21 & 0.709 & 3.36 & 2.61 + @xmath33 & & 0.0873 & 0.548 & 0.0194 & 0.0517 + @xmath35 & & 0.0253 & 0.00823 & 0.0379 & 0.0297 + @xmath34 & & 0.339 & 0.282 & 0.613 & 0.51 + @xmath17 & @xmath36 + @xmath49 140 cm & 0.0333 & 0.137 & 0.0326 & 0.0184 + 140 - 150 cm & 0.0129 & 0.65 & 0.858 & 0.615 + 150 - 160 cm & 0.00177 & 0.86 & 0.0506 & 0.533 + 160 - 170 cm & 0.000643 & 0.00561 & 0.807 & 0.00456 + 170 - 180 cm & 0.0873 & 0.548 & 0.0194 & 0.0517 + @xmath50 180 cm & 0.98 & 0.292 & 0.56 & 0.0386 + @xmath17 & @xmath36 + @xmath49 140 cm & 0.829 & 0.566 & 0.83 & 0.919 + 140 - 150 cm & 7.42 & 1.27 & 0.946 & 1.69 + 150 - 160 cm & 1.52 & 0.802 & 1.09 & 0.569 + 160 - 170 cm & 1.83 & 1.09 & 0.926 & 0.818 + 170 - 180 cm & 0.339 & 0.282 & 0.613 & 0.51 + @xmath50 180 cm & 0.181 & 1.2 & 1.53 & 1.26 + tables [ table5ca ] and [ table5cb ] show the dependence of observables on minimum height in the @xmath47 and @xmath48 ped / m@xmath164 ranges , respectively .
the trends discussed in the main text are still present .
we notice again a tendency of short people ( most probably children ) to walk faster at higher density . @xmath33 and @xmath34 values for minimum height at different densities
are shown in tables [ table5cex1 ] and [ table5cex2 ] @xmath14 & @xmath17 & @xmath36 + @xmath49 140 cm & 21 & 1138 @xmath37 63 ( @xmath38=288 ) & 1034 @xmath37 80 ( @xmath38=368 ) & 693 @xmath37 44 ( @xmath38=201 ) & 616 @xmath37 92 ( @xmath38=421 ) + 140 - 150 cm & 29 & 1067 @xmath37 57 ( @xmath38=304 ) & 876 @xmath37 51 ( @xmath38=275 ) & 671 @xmath37 40 ( @xmath38=218 ) & 420 @xmath37 64 ( @xmath38=346 ) + 150 - 160 cm & 148 & 1104 @xmath37 18 ( @xmath38=224 ) & 837 @xmath37 25 ( @xmath38=304 ) & 648 @xmath37 11 ( @xmath38=128 ) & 395 @xmath37 32 ( @xmath38=390 ) + 160 - 170 cm & 290 & 1162 @xmath37 11 ( @xmath38=187 ) & 880 @xmath37 19 ( @xmath38=318 ) & 688 @xmath37 13 ( @xmath38=217 ) & 409 @xmath37 22 ( @xmath38=379 ) + 170 - 180 cm & 141 & 1259 @xmath37 16 ( @xmath38=188 ) & 878 @xmath37 21 ( @xmath38=253 ) & 718 @xmath37 17 ( @xmath38=198 ) & 364 @xmath37 28 ( @xmath38=331 ) + @xmath50 180 cm & 7 & 1242 @xmath37 69 ( @xmath38=182 ) & 929 @xmath37 73 ( @xmath38=194 ) & 793 @xmath37 45 ( @xmath38=120 ) & 316 @xmath37 100 ( @xmath38=270 ) + @xmath165 & & 9.97 & 1.72 & 2.32 & 1.8 + @xmath33 & & @xmath40 & 0.128 & 0.0422 & 0.11 + @xmath35 & & 0.0733 & 0.0135 & 0.0181 &
0.0141 + @xmath34 & & 0.906 & 0.634 & 1.13 & 0.767 + @xmath14 & @xmath17 & @xmath36 + @xmath49 140 cm & 8 & 1185 @xmath37 57 ( @xmath38=162 ) & 872 @xmath37 150 ( @xmath38=416 ) & 512 @xmath37 68 ( @xmath38=193 ) & 555 @xmath37 190 ( @xmath38=543 ) + 140 - 150 cm & 6 & 1166 @xmath37 130 ( @xmath38=315 ) & 965 @xmath37 170 ( @xmath38=409 ) & 604 @xmath37 73 ( @xmath38=179 ) & 590 @xmath37 190 ( @xmath38=457 ) + 150 - 160 cm & 39 & 1068 @xmath37 23 ( @xmath38=146 ) & 754 @xmath37 28 ( @xmath38=177 ) & 518 @xmath37 27 ( @xmath38=169 ) & 408 @xmath37 52 ( @xmath38=327 ) + 160 - 170 cm & 72 & 1093 @xmath37 20 ( @xmath38=170 ) & 722 @xmath37 16 ( @xmath38=136 ) & 586 @xmath37 14 ( @xmath38=121 ) & 321 @xmath37 24 ( @xmath38=203 ) + 170 - 180 cm & 42 & 1127 @xmath37 24 ( @xmath38=158 ) & 792 @xmath37 26 ( @xmath38=170 ) & 618 @xmath37 26 ( @xmath38=171 ) & 352 @xmath37 44 ( @xmath38=284 ) + @xmath166 & & 1.34 & 3.29 & 2.62 & 2.31 + @xmath33 & & 0.259 & 0.0127 & 0.0368 & 0.0597 + @xmath35 & & 0.0319 & 0.0751 & 0.0608 & 0.054 + @xmath34 & & 0.787 & 1.44 & 0.606 & 1.18 + & @xmath36 + 0 - 0.05 ped / m@xmath1 & @xmath40 & 0.128 & 0.0422 & 0.11 + 0.05 - 0.1 ped / m@xmath1 & @xmath40 & 0.000607 & 0.000112 & 4.09@xmath127 + 0.1 - 0.15 ped / m@xmath1 & 1.84@xmath52 & @xmath40 & 3.34@xmath52 & @xmath40 + 0.15 - 0.2 ped / m@xmath1 & 0.259 & 0.0127 & 0.0368 & 0.0597 + 0.2 - 0.25 ped / m@xmath1 & 0.303 & 0.602 & 0.603 & 0.765 + & @xmath36 + 0 - 0.05 ped / m@xmath1 & 0.906 & 0.634 & 1.13 & 0.767 + 0.05 - 0.1
ped / m@xmath1 & 1.17 & 0.664 & 0.63 & 0.973 + 0.1 - 0.15 ped / m@xmath1 & 0.856 & 1.32 & 1.12 & 1.29 + 0.15 - 0.2 ped / m@xmath1 & 0.787 & 1.44 & 0.606 & 1.18 + 0.2 - 0.25 ped / m@xmath1 & 0.886 & 0.578 & 0.54 & 0.422 + tables [ table5cc ] , [ table5cd ] , [ table5ce ] and [ table5cf ] show the dependence of observables on minimum height for colleagues , couples , families and friends , respectively .
the dependence of observables on height appears to be attenuated when analysed for groups with a fixed relation ( and in particular for couples ) , as shown by the higher @xmath33 values , and , to a lesser extent , lower @xmath34 values .
@xmath14 & @xmath17 & @xmath36 + 150 - 160 cm & 15 & 1135 @xmath37 36 ( @xmath38=141 ) & 732 @xmath37 25 ( @xmath38=98.2 ) & 652 @xmath37 22 ( @xmath38=85.9 ) & 265 @xmath37 31 ( @xmath38=120 ) + 160 - 170 cm & 159 & 1276 @xmath37 12 ( @xmath38=157 ) & 874 @xmath37 21 ( @xmath38=265 ) & 712 @xmath37 13 ( @xmath38=168 ) & 369 @xmath37 28 ( @xmath38=351 ) + 170 - 180 cm & 170 & 1288 @xmath37 12 ( @xmath38=155 ) & 846 @xmath37 15 ( @xmath38=202 ) & 731 @xmath37 12 ( @xmath38=153 ) & 312 @xmath37 18 ( @xmath38=236 ) + @xmath50 180 cm & 13 & 1220 @xmath37 36 ( @xmath38=128 ) & 789 @xmath37 59 ( @xmath38=214 ) & 689 @xmath37 33 ( @xmath38=120 ) & 263 @xmath37 67 ( @xmath38=241 ) + @xmath167 & & 5.03 & 2.2 & 1.49 & 1.63 + @xmath33 & & 0.002 & 0.0881 & 0.217 &
0.182 + @xmath35 & & 0.041 & 0.0183 & 0.0125 & 0.0137 + @xmath34 & & 0.996 & 0.556 & 0.53 & 0.307 + @xmath14 & @xmath17 & @xmath36 + 150 - 160 cm & 20 & 1060 @xmath37 43 ( @xmath38=193 ) & 736 @xmath37 32 ( @xmath38=143 ) & 631 @xmath37 25 ( @xmath38=111 ) & 288 @xmath37 39 ( @xmath38=175 ) + 160 - 170 cm & 60 & 1114 @xmath37 21 ( @xmath38=160 ) & 716 @xmath37 33 ( @xmath38=254 ) & 591 @xmath37 22 ( @xmath38=171 ) & 305 @xmath37 34 ( @xmath38=261 ) + 170 - 180 cm & 15 & 1092 @xmath37 42 ( @xmath38=162 ) & 678 @xmath37 35 ( @xmath38=137 ) & 592 @xmath37 24 ( @xmath38=93.1 ) & 245 @xmath37 39 ( @xmath38=151 ) + @xmath168 & & 0.773 & 0.296 & 0.528 & 0.408 + @xmath33 & & 0.464 & 0.745 & 0.591 & 0.666 + @xmath35 & & 0.0165 & 0.00639 & 0.0114 & 0.00878 + @xmath34 & & 0.321 & 0.414 & 0.249 & 0.249 + @xmath14 & @xmath17 & @xmath36 + @xmath49 140 cm & 33 & 1117 @xmath37 38 ( @xmath38=216 ) & 1062 @xmath37 72 ( @xmath38=411 ) & 570 @xmath37 39 ( @xmath38=222 ) & 746 @xmath37 89 ( @xmath38=509 ) + 140 - 150 cm & 19 & 1122 @xmath37 62 ( @xmath38=270 ) & 832 @xmath37 60 ( @xmath38=262 ) & 636 @xmath37 32 ( @xmath38=139 ) & 410 @xmath37 65 ( @xmath38=285 ) + 150 - 160 cm & 77 & 1107 @xmath37 23 ( @xmath38=204 ) & 835 @xmath37 34 ( @xmath38=302 ) & 583 @xmath37 19 ( @xmath38=163 ) & 466 @xmath37 44 ( @xmath38=387 ) + 160 - 170 cm & 99 & 1080 @xmath37 17 ( @xmath38=170 ) & 831 @xmath37 24 ( @xmath38=241 ) & 580 @xmath37 17 ( @xmath38=166 ) & 467 @xmath37 33 ( @xmath38=332 ) + 170 - 180 cm & 17 & 1053 @xmath37 36 ( @xmath38=149 ) & 833 @xmath37 65 ( @xmath38=268 ) & 566 @xmath37 37 ( @xmath38=153 ) & 458 @xmath37 98 ( @xmath38=403 ) + @xmath169 & & 0.602 & 4.3 & 0.542 & 4.04 + @xmath33 & & 0.661 & 0.00224 & 0.705 & 0.00344 + @xmath35 & & 0.00993 & 0.0668 & 0.00896 & 0.0631 + @xmath34 & & 0.312 & 0.788 & 0.478 & 0.76 + @xmath14 & @xmath17 & @xmath36 + @xmath49 140 cm & 4 & 1129 @xmath37 55 ( @xmath38=109 ) & 611 @xmath37 20 ( @xmath38=39.7 ) & 518 @xmath37 31 ( @xmath38=61.8 ) & 246 @xmath37 68 ( @xmath38=136 ) + 140 - 150 cm & 16 & 1115 @xmath37 88 ( @xmath38=354 ) & 816 @xmath37 44 ( @xmath38=175 ) & 628 @xmath37 43 ( @xmath38=172 ) & 382 @xmath37 80 ( @xmath38=319 ) + 150 - 160 cm & 101 & 1100 @xmath37 20 ( @xmath38=202 ) & 770 @xmath37 21 ( @xmath38=209 ) & 659 @xmath37 10 ( @xmath38=104 ) & 287 @xmath37 27 ( @xmath38=269 ) + 160 - 170 cm & 142 & 1138 @xmath37 15 ( @xmath38=174 ) & 802 @xmath37 18 ( @xmath38=211 ) & 665 @xmath37 12 ( @xmath38=146 ) & 324 @xmath37 23 ( @xmath38=276 ) + 170 - 180 cm & 53 & 1199 @xmath37 23 ( @xmath38=168 ) & 806 @xmath37 19 ( @xmath38=138 ) & 673 @xmath37 18 ( @xmath38=132 ) & 323 @xmath37 32 ( @xmath38=233 ) + @xmath50 180 cm & 2 & 1606 @xmath37 23 ( @xmath38=32.1 ) & 928 @xmath37 61 ( @xmath38=86.6 ) & 805 @xmath37 120 ( @xmath38=171 ) & 329 @xmath37 72 ( @xmath38=102 ) + @xmath170 & & 4.13 & 1.27 & 1.68 & 0.515 + @xmath33 & & 0.0012 & 0.276 & 0.138 & 0.765 + @xmath35 & & 0.0621 & 0.02 & 0.0263 & 0.00819 + @xmath34 & & 2.52 & 5.74 & 2.83 & 0.461 + tables [ table5cg ] , [ table5ch ] and [ table5ci ] show the dependence of observables on minimum height for two females , mixed and two males dyads , respectively .
as discussed in the main text and shown in figure [ fastchild2 ] in the main text , there is a loss of linearity in @xmath17 , but the patterns described in the main text are still present , although partially attenuated , when gender is kept fixed .
@xmath14 & @xmath17 & @xmath36 + @xmath49 140 cm & 7 & 956 @xmath37 74 ( @xmath38=195 ) & 1186 @xmath37 150 ( @xmath38=385 ) & 529 @xmath37 68 ( @xmath38=180 ) & 935 @xmath37 190 ( @xmath38=509 ) + 140 - 150 cm & 21 & 1022 @xmath37 57 ( @xmath38=262 ) & 841 @xmath37 66 ( @xmath38=300 ) & 591 @xmath37 36 ( @xmath38=166 ) & 439 @xmath37 96 ( @xmath38=442 ) + 150 - 160 cm & 114 & 1098 @xmath37 18 ( @xmath38=197 ) & 780 @xmath37 20 ( @xmath38=214 ) & 656 @xmath37 11 ( @xmath38=112 ) & 306 @xmath37 26 ( @xmath38=279 ) + 160 - 170 cm & 104 & 1131 @xmath37 16 ( @xmath38=159 ) & 768 @xmath37 18 ( @xmath38=185 ) & 658 @xmath37 11 ( @xmath38=115 ) & 278 @xmath37 24 ( @xmath38=245 ) + 170 - 180 cm & 6 & 1123 @xmath37 77 ( @xmath38=188 ) & 706 @xmath37 34 ( @xmath38=83.3 ) & 644 @xmath37 26 ( @xmath38=64.8 ) & 213 @xmath37 59 ( @xmath38=144 ) + @xmath171 & & 2.58 & 6.59 & 3.11 & 9.37 + @xmath33 & & 0.0381 & 4.66@xmath52 & 0.0159 & 4.55@xmath53 + @xmath35 & & 0.0401 & 0.0965 & 0.048 & 0.132 + @xmath34 & & 1.08 & 1.66 & 1.08 & 1.86 + @xmath14 & @xmath17 & @xmath36 + @xmath49 140 cm & 14 & 1100 @xmath37 42 ( @xmath38=159 ) & 947 @xmath37 82 ( @xmath38=307 ) & 590 @xmath37 53 ( @xmath38=199 ) & 593 @xmath37 100 ( @xmath38=373 ) + 140 - 150 cm & 9 & 1107 @xmath37 66 ( @xmath38=199 ) & 967 @xmath37 91 ( @xmath38=273 ) & 664 @xmath37 42 ( @xmath38=127 ) & 552 @xmath37 120 ( @xmath38=346 ) + 150 - 160 cm & 99 & 1092 @xmath37 20 ( @xmath38=195 ) & 829 @xmath37 29 ( @xmath38=286 ) & 609 @xmath37 16 ( @xmath38=160 ) & 429 @xmath37 38 ( @xmath38=376 ) + 160 - 170 cm & 210 & 1128 @xmath37 12 ( @xmath38=176 ) & 811 @xmath37 18 ( @xmath38=262 ) & 618 @xmath37 13 ( @xmath38=184 ) & 395 @xmath37 23 ( @xmath38=328 ) + 170 - 180 cm & 37 & 1083 @xmath37 28 ( @xmath38=172 ) & 806 @xmath37 44 ( @xmath38=267 ) & 589 @xmath37 24 ( @xmath38=143 ) & 404 @xmath37 61 ( @xmath38=372 ) + @xmath50 180 cm & 2 & 937 @xmath37 140 ( @xmath38=204 ) & 687 @xmath37 3.2 ( @xmath38=4.56 ) & 573 @xmath37 48 ( @xmath38=67.8 ) & 253 @xmath37 55 ( @xmath38=77.2 ) + @xmath172 & & 1.14 & 1.3 & 0.418 & 1.27 + @xmath33 & & 0.34 & 0.262 & 0.836 & 0.277 + @xmath35 & & 0.0153 & 0.0175 & 0.00569 & 0.0171 + @xmath34 & & 1.08 & 1.09 & 0.751 & 0.945 + @xmath14 & @xmath17 & @xmath36 + @xmath49 140 cm & 18 & 1221 @xmath37 48 ( @xmath38=203 ) & 977 @xmath37 110 ( @xmath38=455 ) & 577 @xmath37 53 ( @xmath38=226 ) & 631 @xmath37 130 ( @xmath38=549 ) + 140 - 150 cm & 9 & 1301 @xmath37 140 ( @xmath38=405 ) & 861 @xmath37 85 ( @xmath38=255 ) & 640 @xmath37 47 ( @xmath38=142 ) & 456 @xmath37 110 ( @xmath38=343 ) + 150 - 160 cm & 21 & 1196 @xmath37 39 ( @xmath38=180 ) & 739 @xmath37 36 ( @xmath38=164 ) & 600 @xmath37 22 ( @xmath38=103 ) & 320 @xmath37 57 ( @xmath38=261 ) + 160 - 170 cm & 184 & 1238 @xmath37 13 ( @xmath38=179 ) & 863 @xmath37 18 ( @xmath38=241 ) & 700 @xmath37 13 ( @xmath38=174 ) & 372 @xmath37 23 ( @xmath38=315 ) + 170 - 180 cm & 219 & 1272 @xmath37 11 ( @xmath38=156 ) & 834 @xmath37 12 ( @xmath38=184 ) & 719 @xmath37 10 ( @xmath38=151 ) & 310 @xmath37 15 ( @xmath38=224 ) + @xmath50 180 cm & 15 & 1271 @xmath37 46 ( @xmath38=178 ) & 808 @xmath37 53 ( @xmath38=207 ) & 705 @xmath37 35 ( @xmath38=134 ) & 272 @xmath37 59 ( @xmath38=229 ) + @xmath173 & & 1.46 & 2.56 & 4.51 & 5.03 + @xmath33 & & 0.202 & 0.0267 & 0.000499 & 0.00017 + @xmath35 & & 0.0156 & 0.0271 & 0.0468 & 0.0518 + @xmath34 & & 0.394 & 0.721 & 0.903 & 0.826 + tables
[ table5 cm ] and [ table5cn ] show the dependence on minimum height of all observables for dyads with minimum age in the 20 - 29 and 50 - 59 year ranges , respectively , showing the effect of removing children from the population .
finally , tables [ table5cex3 ] and [ table5cex4 ] show the dependence of , respectively , minimum height @xmath33 and @xmath34 values on minimum age ranges .
@xmath14 & @xmath17 & @xmath36 + 140 - 150 cm & 2 & 1161 @xmath37 180 ( @xmath38=254 ) & 748 @xmath37 62 ( @xmath38=87.2 ) & 613 @xmath37 41 ( @xmath38=58.5 ) & 398 @xmath37 54 ( @xmath38=75.7 ) + 150 - 160 cm & 75 & 1157 @xmath37 21 ( @xmath38=184 ) & 800 @xmath37 27 ( @xmath38=232 ) & 668 @xmath37 14 ( @xmath38=122 ) & 311 @xmath37 35 ( @xmath38=307 ) + 160 - 170 cm & 188 & 1175 @xmath37 13 ( @xmath38=176 ) & 781 @xmath37 17 ( @xmath38=231 ) & 658 @xmath37 12 ( @xmath38=165 ) & 300 @xmath37 19 ( @xmath38=265 ) + 170 - 180 cm & 95 & 1209 @xmath37 16 ( @xmath38=152 ) & 808 @xmath37 18 ( @xmath38=173 ) & 688 @xmath37 16 ( @xmath38=156 ) & 309 @xmath37 22 ( @xmath38=219 ) + @xmath50 180 cm & 3 & 1262 @xmath37 130 ( @xmath38=223 ) & 974 @xmath37 170 ( @xmath38=300 ) & 625 @xmath37 9.5 ( @xmath38=16.4 ) & 556 @xmath37 220 ( @xmath38=373 ) + @xmath174 & & 1.19 & 0.818 & 0.684 & 0.756 + @xmath33 & & 0.315 & 0.514 & 0.603 & 0.554 + @xmath35 & & 0.0131 & 0.00906 & 0.00759 & 0.00838 + @xmath34 & & 0.567 & 0.906 & 0.482 & 0.962 + @xmath14 & @xmath17 & @xmath36 + 140 - 150 cm & 3 & 1043 @xmath37 41 ( @xmath38=70.2 ) & 784 @xmath37 56 ( @xmath38=97 ) & 746 @xmath37 63 ( @xmath38=109 ) & 190 @xmath37 16 ( @xmath38=28 ) + 150 - 160 cm & 19 & 1067 @xmath37 39 ( @xmath38=171 ) & 757 @xmath37 34 ( @xmath38=146 ) & 611 @xmath37 36 ( @xmath38=156 ) & 334 @xmath37 52 ( @xmath38=228 ) + 160 - 170 cm & 59 & 1162 @xmath37 25 ( @xmath38=191 ) & 830 @xmath37 29 ( @xmath38=226 ) & 664 @xmath37 22 ( @xmath38=165 ) & 371 @xmath37 41 ( @xmath38=315 ) + 170 - 180 cm & 30 & 1241 @xmath37 33 ( @xmath38=182 ) & 862 @xmath37 44 ( @xmath38=238 ) & 635 @xmath37 27 ( @xmath38=148 ) & 436 @xmath37 68 ( @xmath38=371 ) + @xmath175 & & 3.84 & 0.928 & 0.974 & 0.806 + @xmath33 & & 0.0118 & 0.43 & 0.408 & 0.493 + @xmath35 & & 0.0972 & 0.0254 & 0.0266 & 0.0221 + @xmath34 & & 1.12 & 0.502 & 0.886 & 0.687 + @xmath17 & @xmath36 + 0 - 9 years & 0.000332 & 0.143 & 0.662 & 0.127 + 10 - 19 years & 0.311 & 0.822 & 0.478 & 0.926 + 20 - 29 years & 0.315 & 0.514 & 0.603 & 0.554 + 30 - 39 years & 5@xmath53 & 0.595 & 0.00388 & 0.0423 + 40 - 49 years & 0.000142 & 0.489 & 0.00545 & 0.0649 + 50 - 59 years & 0.0118 & 0.43 & 0.408 & 0.493 + 60 - 69 years & 0.091 & 0.23 & 0.24 & 0.182 + @xmath45 70 years & 0.627 & 0.0424 & 0.0506 & 0.107 + @xmath17 & @xmath36 + 0 - 9 years & 3.42 & 1.11 & 2 & 1.04 + 10 - 19 years & 0.883 & 0.346 & 0.519 & 0.275 + 20 - 29 years & 0.567 & 0.906 & 0.482 & 0.962 + 30 - 39 years & 2.32 & 0.943 & 0.702 & 1.28 + 40 - 49 years & 2.48 & 0.671 & 0.993 & 0.944 + 50 - 59 years & 1.12 & 0.502 & 0.886 & 0.687 + 60 - 69 years & 0.89 & 0.436 & 0.645 & 0.597 + @xmath45 70 years & 1.08 & 2.05 & 2.01 & 1.7 +
we consider a few possible statistical indicators of agreement between coders .
cohen s @xmath176 @xcite is a very popular indicator to compare the agreement between two coders , based on the equation @xmath177 where @xmath33 stands for the agreement rate between coders and @xmath178 for the probability of random agreement .
the agreement between pairs of coders according to this statistics is shown in table [ tablecohen ] .
max age + @xmath179 & 0.815 & 0.961 & 0.636 & 0.476 & 0.582 & 0.555 + @xmath180 & 0.923 & 0.978 & 0.728 & 0.808 & 0.839 & 0.866 + @xmath181 & 0.810 & 0.944 & 0.647 & 0.449 & 0.508 & 0.526 +
these results show that in general the agreement is higher for gender , followed by purpose and relation .
the agreement between coders 1 and 3 is similar also concerning age , while the agreement with coder 2 is quite poor in these categories .
although there is no real sound mathematical way to evaluate the absolute value of these numbers , according to popular benchmarks , an agreement between 0.8 and 1 is considered as `` almost perfect '' , an agreement between 0.6 and 0.8 as `` substantial '' , while an agreement between 0.4 and 0.6 is only `` moderate '' @xcite it generalises eq .
[ kappa ] to deal with multiple coders and categories @xcite .
the corresponding values are shown in table [ tablefleiss ] age + 0.849 & 0.961 & 0.669 & 0.289 & 0.332 & 0.300 + we see that , in relative terms , agreement is higher for gender , followed by purpose and relation , and lowest for age . in absolute terms , according to the benchmarks , we have almost perfect agreement in gender and purpose , substantial in relation and `` fair '' ( i.e. , worst than `` moderate '' ) for age indicators , due to the effect of the different coding by coder 2 . anyway ,
if we try to plot the age difference between coders , as in figure [ fzey ] , we see that although disagreement with coder 2 is substantial , it is almost completely limited to a tendency of coder 2 to put pedestrians in a slightly younger category , i.e. the difference in age between the codings is limited .
nevertheless the fleiss indicator does not take in account the magnitude of difference , and is thus not completely adequate to deal with ordered data .
the krippendorff @xmath55 statistics @xcite , that allows for consideration of quantitative differences between coding results , gives the results shown in table [ tablekrip ] . age + 0.849 & 0.961 & 0.669 & 0.709 &
0.730 & 0.729 + krippendorff does not provide any `` magic number '' but suggests to use data with at least @xmath182 ( satisfied by all our categories ) and require @xmath183 , satisfied by purpose and gender , for reliable results ( @xmath55 between 0.667 and 0.8 could be used for `` tentative conclusions '' ) . using popular indicators of coder reliability ,
we have found that , in relative terms , the most reliable coding regards gender , followed by purpose . in absolute terms , according to the krippendorff @xmath55 statistics that can better cope with the nature of our data , we may see that the purpose and gender codings may be considered as enough reliable to provide sound findings , while the relation and age codings are reliable enough for reporting tentative findings .
the analysis based on these indicators provides an estimate on the reliability of coding of pedestrians in different categories .
we may nevertheless use another approach to test the reliability of our findings when based on different coding processes .
since for each category we analyse the values of the observables @xmath8 , @xmath14 , @xmath17 and @xmath36 , we may compare these quantitative results between different coders .
this comparison , which has also the advantage of being based on more mathematically sound statistical indicators ( standard errors , anova analysis ) is performed in section [ codercomp ] , and shows again that for purpose and gender we have an almost perfect quantitative agreement , while for relation and age , although the agreement is less good , the major patterns of behaviour are qualitatively observed regardless of coders .
the results ( on the common subset of data ) for the purpose dependence of all observables between the main coder ( coder 1 ) and the secondary coders are compared in tables [ table1f1 ] , [ table1f2 ] and [ table1f3 ] .
@xmath14 & @xmath17 & @xmath36 + leisure & 136 & 1085 @xmath37 19 ( @xmath38=220 ) & 796 @xmath37 21 ( @xmath38=248 ) & 636 @xmath37 13 ( @xmath38=151 ) & 351 @xmath37 28 ( @xmath38=327 ) + work & 132 & 1257 @xmath37 14 ( @xmath38=157 ) & 829 @xmath37 17 ( @xmath38=196 ) & 723 @xmath37 12 ( @xmath38=143 ) & 303 @xmath37 21 ( @xmath38=241 ) + @xmath184 & & 53.1 & 1.41 & 23.5 & 1.88 + @xmath33 & & @xmath40 & 0.236 & 2.14@xmath127 & 0.171 + @xmath35 & & 0.166 & 0.00529 & 0.0811 & 0.00703 + @xmath34 & & 0.893 & 0.146 & 0.594 & 0.168 + @xmath14 & @xmath17 & @xmath36 + leisure & 151 & 1093 @xmath37 17 ( @xmath38=212 ) & 793 @xmath37 20 ( @xmath38=243 ) & 641 @xmath37 12 ( @xmath38=147 ) & 344 @xmath37 26 ( @xmath38=318 ) + work & 117 & 1269 @xmath37 15 ( @xmath38=159 ) & 837 @xmath37 18 ( @xmath38=196 ) & 728 @xmath37 13 ( @xmath38=146 ) & 306 @xmath37 22 ( @xmath38=243 ) + @xmath184 & & 56.2 & 2.56 & 23.4 & 1.13 + @xmath33 & & @xmath40 & 0.111 & 2.18@xmath127 & 0.289 + @xmath35 & & 0.175 & 0.00954 & 0.081 &
0.00422 + @xmath34 & & 0.927 & 0.198 & 0.599 & 0.131 + @xmath14 & @xmath17 & @xmath36 + leisure & 133 & 1077 @xmath37 19 ( @xmath38=217 ) & 789 @xmath37 22 ( @xmath38=250 ) & 626 @xmath37 13 ( @xmath38=145 ) & 354 @xmath37 29 ( @xmath38=330 ) + work & 133 & 1262 @xmath37 14 ( @xmath38=156 ) & 836 @xmath37 17 ( @xmath38=195 ) & 732 @xmath37 12 ( @xmath38=144 ) & 302 @xmath37 21 ( @xmath38=239 ) + @xmath185 & & 63.6 & 2.93 & 35.6 & 2.13 + @xmath33 & & @xmath40 & 0.0881 & @xmath40 & 0.145 + @xmath35 & & 0.194 & 0.011 & 0.119 & 0.00802 + @xmath34 & & 0.982 & 0.211 & 0.734 & 0.18 + the differences between coders are thus always of one standard error or smaller , and the extremely significant statistical differences in the @xmath17 and @xmath8 distribution ( along with the less significant @xmath36 and @xmath14 ones ) are reported by all coders .
the results ( on the common subset of data ) for the relation dependence of all observables between the main coder ( coder 1 ) and the secondary coders are compared in tables [ table2f1 ] , [ table2f2 ] and [ table2f3 ] .
while all the major trends exposed in the main text are confirmed , quantitative results between coders may sometimes be different ( we refer in particular to the @xmath36 distribution for couples , extremely narrow according to coder 3 ) .
@xmath14 & @xmath17 & @xmath36 + colleagues & 125 & 1256 @xmath37 14 ( @xmath38=154 ) & 829 @xmath37 18 ( @xmath38=196 ) & 725 @xmath37 13 ( @xmath38=142 ) & 301 @xmath37 21 ( @xmath38=239 ) + couples & 28 & 1087 @xmath37 37 ( @xmath38=194 ) & 690 @xmath37 33 ( @xmath38=174 ) & 611 @xmath37 21 ( @xmath38=112 ) & 248 @xmath37 37 ( @xmath38=198 ) + families & 40 & 1051 @xmath37 24 ( @xmath38=153 ) & 864 @xmath37 54 ( @xmath38=341 ) & 594 @xmath37 21 ( @xmath38=134 ) & 492 @xmath37 69 ( @xmath38=438 ) + friends & 56 & 1121 @xmath37 36 ( @xmath38=271 ) & 777 @xmath37 24 ( @xmath38=182 ) & 669 @xmath37 19 ( @xmath38=145 ) & 286 @xmath37 32 ( @xmath38=243 ) + @xmath186 & & 16.4 & 4.19 & 11.8 & 6.12 + @xmath33 & & @xmath40 & 0.00651 & 3.06@xmath53 & 0.0005 + @xmath35 & & 0.167 & 0.0488 & 0.126 & 0.0697 + @xmath34 & & 1.33 & 0.612 & 0.934 & 0.678 + @xmath14 & @xmath17 & @xmath36 + colleagues & 116 & 1267 @xmath37 14 ( @xmath38=156 ) & 839 @xmath37 18 ( @xmath38=197 ) & 729 @xmath37 14 ( @xmath38=147 ) & 308 @xmath37 23 ( @xmath38=244 ) + couples & 44 & 1082 @xmath37 28 ( @xmath38=184 ) & 703 @xmath37 21 ( @xmath38=140 ) & 582 @xmath37 19 ( @xmath38=125 ) & 296 @xmath37 33 ( @xmath38=221 ) + families & 42 & 1054 @xmath37 25 ( @xmath38=164 ) & 894 @xmath37 53 ( @xmath38=341 ) & 651 @xmath37 25 ( @xmath38=163 ) & 451 @xmath37 70 ( @xmath38=457 ) + friends & 66 & 1131 @xmath37 31 ( @xmath38=254 ) & 786 @xmath37 23 ( @xmath38=188 ) & 673 @xmath37 17 ( @xmath38=136 ) & 304 @xmath37 29 ( @xmath38=238 ) + @xmath187 & & 19 & 6.55 & 11.9 & 3.13 + @xmath33 & & @xmath40 & 0.000276 & 2.54@xmath53 & 0.0262 + @xmath35 & & 0.178 & 0.0692 & 0.119 & 0.0344 + @xmath34 & & 1.35 & 0.74 & 1.04 & 0.437 + @xmath14 & @xmath17 & @xmath36 + colleagues & 136 & 1259 @xmath37 14 ( @xmath38=158 ) & 834 @xmath37 17 ( @xmath38=194 ) & 727 @xmath37 13 ( @xmath38=147 ) & 304 @xmath37 21 ( @xmath38=242 ) + couples & 23 & 1070 @xmath37 42 ( @xmath38=204 ) & 624 @xmath37 20 ( @xmath38=96.4 ) & 578 @xmath37 20 ( @xmath38=95.1 ) & 182 @xmath37 17 ( @xmath38=81.2 ) + families & 50 & 1053 @xmath37 24 ( @xmath38=172 ) & 867 @xmath37 44 ( @xmath38=312 ) & 612 @xmath37 20 ( @xmath38=140 ) & 478 @xmath37 59 ( @xmath38=416 ) + friends & 54 & 1084 @xmath37 32 ( @xmath38=235 ) & 780 @xmath37 27 ( @xmath38=196 ) & 663 @xmath37 22 ( @xmath38=159 ) & 298 @xmath37 33 ( @xmath38=245 ) + @xmath188 & & 23.4 & 7.57 & 12.4 & 7.52 + @xmath33 & & @xmath40 & 7.11@xmath52 & 1.36@xmath53 & 7.61@xmath52 + @xmath35 & & 0.213 & 0.0807 & 0.125 & 0.0801 + @xmath34 & & 1.27 & 0.915 & 1.06 & 0.849 + the results ( on the common subset of data ) for the gender dependence of all observables between the main coder ( coder 1 ) and the secondary coders are compared in tables [ table3f1 ] , [ table3f2 ] and [ table3f3 ] , showing that there is basically no difference in the coding of gender .
@xmath14 & @xmath17 & @xmath36 + two females & 55 & 1076 @xmath37 32 ( @xmath38=240 ) & 745 @xmath37 21 ( @xmath38=155 ) & 629 @xmath37 15 ( @xmath38=112 ) & 290 @xmath37 33 ( @xmath38=242 ) + mixed & 86 & 1095 @xmath37 19 ( @xmath38=173 ) & 820 @xmath37 31 ( @xmath38=287 ) & 641 @xmath37 17 ( @xmath38=159 ) & 384 @xmath37 39 ( @xmath38=360 ) + two males & 127 & 1261 @xmath37 16 ( @xmath38=178 ) & 836 @xmath37 17 ( @xmath38=195 ) & 727 @xmath37 13 ( @xmath38=150 ) & 305 @xmath37 22 ( @xmath38=243 ) + @xmath189 & & 27.2 & 3.25 & 12.8 & 2.48 + @xmath33 & & @xmath40 & 0.0404 & 5.09@xmath127 & 0.0855 + @xmath35 & & 0.171 & 0.0239 & 0.0879 & 0.0184 + @xmath34 & & 0.93 & 0.494 & 0.699 & 0.292 + @xmath14 & @xmath17 & @xmath36 + two females & 53 & 1078 @xmath37 33 ( @xmath38=241 ) & 747 @xmath37 22 ( @xmath38=158 ) & 637 @xmath37 15 ( @xmath38=106 ) & 286 @xmath37 32 ( @xmath38=233 ) + mixed & 89 & 1093 @xmath37 18 ( @xmath38=173 ) & 814 @xmath37 30 ( @xmath38=283 ) & 635 @xmath37 17 ( @xmath38=159 ) & 382 @xmath37 38 ( @xmath38=360 ) + two males & 126 & 1263 @xmath37 16 ( @xmath38=177 ) & 838 @xmath37 17 ( @xmath38=194 ) & 728 @xmath37 13 ( @xmath38=150 ) & 306 @xmath37 22 ( @xmath38=244 ) + @xmath189 & & 28.2 & 3.12 & 13.3 & 2.48 + @xmath33 & & @xmath40 & 0.0459 & 3.22@xmath127 & 0.0853 + @xmath35 & & 0.176 & 0.023 & 0.091 & 0.0184
+ @xmath34 & & 0.935 & 0.494 & 0.604 & 0.3 + @xmath14 & @xmath17 & @xmath36 + two females & 55 & 1074 @xmath37 32 ( @xmath38=239 ) & 742 @xmath37 21 ( @xmath38=153 ) & 636 @xmath37 14 ( @xmath38=103 ) & 281 @xmath37 31 ( @xmath38=230 ) + mixed & 89 & 1093 @xmath37 19 ( @xmath38=175 ) & 824 @xmath37 31 ( @xmath38=288 ) & 634 @xmath37 17 ( @xmath38=161 ) & 397 @xmath37 39 ( @xmath38=368 ) + two males & 124 & 1267 @xmath37 16 ( @xmath38=173 ) & 834 @xmath37 17 ( @xmath38=190 ) & 730 @xmath37 13 ( @xmath38=150 ) & 298 @xmath37 21 ( @xmath38=232 ) + @xmath189 & & 30.4 & 3.44 & 14.2 & 3.99 + @xmath33 & & @xmath40 & 0.0336 & 1.36@xmath127 & 0.0196 + @xmath35 & & 0.187 & 0.0253 & 0.0969 & 0.0293 + @xmath34 & & 0.987 & 0.511 & 0.622 & 0.359 + the results ( on the common subset of data ) for the minimum age dependence of all observables between the main coder ( coder 1 ) and the secondary coders are compared in tables [ table4f1 ] , [ table4f2 ] and [ table4f3 ] . sadly , almost no groups with children are present in the common set .
the drop in velocity with age is , on the other hand , confirmed in a statistically significant way by all coders .
@xmath14 & @xmath17 & @xmath36 + 10 - 19 years & 16 & 1157 @xmath37 86 ( @xmath38=343 ) & 715 @xmath37 31 ( @xmath38=123 ) & 653 @xmath37 23 ( @xmath38=92.3 ) & 223 @xmath37 38 ( @xmath38=151 ) + 20 - 29 years & 58 & 1183 @xmath37 28 ( @xmath38=215 ) & 765 @xmath37 28 ( @xmath38=211 ) & 666 @xmath37 20 ( @xmath38=149 ) & 268 @xmath37 33 ( @xmath38=252 ) + 30 - 39 years & 96 & 1186 @xmath37 21 ( @xmath38=203 ) & 817 @xmath37 21 ( @xmath38=211 ) & 689 @xmath37 17 ( @xmath38=166 ) & 327 @xmath37 27 ( @xmath38=262 ) + 40 - 49 years & 41 & 1193 @xmath37 25 ( @xmath38=161 ) & 811 @xmath37 27 ( @xmath38=173 ) & 684 @xmath37 22 ( @xmath38=143 ) & 327 @xmath37 38 ( @xmath38=245 ) + 50 - 59 years & 31 & 1210 @xmath37 29 ( @xmath38=160 ) & 880 @xmath37 46 ( @xmath38=254 ) & 696 @xmath37 29 ( @xmath38=160 ) & 407 @xmath37 65 ( @xmath38=360 ) + 60 - 69 years & 21 & 1017 @xmath37 35 ( @xmath38=160 ) & 869 @xmath37 66 ( @xmath38=304 ) & 671 @xmath37 34 ( @xmath38=156 ) & 401 @xmath37 85 ( @xmath38=388 ) + @xmath45 70 years & 5 & 949 @xmath37 15 ( @xmath38=34.2 ) & 913 @xmath37 170 ( @xmath38=379 ) & 608 @xmath37 28 ( @xmath38=61.7 ) & 551 @xmath37 210 ( @xmath38=470 ) + @xmath190 & & 3.35 & 1.81 & 0.462 & 1.91 + @xmath33 & & 0.00337 & 0.0974 & 0.836 & 0.0789 + @xmath35 & & 0.0715 & 0.0399 & 0.0105 & 0.0421 + @xmath34 & & 1.73 & 0.964 & 0.578 & 1.29 + @xmath14 & @xmath17 & @xmath36 + 0 - 9 years & 2 & 1190 @xmath37 220 ( @xmath38=312 ) & 749 @xmath37 110 ( @xmath38=152 ) & 700 @xmath37 87 ( @xmath38=123 ) & 202 @xmath37 55 ( @xmath38=77.8 ) + 10 - 19 years & 16 & 1169 @xmath37 84 ( @xmath38=334 ) & 682 @xmath37 20 ( @xmath38=80.1 ) & 646 @xmath37 20 ( @xmath38=78.4 ) & 172 @xmath37 18 ( @xmath38=73.3 ) + 20 - 29 years & 107 & 1163 @xmath37 19 ( @xmath38=196 ) & 765 @xmath37 16 ( @xmath38=165 ) & 655 @xmath37 13 ( @xmath38=138 ) & 288 @xmath37 22 ( @xmath38=231 ) + 30 - 39 years & 78 & 1217 @xmath37 24 ( @xmath38=209 ) & 869 @xmath37 28 ( @xmath38=244 ) & 727 @xmath37 21 ( @xmath38=185 ) & 362 @xmath37 33 ( @xmath38=290 ) + 40 - 49 years & 32 & 1181 @xmath37 31 ( @xmath38=176 ) & 855 @xmath37 44 ( @xmath38=249 ) & 645 @xmath37 22 ( @xmath38=124 ) & 418 @xmath37 66 ( @xmath38=373 ) + 50 - 59 years & 24 & 1074 @xmath37 32 ( @xmath38=158 ) & 853 @xmath37 60 ( @xmath38=293 ) & 706 @xmath37 29 ( @xmath38=143 ) & 343 @xmath37 74 ( @xmath38=363 ) + 60 - 69 years & 9 & 1047 @xmath37 42 ( @xmath38=127 ) & 861 @xmath37 100 ( @xmath38=311 ) & 655 @xmath37 45 ( @xmath38=135 ) & 432 @xmath37 120 ( @xmath38=375 ) + @xmath190 & & 2.1 & 3.09 & 2.3 & 2.13 + @xmath33 & & 0.0533 & 0.0061 & 0.0349 & 0.0505 + @xmath35 & & 0.0461 & 0.0663 & 0.0503 & 0.0467 + @xmath34 & & 0.842 & 0.833 & 0.482 & 1.14 + @xmath14 & @xmath17 & @xmath36 + 10 - 19 years & 14 & 1163 @xmath37 98 ( @xmath38=367 ) & 701 @xmath37 31 ( @xmath38=117 ) & 623 @xmath37 35 ( @xmath38=130 ) & 218 @xmath37 48 ( @xmath38=181 ) + 20 - 29 years & 64 & 1157 @xmath37 29 ( @xmath38=236 ) & 758 @xmath37 24 ( @xmath38=194 ) & 658 @xmath37 20 ( @xmath38=158 ) & 274 @xmath37 27 ( @xmath38=220 ) + 30 - 39 years & 50 & 1197 @xmath37 27 ( @xmath38=193 ) & 830 @xmath37 32 ( @xmath38=227 ) & 685 @xmath37 23 ( @xmath38=162 ) & 349 @xmath37 43 ( @xmath38=302 ) + 40 - 49 years & 77 & 1205 @xmath37 19 ( @xmath38=163 ) & 832 @xmath37 23 ( @xmath38=205 ) & 684 @xmath37 16 ( @xmath38=141 ) & 351 @xmath37 33 ( @xmath38=293 ) + 50 - 59 years & 36 & 1205 @xmath37 25 ( @xmath38=152 ) & 820 @xmath37 35 ( @xmath38=207 ) & 722 @xmath37 28 ( @xmath38=168 ) & 300 @xmath37 39 ( @xmath38=233 ) + 60 - 69 years & 20 & 1025
@xmath37 40 ( @xmath38=179 ) & 903 @xmath37 74 ( @xmath38=332 ) & 699 @xmath37 29 ( @xmath38=129 ) & 418 @xmath37 97 ( @xmath38=436 ) + @xmath45 70 years & 7 & 956 @xmath37 26 ( @xmath38=69.1 ) & 881 @xmath37 120 ( @xmath38=326 ) & 605 @xmath37 36 ( @xmath38=96.4 ) & 503 @xmath37 140 ( @xmath38=382 ) + @xmath190 & & 3.69 & 2.04 & 1.33 & 1.67 + @xmath33 & & 0.00153 & 0.0604 & 0.245 & 0.129 + @xmath35 & & 0.0783 & 0.0449 & 0.0296 & 0.0369 + @xmath34 & & 1.74 & 0.755 & 0.732 & 1.09 +
table [ table4a_av ] shows the average age dependence of all observables , and the age dependence of variables @xmath14 , @xmath17 , @xmath36 is also graphically shown in figure [ f4a_av ] , while that of @xmath8 is shown in figure [ f4b_av ] ( left panels ) .
@xmath14 & @xmath17 & @xmath36 + 10 - 19 years & 60 & 1147 @xmath37 34 ( @xmath38=264 ) & 865 @xmath37 43 ( @xmath38=332 ) & 575 @xmath37 20 ( @xmath38=158 ) & 496 @xmath37 57 ( @xmath38=445 ) + 20 - 29 years & 370 & 1181 @xmath37 9.2 ( @xmath38=178 ) & 793 @xmath37 12 ( @xmath38=226 ) & 662 @xmath37 8.1 ( @xmath38=155 ) & 313 @xmath37 14 ( @xmath38=274 ) + 30 - 39 years & 269 & 1213 @xmath37 12 ( @xmath38=199 ) & 831 @xmath37 14 ( @xmath38=234 ) & 670 @xmath37 11 ( @xmath38=174 ) & 360 @xmath37 18 ( @xmath38=302 ) + 40 - 49 years & 195 & 1172 @xmath37 13 ( @xmath38=183 ) & 852 @xmath37 17 ( @xmath38=232 ) & 674 @xmath37 12 ( @xmath38=167 ) & 387 @xmath37 23 ( @xmath38=316 ) + 50 - 59 years & 114 & 1157 @xmath37 18 ( @xmath38=194 ) & 825 @xmath37 20 ( @xmath38=217 ) & 650 @xmath37 15 ( @xmath38=159 ) & 376 @xmath37 30 ( @xmath38=317 ) + 60 - 69 years & 69 & 1032 @xmath37 20 ( @xmath38=168 ) & 875 @xmath37 40 ( @xmath38=332 ) & 635 @xmath37 20 ( @xmath38=163 ) & 467 @xmath37 50 ( @xmath38=416 ) + @xmath45 70 years & 12 & 886 @xmath37 29 ( @xmath38=99.8 ) & 786 @xmath37 79 ( @xmath38=275 ) & 588 @xmath37 19 ( @xmath38=66.6 ) & 385 @xmath37 100 ( @xmath38=363 ) + @xmath191 & & 13.2 & 2.26 & 3.79 & 4.75 + @xmath33 & & @xmath40 & 0.036 & 0.000955 & 8.72@xmath52 + @xmath35 & & 0.0681 & 0.0124 & 0.0206 &
0.0257 + @xmath34 & & 1.67 & 0.275 & 0.598 & 0.603 + ( black and circles ) , @xmath17 ( red and squares ) and @xmath36 ( blue and triangles ) dependence on average ( left ) and maximum ( right ) age .
dashed lines provide standard error confidence bars .
the point at 75 years corresponds to the `` 70 years or more '' slot.,title="fig : " ] ( black and circles ) , @xmath17 ( red and squares ) and @xmath36 ( blue and triangles ) dependence on average ( left ) and maximum ( right ) age . dashed lines provide standard error confidence bars .
the point at 75 years corresponds to the `` 70 years or more '' slot.,title="fig : " ] dependence on average ( left ) and maximum ( right ) age .
dashed lines provide standard error confidence bars .
the point at 75 years corresponds to the `` 70 years or more '' slot.,title="fig : " ] dependence on average ( left ) and maximum ( right ) age .
dashed lines provide standard error confidence bars .
the point at 75 years corresponds to the `` 70 years or more '' slot.,title="fig : " ] table [ table4a_max ] shows the average age dependence of all observables , and the age dependence of variables @xmath14 , @xmath17 , @xmath36 is also graphically shown in figure [ f4a_av ] , while that of @xmath8 is shown in figure [ f4b_av ] ( right panels ) .
@xmath14 & @xmath17 & @xmath36 + 10 - 19 years & 28 & 1141 @xmath37 55 ( @xmath38=292 ) & 733 @xmath37 35 ( @xmath38=186 ) & 628 @xmath37 17 ( @xmath38=92.5 ) & 283 @xmath37 49 ( @xmath38=258 ) + 20 - 29 years & 327 & 1177 @xmath37 9.6 ( @xmath38=174 ) & 789 @xmath37 12 ( @xmath38=225 ) & 660 @xmath37 8.2 ( @xmath38=149 ) & 309 @xmath37 15 ( @xmath38=278 ) + 30 - 39 years & 292 & 1203 @xmath37 12 ( @xmath38=204 ) & 831 @xmath37 14 ( @xmath38=238 ) & 648 @xmath37 10 ( @xmath38=172 ) & 384 @xmath37 19 ( @xmath38=321 ) + 40 - 49 years & 143 & 1201 @xmath37 15 ( @xmath38=181 ) & 849 @xmath37 23 ( @xmath38=275 ) & 680 @xmath37 15 ( @xmath38=176 ) & 377 @xmath37 28 ( @xmath38=336 ) + 50 - 59 years & 179 & 1178 @xmath37 14 ( @xmath38=193 ) & 854 @xmath37 16 ( @xmath38=217 ) & 674 @xmath37 13 ( @xmath38=178 ) & 388 @xmath37 23 ( @xmath38=306 ) + 60 - 69 years & 105 & 1043 @xmath37 17 ( @xmath38=174 ) & 872 @xmath37 30 ( @xmath38=310 ) & 638 @xmath37 16 ( @xmath38=162 ) & 452 @xmath37 40 ( @xmath38=407 ) + @xmath45 70 years & 15 & 927 @xmath37 33 ( @xmath38=128 ) & 760 @xmath37 65 ( @xmath38=254 ) & 575 @xmath37 18 ( @xmath38=67.9 ) & 374 @xmath37 85 ( @xmath38=330 ) + @xmath191 & & 14.2 & 3.37 & 1.97 & 3.7 + @xmath33 & & @xmath40 & 0.0027 & 0.0668 & 0.00122 + @xmath35 & & 0.0731 & 0.0183 & 0.0108 &
0.0201 + @xmath34 & & 1.38 & 0.484 & 0.619 & 0.443 + it may be seen that the results concerning minimum ( section [ ageef ] ) , maximum and average age are quite similar above 20 years .
nevertheless , using minimum age allows us to spot the presence of children below 10 years of age and verify their peculiar behaviour .
tables [ table_hc_av ] and [ table_hc_max ] show the dependence on , respectively , average and maximum height of all observables .
figure figures [ comphfigv ] and [ comphfigvbis ] provide , on the other hand , a graphical comparison for the @xmath8 and @xmath17 observables .
we show these two figures since these observables are mostly growing with height , and so their analysis is easier .
as the figures show , the `` average '' curves are somehow in between the other two curves , with the `` minimum '' curve on the top and the maximum on the bottom , as expected for an observable that grows with height .
this suggests that dyads have , at least regarding height , a behaviour that is an average of the individual ones , and all three indicators should be basically equivalent . in the main text
we choose to use the `` minimum '' height indicator for two reasons : it allows to better identify dyads with children , and it has a sufficient number of events in all occupied height slots .
@xmath14 & @xmath17 & @xmath36 + @xmath49 140 cm & 14 & 1044 @xmath37 52 ( @xmath38=195 ) & 983 @xmath37 98 ( @xmath38=365 ) & 527 @xmath37 39 ( @xmath38=145 ) & 672 @xmath37 130 ( @xmath38=492 ) + 140 - 150 cm & 22 & 1011 @xmath37 36 ( @xmath38=168 ) & 910 @xmath37 81 ( @xmath38=382 ) & 562 @xmath37 39 ( @xmath38=183 ) & 570 @xmath37 100 ( @xmath38=476 ) + 150 - 160 cm & 118 & 1110 @xmath37 23 ( @xmath38=253 ) & 812 @xmath37 24 ( @xmath38=260 ) & 629 @xmath37 14 ( @xmath38=149 ) & 379 @xmath37 31 ( @xmath38=340 ) + 160 - 170 cm & 472 & 1140 @xmath37 8.3 ( @xmath38=181 ) & 821 @xmath37 12 ( @xmath38=250 ) & 646 @xmath37 7.6 ( @xmath38=165 ) & 372 @xmath37 15 ( @xmath38=329 ) + 170 - 180 cm & 421 & 1222 @xmath37 8.7 ( @xmath38=179 ) & 828 @xmath37 11 ( @xmath38=224 ) & 685 @xmath37 8 ( @xmath38=164 ) & 342 @xmath37 14 ( @xmath38=282 ) + @xmath50 180 cm & 42 & 1275 @xmath37 27 ( @xmath38=174 ) & 793 @xmath37 26 ( @xmath38=171 ) & 693 @xmath37 19 ( @xmath38=122 ) & 274 @xmath37 34 ( @xmath38=220 ) + @xmath51 & & 17.9 & 1.95 & 7.31 & 5.74 + @xmath33 & & @xmath40 & 0.0829 & 9.42@xmath53 & 3.1@xmath52 + @xmath35 & & 0.0764 & 0.00894 & 0.0327 & 0.0258 + @xmath34 & & 1.54 & 0.816 & 1.3 & 1.29 + @xmath14 & @xmath17 & @xmath36 + @xmath49 140 cm & 3 & 1172 @xmath37 12 ( @xmath38=21.4 ) & 650 @xmath37 49 ( @xmath38=84.4 ) & 597 @xmath37 23 ( @xmath38=39.1 ) & 189 @xmath37 60 ( @xmath38=104 ) + 140 - 150 cm & 3 & 1051 @xmath37 79 ( @xmath38=136 ) & 611 @xmath37 11 ( @xmath38=19.6 ) & 518 @xmath37 39 ( @xmath38=67.5 ) & 242 @xmath37 33 ( @xmath38=57.2 ) + 150 - 160 cm & 49 & 988 @xmath37 25 ( @xmath38=178 ) & 782 @xmath37 34 ( @xmath38=240 ) & 594 @xmath37 20 ( @xmath38=138 ) & 366 @xmath37 49 ( @xmath38=346 ) + 160 - 170 cm & 336 & 1129 @xmath37 10 ( @xmath38=191 ) & 820 @xmath37 14 ( @xmath38=256 ) & 637 @xmath37 8.3 ( @xmath38=151 ) & 378 @xmath37 19 ( @xmath38=343 ) + 170 - 180 cm & 556 & 1191 @xmath37 8.1 ( @xmath38=191 ) & 830 @xmath37 10 ( @xmath38=237 ) & 671 @xmath37 7.1 ( @xmath38=167 ) & 359 @xmath37 13 ( @xmath38=306 ) + @xmath50 180 cm & 142 & 1250 @xmath37 15 ( @xmath38=183 ) & 843 @xmath37 21 ( @xmath38=252 ) & 677 @xmath37 15 ( @xmath38=182 ) & 365 @xmath37 26 ( @xmath38=313 ) + @xmath51 & & 18.8 & 1.31 & 4.21 & 0.427 + @xmath33 & & @xmath40 & 0.257 & 0.000849 & 0.83 + @xmath35 & & 0.08 & 0.00601 & 0.0191 & 0.00197 + @xmath34 & & 1.44 & 0.929 & 0.881 & 0.555 + dependence on average ( black and circles ) , minimum ( red and squares ) and maximum ( blue and triangles ) height .
dashed lines provide standard error confidence bars .
the points at 135 and 185 cm correspond to the `` less than 140 '' and `` more than 180 '' cm slots . ]
dependence on average ( black and circles ) , minimum ( red and squares ) and maximum ( blue and triangles ) height .
dashed lines provide standard error confidence bars .
the points at 135 and 185 cm correspond to the `` less than 140 '' and `` more than 180 '' cm slots . ]
99 f. zanlungo , t. ikeda , t. kanda , _
potential for the dynamics of pedestrians in a socially interacting group _ , phys .
e , 89 , 1 , 021811 , ( 2014 ) f. zanlungo , d. bri and t. kanda , _ spatial - size scaling of pedestrian groups under growing density conditions _ physical review e 91 ( 6 ) , 062810 ( 2015 ) bri , draen and zanlungo , francesco and kanda , takayuki , _ density and velocity patterns during one year of pedestrian tracking _ , transportation research procedia , 2 , 7786 , 2014 m. schultz , l. rger , f. hartmut and b. schlag , _ group dynamic behavior and psychometric profiles as substantial driver for pedestrian dynamics _
, in _ pedestrian and evacuation dynamics 2012 _ , u. weidmann , u. kirsh and m. schreckenberg , vol ii , pp . 1097 - 1111 ( 2014 ) m. moussad , n. perozo , s. garnier , d. helbing and g. theraulaz , _ the walking behaviour of pedestrian social groups and its impact on crowd dynamics _ , plos one , 5 , 4 , e10047 , ( 2010 ) k. kumagai , i. ehiro , k. ono , _ numerical simulation model of group walking for tsunami evacuees _ , proceedings of the 2016 pedestrian and evacuation dynamics conference t. kanda , m. shiomi , z. miyashita , h. ishiguro , n. hagita , _ an affective guide robot in a shopping mall _ , acm / ieee international conference on human - robot interaction ( hri ) , 173 - 180 ( 2009 ) m. shiomi , f. zanlungo , k. hayashi , t. kanda , _ towards a socially acceptable collision avoidance for a mobile robot navigating among pedestrians using a pedestrian model _ , international journal of social robotics , 6 , 3 , 443455 , 2014 r. murakami , s. morales , s , satake , t. kanda , _ destination unknown : walking side - by - side without knowing the goal _ , proceedings of the 2014 acm / ieee international conference on human - robot interaction , 471478 , 2014 m. costa , _ interpersonal distances in group walking _ , journal of nonverbal behavior , 34 , 1 , 15 - 26 ( 2010 ) f. zanlungo and t. kanda , _ do walking pedestrians stabily interact inside a large group ?
analysis of group and sub - group spatial structure _ , cogsci13 , ( 2013 ) a. gorrini , g. vizzari , s. bandini , _ granulometric distribution and crowds of groups : focusing on dyads _ , 11th conference international traffic and granular flow , delft ( nl ) , 2015 s. bandini , l. crociani , a. gorrini , g. vizzari , _ an agent - based model of pedestrian dynamics considering groups : a real world case study _ , intelligent transportation systems ( itsc ) , 2014 ieee 17th international conference on , 572577 g. kster , m. seitz , f. treml , d. hartmann and w. klein , _ on modeling the influence of group formation in a crowd _ ,
contemporary social science , 6 , 3 , 397 - 414 , ( 2011 ) g. kster , f. treml , m. seitz , and w. klein , _ validation of crowd models including social groups_. in ulrich weidmann , uwe kirsch , and michael schreckenberg , editors , pedestrian and evacuation dynamics 2012 , pages 1051 - 1063 .
springer international publishing , 2014 x. wei , x. lv , w. song , x. li , _ survey study and experimental investigation on the local behavior of pedestrian groups _ , complexity , volume 20 , issue 6 july / august 2015 pages 8797 i. karamouzas , and m. overmars , _ simulating the local behaviour of small pedestrian groups _ , proceedings of the 17th acm symposium on virtual reality software and technology , 183 - 190 , ( 2010 ) y. zhang , j. pettr , x. qin , s. donikian and q. peng , _ a local behavior model for small pedestrian groups _ , computer - aided design and computer graphics ( cad / graphics ) , 2011 12th international conference on , 275 - 281 ( 2011 ) n. bode , s. holl , w. mehner , a. seyfried , _ disentangling the impact of social groups on response times and movement dynamics in evacuations _ , plosone http://dx.doi.org/10.1371/journal.pone.0121227 a. gorrini , l. crociani , c. feliciani , p. zhao , k. nishinari , s. baldini , _ social groups and pedestrian crowds : experiment on dyads in a counter flow scenario _ , proceedings of the 2016 pedestrian and evacuation dynamics conference p. zhao , l. sun , l. cui , w. luo , y. ding , _ the walking behaviours of pedestrian social group in the corridor of subway station _ , proceedings of the 2016 pedestrian and evacuation dynamics conference c. von krchten , frank mller , anton svachiy , oliver wohak , andreas schadschneider , _ empirical study of the influence of social groups in evacuation scenarios _ traffic and granular flow 15 ( springer , 2016 ) w. wang , s. lo , s. liu , j. ma , _ a simulation of pedestrian side - by - side walking behaviour and its impact on crowd dynamics _ , proceedings of the 2016 pedestrian and evacuation dynamics conference j. huang , x. zou , x. qu , j. ma , r. xu , _ a structure analysis method for complex social pedestrian groups with symbol expression and relationship matrix _ , proceedings of the 2016 pedestrian and evacuation dynamics conference f. zanlungo and t. kanda , _ a mesoscopic model for the effect of density on pedestrian group dynamics _ europhysics letters , 111 , 38007 ( 2015 ) g. turchetti , f. zanlungo , b. giorgini , _ dynamics and thermodynamics of a gas of automata _ epl ( europhysics letters ) 78 ( 5 ) , 58003 f. zanlungo , d. bri , t. kanda , _ pedestrian group behaviour analysis under different density conditions _ , transportation research procedia , 2 , 149158 , 2014 r. bohannon , _ comfortable and maximum walking speed of adults aged 20 - 79 years : reference values and determinants _ , age and ageing , 26 , 15 - 19 , 1997 n. bode , _ the effect of social groups and gender on pedestrian behaviour immediately in front of bottlenecks _ , proceedings of the 2016 pedestrian and evacuation dynamics conference l. henderson , d. lyons , it sexual differences in human crowd motion , nature 240 , 353 - 355 ( 08 december 1972 ) ; i. von sivers , f. knzner , g. kster , _ pedestrian evacuation simulation with separated families _ , proceedings of the 2016 pedestrian and evacuation dynamics conference y. feng , d. li , ` _ an empirical study and conceptual model on heterogenity of pedestrian social groups for friend -group and family - group _ , proceedings of the 2016 pedestrian and evacuation dynamics conference f. mller , a. schadschneider , _ evacuation dynamics of asymmetrically coupled pedestrian pairs _ , traffic and granular flow ' 15 ( springer , 2016 ) f. zanlungo , z. ycel , t. kanda , _ the effect of social roles on group behaviour _ , proceedings of the 2016 pedestrian and evacuation dynamics conference d. bri , t. kanda , t. ikeda , t. miyashita _ person tracking in large public spaces using 3-d range sensors _ , ieee transactions on human machine systems , vol 43 , no 6 , pp 522 - 534 , ( 2013 ) http://www.irc.atr.jp/sets/groups/ s. seer , n. brndle , c. ratti , _ kinects and human kinetics : a new approach for studying crowd behavior _ ,
transportation research part c : emerging technologies , 48(0):212 - 228 , 2014 a. corbetta , l. bruno , a. muntean , f. toschi , _ high statistics measurements of pedestrian dynamics _ , transportation research procedia , 2 , 96 - 104 m. knapp , _ nonverbal communication in human interaction _ , cengage learning ( 2012 ) c. kleinke , _ gaze and eye contact : a research review .
_ , psychological bulletin , 100 , 1 , 78 , 1986 m. argyle , j. dean , _ sociometry _ , 289 - 304 , 1965 j. zhang , w. klingsch , a. schadschneider , a. seyfried , _ transitions in pedestrian fundamental diagrams of straight corridors and t - junctions _ , journal of statistical mechanics : theory and experiment , 2011 , 06 , p06004 , 2011 a. corbetta , m. jasper , c. lee , f. toschi , _ continuous measurements of real - life bidirectional pedestrian flows on a wide walkway _ , proceedings of the 2016 pedestrian and evacuation dynamics conference http://www.mext.go.jp/b_menu/toukei/001/022/2004/002.pdf ( in japanese ) z. ycel , f. zanlungo , t. ikeda , t. miyashita , and n. hagita _ deciphering the crowd : modeling and identification of pedestrian group motion _
, sensors , vol . 13 , pp .
875 - 897 , 2013 d. bri , f. zanlungo , t. kanda , _ modelling of pedestrian groups and application to group recognition _ , submitted to mipro 2016 , opatja , croatia r. ash , _ statistical inference , a concise course _ , dover 2011 w. press , s. teukolsky , w. vetterling , b. flannery , _ numerical recipes in c _ , second edition , cambridge university press , 1992 .
we use the gamma function routine of page 214 , the incomplete beta function routine of page 228 , and the f test routine of page 619 , adapted by us to a single tail test .
j. cohen , _ statistical power analysis for the behavioural sciences _ ,
second edition , routledge , 1988 cohen , jacob _ a coefficient of agreement for nominal scales_. educational and psychological measurement 20 ( 1 ) : 3746 ( 1960 ) .
landis , j.r . ; koch , g.g .
( 1977 ) . _ the measurement of observer agreement for categorical data _ biometrics . 33 ( 1 ) : 159174 .
fleiss , j. l. ( 1971 ) _ measuring nominal scale agreement among many raters _
, psychological bulletin , vol .
378382 krippendorff , klaus .
_ reliability in content analysis _
, human communication research 30.3 ( 2004 ) : 411 - 433 . | in recent years , researchers in pedestrian behaviour and crowd modelling have become more and more interested in the behaviour of walking social groups , since these groups represent an important portion of pedestrian crowds , and present peculiar dynamical features .
it is anyway clear that , being group dynamics determined by human social behaviour , it probably depends on properties such as the purpose of the pedestrians , their personal relation , their gender , age , and body size .
we may call these the `` intrinsic properties '' of the group ( opposed to extrinsic ones such as crowd density or environmental features ) . in this work we quantitatively analyse the dynamical properties of pedestrian dyads ( distance , spatial formation and velocity ) by analysing a large data set of
automatically tracked pedestrian trajectories in an unconstrained `` ecological '' setting ( a shopping mall ) , whose relational group properties have been analysed by three different human coders . we observed that females walk slower and closer than males , that workers walk faster , at a larger distance and more abreast than leisure oriented people , and that inter group relation has a strong effect on group structure , with couples walking very close and abreast , colleagues walking at a larger distance , and friends walking more abreast than family members .
pedestrian height ( obtained automatically through our tracking system ) influences velocity and abreast distance , both growing functions of the average group height .
results regarding pedestrian age show as expected that elderly people walk slowly , while active age adults walk at the maximum velocity .
groups with children have a strong tendency to walk in a non abreast formation , with a large distance ( despite a low abreast distance ) .
a cross - analysis of the interplay between these intrinsic features , taking in account also the effect of extrinsic crowd density , confirms these major effects but reveals also a richer structure .
an interesting and unexpected result , for example , is that the velocity of groups with children _ increases _ with density , at least in the low - medium density range found under normal conditions in shopping malls .
children also appear to behave differently according to the gender of the parent . |
in the theoretical study of polymer physics @xcite , computer simulations provide a powerful method to mimic the behavior of polymers covering the range from atomic to coarse - grained scales depending on the problems one is interested in @xcite .
the generic scaling properties of single linear and branched polymers in the bulk under various solvent conditions and polymer solutions at different temperatures have been described quite well by simple coarse - grained lattice models , e.g. self - avoiding walks on the simple cubic lattice , and off - lattice models , e.g. bead - spring model , etc .
a wide variety of computational strategies has been employed to simulate and analyze these models , including conventional ( metropolis ) monte carlo ( mc ) schemes with various types of moves , chain growth algorithms , parallel tempering , and molecular dynamics @xcite . on the one hand , however , as the size and complexity of a system increases , detailed information at the atomic scale may be lost when employing low resolution coarse - graining representations . on the other hand
, the cost of computing time may be too high if the system is described at high resolution .
therefore , more scientific effort has been devoted to developing an appropriate coarse - grained model where the connection to the details on the atomic scale is not lost , but rather atomistic details are suitably mapped on effective potentials on the coarser scale .
such models then can reproduce the global thermodynamic properties and the local mechanical and chemical properties such as the intermolecular forces between polymer chains @xcite .
improving such approach further is still an active area of research . in this work
we deal with linear polymer chains in a melt on the simple cubic lattice .
although coarse - grained lattice models neglect the chemical detail of a specific polymer chain and only keep chain connectivity ( topology ) and excluded volume , the universal behavior of polymers still remains the same in the thermodynamic limit ( as the chain length @xmath2)@xcite .
we consider for our simulations the bond fluctuation model ( bfm ) @xcite where linear chains are located on the simple cubic lattice with bond constraints .
the bfm has the advantages that the computational efficiency of lattice models is kept and the behavior of polymers in a continuum space can be described approximately .
the model thus introduces some local conformational flexibility while retaining the computational efficiency of lattice models for implementing excluded volume interactions by enforcing a single occupation of each lattice vertex .
although much work using this model exists already , only the fully flexible limit of the model has been used exclusively .
this limit does not suffice when one considers a possible mapping of atomistic details to this model , which requires to include some description of chain stiffness as well .
a review of recent bfm studies is given in ref .
when the concentration of polymer solutions is above the overlap concentration denoted by @xmath3 , the excluded volume interactions are screened @xcite .
the average interaction between monomers finally should cancel in a polymer melt since every monomer is isotropically surrounded by other monomers belonging to the same chain or not according to flory s argument .
therefore the polymer chains in a melt behave as ideal chains , where the excluded volume effect is no longer important .
however , a careful investigation of an individual polymer chain in a polymer melt based on the bfm and the bead - spring model ( bsm ) shows that there are noticeable deviations from an ideal chain behavior @xcite .
the deviations are due to the incompressibility constraint of the melt
. therefore , small scale - free corrections to the asymptotic behavior for ideal chains exist and they are irrespective of the stiffness of the chains . in refs .
@xcite , the authors have shown that the corrections must decay as the dimensionless ( ginzburg ) parameter @xmath4 for @xmath5 with the increase of the segments @xmath6 and the effective bond length @xmath7 of the chain . here
@xmath8 is the overall monomer density and @xmath9 the segmental overlap density related to the internal distance @xmath10 .
it is still unclear whether the predicted power - law deviations from unperturbed gaussian behavior can be observed for the internal segment lengths of semiflexible chains in the range @xmath11 by varying the stiffness of chains , i.e. tuning the bending energy @xmath12 .
a very important practical problem for the simulation of melts of very long polymer chains is the construction of a suitable initial configuration .
a naive approach for creating initial configurations of melts with long polymers , would be to switch off initially the excluded volume interactions between monomers ( so that the polymers resemble gaussian chains ) , arrange these chains randomly in a simulation box , and switch on gradually the excluded volume interactions at the final step .
it has been demonstrated by simulations using the bsm in the continuum that this strategy is not feasible since it leads to chain deformations on short length scales @xcite . with increasing chain length this effect
becomes more pronounced due to the fact that self - screening and correlation hole effects , inherent to flexible high polymer melts , are not properly accounted for by this procedure .
this problem was overcome by reducing the density fluctuations of gaussian chains via a
pre - packing " strategy , and then switching on the excluded volume interactions in a quasi - static way ( `` slow push - off '' ) @xcite .
we test a similar strategy of simulating polymer melts based on the lattice model bfm in this work to check whether it helps to generate nearly equilibrated initial configurations through such a procedure .
the outline of the paper is as follows : sec .
[ model ] describes the model and the simulation technique discussing how to generate initial configurations of polymer chains in a melt to equilibrate the system .
[ results ] presents the simulation results at different stages from the initial state to the equilibrium state of fully flexible and moderately stiff polymer chains in a melt .
finally our conclusions are summarized in sec .
[ conclusion ] .
in the standard bond fluctuation model ( bfm ) @xcite a flexible polymer chain with excluded volume interactions is described by a self - avoiding walk ( saw ) chain of effective monomers on a simple cubic lattice ( the lattice spacing is the unit of length ) .
each effective monomer of such a saw chain blocks all eight corners of an elementary cube of the simple cubic lattice from further occupation .
two successive monomers along a chain are connected by a bond vector @xmath13 which is taken from the set \{@xmath14,@xmath15 , @xmath16 , @xmath17 , @xmath18 , @xmath19 } including also all permutations .
the bond length @xmath20 is therefore in a range between @xmath21 and @xmath22 .
there are in total @xmath23 bond vectors and @xmath24 different bond angles between two sequential bonds along a chain serving as candidates for building the conformational structure of polymers . as the stiffness of the chains
is considered , semiflexible chains of @xmath25 monomers in a melt thus are described by saw chains on the simple cubic lattice , with a bending potential @xmath26 where @xmath12 is the bending energy ( @xmath27 for ordinary saws ) , and @xmath28 is the bond angle between the @xmath29 bond vector and the @xmath30 bond vector along a chain .
@xmath31 is of order unity throughout the whole paper . a systematical study of single semiflexible chains based on the bfm @xcite shows that due to bond vector fluctuations and lattice artifacts the initial decay of the bond - bond orientational correlation function deviates from the simple exponential decay for @xmath32 .
therefore , one should be careful of using the bfm for studying rather stiff chains as has already been pointed out also in ref .
@xcite .
it is well understood that linear polymers in a melt can be described by saw chains based on the bfm at a volume fraction @xmath33 @xcite . since in the bfm
, each effective monomer occupies one unit cell , containing @xmath34 lattice sites , the monomer density is therefore defined as @xmath35 . a detailed description of how we prepare the initial configurations is given as follows : the initial configuration of polymer melts containing @xmath36 chains of @xmath25 monomers in a box of size @xmath37 with periodic boundary conditions in all three directions are generated in the following way : * the linear dimension of a simple cubic lattice is set up as @xmath38 with volume fraction @xmath33 .
* at the @xmath39 step the first @xmath36 monomers of the chains are randomly put on the lattice sites without double overlapping . it can be easily done by dividing the box into @xmath36 blocks and putting a single monomer in each block at a randomly chosen position inside the block . *
polymer chains are built like non - reversal random walks ( nrrws ) by adding one monomer at each step until all @xmath36 chains reaching the required chain length @xmath40 .
( we define here the chain length as the number of bonds along the contour of a chain .
the contour length of the chain then is @xmath41 where @xmath42 is the root - mean - square bond length ) .
based on the bfm there are @xmath23 possibilities to place the next monomer for each chain at the @xmath43 step , but at the following steps only one of @xmath44 directions can be selected since an immediately reverse step is not allowed . at this stage ,
the excluded volume effect is switched off completely and @xmath36 nrrws of @xmath45-steps in the simulation box are generated .
+ if the stiffness of the chains is considered , the probability of placing the @xmath30 monomer connected to the @xmath29 monomer is proportional to @xmath46 . *
finally the excluded volume interactions between monomers are considered by applying the local 26 and slithering - snake moves to relax the polymer chains and push off those monomers blocking the same lattice site until all chains satisfy self- and mutual - avoidance .
steepest descent " approach only works for lattice chains since excluded volume constraints can be checked very efficiently , but would not work for continuum chains . due to the large local density fluctuations of randomly overlapping nrrw chains in a melt , a pre - packing procedure to reduce the chain deformations
is also tested before the excluded volume interactions are switched on .
this pre - packing procedure has been successfully performed in the simulations of equilibrating very long polymer chains in melts up to @xmath47 and @xmath48 based on the bead - spring model in the continuum @xcite . in this procedure
all nrrw chains still keep their own structures as rigid bodies , but are rearranged by monte carlo ( mc ) trial moves .
the cost function which describes the average fluctuations of the particle number @xmath49 within a sphere of radius @xmath50 with its center located at any monomer @xmath51 is defined by @xmath52 with @xmath53 where @xmath54 is the total number of monomers in a melt , @xmath55 is the number of monomers inside the @xmath29 sphere of radius @xmath50 , @xmath56 is the distance between monomers @xmath51 and @xmath57 irrespective of the chain connectivity , and @xmath58 is the heaviside step function . a trial move is accepted if the cost function becomes smaller , namely , the local density fluctuation is minimized in this process .
two types of mc moves , translation and pivot - like moves , applied to the center of mass of an individual chain are considered as follows : ( a ) in the translation move , an individual chain ( or its center of mass ) is randomly translated by a vector chosen from the set @xmath59 including all permutations .
( b ) in the pivot - like move , a whole individual chain is transformed by randomly choosing one of the @xmath60 symmetric operators including rotations by @xmath61 or @xmath62 around a random axis and reflections through its center of mass , and inversion at its center of mass .
the center of mass serves as a pivot point here .
note that all monomers are only allowed to sit on the lattice sites , so a chain would need to be shifted slightly after the pivot - like move .
however , the deviation of its center of mass is less than @xmath63 lattice spacing
. for equilibrating the system and measuring interesting physical quantities in the equilibrium runs three types of mc moves , local @xmath64 moves @xcite , slithering - snake moves , and pivot moves , are used in our simulations and briefly described as follows , * local @xmath64 ( l26 " ) move : a monomer is chosen randomly to move to one of the @xmath64 nearest and next nearest neighbor sites surrounding it , i.e , randomly translated by a vector chosen from the set @xmath59 including all permutations .
the bond crossing is allowed during the move .
this is different from the traditional local @xmath65 ( l6 " ) move where a monomer of the chain is chosen at random and moved to the nearest neighbor sites in the six lattice directions randomly . *
slithering - snake move : an end monomer is removed randomly and connected to the other end of the same linear chain by a bond vector randomly chosen from the set of @xmath23 allowed bond vectors . *
pivot move : a monomer is chosen randomly from a linear chain as a pivot point , and the short part of the linear chain is transformed by randomly selecting one of the @xmath66 symmetry operations ( no change ; rotations by @xmath61 and @xmath62 ; reflections and inversions ) . of course ,
any attempted monte carlo move is accepted only if it does not violate constraints of our physical systems , such as excluded volume and bond length constraints .
we first wish to test whether applying the pre - packing process to rearrange the nrrw chains based on the bfm before the excluded volume interaction is switched on helps to prepare the nearly equilibrated initial configurations or not .
we restrict the size of the lattice to be @xmath67 in our simulations .
the total number of effective monomers based on the bfm is therefore @xmath68 .
varying @xmath36 and @xmath25 but keeping @xmath69 fixed the conformational properties of polymer melts including fully flexible chains ( @xmath70 ) and moderately stiff chains ( @xmath71 , @xmath72 , and @xmath73 ) are studied .
results of simulating polymer melts from preparing the initial configuration to the analysis of equilibrated chain structures in a melt are discussed in the next subsection iii a. for @xmath74 .
data are for @xmath75 nrrw chains of @xmath76 monomers in a box of size @xmath67 .
several bending energies @xmath12 are chosen , as indicated . ] in fig .
[ fig - melt - u ] we show the typical time series of the local density fluctuation @xmath77 with @xmath74 for @xmath75 nrrw chains of @xmath76 monomers in a box of size @xmath67 , and for the bending energies @xmath27 ( fully flexible ) , @xmath78 , @xmath72 , and @xmath73 . in the pre - packing process , the radius @xmath50 is first set to @xmath79 in the unit of lattice spacings until all curves describing the local density fluctuation reach a plateau , and then reduced to @xmath80 until all curves reach another plateau again .
the estimates of the collective structure factor @xmath81 for @xmath82 and the local density fluctuation @xmath77 for @xmath74 are of the same order of magnitude ( see sec . [ structure ] ) from the final configurations generated by this two - step pre - packing process .
each monte carlo step contains @xmath36 translation moves and @xmath36 pivot - like moves such that the locations of all chains in the box are rearranged by the trial moves .
each curve shown in fig .
[ fig - melt - u ] represents the results averaged over @xmath83 independent realizations .
we see that the local density fluctuation can be reduced by about an order of magnitude after about @xmath84 monte carlo steps .
( a ) ( b ) + ( c ) ( d ) + it is interesting to know what we have gained by introducing the pre - packing procedure before equilibrating the system
. therefore , we check the process of pushing away the monomers occupying the same unit cells ( eight lattice sites are occupied by one effective monomer in bfm ) by observing the time series of the number of overlapping monomers , @xmath85 , as the excluded volume effect is switched on . for simplicity ,
if a lattice site is occupied by @xmath6 effective monomers , it contributes @xmath86 overlapping monomers to @xmath85 .
the transition process from @xmath75 nrrw chains of @xmath76 monomers to saw chains is observed according to the decrease of @xmath85 shown in fig .
[ fig - melt - overlap ] for @xmath70 , @xmath78 , @xmath72 , and @xmath73 .
two sets of data ( @xmath34 samples for each ) representing the push - off process before ( dashed curve in green color online ) and after ( solid curves in red color online ) the pre - packing process is applied .
here one mc step is a sequence of @xmath69 l26 moves , and @xmath87 slithering - snake moves . in the slithering - snake moves , an end - monomer of each chain
is selected and attempted to move once until all chains have been chosen , and then the same procedure is repeated @xmath25 times .
a trial move is accepted if @xmath88 and @xmath89 where @xmath90 is a random number and @xmath91 .
the latter condition is the metropolis criterion preserving the stiffness of the chains .
if only l26 moves are applied in this push - off process , at the beginning it is very efficient to push off overlapping monomers , but then some chains are trapped in a state containing knots , or a state blocked by other chains .
the slithering - snake moves therefore indeed play an important role at the intermediate stages for the chains escaping from a trapped state . from our observations in fig .
[ fig - melt - overlap ] it seems that applying an additional pre - packing process does not speed up the push - off process .
however , it is still necessary to investigate the conformation of chains in a melt carefully .
in addition , it is interesting to note that pre - packing leads to a significantly faster decrease of @xmath85 with time during intermediate stages of the process .
this feature may be of interest in cases where one does not require that strictly @xmath92 in the initial stage of the averaging .
consider e.g. a variant of the model where monomers overlap is not strictly forbidden ( which corresponds to an infinitely high repulsive energy ) but only leads to a large but finite energy penalty , e.g. @xmath93 .
the probability to accept a move which increases the overlap then is of order @xmath94 , i.e. essentially negligible , and for long chains the statistical properties of the model are essentially the same as for the model with strict excluded volume .
for such a model the overlap energy per monomer reduces after about @xmath95 monte carlo steps from an initial value ( of the order of @xmath96 ) by three orders of magnitude , i.e. @xmath97 , and this small number is reached @xmath98 to @xmath99 times faster with pre - packing than without it . for very large systems ( and complicated models , leading to slow simulation algorithms ) it may hence be an acceptable compromise to allow a small fraction ( @xmath100 or less ) of overlap in the system . in table [ table1 ] we compare the average values of local density fluctuation @xmath77 \{eq .
( [ eq - fluc ] ) } for @xmath74 ( since @xmath101 ) over @xmath83 configures at different stages from the preparation of the initial configurations to the final configurations at the equilibrium state .
two ways of preparing the initial configurations ( sec .
[ model ] ) are compared as follows : ( 1 ) randomly generating weakly - interactive nrrws ( nrrws ) , applying the pre - packing process ( pre - packing , nrrws ) , and switching on the excluded volume effect ( pre - packing+ev , saws ) .
( 2 ) randomly generating weakly - interactive nrrws ( nrrws ) and switching on the excluded volume effect ( ev , saws ) .
we see that for weakly - interactive nrrws the local density fluctuation decreases as the bending energy @xmath12 increases .
since the bond angles tend to become smaller , the distance between non - bonded monomers apart from each other along the chain becomes longer . after the pre - packing process
the fluctuation reduces by about @xmath102 .
once the excluded volume interactions between monomers are switched on , the fluctuation remains almost the same for all cases . [ cols="^,>,>,>,>",options="header " , ] ( a ) ( b ) + ( a ) ( b ) + ( c ) ( d ) + the conformational behavior of polymer chains of size @xmath25 in a melt can also be described by the probability distributions of end - to - end distance @xmath105 and gyration radius @xmath104 , @xmath106 and @xmath107 , respectively .
the probability distribution of @xmath105 for ideal chains is simply a gaussian distribution , @xmath108 there exists exact theoretical prediction of the probability distribution of @xmath104 for ideal chains @xcite , but they are complicated to evaluate
. however , it has been checked @xcite that the same formula suggested by lhuillier @xcite for polymer chains under good solvent conditions in @xmath109-dimensions is still a good approximation for ideal chains , i.e. , @xmath110 \label{eq - prg}\ ] ] where @xmath111 and @xmath112 are ( non - universal ) constants , and the exponents @xmath113 and @xmath114 are linked to the space dimension @xmath109 and the flory exponent @xmath115 by @xmath116 and @xmath117 . here @xmath118 is the des cloizeaux exponent @xcite for the osmotic pressure of a semidilute polymer solution , and @xmath114 is the fisher exponent @xcite characterizing the end - to - end distance distribution .
numerically , the probability distribution of @xmath119 is obtained by accumulating the histogram @xmath120 of @xmath119 over all configurations and all chains of size @xmath25 , given by @xmath121 here @xmath119 stands for @xmath103 or @xmath104 .
note that an angular average over all directions has been included in the accumulating process of the histogram due to spherical symmetry .
thus , the normalized histogram of @xmath119 is given by @xmath122 where @xmath123 is the normalization factor such that @xmath124 in order to compare the probability distributions of @xmath103 and @xmath104 between various different chain sizes @xmath25 and bending energies @xmath12 , instead of @xmath103 and @xmath104 we use the reduced end - to - end distance @xmath125 and the reduced gyration radius @xmath126 .
results of the normalized histogram @xmath127 compared to the theoretical prediction @xmath128 using @xmath129 and eq .
( [ eq - pre ] ) are shown in fig .
[ fig - pre ] . here
the normalization factor @xmath130 for all cases .
we see the nice data collapse for chain sizes @xmath131 and all bending energies @xmath12 in fig .
[ fig - pre]a , b and the data are described by a universal scaling function @xmath132 for ideal chains .
for @xmath133 the data deviates from the master curve due to the finite - size effect .
results of the normalized histogram @xmath134 compared to the theoretical prediction @xmath135 using @xmath136 where @xmath137 is a constant and eq .
( [ eq - prg ] ) are shown in fig .
[ fig - prg ] . here
the normalization factor @xmath123 , parameters @xmath138 and @xmath139 are determined numerically depending on the simulation data , and they depend slightly on the chain size @xmath25 . as @xmath140
they will tend to asymptotic values ( not shown ) . near the peak ( @xmath141 )
, we see the finite - size effect is stronger for the distribution of @xmath142 . the fitting curve @xmath143 with parameters @xmath144 , @xmath145 , and @xmath146 determined by the least - square fit for @xmath76 and @xmath70 and @xmath147 , @xmath148 , and @xmath149 for @xmath76 and @xmath150
are shown in fig .
[ fig - prg]a , b , respectively for comparison .
the distribution @xmath151 has its maximum value , i.e. @xmath152 , and the corresponding gyration radius @xmath153 .
therefore , using eq .
( [ eq - prg ] ) , the logarithm of the rescaled probability distribution as a function of @xmath154 is given by @xmath155 \label{eq - a}\end{aligned}\ ] ] with @xmath156 where one fitting parameter @xmath157 is left .
results of @xmath158 plotted versus @xmath159 are presented in fig .
[ fig - prg]c , d .
we see that due to the finite - size effect , data for fully flexible chains in a melt shown in fig .
[ fig - prg]c , d only start to converge for @xmath160 . as @xmath25 increases ,
the systematical errors need to be taken into account . using the least square fit
, it gives @xmath161 .
it is obvious that the distribution can only be well described by eq .
( [ eq - a ] ) for @xmath162 , similar as that was found for gaussian chains in ref . @xcite . for polymer melts of a fixed chain length @xmath76
, we see that for @xmath163 the distribution remains the same for different stiffnesses , while for @xmath162 the data for @xmath27 are slightly deviated from the data for @xmath164 .
we plot the same distribution , eq . ( [ eq - a ] ) with @xmath165 for comparison .
the collective static structure factor @xmath81 of polymer melts is defined by the total scattering from the center of all monomers inside the box , regardless of whether they are linked along a polymer chain or not , @xmath166 \right \rangle \,.\ ] ] here @xmath54 are the total number of monomers and @xmath167 represents the average over all independent configurations , and over all vectors @xmath168 of the same size @xmath169 . note that since a simple cubic lattice of size @xmath170 with periodic boundary condition is considered for our simulations , only the following @xmath168 are allowed , @xmath171 where @xmath172 for @xmath173 , @xmath21 , and @xmath98 , and @xmath174 .
@xmath81 characterizes a competition between the intramolecular fluctuations ( @xmath175 ) and the intermolecular correlations ( @xmath176 ) , thus one can write @xmath177 with @xmath178 \right\rangle = s_c(q)\ ] ] and @xmath179 \right \rangle\ ] ] where the contributions from the intramolecular fluctuations are simply equivalent to the average of the standard static structure factor of single polymer chains in a melt , @xmath1 .
( a ) ( b ) + ( c ) ( d ) + the collective structure factors @xmath81 which represent the conformations of the whole polymer system at different stages from weakly - interactive nrrw chains to equilibrated saw chains in a melt for @xmath76 and for @xmath70 , @xmath78 , @xmath72 , and @xmath73 are shown in fig .
[ fig - sq - melt512i ] .
results include the initial nrrw chains ( nrrws ) , rearranged nrrw chains through the pre - packing process ( pre - packing ) , initial saw chains obtained from the initial nrrw chains by switching on the excluded volume effect before ( ev ) and after ( pre - packing+ev ) applying the pre - packing procedure , and the equilibrated saw chains in a melt ( equilibrium ) .
each curve shows the result averaged over @xmath83 independent configurations at the intermediate states , while @xmath180 independent configurations are considered for the measurement in equilibrium .
the average static structure factor of single nrrw chains generated initially , @xmath1 , is also included for comparison . in fig .
[ fig - sq - melt512i ] we see that in all cases there is no interactions between nrrws generated in the box at the beginning , so the structure factor shows the same scaling behavior for single nrrws in dilute solution , i.e. @xmath181 . the scaling predictions @xmath182 with @xmath183 for ideal chains and @xmath184 for rigid rods are verified for fully flexible and moderately stiff chains .
after the pre - packing process where nrrw chains are rearranged , the collective structure factor decreases in the low - q regime .
once all overlapping monomers are pushed off there is no difference between the configurations generated by the push - off process with and without the pre - packing in the intermediate and large @xmath185 , but for very small @xmath185 a small discrepancy exists between those data sets .
it is consistent with our previous observation of @xmath186 in fig .
[ fig - rs - melt512 ] that the discrepancy becomes more prominent at larger scales . of polymer melts , plotted vs. @xmath185 on semi - log scales data are for polymer melts consisting of @xmath75 chains of lengths @xmath76 , and for various values of @xmath12 , as indicated . in ( a )
the peak appears at @xmath187 measuring the mean inter - particle distance in the polymer melts , and the second peak at @xmath188 is the so - called bragg peak . ]
( a ) ( b ) + in the following we only focus on polymer melts consisting of @xmath75 chains of @xmath76 monomers in equilibrium .
figure [ fig - sq - melt512 ] shows the results for the scattering from polymer melts in equilibrium .
four different choices of the stiffness characterized by the bending energy @xmath12 are included . at small @xmath185
there is a systematic dependence on @xmath12 , reflecting the ( rather weak ) dependence of the compressibility on @xmath12 .
it is remarkable that the structure factor of large values of @xmath185 is completely independent of the chain stiffness .
as @xmath185 increases , the first peak , the so - called amorphous halo for non - crystalline materials , appears at @xmath189 measuring the mean inter - particle distance ( @xmath190 lattice spacings ) in a polymer melt .
the second peak , the so - called bragg peak , appears at @xmath191 probing the structure on the scale of the monomers ( @xmath192 lattice spacing ) . the scattering from single chains in a melt in equilibrium is shown in fig .
[ fig - melt - sqc512 ] . in fig .
[ fig - melt - sqc512 ] one sees that @xmath193 for small @xmath185 in the guinier regime , then a cross - over occurs to the power law of gaussian coils ( ideal chains ) , @xmath194 with @xmath195 . for moderately stiff chains
one observes also a rigid - rod regime @xmath184 .
the two peaks at @xmath187 and @xmath196 show up in a similar way as that for the collective structure @xmath81 in fig .
[ fig - sq - melt512 ] and are due to the underlying lattice model . in order to clarify whether single chains in a melt behave as ideal chains we show the structure factors @xmath1 in a kratky - plot in fig .
[ fig - melt - sqc512](b ) . the debye function @xcite describing the scattering from gaussian chains , @xmath197 is also presented in fig .
[ fig - melt - sqc512]b for comparison . for fully flexible chains ( @xmath70 ) in a melt
we see the discrepancy between our simulation results and the debye function at the intermediate values of @xmath185 , which agrees with the previous finding @xcite that polymer chains in a melt are not random walks .
while for @xmath70 the deviation from ideality is clearly recognized due to the minimum in the kratky - plot near @xmath198 , for @xmath199 there occurs no minimum any longer ( for @xmath76 ) . while the deviations from gaussian statistics are still obvious from @xmath200 , fig .
[ fig - hs - melt ] , they are not easy to extract from @xmath1 . for @xmath71
we see the kratky plateau over all intermediate regimes , while the regime decreases as @xmath12 increases since the rigid - rod behavior takes over .
( a ) ( b ) + the isothermal compressibility of @xmath201 characterizing the degree of density fluctuations on large scale is a measure for the response of a systems volume to an increase in the system pressure @xcite , @xmath202 , and is related to the structure factor @xmath81 as @xmath203 , @xmath204 where @xmath205 is defined as the dimensionless compressibility " . in fig .
[ fig - sq - melt512 ] we see that as @xmath206 the estimates of the collective structure factor for polymer chains of size @xmath76 in a melt reach a plateau value @xmath207 . due to the finite - size effect and lattice artifact ( @xmath208 )
, we take the same data but plot @xmath209 versus @xmath210 in fig .
[ fig - kappa ] .
we see the fluctuations of data near @xmath211 .
however , the best fit of the straight line going through our data gives the estimate of @xmath212^{-1}$ ] .
results are listed in table [ table3 ] .
we see that @xmath213 $ ] and depends only weakly on the stiffness of the chains .
( a ) ( b ) + in the thermodynamic limit by taking the limit @xmath140 ( but keeping the density @xmath214 fixed ) , then taking the limit @xmath206 , the local density fluctuations in subvolumina @xmath215 @xcite @xmath216 where @xmath217 .
choosing @xmath74 such that @xmath218 , i.e. in the limit @xmath219 , the local density fluctuations estimated from the configurations of polymer melts in equilibrium are listed in table [ table1 ] .
the estimates of @xmath220 $ ] are a little bit higher than the estimates from the collective structure factor , but they are compatible .
we also check the finite - size effect on the estimate of the local density fluctuations .
[ fig - cg ] shows the local density fluctuations @xmath221/\langle n(\sigma ) \rangle$ ] as a function of @xmath50 . the arrow and dashed lines shown in fig .
[ fig - cg ] indicate the estimate of the compressibility @xmath222 at @xmath223 .
we see that the local density fluctuations at certain values of @xmath50 ( @xmath224 ) do not depend on the chain lengths @xmath25 for chains of different stiffnesses , while the discrepancy between the data sets decreases as the stiffness of chains increases ( see fig .
[ fig - cg]a ) . for @xmath225 the local density fluctuations tend to diverge . as the size of polymer melts increases ( @xmath226 increases , but @xmath227 is fixed ) the curves toward to a flat curve for @xmath228 ( see fig .
[ fig - cg]b ) . according to the finite - size scaling the estimate of @xmath229 can be obtained by extrapolating the data of @xmath230 ) to @xmath231 .
we have studied the conformations of polymer chains consisting of fully flexible and semiflexible chains in a melt based on the bond fluctuation model at volume fraction @xmath33 .
the initial configurations of polymer melts are prepared through several procedures .
we first generate nrrws in a box with periodic boundary conditions in all three directions , rearrange nrrws by the pre - packing process to reduce the local density fluctuations or remain nrrws at the same positions , and then push off overlapping monomers blocking the same lattice sites .
we compare the conformations of polymer melts at the end of each process to the result obtained from the equilibrated configurations .
applying the additional pre - packing process seems to have no significant effect on preparing initial configurations of polymer chains in a melt based on the lattice model after the excluded volume interactions are switched on completely .
namely , the estimates of the mean square internal end - to - end distance @xmath186 and the collective structure factors @xmath81 from the configurations generated by the two methods ( denoted by pre - packing+ev , and ev in figs .
[ fig - rs - melt512 ] and [ fig - sq - melt512i ] ) are almost the same , and very close to estimates from the equilibrated configurations .
there is also no difference of the equilibrating time starting from the initial configurations generated by these two methods .
if , however , one accepts a marginally incomplete elimination of excluded volume violations ( @xmath232 for a system of @xmath233 monomers ) one would have significant advantages .
another possibility of testing this pre - packing process for the future work would be to use it as a criterion of putting chains into the box at the beginning .
in our simulations we combine the algorithms of the local 26 moves , slithering - snake , and pivot moves ( instead of double - bridging moves @xcite ) for equilibrating the system of polymer melts . although the total number of monomers @xmath234 which is eight times smaller than @xmath235 in ref .
@xcite based on the same model for @xmath27 , the estimates of the mean square end - to - end distance @xmath236 , the mean square gyration radius @xmath237 , the mean square bond lengths @xmath238 , and the dimensionless compressibility @xmath205 are all in perfect agreement with those results given in ref .
@xcite at fixed chain size @xmath25 .
for fully flexible and moderately stiff chains the ratio @xmath239 as expected for ideal chains . for moderately stiff chains in a melt the internal mean square end - to - end distance @xmath186
is well described by the freely rotating chain model .
results of the probability distributions of reduced end - to - end distance @xmath240 and reduced gyration radius @xmath241 for polymer chains in a melt for various values of @xmath25 and @xmath12 show the nice data collapse , and are described by universal functions , eqs .
( [ eq - pre ] ) and ( [ eq - prg ] ) , for ideal chains . the collective structure factors @xmath81 for the whole polymer melts and the standard structure factor @xmath1 for single chains in a melt are also calculated and compared with the theoretical predictions .
a detailed investigation of @xmath1 in a kratky - plot for @xmath70 , @xmath78 , @xmath72 , and @xmath73 shows that for fully flexible chains in a melt there is a significant deviation from the debye function for gaussian chains at the intermediate values of @xmath185 as found by wittmer et al .
@xcite , while for @xmath71 it is perfectly described by the debye function .
since real polymers ( such as polystyrene ) exhibit some local chain stiffness , it is clear that the deviations from gaussian statistics found for fully flexible chains in melts in the kratky - plot will be very difficult to test experimentally .
however , since only chains up to @xmath76 are considered in our simulations , such a deviation may also occur as the chain length increases .
careful investigation of the structure factor and the mean square internal end - to - end distance for much longer semiflexible polymer chains in a melt will be required to clarify this matter .
the dimensionless compressibility @xmath205 determined by the estimates of the collective structure factors @xmath81 for small @xmath185 and the local density fluctuations @xmath242 within a sphere of radius @xmath50 are compatible from our simulations for all cases .
we hope that the present work will help to the further development of a coarse - graining approach by using the bfm as an underlying microscopic model and will be useful for the interpretation of corresponding experiments searching for excluded volume effect in the scattering function of polymers in melts .
i am indebted to k. binder and k. kremer for stimulating discussions and for carefully reading the manuscript , and also thank l. moreira and w. paul for their helpful discussions .
i thank the max planck institute for polymer research for the hospitality while this research was carried out .
i also thank the john von neumann institute for computing ( nic jlich ) for a generous grant of computer time and the rechenzentrum garching ( rzg ) , the supercomputer center of the max planck society , for the use of their computers . | we use monte carlo simulations to study polymer melts consisting of fully flexible and moderately stiff chains in the bond fluctuation model at a volume fraction @xmath0 . in order to reduce the local density fluctuations , we test a pre - packing process for the preparation of the initial configurations of the polymer melts , before the excluded volume interaction is switched on completely .
this process leads to a significantly faster decrease of the number of overlapping monomers on the lattice .
this is useful for simulating very large systems , where the statistical properties of the model with a marginally incomplete elimination of excluded volume violations are the same as those of the model with strictly excluded volume .
we find that the internal mean square end - to - end distance for moderately stiff chains in a melt can be very well described by a freely rotating chain model with a precise estimate of the bond - bond orientational correlation between two successive bond vectors in equilibrium .
the plot of the probability distributions of the reduced end - to - end distance of chains of different stiffness also shows that the data collapse is excellent and described very well by the gaussian distribution for ideal chains .
however , while our results confirm the systematic deviations between gaussian statistics for the chain structure factor @xmath1 [ minimum in the kratky - plot ] found by wittmer et al .
\{epl * 77 * 56003 ( 2007 ) . } for fully flexible chains in a melt , we show that for the available chain length these deviations are no longer visible , when the chain stiffness is included .
the mean square bond length and the compressibility estimated from collective structure factors depend slightly on the stiffness of the chains . |
the pairing of massive black holes ( mbhs ) is a natural consequence of galaxy mergers in hierarchical structure formation scenarios @xcite . to date , over 30 dual active galactic nuclei ( agns ) have been discovered in merger remnants at redshifts @xmath5 in the deep survey @xcite , over 160 have be reported from studies of double - peaked [ oiii ] profiles ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and nearly 40 from the swift bat survey @xcite .
furthermore , _
observations of the nucleus of ngc6240 have unveiled a mbh binary with separation @xmath6 kpc @xcite .
the coalescence of mbh binaries may become the dominant source of gravitational waves at mhz frequencies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , a sensitivity range probed by the _ laser interferometer space antenna _ ( lisa )
. asymmetries in the configuration of the coalescing binary cause the beaming of gravitational wave radiation in a preferred direction . to conserve linear momentum , the remnant hole recoils in the direction opposite to the radiative flux @xcite .
the recoil or
kick " velocity , @xmath7 , depends on the binary mass ratio and on the magnitude and direction of their spins , and does not depend on the total mass of the binary .
recent advances in numerical relativity @xcite have allowed several groups @xcite to evolve binary black holes from inspiral through coalescence and estimate the resulting recoil velocities as a function of mass ratio and spin @xcite .
these studies show kick velocities in the range from a hundred to a few thousand @xmath8 .
the maximum speed , @xmath9 , is obtained for equal - mass maximally - spinning black holes with anti - aligned spin vectors in the orbital plane of the binary @xcite .
when @xmath7 is lower than the escape speed of the host , the mbh will wander through the galaxy in an orbit that depends on the detailed characteristics of the host potential .
several groups have calculated the trajectories of mbhs in smooth spherical potentials @xcite .
asphericities in the dark matter potential result in highly non - radial mbh orbits , increasing its decay timescale compared to the spherical case @xcite . in gas - rich galaxy mergers , the wandering hole may travel through a rotationally supported gaseous disk that is inhomogeneous , non - axisymmetric , clumpy , and turbulent .
these factors affect the kinematics of the recoiling mbh and its ability to accrete new fuel . in this paper
, we study the orbits of recoiling mbhs in high - resolution galaxy merger remnants that have successfully formed black hole pairs .
these remnants are the endpoint of the 1:1 major merger simulation of @xcite and the 1:4 and 1:10 minor merger simulations of @xcite .
we construct a semi - analytical model based on the host potential formed after the merger and calculate the trajectories of mbhs for arbitrary recoil velocities and inclination angles , accounting for the effect of dynamical friction and gas drag .
this allows us to obtain the characteristic apocenter distances and return times for mbhs recoiling in different environments .
in addition , in order to study the possible effect of the moving hole on the ambient gas , we perform high - resolution @xmath10-body + sph simulations of recoiling mbhs in the major merger galaxy remnant . under the assumption that the mbh binary is able to coalesce efficiently after reaching the minimum length scale resolved by the simulation , the hole
is given a kick and launched into orbit in an environment characterized by a turbulent , actively accreting gaseous disk surrounded by unrelaxed stellar and dark matter components .
the paper is organized as follows .
the probability distribution of kick velocities is reviewed in [ kickvelocity ] .
we describe the main properties of our three host galaxies in [ hosts ] , and our semi - analytical model in [ model ] .
we assess the response of the gaseous disk to the motion of the mbh , together with its detectability as a velocity - offset or spatially - offset agn in [ detectability ] .
finally , we discuss and summarize our results in [ summary ] .
the recoil velocity depends on the binary mass ratio , @xmath11 , on the magnitude of the black hole spin vectors @xmath12 , and on their orientation with respect to the orbital angular momentum vector .
fitting formulae for @xmath7 to fully general relativistic numerical calculations , given a particular binary configuration , have been provided by several authors ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in the formulation of @xcite
: @xmath13 \\
\nonumber & + & \frac{k_s ( q-1 ) \eta^2}{(1+q)^3 } [ q^2\ , a_1 \sin{\alpha_1}\cos(\phi_1 - \phi_1 ) \\
\nonumber & & + a_2 \sin{\alpha_2}\cos(\phi_2 -\phi_2 ) ] , \end{aligned}\ ] ] where @xmath14 is the contribution to the recoil velocity from mass asymmetry , the indices @xmath15 and @xmath16 refer to projections parallel and perpendicular to the orbital angular momentum , and @xmath17 and @xmath18 are orthogonal unit vectors in the orbital plane
. here @xmath19 are the magnitudes of the hole spin vectors , @xmath20 the angles between @xmath21 and the orbital angular momentum vector @xmath22 , @xmath23 is the symmetric mass ratio , @xmath24 are angles between the projection of @xmath21 along the orbital plane and a reference angle , and @xmath25 are constants that depend on the mass ratio .
figure [ schematic ] shows a schematic of the binary geometry prior to coalescence .
the best - fit parameters are @xmath26 , @xmath27 , @xmath28 , @xmath29 , @xmath30 , @xmath31 , and @xmath32 @xcite . .
the solid line represents the mean of the distribution , the dashed lines show 1@xmath33 deviations from the mean , and the dash - dotted line shows the maximum value attainable . the component @xmath34 vanishes in the case @xmath35 . ] to obtain the recoil velocity distribution as a function of @xmath36 and the probability distribution function at a given @xmath36 , we sample the equation above using a method similar to that of @xcite . for each value of the binary mass ratio we carry out @xmath37 realizations where the black hole spins have random orientations ( with @xmath38 ) or
are aligned with the angular momentum vector ( @xmath35 ) . the angle @xmath39 and spin magnitudes are randomly sampled from a uniform distribution in the range @xmath40 and @xmath41 , respectively .
( _ solid line _ ) , @xmath42 ( _ dashed line _ ) , @xmath43 ( _ dotted line _ ) and @xmath44 ( _ dash - dotted line _ ) . for @xmath35 , @xmath34 vanishes and the pdf is null . ]
0.1 cm figure [ distributions ] shows the distributions of @xmath7 , @xmath45 , and @xmath46 in the case where the black hole spins are either randomly oriented or aligned parallel to the the angular momentum vector .
the maximum recoil velocity , @xmath47 , is attained in the direction parallel to @xmath22 for maximally spinning black holes with anti - aligned spins oriented along the orbital plane , and @xmath44 .
the mean @xmath7 exceeds @xmath48 for mass ratios @xmath49 . if spins are randomly oriented , then the maximum recoil velocity attainable in the orbital plane is smaller than the component along the orbital angular momentum .
the maximum kick in the case @xmath35 is of the order of @xmath50 and is oriented in the orbital plane ( @xmath51 ) .
figure [ pdf ] shows the probability density distributions ( pdf ) of @xmath7 , @xmath34 , and @xmath52 at a given mass ratio , in cases where the spins are randomly oriented or aligned with the orbital angular momentum vector .
the pdfs are calculated for mass ratios @xmath36=0.1 , 0.33 , 0.25 , and 1 . for randomly oriented spins , the peak of the @xmath7 pdf shifts towards lower velocities for lower mass ratios , and
the distribution broadens at higher mass ratios ( see also * ? ? ?
we study the orbits of recoiling mbhs in three galaxy remnants that were previously simulated at high resolution , and in which the mbhs of the progenitor galaxies were shown to successfully formed close pairs .
the remnants are the endpoint of the 1:1 gas - rich major merger described in @xcite , and of the 1:4 and 1:10 gas - rich minor mergers of @xcite .
figures [ host ] and [ rotation ] show a 2-d rendering of the hosts projected density and the rotation curves of all components gas , stars , and dark matter of the three remnants .
our major merger remnant is part of the suit of simulations carried out @xcite .
these simulations tracked the formation of a mbh binary down to parsec scales following the collision between two spiral galaxies , and were run using the massively parallel @xmath10-body + smoothed particle hydrodynamics ( sph ) code gasoline @xcite .
initially , two equal mass galaxies composed of a spherical dark matter halo , an exponential disk , a spherical stellar bulge , and a central mbh , were placed on a parabolic , coplanar , and prograde - prograde orbit .
the galaxies have a virial mass of @xmath53 and a baryonic disk mass of about @xmath54 , with @xmath55 of the disk mass in a cold gas component ( @xmath56 ) and the rest in stars ( see @xcite for details on the structural parameters of the models ) .
radiative cooling and star formation were implemented until the late stages of the simulation , when the dark matter halos had nearly merged and the baryonic cores were separated by about 6 kpc . at this point , in the reference run of @xcite ,
the inner 30 kpc of the simulation volume was refined with the technique of particle splitting to achieve a spatial resolution of 2 pc .
the star and dark matter particles were not splitted to limit the computational burden .
the calculation that we use in this paper has a softening of 10 pc and a mass resolution in the gas component of @xmath57 after splitting ( for a total of @xmath58 sph particles ) .
we use this run rather than the highest resolution simulation presented in @xcite because this resolution matches those of the minor merger runs also used in this work , and because it allows to perform more recoil experiments at a reasonable computational cost ( @xcite verified the convergence in the mass distribution of the merger remnants between the different resolutions ) .
the radiation physics in the refined region was modeled via an effective equation of state , an ideal gas with adiabatic index @xmath59 .
calculations that include radiative transfer have shown that this simple treatment approximates well the thermodynamics of a solar metallicity gas heated by a starburst over a wide range of densities @xcite . by the end of the simulation , @xmath60 gyr after the merger ,
tidal tails continue to evolve in the outer parts of the remnant and over 60% of the gas has been funneled to the central region , forming a compact nuclear gaseous disk of mass @xmath61 , radial length @xmath62 pc , and scale height @xmath63 pc .
the inner disk is turbulent and rotationally supported .
the two mbhs have sunk down from about 40 pc to a few parsecs in less than a million years , are gravitationally bound to each other , and their orbital decay is controlled by gas drag in the supersonic regime , not dynamical friction against the stellar background @xcite . figure [ rotation ] ( left panel ) shows the rotation curve for the different components of the galaxy remnant at the endpoint of the merger .
the gaseous component dominates the dynamics of the inner 300 pc of the remant galaxy .
the escape speeds from the center in the perpendicular ( @xmath64 ) and parallel ( @xmath65 ) directions to the angular momentum vector of the gaseous disk are @xmath66 and @xmath67 , respectively .
the hierarchical @xmath68cdm structure formation scenario predicts a high frequency of minor galaxy mergers @xcite .
@xcite run a set of collisionless ( dry " , with @xmath69 ) and gasdynamical ( wet " ) minor mergers with the same gas fraction in the primary and secondary galaxies , either @xmath70 or 0.3 . as in the major mergers ,
the primary galaxy is a milky way - mass galaxy sampled with @xmath37 particles dark matter particles and ( initially ) @xmath71 star and @xmath72 gas particles .
the secondary galaxy is obtained by rescaling the primary galaxy in mass by 1/4 or 1/10 of its mass , according to the scaling laws of @xcite for cdm galaxies , and is sampled by an equivalent number of particles as the primary . in the 1:10 merger , which is assumed to take place at @xmath73
since this is an ideal redshift window for lisa , the masses , radii and velocities of the galaxies are scaled further by a factor @xmath74 as expected in cdm @xcite .
_ ( b ) _ remnant of a 1:4 merger simulation with initial gas fraction @xmath70 @xcite . _ ( c ) _ remnant of a 1:10 merger simulation with initial gas fraction @xmath75 @xcite .
the scaling is logarithmic and the highest and lowest values of the scale are given in units of atoms @xmath76 .
note that the lower central density in the 1:4 merger remnant , which manifests as a shallower profile of the inner rotation curve inside a few hundred parsecs relative to the 1:10 merger case , is an artifact of the larger gravitational softening adopted in this run.,scaledwidth=40.0% ] while no mbh pair was formed in the collisionless runs ( the two holes were defined as a pair " if their relative orbit shrunk down to a separation equal to twice the softening , which sets the resolution limit ) , the presence of a gaseous component allows the formation of a mbh pair because the massive gaseous cores in which the mbhs are embedded sink efficiently by dynamical friction delivering the holes at the center of the remnant .
furthermore , the higher background density arising as inflows driven by tidal torques associated with bar instabilities and shocks push gas to the center of the galaxies during the merger and enhance dynamical friction in the late stages of the orbital decay @xcite . here
we consider two such dissipational runs : a 1:4 merger with initial gas fraction @xmath70 , and a 1:10 merger with @xmath75 .
both resulted in the formation of mbh pairs separated by twice the softening length ( 200 pc for the 1:4 merger and 40 pc for 1:10 ) , and it is expected that the holes will subsequently merge on a timescale @xmath77 1 gyr ( see * ? ? ? * ) .
the galaxies used in the simulation were initialized with a mixture of primordial metallicities .
the runs were performed with the gasoline code , including this time radiative cooling , atomic line cooling ( by h and he ) , and cooling via collisions , star formation , and blast - wave supernova feedback until the end of the simulation , rather than switching to an effective equation of state . the resulting interstellar medium ( ism )
density distribution is clumpier than in the case of the major mergers that adopt an effective equation of state and neglect supernova explosions .
0.3 cm the most important difference between the major and the minor mergers , however , is that bar - driven torques are much weaker in the latter case .
this is especially true in the primary galaxy , which is only weakly destabilized by the secondary : as a result , gas inflows are much weaker , and @xmath78 of the gas that was originally at large radii reaches the central hundred parsec region by the end of the merger , as opposed to more than @xmath79 in the case of major mergers .
figure [ rotation ] ( middle and right panels ) shows the rotation curves of the minor merger remnants .
they include gas that will eventually form stars ; furthermore , heating and stirring by supernova feedback will make the gas more pressure - supported and extended .
therefore , we expect the mbh return time in minor mergers to be largely determined by dynamical friction against the stars and the dark matter . in addition
, the lower central densities imply a lower escape speed , so we expect that holes with a given kick velocity will travel larger distances in this case compared to a major merger .
the escape speeds from the center in the perpendicular ( @xmath64 ) and parallel ( @xmath65 ) directions to the angular momentum vector of the gaseous disk are @xmath80 and @xmath81 in the 1:4 merger , respectively , and @xmath82 and @xmath83 in the 1:10 merger .
we performed six simulations of recoiling mbhs in the 1:1 merger remnant of @xcite .
a mbh binary of total mass @xmath84 was replaced with a single black hole of the same mass .
the hole was given a kick of @xmath7 = 400 , 800 , and 1,200 @xmath8 at an angle @xmath85 or @xmath86 relative to the angular momentum of the disk . in section [ kickvelocity ]
we showed that maximum recoil velocities are of order @xmath50 if spins are aligned with the orbital angular momentum of the binary . here
, we make no assumptions regarding the spin orientation or the orientation of the binary orbital plane relative to the gaseous disk , since the main objective of the simulations is to calibrate the semi - analytical model described in section [ model ] .
however , these configurations will enter in the calculation of the probability of observing recoiling mbhs as off - nuclear agn during their wandering phase ( see section [ detectability ] ) . for comparison , we have also run a simulation without the mbh .
our gas mass resolution is @xmath87 , for a total of @xmath88 sph particles , and the gravitational softening is 10 pc for the mbh as well as the sph particles .
the hole was kicked starting from where the binary was , at the center of mass of the gas distribution located some 20 pc away from the center of the potential .
the simulations were run using gasoline , and were restarted using the same thermodynamical parameters as in the refined part of the @xcite simulation , i.e. the sph particles were treated as an ideal gas with adiabatic index @xmath59 . since only a small fraction of the nuclear gas turns into stars during the timescale of the simulations , we neglect star formation to reduce the computational burden .
we run the simulations for up to 50 myr , a compromise between obtaining a good calibration for our semi - analytical models and limit cpu time .
it has been shown by @xcite that mbhs moving in nuclear disks with varying mass fraction of stars produce a drag ( via dynamical friction ) that is almost indistinguishable from the purely gaseous case .
it is gas thermodynamics instead that determines the magnitude of frictional drag in both purely gaseous or star forming disks , as it sets the density and temperature distribution of the gaseous disk @xcite . in major merger remnant .
the panels show a hole kicked with _ ( a ) _ @xmath89 and @xmath90 ( parallel to the gaseous disk plane ) , _ ( b ) _ @xmath89 and @xmath85 , _ ( c ) _ @xmath91 and @xmath90 , and _
@xmath91 and @xmath85 .
the colors indicate time along the orbit , from @xmath92 ( _ brown _ ) to @xmath93 ( _ yellow _ ) .
the runtime of the simulation is @xmath94 myr for @xmath89 and @xmath95 myr for @xmath91 .
the black dots mark 1 myr intervals .
orbits are plotted with respect to the deepest point in the disk potential . ]
0.3 cm it is important to emphasize that , in a model with an effective equation of state , the temperature should not be associated with the physical temperature of a specific phase of the interstellar medium in the nuclear disk , e.g. the molecular gas that makes up most of the disk mass @xcite , but rather with an effective pressure that provides a realistic match to the pressure scale height of the disk . in a realistic case
the cold molecular phase is embedded in a hot ionized medium that fills most of the volume and sets the actual pressure scale height .
the pressure scale height of our nuclear disk is @xmath9615 pc at the beginning of the simulations , and @xmath9621 pc at @xmath9720 myr , and compares well with the pressure scale height of 23 pc observed in mrk 231 @xcite .
likewise , the effective sound speed used throughout the paper ( calculated directly from the simulations ) should be close to the sound speed of the hot , diffuse phase ( the cold molecular phase has a sound speed of @xmath98 ) .
the peak of the sound speed in the major merger is @xmath50 , while it is @xmath99 and @xmath100 in the 1:4 and 1:10 mergers , respectively .
figure [ orbits ] shows the resulting orbits for mbhs kicked with @xmath89 and @xmath91 in a direction that is either along the disk plane ( cases _ a _ and _ c _ ) or perpendicular to it ( cases _ b _ and _ d _ ) .
the boxes have varying lengths and the orbits are centered on the point where the potential is the deepest .
the orbit for the case @xmath91 and @xmath90 is shown again in figure [ orbit ] , together with the evolution of the background gas number density and specific internal energy .
to obtain a large sample of mbh orbits at relatively low computational cost , we have constructed a semi - analytical model based on our numerical results .
we can then follow the trajectories of mbhs until they return to the center of their host , on timescales that can be much longer than the time it is practical to run a full @xmath10-body + sph simulation , particularly for larger kick velocities .
the motion of a mbh in a multi - component potential can be studied analytically by integrating numerically its equation of motion under the action of a conservative force @xmath101 and a damping frictional term : @xmath102 where the total potential is @xmath103 .
the gradient of the potential is computed directly from the simulation data .
the dynamical friction term for an object of mass @xmath104 moving with relative velocity @xmath105 through a gaseous medium of density @xmath106 is given by @xmath107 the velocity integral , @xmath108 , depends on the mach number @xmath109 , where @xmath110 is the sound speed . in our model , the local gas density and sound speed are measured directly from the simulation data in a sphere of radius @xmath111^{1/3}\ ] ] around the mbh , where @xmath112 is the average density at a distance @xmath113 from the host center and we choose a reference radius @xmath114pc . as for the shape of the velocity integral
, we adopt the hybrid formulation of @xcite , who combined the prescriptions of @xcite and @xcite into : @xmath115 \\ ( 0\leq \mathcal{m } \leq 0.8);\\ 1.5 \ln \lambda \left [ { \rm erf } \left ( \frac{\mathcal{m}}{\sqrt{2}}\right ) -\sqrt{\frac{2}{\pi } } \mathcal{m } \exp\left ( { -\frac{\mathcal{m}^2}{2}}\right ) \right ] \\ ( 0.8<\mathcal{m } \leq 1.5);\\ 0.5\ln \left ( 1-\frac{1}{\mathcal{m}^2}\right ) + \ln\lambda \\ ( \mathcal{m } > 1.5 ) . \\ \end{array}\right.\end{aligned}\ ] ] the coulomb logarithm is set to @xmath116 @xcite . in a collisionless medium ,
the velocity integral depends on the local velocity distribution of the stars or dark matter .
if the medium is isotropic and the velocity distribution is maxwellian , then @xmath117 , where @xmath33 is the one - dimensional isotropic velocity dispersion . equation ( [ df ] ) then becomes the chandrasekhar dynamical friction formula @xcite , where @xmath118 and @xmath106 is the stellar or dark matter density . in a recent merger ,
however , the velocity distribution of the stellar and dark component will deviate from maxwellian .
we therefore calculate the velocity integral numerically from the actual velocity space distribution of the particles in the simulations , @xmath119 where @xmath120 is the number of particles with velocities between @xmath121 and @xmath122 in the vicinity of the hole .
the equation of motion of the mbh is integrated numerically using an adaptive adams - bashforth - moulton algorithm .
0.3 cm figure [ model_comp ] shows a comparison between the orbits of a mbh in our numerical simulations versus the semi - analytical model .
the hole was kicked with @xmath123 in the gaseous disk of the major merger remnant .
the trajectory is shown in projection in the left and middle panels , while the right panel depicts the radial evolution of the orbit . over the timescale covered by the simulation ,
the semi - analytical model is clearly able to reproduce both the amplitude of the oscillations ( dictated by the potential ) and the decay by friction .
figure [ apocenter ] shows the first apocenter distance @xmath124 , estimated from our semi - analytical approach , for mbhs recoiling in our 1:1 , 1:4 , and 1:10 merger remnants along a direction that is either parallel or perpendicular to the disk angular momentum vector .
the black hole mass was scaled in each case to match the @xmath125 relation , yielding @xmath126 for the 1:1 , 1:4 , 1:10 mergers , respectively . at supersonic speeds
, gas drag plays an important role in damping the mbh orbit in the 1:1 remnant . shortly after the kick , when the hole s is moving at mach number @xmath127 , gas drag accounts for over 70% of the total frictional force . in the deep potential
well of the remnant , the gas density can reach @xmath72 particles @xmath76 , and mbhs kicked in the plane of the disk with initial velocities as high as @xmath1 reach apocenter distances of only 0.3 kpc .
as @xmath7 approaches the escape speed along this direction , i.e. @xmath128 , the first apocenter distance reaches @xmath129 kpc .
mbhs recoiling in the direction parallel to the disk angular momentum vector can reach the low - density outer regions for slightly lower initial kicks , and therefore reach larger apocenter distances .
a hole with @xmath130 reaches @xmath131 kpc and has a return timescale that is close to 10 gyr .
minor mergers have significantly lower central density profiles , and apocenter distances can be significantly larger even for modest recoil speeds . in the @xmath132 merger remnant , a kick of @xmath133 perpendicular to the disk can displace the hole by @xmath134 kpc . the same initial speed yields a maximal separation of @xmath135 kpc in the @xmath136 case .
three main methods have been proposed to detect a recoiling hole : 1 ) through the imprint of the moving mbh on the gaseous medium of the host galaxy ; 2 ) as a kinematically - offset agn , as the broad line region ( blr ) carried by the hole is displaced by a velocity @xmath137 from the narrow starlight associated with the host galaxy ; 3 ) as a spatially - offset agn , where the accreting mbh is observed to be displaced by a distance @xmath138 from the host s nucleus .
@xcite recently carried out a suite of numerical simulations aimed at studying the detectability of recoiling mbhs in the hot gas of elliptical galaxies .
the host galaxy was modeled as a spherical hernquist sphere @xcite , with a mass resolution of @xmath139 and a polytropic equation of state with @xmath140 .
they found that , at high mach numbers , the shocked gas can produce an x - ray signature that may be detectable with the _ chandra _ satellite in nearby galaxies .
simulations of recoiling holes surrounded by a thin circumbinary disk @xcite show that the likelihood of observable emission is low and depends strongly on the orientation of the kick relative to the thin disk , on the magnitude of the kick , and on the mass of the disk . in the numerical simulations described in section [ simulations ] ,
the inner disk is rotationally supported ( @xmath141 ) , turbulent ( @xmath142 at @xmath143 pc ) , and actively accreting material from the outer regions .
dense columns of cooler gas reach the nuclear region , constantly changing its internal energy profile .
we find that the ratio between the integrated energy deposition of the mbh ( associated with energy losses from gas frictional drag ) and the integrated change in internal energy by the gas ( measured in the simulation without mbh ) is very small , @xmath144 . as a result ,
overdensities produced in the gas by the recoiling mbh are rapidly weakened .
in contrast with the results of @xcite , the removal of the central hole from the disk potential does not have a significant effect on the central density profile , since the mass of the hole amounts to only @xmath145 of the mass of the nuclear disk . at the end of the major merger @xmath146 mbh simulation ,
the gas density and internal energy profiles ( fig . [ orbit ] ) are nearly undistinguishable from those in the no - mbh run . by contrast , in the 1:10 merger , the integrated change in gas internal energy is comparable to the orbital energy lost by the mbh over a few tens of orbits .
high mach numbers can be reached at lower recoil speeds since the sound speed profile rises slowly to reach @xmath147 at 50 pc and then stays constant out to 100 pc ( whereas in the major merger the sound speed reaches a peak of @xmath148 at 10 pc and then drops to @xmath149 at 100 pc ) .
this suggests that minor merger remnants may be suitable sites for the detection of recoiling mbhs through their interaction with ambient material .
the detectability of mbhs as offset agns depends on the hole s phase - space location at the time of observation and on its ability to accrete material and shine at that location .
we assume that , upon ejection , the hole carries a punctured accretion disk ( e.g. * ? ? ?
* ) and an associated blr that is visible for a time comparable to the viscous lifetime of the disk . the mass carried by the disk is @xcite @xmath150 where @xmath151 is the viscosity parameter in units of the fiducial value @xmath152 , @xmath153 is the mbh mass in units of @xmath154 , @xmath155 , @xmath156 is the matter - to - radiation conversion efficiency in units of 10% , @xmath157 is the hole s accretion rate in units of the eddington rate @xmath158 ( we use @xmath159 ) , and @xmath160 is the kick velocity in units of @xmath161 . the condition @xmath162 requires @xmath163 when the above equation is not satisfied , we impose @xmath164 : this yields an agn lifetime equal to @xmath165 .
the condition @xmath166 also sets the maximum outer radius of the disk @xmath167 where @xmath168 is the schwarzschild radius of the hole . at large radii ( @xmath169 pc ) , gas in the disk may be gravitationally unstable and prone to star formation @xcite .
however , the maximum disk size considered in our model is @xmath170 pc , and therefore we neglect this effect in our analysis .
we also find that ram pressure stripping will not affect the accretion disk as the mbh travels through the host . for a hole moving with velocity @xmath105 relative to a medium of density @xmath106 , we can calculate the stripping radius @xmath171 by equating the ram pressure force @xmath172 to the restoring force of the disk , using the density profile of @xcite . in the cases considered here , @xmath173 , and
therefore the disk is able to survive . because precise measurements of the peaks of broad emission lines in quasar spectra are challenging , the observation of potential recoil candidates as kinemtically - offset agns whose broad emission lines are substantially shifted in velocity relative to the narrow - line gas left behind @xcite is biased toward large kicks . in the following
we shall assume that a recoiling hole is detectable as a kinematically - offset agn if its velocity shift , @xmath137 , is above the detection threshold @xmath174 .
at every time step of our numerical integration of the hole s orbit , we calculate the instantaneous accretion rate , the associated luminosity , and update the mbh mass . when the initial bound disk is exhausted , the blr is assumed to vanish and the recoiling mbh is no longer recognizable as such from kinematics measurements . from this point on
, the mbh is allowed to accrete material from the ism .
the fueling rate depends on the local density and sound speed of the gaseous medium and on the speed of the hole following the bondi - hoyle formula : @xmath175 where @xmath176 is the local gas number density of the ism in units of @xmath76 , @xmath177 is expressed in units of @xmath161 , and @xmath105 is the instantaneous mbh velocity relative to the host .
the mass accretion rate in this phase is given by the minimum between the bondi rate and the eddington rate , and the bolometric luminosity is then @xmath178 to be detected as a spatially - offset agn , a recoiling accreting hole must be both above a threshold spatial separation @xmath179 from the host center as well as above a luminosity threshold @xmath180 .
we set @xmath179 equal to roughly twice the resolving power of the telescope 0.2
`` for the _ hubble space telescope ( hst ) _ and the _ james webb space telescope ( jwst ) _ , and 1 '' for _
chandra_. for reference , at redshift @xmath181 ( 0.1 ) an offset of 10 kpc corresponds to an angular separation of 1.25 `` ( 5.4 '' ) . in the following
, we set the threshold x - ray luminosity to @xmath182 where @xmath183 is the on the on - axis _
limit and @xmath184 is the luminosity distance , and use a bolometric to x - ray correction of @xmath185 if @xmath186 and @xmath187 otherwise @xcite .
the optical b - band luminosity threshold @xmath188 is calculated using a limiting flux of @xmath189 from observed broad - line agn @xcite and the b - band bolometric correction is taken from @xcite .
both x - ray and optical thresholds are consistent with observational limits from multi - wavelength studies of agn ( e.g * ? ? ?
recoiling with @xmath190 along the nuclear disk of the major merger remnant at @xmath191 .
the mbh undergoes disk accretion and is shining with an x - ray luminosity @xmath192 along the thin red line , and is powered by ism accretion in the orange region . ]
0.3 cm figure [ phase ] shows a breakdown of the orbital phase diagram of a mbh of mass @xmath193 recoiling with velocity @xmath190 in the plane of the disk of the major merger remnant at @xmath191 .
fiducial regions where @xmath194 , @xmath195 , and @xmath196 are marked in the velocity - distance plane . in this example @xmath197 correspond to 1 kpc .
the mbh undergoes disk accretion and is shining with an x - ray luminosity above @xmath198 along the thin red line : it is therefore detectable as a kinematically offset agn when it is still close to the center , along the portion of the red line where @xmath199 .
the mbh is instead detectable as a spatially - offset agn in the low - velocity region where it accretes rapidly from the ism ( orange ) or from the initial disk ( red ) and where @xmath200kpc at this redshift .
note that large kinematic offsets occur near pericenter and can be observable for a shorter timescale than large spatial offsets , which occur near apocenter . in the following , we study in details the detectability of recoiling mbhs in two cases .
+ . _ dashed cyan curves _ : total time during which the mbh has velocity offset @xmath199 , i.e. above the threshold for the detectability of kinematically offset agns . _ solid cyan curves _ : total time over which the mbh has a velocity offset @xmath199 while shining with an optical luminosity @xmath201 ( for @xmath202 ) .
the dashed blue and solid blue curves use a threshold for detection of kinematically offset agns of @xmath203 .
_ solid black curve _ : lifetime of the accretion disk carried by the hole as a function of recoil velocity . ] 0.3 cm .
a best - case scenario for the detectability of recoiling mbhs as offset agns , in which : ( a ) the spins are randomly oriented with respect to the orbital angular momentum prior to the merger of the mbh binary ( a possible mechanism for misalignment of the spins is chaotic accretion , see @xcite ) ; ( b ) the orbital plane of the binary is randomly oriented with respect to the plane of the gaseous disk ( e.g. because of turbulent motions at sub - pc scales that may be uncorrelated with the angular momentum of the gaseous disk ) ; and ( c ) 1:1 , 1:4 , and 1:10 galaxy merger remnants host mbh binaries of mass ratios @xmath204 , respectively . in a recent study
, @xcite showed that during a 1:10 merger the primary galaxy may induce strong tidal torques onto the satellite that can result in the rapid growth of the smaller black hole , yielding in some cases binary mass ratios of @xmath205 . in simulations where the gas fraction was set to @xmath75 and
the merger was coplanar , a setup similar to the our initial 1:10 merger , the final mass ratio of the binary was found to be @xmath43 .
+ . a worse - case scenario for the detectability of recoiling mbhs as offset agn , in which : ( a ) the spins are aligned with the orbital angular momentum vector of the binary prior to the merger event .
studies of the spin evolution of mbhs in a gaseous medium show that spins tend to align in gaseous environments due to frame dragging , yielding maximum recoil velocities of @xmath206 @xcite .
this effect drastically decreases the probability of observing both kinematically and spatially offset agns in gaseous environments ; ( b ) the orbital plane of the binary is aligned with the gaseous disk .
this assumption , together with ( a ) , implies that all recoils occur in the plane of the disk ; and ( c ) 1:1 , 1:4 , and 1:10 galaxy merger remnants host mbh binaries of mass ratios @xmath207 . in both cases
we consider a random uniform distribution of spin magnitudes in the range @xmath208 .
conservation of linear momentum also requires that the mass of the disk carried by the hole reduces the initial kick velocity to an effective value @xmath209 we account for this effect when estimating the detection probabilities associated with recoiling mbhs below .
several candidates of velocity - offset agns have been discussed in the literature ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , but none has been positively confirmed to date .
the study of @xcite shows that there is only scarce ( if any ) evidence for recoiling mbhs in _
sloan digital sky survey _ quasars . while shining with an optical luminosity @xmath201 ( for @xmath202 ) . _ blue curves : _ same for @xmath210 .
both probabilities were computed assuming our best - case scenario case 1 ( see text for details ) . ] 0.3 cm to assess the detectability of recoiling mbhs as kinematically offset agns , we have generated a library of orbits using the semi - analytical model described in section [ model ] , for kick velocities lower than the escape speed from the host and return timescales shorter than the hubble time .
we have computed the accretion rate and the resulting optical and x - ray luminosities along the trajectory of the recoiling holes .
figure [ timescales_dv ] shows the return timescales of recoiling mbhs in the major merger remnant , for different kick velocities in the plane of the nuclear disk and perpendicular to it .
the figure also shows the lifetime of the initial disk carried by the hole , the total time spent by the hole above the threshold velocity offsets @xmath211 and @xmath212 , as well as the detectability window as a velocity - offset agn , i.e. the total time during which the mbh is both above the velocity and the luminosity thresholds ( at @xmath202 ) .
the detectability window of a recoiling hole depends on the kinematic profile of its orbit . for moderate kick velocities , where the mbh can undergo many pericenter passages before the initial disk is exhausted ,
more eccentric orbits yield higher relative velocities , making the mbh observable as a kinematically - offset agn .
when the mbh is kicked in the direction perpendicular to the disk , the return timescale is longer and orbits are less eccentric , yielding generally shorter observability windows for large kinematic displacements . in our major merger , a mbh launched with recoil velocity @xmath91 in the @xmath90 direction , will return to the center of the galaxy in about 0.3 gyr , and would be observable as a kinematically - offset agn for only 0.2 myr .
the detectability window decreases by a factor of 4 in the @xmath85 case for the same recoil speed . in the minor mergers
the threshold velocity is reached only for kicks with @xmath213 , which are are not shown here .
the probability of observing a recoiling hole as a velocity - offset agn with @xmath214 along the line of sight ( los ) can be written as @xmath215 where @xmath216 is the probability that a hole with initial kick @xmath7 has an velocity offset @xmath217 while moving along the los at the time of observation and shining with bolometric luminosity above the threshold @xmath180 ( assuming no obscuration ) . here , @xmath218 is the probability that a coalescing mbh binary with spin vectors @xmath219 and mass ratio @xmath36 recoils with velocity @xmath220 . in terms of the relevant timescales , @xmath221 \\ \nonumber & \times & p_{{{v_{\rm kick } } } } ( \vec{a}_1,\vec{a}_2,q),\end{aligned}\ ] ] where @xmath222 is the integrated time during which @xmath217 along the los , and @xmath223 .
we have evaluated equation ( [ pdv ] ) for a range of recoil velocities , converting each value of @xmath7 into an effective recoil velocity @xmath224 ( using equation [ veff ] ) , and interpolating the values of @xmath225 using the timescales shown in figure [ timescales_dv ] . the resulting probability is shown in figure [ prob_case1_dv ] under the assumptions of our best - case scenario case 1 ( random alignment of the binary orbital plane and spin ) .
the probability peaks at @xmath226 to a maximum value of only @xmath227 for @xmath228 , and is nearly 2 orders of magnitude smaller for @xmath210 .
we calculated the same probability in case 2 , where the spins of the binary are aligned with the orbital angular momentum vector , the orbital plane is aligned with the gaseous disk , and all kicks occur in the plane of the disk . in this scenario
the maximum recoil velocity is @xmath229 and the holes are unobservable as kinematically - offset agns .
0.3 cm we can use a similar method to that described above to estimate the observability of recoling mbhs as spatially - offest agns .
figure [ timescales_dx ] shows the integrated time during which the mbh is displaced by a threshold distance of @xmath230 and @xmath231 at @xmath202 ( corresponding to twice the resolution of _ hst / jwst _ and _ chandra _ , respectively ) , as well as the integrated time during which the mbh is both displaced by @xmath179 and is shining with @xmath232 .
the upper , middle , and lower panels show results for the 1:1 , 1:4 , and 1:10 mergers , respectively . in the major merger ,
only large recoil velocities ( of the order the escape speed ) can displace the mbh by @xmath179 when the kick occurs in the plane of the gaseous disk .
the hole can shine for timescales far longer than the lifetime of the accretion disk carried along , since there is a large supply of gas in the galactic disk for fueling near apocenter , where the mbh spends most of its wandering time . in the minor mergers ,
the potential wells are shallower and large apocenter distances can be reached at low kick velocities ( @xmath233 ) . for trajectories having low inclinations relative to the gaseous disk
, bondi accretion can be significant and the mbh can wander sufficiently far from the center of the host to be detectable as an offset x - ray agn at @xmath234 .
the detectability window typically ranges from a few to tens of myr .
the probability of observing a recoiling mbh as a spatially - offset agn with displacement @xmath235 perpendicular to the los can be written as @xmath236 where @xmath237 is the probability that the hole is shining with resulting luminosity @xmath238 and is displaced by a projected distance larger than the threshold @xmath179 . in terms of the relevant timescales , @xmath239 \\ \nonumber & \times & p_{{{v_{\rm kick } } } } ( \vec{a}_1,\vec{a}_2,q).\end{aligned}\ ] ] and @xmath240 . _
dashed lines : _ @xmath241 and @xmath242 .
all probabilities were computed assuming our best - case scenario case 1 ( see text for details ) . ]
0.3 cm figure [ prob_case1_dr ] shows the probability @xmath243 that a mbh could be observed as a spatially - offset agn during its wandering time assuming case 1 . along both orientations considered here ( on and off the plane of the gaseous disk ) ,
the probability peaks at @xmath244 for optical detection , @xmath245 , and is higher for the 1:10 merger where the recoiling hole reaches larger apocenter distances .
the probability that such a source could be observed by _
instead drops by nearly an order of magnitude . for the 1:1 merger remnant the peak of the distribution is @xmath244 in the @xmath90 case .
the distribution broadens in the @xmath85 case peaks at a lower value since the mbh spends a considerable fraction of its return time in a non - active phase .
we have carried out the same calculation in case 2 , and find that the likelyhood of observing spatially - offset agns becomes negligible .
there are several reasons for this : ( 1 ) when spins are aligned , the maximum recoil velocity is @xmath206 ; ( 2 ) at low recoil speeds the mass carried by the disk is @xmath164 , and the effective recoil velocity is @xmath246 ; and ( 3 ) recoil velocities in the direction perpendicular to the gaseous disk are not allowed since @xmath35 .
we have studied the orbital evolution of recoiling mbhs in post - merger environments via high - resolution @xmath10-body + sph simulations and a semi - analytical model that combines the results of merger simulations with a prescription for mbh accretion .
our results can be summarized as follows : 1 .
we find little evidence for an observable imprint of a recoiling , non - active mbh in the gaseous turbulent medium of our major merger remnant . in our simulation ,
the total energy deposited by the moving hole is only @xmath145 of the change of internal energy of the gas , even at high mach numbers .
the contribution to the thermodynamics of the ambient gas becomes significant in minor merger remnants instead , potentially allowing for an electromagnetic signature of mbh recoil .
we have built a library of mbh orbits in three gas rich galaxy mergers , allowing the holes to carry a punctured accretion disk and the associated blr upon ejection , and to accrete from the ism when the disk is exhausted . in the major merger remnant , the combination of both deeper central potential well and drag from high - density gas confines even mbhs with kick velocities as high as @xmath1 to within 1 kpc from the host s center . in minor merger remnants
gas drag is weaker , the potential is shallower , and smaller recoil speeds can displace the hole to larger distances .
we have dissected the orbits of recoiling holes into regions where they may be detectable as kinematically and/or or spatially offset agns .
kinematically - offset agns carry their accretion disk and associated blr for a timescale @xmath247 and can be observable as such when their broad emission lines have a velocity shift @xmath248 relative to the narrow - line gas left behind . in the major merger remnant , moderate recoil velocities ( @xmath2 )
provide the most likely scenarios for the detection of kinematically offset nuclei , as the initial disk can fuel the mbh during several pericenter ( high - velocity ) passages .
the convolution of the total fractional time spent as a kinematically - offset agn with the probability that a coalescing binary receives a given kick velocity yields the overall probability of observing recoiling mbhs as kinematically - offset agns .
this is small , of order @xmath249 , even in the optimistic case ( our case 1 " ) where the spins of the binary holes are randomly oriented relative to the orbital angular momentum vector and the orbital plane of the binary is randomly oriented relative to the galactic disk .
it is totally negligible when the spins are aligned instead ( e.g. * ? ? ?
* ) , and the maximum recoil speed drops to @xmath206 .
4 . at low redshift
, spatially - offset agn can be observed with both _
chandra _ and _ hst / jwst _ for a few myr in major mergers for high recoil speeds ( @xmath250 ) . in minor mergers ,
the detectability window shifts to lower recoil speeds ( @xmath233 ) and extends to up to a few 100 myr in some cases . under the case 1
optimistic scenario , the probability of observing spatially - offset agns is larger ( by two dex ) in the 1:10 merger than in the 1:1 merger , since in the former larger offsets occur for lower , more likely kicks .
the peak of the probability distribution ( @xmath251 ) occurs for @xmath252 and optical detection , and is lower by about one dex in the case of x - ray detection .
our semi - analytical model has been calibrated using a limited set of numerical simulations .
the choice of a gas fraction of @xmath55 for the major merger is rather conservative ; a larger gas content , quite likely at high redshift , would cause increased drag on the mbh and reinforce our conclusions . for the 1:10 minor merger , representative of some of the most common types of mergers during the hierarchical assembly of galaxies
, we have assumed a larger gas fraction , 30% , as in this case the pairing of the two holes and their coalescence is significantly faster .
such a choice is still on the conservative side ; the typical gas fraction in galaxies with masses in the range @xmath253 ( corresponding to the virial masses of the primary and secondary galaxies in the 1:10 mergers , respectively ) is in the range 30 - 50% at low redshift @xcite and would be even higher at early epochs . as for the mass ratio , our simulations sample quite well the regime in which the dynamical friction timescale is short enough for the galaxies to merge in a few orbits . at larger mass ratios
the satellite galaxies will suffer severe tidal mass losses as dynamical friction reduces their pericentre distance : this will increase substantially their decay time @xcite and inhibit the formation of a mbh binary .
a limitation of the present study is that our simulations do not capture all the phases of the ism : gas is either warm , with temperatures close to @xmath254 k , or hot ( @xmath255 k ) as a result of heating by supernovae explosions or shocks in the final collision of the merging galaxy cores .
the cold molecular phase that is believed to represent a significant fraction of the mass of the nuclear disk ( e.g. * ? ? ?
* ) is missing in our current calculations . at temperatures of hundreds of kelvins in the presence of a starburst @xcite , molecular material
may be arranged in a thin and dense sheet the midplane of the nuclear disk ( e.g. * ? ? ?
mbhs recoiling along the plane may then suffer more dynamical friction , and one may expect a greater difference between the fate of in - plane and off - plane recoils than seen in our study .
the finite resolution of our numerical simulations may also affect our results .
in particular , the resolution of our 1:4 merger ( 200 pc ) results in a sharp decline of the gas distribution in the remnant nucleus . to quantify this effect
, we have rescaled the central gas and stellar densities to their corresponding values at 200 pc .
we find that the recoiling mbhs in the 1:4 merger reach smaller apocenter distances and wander for even shorter timescales .
the effect is more dramatic for kicks below @xmath50 . since mbhs with these velocities are not observable as neither kinematically nor spatially - offset agns , our detection probabilities remain largely unaffected .
another important limitation of our semi - analytical modelling is the neglect of agn feedback , which acts to reduce the efficiency of frictional drag by heating the gas surrounding the recoiling mbh .
recent work by @xcite shows that in simulations where agn feedback is fully included , the mbh creates a hot low - density ( instead of a high - density ) wake that expands away from the hole .
the mbh describes a precessing elliptical orbit that tend to circularize , and decays towards the center of the remnant on a longer return timescale . to get a simple estimate of the effect of agn feedback on our calculations we have assumed in our semi - analytic model that at each time - step a fraction equal to 5% of the radiative energy produced by the accreting mbh goes directly towards increasing the local thermal energy of the gas .
we find that in dense regions where the hole is accreting at high bondi rates , the reduction of the effective mach number from such energy deposition lengthen the return timescale by 15% in the major merger for @xmath256 and 20% for @xmath257 in the 1:10 merger .
the time during which the mbhs are observable as spatially offset agn increases by 6% and 9% respectively .
the overall detection probability however , remains low .
furthermore , in our detectability calculations we have not included obscuration by dust .
because theoptical and x - ray reddening does not depend on the orientation of the host galaxy @xcite , it is believed that most of the obscuration occurs in a region close to the mbh , with optical light being more strongly affected that x - rays and infrared wavelengths ( e.g. * ? ? ?
* ; * ? ? ?
in particular , the recent study of @xcite shows that the _ swift _ bat x - ray survey has revealed roughly 5 - 6 times more dual agns than the sloan digital sky survey ( sdss ) due to optical extinction and dilution by star formation . as a result
, the detection probability is further reduced for _ hst / jwst _ sources . while we have pointed out the low likelihood that moderate - to - large recoils can be observed , one possibly observable consequence of this phenomenon is the existence of a population of hot - dust - poor agns due to the removal of the mbh and associated blr from the surrounding hot dust region .
nearly @xmath55 of spectroscopically confirmed type 1 agns samples @xcite , and xmm - cosmos ( elvis et al .
2010 in prep ) show a relatively weak infrared bump , associated with dust emission . the number of these objects has been shown to increase with redshift , from 6% at @xmath258 to @xmath259 at @xmath260 @xcite .
they find that the covering factor of these agn is only 2 - 30% , significantly smaller than the 75% predicted by the unified model .
since these agn are in the redshift range @xmath261 , these hot - dust - free agn are not recently born agn which have not time to form a dusty torus @xcite . based on their spectral resolution ( 2 )
, a @xmath262 source could have a relative velocity of @xmath263 .
we thank the anonymous referee for thoughtful comments that substantially improved our paper .
this research was funded by nasa through grants nnx09aj34 g ( p.m. ) and by nsf through a graduate student research fellowship ( j.g ) .
simulations were carried out on the pleiades supercomputer at nasa ames , and some results were rendered using the splash code @xcite . | the asymmetric emission of gravitational waves produced during the coalescence of a massive black hole ( mbh ) binary imparts a velocity kick " to the system that can displace the hole from the center of its host .
here we study the trajectories and observability of mbhs recoiling in three ( one major , two minor ) gas - rich galaxy merger remnants that were previously simulated at high resolution , and in which the pairing of the mbhs had been shown to be successful .
we run new simulations of mbhs recoiling in the major merger remnant with mach numbers in the range @xmath0 , and use simulation data to construct a semi - analytical model for the orbital evolution of mbhs in gas - rich systems .
we show that : 1 ) in major merger remnants the energy deposited by the moving hole into the rotationally supported , turbulent medium makes a negligible contribution to the thermodynamics of the gas .
this contribution becomes significant in minor merger remnants , potentially allowing for an electromagnetic signature of mbh recoil ; 2 ) in major merger remnants , the combination of both deeper central potential well and drag from high - density gas confines even mbhs with kick velocities as high as @xmath1 within 1 kpc from the host s center ; 3 ) kinematically offset nuclei may be observable for timescales of a few myr in major merger remnants in the case of recoil velocities in the range @xmath2 ; 4 ) in minor mergers remnants the effect of gas drag is weaker , and mbhs with recoil speeds in the range @xmath3 will wander through the host halo for longer timescales . when accounting for the probability distribution of kick velocities , however , we find that the likelihood of observing recoiling mbhs in gas - rich galaxy mergers is very low , typically below @xmath4 . |
off - nuclear ( non - agn ) point - like luminous x - ray sources , which exceed the eddington luminosity of a usual stellar - mass compact object like neutron stars and galactic binary blackholes ( @xmath5 @xmath6 $ ] ) in nearby galaxies have been recognized in many galaxies @xcite .
the luminosity distribution of these sources extends up to a few @xmath7 @xmath6 $ ] .
the most luminous one has been found in the nuclear region of the starburst galaxy @xcite reaching @xmath8 @xmath6 $ ] .
the location of this source was recently verified to be offset from the dynamical center of the galaxy by _ chandra _ @xcite .
one interpretation of these x - ray sources is an accretion onto intermediate - mass blackholes of @xmath9 @xmath10 , while there are other possibilities including beaming , super - eddington accretion , and a young supernova remnant in a dense interstellar medium .
probably they do not form a single class of x - ray sources .
the spectra of many of these sources can be described by a multicolor disk blackbody model with or without additional hard tail @xcite .
the inner disk temperatures of the sources studied by @xcite are @xmath11 [ kev ] . since these temperatures are higher than that from an intermediate ( @xmath12 ) blackhole accretion , they proposed a picture of a spinning ( kerr ) blackhole with a much smaller inner - disk radius than that of a non - spinning ( schwarzschild ) one .
the luminous x - ray source in the star - forming dwarf galaxy holmberg ii ( ugc 4305 ) is particularly peculiar among these cases in terms of the nature of the host galaxy and the location of the x - ray source .
the galaxy is nearby at a distance of 3.05 mpc , which was derived from cepheid observations @xcite .
it is a dwarf irregular galaxy in the m81 group with an on - going star - formation activity . like in other galaxies of this class
, numerous hii regions are scattered throughout the galaxy @xcite .
the bright x - ray source , which has been cataloged in the _ rosat _ bright survey ( rbs ) @xcite , is one of the most luminous x - ray source seen in dwarf galaxies ( @xmath13 @xmath14}$ ] ) .
it was observed three times with _
rosat _ pspc as pointed observations and once with hri .
@xcite ( hereafter zgw ) studied the _
data on this x - ray source .
their main conclusions were : 1 ) the x - ray source seems point - like and located near the edge of the galaxy .
its position is consistent with one of the compact hii region complexes of the galaxy considering an uncertainty of @xmath15 arcseconds for the _ rosat _ absolute astrometry from its star tracker .
a chance spatial co - incidence of this hii - region to an unrelated x - ray source of this brightness is highly unlikely ( @xmath16 ) .
2 ) the source is time variable on scales of days and years .
3 ) the spectrum is soft and can be fitted with a steep power - law ( @xmath17 ) or a thermal plasma ( raymond - smith ) of @xmath18 kev .
time variability suggests an accretion onto a compact object , but the luminosity is 2 - 4 orders of magnitude larger than typical x - ray binaries . there is a radio - knot which is co - spatial with the hii - region in the x - ray source region .
it is one of the three showing radio spectra suggestive of a non - thermal emission @xcite , suggesting a recent supernova or multiple supernovae .
however , currently there is no indication of a special character in the optical - radio data in this particular hii - region which is presumably associated with the compact x - ray source compared with other knots in the galaxy . as a part of our study of a sample of apparent non - agn galaxies which have compact luminous x - ray sources , selected from the rbs ,
we have made further x - ray studies of holmberg ii .
in particular , we have obtained a 20 [ ks ] _ asca _ observation of this x - ray source in order to search for the hard x - ray signature of an agn - like activity . in this paper , we report the results from our data analysis of our _ asca _ observation and of the _ rosat _ data from archives .
the scope of the paper is as follows . in sect .
[ sec : obs ] , we summarize the _ asca _ and _ rosat _ observations and data used in our analysis . in sect .
[ sec : spec ] we report our new spectral analysis of the x - ray source by a joint _ asca_-_rosat _ pspc analysis . in sect .
[ sec : hri ] , we searched for an extension by applying a wobble - correction to the _ rosat _ hri data .
long and short - term x - ray light curves are examined in sect .
[ sec : lc ] .
finally the results are discussed in sect .
[ sec : disc ] .
holmberg ii was observed with _
rosat _ five times .
firstly , it was bright in the _ rosat _ all - sky survey ( rass ) @xcite and cataloged in the rbs @xcite .
it was observed three times with _ rosat _
pspc pointed observations in 1992 - 1993 for 3.7 - 11 [ ks ] each .
also it was observed with _ rosat _
hri for 7.8 [ ks ] .
the _ rosat _
pointed data have been retrieved from the high energy astrophysics archival research center ( heasarc ) located at nasa goddard space flight center .
the asca observation of holmberg ii has been made with a pointing at the x - ray source in 1999 as one of the accepted targets from our proposal in the _ asca _ eao-6 ( ao-7 ) program ( pi = lehmann ) .
the brightness of the source is such that we can use the sis faint mode for both the medium and high bit rates with the 1-ccd mode .
thus we chose to use the 1-ccd / faint modes throughout .
for the subsequent analysis , we use the bright2 mode data converted from the faint mode . for gis
, we have used the ph mode as usual . the log of _ rosat _ and _ asca _ pointed observations is shown in table [ tab : log ] .
extraction and preparation of spectra and response matrices have been made using tools included in ftools 5.0 and subsequent spectral fittings have been made using xspec 11 .
the source spectra have been extracted from event files which have been screened using the standard screening criteria .
the source - spectrum extraction radii were 6 and 3 arcminutes for the gis and sis respectively .
the background spectra have been extracted from off - source areas from the same detector , excluding 7.5 and and 3.7 arcminutes from the x - ray source for the gis and sis data respectively .
the background subtraction have been made by scaling the background spectra by the areas of extraction , as automatically made by xspec . because of the brightness of the source , the analysis is little affected by the uncertainties related to background scalings .
the two gis and two sis spectra are co - added respectively .
the background spectra and response files have been generated respectively for the summed gis 2 + 3 and sis 0 + 1 spectra .
the _ rosat _ pspc spectra have been extracted using an extraction radius of 1 arcminute .
background spectra have been extracted from an annular region around the source , between 1.25 and 3.75 arcminutes .
one spectrum has been made for each of the three pspc pointed observation ( table [ tab : log ] ) .
we see no evidence of a spectral change among these three datasets ( see below ) .
thus the three spectra for these three sequences have been co - added for the joint analysis with the _ asca _ data .
the _ asca _ and _ rosat _ spectra have been re - binned such that each bin contains at least 25 counts in order to utilize the @xmath19 statistics during the spectral analysis .
the source was bright enough for this re - binning process without sacrificing the resolution .
the channel energy ranges used for the analysis are 0.15 - 2 kev , 0.6 - 10 kev , and 1.0 - 7 kev for the _ rosat _ pspc , _ asca _ gis , and _ asca _ sis , respectively . because of the calibration uncertainties related to the radiation damage , we have ignored _ asca _ sis channels below 1 kev .
firstly , we have verified the basic fitting results of zgw with joint fits to the three pspc spectra , where all model parameters are joined except for the global normalization . as observed by zgw , slight variations of the global normalization have been observed .
this is discussed in sect .
[ sec : lc ] .
when all parameters are separately fit for the three datasets , the best fit parameters are always consistent with the joint ones within errors .
this warrants that we can use the summed spectra for _ pspc _ for further analysis .
next we have made a spectral analysis of _ asca _ only ( gis & sis ) . we have tried to fit the spectra with three models : ( a1 ) a single power - law , ( a2 ) a thermal plasma using the xspec _ mekal _ model , and ( a3 ) a disk blackbody @xcite ( xspec model _ diskbb _ ) , where `` a '' stands for _ asca_. the best - fit parameters are shown in table [ tab : spec ] and fit residuals are shown in fig .
[ fig : ascaresid ] .
the single power - law of photon index @xmath0 gave the best fit without inclusion of any absorption .
the photon indices for separate fits to the gave @xmath20 and @xmath21 for gis and sis , respectively ( @xmath22 fixed to 0 ) .
the agreement verifies the goodness of our background subtraction .
the 90% upper - limit to the column density of the absorbing gas was @xmath23 @xmath24 $ ] .
the thermal plasma model with @xmath1 [ kev ] also gave a good fit with @xmath25 .
we have obtained a heavy element abundance of @xmath26 , which is marginally consistent with the o / h ratio of @xmath27 ( relative to solar ) for this galaxy @xcite .
however note that the metallicity of the source of x - ray emission does not necessarily be equal to that of some average for the hii regions scattered throughout the galaxy and that the hard x - ray mainly measures the iron abundance , which does not necessarily match with that of oxygen .
the multicolor disk blackbody model failed to fit the _ asca _ data as shown in fig .
[ fig : ascaresid](c ) .
thus the disk - blackbody component does not dominate the hard x - ray emission observed by our _ asca _ observation , unlike the cases of , e.g. , source 1 and source b studied by @xcite .
the power - law index of the _ asca _ spectrum is much harder than that of _ rosat _ pspc spectrum ( @xmath28 ; zgw ) , which has response in @xmath29 kev . finally , we have made a joint spectral analysis using all the _ rosat _ pspc and _ asca _ gis / sis data .
we have summed all the three pspc spectra for the joint spectral fits .
we have joined all the model parameters except for the global normalization as before . because the _ rosat _ spectrum , covering a lower energy range , is much softer than that of _ asca _ , we have tried models with hard power - law ( @xmath30 ) with some soft excess and an absorbing column . adding
a softer power - law component did not give a satisfactory fit .
as the soft component , we have tried the thermal plasma and disk blackbody models .
as shown in table [ tab : spec ] , both `` power - law+thermal '' ( j1 ) and `` power - law+disk blackbody '' ( j2 ) models gave reasonable fits .
note that the abundance of the thermal plasma have been fixed to 0.4 ( see above ) for j1 .
when this was a free parameter , the fit could not constrain it . in both j1 and j2 ,
the absorbing column was @xmath31 @xmath32}$ ] , which exceeds the galactic value implied by the 21 cm data of @xmath33 @xcite , indicating an absorbing component within holmberg ii .
the soft excess component , which can either be described as a @xmath3 [ kev ] thermal plasma or a @xmath34 [ kev ] disk blackbody spectrum , has 20 - 30% of the total 0.5 - 2 kev total luminosity .
the _ rosat _ and _ asca _ spectra are shown with folded models and residuals in fig .
[ fig : spec ] for j1 .
the total luminosity of the x - ray source is @xmath35 @xmath6 $ ] in the 0.1 - 10 kev after correcting for absorption ( gis normalization ) .
this corresponds to an eddington luminosity of @xmath36 $ ] . for the fit j2 ,
the gis normalization of the disk blackbody component gives @xmath37 @xmath38 $ ] , where @xmath39 is the inner disk radius , @xmath40 is the viewing angle and the 90% confidence error was searched for by allowing all other variable parameters to vary . assuming @xmath41 , where @xmath42 is the schwarzschild radius
, this corresponds to a blackhole mass of 10 - 25 @xmath43 $ ] , while @xcite pointed out that the @xmath39 estimation based on the disk blackbody within xspec is likely to be underestimated .
further discussion is made in sect .
[ sec : disc ] we have also tried a two - thermal plasma model fit ( j3 ) , where the relative abundances of two thermal plasma have been fixed to equal .
this model gave an acceptable fit but with a hard thermal component ( thermal@xmath44 ) with @xmath45 [ kev ] and a soft thermal component ( thermal@xmath46 ) with @xmath47 [ kev ] .
in this case , the fitted abundance was very low @xmath48 . we did not find a satisfactory fit for the model where the hard component is a thermal plasma and the soft component is a power - law .
one caveat on our joint fitting analysis is a possible cross - calibration problem .
@xcite found that the _ rosat _ pspc spectrum is much softer than that of _ asca _ sis for the simultaneous observations of ngc 5548 ( @xmath49 for _ rosat _ pspc versus @xmath50 for _ asca _ sis ) .
this might have been partially be caused by the soft excess combined with the spectral resolution of pspc or there may be real cross - calibration problems .
the origin of this discrepancy is inconclusive . in our case ,
the difference of @xmath51 between the _ rosat _ pspc and _ asca _ gis spectra is much larger than that @xcite observed for a similar _ asca _ index .
thus there certainly is a soft component over the extrapolation of the hard power - law even in case the discrepancy observed by @xcite was solely caused by calibration problems .
there are wobble - phase dependent systematic errors with the aspect solution for the _ rosat _ , which could lead to a detection of spurious extended component in the hri data @xcite .
this has lead to zgw make a conservative conclusion that they had detected no extension in the hri image .
we have applied a correction for this effect by creating images in 10 wobble - phases , centering each of these images independently , and re - constructing the image using the centering information @xcite .
the original and corrected hri images are shown in fig .
[ fig : wcorr ] .
the faint knot seen in the nw in the original hri image has disappeared in the corrected image .
this reconstruction worked very well down to an hri count rate of @xmath52 [ cts s@xmath53 for a large number of stars , but can be even used for fainter sources @xcite .
the radial profile of the wobble - corrected hri image has been compared with the theoretical hri psf from @xcite and the re - calibrated hri psf in fig .
[ fig : hriprof ] .
the re - calibrated hri psf was derived from wobble - corrected hri images of 21 stars from the rbs @xcite .
the re - calibrate hri psf shows a slightly deviation from the theoretical psf in the radial range between 10 and 30 arcseconds .
the comparison of the radial profile with both psf s indicates an extended component .
subtracting the re - calibrated psf , @xmath54-ray emission is in the extended component and most of the excess comes from a radius of @xmath55 , corresponding to @xmath56 [ pc ] .
the _ asca _ gis flux in the 0.5 - 2 kev band calculated using the spectral model j1 have been compared with the previous _ rosat _ observations . because _
gis is sensitive only above 0.6 kev , while _ rosat _ hri has no spectral resolution and sensitive in 0.2 - 2 kev , we have made separate long - term light curves for the 0.2 - 2 and 0.5 - 2 kev bands .
these are shown in fig .
[ fig : llcurve ] .
the fluxes from _ rosat _ are slightly different from fig .
3 . of zgw probably because of the difference in spectral models . as seen in fig .
[ fig : llcurve ] , there is a 30% decrease in flux between the first ( april 1992 ) and third ( march 1993 ) pspc pointed observations and a factor of @xmath57 increase towards the hri observation ( oct , 1994 ) .
the source brightness came back to a flux close to the pspc observation during the _ asca _ observation in 1999 .
this does not necessarily mean that the source is variable at a time scale of `` years '' , rather it means that there is a variability at a timescale longer than the elapsed time of a single observation ( @xmath58[s ] ) .
we also searched for evidence of time variability in the _ asca _ data .
first , we have defined the combined good time intervals ( gti ) , which consist of the intersection of gti s from all the sis and gis detectors .
we have used channels corresponding to 0.6 - 10 kev for the gis s and 0.7 - 7 kev for the sis s respectively . for each detector ,
binned source and background light curves have been extracted using a 1024 second binsize .
the extraction regions are the same as those of the spectral analysis . for each time bin , only the interval overlapped with the combined gti have been used to calculate the countrate .
the background - subtracted light curves have been co - added to make the final light curve .
there is no significant correlation between the background - subtracted source light curve and the background light curve .
this also warrants goodness of the background subtraction .
the resulting light curve is shown with the best @xmath19-fit constant value in fig .
[ fig : ascalc ] .
no point is more than 2@xmath59 away from the constant value and @xmath60 .
thus there is no evidence for variability at time scales between @xmath61 and @xmath62 seconds in the _ asca _ data .
separate light curves for the hard ( @xmath63 [ kev ] ) and soft ( @xmath64 [ kev ] ) bands did not show any sign of variability either .
this is in contrast to the _ rosat _ light curves reported by zgw , which show a convincing case of gradual decrease of flux by a factor of @xmath65 over @xmath66 seconds during the second pointed pspc observation and less convincing indications of shorter timescale variabilities .
our analysis suggests at least two - components of the x - ray emission . an extended and a point - like component
have been revealed by the spatial analysis .
the variability also provides evidence for at least a compact component .
_ asca_/_rosat _
spectral analysis suggests a non - thermal power - law component plus a warm thermal component or a moderately low temperature disk blackbody component .
the overall spectrum can also be described as a superposition of two thermal components with a very low metallicity . however , the variability observed in this object argues against total thermal origin of the x - ray source .
the x - ray source at the edge of holmberg ii shows unusual characters , even given the large number of off - nuclear x - ray sources in nearby galaxies at similar luminosities . while most of the x - ray sources studied by @xcite reside in the nuclear region , this is at the very edge of a dwarf galaxy . unlike a number of off - nuclear sources studied by @xcite , hard x - ray emission is not dominated by a disk blackbody component with @xmath67 [ kev ] .
if the soft excess component comes from a disk blackbody emission , the implied blackhole mass from the normalization of this component is 10 - 25 @xmath43 $ ] assuming a schwarzschild blackhole ( see sect .
[ sec : spec ] ) .
however , estimating the mass from the temperature @xmath34 [ kev ] using eq .
( 12 ) of @xcite , we obtain @xmath68 @xmath69 $ ] .
this large discrepancy argues against the disk - blackbody interpretation of the soft excess , even though we consider the underestimation of the mass pointed out by @xcite , nearly edge on viewing angle , and spinning of the blackhole .
one of the most surprising results from our analysis is that these two ( or more ) bright x - ray components co - exist in a compact region in one of the numerous hii regions and no other x - ray source with a comparable brightness exists in other part of the galaxy .
all other x - ray sources scattered throughout the galaxy , which are either supernova remnants or x - ray binaries , have luminosities of @xmath70 @xmath71}$ ] @xcite .
thus it is unlikely that the two ( or more ) components have independent origins .
the possible explanations of this x - ray source with observed multiple components are : ( a ) : : a supernova remnant ( snr ) or a composite of snrs ( the @xmath72 0.3 [ kev ] extended thermal component ) feeds an intermediate mass blackhole with gas .
( b ) : : an accretion onto a stellar - mass compact object , e.g. , similar to galactic blackhole x - ray binaries , beaming towards our line of sight , is embedded in an snr or a composite of snrs .
( c ) : : the non - thermal power - law and the soft excess components ( possibly blackbody emission from the accretion disk around an intermediate - mass blackhole ) come from the same compact region while the extended component is caused by scattering / reflection of the central source .
the flat radio spectrum @xcite in the region of this x - ray source shows an existence of a supernova remnant or a composite of supernova remnants .
this supports the possibilities ( a ) or ( b ) .
however , the extended component is too luminous for thermal emission from a single ( or a few ) supernova remnant(s ) ( snr ) , considering its size of @xmath73 pc and its 0.5 - 2 kev luminosity of a few @xmath74 @xmath6 $ ] .
a single snr can only be at this luminosity when it is very young ( @xmath75 years ) @xcite and thus it can not have grown into this scale size nor can feed the intermediate - mass blackhole located 200 [ pc ] away .
thus more than a few supernova remnants are needed to make the extended structure , which also feeds the blackhole causing the variable , non - thermal emission . from our current data , we are not able to distinguish whether the central source represents an accretion onto an intermediate - mass blackhole or a stellar mass compact object ( e.g. x - ray binaries ) with beaming towards our line of sight .
the possibility ( c ) is not likely for the following reasons .
the possibility that electron - scattering by surrounding diffuse gas is the origin of the extended emission has the following difficulty .
scattering gas with a column density of @xmath76 @xmath77 $ ] is required to account for the @xmath78 extended component .
the gas would cause heavy photoelectric absorption unless the scattering diffuse gas is highly ionized @xcite .
if the gas was thermally ionized , the temperature of the gas must be greater than a few million degrees [ k ] in order that the absorption features due to a column density of @xmath79 @xmath77 $ ] is not visible in the _ asca _ spectrum . for demonstration , we assume a uniform sphere of gas of 200 [ pc ] in radius filled with hot ionized gas with a column density of @xmath79 @xmath77 $ ] from the central source .
such a sphere of thermal gas would produce x - ray emission which is several orders of magnitude larger than observed .
photoionization also can not be a valid explanation , because the lower limit of the ionization parameter required to be consistent with the _ asca _ spectrum also requires several orders of magnitude larger luminosity for the ionizing source .
we also consider a picture where the surfaces of optically - thick molecular clouds scattered throughout the region ( giant molecular clouds where star forming activities are embedded ) reflect the radiation from the central source , while our line of sight to the central source is not blocked by any of them .
this picture also has a similar difficulty . in order to attain sufficient reflectivity in the _ rosat _ band ( @xmath29 ) [ kev ]
, the surface of the clouds have to be highly ionized . using the xspec model `` pexriv '' @xcite
, we estimate that an ionization parameter of @xmath80 a few hundred @xmath81 $ ] is required to have the reflectivity of more than 20 - 30% in this band . for a central ionizing source luminosity of @xmath82
a few @xmath7 @xmath83 $ ] , at a distance of @xmath84 [ pc ] , this ionization parameter corresponds to a gas density at the surfaces of the clouds of @xmath85 @xmath86 $ ] .
this is much lower than that of a usual interstellar medium and scattered clouds of this density can not be optically thick .
if the intermediate - mass blackhole picture of this x - ray source is the case , there is a question of its formation .
@xcite discussed possible origins of the off - nucleus `` intermediate - mass '' blackholes presumably responsible for off - nuclear luminous x - ray sources and their preferred model was multiple mergings of stellar remnants ( stellar - mass blackholes and neutron stars ) .
@xcite found a molecular superbubble in the vicinity of a similar off - nuclear x - ray source in m82 , suggesting that an intensive star formation is connected to the formation of an intermediate - mass blackhole .
this may be the case for the compact component of the x - ray source discussed here because of the intense star - forming activity in the region .
however , the question remains as to why there is only one such x - ray source in this galaxy and not at the locations of other numerous hii regions .
probably the x - ray emission is a short - lived phenomenon , where an intermediate mass blackhole ( similar things may exist in many star - forming regions ) is fed by the passing of a shell of dense gas induced by multiple supernova explosions .
chandra ( scheduled as a gt target by murray ) and xmm - newton ( gt target by watson ) observations of this source will certainly improve our knowledge of this peculiar x - ray source .
the chandra observation can make spatially resolved x - ray spectroscopy and reveal whether the non - thermal component corresponds to the point - source and thermal to the extended component or it has a more complicated structure .
the xmm - newton observation would give a better spectral information with a much better statistics and much less cross - calibration uncertainties in the @xmath87 kev region , where multiple components are apparent .
it will also give variability information at short timescales .
we summarize the main conclusions of our analysis : 1 . new _ asca _ and archival _ rosat _ data of the x - ray source in holmberg ii have been analyzed .
2 . the combined _ asca _ and _ rosat _ spectrum shows can be described as a @xmath88 power - law and a soft excess which can either be described as a @xmath3 [ kev ] thermal plasma or @xmath34 disk blackbody .
we argue against the disk blackbody interpretation of the soft excess component based on the inconsistency between the luminosity and the temperature .
the wobble - corrected _ rosat _ hri image indicates @xmath89 of the x - ray emission is extended at a scale of @xmath9010@xmath91 ( @xmath9225% ) .
variability in the _ rosat _ data shows that the point - like component probably comes from accretion onto a compact object .
it is remarkable that this multiple - component x - ray source is the only one with a similar strength among the numerous hii regions in the galaxy with a similar nature .
it is natural to suppose that these two components are related .
possible explanations of the nature of the x - ray source include multiple supernova remnants ( extended component ) feeding the accretion onto an intermediate - mass blackhole .
we can not , however , exclude the possibility that the central source is a beamed radiation from an accretion onto a stellar mass ( usual x - ray binaries ) object .
we argue against a picture where electron scattering or reflection by scattered optically - thick clouds of the central source makes the extended x - ray emission . this research have made use of data and software obtained from the high energy astrophysics archival research center ( heasarc ) located at nasa goddard space flight center .
the authors appreciate the effort of _ asca _ and _ rosat _ teams for having created and operated these superb observatories .
the authors thank the referee , hironori matsumoto for his careful review and useful comments .
anders e. & grevesse n. 1989 , geochimica et cosmochimica acta 53 , 197 colbert , e. j. m. & mushotzky , r. f. 1999 , , 519 , 89 crawford , c. s. , lehmann , i. , fabian , a. c. , bremer , m. n. , & hasinger , g. 1999 , , 308 , 1159 david , l.p.,harnden , f.r . , kearns , k.e . , & zombek , m.v . _ the rosat high resolution imager ( hri ) _
( cambridge : sao ) dickey , lockman 1990 , araa 28 , 215 elvis , m. , fassnacht , c. , wilson , a. s. & briel , u. 1990 , , 361 , 459 fabbiano , g. ara&a , 28 , 215 harris , d. e. , silverman , j. d. , hasinger , g. & lehmann , i. 1998 , , 133 , 431 hodge p. , strobel n.v .
, kennicutt r.c .
1994 , pasp , 106 , 309 hoessel , j. g. , saha , a. & danielson , g. e. 1998 , , 115 , 573 hunter , d. a. & gallagher , j. s. 1985 , , 58 , 533 iwasawa , k. , fabian , a. c. & nandra , k. 1999 , , 307 , 611 kerp j. & walter f. et al .
2000 in " x - ray astronomy 2000 , eds .
r. giacconi , l. stella , s. serio , asp conf .
series , in press ( astro - ph/0012255 ) lehmann i. , et al .
1999 in `` highlights in x - ray astronomy '' , mpe report 272 , ( garching : mpe ) , 209 magdziarz , p. & zdziarski , a. a. 1995 , , 273 , 837 makishima , k. et al .
2000 , , 535 , 632 matsumoto , h. & tsuru , t. g. 1999 , , 51 , 321 matsumoto , h. et al .
2001 , apjl 547 , l25 matsushita , s. , kawabe , r. , matsumoto , h. , tsuru , t. g. , kohno , k. , morita , k. , okumura , s. k. & vila - vilar , b. 2000 , , 545 , l107 mitsuda , k. et al .
1984 , , 36 , 741 morrison , r. & mccammon , d. 1983 , , 270 , 119 morse , j. a. 1994 , , 106 , 675 ptak a. & griffiths r.e 1999 , , 517 , l85 puche d. , westpfahl d.j , brinks e. 1992 , , 103 , 1841 read , a. m. , ponman , t. j. & strickland , d. k. 1997 , , 286 , 626 schlegel , e. m. 1994 , , 424 , l99 schlegel e. m. et al .
1999 , aj 118 , 2689 schwope a. d. , hasinger g. , lehmann i. et al .
2000 , astron .
nachr . , 321 , 1 taniguchi , y. , shioya , y. , tsuru , t. g. & ikeuchi , s. 2000 , , 52 , 533 tongue , t. d. & westpfahl , d. j. 1995 , , 109 , 2462 voges , w. , et al .
1999 , , 349 , 389 wilson , a. s. , elvis , m. , lawrence , a. & bland - hawthorn , j. 1992 , , 391 , l75 zezas a.l . , georgantopoulos i. , ward m.j . 1999 , , 308 , 302 ( zgw ) pspc & rp600140n00 & 14-apr-1992 & 7.3 + & rp600431n00 & 29-oct-1992 & 11.6 + & rp600431a00 & 14-mar-1993 & 3.7 + _ rosat _ hri & rh600745n00 & 17-oct-1994 & 7.8 + _ asca _
gis(ph ) & ad77075000 & 21-oct-1999 & 16.5 + _ asca _ sis ( 1ccd / faint ) & & & 16.5(s0)/16.3(s1 ) + * pl * & @xmath93 ; @xmath94 ; @xmath95 ; + ( a1 ) & @xmath96 ; @xmath97 + + * thermal * & @xmath93 ; @xmath98 ; @xmath99 ; + ( a2 ) & @xmath100 ; @xmath101 ; @xmath102 + + disk bb & @xmath93 ; @xmath94 ; @xmath103 ; + ( a3 ) & @xmath104 ; @xmath105 ; + + * ( pl+thermal ) @xmath106 abs . * & @xmath93 ; @xmath94 ; @xmath107 ; @xmath108 ; + ( j1 ) & @xmath109 ; @xmath110;@xmath111 ; @xmath112 ; + & @xmath113 ; @xmath114 + + * ( pl+disk bb)@xmath106 abs . * & @xmath93 ; @xmath115 ; @xmath116 ; @xmath117 ; + ( j2 ) & @xmath109 ; @xmath118 ; @xmath119 + & @xmath120 ; @xmath121 + + * ( thermal@xmath44+thermal@xmath46)@xmath106 abs . * & @xmath93 ; @xmath94 ; @xmath122 ; @xmath123 ; + ( j3 ) & @xmath124 ; @xmath125 ; @xmath126 ; + & @xmath127 ; @xmath128 ; @xmath129 + | we report the results of the analysis of new _ asca _ observations and archival _ rosat _ data of the compact luminous x - ray source found at the edge of the nearby star - forming dwarf galaxy ( ) in the group . we have found a number of new features in the x - ray properties of this source . our new _ asca _
spectrum revealed that the x - ray emission extends to the hard band and can be best described by a power - law with a photon spectral index @xmath0 while a @xmath1 [ kev ] thermal plasma with a low abundance ( @xmath2 ) is also acceptable .
the _ asca _
spectrum does not fit with a multi - color disk blackbody , unlike some off - nucleus x - ray sources with similar luminosities .
the joint _
asca_-_rosat _
spectrum suggests two components to the spectrum : the hard power - law component and a warm thermal plasma ( @xmath3 [ kev ] ) . an additional absorption over that of our galaxy
is required .
the wobble correction of the _ rosat _ hri image has clearly unveiled the existence of an extended component which amounts to @xmath4 of the total x - ray emission .
these observations indicate that there are more than one component in the x - ray emission .
the properties of the point - like component is indicative of an accretion onto an intermediate mass blackhole , unless a beaming is taking place .
we argue that the extended component does not come from electron scattering and/or reflection by scattered optically - thick clouds of the central radiation .
possible explanations of this x - ray source include multiple supernova remnants feeding an intermediate - mass blackhole . |
the main results of this paper are an explicit construction of a bicategory whose objects are entwinings and of a homomorphism of bicategories from that bicategory to r. street s bicategory of corings ( a straightfoward analogue of the 2-category of ( co)monads from @xcite ) .
the results in this manuscript hold in one of the following two generalities . in the first case we consider entwinings @xmath0 between @xmath1-algebras @xmath2 and @xmath1-coalgebras @xmath3 over ( variable ) commutative unital ring @xmath1 ( @xcite ) ; in the latter case the entwinings are between @xmath1-rings @xmath2 ( monoids in @xmath4 ) and @xmath1-corings @xmath3 ( comonoids in @xmath4 ) over ( variable ) not necessarily commutative ring @xmath1 .
i will talk about `` algebras '' and `` coalgebras '' over commutative rings but will be careful about the sides for the bimodules even when over the ground ring @xmath1 , @xmath5 etc .
, so that all our calculations seem identical in both cases . @xmath6 does not need to be invertible for our purposes .
one can go to the third generality , working with the internal entwining structures in monoidal categories ( cf.@xcite ) .
in fact , in familiar cases of our interest , the monoidal categories have obvious coherences for associativity so we do not write them in our statements and proofs . on the other hand
if the coherences are indeed algebraically nontrivial , then the statements here are more complicated and somewhat more interesting .
when the preprint version 1 of this article has been posted , g. bhm has kindly called my attention to the following argument : distributive laws ( @xcite ) are just monads in a 2-category of monads in the sense of formal monad theory ( @xcite ) of r. street , and in particular they themselves make a 2-category ; the analogue can be easily written out for mixed distributive laws betwen a monad and a comonad ; it is not published in detail , but it is widely known among the experts that the formal monad theory can be extended to bicategorical setup , instead of strict 2-categories ; finally entwinings are _ mixed _ distributive laws in the setup of the bicategory of rings and bimodules ; regarding that in a bicategory we can do 2 dualizations ( inverting 1-cells and 2-cells ) there are thus 4 natural bicategories of entwinings .
our construction is explicit and from scratch and does not use this chain of constructions and translations of data ( explicit descriptions are its merit but also its conceptual deficiency ) .
the bicategory introduced here is also analogous to the 2-category of distributive laws between actions of a fixed monoidal category @xmath7 on variable category @xmath8 and monads in @xmath8 ( such distributive laws were studied in our earlier preprint @xcite and the 2-category they form in @xcite ) ; such distributive laws can be identified with @xmath7-equivariant monads , that is , monads in the 2-category @xmath9 of @xmath7-actegories , colax @xmath7-equivariant functors and their natural transformations .
finally , one can study the distributive laws between actions of two different variable categories on their common target category .
if we restrict in the latter case to the invertible distributive laws , then we call such data biactegories ( not a typo ! ) ; in work in progress @xcite we introduce and study a tensor product of biactegories using certain pseudcoequalizer ( the induction pseudofunctor for actegories can be viewed as a special case of that pseudocoequalizer ) with motivation in associating 2-vector bundles to principal bundles with structure 2-group . finally , biactegories make a tricategory @xmath10 which is a categorification of the bicategory of bimodules @xmath11 .
of @xmath12 are the entwinings @xmath13 . here
ring @xmath1 may vary ! @xmath14 satisfies the usual two pentagons and two triangles ( @xcite ) which we do not write here ; these data are implicit : multiplication @xmath15 , unit @xmath16 , comultiplication @xmath17 , counit @xmath18 . of @xmath12
are the triples @xmath19 , where @xmath20 is a @xmath21-bimodule , and @xmath22 , and @xmath23 are maps of @xmath21-bimodules , for which the following 5 diagrams commute : the hexagon : @xmath24^{\chi\otimes m}\ar[r]^{d\otimes\alpha } & d\ots m\otr a\ar[r]^{\gamma\otimes a } & m\otr c\otr a\ar[d]^{m\otimes\psi } \\ b\ots d\ots m \ar[r]^{b\otimes\gamma } & b\ots m\otr c \ar[r]^{\alpha\otimes c } & m\otr a\otr c } \ ] ] the pentagon for the @xmath21-bimodule map @xmath25 : @xmath26^{b\otimes\alpha}\ar[d]^{\mu^n\otimes m } & b\ots m\otr a\ar[r]^{\alpha\otimes a } & m\otr a\otr a\ar[d]^{m\otimes\mu^a}\\ b\ots m\ar[rr]^\alpha & & m\otr a } \ ] ] the pentagon for the @xmath21-bimodule map @xmath27 : @xmath28^\gamma\ar[d]_{\delta^d\otimes m } & & m\otr c\ar[d]^{m\otimes\delta^c}\\ d\ots d\ots m\ar[r]^{d\otimes\gamma } & d\ots m\otr c\ar[r]^{\gamma\otimes c } & m\otr c\otr c } \ ] ] and the two triangles : @xmath29_{\eta^b\otimes m}\ar[dr]^{m\otimes\eta^a } & & d\ots m\ar[rr]^\gamma\ar[dr]_{\epsilon^d } & & m\otr c\ar[dl]^{\epsilon^c}\\ b\ots m\ar[rr]^\alpha & & m\otr a&&m } \ ] ] notice that in these diagrams we did not bother inserting the brackets and associativity isomorphisms inherited from the bicategory of bimodules .
a pedantic reader will easily correct this .
@xmath30 are the @xmath21-bimodule maps @xmath31 such that the following two squares commute : @xmath32^\alpha\ar[d]^{b\otimes\theta}&m\otr a\ar[d]^{\theta\otimes a } & & d\ots m\ar[r]^\gamma\ar[d]^{d\otimes\theta}&m\otr c\ar[d]^{\theta\otimes c } \\ b\ots n\ar[r]^\beta & n\otr a&&d\ots n\ar[r]^\delta & n\otr c & & } \ ] ] given a diagram of morphisms of entwinings @xmath33^{(m,\alpha,\gamma ) } & ( s , b , d,\chi)\ar[r]^{(p,\sigma,\tau ) } & ( u , e , g,\lambda)\ar[r]^{(q,\rho,\nu ) } & ( v , f , h,\xi ) } \ ] ] define the composition ( up to coherences again ) @xmath34 if we should insert the coherences from the underlying bicategory of bimodules , instead of @xmath35 write the composition of 5 maps @xmath36 where the isomorphisms @xmath37 are the components of the associtivity coherence of @xmath11 . the coherences from the underlying bicategory @xmath11 play the role of coherence in @xmath12 as well , which will be denoted by the same letter @xmath38 : @xmath39 in other words , @xmath40 ( in particular , the pentagon for the coherence follows from the pentagon for @xmath38 in @xmath11 ) .
we need to check that this definition is meaningful .
the components of @xmath38 should be 2-cells in @xmath11 , which are in fact 2-cells in @xmath12 , i.e. two naturality squares commute , first of which is @xmath41^{``\rho(\sigma\alpha ) '' } \ar[d]_{f\otu a_{q , p , m } } & & ( q\otu(p\ots m))\otr a\ar[d]^{a_{q , p , m}\otr a } \\ f\otv((q\otu p)\ots m))\ar[rr]^{``(\rho\sigma)\alpha '' } & & ( ( q\otu p)\ots m))\otr a , } \ ] ] where the horizontal arrows just symbollically denoted @xmath42 correspond to the analogue of @xmath25 for the triple composition , which do involve @xmath43 but also various coherences .
we will show that this square commutes , while another square similar to ( [ eq : triplesq ] ) , but for the analogue of @xmath27 is commutative as well , with similar proof left to the reader . keeping track of coherences ,
the analogue of @xmath25 for a single composition is a composition of 5 maps , thus if for one of them we insert another composition of 5 maps , we have the composition of 9 maps . in the diagram showing that ( [ eq : triplesq ] ) commutes , and where maps @xmath44 are explicit compositions of 9 maps each , we will skip tensor product signs , and enclose the corners of ( [ eq : triplesq ] ) in boxes for emphasis : @xmath45^{\,\,\,\,\,\,a_{f , q , pm}}\ar[d]^{fa_{q , p , m } } & ( fq)(pm)\ar[r]^{\rho ( pm)}\ar[d]^{a_{fq , p , m } } & ( qe)(pm)\ar[r]^{a_{q , e , pm}}\ar[d]^{a_{qe , p , m } } & q(e(pm))\ar[r]^{q a_{e , p , m } } & q((ep)m)\ar[dd]_{q(\sigma m)}\ar[dl]_{a_{q , ep , m } } \\
\fbox{$f((qp)m)$}\ar[d]^{a_{f , qp , m } } & ( ( fq)p)m\ar[r]^{(\rho p)m } & ( ( qe)p)m\ar[r]^{a_{q , e , p}m } & ( q(ep))m\ar[d]_{(q\sigma)m } & \\
( f(qp))m\ar[ru]_{a_{f , q , p}m } & & & ( q(pb))m\ar[d]_{a_{q , p , bm } } & q((pb)m)\ar[dd]_{q a_{p , b , m}}\ar[l]_{a_{q , pb , m } } \\ \fbox{$((qp)m)a$ } & ( qp)(ma)\ar[l]_{\,\,\,\,a_{qp , m , a } } & ( qp)(bm)\ar[l]_{(qp)\alpha } & ( ( qp)b)m\ar[l]_{a_{qp , b , m } } & \\ & \fbox{$(q(pm))a$}\ar[lu]^{a_{q , p , m}a } & q((pm)a)\ar[l]^{\,\,\,\,a_{q , pm , a } } & q(p(ma))\ar[l]^{q a_{p , m , a}}\ar[ull]^{a_{q , p , ma } } & q(p(bm))\ar[llu]_{a_{q , p , bm}}\ar[l]^{q(p\sigma ) } } \ ] ] the diagram is commutative as it splits into 3 naturality squares and 4 associativity pentagons . a similar diagram may be written for @xmath27 instead of @xmath25 .
the composition formula ( [ eq : compmorentw ] ) indeed defines a morphism of entwinings . with skipping tensor product signs in a big diagram
i just draw the commutative diagram for showing the hexagon for the composition .
m)}\ar[rd]_{g\sigma m } \ar[ddd]_{\lambda pm } & & gpma\ar[rr]^{(p\gamma \circ \tau m)a}\ar[rd]^{\tau ma } & & pmca\ar[ddd]^{pm\psi } \\ & gpbm\ar[ur]^{gp\alpha}\ar[r]_{\tau bm } & pdbm\ar[r]_{pd\alpha}\ar[d]^{p\chi m } & pdma\ar[ur]_{p\gamma a } & \\ & epdm\ar[dr]_{ep\gamma}\ar[r]^{\sigma dm } & pbdm\ar[r]^{pd\gamma } & pbmc\ar[dr]^{p\alpha c}&\\ egpm\ar[ur]^{e\tau m}\ar[rr]_{e(p\gamma\circ \tau m ) } & & epmc\ar[ur]_{\sigma mc}\ar[rr]_{(p\alpha\circ\sigma m)c } & & pmac } \ ] ] horizontal and vertical composition of 2-cells is simply given by the horizontal and vertical composition of the underlying morphisms of bimodules .
one checks that the composed 2-cells are indeed 2-cells , i.e. satisfy the two squares ( [ eq : twosquares ] ) .
in this article i work with the old street s version @xmath47 of the bicategory of corings .
it is summarized in the end of the article @xcite but it is just a variant of a construction in @xcite ) .
there is a different ( variant of the ) bicategory of corings also defined in @xcite ( could be obtained using certain dualization if compared to @xmath47 ) and studied in more detail .. 06 in let us now define the morphism of bicategories composed coring @xmath48 : standard `` composed comonad '' formula well known in coring setup : @xmath13 gives rise to the composed @xmath2-coring @xmath49 where @xmath50 is the composition @xmath51 and similarly for @xmath52 : @xmath53 .
let us define @xmath54 .
the underlying module is @xmath55 , it is a left @xmath56-module via action @xmath57 and a right @xmath2-module via action @xmath58 .
it is easy to check that the two actions are compatible , making @xmath59 a @xmath60-bimodule : @xmath61_{b\otimes m\otimes \mu^a } \ar[rr]^{\alpha\otimes a\otimes a } & & m\otimes a\otimes a\otimes a\ar[d]_{m\otimes a\otimes\mu^a } \ar[rr]^{m\otimes\mu^a\otimes a } & & m\otr a\otr a\ar[d]_{m\otimes\mu^a } \\ b\ots m\otr a\ar[rr]^{\alpha\otimes a } & & m\otr a\otr a\ar[rr]^{m\otimes\mu^a } & & m\otr a } \ ] ] notice that the definition of the @xmath60-module structure on @xmath59 implies that the diagram @xmath62^{\alpha\otimes a}\ar[d]^{\pr } & & m\otr a\otr a\ar[d]^{\pr } \\
b\otb(m\otr a)\ar[r]^{\cong } & m\otr a\ar[r]^\cong&(m\otr a)\ota a } \ ] ] commutes . :
the triple @xmath63 maps to the pair @xmath64 where @xmath55 is a @xmath60-bimodule as above and for the map @xmath65 , one first defines an auxiliary map @xmath66 as the composition @xmath67 one checks that @xmath68 is a map of @xmath60-bimodules .
the fact that @xmath68 is a map of left @xmath56-modules essentially boils down to the pentagon for map @xmath69 .
i will skip the tensor signs in drawing the commutative diagram showing this : @xmath70^{bb\gamma a } \ar[dd]^{\mu^b dma } & bbmca\ar[r]^{b\gamma ca}\ar[dd]^{\mu^b mca } & bmaca\ar[r]^{bma\psi}\ar[d]_{\alpha aca } & bmaac\ar[d]^{\alpha aac } \\ & & maaca\ar[d]_{m\mu^a ca}\ar[r]^{maa\psi } & maaac\ar[d]^{m\mu^a ac } \\ bdma\ar[r]_{b\gamma a } & bmca \ar[r]_{\alpha ca } & maca\ar[r]_{ma\psi } & maac } \ ] ] similarly , the fact that @xmath68 is a map of right @xmath2-modules similarly essentially boils to the pentagon for entwining @xmath6 : @xmath71^{b\gamma aa}\ar[dd]^{bdm\mu^a } & bmcaa\ar[r]^{\alpha caa}\ar[dd]^{bmc\mu^a } & macaa\ar[r]^{ma\psi a}\ar[dd]^{mac\mu^a } & maaca\ar[d]^{maa\psi } \\ & & & maaac\ar[d]^{ma\mu^a c}\\ bdma\ar[r]^{b\gamma a } & bmca\ar[r]^{\alpha ca } & maca\ar[r]^{ma\psi } & maac } \ ] ] let @xmath72 and @xmath73 be the canonical projections .
now once we defined @xmath74 we prove that there is a unique map @xmath75 such that if @xmath76 and then define @xmath77 .
showing this is a longer naturality calculation , involving the hexagon for the map @xmath78 , the pentagon for @xmath25 , the pentagon for the entwining @xmath6 and 4 or 6 naturality squares ( depending on the way of defining @xmath68 ) .
indeed , start with @xmath79 and acts with middle @xmath56 either to the first or the second tensored pair . after that
apply @xmath68 .
thus , omiting the tensor ( over the ground rings ) signs , the composition @xmath80 equals the composition @xmath81 ( warning : if one parallely truncates the tail of the two chains of maps , one does not get an identity ! ) @xmath82^{bd\alpha a}\ar[d]^{b\chi ma } & bdmaa\ar[r]^{b\gamma aa } & bmcaa\ar[r]^{bmc\mu^a}\ar[dr]^{\alpha caa}\ar[d]_{bm\psi a } & bmca\ar[r]^{bm\psi}\ar[dr]^{\alpha ca } & bmac\ar@/^2pc/[ddd]_{\alpha ac } \\
bbdma\ar[r]^{bb\gamma a}\ar[dd]^{\mu^b dma } & bbmca\ar[r]^{b\alpha ca}\ar[dd]^{\mu^b mca } & bmaca\ar[d]_{\alpha aca } & macaa\ar[r]_{mac\mu^a}\ar[ld]^{ma\psi a } & maca\ar[dd]_{ma\psi } \\ & & maaca\ar[d]_{m\mu^a ca}\ar[r]_{maa\psi}&maaac\ar[dr]_{m\mu^a ac } & \\ bdma\ar[r]_{b\gamma a } & bmca\ar[r]_{\alpha ca } & maca\ar[rr]_{ma\psi } & & maac } \ ] ] in the situation @xmath83^{(m,\alpha,\gamma ) } & & ( s , b , d,\chi)\ar[rr]^{(p,\sigma,\tau ) } & & ( u , e , g,\lambda ) } \ ] ] consider the diagram of corings @xmath84^{(\mm,\zeta^m ) } & & \dd\ar[rr]^{(\pp,\zeta^p ) } & & \ee } \ ] ] where the corings are @xmath85 , @xmath86 , @xmath87 , ( over @xmath88 and @xmath89 , respectively ) and @xmath90 and @xmath91 are the corresponding bimodules . _
_ in fact , we shall prove the commutativity of a diagram in which the upper row is the representative of the map @xmath65 at the level of the tensor products over @xmath5 and @xmath1 , but in the lower row we indeed have the equivalence classes .
the diagrams are a bit more complicated so we will in addition to skipping the tensor products over @xmath5 and @xmath1 , also abbreviate the signs for the tensor products over @xmath2 and @xmath56 by a dot @xmath92 ( the modules involved gurantee that the meaningful choice between @xmath93 and @xmath94 is unique ) . for example , @xmath95 means @xmath96 .
one also needs to be careful `` cancelling '' @xmath56 and @xmath2 in tensor products over @xmath56 and @xmath2 respectively .
carefully distinguish the following two maps ( and their analogues ) .
the first is the natural projection from the tensor product over @xmath5 to the tensor product over @xmath97 , say @xmath98 ( also sometimes shortly denoted @xmath99 ) and another is inserting the unit over @xmath56 , say the map @xmath100 . thus diagram ( [ eq : streetonecell ] ) may be expanded to @xmath101^{b\gamma a}\ar[d]^{b\delta^d ma } & & bmca\ar[r]^{bm\psi}\ar[d]_{bm\delta^c a } & bmac\ar[r]^{\alpha ac}\ar[rd]^{bma\delta^c } & maac\ar@/^2pc/[dd]^{ma\delta^c } \\
bddma\ar[r]^{bd\gamma a}\ar[d]^{bd\cdot \eta^b dma } & bdmca\ar[r]^{b\gamma ca}\ar[d]^{bd\cdot \eta^b mca}\ar[dr]^{bdm\psi } & bmcca\ar[r]^{bmc\psi } & bmcac\ar[r]^{bmc\psi } & bmacc\ar[d]_{\alpha ac } & \\
bd\cdot bdma\ar[r]^{bd\cdot b\gamma a}\ar[dd]^\pr & bd\cdot bmca\ar[d]_{bd\cdot bm\psi } & bdmac\ar[d]^{bd\cdot m\eta^a ac}\ar[dl]^{bd\cdot\eta^b mac } \ar[ru]^{b\gamma ac}\ar@/^2pc/[rd]^{\pr } & & maacc\ar[d]_{\pr } & \\ & bd\cdot bmac\ar[r]_{bd\cdot\alpha ac } & bd\cdot maac\ar[d]_\pr & bd\cdot mac\ar[r]^{\zeta c}\ar[dl]^\pr & ma\cdot acc\ar[d]_\pr & \\ bd\cdot bd\cdot ma\ar[rr]^{bd\cdot\zeta } & & bd\cdot ma\cdot ac\ar[rr]^{\zeta\cdot ac } & & ma\cdot ac\cdot ac & } \ ] ] where one can directly observe the commutativity of all smallest circuits and hence of the entire diagram .
in addition to the pentagon ( [ eq : streetonecell ] ) , we need to check the compatibility of the map @xmath65 with the counits of the corings involved : @xmath102 .
this follows by the calculation for the representatives , namely the diagram @xmath103^{b\otimes\gamma\otimes a } \ar[d]_{b\otimes\epsilon^d\otimes m\otimes a } & b\ots m\otr c\otr a\ar[r]^{\alpha\otimes c\otimes a } \ar[dl]_{b\otimes m\otimes\epsilon^c\otimes a } & m\otr a\otr c\otr a\ar[d]^{m\otimes a\otimes \psi } \ar[dl]_{m\otimes a\otimes\epsilon^c\otimes a } & \\
b\ots m\otr a\ar[r]^{\alpha\otimes m\otimes a}\ar[d]^\pr & m\otr a\otr a\ar[dr]^\pr & m\otr a\otr a\otr c \ar[l]_{m\otimes a\otimes a\otimes\epsilon^c } \\ b\otb ( m\otr a)\ar[r]^{\cong } & m\otr a&(m\otr a)\ota a\ar[l]^\cong & } \ ] ] commutes .
this finishes the proof . :
if @xmath104 then @xmath105 .
one sees easily that @xmath106 is indeed a map of @xmath60-bimodules .
we will just draw the diagram for the left @xmath56-equivariance : @xmath107_{b\otimes\theta\otimes a}\ar[rr]^{\alpha\otimes a } & & m\otr a\otr a\ar[d]_{\theta\otimes a\otimes a}\ar[rr]^{m\otimes\mu^a } & & m\otr a\ar[d]^{\theta\otimes a}\\ b\ots n\otr a\ar[rr]^{\alpha\otimes a } & & n\otr a \otr a\ar[rr]^{n\otimes\mu^a } & & n\otr a } \ ] ] for a fixed domain and codomain 1-cells , this tautological map is injective , but not the surjective map , because the property that the composition is a 2-cell in @xmath47 is weaker than the property that @xmath106 is a 2-cell in @xmath12 . in @xmath47 case just the external square in @xmath103^{b\otimes\gamma\otimes a }
\ar[d]_{b\otimes d\otimes\theta\otimes a } & b\ots m\otr c\otr a\ar[r]^{\alpha\otimes\psi } \ar[d]_{b\otimes \theta\otimes\psi } & m\otr a\otr a\otr c\ar[d]^{\theta\otimes a\otimes a\otimes c}\\ b\ots d\ots n\otr a\ar[r]_{b\otimes\delta\otimes a } & b\ots n\otr c\otr a\ar[r]_{\gamma\otimes\psi } & n\otr a\otr a\otr c } \ ] ] commutes
. the commutativity of the right - hand square is implied form one of the squares in the axioms for @xmath108 ( @xmath108 vs. @xmath27 ) , while the left - hand square is actually the pasting of another such square ( @xmath108 vs. @xmath25 ) and of a naturality square for the tensoring with @xmath6 .
we need to check the functoriality .
thus consider again the diagram ( [ eq : diagmorentw ] ) of morphisms of entwinings and the compositions ( [ eq : compmorentw ] ) .
@xmath109 versus @xmath110 up to coherences ( some of which we already skipped ) , the two answers should agree ; the additional coherences make the functoriality true up to invertible 2-cell ( pseudofunctoriality ) . if we look at the underlying module , this is obvious : @xmath111 , and the agreement for @xmath65-s is @xmath112^{\zeta^{p\otimes m}}\ar[d]^\cong & & & ( \pp\otb\mm)\ota\cc\ar[d]^\cong\\ ( \ee\ote\pp)\otb\mm\ar[r]^{\zeta^p\otb\mm } & ( \pp\otb\dd)\otb\mm\ar[r]^\cong & \pp\otb(\dd\otb\mm)\ar[r]^{\pp\otb\zeta^m } & \pp\otb(\mm\ota\cc ) } \ ] ] what expands into the diagram @xmath113^{e\tau bma}\ar[d]_\pr & epdbma\ar[r]^{\sigma\chi ma}\ar[d]^{\pr } & pbbdma\ar[r]^{pbb\gamma ma}\ar[d]^\pr & pbbmca\ar[r]^{pb\alpha\psi}\ar[d]^\pr & pbmaac\ar[d]^\pr \\ egpb.ma\ar[dd]_\cong\ar[r]^{e\tau b.ma } & epdb.ma\ar[d]^{\sigma db.ma}\ar@/_4pc/[dd]_{\cong } & pb.bdma\ar[d]^\cong\ar[r]^{pb.b\gamma a } & pb.bmca\ar[dd]^\cong\ar[r]^{pb\alpha ca}&pb.maac\ar[d]^\cong \\ & pbdb.ma\ar[r]^\cong&pbdma\ar[dr]^{pb\gamma a}&&pmaac
\\ egpma\ar[r]^{e\tau ma}&epdma\ar[ru]^{\sigma dma}\ar[r]^{ep\gamma a } & epmca\ar[r]^{\sigma mca}&pbmca\ar[r]^{p\alpha ca}&pmaca\ar[u]_{pma\gamma } } \ ] ] the only detail requiring explanation is the commutativity of the hexagon below the second arrow in the upper row . to show that it commutes one
needs to expand it by inserting @xmath114 in the middle of the hexagon with a projection to @xmath115 and map @xmath116 to @xmath117 and also map @xmath118 from the vertex @xmath119 .
then the lower right corner of the split hexagon commutes essentially by the compatibility of the unit @xmath120 with @xmath121 .
thus we obtained
the operation of producing the lifting monad may be also promoted to a canonical morphism of bicategories from the bicategory of entwinings to the 2-category of monads which act in the categories of right comodules over variable coalgebras over variable rings .
this is in a complete analogy to one of the results in my earlier article @xcite , so i will not bother writing details here .
i thank t. brzeziski for the encouragement and pointing out an important error in an early version .
the article has been written at irb , zagreb ( partial support of croatia / mses national projects 037 - 0372794 - 2807 , 098 - 0000000 - 2865 ) and max planck institute for mathematics in bonn whom i thank for excellent working condition .
my trips to bonn have been partly funded by daad / mses bilateral project . | we define a bicategory in which the 0-cells are the entwinings over variable rings .
the 1-cells are triples of a bimodule and two maps of bimodules which satisfy an additional hexagon , two pentagons and two ( co)unit triangles ; and the 2-cells are the maps of bimodules satisfying two simple compatibilities .
the operation of getting the `` composed coring '' from a given entwining , is promoted here to a canonical morphism of bicategories from a bicategory of entwinings to the street s bicategory of corings .
entwining , coring , distributive law , bicategory , bimodule msc classification : 16w30 , 18c15 , 18d05 |
after the discovery of super - massive black holes ( smbhs ) at the center of every massive galaxy in the local universe ( e.g. , @xcite ) , the issue of how these smbhs formed and evolved over cosmic history has become one of the major unanswered questions in observational cosmology .
luminosities of agns reflect the mass accretion rates of their smbhs , therefore the luminosity function of agns and its cosmological evolution reflects the growth history of smbhs through accretion .
hard x - ray selection is an effective selection method for agns , it can detect obscured agns without serious bias , except for compton - thick agns , and can efficiently sample agns without contamination by a large number of star - forming galaxies . the hard x - ray luminosity function of agns and its evolution as a function of redshift have been examined with combining x - ray selected agn samples with various survey depths and areas @xcite .
based on the cosmological evolution of the hard x - ray luminosity function , the average accretion growth curves of smbhs are inferred @xcite . in order to construct a large sample of x - ray - selected agns , we have conducted deep multi - wavelength imaging observations covering the fuv to mir and deep optical and nir multi - object spectroscopic observations of the x - ray sources found in the subaru _ xmm - newton _ deep survey ( sxds ) @xcite .
the x - ray images obtained with _ xmm - newton _ cover a 1.3 deg@xmath6 area down to @xmath7 erg s@xmath8 @xmath9 in the 0.52 kev band and @xmath10 erg s@xmath8 @xmath9 in the 210 kev band @xcite .
the @xmath11 relation of the x - ray sources has a knee at a hard x - ray flux of @xmath12 erg s@xmath8 @xmath9 and the contribution to the cosmic x - ray background is maximum in this flux range @xcite .
several hard x - ray surveys covering @xmath13deg@xmath6 area in this flux range have been conducted @xcite . among the x - ray surveys with similar depth and area ,
sxds is notable in having deep multi - wavelength data in the wavelength range from far - uv to mid - ir . in this paper
, we describe the x - ray sample in section 2 , and the multi - wavelength imaging data in section 3 .
the 3 @xmath14 detection limits of the imaging data in the @xmath2- , @xmath15- , and 3.6 bands reach 27.4 mag , 25.3 mag , and 24.5 mag in the ab magnitude system , respectively .
they are expected to be deep enough to detect counterparts of almost all of the x - ray sources at the flux limit : 90% of the x - ray sources found in a wide area _
survey at a flux limit of @xmath7 erg s@xmath8 @xmath9 in the 0.52.0 kev band have a counterpart brighter than 26.4 mag ( @xmath2-band ) , 23.6 mag ( @xmath15-band ) , and 22.8 mag ( 3.6-band ) @xcite .
the positional uncertainties of the _ xmm - newton _ x - ray sources are larger than those of _
chandra _ sources , and they are not small enough to securely identify the faintest optical counterparts ( @xcite ; @xcite ) . in order to quantitatively select the counterparts , we apply the likelihood ratio method @xcite .
we apply the method not only to the deep optical source catalog but also to the deep 3.6 , nuv , and 24 source catalogs to improve the reliability of the counterpart selection .
the identification procedure with the multi - wavelength source catalogs is described in section 4 . in order to reveal the nature of the x - ray sources ,
we have been conducting a large campaign of spectroscopic observations in the nir as well as in the optical wavelength range . because obscured agns at @xmath16 can be significant contributor to the x - ray source population , nir spectroscopy is effective in revealing their nature by detecting strong rest - frame optical emission lines .
the spectroscopic follow - up observations are described in section 5 . even with the deep optical and nir spectroscopic observations , a significant fraction of the x - ray sources are fainter than the spectroscopic limit . in order to overcome
the limitation of the spectroscopic identification , photometric redshift estimation with a deep imaging data set covering wide wavelength range is important @xcite .
the multiwavelength photometry and photometric redshift evaluation are described in section 6 . in section 7
, we discuss the x - ray properties and host galaxy stellar mass of the x - ray selected agns .
utilising the multi - wavelength photometry , we examine the scatter of the seds of the x - ray agns , and compare the scatter with that of non - x - ray galaxies with similar redshifts and stellar masses . throughout the paper , cosmological parameters h@xmath17 = 70 km s@xmath8
mpc@xmath8 , @xmath18 , and @xmath19 are used .
the magnitudes are based on ab magnitude system @xcite .
using the large sample of x - ray selected agns , @xcite derive the comoving space density and obscured fraction of agns at @xmath20 . furthermore , @xcite examine the cosmological evolution of the x - ray agn luminosity function including our sample .
the black hole mass and eddington ratio distribution functions of the x - ray selected agns at @xmath21 are discussed in @xcite , using the optical and nir spectra of the x - ray selected agns .
the sxds field is centered on and .
the field is selected to be observable with 8 - 10 m class optical telescopes in both of the northern and the southern hemispheres .
the field was observed with _
xmm - newton _ with one central @xmath22 diameter field with a 100ks exposure and six flanking fields with 50ks exposure time each @xcite .
these seven pointings cover a 1.3 deg@xmath6 field .
details of the observations and data reduction procedures are described in @xcite . in total ,
866 and 645 sources are detected above likelihood of 7 in the 0.52 kev ( soft ) and 210 kev ( hard ) bands , respectively , through source detection utilising a maximum likelihood fit method .
the spatial distribution of the sources is shown in figure [ fig : sxds_fov ] ( small dots ) .
a likelihood ( @xmath23 ) of 7 corresponds to the probability of being a random background noise , @xmath24 of @xmath25 ( @xmath26 ) .
the 0.52 kev and 210 kev band x - ray flux distributions of the x - ray sources are shown in figure [ fig : flux_dist ] .
the 0.52 kev ( 210 kev ) flux is calculated by multiplying the 0.52 kev ( 210 kev ) count rate with 0.167 ( 0.838)@xmath27 [ erg s@xmath8 @xmath9]/[cts ksec@xmath8 ] @xcite .
these are the conversion factors for a power - law spectrum with @xmath30 , which is the typical apparent x - ray spectral slope of a non - obscured agn . in this paper ,
only x - ray sources in the region covered with deep optical imaging data described in the next section are considered .
the region is shown as 5 rectangles with thin solid lines .
some of the x - ray sources whose optical images are affected by halo or electron leakage from very bright objects are removed as well .
a total of 781 ( 584 ) soft ( hard)-band x - ray sources with likelihood above 7 are covered .
hereafter these sources are referred as the soft-(hard-)band sample . considering sources common to both samples
, there are 945 unique x - ray sources , we refer the combined sample as the total sample .
the flux distributions of the samples are shown with thick histograms in figure [ fig : flux_dist ] .
the corresponding survey area for the soft-(hard-)band sample is shown in the left ( right ) panel of figure [ fig : sxds_areacurve ] .
the solid line represents the survey area of the entire sample of @xcite , and the dotted line shows the survey area of the sample considered in this paper .
deep optical imaging observation in the field was conducted with the optical wide - field @xmath31 mosaic ccd camera , suprime - cam , attached to the prime - focus of the 8.2 m subaru telescope .
the observation was conducted as a subaru observatory project from september 2002 to january 2004 .
imaging data in five overlapping fovs in the @xmath32 , @xmath33 , @xmath1 , @xmath2 , and @xmath34 bands was obtained .
the thin solid rectangles in figure [ fig : sxds_fov ] indicate the area covered .
the total exposure times for each field range 5.5 - 5.8h ( @xmath32 ) , 4.9 - 5.3h ( @xmath33 ) , 3.9 - 4.1h ( @xmath1 ) , 5.2 - 10.8h ( @xmath2 ) , 3.1 - 5.2h ( @xmath34 ) .
for details of the deep imaging data , see @xcite . in these images
the fwhm of stellar objects is @xmath35 .
we apply a gaussian convolution to make the resulting psfs in the 5 bands constant for each region .
we use the photometric zero - point provided in @xcite .
the typical 2 @xmath14 detection limits in @xmath36 diameter apertures are determined by measuring the standard deviation of fluxes within a fixed aperture at positions where no object is detected .
the limits are 28.7(@xmath32 ) , 28.1(@xmath33 ) , 28.1(@xmath33 ) , 28.0(@xmath2 ) , and 26.9(@xmath34 ) mag . correcting for the aperture loss for point sources , the typical magnitude limits are 28.1(@xmath32 ) , 27.6(@xmath33 ) , 27.6(@xmath1 ) , 27.6(@xmath2 ) , and 26.6(@xmath34 ) mag for the total magnitudes of point sources .
the limiting magnitudes in the aperture and for the total magnitudes are summarised in table [ tab : imagelist ] . for other bands ,
we derive the limiting magnitudes for aperture and total magnitudes in the same way .
we use the suprime - cam @xmath1-band astrometry as the astrometric reference frame in this paper .
astrometric calibration is carried out in the @xmath1-band images using the 2mass point source catalog .
the rms residual of the fitting is typically @xmath37 .
the multi - wavelength data described below are aligned to the @xmath1-band image using commonly detected stellar objects .
the entire survey region is covered with the canada france hawaii telescope legacy survey ( cfhtls ) .
we use the @xmath38-band data from the cfhtls archive .
the typical 3.0@xmath14 total - magnitude limit for a point source is 26.11 mag .
part of the survey region , the southern fov out of the 5 suprime - cam fovs , is covered by a deeper @xmath39band imaging observation with suprime - cam with 3.7 hours integration .
details of these data are described in @xcite .
the 3.0@xmath14 limiting total magnitude of a point sources is 26.46 mag .
the central @xmath40 area of the survey region is also covered by a deep imaging data obtained with the mosaic - ii camera attached to the ctio blanco 4 m telescope ( @xcite ; pi .
n.padilla ) . in this observation ,
the sloan digital sky survey ( sdss ) @xmath41-band filter is used .
the region covered is shown with a thin dashed line in figure [ fig : sxds_fov ] .
the total integration time is 10h and the 3.0@xmath14 limiting total magnitude of a point source is 26.29 mag . because the filter systems used in the three @xmath38-band images are different , we adjust the suprime - cam and mosaic - ii magnitudes to the photometric system used in the megacam . using stars in the overlapping regions
, we fit the magnitude difference between the suprime - cam ( mosaic - ii ) and megacam photometry with a linear function of the @xmath42 color , and apply the function to the suprime - cam ( mosaic - ii ) photometry .
we check the photometric zero - point of the megacam data using the @xmath32 and @xmath1 photometry from the suprime - cam data and the @xmath42 and @xmath43 color - color diagram of stellar objects .
the distribution of the stellar objects on the diagram is consistent with the colors derived with the seds of stars from @xcite . for bright objects which are saturated in the deep optical images , their optical magnitudes are determined with shallow optical images taken with wide field camera attached to issac newton telescope .
@xmath38- , @xmath44- , @xmath1- , @xmath45- , @xmath46-bands data are available .
the @xmath44 , @xmath1 , @xmath45 , @xmath46 optical magnitudes are adjusted to the suprime - cam @xmath32 , @xmath1 , @xmath2 , @xmath34 system with magnitude offset and its color dependence derived with stars in the overlapping region . a large part of the sxds region is covered by the ultra deep survey of ukirt infrared deep sky survey ( ukidss uds ; @xcite ) .
the coverage of this survey is shown with the thick solid rectangle in figure [ fig : sxds_fov ] .
archival image data from data release 8 ( ukidss dr8 ) is used in this paper .
the stellar image size is @xmath47 .
we use the photometric zero - point provided in the data release .
the 3 @xmath14 detection limit of the image for a point source is measured to be 25.40 , 24.82 , 25.30 mag for @xmath48- , @xmath49- , and @xmath15 bands , respectively , with a @xmath50 diameter aperture photometry and aperture correction for point sources . for x - ray sources outside of the ukidss uds region , we do not have nir photometric data .
almost the entire fov of the sxds is covered by the _
observation with the infra - red array camera ( irac ) instrument obtained in the course of the _ spitzer _ wide - area infra - red extragalactic ( swire ) legacy survey project ( @xcite ; @xcite ) .
seven sources in the total sample are outside of the filed covered .
we use publicly available mosaiced images of irac ( dr2 2005 - 06 - 03 ; @xcite ) .
the fwhm of the psfs in the irac images are determined to be 1.9 , 2.0 , 1.9 , and 2.2@xmath51 by combining images of @xmath52 stars in the fov , and the corresponding aperture corrections for a @xmath53 diameter aperture are 0.33 , 0.36 , 0.54 , and 0.66 mag for 3.6 , 4.5 , 5.8 , and 8.0 bands , respectively @xcite .
we determine the photometric zero - points of the mosaiced irac images using @xmath54 , @xmath55 , @xmath56 , and @xmath57 colors of bright blue stars following @xcite ; values of 0.04 , 0.02 , 0.00 , and 0.00 mag are used for stars with @xmath58 colors bluer than 0.3 mag based on a theoretical stellar sed model ( table 5 of @xcite ) .
near - infrared photometry of stellar objects in the sxds field is taken from the two micron all - sky survey point source catalog ( 2mass psc ; @xcite ) .
@xmath15-band magnitudes in the 2mass psc are determined with a profile fitting method .
we select blue stars with @xmath59 mag and examine their average @xmath54 , @xmath55 , @xmath56 , and @xmath57 colors .
photometry in the irac bands are determined with a @xmath53 diameter aperture with an aperture correction derived from the psf .
we determine the photometric zero - points to match the expected colors from the theoretical stellar sed model .
the 3 @xmath14 detection limit of the image for point sources is measured to be 22.77 , 22.07 , 19.97 , and 19.91 mag for the 3.6 , 4.5 , 5.8 , and 8.0 bands , respectively , with a @xmath53 aperture photometry diameter and assumption for the aperture correction for point sources .
comparing the coordinates of objects detected in both of the @xmath1-band and 3.6-band , we found a @xmath60 offset in the right ascension direction between the deep suprime - cam @xmath1-band and the irac 3.6-band images .
we shift the irac image to match the coordinates of the suprime - cam image . after the correction
, the difference between the @xmath1-band and 3.6-band coordinates is @xmath61 r.m.s .
, which is thought to be dominated by the positional uncertainty of the 3.6 source .
a significant part of the sxds region is covered by one of the _ spitzer _ legacy programs , _ spitzer _ ukidss ultra deep survey ( spuds ) .
irac observations cover the region shown in gray in figure [ fig : sxds_fov ] . the photometric zero - point of the spuds data
is adjusted to swire photometry with overlapping objects .
the typical 2.0@xmath14 detection limit for @xmath62 aperture is 25.26 , 24.83 , 22.99 , and 23.19 mag and the detection limits for the point source are estimated to be 24.49 ( @xmath63 ) , 24.03 ( @xmath63 ) , 21.69 ( @xmath63 ) , 21.77 ( @xmath64 ) mag for 3.6 , 4.5 , 5.8 , and 8.0 bands , respectively .
the _ spitzer _ observation with the multiband imaging photometer ( mips ) instrument in the swire legacy survey ( @xcite ; @xcite ) also covers almost the entire area of the sxds .
we use publicly available mosaiced images of mips data in 24 band ( dr5 2006 - 12 - 23 ) .
the fwhm of psf of the mips image is @xmath65 , and aperture correction for a @xmath66 diameter aperture is 1.23 mag .
we use the photometric zero - point provided in the data release .
the 2@xmath14 detection limit of the data in the aperture is 20.21 mag , and the 4@xmath14 detection limit for the total magnitude of stellar objects is 18.22mag . a large part of the sxds region is also covered by the mips observation of spuds . the coverage is shown with the thick dashed line in figure [ fig : sxds_fov ] .
the 2@xmath14 detection limit in a @xmath66 diameter aperture is 22.06 mag , and the 4@xmath14 detection limit in total magnitude of stellar objects is 20.07mag .
deep imaging survey ( dis ) of the galaxy evolution explorer ( _ galex _ ; @xcite ) covers almost the entire area of the sxds in the fuv ( effective wavelength of 1516 ) and nuv ( effective wavelength of 2267 ) bands .
we use the reduced data obtained from the _ galex _ archive .
the tilenames of xmmlss_03 ( 29189s , 30289s ) , xmmlss_04 ( 27306s , 28561s ) , xmmlss_07 ( 27440s , 27520s ) , and xmmlss_09 ( 30606s , 32354s ) overlap with the sxds .
the numbers in the parenthesis show the total exposure times in fuv and nuv bands , respectively , for each tile .
the fwhm of the reduced images are @xmath67 and @xmath68 in the fuv and nuv images , respectively .
we use the photometric zero - point provided in @xcite ; 18.82mag and 20.08mag for fuv and nuv bands .
the aperture corrections for point sources are 0.36 and 0.59 mag for fuv and nuv bands for a @xmath69 diameter aperture @xcite .
the typical point source limiting magnitudes estimated with the @xmath70 aperture photometry reach down to 25.52 ( 4@xmath14 ) and 25.15 ( 3@xmath14 ) mag in the fuv and nuv bands , respectively .
for each x - ray source the statistical error in the position estimated through the maximum likelihood fit , @xmath71 , is given in @xcite .
when we match the x - ray and @xmath1-band sources , we use an uncertainty radius , @xmath72 , which is determined by the square of the @xmath71 and @xmath73 added in quadrature .
the latter term represents any positional uncertainties other than @xmath71 , i.e. any systematic uncertainty in the positions of the @xmath1-band sources and x - ray sources .
the term is estimated through the scatter of positional differences of x - ray sources whose @xmath71 is less than @xmath74 and their candidate @xmath1-band counterparts .
sample images of three x - ray sources in the @xmath1-band are shown in the second column of figure [ fig : identification_sample ] .
almost all of the x - ray sources have at least a @xmath1-band source in the @xmath75 radius above the detection limit of the @xmath1-band images .
some x - ray sources have multiple @xmath1-band sources within their error radii .
upper left panel of figure [ fig : coord_diff ] shows radial distribution of @xmath1-band sources down to 27.0 mag from the x - ray positions .
the dotted line shows the average number density of sources outside of x - ray uncertainty area .
the excess sources above the line statistically represent the counterparts of the x - ray sources .
the average number density corresponds to the expected number of the contaminating sources .
if we consider the region within @xmath76 down to @xmath77 mag , the ratio between the number of excess and contaminating sources is 0.30 .
it is thus important to select the most probable counterpart for each x - ray source . at first
, we identify the x - ray sources with sources detected in the @xmath1-band .
we construct a catalog of @xmath1-band sources by applying sextractor @xcite to the deep suprime - cam @xmath1-band images . in order to quantitatively select the most probable identification , we utilise the likelihood ratio defined as the probability of an x - ray source @xmath45 with x - ray flux @xmath78 having a counterpart @xmath79 with magnitude @xmath80 and distance @xmath81 divided by the probability of it being a chance match ( @xcite ; @xcite ; @xcite ) .
the likelihood ratio is calculated as : @xmath82 assuming the spatial distribution of the counterparts in the x - ray error circles follows the gaussian distribution .
@xmath83 is a cumulative number density of optical objects brighter than @xmath80 .
@xmath84 is the probability of the @xmath45-th x - ray source has its counterpart brighter than @xmath80 .
if the probability of being a counterpart for the x - ray source exceeds the probability of being a chance match , the likelihood ratio is greater than 1 .
the number density of optical objects is calculated from the suprime - cam @xmath1-band image directly without correction for detection incompleteness .
we determine @xmath84 by examining the distribution of @xmath85 for the x - ray sources , and assume that the distribution is constant in the sxds x - ray flux range .
we consider @xmath1-band sources within 5.0@xmath86 of each x - ray source .
firstly , we calculate @xmath87 with @xmath88 and select the @xmath1-band object with the highest likelihood ratio for each x - ray source as the initial guess for the @xmath1 primary candidate . based on the @xmath1-band magnitudes of the identified objects , we determine @xmath84 .
the @xmath89 value is calculated with @xmath90 in the ab magnitude system .
the half - maximum wavelength range of the @xmath1-band filter response ( 59507085 ) of the suprime - cam is used in the calculation .
we use the 0.510 kev flux , which is derived by summing the flux in the 0.52 kev and 210 kev bands , as the x - ray flux @xmath78 .
we call the @xmath1-band object with the highest likelihood ratio for a given x - ray source the @xmath1 primary candidate for the source .
the reliability of the identification is evaluated as the probability that a @xmath1-band object @xmath79 is the counterpart of an x - ray source @xmath45 .
the reliability is given by @xmath91 @xcite .
@xmath92 represents the probability of finding an @xmath1-band counterpart above the detection limit @xmath93 .
we use a value of @xmath92 of 0.8 for all x - ray sources for simplicity .
the cumulative distribution of the reliability of the @xmath1 primary candidate is shown in figure [ fig : reliable_hist_all ] with thick and thin black solid lines for the soft- and hard - band samples , respectively .
a significant fraction of the x - ray sources have low reliability ; 30% of the x - ray sources have primary optical counterpart with reliability smaller than 0.6 .
x - ray sources that have the @xmath1 primary candidate with low - reliability have a few @xmath1-band objects with similar @xmath1-band magnitudes and distances from the x - ray position .
these are the problematic cases and the fraction of such cases in the total sample is similar to those in _ xmm - newton _ survey fields with similar depth @xcite . in order to quantitatively examine the reliability of the optical identification process , we conduct monte - carlo simulations of the optical identification procedure .
we use the large catalog of the optical objects in the fields made with the deep optical imaging .
we regard each optical object as an x - ray source and add pseudo x - ray flux information and simulate the x - ray coordinates by adding positional uncertainty to the optical coordinates .
we use @xmath12 erg s@xmath8 @xmath9 in the total band for the x - ray flux .
we assume that the positional error follows a gaussian distribution .
we simulated a positional error @xmath72 of @xmath94 to @xmath95 with a @xmath96 step .
we apply the same optical identification process described above to the pseudo - x - ray source catalog . the identification process recovers more than 90% of the counterparts for the pseudo - x - ray sources with @xmath97 , and the fraction decreases down to 40% for the pseudo - x - ray sources with @xmath98 and counterparts of @xmath99mag .
it is important to evaluate the fraction of correct identifications among the @xmath1-band primary candidates . for pseudo - x - ray sources with @xmath97 and primary candidate of @xmath100mag , 95% of the primary candidates
are the correct identification .
on the other hand , only 50% of the primary candidate with @xmath101 mag are correct for pseudo - x - ray sources with @xmath102 .
the results also show that even among the optical identifications with a bright optical counterpart , @xmath103 mag , there can be non - negligible fraction of incorrect identifications . based on the simulations
, we can estimate the probability of mis - identification for each x - ray source with @xmath72 and the @xmath1-band magnitude of the primary candidate . summing the probability of the mis - identification for all of the x - ray sources
, we calculate the expected number of mis - identification to be 102 among the 945 sources in the total sample .
the identification only with the @xmath1-band information is clearly affected by mis - identification . for the primary candidate identified in each band .
black solid curves are for the @xmath1-band identification .
red dotted lines are for the 3.6 identification .
blue dot - dashed lines are for 24 identification .
green dashed lines are for nuv identification .
the thick ( thin ) lines are for the hard - band ( soft - band ) sample.,width=283 ] most of the expected mis - identifications are thought to be associated with x - ray sources with large @xmath71 and faint optical counterparts . because optically - faint x - ray sources have on - average a red optical - mir color ( @xcite ; @xcite ;
@xcite ) ; identification in the mir - band can be more reliable ( @xcite ; @xcite ) .
upper middle and upper right panels of figure [ fig : coord_diff ] show the radial distribution of 3.6 sources from the centers of the x - ray sources . for the uncertainty radius , @xmath104 , we use root sum squared of the @xmath71 and @xmath73 .
the latter term is derived same way for the @xmath1-band sources . if we take the ratio between the numbers of excess and contaminating sources within @xmath105 , the ratios are 2.2 and 0.9 at the depth of the swire and spuds coverages , respectively .
the ratios are higher than the ratio in the @xmath1-band , and it is expected that more reliable identification can be made in the 3.6-band .
@xcite really show that 3.6 identification provides a better counterpart assignment than optical only identification for _ xmm - newton _ x - ray sources .
they checked the @xmath1 and 3.6 identifications of _ xmm - newton _ x - ray sources with the source positions measured with _
chandra _ , and found 20% and 4% different identifications from _ chandra _ for the @xmath1 and 3.6 primary candidates , respectively . in order to improve the reliability of the identification
, we apply the 3.6 identification utilising the swire and spuds 3.6 images .
first , we construct a 3.6 object catalog applying sextractor @xcite to the 3.6 images with detection criteria of 3 connected pixels above 2@xmath14 of the background .
we use mag_auto as the 3.6 magnitude .
then , we identify the x - ray sources with the 3.6 objects in the same way for the @xmath1-band objects using @xmath87 and @xmath106 determined with the 3.6 object list .
we consider 3.6 sources within 5@xmath107 of each x - ray source .
the number count of the 3.6 objects are derived from the source number density in the 3.6 source catalog .
we select 3.6 primary candidates with @xmath108 , then we derive @xmath109 with the 3.6 magnitudes of the primary candidates .
finally , we conduct the 3.6 identification with the derived @xmath109 .
the distribution of the reliability of the 3.6 primary candidate is shown in figure [ fig : reliable_hist_all ] with the red dotted lines .
the fraction of the high reliability identifications among the 3.6 identifications is higher than that among the @xmath1-band identifications : 80% of the sources have a primary candidate with @xmath110 .
we evaluate the expected number of mis - identifications among the 3.6 primary candidates using a monte carlo simulation as for the @xmath1-band identification .
the mis - identification rate for all of the x - ray sources is 44 among the 945 x - ray sources .
the fraction of expected mis - identifications is consistent with that reported in @xcite .
the expected number of mis - identifications is significantly reduced from that of the @xmath1-band identification .
the psf size of the 3.6-band irac image is larger than that of the @xmath1-band image , but as can be seen in figure [ fig : identification_sample ] , sufficiently small to pinpoint which is more reliable counterpart among the multiple @xmath1-band sources that are in the x - ray uncertainty area .
we use 3.6 identification to pinpoint the optical counterpart , and we refer the position of the selected optical counterpart as the coordinate of the counterpart . similar to the 3.6 identification , identifications in the @xmath15-band can be effective in selecting optically - faint and red counterparts .
additionally the psf size of the ukidss uds @xmath15-band image is significantly smaller than that of the irac 3.6 image , therefore it is possible that true counterparts that are blended with contaminating sources in the 3.6 image can be resolved in the @xmath15-band image . using the ukidss uds @xmath15-band image
, we check the difference of the @xmath15-band identifications from those in the @xmath1 and 3.6 bands .
the @xmath15-band identifications basically follow the @xmath1 and 3.6 identifications ; among 762 x - ray sources covered in the @xmath15-band image , 636 and 650 sources have @xmath15-band primary candidate consistent with primary candidate in the @xmath1 and 3.6 identifications , respectively .
if we remove @xmath15-band primary candidate which are located far outside of the x - ray error circle , there are 11 @xmath15-band primary candidates that are different from both of the @xmath1-band and 3.6 identifications .
all of the 11 @xmath15-band primary candidates have reliability lower than 0.5 , i.e. there are other @xmath15-band source with similar likelihood , and the identifications are marginal . considering the similarity to the @xmath1 and 3.6 identifications and the limited coverage of the @xmath15-band image , we do not refer the @xmath15-band identifications in the final selection of the primary counterparts .
determined in the 3.6 ( top ) , nuv ( middle ) , and 24 ( bottom ) bands identification vs. @xmath1-band identification .
red filled circles represent objects where all 4 identifications are the same .
red open circles have matched identifications in the @xmath1 and 3.6 bands .
blue filled squares and blue open squares are identified either in @xmath1 or 3.6 bands .
blue crosses are identified in the nuv or 24 bands .
black open squares and black open circles indicate objects matched with radio and variability agn candidates , respectively .
, width=283 ] utilising the multi - wavelength dataset of the sxds , we can further improve the reliability of the identification .
luminous blue broad - line agns at low to intermediate redshifts are reliably identified with strong nuv emission . on the other hand ,
obscured narrow - line agns show strong 24 emission . in the bottom panels of figure [ fig : coord_diff ] ,
radial distributions of nuv and 24 sources are shown .
the 1@xmath14 uncertainty radius , @xmath111 and @xmath112 , are calculated with root sum of squares of @xmath71 and the additional term associated with the systematic x - ray source and nuv or 24 position uncertainty .
we use @xmath113 and @xmath114 as the latter term for nuv and 24 identifications , respectively .
in the same way for @xmath1 and 3.6-bands , if we consider within @xmath115 or @xmath116 , the ratio between the number of excess and contaminating sources are 2.5 , 22.0 , and 8.3 for nuv down to 25.1 mag , 24 down to 18.1 mag ( swire depth ) and 24 down to 20.0 mag ( spuds depth ) , respectively .
thus most of the associated 24 sources are expected to be counterparts of the x - ray sources .
some contaminations are expected for nuv sources , see discussions on the distributions of reliability below .
therefore , we utilise nuv and 24 identifications as well as the @xmath1 and 3.6 identifications to determine the primary counterpart of each x - ray source . we refer a primary object selected in each identification as the primary candidate , and the best candidate in the combined identifications as the primary counterpart .
we perform nuv and 24 identifications in the same way as for the @xmath1 and 3.6 identifications .
first , we construct object catalogs from the nuv and 24 images with sextractor .
we use the nuv and 24 magnitude determined in @xmath70 and @xmath66 diameter apertures with aperture corrections of 0.59 and 1.23 mag determined with average psf , respectively .
the number densities of nuv and 24 sources are directly derived from the source list made by applying sextractor to the sxds dataset . for the nuv and 24 identifications
, we do not consider @xmath117 and set @xmath118 for simplicity .
we calculate @xmath106 in the same way for @xmath1 and 3.6 identification .
as shown in figure [ fig : identification_sample ] , the psf sizes of the nuv and 24-band images are significantly larger than that of the @xmath1-band images .
the psf sizes are still sufficiently small to pinpoint one @xmath1-band counterpart in most of the cases .
we apply eye - ball check for the @xmath1-band images of the nuv and 24 sources .
same for 3.6 identification , even if we use nuv or 24 identification , we refer the position of the selected optical counterpart as the coordinate of the counterpart .
the distribution of the @xmath106 of nuv and 24 identifications are shown with green dashed and blue dot - dashed lines , respectively , in figure [ fig : reliable_hist_all ] .
the nuv identification is good for blue luminous broad - line agns , and higher reliability identification can be achieved than @xmath1 identification for more than 40% of the x - ray sources .
however , another 40% of the sources have multiple nuv sources with similar magnitude in the uncertainty area giving a reliability less than 0.5 . in the 24 band
we achieve highly reliable identification , @xmath110 , for 80% of the sources . for the rest of the x - ray sources ,
the 24 data is not deep enough to detect the counterparts of the x - ray sources and the fraction is saturated at 80% .
the higher reliability of the nuv and 24 identifications can partly be caused by the blending effect due to the large psf size in the two bands .
the counterparts of the x - ray sources are bright in the nuv and 24 bands , and fainter nuv or 24 band sources that are contaminating and are located at the x - ray source position by chance can be hidden due to blending with the brighter true counterpart .
such effect of blending is discussed for 3.6 sources in @xcite . additionally , candidates of agns are selected in radio and variability surveys in sxds , and we utilise this information in the final identification . a large part ( 0.8 deg@xmath6 ) of the sxds field is covered by the very large array ( vla ) radio survey observation at 1.4ghz down to 100@xmath119jy @xcite . in total , 505 radio sources are detected , and 480 sources are within the x - ray coverage region . among the 945 x - ray sources in the total sample , 789 of them are covered in the 100@xmath119jy sample . if we consider 1@xmath14 positional matching uncertainty , @xmath120 with the root sum of squares of @xmath71 and @xmath121 , there are 62 radio sources associated with the x - ray sources within 4.0@xmath122 .
most of the broad - line agns detected in the radio survey are also detected in x - ray ; among 25 broad - line agns detected in the vla survey @xcite , 21 of them are in the x - ray survey region , and 17 of them are associated with x - ray sources .
three radio broad - line agns without x - ray detection are fainter than @xmath2=23.5 mag , and are likely to be below the x - ray detection limit .
the remaining radio broad - line agn without x - ray detection are located in a gap of the x - ray detectors , and the agn can be missed due to bad sensitivity in the region .
variable agn candidates are selected with @xmath2-band suprime - cam imaging data @xcite .
there are 936 variable objects detected down to magnitude of the varying component of @xmath123 in the overlapping region with the x - ray images @xcite , and 193 of them are matched with x - ray sources .
finally the primary counterpart is determined utilising all of the @xmath1 , 3.6 , 24 , and nuv identifications .
radio and variability agn candidates are also considered during the selection of the primary counterpart . among the 945 x - ray sources ,
213 sources have matched primary candidate considering all 4 identifications , and we select this object as the primary counterpart .
404 x - ray sources have only one object which have likelihood ratio larger than 1 either in the 4 identifications within the x - ray uncertainty area , and we select the object as the primary counterpart . for the remaining x - ray sources ,
there are more than 2 objects with @xmath124 within the x - ray uncertainty area .
we multiply the likelihood ratios of the 4 identifications for each source , and if there is an object with a likelihood ratio product larger than 1000 , we select this object as the primary counterpart .
there are 112 x - ray sources that meet the criteria . of the remaining 183 x - ray sources ,
90 x - ray sources have a primary candidate selected both in the @xmath1 and 3.6 identifications . among them , we select 77 of them as the primary counterpart . for 13 x - ray sources , there is another primary candidate selected at 24 or in the nuv , and the candidate is selected as the primary counterpart .
the other 93 x - ray sources have different primary candidates in the @xmath1 and 3.6 identifications .
for such source , we determine the primary counterpart comparing the likelihood ratios in the 4 identifications .
we select 52 , 24 , 20 , and 6 primary candidates with the 3.6 , @xmath1 , 24 , and nuv identifications , respectively . the distribution of the reliability of the primary counterparts in the @xmath1 , 3.6 , 24 , and nuv identifications are shown in figure [ fig : multi_reliability ] .
almost all of the primary counterparts have high reliability in one of the 4 identifications .
the list of the primary counterparts is shown in table [ tab : xrayid ] . in the list , the x - ray identification number ( column 1 ) , x - ray ra and dec .
position ( column 2 and 3 ) , x - ray 1@xmath14 uncertainty radius , @xmath71 ( column 4 ) , and likelihood in the 0.5 - 2 kev and 2 - 10 kev bands ( column 5 and 6 ) are shown .
columns 7 to 10 provide the information from the multi - wavelength identification process in the @xmath1 ( 7 ) , nuv ( 8) , 3.6 ( 9 ) , and 24 ( 10 ) bands .
flag 1 means the primary counterpart is the primary candidate in the band , flag @xmath125 means the counterpart is the primary candidate but the likelihood is less than 1.0 , and flag @xmath126 means the counterpart is not the primary candidate in the band .
column 11 and 12 show the matching with the 1.4ghz radio source and variability agn candidates . the ra and dec . of the primary counterpart
is given in columns 13 and 14 .
the multi - wavelenth images of the x - ray sources are available on - line .
the identifications discussed above assume the x - ray source is a single object .
this is not the case for clusters of galaxies , which are non - negligible population among extra - galactic x - ray sources .
some of the identifications with low - reliability can be caused by the multiple galaxies associated with an x - ray source that is associated with a cluster of galaxies .
@xcite constructed catalogs of groups and clusters of galaxies in the sxds field through a wavelet detection of extended x - ray emission for the _ xmm - newton _ imaging dataset .
they cross - match their candidates with the extended sources in @xcite , and 14 sources in @xcite are matched . among them , we regard 10 extended x - ray sources as candidate of clusters of galaxies as listed in table [ tab : cluster ] . among the remaining 4 sources , we regard 3 sources as candidates of agns .
sxds0051 is cross - identified with sxdf47xgg at @xmath127 , but sxds0051 is significantly detected in the 2.010 kev band and the observed 0.52 kev and 24.5 kev hardness ratio , hr2 , is relatively large ( @xmath128 ) .
thus we regard the x - ray source as an agn candidate .
sxds0396 is cross - identified with sxdf64xgg at @xmath129 , but sxds0396 is significantly detected in the 2.010 kev band and hr2 ( @xmath130 ) implies a power - law with photon index of 1.6 .
therefore , we regard the x - ray source as an agn candidate .
the source is also close to sxds0386 .
sxds0934 is cross - identified with sxdf06xgg at @xmath131 , but the primary optical counterpart is identified with a broad - line agn at @xmath132 , therefore we regard the agn as the x - ray source . sxds0784
matched with sxdf07xgg has a detection likelihood less than 7 in both of the soft and hard x - ray bands , and it is not included in the total sample .
we also found 4 additional candidates of clusters of galaxies with an excess of objects with similar @xmath133 and @xmath134 colors within @xmath135 from the x - ray source position .
these are also listed in table [ tab : cluster ] .
the median photometric redshift of objects within @xmath136 from the x - ray position is used as the photometric redshift for the candidates . in the table ,
the x - ray source i d ( column 1 ) , hr2 ( 2 ) , spectroscopic redshift ( 3 ) , and photometric redshift ( 4 ) are summarised .
four x - ray sources , sxds0416 , sxds0390 , sxds0404 , and sxds0402 , are associated with a bright nearby galaxy kug 0214@xmath0057 at @xmath137 @xcite .
they are strong ultra - luminous x - ray sources ( ulxs ) candidates .
we conducted a spectroscopic campaign observation with a multi - slit optical spectrograph , focas on the subaru telescope from january 2001 to november 2005 .
we targeted the primary counterparts with the highest priority , and in addition we observed secondary counterparts if possible .
we put the slits at the positions of the primary counterparts determined in the optical images , even for primary counterparts selected by the identifications in the 3.6@xmath119 m , 24@xmath119 m , or nuv band .
the situation is the same for following spectroscopic observations .
the fov of the instrument is a 6@xmath138 diameter circle . in total 50 fields
were observed .
about three quarters of the fields were selected to cover as many as x - ray sources as possible .
the remaining fields were chosen to cover as many as @xmath139 red sequence galaxies as possible @xcite and to cover a candidate @xmath140 cluster of galaxies found by narrow - band surveys @xcite . in such fields , x - ray sources
are observed as a filler targets for the masks . in the observations ,
we use a 150r grating with sy47 order - cut filter to cover 4700 to 9400 .
a slit width of @xmath141 is used and the resulting spectral resolution is 20 .
hereafter , we use the spectral resolution measured with night sky emission lines obtained simultaneously with the target spectrum . the spectral resolution for
the target can be higher because the slit illumination by a stellar object can be more concentrated than the smooth illumination by night sky emission . in later observations , we use 300b grating with order - cut filter sy47 or without the filter . with the order - cut filter
the spectral coverage is 4700 to 9200 . without the filter
the spectral coverage in the blue extends down to 3800 , but 2nd order light can contaminate wavelength range above 7600 . a slit width of @xmath141 is used and the resulting spectral resolution is 13 .
the central coordinate , observing mode , and the exposure time of each field are summarised in table [ tab : focas_fovs ] . during the observation
, a few additional objects were observed in long - slit mode with 300b grating with sy47 order - cut filter .
the data are reduced with a script based on iraf .
the procedure follows usual analysis of long - slit observations .
the bias counts are subtracted using the overscan area of the ccd , and a flat - fielding correction is applied with a dome flat frame taken after the observation .
wavelength calibration is performed using a thar lamp spectrum .
we did not dither the object position along the slit during the observation , and the sky subtraction is performed by using the sky spectra obtained on either side of an object .
the flux calibration of the reduced spectra is carried out utilising the spectra of spectroscopic standard stars taken with a @xmath142-width slit .
feige 110 and g191b2b are used as the standard stars .
the standard stars are obtained with and without sy47 order - cut filter , in order to evaluate the contamination of the 2nd order light at the red end of the wavelength coverage . the sensitivity curve for each night
is determined with the standard stars .
furthermore , we adjust the normalization of the spectra of the targets using their photometric data . after the sensitivity correction , we also applied correction for atmospheric absorption features , a - band and b - band , using the absorption profile of the bands derived with a bright object in the same mask .
spectroscopic follow - up observations for bright optical counterparts with @xmath143 mag were obtained with the 2df 400 multi - fibre optical spectrograph on anglo - australian 3.9 m telescope on 3 - 6 nov .
2002 in director s discretionary time .
the integration time ranges from 6 min for the brightest objects to 3 hours for the faintest objects .
the 2df has 2 spectrographs , each spectrograph covers 200 fibres .
we used a 270r grating with wavelength range of 4400 - 9100 for the first spectrograph , and 316r grating with wavelength range of 4700 - 8500 for the second spectrograph .
the spectral resolutions are 14 and 12 and the spectral samplings are 4.8 and 4.1 pixel@xmath8 for the former and latter spectrographs . we observed a relatively bright star in each fibre configuration in order to correct wavelength dependence of the sensitivity .
the data are reduced in the standard way for the 2df data using the _ 2dfdr _ software .
some of the radio and x - ray sources in the sxds field were observed with the visible multi - object spectrograph ( vimos ) on the very large telescope ( vlt ) ( p074.a-0333 ; @xcite ) .
the mr - orange grating and the gg475 order - cut filter were used and a spectral resolution of 10 in the wavelength range between 4800 and 10,000 is achieved .
27 slit masks were prepared to cover @xmath144% of the radio and x - ray sources in the sxds .
@xmath14575% of the 27 masks were observed with one 2700-s and two 1350-s exposures .
additionally , two fovs of vimos were observed in the course of spectroscopic observations of extended ly@xmath146 emitters by @xcite ( p074.a-0524 ) .
the hr - orange grism and the gg435 order - cut filter were used . the wavelength range between 5000 and 6800
was covered with 2.8 resolution .
a deep spectroscopic survey for @xmath15-band selected galaxies in the ukidss uds region was conducted with the vlt vimos and vlt focal reducer and low dispersion spectrograph 2 ( fors2 ) instruments in the udsz eso large programme ( 180.a-0776 , pi : almaini ; @xcite ; @xcite ) . in the vimos observations ,
the lr - blue and lr - red grisms were used for each mask and wavelength range between 3700 - 9500 was covered . a spectral resolution of 35
was achieved at 7000 .
the fors2 observations were conducted with a 300i grism covering 6000 - 10000 and with spectral resolution of 12 .
in the course of the optical spectroscopic survey of lyman alpha emitters and lyman break galaxies in the sxds field , the primary counterparts of the x - ray sources were observed as filler objects in the masks .
a multi - object spectroscopic observation with the deep imaging multi - object spectrograph ( deimos ) instrument attached to the keck ii telescope was also carried out @xcite .
the observations used the 830 g grating and the gg495 filter .
the 5700 - 9500 wavelength range was covered with a spectral resolution of 2.6 .
further spectroscopic survey observations were made with the inamori - magellan areal camera and spectrograph ( imacs ) on the magellan 6.5 m baade telescope @xcite . in the imacs observations , the f/2 camera with @xmath147 diameter fov and
300 lines mm@xmath8 grism with central wavelength of either 6700 or 8000 were used .
wavelength ranges of from 4700 to 8000 ( from 6200 to 9600 ) were covered with 7 ( 6 ) resolution and 1.34 pixel@xmath8 ( 1.25 pixel@xmath8 ) by the former ( latter ) setup .
x - ray sources are also observed as filler targets during spectroscopic observations of the balmer - break galaxies in the field with imacs on magellan baade telescope ( pi .
n.padilla ; @xcite ) . in the observation ,
the f/2 camera with @xmath147 diameter fov and 200 lines mm@xmath8 grism with central wavelength of 6600 was used . the wavelength range from 4500 to 9000
is covered with 10 spectral resolution and 2.0 pixel@xmath8 sampling .
some qso candidates identified with the kx selection technique were spectroscopically observed with the aaomega spectrograph on the aat 3.9 m telescope and x - ray agns are also observed as filler targets @xcite .
the 580v and 385r gratings were used in the dual - beam spectrograph of aaomega and the wavelength range from 3700 to 8600 was covered simultaneously .
spectral resolution is 5.0 with spectral sampling of 3.4 pixel@xmath8 .
in addition to the optical spectroscopic observations , we conducted intensive nir spectroscopic observations with the fiber multi object spectrograph ( fmos ) on the subaru telescope @xcite .
this instrument can observe up to 200 objects simultaneously over a 30@xmath148 diameter fov in the cross - beam switching mode with two spectrographs , irs1 and irs2 .
the fibers have an aperture of @xmath149 diameter on the sky .
the spectrographs cover the wavelength range between 9000 and 18000 with spectral resolution of @xmath150 at @xmath151 m in the low - resolution mode .
a total of 851 sources were observed in observations of 22 fovs with this setup during guaranteed , engineering and open - use ( s11b-098 silverman et al . and s11b-048 yabe et al . )
time observations .
the central coordinate , observation period , integration time , and observing mode of each fov are summarised in table [ tab : fmos_fovs ] .
the nir spectra are reduced with the pipeline data reduction software , fibre - pac @xcite .
the resulting one - dimensional spectra are corrected for atmospheric absorption and sensitivity dependence on wavelength using relatively faint f - g type stars observed simultaneously with the targets .
the primary counterparts of 31 x - ray sources with a bright stellar object show optical spectra consistent with galactic stars . we identify them as galactic stars .
the i d number and hr2 of the x - ray sources are given in table [ tab : star ] .
almost all of them have a small hr2 value , which is consistent with the thermal x - ray emission from galactic stars .
hereafter , we exclude x - ray sources that are selected as candidates of clusters of galaxies ( section 4.4 ) , ulxs ( section 4.5 ) , and galactic stars , and refer to the remaining x - ray sources as agn candidates
. there are 896 agn candidates in the total sample as summarised in table [ tab : number_summary ] . among the primary counterparts of the 896 agn candidates , 597 counterparts
are observed in the optical spectroscopic observations discussed above , as is summarised in table [ tab : number_summary ] . as a first step , we determine the redshift of each object by examining its optical spectrum and fitting gaussian profiles to the observed emission and absorption lines .
a large fraction of objects show multiple features .
if a narrow emission line is detected , we determine the redshift with that line .
for example , if [ o ] @xmath1523727 emission line is detected , we determine the redshift of the object with that emission line .
for low - redshift objects , the optical spectrum covers the h@xmath146 and h@xmath153 wavelength range and shows strong [ o ] @xmath1525007 emission line .
we determine their redshifts with the [ o ] @xmath1525007 line . for high - redshift objects ,
optical spectrum covers the rest - frame wavelength range below mg @xmath1522800 emission line .
if that emission line is detected , it is used to determine the redshifts , otherwise , c ] @xmath154 or c @xmath155 emission lines are used .
for objects only showing absorption lines , we determine their redshift with the ca hk doublet absorption lines at 3969 and 3934 , if these absorption lines are detected .
we regard a permitted emission line with velocity width larger than 1000 km s@xmath8 ( after deconvolving the instrumental profile ) as a broad - emission line .
this threshold is narrower than the typical criteria for broad - line agns ( 1500 or 2000 km s@xmath8 ) .
we set the threshold considering the distribution of the fwhm of the broad - line agns in the local universe @xcite . although the threshold is narrower , most of the broad emission lines are broader than 1500 km s@xmath8 and the distribution of the fwhm of the broad - emission lines , especially the lack of the broad - emission lines below 1500 km s@xmath8 , is similar to that of optically - selected broad - line agns in the sdss as discussed in @xcite .
some objects only show one either broad or narrow emission line in the observed optical wavelength range .
for such objects , we identify most of the single narrow - emission lines with [ o ] @xmath1523727 , considering the absence of any other emission line .
if the identification is not conclusive due to the narrow wavelength coverage of a spectrum , we identify the single narrow - emission line , utilising the photometric redshift described in the next section . for objects with one prominent broad emission line ,
we mostly identify it with mg @xmath1522800 , consistent with the absence of another broad emission line in the observed wavelength range .
the identifications with a single broad - emission line in the optical are checked with the nir spectra covering rest - frame optical wavelength range .
intensive spectroscopic observations in the nir wavelength range cover 851 of the total x - ray sources as summarised in table [ tab : number_summary ] .
spectroscopic identification in the nir spectra is conducted in the same manner as for the optical spectra . in the nir spectra ,
if one narrow or broad emission line is detected , almost all of them are identified with the h@xmath146 emission line .
we identify some single narrow - emission lines with [ o ] @xmath1525007 , considering the photometric redshift . in such cases , [ o ]
@xmath1524959 is presumed to be weaker than the detection limit . among the objects that are observed but not identified in the optical spectroscopy ,
23 objects are identified with nir spectroscopic observations with fmos .
they are agns at @xmath156 , and 15 of them are narrow - line agns , which are mostly identified through their strong h@xmath146 or [ o ] @xmath1525007 lines .
objects that show broad - emission in at least one of the observed permitted emission lines are referred with `` broad - line '' agn , hereafter .
it needs to be noted that the criteria is not homogeneous and depends on the rest - frame wavelength coverage of the optical and nir spectroscopic observations .
for example , some objects are identified as `` broad - line '' agn with one prominent broad h@xmath146 emission line , but the same objects can be identified as `` narrow - line '' agn , if spectra covering h@xmath146 line are not available .
furthermore , an object with broad h@xmath146 emission lines can be type 1 , 1.5 , 1.8 or 1.9 seyferts .
on the other hand , objects without broad - emission lines are referred with `` narrow - line '' agns .
considering the x - ray luminosity distribution of the sample , their x - ray emission is thought to be dominated by the agn component , and we ignore contributions to x - ray sources from star - forming galaxies in the agn candidates . utilising the optical and nir spectra ,
586 out of the 896 agn candidates ( 65% ) are spectroscopically identified .
the fraction of spectroscopically identified objects in the entire agn candidate sample is shown in figure [ fig : sxds_idrate ] as a function of radius from the center of the survey field .
if we limit the sample to primary counterparts brighter than @xmath157 mag , the identification rate reaches 86% .
the spectroscopic identification of each agn candidate is summarised in table [ tab : multi3 ] .
column 22 shows the spectroscopic redshift , and column 23 describes the classification .
bla and nla represent broad - line and narrow - line agns , respectively .
nid indicates the object is observed in the optical but can not be identified with the optical spectrum .
nsp indicates the object is not observed in the optical .
nid and nsp do not refer the availability of nir spectrum .
mag , for all x - ray sources , and for x - ray sources with @xmath158 mag.,width=283 ]
for x - ray sources without spectroscopic identification , we determine their photometric redshifts . in order to determine photometric redshifts reliably , we conduct multi - wavelength photometry of each primary counterpart paying careful attention to the difference in the psf size and shape . in the photometry , the @xmath2-band image taken with suprime - cam
is used as the reference frame for the object detection .
sextractor is used with a detection threshold of 3 connected pixels above 2.0 @xmath14 . the aperture magnitude determined with a variable elliptical aperture ( mag_auto )
is used as the @xmath2-band total magnitude of an object .
we use phot_autoparams of ( 2.5 , 3.5 ) to determine the variable elliptical aperture . with the parameters ,
94% of total flux will be covered within the variable elliptical aperture @xcite .
using the @xmath2-band detected catalog , optical colors of the object are measured in @xmath142 diameter aperture using the multi - image mode of sextractor .
the total magnitude in each optical band is determined with a fixed aperture color and the aperture correction determined in the @xmath2-band , i.e. the difference between @xmath2-band mag_auto and @xmath2-band mag_app . for photometry in wavelengths other than the optical bands , we measure a fixed aperture color between a selected band and @xmath2-band using the psf - matched @xmath2-band image whose psf is matched to the image in the selected band .
the fixed aperture color is converted into the total magnitude in the selected band using the difference between the @xmath2-band mag_auto in the original @xmath2-band image and the aperture magnitude measured in the psf - matched image .
we try to conserve the fixed aperture color , and assume the light profile in the selected band follows that in the @xmath2-band .
the details are described as follows .
at first we subtracted the background component of each image using the object detection software , sextractor .
the background is determined locally with median filtering in a 48@xmath51 box .
for very bright and extended objects with sizes larger than 30@xmath51 , i.e. nearby galaxies , the background removal subtracts the extended halo component , thus the photometry is not reliable for such objects .
background - subtracted images are magnified and aligned to suprime - cam @xmath2-band imaging data .
coordinate conversions are determined with positions of common bright stars using the * geomap * command in iraf with a 2nd order polynomial and the images are registered with the * gregister * command .
fluxes of objects in the original images are conserved .
the psf size of the suprime - cam @xmath2-band image is the smallest amongst the multi - wavelength data .
we make the psf - matched @xmath2-band images by convolving with a kernel which represents the psf difference between the @xmath2-band and a given band .
the determination of the kernel and the convolution is performed with the * psfmatch * command in iraf . after matching the psf size and shape
, we measure colors of objects with the multi - image mode of sextractor .
colors in the @xmath2-band and a given band are measured from the images in the band and the psf - matched @xmath2-band at the position in the original @xmath2-band image . the total magnitude in the band
is derived from the aperture photometry and the aperture correction derived from the difference between the mag_auto in the original @xmath2-band image and the mag_app in the psf - matched @xmath2-band image .
different aperture diameters are used for different datasets ; @xmath70 diameter for _ galex _ fuv and nuv , @xmath142 diameter for @xmath38-band images , and ukidss @xmath48 , @xmath49 , and @xmath15-bands , @xmath62 diameter for the irac images . the uncertainties in the photometric measurement are determined as the 1@xmath14 uncertainty of the flux measurement in the aperture magnitude assuming a background limited regime . for bright objects ,
the systematic uncertainty dominates the measurement uncertainty , and constant values , 0.01 - 0.05 mag depending on wavelength , are added to represent such systematic uncertainties . in order to obtain photometric data for primary counterparts that are not detected in
the @xmath2-band selected catalog , we also construct @xmath32- , @xmath1- , @xmath33- , @xmath34- and 3.6 band selected catalogs using the images in each band as the reference frame .
we apply the same procedures as for the @xmath2-band selected catalog .
the results of the multi - wavelength photometry are summarised in table [ tab : multi1 ] , [ tab : multi2 ] , and [ tab : multi3 ] . the total magnitude and associated error for each band
are given .
an error of 99.99 indicates the associated magnitude is an upper limit .
a total magnitude of 99.99 indicates the object is not covered in the imaging data of that band .
the values in the tables are corrected for galactic extinction .
the applied corrections are summarised in table [ tab : imagelist ] . in the photometric redshift determination
, we consider the stellarity of the primary counterpart .
the stellarity is determined in the deep @xmath2-band image . for objects brighter than @xmath159mag , the cls_star parameter from sextractor
is used . an object with cls_star larger than 0.90
is classified as stellar . for fainter objects ,
we use the fwhm in the @xmath2-band image to define stellar objects .
if the @xmath2-band fwhm is smaller than @xmath160 , an object is classified as stellar . for bright objects with saturation in their peak , the stellarity measurements are not reliable . in such cases
we check the optical images by eye , and assign the corrected stellarity in the catalog .
the resulting stellarity is also listed in column 21 of table [ tab : multi3 ] .
the photometric redshift of each primary counterpart is determined with the hyperz code @xcite , which determines photometric redshifts via template fitting with @xmath161 minimization of the multi - wavelength photometry ranging from 1500 to 8.0 .
we use two sets of templates , one represents galaxy seds and the other one represents qso seds . for each object , we determine photometric redshifts from the galaxy and qso templates separately . for the galaxy templates , we use the spectral evolutionary models of @xcite with miller & scalo initial mass function ( imf ) @xcite and evolving metallicity .
it needs to be noted that we assume a salpeter imf @xcite in the determination of stellar mass of host galaxies of narrow - line agns in section 7.2 in order to compare with results in literature .
the difference in the estimated photometric redshift caused by the difference in the applied imf is expected to be negligible . in the hyperz code ,
burst , constant , and @xmath162 models with @xmath163gyr to 30gyr are available .
the dust extinction curve of star - forming galaxies @xcite is utilised with @xmath164mag to @xmath165mag in step of @xmath166 .
the galaxy templates only represent the stellar continuum component , and neither the excess emission of an agn hot dust component nor pah emission are included , although such emission can affect the rest - frame wavelength range above 2@xmath119 m .
instead of including such a component in addition to the galaxy sed model , we increase the irac photometric errors to compensate . for qso templates
, we construct our own qso template by combining average uv - optical average seds from sdss and nir - mir seds of qsos from infrared space observatory ( _ iso _ ) observations .
we use average qso seds in @xcite in the wavelength range between 1200 and 7000 .
a total of 6 average qso seds are constructed depending on the uv - optical color .
we extend the sed to the shorter wavelength range by extrapolating the continuum component in the above range with a power - law . in the longer wavelength range above 1
, we used three qso ir seds from @xcite .
the seds are constructed from the _ iso _ observations of broad - line qsos . by adjusting the normalization ,
we connect the six optical and the three ir seds smoothly and make 18 qso templates covering uv to mir wavelength range . because the observed slopes in the uv - optical and ir wavelength ranges are independent
, we consider seds with blue uv - optical sed and red ir sed , and vice versa .
the templates are compared with the observed seds of spectroscopically - identified broad - line agns in figure [ fig : sed_plot_type1 ] .
all of the seds are normalized at 1 in the rest - frame .
the red and blue dots represents seds of x - ray agns with absorption characterised by a hydrogen column density , @xmath167 , larger than and smaller than @xmath168 @xmath9 , respectively .
the derivation of @xmath167 is described in section 7.1 .
the templates enclose the entire range of seds of broad - line agns , although some agns with @xmath169 show redder sed in the uv to optical wavelength range .
their seds are similar to the galaxy seds ( see section 7.3 ) .
the templates also cover the variety of the seds observed in the x - ray - selected broad - line agns in the cosmos survey @xcite . in the template fitting ,
two additional constraints are applied in addition to the @xmath161 minimization .
the first one is that an object with stellar morphology in the @xmath2-band image is assumed to be at @xmath170 . in the left panel of figure
[ fig : sxds_cstar_i ] , the @xmath2-band magnitudes of the spectroscopically - identified objects are shown as a function of redshift .
stellar objects are shown with asterisks and crosses in this figure . out of 261 stellar objects ,
209 are spectroscopically identified .
a total of 182 and 27 of them are identified with broad - line and narrow - line agns at @xmath170 , respectively .
stellar broad - line agns show bright nuclei and their observed optical light is dominated by the nuclear component with stellar morphology .
it should be noted that this constraint is only applicable to the x - ray samples with similar depth and area to sxds . as shown in section 7.1 , in the limited area of the sxds
, there is a correlation between redshifts and luminosities , therefore luminous ( @xmath171 ) broad - line agns , i.e. qsos , whose optical morphologies are dominated by the nuclear component , statistically appear only above @xmath170 .
such correlation is a typical of flux - limited surveys .
less - luminous broad - line agns are found at lower - redshifts and they show extended morphologies due to their host galaxies .
stellar narrow - line agns are not resolved possibly due to the fact that they are at high redshifts .
an additional consideration is the absolute magnitude range of the galaxy and qso templates .
taking into account the absolute magnitude ranges of spectroscopically identified agns , we limit the @xmath46-band absolute magnitude range of the galaxy ( qso ) template to lie between @xmath172 and @xmath173 mag ( @xmath174 and @xmath175 mag ) . in section 7.3 , it will be shown that the rest - frame optical to nir continuum of narrow - line agns are dominated by stellar continuum of their host galaxies . moreover ,
the range of the estimated stellar mass of the host galaxies of spectroscopically - identified narrow - line agns is limited ( see section 7.2 )
. such limited mass range of x - ray agn host galaxies has already been reported ( e.g. @xcite ) .
the absolute magnitude range of galaxy template corresponds to the stellar mass range of @xmath176 m@xmath177 with the galaxy sed templates used in section 7.2 .
we introduce the faint limit of the qso absolute magnitude range , because the host galaxy component will dominate the total magnitudes if nuclear component is fainter than the limit .
these consideration is necessary to reduce the number of outliers in photometric redshift estimation .
once we determine photometric redshifts with both of the qso and galaxy templates for each agns , we compare the minimum @xmath178 with qso templates , @xmath179 , and that with galaxy templates , @xmath180 , to chose the best photometric redshift for each agns . in the left panel of figure [ fig : hyperz_chi2 ] ,
@xmath179 and @xmath180 are compared for spectroscopically - identified agns .
the redshift is not fixed to their spectroscopic redshift during the fitting process , in order to include the effect of redshift uncertainty .
narrow - line agns ( red circles and crosses ) on average better fitted with galaxy templates , and broad - line agns ( blue squares and asterisks ) on average better fitted with qso templates .
crosses and asterisks represent agns with stellar morphology .
because the number of the galaxy templates is larger than that of the qso templates and the average @xmath178 value for extended narrow - line agns with the galaxy templates is smaller than that for stellar broad - line agns with qso - templates , we do not chose the best photometric redshift just by comparing @xmath179 and @xmath180 , but consider the stellarity as well as the @xmath178 values . considering the distribution in the panel , we regard the qso templates as better if @xmath179 is smaller than @xmath181 for stellar objects . on the contrary , for extended objects ,
we regard the gal - template is better if @xmath180 is smaller than @xmath182 .
these selection criteria are shown with the dashed line in figure [ fig : hyperz_chi2 ] .
based on the comparison , we constrain not only their redshifts but also their agn types .
we classify an object whose photometric redshift is determined with the galaxy ( qso ) templates as narrow - line ( broad - line ) agn , hereafter .
we can examine the reliability of the photometric classification with the template fitting results for the agns with spectroscopic classification .
as described above , we do not fix their redshifts with spectroscopic redshifts during the fitting process . among the 302 narrow - line ( 284 broad - line )
agns with spectroscopic identification , 267 ( 206 ) of them are correctly classified as a narrow - line ( broad - line ) agn in the photometric classification . because less luminous broad - line agns have extended morphology , a fraction of them are fitted better with the galaxy templates and photometrically classified as narrow - line agns .
their optical continuum can be dominated by the host galaxy component , see section 7.3 .
those mis - classifications lower the rate of correct classification for the broad - line agns .
if we concentrate on the 180 broad - line agns with stellar morphology , 170 are correctly classified as a broad - line agn in the photometric classification .
the derived photometric redshifts and classification are summarised in columns 24 and 25 of table [ tab : multi3 ] , respectively .
column 26 describes flags for the photometric redshift determination : flag=1 means the reduced @xmath161 is larger than 10 , flag=2 means photometric redshift is determined with unreliable photometric data for a bright and extended galaxy , and flag=3 means photometric redshift is determined with less than four bands .
the number of x - ray agns with each flag are summarised in table [ tab : number_summary ] . for all agns ( top panel ) , broad - line agns ( middle ) , and narrow - line agns ( bottom).,width=283 ] the reliability of the photometric redshifts
is examined by comparing them with the spectroscopic values .
the comparison of the photometric and spectroscopic redshifts of spectroscopically - identified agns is shown in figure [ fig : hyperz_zsp_zph ] .
we separate the sample into 4 categories using the stellarity and the best - fit template .
the difference between the spectroscopic and photometric redshifts is defined as @xmath183 , and the bottom part of the panels show @xmath184 .
photometric redshifts for objects that are fitted better with galaxy templates have a small scatter ( shown in left panels ) , but the scatter is larger for objects fitted better with qso templates ( shown in right panels ) .
the distribution of @xmath184 is shown in figure [ fig : specz_photz_hist ] .
there is no systematic difference between the spectroscopic and photometric redshifts ; the median of of @xmath185 is @xmath186 for all the agns in the total sample , and @xmath187 and @xmath188 for broad - line and narrow - line agns , respectively .
the scatter of the difference is examined with the normalized median absolute deviation ( nmad ; @xmath189 ) following @xcite , @xmath190 @xmath191 corresponds to the standard deviation of a gaussian distribution , and is less sensitive to the outliers than the usual definition of the standard deviation @xcite . for the total sample ,
@xmath191 is 0.098 , which is larger than that of the photometric redshift estimations for x - ray - selected agns with medium band filters ( @xcite ; @xcite ; @xcite ) . for example , @xcite reports @xmath191 of 0.059 for a blind test sample .
because we do not train the photometric redshift code with seds obtained in the spectroscopic redshift sample , our @xmath191 should be compared with their blind test results .
the larger @xmath191 can be explained by the lack of the photometric data with medium band filters .
the @xmath191 for broad - line agns ( 0.109 ) is larger than that for narrow - line agns ( 0.072 ) , as expected from figure [ fig : specz_photz_hist ] .
this is because there is no strong feature in the seds of the broad - line agns except for the break below ly@xmath146 .
since the break can only be covered by deep optical observations for agns at @xmath192 , the discrepancy between @xmath193 and @xmath194 is smaller for agns at @xmath192 .
photometric redshift determination is crucial to understand the nature of the x - ray - selected agns .
the redshift distribution of the agn sample is summarised in figure [ fig : redshift_hist ] . for narrow - line agns ,
the redshift distribution of spectroscopically - identified objects is different from that of objects without spectroscopic identification : they are located at higher redshifts on average .
as shown in the right panel of figure [ fig : sxds_cstar_i ] , among the 310 primary candidates without spectroscopic identification , 258 objects have extended morphology .
photometric redshift estimation suggests that they are narrow - line agns at @xmath195 .
the green lines in the figure indicate the redshift evolution of the @xmath2-band magnitude of 10gyr ( solid ) and 2gyr ( dashed ) single burst stellar population with stellar mass of @xmath196 ( upper ) and @xmath197 ( lower ) .
the @xmath2-band magnitudes of most of the spectroscopically - identified narrow - line agns are enclosed within the green lines , and the @xmath2-band magnitudes of narrow - line agns with photometric redshifts are also enclosed within the lines .
the photometric redshift results suggest that there are 45 broad - line agn candidates that have not been spectroscopically identified , and 86% of the broad - line agns are thought to be spectroscopically identified . among broad - line agn candidates without spectroscopic identification , 7 objects have a photometric redshift above 3 , where photometric redshift estimation is reliable even for broad - line agns .
vs. @xmath198 estimated simply assuming a power - law with @xmath30 .
blue squares and red circles represent broad - line and narrow - line agns , respectively .
open symbols represent objects that are not detected in the 210 kev band and @xmath198 is estimated from the 0.52 kev count rate.,width=283 ] in order to examine the physical properties of the x - ray agns , at first , we evaluate the intrinsic 210 kev luminosity , @xmath198 ( erg s@xmath8 ) , and the column density of photo - electric absorption to the nucleus , @xmath167 ( @xmath9 ) , of each agn . in order to evaluate @xmath198 ,
it is necessary to correct for the effect of photo - electric absorption using the @xmath167 value .
the number of x - ray photons of each agn in the current sample is not sufficient to evaluate @xmath167 for individual agn accurately with spectral fitting .
thus we estimate @xmath167 of each agn from its hr2 value .
it needs to be noted that the relation between the hr2 and @xmath167 is non - linear @xcite , and the uncertainty associated with @xmath167 can be large , especially around hr2 of @xmath199 . furthermore , the relation varies with redshift : at higher redshifts @xmath167 is steeper function of hr2 , and thus the uncertainty is larger than the agns at low redshifts with same uncertainty of hr2 . due to these limitations ,
a statistical analysis , like that applied in @xcite and @xcite , is necessary to derive the distribution function of @xmath167 and the hard x - ray luminosity function .
therefore , in this paper , we use @xmath198 and @xmath167 to evaluate average properties of a population of agns ( section 7.2 ) and to construct statistical sample of agns sub - divided by their x - ray properties ( section 7.3 ) . in order to estimate @xmath167 from hr2 ,
we assume an intrinsic shape of the x - ray spectrum of agns following @xcite . the intrinsic x - ray spectrum of agns is modeled by a combination of a power - law component with high - energy cut off @xmath200 and a reflection component . a photon index @xmath201 of 1.9 and
cutoff energy @xmath202 of 300 kev are assumed .
we calculate the reflection component with the `` pexrav '' @xcite model in the xspec package applying a solid angle of 2@xmath203 , an inclination angle of @xmath204 , and solar abundance of all elements .
the strength of the reflection component is about 10% of the direct component just below 7.1 kev . for the assumed intrinsic x - ray spectrum
, we ignore any scattered light component and the luminosity dependence of the strength of the reflection component for simplicity .
we only consider @xmath205 up to 24.0 . for compton - thick agns
whose x - ray spectrum is affected by a scattered component , @xmath167 can be underestimated due to the soft x - ray spectrum of that component . however , the number of such agns should be small in the x - ray flux range covered by the current sample @xcite .
if hr2 is smaller than that expected for @xmath201=1.9 , we determine the corresponding photon index by assuming no absorption .
if an object is only detected in the soft - band , we assign @xmath201=1.9 and we neglect their @xmath205 . for objects only detected in the hard - band ,
we assign @xmath206 and @xmath201=1.9 .
@xmath198 is derived from the observed 210 kev count - rate assuming the intrinsic spectrum discussed above and correcting for the effect of the photo - electric absorption with the best estimated @xmath167 . for objects only detected in the 0.52 kev band ,
the count - rate in that band is used instead of the 210 kev count - rate to derive @xmath198 .
the best - estimated @xmath198 are compared with @xmath198 estimated simply assuming typical apparent power - law spectra of non - absorbed agns with @xmath30 in figure [ fig : sxds_lum_comp ] .
we use the count - rate in the 210 kev band in the estimation , but for objects only detected in the 0.52 kev band , we use the count - rate in the 0.52 kev band and show them with open symbols . because the observed 210 kev energy range covers higher energy range in the object rest - frame , for most of the x - ray sources , the amounts of absorption correction are not large .
therefore , the statistical conclusions derived using the best - estimated @xmath198 in later sections do not significantly change even if we use the @xmath198 derived simply with @xmath30 . the best - estimated @xmath198 and @xmath167 are shown as a function of redshift in the left and right panels of figure [ fig : sxds_red_lum ] . in the right panel , we show the uncertainties of @xmath205 for objects with @xmath207 .
the uncertainties are determined by medians of uncertainties for objects with @xmath208 at @xmath209 , @xmath210 , @xmath211 , and @xmath212 .
uncertainty of each object is determined with the uncertainty of hr2 .
the sxds agn sample covers wide range of the hard x - ray luminosity , from @xmath213 to @xmath214 erg s@xmath8 , and encloses the knee of the hard x - ray luminosity function at @xmath215 ( @xmath216 at @xmath217 and @xmath218 at @xmath219 ; @xcite ) .
there is an apparent relation between the redshift and luminosity due to the flux limit of the sample .
the distribution of @xmath167 is broadly consistent with the optical classification on the basis of the existence of broad - emission lines or by sed modeling : broad - line agns have smaller @xmath167 on average . in the redshift range below 0.6 , almost all of the agns with @xmath220 lack broad - emission lines . for these agns , it is likely that the broad - emission line component may be diluted by the host galaxy component . in table
[ tab : xrayprop ] , columns 2 and 3 show the redshift and hr2 values assumed to derive the best - estimated @xmath205 , @xmath201 , and @xmath221 .
derived values are shown in columns 4 - 6 .
@xmath0 in the column 4 means @xmath205 is not given because hr2 is smaller than expected for a power - law of @xmath201=1.9 or the object is only detected in the soft - band .
we derive the stellar mass of host galaxies of narrow - line agns , @xmath3 , by assuming that their uv to nir seds are dominated by the host galaxy component .
the similarity of the seds of narrow - line agns to those of non x - ray galaxies ( e.g. @xcite ) supports this assumption .
the similarity will be discussed further in section 7.3 .
we apply sed template fitting with a fixed redshift which is determined either spectroscopically or photometrically , and we use larger number of sed templates than those used in the photometric redshift estimation in order to cover the variety of the seds of galaxies accurately . furthermore
, the effect of hot dust component above 1 wavelength range is considered in the templates as described below .
we construct the sed templates using pegase.2 stellar population synthesis code @xcite . a salpeter imf with mass range of 0.1 - 120 @xmath224 is assumed .
sed templates with burst and continuous star formation histories are constructed with a simple exponentially declining star formation rate proportional to @xmath225 .
@xmath226 represents the age of the stellar system and @xmath162 represents the decay time scale of the star formation .
we use @xmath162 of 0.5gyr and 2.0gyr for burst and continuous models .
we use ages of 1.0 , 2.0 , 2.5 , 3.0 , 3.5 , 4.0 , 4.5 , 5.0 , 6.0 , 8.0 , and 10.0 gyrs for the burst model and 0.5 , 1.0 , 2.0 , 3.0 , 4.0 , 5.0 , 6.0 , 8.0 , and 10.0 gyrs for the continuous model .
we considered three metallicities , 0.5 , 1.0 , and 2.0 times the solar metallicity .
we apply reddening with the calzetti extinction curve @xcite and consider color excess @xmath227 with 0.1 mag steps up to 0.9 mag .
in addition to the simple exponentially declining and continuous models , we construct templates by combining the two models with a stellar mass ratio of 0.0 , 1.0 , and 50.0 , representing the mixture of the bulge and disk systems of galaxies . in the wavelength range above 1 ,
a significant fraction of narrow - line agns show an excess compared to the galaxy templates due to hot dust component . in order to reproduce the excess , we construct galaxy templates with an additional power - law component representing the excess above 1 .
figure [ fig : sxds_sedfit_fix_sel ] shows samples of the sed fitting results for narrow - line agns at @xmath139 in order of the strength of the excess component from the top - left panel with no excess to the bottom - right panel with the strongest excess .
the excess components with various strength can be seen above observed wavelength of 2 are well reproduced with the best - fit templates .
furthermore , for galaxies at @xmath228 , the 8.0 photometry is affected by the 8 pah feature , which is not included in the galaxy templates .
we do not use 8.0 photometry for galaxies at @xmath228 .
we do not consider the constraints on the absolute magnitude in the fitting .
, green squares are for @xmath229 , and red circles are for @xmath230 .
spectroscopically identified narrow - line agns are plotted with large open symbols , and photometric narrow - line agn candidates are plotted with small filled symbols .
photometric candidates with phot - z flags of 0 or 1 are plotted .
big crosses indicate the median values in the three redshift ranges .
solid lines indicate the stellar mass and 210 kev luminosity relation assuming the @xmath231-@xmath232 relation from @xcite and bolometric correction of @xcite . from bottom to top , an eddington ratio of 1 , 0.1 , 0.01 , and 0.001 are assumed .
dotted and dashed lines indicate the relations derived with the @xmath231-@xmath233 relations of @xcite and @xcite , respectively , for an eddigton ratio of 1.,width=283 ] for narrow - line agns .
the solid histogram is for the entire sample , and blue dotted , green dashed , and red dot - dashed histograms represent the distribution of @xmath234 , @xmath235 , and @xmath236 narrow - line agns , respectively .
the numbers for the entire sample are divided by 2 for clarity.,width=283 ] derived from @xmath198 and @xmath237-@xmath238 relation @xcite and @xmath231 derived from @xmath232 and local @xmath231-@xmath232 relation @xcite .
symbols are coded with redshift ; blue triangles are for @xmath239 , green squares are for @xmath229 , and red circles are for @xmath230 .
spectroscopically identified narrow - line agns are plotted with large open symbols , and photometric narrow - line agn candidates are plotted with small filled symbols .
photometric candidates with phot - z flags of 0 or 1 are plotted .
big crosses indicate the median values in the three redshift ranges .
dotted line indicate the redshift dependence of the difference with cosmological evolution of the @xmath231-@xmath232 relation with @xmath240 @xcite.,width=283 ] the resulting stellar masses of host galaxies of narrow - line agns are shown in figure [ fig : sxds_lum_mass ] .
although the sxds sample covers a wide range in luminosity and redshift , the estimated @xmath232 values are remarkably constant . if we divide the sample into 3 redshift bins with @xmath234 ( low ) , @xmath235 ( mid ) , and @xmath236 ( high ) , the median @xmath232 increases by 0.18 dex from the low to the high redshift bins , although the median @xmath198 increases by 1.4 dex .
the distributions of the @xmath232 in the 3 redshift bins are compared with each other in figure [ fig : sxds_mass_hist ] .
the probabilities that a pair of distributions is drawn from the same sample are examined with the kolmogorov - smirnov test , and they are 1.0% , 4.7% , and 28% with the low - mid , low - high , and mid - high redshift samples .
the mass range of the host galaxies is broadly consistent with those for various x - ray selected samples of narrow - line ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) and broad - line agns ( e.g. , @xcite ; @xcite ; @xcite ) at redshifts up to 4 .
it needs to be noted that we consider the absolute magnitude range of the galaxy template in the photometric redshift determination .
the absolute magnitude range can affect the distribution of @xmath232 of narrow - line agns without spectroscopic redshifts , however , only 15 out of the 259 narrow - line agns without spectroscopic redshifts are significantly affected by the faint limit of the range and have absolute magnitude close to the limit .
some narrow - line agns have @xmath232 less than @xmath241 m@xmath177 .
their stellar mass estimation would be affected by scattered nuclear light in the uv . such contamination with a blue continuum
possibly makes the sed appear similar to a younger stellar population with a smaller mass - to - light ratio , and the estimated stellar mass can thus appear smaller even with the rest - frame nir photometry constraints .
the solid lines in figure [ fig : sxds_lum_mass ] indicate the relationship between the stellar mass and x - ray luminosity if we assume the local black hole and host - galaxy stellar mass , @xmath231-@xmath232 , relation from @xcite and the agn bolometric correction for 210 kev x - ray luminosity from @xcite .
the lines from bottom to top correspond to an eddington ratio of 1 , 0.1 , 0.01 , and 0.001 .
@xcite derive the relationship for broad - line agns following @xcite with a chabrier imf @xcite .
we convert their relation to that with the salpeter imf , considering that @xmath232 is 0.255 dex heavier with the salpeter imf .
we also plot the @xmath3 and @xmath198 relation based on the black hole and bulge stellar mass , @xmath231-@xmath242 , relations from @xcite and @xcite with dotted and dashed lines , respectively .
they are consistent with the relation derived with @xmath231-@xmath232 relation of @xcite around @xmath243 m@xmath177 , and they show steeper dependence on @xmath198 than the relation . the difference below @xmath243 m@xmath177 can be explained with the difference between @xmath232 and @xmath242 .
the constant @xmath232 of narrow - line agns with wide luminosity and redshift ranges implies higher @xmath237 in the higher redshift bins ; significant fraction of narrow - line agns at @xmath244 possibly have @xmath198 close to or above the eddington limit . on the other hand ,
the distribution of narrow - line agns on the @xmath232-@xmath198 plane can be explained with the cosmological evolution of the @xmath231-@xmath232 relationship with higher @xmath245 at higher redshifts ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) , and @xmath237 for the narrow - line agns at high - redshifts may be overestimated . assuming that the @xmath237 and @xmath238 relation that is derived with broad - line agns at @xmath246 in the same sxds x - ray sample ( eq.(16 ) of @xcite ) is applicable to the narrow - line agns
, we estimate @xmath247 of the narrow - line agns from @xmath198 and the bolometric correction of @xcite .
figure [ fig : sxds_red_massratio ] shows the difference between the estimated @xmath247 and the @xmath248 , which is calculated from @xmath232 and the local @xmath231-@xmath232 relationship from @xcite .
the difference implies the cosmological evolution of the @xmath231-@xmath232 relationship with larger @xmath231 compared to @xmath232 at higher redshifts can explain the observed distribution on the @xmath232-@xmath198 plane : the @xmath234 narrow - line agns have @xmath249 dex smaller @xmath231 than that expected from @xmath232 , and the @xmath230 narrow - line agns have @xmath250 dex larger @xmath231 .
if we simply fit the distribution , cosmological evolution of @xmath231-@xmath232 relationship with @xmath251 is required .
if we take at face value , the derived cosmological evolution of @xmath231-@xmath232 relation is stronger than that in @xcite shown with dotted line in the figure . in order to derive the cosmological evolution of @xmath231-@xmath232 relationship reliably
, we need to consider the bias associated with the luminosity - limited sample @xcite .
such detailed analysis is beyond the scope of this paper , and will be discussed somewhere else . utilising the multi - wavelength photometric data of the large sample of x - ray selected agns
, we examine the scatter of their uv - mir seds as a function of @xmath198 and @xmath167 , and compare the range of the scatter with that of galaxies without x - ray agn activity .
we choose galaxies which have the same mass and redshift distributions as the x - ray agns .
because x - ray agns only appear among massive galaxies , it is important to compare the properties of agns with those of galaxies without x - ray agn activity in the same mass range ( @xcite ; @xcite ) . we construct a sample of galaxies without x - ray agn activity in the overlapping region of the deep suprime - cam optical and uds nir imaging data . as a first step the photometric redshift
is determined with the fuv to mir photometric data for each object in the same way as for the x - ray agns as described in section 6 .
we do not apply a qso template to the galaxies and do not consider the constraints on the absolute magnitude .
once photometric redshifts are determined , we evaluate the stellar masses of the galaxies applying the model described in section 7.2 . in order to construct the sample of galaxies without x - ray activity , we select galaxies that have similar stellar mass and redshift distribution of the x - ray agns . for each x - ray agn , we select galaxies without x - ray detection within @xmath252 of the redshift of the agn .
we assume broad - line agns follow the same host galaxy stellar mass distribution shown in figure [ fig : sxds_mass_hist ] . within the redshift range ,
we randomly select 14 galaxies for each agn following the mass distribution of 1 in the range @xmath253 , 2 in the range @xmath254 , 5 in the range @xmath255 , 5 in the range @xmath256 , and 1 in the range @xmath257 .
the mass distribution roughly reproduces that of the total sample of the narrow - line agns . in the redshift range between 1.0 and 2.0 , there are 12248 ( @xmath253 ) , 9140 ( @xmath254 ) , 5899 ( @xmath255 ) , 2557 ( @xmath256 ) , and 310 ( @xmath257 ) galaxies that are not detected in x - ray in the mass range shown in the parenthesis .
we allow overlap in the selection , and a galaxy can be selected multiple times .
we refer to this sample of galaxies as the matched sample , hereafter .
the seds of x - ray agns are shown with dots in figure [ fig : sed_plot_nh1 ] . from top to bottom , the sample is subdivided by x - ray luminosity : @xmath258 , @xmath259 , and @xmath260 , and from left to right , the sample is subdivided by x - ray absorption to the nucleus : @xmath261 and @xmath262 .
blue and red symbols represent broad- and narrow - line agns , respectively .
all of the seds are normalized at 1 , and each photometric point is located at its rest - frame effective wavelength .
it is clear that the luminous and less - absorbed agns show seds within the range of the seds of the optically - selected qsos ( green dashed lines ) .
less - luminous or absorbed agns show redder uv - optical seds , which are similar to those of galaxies in the matched sample .
the thick black solid and dotted lines in the figure show the median ( 50%-ile ) value of the 1-normalized seds of the galaxies in the matched sample .
we construct the matched sample for each luminosity bin . in order to determine the median value at each wavelength , we linearly interpolate the photometric data points of each galaxy in the @xmath263 - @xmath264 plane . in the percentile calculation
, we include photometric data points which are upper limits . in the wavelength range
where less than 50% of galaxies in the sample are detected , the derived median value is an upper limit , and we plot this with the dotted line .
the thin black solid and dotted lines represent 10%-ile and 90%-ile from top to bottom at each wavelength range . in order to quantitatively
compare the range of the seds , we also construct the percentile seds of x - ray agns and show them in figure [ fig : sed_plot_nh2 ] .
the sed of each agn is interpolated in the same way as for the galaxies , and is normalized at 1.0 . at each wavelength
we derive from brightest 10%-ile to 90%-ile seds with 10% interval . if the fraction of detected objects at a certain wavelength is less than the corresponding percentile , we regard the percentile sed as an upper limit and plot the sed with dotted lines . at wavelengths below 1 where the seds of broad - line agns are dominated by a blue power - law component ,
most of the broad - line agns have bluer seds than the galaxies in the matched sample , especially in the upper - left highest luminosity and lowest absorption bin : 80% of the x - ray agns have bluer seds than the median sed of the galaxies in the matched sample .
the almost constant median sed of the most luminous less - absorbed agns is consistent with the median pre - host subtraction seds of broad - line agns derived with the cosmos dataset ( @xcite ) .
if we compare 90%-ile range of @xcite , the upper envelope is consistent , but the current sample has larger scatter in the lower envelope due to the inclusion of narrow - line agns .
the seds of x - ray agns become redder with lower luminosity or higher @xmath167 . in the lowest luminosity bin ,
even with the low @xmath167 , the median sed of the agns is similar to that of galaxies in the matched sample .
as can be seen in figure [ fig : sed_plot_nh1 ] , almost all of them are narrow - line agns , and they do not show strong broad - emission lines in their observed spectra .
the median seds of the absorbed ( @xmath265 ) x - ray agns have similar seds to those of the galaxies in the matched sample .
for example , the rest - frame @xmath266 color of the median sed of agns with @xmath265 and @xmath259 is only 0.05 mag bluer than the galaxies in the matched sample .
the similarity of the @xmath266 colors of the agn host galaxies and the galaxies in the matched sample is consistent with the results obtained in the _ chandra _ deep fields @xcite , and does not support the distribution of agn host galaxies in a distinct region of color - magnitude space ( e.g. , @xcite ) . in the wavelength range above 1 ,
the seds of the agns are thought to be dominated by @xmath267 1,000k hot thermal emission from a dusty torus , and are roughly constant on the @xmath268 plane . on the other hand ,
the seds of galaxies without x - ray agn activity are dominated by the stellar continuum , which shows a 1.6 bump and continuous decrease at longer wavelengths .
pah 8 and cold - dust emissions from star - forming regions contribute in the wavelength range longward of 5 .
as can be seen in figure [ fig : sed_plot_nh1 ] , agns with @xmath269 are distributed around a constant sed above 1 for both the @xmath261 and @xmath270 samples , and the seds are consistent with the observed range of broad - line agns shown with green dashed lines . for agns with lower luminosity , the contribution from the hot dust component decreases and the seds are bluer and closer to those of galaxies peaked at around 1.6 . the galaxies in the matched sample
do not show an excess component above 1 in the lowest luminosity bin , which samples galaxies in the low redshift range , shown in the bottom panels . on the contrary in the highest luminosity bin , which samples galaxies in the higher redshift range , shown in the top panels , the median sed of the galaxies in the matched sample show a constant sed in this wavelength range . as a result , even though in the middle and high luminosity ranges more than 70 - 80% of x - ray agns show excess component compared to the median sed of the galaxies in the matched sample , the 10%-ile seds of x - ray agns are similar to those of the galaxies in the matched sample .
several agn selection methods based on the detection of hot thermal emission by multi - band photometric data are proposed with the _ spitzer _
( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) , _ akari _ ( @xcite ; @xcite ) , and _ wide - field infrared survey explorer ( wise ) _ ( @xcite ) dataset .
the similarity of the seds of agns with @xmath271 to those of the galaxies in the matched sample suggests that agn selection based on detecting @xmath272 excess emission above the stellar continuum can be contaminated by galaxies without x - ray agn activity at @xmath273 ( e.g. @xcite ; @xcite ; @xcite ) .
it needs to be noted that the seds above rest - frame 3 in the current study are based on broad - band photometry only at 8 and 24 , and detailed continuum structure can be missed .
the constant seds of galaxies without x - ray agn activity at @xmath273 can be caused by high star - formation activity associated with massive galaxies whose star - formation activity is strongly affected by dust .
the cosmological evolution of the @xmath232 - star - formation rate relation suggests galaxies with @xmath196 at @xmath274 and @xmath275 have 7 and 24 times higher star - formation rate than those at @xmath276 on average @xcite .
furthermore , star - formation activity in massive galaxies is strongly affected by dust @xcite .
@xcite show seds of ir - excess galaxies can be reproduced with a sed of local star - forming galaxy with additional reddening with @xmath277 mag . a larger contribution from thermal dust emission than for local massive galaxies can explain the constant seds of the non - x - ray galaxies at @xmath273 @xcite .
such evolution of seds of galaxies may affect the discussion of mir seds of torus components of the broad - line agns @xcite .
we report the multi - wavelength identification of the x - ray sources found in the sxds field using deep imaging data covering the wavelength range between far - uv to mid - ir . among the 1045 x - ray sources detected in the 7 fovs of _ xmm - newton _
@xcite , we consider 945 x - ray sources that are detected above likelihood 7 in either of the soft- or hard - band and covered by the deep suprime - cam imaging observations .
we select a primary counterpart for each x - ray source applying the likelihood ratio method to @xmath1-band , 3.6 , near - uv , and 24 source catalogs .
we also consider the matching catalogs with agn candidates selected in 1.4ghz radio and @xmath2-band variability surveys .
once candidates of galactic stars , ulxs in a nearby galaxy , and clusters of galaxies are removed there are 896 agn candidates in the sample .
we conduct spectroscopic observations of 596 and 851 primary counterparts in the optical and nir wavelengths , respectively , and spectroscopically identify 65% , 597 , of the total sample . for the remaining x - ray agn candidates ,
we evaluate their photometric redshift with photometric data in 15 bands . in the photometric redshift determination
, we apply sed templates of qsos and galaxies separately to each primary counterpart , and select the best - fit photometric redshift considering the @xmath2-band stellarity .
utilising the multi - wavelength photometric data of the large sample of x - ray selected agns , we examine @xmath3 of the host galaxies of the narrow - line agns .
the median @xmath3 in the three redshift bins , @xmath234 , @xmath235 , and @xmath236 , only increases 0.18 dex from the low to the high redshift bins , although the median @xmath198 increases 1.4 dex in the redshift range .
if we consider the @xmath231-@xmath3 relation in the local universe , the constant @xmath3 implies higher @xmath237 for luminous agns at high redshifts . on the other hand ,
if we assume the @xmath237 and @xmath238 relation derived for broad - line agns , we can estimate @xmath231 of the narrow - line agns .
the constant @xmath3 can also be explained with strong redshift evolution of @xmath231-@xmath3 relation from that determined in the local universe .
finally , the scatter of the rest - frame uv - mir seds of x - ray agns is examined as a function of @xmath198 and @xmath167 .
the scatter is compared with the scatter of the seds of the galaxies without x - ray detection .
the comparison sample of galaxies are selected to have redshift and @xmath3 distributions matched to the x - ray agns .
the uv - nir seds of x - ray agns with @xmath265 are remarkably similar to those of the galaxies in the matched sample . in the nir - mir wavelength range ,
the median seds of x - ray agns are redder than those of the galaxies in the matched sample , and luminous agns show constant median sed in the wavelength range , suggesting the existence of the hot dust component .
however , the scatter of the seds of the luminous x - ray agns broadly overlap with that of the galaxies in the matched sample .
the galaxies in the matched sample of the luminous agns also show a more constant sed than the galaxies in the matched sample of the less - luminous agns , which are at lower redshifts than the luminous agns on average .
a higher star - formation rate in massive galaxies at @xmath278 can explain the constant sed of the galaxies in the matched sample . due to the similarity of the seds of x - ray agns to those of the galaxies in the matched sample
, it is expected that agn selection criteria looking for excess in the nir - mir wavelength range can be significantly contaminated by non - agn galaxies .
lrrrl & n@xmath279 & n@xmath280 & n@xmath281 & comment + x - ray sources & 1045 & 866 & 645 & listed in @xcite + x - ray sample & 945 & 781 & 584 & within deep suprime - cam image coverage + covered by 1.4ghz radio survey & 789 & 655 & 497 & covered in @xcite + clusters of galaxies candidates & 14 & 14 & 1 & see table [ tab : cluster ] + galactic star candidates & 31 & 30 & 4 & see table [ tab : star ] + ulx candidates & 4 & 4 & 3 & see section [ sec : ulx ] + agn candidates & 896 & 733 & 576 & + & & & & + + optical spec . observed & 597 & 523 & 399 & + nir spec .
observed & 851 & 704 & 548 & + spec .
identified & 586 & 514 & 397 & + photz flag=1 & 1 & 1 & 0 & + photz flag=2 & 10 & 10 & 7 & + photz flag=3 & 12 & 6 & 7 & + _ galex _ & fuv & 26.64 & 25.52(4@xmath14,0.36 ) & 0.18 & deep imaging survey + & & & & & + _ galex _ & nuv & 26.18 & 25.15(3@xmath14,0.59 ) & 0.17 & deep imaging survey + & & & & & + subaru / suprime & @xmath38 & 27.39 & 26.46(3@xmath14,0.49 ) & 0.11 & south fov only + ctio / mosaic - ii & @xmath38 & 26.86 & 26.29(3@xmath14,0.13 ) & 0.11 & + cfht / mega - cam & @xmath38 & 26.65 & 26.11(3@xmath14,0.10 ) & 0.11 & cfhtls wide + int / wfcam & @xmath38 & & & 0.11 & only used for bright objects + & & & & & + subaru / suprime & @xmath32 & 28.50 - 28.83 & 27.91 - 28.24(3@xmath14,0.15 ) & 0.09 & + int / wfcam & @xmath44 & & & & only used for bright objects + & & & & & + subaru / suprime & @xmath33 & 28.09 - 28.22 & 27.55 - 27.68(3@xmath14,0.10 ) & 0.07 & + & & & & & + subaru / suprime & @xmath1 & 27.95 - 28.18 & 27.38 - 27.61(3@xmath14,0.13 ) & 0.07 & + int / wfcam & @xmath1 & & & & only used for bright objects .
+ & & & & & + subaru / suprime & @xmath2 & 27.91 - 28.10 & 27.30 - 27.49(3@xmath14,0.17 ) & 0.06 & + int / wfcam & @xmath2 & & & & only used for bright objects .
+ & & & & & + subaru / suprime & @xmath34 & 26.83 - 27.08 & 26.25 - 26.50(3@xmath14,0.14 ) & 0.04 & + int / wfcam & @xmath34 & & & & only used for bright objects .
+ & & & & & + ukirt / wfcam & @xmath48 & 26.06 & 25.40(3@xmath14,0.22 ) & 0.02 & ukidss / uds - dr8 + & & & & & + ukirt / wfcam & @xmath49 & 25.48 & 24.82(3@xmath14,0.22 ) & 0.02 & ukidss / uds - dr8 + & & & & & + ukirt / wfcam & @xmath15 & 25.96 & 25.30(3@xmath14,0.22 ) & 0.01 & ukidss / uds - dr8 + & & & & & + _ spitzer_/irac & 3.6 & 23.54 & 22.77(3@xmath14,0.33 ) & 0.01 & swire + _ spitzer_/irac & 3.6 & 25.26 & 24.49(3@xmath14,0.33 ) & 0.01 & spuds + & & & & & + _ spitzer_/irac & 4.5 & 22.87 & 22.07(3@xmath14,0.36 ) & 0.00 & swire + _ spitzer_/irac & 4.5 & 24.83 & 24.03(3@xmath14,0.36 ) & 0.00 & spuds + & & & & & + _ spitzer_/irac & 5.8 & 21.27 & 19.97(4@xmath14,0.54 ) & 0.00 & swire + _ spitzer_/irac & 5.8 & 22.99 & 21.69(4@xmath14,0.54 ) & 0.00 & spuds + & & & & & + _ spitzer_/irac & 8.0 & 21.17 & 19.91(4@xmath14,0.66 ) & 0.00 & swire + _ spitzer_/irac & 8.0 & 23.19 & 21.77(4@xmath14,0.66 ) & 0.00 & spuds + & & & & & + _ spitzer_/mips & 24 & 20.21 & 18.22(4@xmath14,1.23 ) & 0.00 & swire + _ spitzer_/mips & 24 & 22.06 & 20.07(4@xmath14,1.23 ) & 0.00 & spuds + we warmly thank the staff members of the subaru telescope for their support during the observations .
we would like to thank the anonymous referee for valuable comments that improved this paper .
ma is supported by jsps grant - in - aid for young scientist ( b ) ( 18740118 ) , and ma and hh are supported by grant - in - aid for scientific research ( b ) ( 21340042 ) .
hh thanks the grant - in - aids for challenging exploratory research ( 24650145 ) from jsps , which partially supported this research .
this publication makes use of data products from the two micron all sky survey , which is a joint project of the university of massachusetts and the infrared processing and analysis center / california institute of technology , funded by the national aeronautics and space administration and the national science foundation .
based on observations obtained with megaprime / megacam , a joint project of cfht and cea / irfu , at the canada - france - hawaii telescope ( cfht ) which is operated by the national research council ( nrc ) of canada , the institut national des science de lunivers of the centre national de la recherche scientifique ( cnrs ) of france , and the university of hawaii .
this work is based in part on data products produced at terapix available at the canadian astronomy data centre as part of the canada - france - hawaii telescope legacy survey , a collaborative project of nrc and cnrs .
aird , j. , nandra , k. , georgakakis , a. , laird , e.s . ,
steidel , c.c . ,
sharon , c. 2008 , , 387 , 883 aird , j. , et al .
2010 , , 401 , 2531 akiyama , m. 2005 , , 629 , 72 alexander , d.m . , et al .
2011 , , 738 , 44 alonso - herrero , a. , p@xmath282rez - gonz@xmath283lez , p.g . ,
rieke , g.h . ,
alexander , d.m . ,
rigby , j.r , papovich , c. , donley , j.l .
, & rigopoulou , d. 2008 , , 677 , 127 barger , a.j . , cowie , l.l . ,
mushotzky , r.f . , yang , y. , wang , w .- h . , steffen , a.t . ,
capak , p. 2005
, , 129 , 578 bennert , v.n . , auger , m.w .
, treu , t. , woo , j .- h . , &
malkan , m.a .
2011 , , 742 , 107 bertin , e. & arnouts , s. 1996 , , 117 , 393 bolzonella , m. , miralles , j .-
pello , r. , 2000 , , 363 , 476 bongiorno , a. , et al .
2014 , , 443 , 2077 bradshaw , e.j . ,
almaini , o. , hartley , w.g . , et al .
2013 , , 433 , 194 brammer , g.b . ,
van dokkum , p.g . , & coppi , p. 2008
, , 686 , 1503 brusa , m. , et al .
2009 , , 507 , 1277 brusa , m. , et al .
2010 , , 716 , 348 bruzual , a.g .
, & charlot , s. , 1993 , , 405 , 538 calzetti , d. , armus , l. , bohlin , r.c . ,
kinney , a.l . ,
koornneef , j. , storchi - bergmann , t. 2000 , , 533 , 682 cappelluti , n. , et al .
2009 , , 497 , 635 cardamone , c.n . ,
urry , c.m . ,
damen , m. , van dokkum , p. , treister , e. , labbe , i. , virani , s. , lira , p. , & gawiser , e. 2008 , , 680 , 130 cardamone , c.n . , et al .
2010 , , 189 , 270 chabrier , g. 2002 , , 567 , 304 chiappetti , l. , et al .
2005 , , in press , astro - ph/0505117 civano , f. , et al . 2012 , , 201:30(21pp ) cowie , l.l . ,
garmire , g.p . ,
bautz , m.w . ,
barger , a.j . ,
brandt , w.n .
, hornschemeier , a.e .
2002 , , 566 , 5 daddi , e. , et al .
2007 , , 670 , 173 de ruiter , h.r . ,
arp , h.c . ,
willis , a.g . 1977 ,
a&as , 28 , 211 decarli , r. , falomo , r. , treves , a. , labita , a. , kotilainen , j.k . , & scarpa , r. 2010 , , 402 , 2453 diaz tello , j. , et al .
2013 , , 771:7 donley , j.l . ,
rieke , g.h . , rigby , j.r .
, & p@xmath282rez - gonz@xmath283lez , p.g .
2005 , , 634 , 169 donley , j.l . , et al .
2012 , , 748 , 142 ebrero , j. , carrera , f.j .
, page , m.j . , silverman , j.d . , barcons , x. , cebellos , m.t .
, corral , a. , della cecca , r. , watson , m.g .
2009 , , 493 , 55 elbaz , d. , et al .
2011 , , 533 , 119 elvis , m. , et al .
2009 , , 184 , 158 elvis , m. , et al .
2012 , , 759:6(20pp ) feruglio , c. , et al .
2008 , , 488 , 417 finoguenov , a. , et al .
2010 , , 403 , 2063 fioc , m. , & rocca - volmerange , b. 1997 , , 326 , 950 fiore , f. , et al .
2003 , , 409 , 79 fiore , f. , et al .
2008 , , 672 , 94 furusawa , h. et al .
2008 , , 176 , 1 georgakakis , a. , rowan - robinson , m. , nandra , k. , digby - north , j. , p@xmath282rez - gonz@xmath283lez , p.g . , &
barro , g. 2010 , , 406 , 420 gunn , j.e .
, stryker , l.l .
1983 , , 52 , 121 hanami , h. , et al .
2012 , , 64 , 70 hao , l. , et al .
2005 , , 129 , 1783 hao , h. , et al .
2014 , , 438 , 1288 h@xmath284ring , n. , & rix , h .- w .
2004 , , 604 , l89 hasinger , g. , miyaji , t. , schmidt , m. , hiroi , k. , ueda , y. , akiyama , m. , watson , m.g .
2012 , , 758 , 49 ivison , r. , et al .
2004 , , 154 , 124 iwamuro , f. , moritani , y. , yabe , k. , et al . , 2012 , , 64 , 59 jahnke , k. , bongiorno , a. , brusa , m. , et al .
2009 , , 706 , 215 kim , d.w . , et al .
2004 , , 600 , 59 kimura , m. , maihara , t. , iwamuro , f. , et al .
, 2010 , , 62 , 1135 kiuchi , g. , ohta , k. , akiyama , m. 2009 , , 696 , 1051 koekemoer , a.m. , et al .
2004 , , 600 , 123 kormendy , j. , & richstone , d. 1995 , , 33 , 581 lacy , m. , et al .
2004 , , 154 , 166 lacy , m. , et al .
2005 , , 161 , 41 la franca , f. , et al . 2005 , , 635 , 864 laird , e.s . , et al .
2009 , , 180 , 102 lawrence , a. , et al .
2007 , , 379 , 1599 leipski , c. , et al .
2005 , a&a , 440 , l5 loaring , n.s . ,
2005 , astro - ph/0507408 lonsdale , c. j. , et al .
2003 , , 115 , 897 lonsdale , c.j . , et al . 2004 , , 154 , 54 luo , b. , et al .
2010 , , 187 , 560 lusso , e. , hennawi , j.f . ,
comastri , a. , zamorani , g. , richards , g.t . , vignali , c. , treister , e. , schawinski , k. , salvato , m. , & gilli , r. 2013 , , 777:86(28pp ) magdziarz , p. , & zdziarski , a. a. 1995 , , 273 , 837 mainieri , v. , et al .
2011 , , 535 , a80 marconi , a. , & hunt , l.k .
2003 , , 589 , l21 marconi , a. , risaliti , g. , gilli , r. , hunt , l.k . ,
maiolino , r. , & salvati , m. 2004 , , 351 , 169 mateos , s. , et al .
2012 , , 426 , 3271 mclure , r.j . ,
jarvis , m.j .
, targett , t.a . ,
dunlop , j.s .
, best , p.m. 2006 , , 368 , 1395 mclure , r.j . , pearce , h.j . ,
dunlop , j.s .
, et al .
2013 , , 428 , 1088 merloni , a. , et al .
2010 , , 708 , 137 mignoli , m. , et al .
2004 , , 418 , 827 miller , g.e . , & scalo , j.m .
1979 , , 41 , 513 morokuma , t. , et al .
2008 , , 676 , 121 morokuma , t. , et al .
2008 , , 676 , 163 morrissey , p. , conrow , t. , barlow , t.a . , et al .
2007 , , 173 , 682 murray , s.s . ,
2005 , astro - ph/0504084 nandra , k. , georgakakis , a. , willmer , c.n.a . ,
cooper , m.c . ,
croton , d.j . ,
davis , m. , faber , s.m . ,
koo , d.c . ,
laird , e.s . , & newman , j.a .
2007 , , 660 , l11 nobuta , k. , et al .
2012 , , 761 , 143 oke , j.b .
1974 , , 27 , 210 ouchi , m. , shimasaku , k. , akiyama , m. , et al .
2005 , , 620 , 1 ouchi , m. , shimasaku , k. , akiyama , m. , et al .
2008 , , 176 , 301 ouchi , m. , et al .
2009 , , 696 , 1164 pannella , m. , et al .
2014 , , in press , arxiv:1497,5972 peng , c.y . ,
impey , c.d . ,
rix , h-.w . ,
kochanek , c.s . ,
keeton , c.r . ,
falco , e.e .
, leh@xmath283r , j. , mcleod , b.a .
2006 , , 649 , 616 richards , g.t . , et al .
2003 , , 126 , 1131 rosario , d.j . , et al .
2013 , , 763 , 59 sani , e. , marconi , a. , hunt , l.k . , & risaliti , g. 2011 , , 413 , 1479 saito , t. , shimasaku , k. , okamura , s. , ouchi , m. , akiyama , m. , yoshida , m. , ueda , y. 2008 , , 675 , 1076 salpeter , e. 1955 , , 121 , 161 salvato , m. , et al .
2009 , , 690 , 1250 schramm , m. , & silverman , j.d .
2013 , , 767 , 13 schulze , a. , & wisotzki , l. 2014 , , 438 , 3422 sekiguchi , k. , et al .
2005 , multiwavelength mapping of galaxy formation and evolution , p.82 silverman , j.d . , et al .
2008 , , 679 , 118 silverman , j.d . , et al .
2009 , , 696 , 396 simpson , c. , martinez - sansigre , a. , rawlings , s. , ivison , r. , akiyama , m. , sekiguchi , k. , takata , t. , ueda , y. , watson , m. 2006 , , 372 , 741 simpsion , c. , rawlings , s. , ivison , r. , et al .
2012 , , 421 , 3060 skrutskie , m.f . ,
cutri , r.m .
, stiening , r. , et al .
2006 , , 131 , 1163 smail , i. , sharp , r. , swinbank , a.m. , akiyama , m. , ueda , y. , foucaud , s. , almaini , o. , croom , s. 2008 , , 389 , 407 stern , j. , & laor , a. 2012 , , 423 , 600 stern , d. , et al .
2005 , , 631 , 163 surace , j.a . , et al . 2005 , `` the swire data release 2 : image atlases and source catalogs for elais - n1 , elais - n2 , xmm - lss , and the lockman hole '' sutherland , w. , & saunders , w. 1992 , , 259 , 413 takagi , t. , et al .
2010 , , 514 , 5 ueda , y. , akiyama , m. , ohta , k. , miyaji , t. 2003 , , 598 , 886 ueda , y. , watson , m.g . , stewart , i.m . ,
akiyama , m. , & schwope , a.d .
2008 , , 179 , 124 ueda , y. , akiyama , m. , hasinger , g. , miyaji , t. , watson , m.g .
2014 , , 786 , 104 watson , m. , roberts , t.p .
, akiyama , m. , ueda , y. 2005 , , 437 , 899 wolstencroft , r.d . ,
savage , a. , clowes , r.g . , macgillivray , h.t . ,
leggett , s.k . , kalafi , m. 1986 , , 223 , 279 xue , y.q . ,
brandt , w.n . ,
luo , b. , rafferty , d.a . ,
alexander , d.m . ,
bauer , f.e . ,
lehmer , b.d . ,
schneider , d.p . , & silverman , j.d .
2010 , , 720 , 368 yamada , t. , et al .
2005 , , 634 , 861 yamada , t. , et al .
2009 , , 699 , 1354 yang , y. , mushotzky , r.f . ,
steffen , a.t . ,
barger , a.j .
, cowie , l.l .
2004 , , 128 , 1501 yan , h. , windhorst , r.a . , rottgering , h.j.a .
, cohen , s.h . ,
odewahn , s.c . ,
chapman , s.c . , & keel , w.c .
, 2003 , , 585 , 67 yencho , b. , barger , a.j .
, trouille , l. , winter , l.m .
2009 , , 698 , 380 yoshida , m. , shimasaku , k. , ouchi , m. , sekiguchi , k. , furusawa , h. , okamura , s. 2008 , , 679 , 269 zheng , w. , et al .
2004 , , 155 , 73 | we report the multi - wavelength identification of the x - ray sources found in the subaru @xmath0 _ xmm - newton _ deep survey ( sxds ) using deep imaging data covering the wavelength range between the far - uv to the mid - ir .
we select a primary counterpart of each x - ray source by applying the likelihood ratio method to @xmath1-band , 3.6 , near - uv , and 24 source catalogs as well as matching catalogs of agn candidates selected in 1.4ghz radio and @xmath2-band variability surveys .
once candidates of galactic stars , ultra - luminous x - ray sources in a nearby galaxy , and clusters of galaxies are removed there are 896 agn candidates in the sample .
we conduct spectroscopic observations of the primary counterparts with multi - object spectrographs in the optical and nir ; 65% of the x - ray agn candidates are spectroscopically - identified . for the remaining x - ray agn candidates , we evaluate their photometric redshift with photometric data in 15 bands . utilising the multi - wavelength photometric data of the large sample of x - ray selected agns
, we evaluate the stellar masses , @xmath3 , of the host galaxies of the narrow - line agns .
the distribution of the stellar mass is remarkably constant from @xmath4 to @xmath5 .
the relation between @xmath3 and 210 kev luminosity can be explained with strong cosmological evolution of the relationship between the black hole mass and @xmath3 .
we also evaluate the scatter of the uv - mir spectral energy distribution ( sed ) of the x - ray agns as a function of x - ray luminosity and absorption to the nucleus .
the scatter is compared with galaxies which have redshift and stellar mass distribution matched with the x - ray agn .
the uv - nir seds of obscured x - ray agns are similar to those of the galaxies in the matched sample . in the nir - mir range , the median seds of x - ray agns are redder , but the scatter of the seds of the x - ray agn broadly overlaps that of the galaxies in the matched sample . |
0.25 in studies of low surface brightness ( lsb ) galaxies have revealed a population of relatively unevolved systems compared to the traditional hubble sequence of spirals ( bothun et al 1997 ) .
although the overall properties of lsb galaxies span a wide range in color , gas content , and metallicity ( see oneil et al 1997 ; mcgaugh & de blok 1997 ) , there is a significant subset that have extremely blue colors , high gas contents , and low metallicities .
these are obvious candidates for being the most unevolved systems in the nearby universe . among these kinds of systems ,
ugc 12695 may be the best example of a very unevolved and quite young lsb galaxy .
previous studies of ugc 12695 have shown it to have a very blue color , low metallicity , high gas mass fraction , and unique morphological structure ( mcgaugh & de blok 1997 ; mcgaugh , schombert , & bothun 1995 ; mcgaugh 1994 ; mcgaugh & bothun 1994 ; klein , et al 1992 ; schneider , et al 1990 ; bottinelli , et al 1990 ; lewis 1987 ) .
this galaxy is relatively isolated with its nearest neighbor ( ugc 12687 ) residing 11.2 arcminutes ( @xmath0 10 galaxy diameters ) away . because of its intriguing properties , ugc 12695 was a high priority target in our hubble space telescope ( hst ) wide field planetary camera-2 ( wfpc2 ) survey of selected lsb galaxies to further study their structure and evolution . in this paper
, we report on the remarkable nature of this galaxy as revealed by these new observations .
section 2 describes the observations and data reduction techniques .
section 3 describes the global properties of ugc 12695 , while section 4 discusses some of the individual star forming regions of the galaxy . in section 5
we discuss the background galaxies discovered within the wfpc2 images , and finally in section 6 we consider the implications of our discoveries .
eight hst wfpc2 images were taken of ugc 12695 on january 7 , 1997 , using all four wfpc2 cameras . the nucleus of ugc 12695 was positioned in the wf3 field .
initial image processing occurred using the stsdas pipeline , followed by additional reduction to eliminate cosmic rays and then to combine the images .
the wfpc2 consists of three wide field cameras and one planetary camera .
the wide field cameras have a focal ratio of f/12.9 and a field of view of 80 x 80 with each pixel subtending 0.0996 arcsec .
the three cameras form an l - shape , with the planetary camera completing the square .
the planetary camera has a focal ratio of f/28.3 , a scale of 0.0455 arcsec / pixel , and an overall field of view of 36 x 36 .
each camera has an 800 x 800 pixel silicon ccd with a thermo - electric cooler to suppress dark current .
the wfpc2 has two readouts formats
single pixel resolution ( full mode ) and 2x2 pixel binning ( area mode ) .
all images taken through the f814w ( 814 ) filter were taken in full mode while all images taken through the f300w ( 300 ) filter were taken in area mode .
images of ugc 12695 were taken through both the 814 and the 300 filter .
the 814 filter is a broadband filter with @xmath6 = 7924 and @xmath7 = 1497 .
it is designed to be similar to the johnson i - band filter .
the 300 filter has @xmath6 = 2941 and @xmath7 = 757 , and is designed to be similar to the johnson u - band filter .
all images had 500s exposure times .
as the sensitivity level through the 814 images was considerably higher than through the 300 images ( due to the ccd response ) , and ugc 12695 tended to be much brighter through the 814 , all galaxy identification and analysis was done with the 814 images .
when possible , photometry for a galaxy was also done through the 300 filter .
sky flat fields of the sunlit earth were taken through each filter and routinely calibrated against an internal flat field calibration system .
the internal system consists of two lamps ( optical and uv ) illuminating a diffuser plate .
the internal flats are used to monitor and correct for changes in the flat fields .
dark frames are averages of ten calibration images taken over the space of two weeks .
the intrinsic dark rate of the wfpc2 ccds is @xmath10.01 e@xmath8/pixel / sec .
the calibration dark fields were scaled to the exposure time of each image .
a bias field was generated for each image using extended register pixels which do not view the sky .
the data reduction process was as follows : first , all known bad pixels were removed , using the static mask reference file . the bias level , calculated as described above ,
was then removed from each frame .
the bias image , generated to remove any position - dependent bias pattern , was then subtracted from the image , as was the dark field image ( described above ) .
flat field multiplication was then performed , using the fields described above .
all the preceding image calibration was performed at stsci using the standard wfpc2-specific calibration algorithms ( the pipeline ) .
four images of 500s each were taken through each filter .
after the images were reduced , they were inspected for obvious flaws such as filter ghosts or reflections . each frame
was then shifted , registered and combined , using the stsdas crrej task to eliminate cosmic rays and other small scale flaws ( @xmath9 = 10,8,6 ) .
the resultant images were checked by eye to ensure any registration errors were under 0.5 pixel .
the intensities were then reduced by a factor of 4 ( using the iraf imarith procedure ) to give the resultant image with mean intensity value from the four combined images .
finally , the images were mosaicked , using the stsdas wmosaic task to create a complete wfpc2 image which is shown in figure [ fig : wfpc ] . the total flux was conserved in the mosaic step .
background galaxy identification was done within the confines of the iraf environment .
each image was enlarged by a factor of four and scanned by eye for nonstellar objects . by examining both processed images available for each field a minimum of four times a list was compiled of all possible nonstellar objects in each field which had a minimum diameter of roughly 5 pixels ( 0.5 for the wide field cameras , and 0.23 for the planetary camera ) .
automated galaxy search techniques were not employed for this image , as one ( focas ) had been used on other wfpc2 images and shown to be less reliable than the by - eye search technique ( oneil , et al 1998 ) .
all objects on the list then had their appearance checked against their image in one of the uncombined frames to insure no errors had occurred during the image processing phase ( e.g. image registration errors ) .
remaining objects were considered potential galaxies and left on the list .
it should be noted that because of the low sensitivity in the 300 frames , all galaxy identification was done in the 814 frames .
most of the galaxies were not visible in the 300 frames .
the zeropoints for each field were taken from the photflam value given in the image headers .
the zeropoints from the stmag system ( the space telescope system based on a spectrum with constant flux per unit wavelength set to approximate the johnson system at v ) , is @xmath10 for the 814 filter , the resultant zeropoint was 22.93 , while for the 300 image the zeropoints were 19.43 ( for the pc image ) , and 19.46 ( for the wf images ) .
conversion of the zeropoints for the vega system was done by comparing the stmag values of various objects within the field with the known values in both u and i found from michigan - dartmouth - mit 1.3 m telescope images taken of ugc 12695 .
( see mcgaugh 1992 for details on the ground based images . )
the determined conversion factors were then ( i @xmath11 814 ) = @xmath111.44 @xmath5 0.05 , ( u @xmath11 300 ) = 0.04 @xmath5 0.05 .
it should be noted that this conversion is from the stmag system to the johnson - cousins bands , not from the ` vega ' system used by holtzman et al ( 1995 ) , and others .
the difference in zeropoint conversions to johnson - cousins i - band between these two magnitude systems is substantial for objects with the colors of galaxies .
appendix a provides more information and documentation on this zeropoint difference .
the peak intensity for each galaxy was found and ellipses were fit around that point to obtain the intensity in each annulus using the modified gasp software ( cawson 1983 ; bothun et al 1986 ) . in the amorphous galaxies , like ugc 12695 , the physical center , estimated by centroiding with respect to fitted isophotes , was chosen . in the cases of interacting ( or overlayed ) galaxies ,
the competing galaxy was masked , allowing for a surface brightness profile to be obtained for each of the involved galaxies , when possible .
the core of the point spread function has a radius of 0.1 for the planetary camera , and 0.2 for the wide field camera , and the surface brightness profiles can not be trusted below that .
the average sky - subtracted intensity within each ( annular ) ellipse was found and calibrated with the photometric zeropoint .
exponential surface brightness profiles were then derived using @xmath12 where @xmath13 is the central surface brightness of the disk in linear units ( @xmath14 /pc@xmath15 ) , and @xmath16 is the exponential scale length in arcsec .
this can also be written ( the form used for data analysis ) as @xmath17 where @xmath18 is the central surface brightness in mag arcsec@xmath19 .
galaxy inclination was found using the gasp software to determine the major and minor axis at each isophote .
the inclination angle is @xmath20 due to the asymmetric nature of most of these galaxies the use of a flattening term ( @xmath21 ) as part of the inclination derivation does not seem appropriate .
we estimate that _
i _ is accurate only to within @xmath5 5@xmath22 when determined in this manner .
with a diffuse , y - shaped nucleus and a faint spiral arm , ugc 12695 has a unusual morphology , which is apparent even from ground based images .
additionally , this imaging has shown what appears to be a number of bright star forming knots spread throughout the galaxy ( figure [ fig : gnd ] ) .
morphological classification on these images would suggest that ugc 12695 is not a well - defined spiral and would therefore most likely be classified as a type i irregular galaxy , with sporadic locations of star formation .
we emphasize , however , that ugc 12695 is considerably more massive than the prototypical examples defined by the lmc and ngc 4449 .
furthermore , the hst data has shown that appearances are deceiving as most of these `` star forming knots '' are , in fact , background galaxies clearly shining through the main body of ugc 12695 .
this discovery changes our view of this galaxy considerably .
table [ tab : glob ] lists the global properties of ugc 12695 found primarily from previous studies . in the first row ,
columns 1 and 2 give the coordinates ( ra and dec ) in the j2000 coordinate system found using the wfpc2 image and the stsdas metric task .
column 3 lists the heliocentric radial velocity obtained from the 21-cm observations of bothun et al ( 1985 ) .
column 4 gives our assumed distances based on a hubble constant of 75 mpc / km / s and a virgocentric infall velocity of 300 km s@xmath23 .
column 5 gives the total hi mass of the galaxy , with @xmath24 and column 6 gives the total dynamical mass of ugc 12695 ( @xmath25 ) ) . in column 7
is listed the gas mass to luminosity ratio , while column 8 provides the gas mass fraction for the galaxy , with @xmath26 .
( values listed in columns 68 come from mcgaugh 1992 . )
column 9 lists the exponential scale length in kpc ( mcgaugh & bothun 1994 ) . in column 10 is the galaxy s absolute magnitude ( mcgaugh 1992 ) .
column 11 lists the galaxy s central surface brightness , determined from equation [ eq : mue ] ( mcgaugh & bothun 1994 ) .
finally , columns 12 - 15 give the galaxy s apparent magnitude and total ( luminosity weighted ) colors , also from mcgaugh & bothun ( 1994 ) .
examination of table [ tab : glob ] shows that ugc 12695 is quite blue and rich in gas , suggesting that it is now undergoing one of its first episodes of star formation .
the chance to study this phenomena at significantly higher resolution motivated the hst observations .
the first surprising discovery from the wfpc2 image of ugc 12695 was that many of the objects believed from ground based observations to be stellar knots contained within ugc 12695 are actually background galaxies which are showing _ through _ both the galaxy s disk and outer nucleus .
the significance of this discovery is two - fold first the number of putative star forming regions has been greatly reduced and second ugc 12695 is a remarkably transparent galaxy , even in its nuclear regions . as all prior photometry of ugc
12695 was done assuming the background galaxies to be part of ugc 12695 , we re - determined the galaxy s surface photometry with the galaxies masked ( and the masked pixels replaced with the average values of the surrounding pixels ) .
the results of re - determining ugc 12695 s photometry with the background galaxies masked are shown in the first two rows of table [ tab : col ] . in the first row ( a )
we show the colors derived for ugc 12695 from ground based images , with the background galaxies included in the photometry , while row b shows the colors of ugc 12695 once these galaxies are masked .
( see table [ tab : backg ] to see which galaxies were masked during this process and which are too far from the nucleus of ugc 12695 to have been included in any of the photometric calculations . )
the significance of masking the background galaxies is immediately apparent masking the background galaxies ( which are primarily red , see below ) has made this already very blue galaxy significantly bluer .
in fact , the resultant u - i color of @xmath0 - 0.2 , may make ugc 12695 the bluest known galaxy .
it is remarkable that this blue color is achieved without many obvious large scale regions of star formation being present .
it is thus the smooth , diffuse light of this galaxy , coupled with small scale regions of star formation , which is very blue .
this assertion is documented below .
taking advantage of the high resolution of the wfpc2 images , we further explored ugc 12695 s global properties by masking the six h-@xmath16 star forming regions listed in table [ tab : hcol ] and determined from ground based images taken at the michigan - dartmouth - mit 1.3 m mcgraw - hill telescope .
row c of table [ tab : col ] shows the colors of ugc 12695 with both the known background galaxies and the h-@xmath16 regions removed .
significantly , no measurable difference between rows b and c can be noted .
additionally , figure [ fig : sbprof ] shows the ( 814 band ) surface brightness profile of ugc 12695 with all background galaxies and h-@xmath16 regions included in the analysis ( solid line ) , ugc 12695 with all the definite background galaxies removed ( dashed line ) , and ugc 12695 with both the background galaxies and all the h-@xmath16 regions removed ( dash - dotted line ) .
no differences are apparent in the plots until after r=20 , and no difference in the overall profile of the galaxy can be seen at any radius .
the discovery of background galaxies shining through ugc 12695 dramatically impacts the galaxy s nuclear ( inner 10 ) color .
it should be noted , though , that as these galaxies appear primarily through the 814 ( i band ) filter , and mostly ` dropped out ' when viewed through either the ( ground based ) b band or the 300 images , the removal of the galaxies does not affect the overall magnitude , surface brightness , gas fraction , etc . of ugc 12695 listed in table [ tab : glob ] and found using the blue luminosity .
hence , the primary effect of the mistaken inclusion of these background galaxies in prior studies is in the total i - band luminosity and the derived u - i or v - i colors .
removal of these galaxies significantly lowers these values .
the high resolution of the hst wfpc2 cameras allows for an unprecedented view of the stellar structure of ugc 12695 .
examining the 814 image of ugc 12695 ( figure [ fig : wfpc](a ) ) shows a number of distinct star forming regions surrounding a fairly diffuse , irregular nucleus .
additionally , a spiral arm can be seen encircling the nucleus , which consist of disjoint stellar knots .
figure [ fig : nuc ] shows a greyscale image of the inner 10 of ugc 12695 .
highly non - circular in appearance , the nucleus of ugc 12695 consists of distinct regions of star formation clustered primarily in the northern ( lower ) portion .
other than these localized regions of star formation , even the nuclear regions of ugc 12695 are transparent , to the point were a spiral background galaxy can be identified clearly ( bottom edge of figure [ fig : nuc ] ) .
we do note , however , that this nuclear region is significantly redder than the galaxy as a whole ( see table 2 ) and thus likely contains the bulk of any old stellar population that might be present .
the lsb nature of this galaxy , however , strongly indicates that this older population is distributed in a very diffuse manner and is certainly not concentrated into a central bulge or bar . a number of regions of h-@xmath16 emission are also present within ugc 12695 , many of which are visible in the ground based images ( figure [ fig : gnd]e ) .
photometry of these areas is given in table [ tab : hcol ] , with the detailed areas studied shown in figure [ fig : upict ] .
the majority of these areas are fairly small , with r@xmath27 @xmath12.6 ( 0.8 kpc ) .
( the only exception is area e , which has a radius of 6.4 ( 2 kpc ) . )
the colors of these regions are extremely blue , with 0.1 @xmath1 ( u - i ) @xmath1 0.5 , indicating a significant level of current star formation .
mcgaugh ( 1992 ) performed a metallicity study on a number of ugc 12695 s h-@xmath16 regions .
the slit positions used are shown as white lines in figure [ fig : upict ] , while the results from that work are given in table [ tab : metal ] .
slit one ( s1 ) runs east - west through the northern h-@xmath16 areas , while slit two ( s2 ) runs roughly north - south through the southern areas .
the areas labeled s2a3 ( unfortunately ) lies on an h-@xmath16 area which is overshadowed in the 814 image by a bright background galaxy .
the metallicities of these areas are low , around 0.4@xmath28 . combined with the extremely blue colors of the regions of h-@xmath16 regions , this indicates that these regions are likely young starburst areas which have not yet enriched much of the gas
. a further look at two of these regions of h-@xmath16 emission ( areas b & c ) is given in figure [ fig : halp ] .
we have determined the 814 - 300 colors of localized regions within these h-@xmath16 areas , with the results listed in table [ tab : halp ] .
regions 1 - 8 within the h-@xmath16 area were chosen to encompass various regions of localized starburst .
the colors of these regions are exceptionally blue , and thus are likely comprised primarily of young stellar types .
significantly , though , region 8 which , although clearly associated with the h-@xmath16 areas , does not encompass a regions of apparent significant star forming activity , also has an 814 - 300 color which is at least 1 magnitude bluer than ugc 12695 s nuclear region .
it is thus clear that the majority of the current star formation in ugc 12695 is happening within very localized ( small spatial scale ) , non - centralized , star forming regions which are dispersed throughout the galaxy .
as these are short exposure images of the nuclear region of a nearby galaxy , the existence of any background galaxies in our image was a large surprise . the fact that we discovered 21 potential background galaxies is proof both of the fairly transparent nature of ugc 12695 as well as the excellent ability of hst wfpc observations to detect distant galaxies no matter where its pointed .
details of the galaxy identification and analysis techniques are given in section 2 of this paper , while the results , as applied to these background galaxies , are listed in table [ tab : backg ] and described below .
the 814 band detection limit was 24.5 mag .
column 1 : the name given to each of these galaxies ( none have been previously identified ) .
the numbering system is based on all objects identified in the image , including foreground stars , image flaws , and regions of h-@xmath16 emission .
columns 2 and 3 : ra and dec in j2000 coordinates , determined using the stsdas metric task .
column 4 : the 814 - 300 color of the galaxies . if the galaxy could not be identified in the 300 image , a maximum apparent brightness ( magnitude ) was determined through assuming the galaxy would have been identified if its radius exceeded 5 pixels and it had an average brightness greater than 3@xmath9 above the sky .
( sky brightness was determined for each of the four wfpc2 frames separately . ) .
column 5 : the u - i color of each galaxy , as described in section 2 . columns 6 and 7 : the total magnitude of the galaxy through both the 814 and 300 filters .
magnitudes are corrected for galactic extinction ( treating the 814 filter as a johnson i band filter and the 300 filter as a johnson u band filter ) but not for inclination or redshift ( k - correction ) .
column 8 : the total integrated magnitude of the galaxy through the 814 band using @xmath29 where @xmath16 is the exponential scale length in arcsec .
if an exponential profile was not fit to a particular galaxy s surface brightness profile this column is left blank .
it should be noted that on occasion ( i.e. u2 - 18 ) the magnitude in this column is considerably brighter than that given in column 7 . in these cases
the maximum aperture size used in calculating the total magnitude ( column 7 ) was smaller than the actual galaxy size , typically due to interference from a neighboring galaxy or from ugc 12695 .
the integrated magnitude given in this column is thus the more accurate .
column 9 : the exponential scale length of the galaxy , found as described in section 2 , equation [ eq : mue ] .
column 10 : the central surface brightness of the galaxy in mag arcsec@xmath19 , as defined in section 2 , equation [ eq : mue ] .
when a surface brightness profile was found for a galaxy , yet no line was fit to it , the central surface brightness is estimated .
column 11 : the inclination corrected central surface brightness in mag arcsec@xmath19 . @xmath30 where the inclination used is listed in column 13 . note that this is a geometric path length correction which assumes no dust , and therefore may not be accurate for these galaxies .
column 12 : the major axis radius in arcsec as measured at the @xmath31= 25.0 mag arcsec@xmath19 isophote .
if the surface brightness profile errors exceeded 0.25 mag arcsec@xmath19 before @xmath31= 25.0 mag arcsec@xmath19 , then the largest accurate radius is given .
column 13 : the inclination angle ( in degrees ) as found by the gasp software ( equation [ eq : inclin ] ) .
the angle is correct to within @xmath5 5@xmath22 .
( the error is primarily due to the galaxies lack of azimuthal symmetry ) .
eight of the galaxies were large and bright enough to be identified within the mdm images .
these galaxies are listed in table [ tab : gndgal ] , along with the galaxies colors , as determined from the ground - based images .
as these are distant galaxies , none could be reliably identified through the b band image , and a few could not even be identified in the v band image . in these cases
the maximum apparent magnitude was determined for the ` drop - out ' band , as described above , using a minimum detection radius of 3 pixels .
the majority of the galaxies discovered have sizes @xmath11.5 , making a morphological description difficult at best .
a few of the galaxies , though , are fairly large and thus could be examined more closely .
figure [ fig : backg ] shows the four largest background galaxies discovered in our image .
all four appear to be well - formed spiral galaxies , with little of the irregular morphology evident in deeper hst wfpc2 surveys such as the hubble deep field ( williams et al 1996 ) .
followup spectroscopy of these galaxies is clearly desirable .
comparing the values of these background galaxies with the colors , sizes , and magnitudes of both the medium and deep hst surveys leads to the conclusion that the background galaxies in our survey lie at a redshift of 0.5 @xmath1z @xmath11.5 .
the radii of the galaxies in the medium - deep survey ( mds ) lie between 1.2 2.2 , while the v @xmath11 i color of the mds galaxies is 0.1 @xmath1v @xmath11 i @xmath13.0 , with @xmath32v @xmath11 i @xmath33= 1.2 , and the survey ranging from 20.0 @xmath1m@xmath34 @xmath122.0 . the background galaxies in our survey thus lie at the red edge of this color spectrum and are typically smaller ( in apparent size ) than those in the mds . as the mds galaxies were determined to lie between 0.5 @xmath1z @xmath11.0 , we conclude the background galaxies in our survey lie at least that distant , or at 0.5 @xmath1z @xmath11.5 . in order to more fully understand the morphology and distribution of these galaxies , and to thereby gain a firmer understanding of galaxy morphological evolution , a more detailed look at these galaxies through multiple filters and at longer exposure times ,
is definitely warranted .
ugc 12695 is a remarkable galaxy .
it has an exceptionally high gas mass fraction , very low metallicity , a strikingly diffuse morphology with perhaps the bluest colors known for a galaxy , and an exceedingly transparent nature .
combined , these attributes indicate ugc 12695 is a highly unevolved galaxy . since ugc 12695 is at rather low redshift ( [email protected] )
then its properties indicate that some potentials may well have late collapse and formation timescales ( such as may also be the case in ngc 1705 , another apparently young galaxy - see meurer et al 1992 ) .
the discovery of apparently young galaxies at low redshift has considerable importance to understanding galaxy evolution . as its unlikely
there is anything unique about ugc 12695 , one would reasonably expect objects like it to exist at any redshift
. however , the lsb nature of ugc 12695 makes it difficult to detect .
indeed it seems ironic that an apparently unevolved galaxy would itself difficult to detect because of its low surface brightness .
this obviously has important implications for detecting young galaxies at any redshift .
understanding the evolution of ugc 12695 might shed considerable light into galaxy formation scenarios as a whole .
unlike most galaxies , ugc 12695 also has the advantage of being relatively isolated , with its only nearest neighbor located at a projected distance of @xmath0200 kpc .
ugc 12695 is likely to have been influenced minimally ( if at all ) by other galaxies .
clues to the formation of ugc 12695 may be found in its structural peculiarities .
the galaxy s low overall density is manifest in its dynamical mass - to - total size , as well as its transparent nature .
the inferred low mass density of ugc 12695 ( and lsb disks in general - see de blok and mcgaugh 1997 ) may mean that stellar orbits are not well - defined in this potential thus resulting in the observed non - circular nature of even the inner nucleus and the fact that the majority of the star formation is taking place in non - centralized , yet strongly localized regions spread throughout the diffuse galaxy body . whatever is the case ,
the overall properties indicate that ugc 12695 is still highly unevolved .
the clear fact that this object has escaped our attention , due to its diffuse nature , causes concern over how many other examples of this phenomenon in the nearby universe have been missed to date .
this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
the calibration we have used in converting our 814 data to the johnson - cousins system is substantially different from the values given in holtzmann et al ( 1995 ) .
the primary reason for this is because our instrumental zeropoint is based on the st magnitude system while the holtzmann et al values use vega to define the zeropoint .
as detailed by whitmore in the wfpc2 photometry manual ( stsci internal publication ) the difference in conversion factors is substantial .
this is because the 814 filter is a close approximation to johnson - cousins i , but the definition of st mags , for a star of color corresponding to k0iii ( a reasonable approximation to the sed of a galaxy ) is different by 1.2 magnitudes than if vega were used as the zeropoint .
since this difference may not be widely known in the community , unless the wfpc2 photometry documentation has been read carefully ( specifically chapter 41 ) , we give a brief synopsis here .
both the 300 and 814 filters are wide bandpass filters , designed to maximize the throughput , and were not designed to specifically match a standard filter bandpass .
the stmag system exists with the true filter design in mind and is based on a spectrum with constant flux per unit wavelength .
conversion from the dn to the stmag system is thus straightforward : @xmath35 where photflam is a value obtained by stsci using standard star calibrations with the four wfpc2 chips , and takes into account distortions , plate scale , etc . because the stmag system is not a standard magnitude system , though , conversion from the dn is often instead done to a more conventional system based on vega s spectrum ( i.e. holtzmann et al 1995 ; driver , windhorst , & griffiths 1995 ) .
conversion from the stmag system to the vega system is typically done through 814@xmath36 @xmath11 814@xmath37 @xmath38 -1.226 , 300@xmath36 @xmath11 300@xmath37 @xmath38 -0.05 .
whichever of these systems is chosen , though , additional corrections need to be made to convert the data to johnson - cousins colors . before discussing the additional corrections necessary for conversion to the johnson - cousins system , an additional factor needs to be considered .
wfpc2 images can be taken with two different gain settings , 7 e@xmath8/adu and 14 e@xmath8/adu .
the calculations done by holtzman et al ( 1995 ) are for the gain=14 setup , while the data contained in this paper , as well as for the majority of the recent wfpc2 data was taken with the gain=7 setting .
this results in a difference of roughly -2.5log(2 ) = -0.75 , but of course has a further dependence on the spectral energy distribution of the source .
our ground - based i - band data , which is used to convert the 814 data to the johnson - cousins system , was obtained through the standard kpno filter set .
this i - band filter has @xmath39 = 8290 @xmath5 40 and @xmath40(fwhm ) = 1950 .
the traditional johnson i - band filter has @xmath39 9000 and @xmath40(fwhm ) = 2400 . because ccds have considerably more red sensitivity than the old s-20 phototube , the peak wavelength of the i - band filter used for ccd observations is significantly bluer . the 814 filter used in the hst observations has @xmath39 = 8386 , @xmath32@xmath41@xmath33= 8269 , @xmath40(fwhm ) = 1758 , values that are similar to the i - band filter used from the ground .
had both the ground - based and hst data been calibrated against vega , we would have recovered the holtzmann et al ( 1995 ) conversion . the much larger conversion that we found ( e.g. 1.4 mags ) is due almost entirely to the use of the stmag system for our wfpc2 data with a gain setting of 7 instead of 14 ( again this is all discussed in chapter 41 of the wfpc2 photometry manual ) .
bothun , g. , impey , c. , & mcgaugh , s. 1997 , pasp 109 , 745 + bothun , g.d .
mould , jeremy r. , caldwell , nelson , & macgillivray , harvey t. 1986 , aj , 92 , 1007 + bothun , g.d . , beers , t. , mould , jeremy r. , and huchra , j. 1985 , aj 90 , 2487 + bottinelli , l. , gouguenheim , l. , fouque , p. , & paturel g. 1990 , a&as , 82 , 391 + cawson , m. 1983 , ph.d .
thesis , university of cambridge + driver , simon p. , windhorst , roger a. , & griffiths , richard e. 1995 apj 453 , 48 + de blok , w.j.g . , & mcgaugh , s. 1997 , mnras , 290 , 533 + holtzman , j. , burrows , c. , casertano , s. , hester , j. , trauger , j. , watson , a. , & worthey , g. 1995 pasp 107 , 1065 + klein , u. , giovanardi , c. , altschuler . d.r . , & wunderlich , e. 1992 , a&a , 255 , 49 + lewis b.m .
1987 , apjs , 63 , 515 + mcgaugh , s.s . , 1994 , apj , 426 , 135 + mcgaugh , s.s . , 1992 , ph.d .
thesis , university of michigan , ann arbor + mcgaugh , s.s .
, & de blok , w.j.e .
1997 , apj , 481 , 689 + mcgaugh , s.s . , schombert , james e. , & bothun , g.d . 1995 ,
aj , 109 , 2019 + mcgaugh , s.s . , & bothun , g.d .
1994 , aj , 107 , 530 + meurer , gerhardt r. , freeman , k. c. , dopita , michael a. , & cacciari , carla 1992 , aj , 103 , 60 + oneil , k. , bothun , g.d . , &
impey , c.d .
1998 , in preparation + oneil , k. , bothun , g.d . , schombert , j. , cornell , m. & impey , c. 1997 aj , 114 , 2448 + schneider , s.e . , thuan , trinh x. , magri , christopher , & wadiak , james e. 1990 , apjs , 72 , 245 + williams , robert , et al 1996 aj , 112 , 1335 + figure [ fig : wfpc ] .
hst wfpc2 image of ugc 12695 taken through the 814 ( i band ) filter ( a ) , and the 300 ( u band ) filter ( b ) with a 2000s exposure time . 0.2 in figure [ fig : gnd ] .
images of ugc 12695 taken using the mdm 1.3 m telescope through the i , v , b , u and continuum subtracted h-@xmath16 bands ( figure [ fig : gnd ] a - e , respectively ) .
the images are 124 across , with north up and east to the left .
figure [ fig : gnd]e also has the regions of h-@xmath16 emission studied in section 4 labeled .
see mcgaugh ( 1992 ) for details on these images . 0.2 in figure [ fig : sbprof ] .
surface brightness profile of ugc 12695 through the 814 filter .
the solid line is the surface brightness profile with all background galaxies and h-@xmath16 regions included in the analysis , the dashed line is ugc 12695 with all the background galaxies ( except u2 - 74 , see text ) removed , and the dash - dotted line is ugc 12695 with both the background galaxies and all the h-@xmath16 regions removed .
( see section 3 . )
the errors in the surface brightness profiles are low , staying below 0.20 mag arcsec@xmath19 up to a radius of 40 , after which the errors reach 0.30 - 0.50 mag arcsec@xmath19 for the last 3 plotted points .
this error comes primarily from uncertainties in the sky background .
our measured value is 21.6 counts / pixel with a mean error of 0.1 counts / per pixel .
this maps onto a mean surface brightness through the 814 filter of 21.5 mag arcsec@xmath19 with a one sigma isophotal detection at 27.4 mag arcsec@xmath19 ( e.g. 0.1 counts / pixel ) .
hence , we can do reliable isophotal detection down to 6 magnitudes below sky with this data . 0.2 in figure [ fig : nuc ] .
expanded view of the nucleus of ugc 12695 through the wfpc2 814 filter .
the image is 20 x 20 , with the same orientation as figure [ fig : wfpc](a ) . 0.2 in figure [ fig : upict ] .
image showing the regions of h-@xmath16 emission discussed in section 4 .
the circled regions correspond to the h-@xmath16 regions listed in table [ tab : hcol ] , while the white lines correspond to the spectral slits used in mcgaugh ( 1994 ) and presented in table [ tab : metal ] . 0.2 in figure [ fig : halp ] .
images of the h-@xmath16 areas b and c discussed in section 4 .
the image is 1 across .
this image is the inverse ( across the vertical axis ) of figure [ fig : upict ] . 0.2 in figure [ fig : backg ] .
images of the four largest background galaxies through the 814 filter .
the images are of u2 - 22 , u2 - 23 , u2 - 39 , and u2 - 45 ( figure [ fig : backg]a d , respectively ) . table [ tab : glob ] .
the global properties of ugc 12695 . 0.2 in table [ tab : col ] .
colors of ugc 12695 determined from the 814 and 300 wfpc2 images . 0.2 in table [ tab : hcol ] .
colors of the h-@xmath16 regions of ugc 12695 . 0.2 in table [ tab : metal ] .
metallicity of select h-@xmath16 regions in ugc 12695 ( mcgaugh 1994 ) . 0.2 in table [ tab : halp ] .
photometry of select regions within the h-@xmath16 areas b and c. 0.2 in table [ tab : backg ] . photometric and structural parameters derived for the background galaxies . 0.2 in table [ tab : gndgal ] .
colors of eight of the background galaxies identified in the ground based ( mdm ) images .
ccccccccccc + & * ra * & * dec * & * v ( km / s ) * & * d ( mpc ) * & * m@xmath42 * & * m@xmath43 * & * m@xmath42/l@xmath44 * & * f@xmath45 * & * h ( kpc ) * & + & ( 1)&(2)&(3)&(4)&(5)&(6)&(7)&(8)&(9 ) + + & 23:36:02.0 & 12:52:32 & 6182 & 4590 & 9.62 & 9.98 & 1.28 & 0.62 & 8.4 & + + + & & * m@xmath44 * & * @xmath46(0 ) * & * m@xmath44 * & * b - v * & * u - b * & * v - i * & + & & ( 10)&(11)&(12)&(13)&(14)&(15 ) + + & & -18.92 & 23.8 & 15.53 & 0.26 & -0.11 & 0.95 + + cccccccccccccc + & & & & & + + & * 814 * & * 300 - 814 * & * u - i * & & * 814 * & * 300 - 814 * & * u - i * & & * 814 * & * 300 - 814 * & * u - i * + a : & 22.215 & -0.541 & 2.159 & & 19.426 & -0.440 & 2.260 & & 18.069 & -2.750 & -0.092 & + b : & 22.215 & -0.541 & 2.159 & & 19.509 & -0.977 & 1.723 & & 18.145 & -2.879 & -0.179 & + c : & 22.215 & -0.541 & 2.159 & & 19.511 & -0.970 & 1.730 & & 18.145 & -2.883 & -0.183 & + + + + + + + + ccccccccc + + * name * & * ra * & * dec * & * 300 - 814 * & * u - i * & r & 300@xmath47 & 814@xmath47 & r@xmath48 + & ( j2000 ) & ( j2000 ) & & & ( ) & & & ( ) + + + a & 23:35:59.68 & 12:52:31.4 & -2.167 & 0.533 & 0.63 & 21.194 & 22.511 & 1.09 + b & 23:26:04.21 & 12:52:25.5 & -2.337 & 0.363 & 1.53 & 19.284 & 19.831 & 2.15 + c & 23:36:04.09 & 12:52:29.4 & -2.627 & 0.077 & 2.62 & 17.964 & 19.941 & 3.31 + d & 23:36:04.25 & 12:52:39.9 & -2.607 & 0.093 & 1.12 & 20.374 & 22.981 & 1.12 + e & 23:36:03.98 & 12:52:46.5 & -2.467 & 0.223 & 6.30 & 17.514 & 19.981 & 6.30 + f & 23:36:04.33 & 12:52:47.3 & -2.557 & 0.143 & 2.37 & 19.724 &
22.281 & 2.37 + + + + + cccc + + region & log(o / h ) & log ( n / h ) & log(ne / h ) + + + s1a1 & -4.18 & -6.52 & -4.84 + s1a2 & -3.98 & -5.66 & -4.76 + s1a3 & -4.01 & -5.81 & -4.86 + s2a1 & -4.08 & -5.81 & -4.83 + s2a2 & -4.09 & -5.62 & -4.96 + s2a3 & -3.92 & - & -4.95 + + cc + region 1 : & 300 - 814 = -2.499 + region 2 : & 300 - 814 = -2.481 + region 3 : & 300 - 814 = -2.950 + region 4 : & 300 - 814 = -2.555 + region 5 : & 300 - 814 = -2.633 + region 6 : & 300 - 814 = -5.076 + region 7 : & 300 - 814 = -2.537 + region 8 : & 300 - 814 = -1.820 + inner 2 ( nucleus ) : & 814 - 300 = -0.541 + inner 10 ( nucleus ) : & 814 - 300 = -0.977 + + ccccccccccccc + + name & ra & dec & 300 - 814 & u - i & 300@xmath47 & 814@xmath47 & 814@xmath47(@xmath16 ) & @xmath16 ( ) & @xmath18 & @xmath18@xmath49 & r@xmath50 ( ) & _ i _ + ( 1)&(2)&(3)&(4)&(5)&(6)&(7)&(8)&(9)&(10)&(11)&(12)&(13 ) + + + u2 - 8 & 23:36:01.64 & 12:53:27.3 & @xmath51 - 1.02 & @xmath511.68 & @xmath5122.56 & 23.581 & 23.090 & 0.380 & 22.987 & 23.487 & 0.894 & 50.9 + u2 - 14 & 23:35:58.22 & 12:52:59.7 & @xmath51 - 1.38 & @xmath511.32 & @xmath5121.18 & 22.561 & 22.986 & 0.246 & 21.932 & 22.263 & 1.415 & 42.5
+ u2 - 17 & 23:36:02.66 & 12:53:34.3 & @xmath51 - 0.33 & @xmath512.37 & @xmath5121.18 & 21.511 & 21.212 & 0.632 & 22.209 & 22.719 & 1.823 & 51.3 + u2 - 18 & 23:35:58.48 & 12:53:49.9 & @xmath511.62 & @xmath514.32 & @xmath5121.18 & 19.561 & 18.164 & 1.033 & 20.223 & 20.755 & 4.127 & 52.0 + u2 - 20 & 23:36:02.06 & 12:53:06.4 & @xmath51 - 1.40 & @xmath511.31 & @xmath5121.18 & 22.571 & 22.165 & 0.539 & 22.817 & 22.878 & 1.207 & 19.1 + u2 - 22 & 23:36:00.07 & 12:52:36.2 & -0.867 & 1.833 & 20.864 & 20.991 & 20.317 & 0.800 & 21.828 & 21.891 & 1.884 & 19.3 + u2 - 23 & 23:35:59.33 & 12:52:40.1 & -0.517 & 2.183 & 21.444 & 20.461 & 18.694 & 1.362 & 21.360 & 22.560 & 4.277 & 70.7 + u2 - 24 & 23:35:58.13 & 12:53:56.5 & @xmath51 - 2.89 & @xmath51 - 0.19&@xmath5121.18 & 24.071 & 23.70 & 0.297 & 23.058 & 23.911 & 0.586 & 62.9 + u2 - 36 & 23:36:04.35 & 12:52:59.9 & @xmath51 - 0.95 & @xmath511.75 & @xmath5121.19 & 22.141 & 21.220 & 0.537 & 21.864 & 22.185 & 1.270 & 41.9 + u2 - 39 & 23:36:03.58 & 12:52:39.1 & @xmath511.28 & @xmath513.98 & @xmath5121.19 & 19.911 & 18.475 & 1.314 & 21.063 & 22.465 & 5.596 & 74.0 + u2 - 41 & 23:36:03.94 & 12:53:04.1 & @xmath510.92 & @xmath513.62 & @xmath5121.19 & 20.271 & 20.301 & 0.086 & 16.969 & 17.564 & 1.512 & 54.5 + u2 - 45 & 23:36:02.36 & 12:52:41.0 & @xmath510.70 & @xmath513.40 & @xmath5121.19 & 20.491 & 19.437 & 0.863 & 21.111 & 21.971 & 0.863 & 63.1 + u2 - 46 & 23:36:06.68 & 12:52:27.1 & - & - & - & - & - & - & - & - & - & - + u2 - 48 & 23:36:07.16 & 12:53:04.7 & @xmath51 - 1.23 & @xmath511.47 & @xmath5121.19 & 22.421 & 22.634 & 0.367 & 22.515 & 22.845 & 0.367 & 42.4 + u2 - 53 & 23:36:03.07 & 12:53:37.1 & - & - & - & - & - & - & - & - & - & - + u2 - 65 & 23:36:01.84 & 12:51:53.4 & @xmath51 - 3.30 & @xmath51 - 0.59&@xmath5121.12 & 24.411 & 24.173 & 0.199 & 23.448 & 22.667 & 0.199 & 60.9 + u2 - 67 & 23:36:04.74 & 12:51:59.4 & @xmath51 - 1.96 & @xmath510.74 & @xmath5121.12 & 23.081 & 22.971 & 0.408 & 23.022 & 23.830 & 0.948 & 61.6 + u2 - 68 & 23:36:04.27 & 12:51:04.1 & @xmath51 - 0.34 & @xmath512.36 & @xmath5121.12 & 21.461 & 21.405 & 0.363 & 21.099 & 21.654 & 2.219 & 53.1 + u2 - 70 & 23:36:04.21 & 12:51:19.7 & @xmath51 - 1.66 & @xmath511.04 & @xmath5121.12 & 22.781 & 22.841 & 0.206 & 21.400 & 21.593 & 1.109 & 33.2 + u2 - 72 & 23:36:01.22 & 12:51:59.1 & @xmath51 - 1.21 & @xmath511.49 & @xmath5121.12 & 22.331 & 21.150 & 0.726 & 22.449 & 22.927 & 1.394 & 29.9 + u2 - 74 & 23:36:02.09 & 12:52:20.2 & @xmath510.36 & @xmath513.06 & @xmath5121.12 & 20.761 & 21.283 & 0.454 & 21.565 & 22.290 &
2.050 & 59.1 + + + + + + + + + cccccc + + & * name * & * b - v * & * v - i * & * radius ( ) * & + + + & u2 - 17 & - & @xmath511.92 & 1.82 & + & u2 - 22 & @xmath510.88 & 1.40 & 1.10 & + & u2 - 23 & @xmath510.00 & 1.91 & 0.95 & + & u2 - 36 & - & @xmath511.29 & 1.27 & + & u2 - 39 & @xmath52 - 0.12 & 0.16 & 5.50 & + & u2 - 41 & @xmath510.49 & 3.07 & 1.55 & + & u2 - 45 & & 1.75 & 0.86 & + & u2 - 72 & @xmath51 - 1.34 & 2.84 & 1.40 & + + + & + + | utilizing the f814w and f300w filters , short exposure hubble space telescope wide field planetary camera-2 ( wfpc2 ) images were taken of ugc 12695 a nearby ( z @xmath0 0.021 ) low surface brightness disk galaxy .
ugc 12695 has an unusual morphology , consisting of a y - shaped nucleus surrounded by a faint spiral arm with a number of bright h ii regions interspersed throughout the galaxy .
surface photometry indicates the majority of recent star formation in this galaxy occurred in these very localized regions , most of which have a radius of @xmath12 . this uneven stellar distribution combined with the galaxy s
overall extremely blue color and low metallicity indicates ugc 12695 is an unevolved galaxy .
some of the structural peculiarities of this galaxy arise because a number of background galaxies , previously thought to be morphological components of this galaxy , are showing through both the outer nucleus and spiral arms of ugc 12695 .
surface photometry of these galaxies shows them to be fairly small ( @xmath2 1.8 ) disk galaxies with total magnitudes ranging from 19.6 through 24.2 , and central surface brightnesses between 20.2 mag / arcsec@xmath3 .
when possible , the u , b , v , and i colors of these galaxies were measured using ground - based images , which show the galaxies to be fairly red indicating they are likely at redshifts z @xmath4 0.5 .
inclusion of them in the photometry of ugc 12695 makes the galaxy appear significantly redder .
when these galaxies are masked out , the resultant u - i color of ugc 12695 is -0.2 @xmath5 0.1 , making it perhaps the bluest galaxy every measured in this color system and confirming its nature as a very unevolved galaxy at low redshift .
spectroscopy of these background galaxies , through the transparent disk of ugc 12695 , may help to clarify its chemical evolution and current heavy element content .
.2 in * keywords * galaxies : evolution ; galaxies : star formation ; galaxies : individual ( ugc 12695 ) ; galaxies : photometry ; galaxies : spiral galaxies : stellar content 0.5 in 0.5 in 0.5 in received ' '' '' ; accepted ' '' '' + = 24pt |
although described by the same line element , the classical interior of the schwarzschild - tangherlini @xcite black hole has entirely different properties than the region outside the event horizon and can be better understood as some sort of the anisotropic and nonstatic universe @xcite . this interpretation ( not mandatory )
is helpful when one is forced to abandon usual , i.e. , referring to the external world , interpretation of the coordinates , metric potentials and so on . in the @xmath0-dimensional case
much work have been done in this direction and we have a good understanding of the geometry and dynamics of the classical interior ( see e.g. , refs .
@xcite and the references therein ) . on the other hand , less
is known about quantum processes inside black holes and their influence upon the background geometry . in the recent paper @xcite we have studied influence of the quantized fields on the static spacetime of the schwarzschild - tangherlini black hole using the semi - classical einstein field equations . since the stress - energy tensor constructed in that paper
functionally depends on the metric , one has a rare opportunity to analyze and compare the quantum corrections to the black hole characteristics ( and the geometry itself ) calculated for various spacetime dimensions .
the purpose of this paper , which is a natural continuation of refs .
@xcite , is to extend the study of the quantized fields to the interiors of the higher - dimensional black holes .
it should be noted however , that now there are problems that do not appear in the external region .
the first one is the problem of the central singularity and its closest vicinity .
it is evident that the semicalssical einstein field equations can not be trusted there .
the second difficulty is to some extend related to the previous one and may be stated as follows .
the effective action of the quantized fields for @xmath6 ( @xmath7 is the coordinate of the event horizon ) has been constructed for the positive - definite metric signature .
once the stress - energy tensor is calculated it can be transformed to the physical spacetime by analytic continuation . in the exterior region
it is the familiar wick rotation , which affects only the time coordinate . on the other hand ,
inside the event horizon the problem is more complicated .
the classical @xmath8dimensional solution describing interior of the schwarzschild - tangherlini black hole with the event horizon located at @xmath9 is given by the line element @xmath10^{-1 } dt^{2 } + \left[\left(\frac{r_{+}}{t}\right)^{d-3}-1\right ] dx^2 + t^{2 } d\omega^{2}_{d-2 } , \label{line_classical}\ ] ] where @xmath11 is a metric on a unit @xmath12-dimensional sphere .
since only in @xmath0 case there is a simple linear relation between the mass and @xmath7 in the present paper we use ( almost ) exclusively the latter .
the radius of the event horizon of the schwarzschild - tangherlini black hole characterized by the mass @xmath13 is given by @xmath14 where @xmath15 is @xmath16-dimensional newton constant and @xmath17 is the volume of the unit @xmath12-dimensional sphere .
if the @xmath12-dimensional sphere is covered by a standard `` angular '' coordinates @xmath18 the metric @xmath19 can be written in the form @xmath20.\ ] ] now , in order to construct the positive - definite metric let us replace @xmath21 by @xmath22 @xmath23 by @xmath24 and @xmath7 by @xmath25 the metric thus becomes : @xmath26^{-1 } dt^{2 } + \left[\left(\frac{r_{+}}{t}\right)^{d-3}-1\right ] dx^2 + t^{2 } d\omega^{2}_{d-2 } , \label{line_classical}\ ] ] where @xmath27.\ ] ] note that our transformation differs form that of ref
. @xcite , which results in the _
negative-_definite metric .
having euclidean version of the geometry of the black hole interior on can construct the one - loop approximation to the effective action of the quantized massive fields in a large mass limit .
indeed , for a sufficiently massive fields , i.e. , when the compton length , @xmath28 associated with the mass of the field , @xmath29 is much smaller than the characteristic radius of the curvature of the spacetime @xmath30 the contribution of the vacuum polarization to the effective action dominates and the contribution of real particles is negligible .
one can therefore make use of the schwinger - dewitt asymptotic expansion that approximates the effective action @xmath31 this approach has been successfully applied in a number of interesting cases and the background spacetimes range from black holes @xcite to cosmology @xcite and from wormholes @xcite to topologically nontrivial spacetimes @xcite . for the purposes of the present paper , the most relevant are the results presented in refs .
@xcite in ref .
@xcite it has been shown that the approximate one - loop effective action @xmath32 of the quantized massive scalar field in a large mass limit can be constructed from the ( asymptotic ) schwinger - dewitt representation of the green function and in the lowest order it can be written in the following form : @xmath33}{(m^{2})^{k - d/2}}\gamma\left(k-\frac{d}{2}\right ) , \label{wreg}\ ] ] where @xmath34 and @xmath35 denote the floor function , i.e. , it gives the largest integer less than or equal to @xmath36 here @xmath37 $ ] is the coincidence limit of the @xmath38-th hadamard - dewitt coefficient constructed from the riemann tensor , its covariant derivatives up to @xmath39-order and contractions .
for the technical details concerning construction of the hadamard - dewitt coefficients the reader is referred to refs .
the ( regularized ) stress - energy tensor can be calculated from the standard definition @xmath40 there is one - to - one correspondence between the order of the wkb approximation and the order of the schwinger - dewitt expansion .
for example , the sixth - order wkb approximation is equivalent to @xmath41 term in @xmath0 and to @xmath42 in @xmath4 whereas for the analogous results in @xmath2 and @xmath43 the eight - order wkb approximation is required .
on general grounds one expects that the lowest - order ( nonvanishing ) term of the schwinger - dewitt expansion is the most important .
the condition @xmath44 ( with the physical constants reinserted ) leads to @xmath45 where @xmath46 and @xmath21 is given in seconds employed in the @xmath0-dimensional back reaction calculations reported in ref .
for example , taking @xmath47 @xmath7 equal to the schwarzschild radius of the sun and @xmath48 kg one has @xmath49 which is many orders of magnitude smaller than the coordinate time of the event horizon .
it follows than that in our calculations we can go fairly close to the central singularity .
note that the coordinate time goes form @xmath50 to 0 . in the rest of the paper
we use the geometric units and the adopted conventions are those of misner , thorne and wheeler @xcite . the paper is organized as follows . in sec .
[ class ] we study some aspects of the classical interior of the @xmath16-dimensional schwarzschild - tangehrlini black holes . in sec .
[ zwr ] we construct and formally solve the @xmath16-dimensional semiclassical einstein field equation and analyze the problem of the finite renormalization . in sec .
[ gen_cons ] we show how to construct the appropriate measure of anisotropy and investigate the two useful scalars : the kretschmann scalar and the karlhede scalar . finally , taking the stress - energy tensor of the quantized massive scalar field , in sec .
[ sec_back ] we study the semiclassical equations and analyze the influence of quantum perturbations on the black hole interior for @xmath51
to gain a better understanding of the classical interior of the schwarzschild - tangherlini black hole let us introduce the proper time @xmath52^{1/2 } } \label{tau_plus}\ ] ] and , in the neighborhood of a point @xmath53 the locally euclidean coordinates @xmath54 near the singularity the schwarzschild - tangherlini metric can be approximated by the kasner metric @xmath55 where @xmath56 it can easily be checked that both kasner conditions are satisfied .
indeed , @xmath57 and @xmath58 on the other hand , near the event horizon the schwarzschild - tangherlini metric asymptotically approaches @xmath59 where @xmath60 once again it is the kasner metric with @xmath61 and vanishing remaining kasner exponents .
finally observe , that the line element ( [ kasner2 ] ) can be formally obtained from the rindler solution @xmath62 by using the complex coordinate transformation .
now , let us consider two points at the same coordinate instant separated by @xmath63 while the coordinate distance remains constant the physical distance between two points on the @xmath64 coordinate line is given by @xmath65^{1/2 } dx\nonumber \\ & = & \left[\left(\frac{r_{+}}{t}\right)^{d-3}-1 \right]^{1/2 } \delta x\end{aligned}\ ] ] and grows as the coordinate time decreases .
on the other hand , the proper distance between two points separated by @xmath66 is given by @xmath67 and it decreases as the coordinate time goes from @xmath7 to 0 .
this behavior is independent of the dimension @xmath68 let us return to the proper time : it should be noted that taking a positive sign of the root of the equation @xmath69 as it has been done in eq .
( [ tau_plus ] ) , the proper time monotonically grows with the coordinate time @xmath70 conversely , taking the negative root , the proper time increases as the coordinate time goes from @xmath7 to 0 .
since the functional relations between @xmath71 and @xmath21 are not very illuminating we present them graphically ( fig i ) , demanding for both types of the universe @xmath72 as @xmath73 this can always be done by suitable choice of the integration constant .
the universe inside the event horizon ( in both time scales ) has a finite lifetime . from the results collected in table
i one sees that @xmath74 decreases with the dimension . as a function of the coordinate time @xmath21 for @xmath75
the upper branches are plotted for a negative root of ( [ proper_time ] ) whereas the lower ones for a positive root .
the arrows indicate the direction of the flow of the coordinate time .
top to bottom ( for the upper branches ) the curves are plotted for @xmath76 , width=415 ] .the lifetime of the interior universe . [ cols="^,<,^",options="header " , ] let us analyze how the growth of the curvature are affected by the quantum processes . and
since the classical part of the kretschmann scalar is positive , one concludes that the growth of curvature ( as @xmath21 decreases ) is weakened if the perturbation is negative .
inspection of the figs .
[ rys3 ] and [ rys4 ] as well as the table [ tta3 ] shows that initially , regardless of the dimension , @xmath77 is always positive for the both types of couplings .
this is very important message as it refers to the region where the quality of the approximation is expected to be high .
a few words of comment are in order here .
first , it should be emphasized once more that although we have plotted functions @xmath78 and @xmath79 for all allowable values of the coordinate time the approximation certainly does not work in the region close to the central singularity .
therefore our results show the tendency in behavior of the quantum corrections as the central singularity is approached ( which can be misleading ) rather than their actual run .
of course the answer to the question of how close the singularity can be approached depends on many factors , such as dimension , the ` radius ' of the event horizon and the type of the quantized field .
each case should be analyzed separately .
the second observation is less obvious and is , roughly speaking , related to the question of how long the first - order approximation dominates the higher - orders terms inside the event horizon .
once again this problem goes to the very core of the schwinger - dewitt asymptotic expansion . and once again there is no better answer than to recall its principal assumptions . finally , let us observe that although the quantum corrections caused by a solitary field is expected to be small in the domain of applicability of the approximation , they can be made arbitrarily large for a large number of fields .
[ h ] on @xmath80 for zero perturbation to the anisotropy @xmath81 of the interior of the @xmath0 ( dashed line ) and @xmath4 ( solid line ) schwarzschild - tangherlini black hole . above the curves ,
the quantum corrections make the black hole interior more anisotropic , whereas below the curves the spacetime becomes more isotropic.,title="fig:",width=415 ] [ h ] on @xmath80 for zero perturbation to the anisotropy @xmath81 of the interior of the @xmath2 ( dashed line ) and @xmath43 ( solid line ) schwarzschild - tangherlini black hole . above the curves ,
the quantum corrections make the black hole interior more isotropic , whereas below the curves the spacetime becomes more anisotropic.,title="fig:",width=415 ] on @xmath80 for zero perturbation to the kretschmann scalar of the interior of the @xmath0 ( dashed line ) and @xmath4 ( solid line ) schwarzschild - tangherlini black hole . above the curves , the quantum corrections to the scalar @xmath82 are positive , whereas below the curves they are negative .
, width=415 ] on @xmath80 for zero perturbation to the kretschmann scalar of the interior of the @xmath2 ( dashed line ) and @xmath43 ( solid line ) schwarzschild - tangherlini black hole . above the curves , the quantum corrections to the scalar @xmath82 are negative , whereas below the curves they are positive .
, width=415 ]
in this appendix we list the @xmath83 and @xmath84 components of the stress - energy tensor for @xmath85 the angular components are not displayed as they can easily be calculated from the covariant conservation equation .
@xmath86 @xmath87 @xmath88 @xmath89 @xmath90 @xmath91 @xmath92 @xmath93 @xmath94 @xmath95 @xmath96 @xmath97
here we list the coefficients of the functions @xmath98 and @xmath100 ( see eqs . ( [ f1 ] -[hlast ] ) ) .
the integration constants @xmath101 are left unspecified and should be determined from the physically motivated boundary conditions .
all quantities considered in the main text , such as @xmath102 @xmath103 @xmath82 and @xmath104 are independent of the integration constant @xmath105 @xmath106 @xmath107 @xmath108 @xmath109 @xmath110 @xmath111 @xmath112 @xmath113 @xmath114 @xmath115 @xmath116 @xmath117 @xmath118 @xmath119 @xmath120 @xmath121 @xmath122 @xmath123 @xmath124 @xmath125 | the first - order semiclassical einstein field equations are solved in the interior of the schwarzschild - tangherlini black holes .
the source term is taken to be the stress - energy tensor of the quantized massive scalar field with arbitrary curvature coupling calculated within the framework of the schwinger - dewitt approximation .
it is shown that for the minimal coupling the quantum effects tend to isotropize the interior of the black hole ( which can be interpreted as an anisotropic collapsing universe ) for @xmath0 and @xmath1 whereas for @xmath2 and @xmath3 the spacetime becomes more anisotropic .
similar behavior is observed for the conformal coupling with the reservation that for @xmath4 isotropization of the spacetime occurs during ( approximately ) the first @xmath5 of the lifetime of the interior universe . on the other hand ,
we find that regardless of the dimension , the quantum perturbations initially strengthen the grow of curvature and its later behavior depends on the dimension and the coupling .
it is shown that the karlhede s scalar can still be used as a useful device for locating the horizon of the quantum - corrected black hole , as expected . |
over the last decade , there have been a growing number of high - angular - resolution studies using observations from submm / mm and cm interferometers , aimed at understanding morphological and kinematical features associated with the massive star formation process , such as molecular outflows , possible rotating disks or envelopes , and inward accretion flows ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) .
more recently , studies based on observations from the submillimeter array ( sma ) , a combined power of high - angular - resolution and broad frequency coverage , have begun to shed light on the physical and chemical diversities in high - mass star - forming cores at different evolutionary stages ( e.g. , rathborne et al . 2008 ; beuther et al . 2009 ; zhang et al .
these efforts provided a wealth of information that greatly advanced our knowledge of massive star formation . in this paper
we present a high - angular - resolution study of the high - mass star - forming region hh 8081 , which has a bolometric luminosity of @xmath1 @xmath2 at an adopted distance of 1.7 kpc @xcite .
this region is distinguished by a 5.3 pc radio jet that has an opening angle of @xmath31@xmath4 , known as the best collimated radio jet driven from a high - mass protostar @xcite .
recent _ spitzer _
8 @xmath5 m imaging reveals a bi - conical outflow cavity surrounding the radio jet @xcite .
the jet powering source has been detected from mid - infrared to cm wavelengths @xcite , yet some basic properties of the source , such as physical and chemical conditions of the circumstellar matter , remain unclear . here
we present a detailed study of the dust continuum and molecular line emission of the radio jet powering source , and the discovery of an exceptional co outflow containing fast moving molecular bullets .
the sma observations toward the phase center ( r.a .
, decl.)@xmath6 were undertaken with eight antennas on 2008 june 30 in the compact configuration and on 2008 september 9 in the sub - compact configuration , both under good weather conditions . particularly the sub - compact data were taken with @[email protected] . the 4 ghz correlator was configured to cover rest frequencies of 219.2 to 221.2 ghz in the lower sideband and 229.2 to 231.2 ghz in the upper sideband . in the compact configuration , we observed uranus , 3c454.3 , and 3c273 for bandpass calibration , callisto for flux calibration , and j1911 - 201 to monitor the time dependent gain . in the sub - compact configuration , observations of 3c454.3 , uranus , and j1733 - 130 and j1911 - 201
were used for the bandpass , flux , and gain calibrations , respectively .
the data were calibrated using the idl mir package and then exported to miriad for further processing .
self - calibration was performed on continuum and the solutions were applied to lines as well .
the two data sets were jointly imaged with the robust parameter set to 0.5 , resulting in a synthesized beam of @xmath3@xmath8@xmath9@xmath10 at fwhm .
figure [ fig1 ] shows the 1.3 mm continuum map and the integrated intensity maps of lines detected from a variety of molecular species .
the mm continuum reveals two dominant peaks mm1 and mm2 ; mm1 coincides well with the radio source that powers the hh 8081 radio jet and mm2 is spatially associated with h@xmath11o masers and a very weak radio source @xcite . in addition ,
emission lines from high - density tracers , such as ch@xmath12oh , ch@xmath12cn , h@xmath13co , ocs , and hnco , clearly reveals another dense core mc ( figures [ fig1]b[fig1]f ) , although the mm continuum map only shows a weak enhancement toward this core . in figure [ fig1]b the ch@xmath12oh ( 8@xmath14@xmath15 emission toward mc is eminently bright with a peak flux of 6.4 jybeam@xmath0
this narrow spectral component at a lsr velocity of @xmath16 kms@xmath0 appears to be a maser emission .
@xcite detected a 44 ghz methanol maser in the vicinity of mc , which is consistent with our 229.8 ghz detection as both masers are class i transitions @xcite .
apart from the maser emission , most thermal line emission in figure [ fig1 ] appears brighter toward mm1 . in particular lines of ch@xmath12oh ( 15@xmath17@xmath18 and hnco @xmath19@xmath20 and those from so@xmath11 isotopologues
are only seen in mm1 .
we will discuss these diversities in section [ divers ] .
mm1 has an integrated flux of 1.2 jy at 1.3 mm continuum , in agreement with the bima measurement of 0.97 jy at 1.4 mm and is completely dominated by the dust emission @xcite . the mm continuum emission from mc
does not appear to be very centrally peaked and is blended with the mm2 emission .
therefore we can not obtain a reliable estimate of the continuum flux for mc alone .
the measured flux toward mm2 and mc amounts to 0.8 jy and is apparently dominated by mm2 . from the continuum fluxes and assuming thermal equilibrium between the gas and dust hence adopting the temperatures derived from the ch@xmath12cn emission ( section [ divers ] ) , we obtain a gas mass of 634 @xmath21 for mm1 and a total mass of 1265 @xmath21 for mm2 and mc , for a dust opacity @[email protected] @xmath23 g@xmath0 or an emissivity index @xmath24@xmath312 @xcite .
figure [ co ] shows the high velocity co ( 21 ) emission integrated over every 20.4 kms@xmath0 from @xmath25 to @xmath26 kms@xmath0 in the blueshifted lobe and from 28.2 to 109.2 kms@xmath0 in the redshifted lobe .
the emission at lower velocities ( @xmath27@xmath2815 kms@xmath0 with respect to the systemic velocity of 12 kms@xmath0 for mm2 ) suffers from side lobes and missing flux due to inadequate @xmath29 coverage , hampering the interpretation of the morphology and kinematics , hence we focus on the high velocity co emission in this work .
the most remarkable feature in figure [ co ] is a jet - like outflow emanating from mm2 to the southeast ( se ) , which exhibits extremely high velocity ( ehv ) emission up to @xmath25 kms@xmath0 for the blueshifted lobe and @xmath30 kms@xmath0 for the redshifted lobe above a 3@xmath31 level of @xmath360 mjybeam@xmath0 per 1.2 kms@xmath0 .
this outflow shows a knotty appearance with multiple clumps aligned along the flow axis . in addition , two more collimated flows , the northwest and northeast flows , are seen in figure [ co]e ; the former apparently originates from mc , whereas it is difficult with the existing data to unambiguously assign a driving source to the latter .
these two flows both extend well beyond the sma primary beam of @xmath354@xmath32 at 230 ghz , thus are not further discussed in this work .
figure [ so ] shows the so ( @xmath33@xmath34 ) emission integrated over five velocity intervals .
the blueshifted emission arising from the se outflow is detected as an ehv clump in figure [ so]a and a low to high velocity clump in figure [ so]b . in figure
[ so]c the blueshifted low velocity emission mostly traces a wide - angle outflow surrounding the radio jet axis . in figure
[ so]d the redshifted low velocity emission reveals two compact clumps coincident with mm1 and mc , as well as extended structures probably attributed to outflows from mm1 and/or mc . in figure
[ so]e , a redshifted clump is detected at very high velocities ; this structure can also be identified in co in figures [ co]e and [ co]f , and is apparently associated with mc .
the ch@xmath12cn ( 12@xmath3511@xmath35 ) emission is a good thermometer of dense gas surrounding high - mass protostars . assuming that all the @xmath36 components are in local thermodynamical equilibrium and tracing the same gas
, we perform a simultaneous fitting , taking into account the optical depth effect , to all the detected components by means of grid search @xmath37 minimization . from figure [ ch3cn ]
, the best fitting models agree well with the observations , and yield gas temperatures of 110@xmath38 k for mm1 , 35@xmath39 k for mm2 , and 45@xmath40 k for mc .
in addition to the ch@xmath12oh ( 8@xmath147@xmath41)@xmath42 line shown in figure [ fig1]b , thermal emission in ch@xmath12oh ( 8@xmath437@xmath44)@xmath42 and ( 3@xmath454@xmath46)@xmath42 is detected in all the three cores but appears much brighter in mm1 .
the ch@xmath12oh ( 15@xmath1716@xmath47)@xmath42 emission , however , is seen in mm1 alone ( figure [ fig1]g ) .
this is not surprising since this transition has an upper energy level ( @xmath48 ) of 374 k , dramatically higher than 3996 k for the other three transitions .
as indicated by the higher temperature inferred from the ch@xmath12cn emission , only the mm1 core possesses sufficient hot gas to populate the high - excitation ch@xmath12oh .
hnco shows a diversity similar to ch@xmath12oh ; the hnco ( 10@xmath499@xmath50 ) emission is seen in both mm1 and mc , and probably mm2 as well , but the hnco ( 10@xmath519@xmath52 ) emission is detected in mm1 alone .
lines of hnco ( @xmath53 ) are suggested as a good probe of the far - infrared field since its @xmath54@xmath550 ladders are populated primarily by far - infrared radiation rather than collisions with h@xmath11 @xcite . the warm / hot dust in mm1
is expected to produce far - infrared photons responsible for the population of the @xmath54=1 hnco , giving rise to the hnco ( 10@xmath519@xmath52 ) emission , however , the cooler dust in mm2 and mc would not provide sufficient far - infrared radiation to populate appreciable hnco at @xmath54@xmath550 .
the emission from so@xmath11 and its isotopologues exhibits a significant diversity among the three cores . besides the so@xmath11 ( 11@xmath5612@xmath57 )
emission shown in figure [ fig1]i , the lines of so@xmath11 ( 22@xmath5823@xmath59 ) , @xmath60so@xmath11 ( 11@xmath6110@xmath49 ) , @xmath60so@xmath11 ( 4@xmath623@xmath63 ) , @xmath60so@xmath11 ( 22@xmath6422@xmath65 ) , and @xmath66so@xmath11 ( 11@xmath6110@xmath49 ) , are clearly detected in mm1 . however
, none of them is seen in mc or mm2 .
the detected transitions have @xmath48 ranging from 19 to 353 k , hence the temperature variation among the cores can not solely account for the non - detection of any transition in mc or mm2 . in addition
, the so@xmath11 chemistry does not seem to be much dependent on luminosity or mass of the central forming star , as bright so@xmath11 emission was detected in low - mass star - forming regions as well ( e.g. , iras 16293 - 2422 , * ? ? ?
it is therefore the chemical effect that plays a central role in shaping the appearance of so@xmath11 in these cores . if so@xmath11 forms via the reaction of so@xmath67oh@xmath68so@xmath11 , as adopted by @xcite and @xcite , the non - detection of so@xmath11 points to the deficiency of oh in mc and mm2 , since bright so emission is detected toward all the three cores
this is plausible considering that the uv induced photodissociation of h@xmath11o may be the primary source of oh in mm1 @xcite and the formation of other molecules detected in mc and mm2 does not necessarily involve the oh radical .
the rich emission lines from so@xmath11 and its isotopologues seem to probe a more evolved stage of mm1 , in agreement with recent sulphur chemistry models @xcite . with a gas temperature of 110 k and moderately rich molecular line emission , mm1 seems to be at an early hot core phase @xcite .
of the two younger cores mm2 and mc , the latter is likely more evolved with its more active molecular line emission .
one issue to be addressed toward mc is that the core appears relatively faint in the dust continuum , which may be caused by the deficiency of dust or by a flattened density profile hence a large fraction of the emission being filtered out by the interferometer . with the existing data we can not shed more light on this issue .
however , its hot - core - like line emission is very likely excited by a high - mass protostar .
although nearby low - mass protostars my excite emission in complex organic molecules such as ch@xmath12oh and ch@xmath12cn ( e.g. * ? ? ?
* ; * ? ? ?
* ) , the emission may not be detectable at much larger distances .
quantitatively , if the dust and gas reach thermal equilibrium , the luminosity of the central source can be estimated based on the heating of the dust @xcite . for dust temperatures of 35110 k at scales of @xmath69 and a typical dust emissivity index of 1.5 , the luminosity of the central star ranges from 2@xmath70@xmath71 to 4@xmath70@xmath72 @xmath2 .
thus , the central protostars in the three cores are likely to be massive .
to obtain a more reasonable assessment of the mass and energetics of the se outflow , we calculate the gas mass down to @xmath27@xmath36 kms@xmath0 ; but to alleviate contamination from ambient gas and unrelated features we only take into account an area closely encompassing the jet - like structure . following the procedure of @xcite , we derive the outflow mass , momentum , dynamical timescale , mass rate , and momentum rate of 0.22 @xmath21 , 4.9 @xmath21 kms@xmath0 , 2.2@xmath70@xmath71 yr , @xmath73 @xmath21yr@xmath0 , and 2.2@xmath74 @xmath21kms@xmath0yr@xmath0 , respectively . considering the missing flux and optical depth effect at low velocities ,
the derived mass and energetics may represent lower limits .
the overlapping blue- and redshifted emission along the jet - like structure suggests that the flow axis is strongly inclined to the plane of sky . in projection the flow axis appears to wiggle over a small angle of @xmath2820@xmath4 .
if the flow axis wiggles along the line of sight as it does in the plane of sky , the inclination angle is likely about 10@xmath4 , half of the wiggle angle .
the outflow mass and momentum rates corrected for the inclination angle would amount to 6@xmath70@xmath73 @xmath21yr@xmath0 and 7@xmath75 @xmath21kms@xmath0yr@xmath0 , respectively .
the one - sided appearance of the outflow is probably ascribed to the peculiar environment . in projection
the central source mm2 is close ( @xmath3@xmath76 ) to the hh 8081 radio jet , which is surrounded by a biconical cavity .
the northwestern side of mm2 is likely deficient in dense gas and the outflow is expanding into the radio jet cavity , thus can not create appreciable co emission .
figures [ pmv]a and [ pmv]b show , respectively , the mass - velocity ( mv ) and position velocity ( pv , cut along the major axis ) diagrams of the se outflow .
the most interesting characteristics in the two diagrams are the distinct ehv features that appear as abrupt bumps in the mv diagram ( denoted by arrows ) and as enhanced condensations in the pv diagram ( denoted by dashed circles ) . this kind of compact and fast moving features are recognized as `` molecular bullets '' in low - mass class 0 protostellar outflows @xcite .
the bullets in the se outflow are , to our knowledge , the first discovery of molecular bullets originating from a high - mass star - forming core .
the bright b1 and b2 bullets have sizes of @xmath30.040.05 pc and masses of 10@xmath77 @xmath21 .
however , recent multi - transition sio and co observations yield hot gas ( @xmath55100 k ) in low - mass molecular bullets @xcite ; if this is the case for the b1 and b2 bullets , their masses would reach @xmath7810@xmath79 @xmath21 .
it is also worth noting that the b1 bullet is clearly detected in so ( @xmath33@xmath34 ) , suggesting shock activity in this bullet ( figure [ so]a ) . from its velocity profile shown in figure [ pmv]c , the b2 bullet has the bulk material moving at highest velocities , with a gradual decrease toward the systemic velocity .
the b1 bullet exhibits a similar but less prominent profile .
the velocity structure of the two bullets is in direct contrast to what is expected if the bullets were created in situ from entrained or swept - up ambient medium , in which case the profile would have a steep decrease toward the systemic velocity ( e.g. , * ? ? ?
* ; * ? ? ?
it is more likely that the majority of the bullet material is ejected from the close vicinity of the central protostar . in this scenario
the presence of multiple molecular bullets directly manifests an episodic nature of the outflow .
@xcite resolved h@xmath11o maser spots toward mm2 into two groups , with one group aligned in a linear distribution and the other located to the ne .
although the h@xmath11o maser spots are seen very close to mm2 ( @xmath3500 au in projection ) , we find that in both velocity and orientation the ne and linear maser groups agree well with the b1 and b2 bullets , respectively .
this further supports the scenario that the bullets are ejected from the close vicinity of the central protostar .
a sketchy and tempting interpretation is that the episodic ejections from the central protostar or star - disk system collides with the existing flow , re - accelerating the flow material to form the fast moving bullets ; the maser emission , which is pumped under critical conditions , signifies the collision sites . future proper motion measurements of the h@xmath11o maser spots and multi - transition observations of the bullets may place more stringent constraints on the excitation mechanisms of the two extreme phenomena .
high velocity , jet - like molecular outflows are often seen in low - mass class 0 protostars ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) ; in theoretical models they are physically linked to the earliest - phase disk accretion within a dense envelope @xcite .
however , jet - like outflows with high velocity emission are still rarely identified toward high - mass star - forming regions ( e.g. , * ? ? ?
based on single - dish co observations , @xcite report high velocity bullets and jets toward the well - known orion - kl outflow , but recent interferometric observations show a fan - shaped or explosive morphology of the outflow @xcite .
the se outflow has a well - defined jet - like morphology , reaching an order of 100 kms@xmath0 in both blue- and redshifted lobes , thus apparently favors a disk - mediated accretion onto the central protostar .
the ehv emission may trace part of the outer layer of an underlying wind .
moreover , the outflow contains multiple molecular bullets , indicating an episodic mass - loss process and further pointing to an episodic nature of the central accretion process .
we report the discovery of jet - like co outflows in the hh 8081 star - forming region . in particular , the outflow arising from core mm2 exhibits blue- and redshifted line wings close to 100 kms@xmath0 from the cloud velocity , and molecular bullets , which points to an episodic , disk - mediated accretion for massive star formation .
three dense cores identified in continuum and molecular line emission demonstrate physical and chemical diversities which suggest an evolutionary sequence that mm2 is younger than mc , which is in turn younger than the hot molecular core mm1 .
aspin , c. et al .
1994 , , 292 , l9 bachiller , r. , cernicharo , j. , martn - pintado , j. , tafalla , m. , & lazareff , b. 1990 , , 231 , 174 bachiller , r. , guilloteau , s. , dutrey , a. , planesas , p. , & martin - pintado , j. 1995 , , 299 , 857 beuther , h. & nissen , h. d. 2008 , , 679 , 121l beuther , h. , schilke , p. , gueth , f. , mccaughrean m. , andersen , m. , sridharan , t. k. , & menten , k. m. 2002 , , 387 , 931 beuther , h. , zhang , q. , bergin , e. a. , & sridharan , t. k. 2009 , , 137 , 406 blake , g. a. , van dishoek , e. f. , jansen , d. j. , groesbeck , t. d. , & mundy , l. g. 1994 , , 428 , 680 bottinelli , s. et al .
2004 , , 615 , 354 cesaroni , r. , neri , r. , olmi , l. , testi , l. , walmsley , c. m. , & hofner , p. 2005
, , 434 , 1039 chandler , c. j. , brogan , c. l. , shirley , y. l. , & loinard , l. 2005 , , 632 , 371 charnley , s. b. 1997 , , 481 , 396 churchwell , e. , wood , d. , myers , p. c. , & myers , r. v. 1986 , , 305 , 405 doty , s. d. , van dishoek , e. f. , van der tak , f. f. s. , & boonman , a. m. s. 2002 , , 389 , 446 fendt , c. 2009 , , 692 , 346 gmez , y. , rodrguez , l. f. , girart , j. m. , garay , g. , & mart , j. 2003 , , 597 , 414 gmez , y. , rodrguez , l. f. , & mart , j. 1995 , , 453 , 268 gueth , f. & guilloteau , s. 1999 , , 343 , 571 guilloteau , s. , bachiller , r. , fuente , a. , & lucas , r. 1992 , , 265 , l49 hatchell , j. , fuller , g. a. , & ladd , e. f. 1999 , , 346 , 278 hildebrand , r. h. 1983 , , 24 , 267 keto , e. & wood , k. 2006 , , 637 , 850 kurtz , s. , cesaroni , r. , churchwell , e. , hofner , p. , & walmsley , c. m. 2000 , in protostars & planets iv , ed .
mannings , v. , boss , a. p. , & russell , s. s. ( tucson , az : univ .
arizona press ) , 299 kurtz , s. , hofner , p. , & lvarez , c. v. 2004 , , 155 , 149 lee , c .- f .
, ho , p. t. p. , palau , a. , hirano , n. , bourke , t. , l. , shang , h. , & zhang , q. 2007 , , 670 , 1188 mart , j. , rodgrguez , l. f. , & reipurth , b. , 1993 , , 416 , 208 mart , j. , rodgrguez , l. f. , & reipurth , b. , 1995 , , 449 , 184 mart , j. , rodgrguez , l. f. , & reipurth , b. , 1998 , , 502 , 337 mart , j. , rodgrguez , l. f. , & torrelles , j. m. 1999 , , 345 , l5 nisini , b. , codella , c. , giannini , t. , santiago garcia , j. , richer , j. s. , bachiller , r. , & tafalla , m. 2007 , , 462 , 163 qiu , k. , et al .
2008 , , 685 , 1005 qiu , k. , zhang , q. , beuther , h. , & yang , j. 2007 , , 654 , 361 qiu , k. , zhang , q. , wu , j. , & chen , h. 2009 , , 696 , 66 rathborne , j. m. , jackson , j. m. , zhang , q. , & simon , r. 2008 , apj , 689 , 1141 rodrguez , l. f. , moran , j. m. , ho , p. t. p. , & gottlieb , e. w. 1980 , , 235 , 845 rodrguez - franco , a. , martn - pintado , j. , & wilson , t. l. 1999 , , 344 , l57 scoville , n. z. & kwan , j. 1976 , , 206 , 718 shang , h. , allen , a. li , z .- y .
, liu , c .- f . ,
chou , m .- y . , anderson , j. 2006 , , 649 , 845 shepherd , d. s. , watson , a. m. , sargent , a. i. , & churchwell , e. 1998 , , 507 , 861 stecklum , b. , feldt , m. , richichi , a. , calamai , g. , & lagage , p. o. 1997 , , 479 , 339 slysh , v. i. , kalenskii , s. v. , & valtts , i. e. 2002 , astronomy reports , 46 , 49 tappe , a. , lada , c. j. , black , j. h. , & muench , a. a. 2008 , , 680 , l117 zapata , l. , rodrguez , l. f. , ho , p. , menten , k. , & schmid - burgk , j. 2009 , poster presented at the millimeter and submillimeter astronomy at high angular resolution meeting , taipei , taiwan zhang , q. , ho .
, p. t. p. , wright , m. c. h. , & wilner , d. j. 1995 , , 451 , l71 zhang , q. , hunter , t. r. , beuther , h. , sridharan , t. k. , liu , s .- y .
, su , y .-
n . , chen , h .-
, & chen , y. 2007 , , 658 , 1152 zhang , q. , hunter , t. r. , & sridharan , t. k. 1998 , , 505 , l151 zhang , q. , wang , y. , pillai , t. , rathborne , j. 2009 , , 696 , 268 zinchenko , i. , henkel , c. , & mao , r. q. 2000 , , 361 , 1079 | we present submillimter array 1.3 mm waveband continuum and molecular line observations of the hh 8081 high - mass star - forming region .
the dust continuum emission reveals two dominant peaks mm1 and mm2 , and line emission from high - density tracers suggests the presence of another core mc . molecular line emission from mm1 , which harbors the exciting source of the hh 8081 radio jet , yields a hot molecular core at a gas temperature of 110 k. the two younger cores mm2 and
mc both appear to power collimated co outflows .
in particular , the outflow arising from mm2 exhibits a jet - like morphology and a broad velocity range of 190 kms@xmath0 .
the outflow contains compact and fast moving molecular clumps , known as `` molecular bullets '' first discovered in low - mass class 0 protostellar outflows .
these `` bullets '' can not be locally entrained or swept up from the ambient gas , but are more likely ejected from the close vicinity of the central protostar .
the discovery of this remarkable outflow manifests an episodic , disk - mediated accretion for massive star formation . |
our analysis of human mobility is based on a data set of 230 volunteers six - week travel diaries in frauenfeld , switzerland @xcite .
this data set contains the volunteers personal information , including age , job and sex , and 36761 trip records . by calculating the spherical distance between the origin and destination from their longitudes and latitudes ,
we can obtain the length of each trip ( see details about data in * methods * ) .
we first measure the individual displacement distributions from the data set .
1(a)-1(c ) show three typical individuals displacement distributions ( table s1 presents all volunteers displacement distributions ) , from which we can not find any universal scaling properties . indeed , when we use the _ kolmogorov - smirnov _
test @xcite to test whether the distributions fit power laws , we find that 87.8% of the individuals can not pass the test ( statistical validation results are listed in table s2 , and the details about _ kolmogorov - smirnov _ test are shown in * methods * ) . this result strongly suggests the absence of scaling laws in human travel at the individual level . to reveal the underlying structure of individual trips , we assign to each individual a mobility network , in which nodes denote locations visited by individuals , edges represent the trips between nodes and edge weight is defined as the number of corresponding trips @xcite . figs .
1(d)-1(f ) show three typical individuals mobility networks ( all networks are presented in table s1 ) . as shown in fig .
1 and table s1 , for most students and employees , their edge weights are highly heterogeneous . for each individual
, we call the trip corresponding to the edge with the largest weight the _ dominant trip _ and define the domination ratio @xmath0 as the ratio of the weight of the dominant trip to the total weight . fig . 2 reports the distribution of domination ratios for different groups of individuals , from which we can see that the student group has the largest @xmath0 on average and the employees average domination ratio is smaller than that of the students but larger than that of the other group .
the difference of @xmath0 results from the fact that students and employees frequently travel between homes and schools / workplaces in working days but retirees or homemakers do not have to do so .
the peak values in the displacement distributions of students and employees are thus usually determined by the lengths of their dominant trips . because the lengths of dominant trips are not necessarily small , the displacement distribution for an individual is usually not right - skewed and is far different from a power law .
in addition , the significant role of the dominant trip indicates that an individual s traveling process in general can not be characterized by the lvy flight @xcite or truncated lvy flight @xcite .
* individual mobility patterns . *
( a - c ) displacement distributions for three typical individuals ( ( a ) a student , ( b ) an employee , ( c ) a retiree ) , where the peak values for the student and the employee result from the trips between two most frequently visited locations .
( d - f ) mobility networks for the three individuals , where the area of a node is proportional to its number of visits and the width of an edge is proportional to its weight.,width=325 ] * distribution of the domination ratios . * ( a ) population .
( b ) student group .
( c ) employee group .
( d ) others .
@xmath1 is the number of group members , and @xmath2 is the average domination ratio.,width=325 ] * displacement distribution @xmath3 of the aggregated data . *
the solid line indicates a power law with an exponential cutoff .
the data were binned using the logarithmic binning method ( see * methods * for details).,width=325 ] * cumulative displacement distributions for a single mode of transportation . * ( a ) 12,028,929 taxi passenger trajectories in beijing .
( b ) 46,541 car trips in detroit .
( c ) 783,210 bus trips in shijiazhuang .
( d ) 205,534 air - flight travels in us.,width=325 ] the aggregated displacement distribution of individuals ( see fig .
3 ) is well approximated by a power law with an exponential cutoff @xmath4 ( the fitness significance @xmath5-value by the _ kolmogorov - smirnov _
test @xcite is 1.000 and the standard _ kolmogorov - smirnov _ distance @xmath6 is 0.039 , see * methods * and fig .
s1 for details ) , which is similar to those observed for bank notes @xcite and mobile phone users @xcite .
as shown above , this scaling property is not a simple combination of many analogous individuals .
we assume that the total travel cost is @xmath7 , the number of trips with cost @xmath8 is @xmath9 .
according to the _ maximum entropy principle _
@xcite , the two constraints , @xmath10 and @xmath11 , lead to the solution @xmath12 , where @xmath13 is the average travel cost .
denote the density of trips with cost @xmath14 by @xmath15 , then @xmath16 .
the travel cost is commonly approximated as the weighted sum @xmath17 , where @xmath18 and @xmath19 are two coefficients , and @xmath20 and @xmath21 are the costs involving time and money , respectively .
previous empirical studies have suggested that the monetary cost is approximately proportional to the travel distance as @xmath22 @xcite , while the travel time approximately obeys a hybrid form @xmath23 @xcite , where @xmath24 , @xmath25 , @xmath26 and @xmath27 are coefficients . the logarithmic term results from the mixture of modes of transportation @xcite .
apparently , people move faster when traveling longer distances : we walk from classroom to office but take an airplane from us to china .
figure s2 reports the statistics related to travel times of the data set used in this paper .
although the data set is not large enough and contains some noisy points , overall speaking , the travel time @xmath20 grows in a hybrid form as mentioned above , with @xmath28 and @xmath29 . integrating the aforementioned terms
, we obtain the displacement distribution @xmath30 , where @xmath31 and @xmath32 .
when @xmath33 is large , the distribution is approximated as a power - law with an exponential cutoff .
indeed , for the real data , @xmath34 and @xmath35 , so @xmath36 for @xmath37 , that is , the term @xmath38 can be neglected .
as shown in fig .
s3 , the corresponding fitting line is very close to a power law with an exponential cutoff ( but with a slightly higher power - law exponent 1.38 ) .
a direct corollary of maximum entropy principle is that the displacement distribution should follow an exponential form if it only accounts for trips from a single mode of transportation because in that case , @xmath39 .
this corollary gets supportive evidences from a number of empirical studies on disparate systems @xcite ( bazzani _ et al .
_ @xcite observed a slight deviation from the exponential law ) .
4 reports empirical cumulative distributions for taxi trajectories in beijing @xcite , car trips in detroit ( downloaded from _ www.semcog.org_ ) , bus trips in shijiazhuang ( collected by the authors ) and air flights in the us @xcite .
the probability density distributions are shown in fig .
s4 . all distributions can be well characterized by exponential - like functions .
the general lessons that we learned from the present analysis could be used to refine our knowledge of human mobility patterns .
the displacement distributions for aggregated data usually display power - law decay with an exponential cutoff . meanwhile , there are examples ranging from taxi trips to air flights in which the displacement distributions are exponential . in these examples ,
every displacement distribution is generated by trips involving a single mode of transportation , which corresponds to a linear relation between the travel cost and distance and eventually results in an exponential displacement distribution according to the principle of maximum entropy . in a word
, we believe the travel cost is one main reason resulting in the regularities in aggregated statistics .
the present results suggest that the form ( power law or exponential or other ) of deterrence function in the gravity law for human travel @xcite may be sensitive to the modes of transportation under consideration .
this study warns researchers of the risk of inferring individual behavioral patterns directly from aggregated statistics .
analogously , the temporal burstiness of human activities is widely observed , and the researchers are aware of the fact that the aggregated scaling laws could either be a combination of a number of individuals , each of whom displays scaling laws similar to the population @xcite , or the result of a mixture of diverse individuals , most of whom exhibit far different statistical patterns than the population @xcite . in comparison ,
such issues are less investigated for spatial burstiness .
in particular , experimental analyses on individuals has rarely been reported .
determining whether the displacement distribution of an individual follows a power - law distribution will require further data and analysis .
it is already known to the scientific community that a number of poissonian agents with different acting rates can make up a power - law inter - event time distribution at the aggregated level @xcite , and very recently , proekt _ et al . _
@xcite showed that the aggregated scaling laws on inter - event time distribution may be resulted from different time scales .
@xcite have applied similar ( yet different ) idea in explaining the aggregated scaling laws in walking behavior .
although being mathematically and technically different , this work embodies some similar perspectives , because the different transportation modes indeed assign different scales onto space : the world becomes smaller by air flights while a city is really big by walking .
elegant analogy between temporal and spatial human behaviors will benefit the studies of each other .
many known mechanisms underlie the scaling laws of complex systems @xcite , including rich get richer @xcite , good get richer @xcite , merging and regeneration @xcite , optimization @xcite , hamiltonian dynamics @xcite , stability constraints @xcite , and so on . the individual mobility model by song _
@xcite is a typical example embodying the rich get richer mechanism .
we have implemented such model . as shown in fig .
s5 , the exploration and preferential return model can well reproduce the diversity of individual mobility patterns .
in addition , for this model , the gibbs entropy of the displacement distribution at the individual level increases continuously due to the increasing number of locations as well as links connecting location pairs . however , the exploration and preferential return model does not explain why the lengths of exploration trips should follow a power law , which is a core assumption leading to the power - law - like aggregated displacement distribution .
therefore , our work has complemented ref .
@xcite and other related works in two aspects : ( i ) providing supportive empirical observation at the individual level ; ( ii ) providing alternative explanation on the emergence of scaling in aggregated displacement distribution . very recently , from the analysis on mobility patterns in an online game , szell _ et al .
_ @xcite observed a characteristic jump length and guessed that the existence of the characteristic length may be due to the single mode of transportation .
the present theory could explain their observation since a jump in such online game costs time that is proportional to the jump length .
* data description . *
this work was performed using a travel survey data set that contains 230 volunteers six - week travel diaries in frauenfeld , switzerland @xcite .
the survey was conducted among 230 volunteers from 99 households in frauenfeld and the surrounding areas in canton thurgau from august to december 2003 .
the volunteers reported their daily travel by filling out ( paper and pencil based ) self - administrated questionnaire day by day in a six - week period .
each reported trip includes the information of origin , destination and purpose .
the origin and destination of a trip were geocoded by longitude and latitude .
the quality of the geocoding is very high - with 60% of trips captured within 100 m of their true origins and destinations and
90% within 500 m. the purpose of trip was classified into work , shopping , education , home , leisure , business and other .
the data has been cross - checked to ensure the consistency and filtered to remove outliers as well as unclear and omitted destination addresses . the final cleaned data set includes 36761 trip records .
besides , the data set also contains socio - demographic characteristics of the volunteers personal information such as age , job and sex . * kolmogorov - smirnov ( ks ) test . * given an observed distribution @xmath40 , we firstly assume that it obeys a certain form @xmath41 , with a set of parameters @xmath42 , whose values are estimated by using the maximum likelihood method @xcite .
the standard ks distance is defined as the maximal distance between the cumulative density functions of the observed data @xmath43 and the fitting curve @xmath44 , namely @xmath45 .
we independently sample a set of data points according to @xmath44 , such that the number of sampled data points is the same as the number of observed data points , and then calculate the maximal distance ( denoted by @xmath46 ) between @xmath44 and the cumulative density function of the sampled data points .
the @xmath5-value is defined as the probability that @xmath47 . in this paper
, we always implement 1000 independent runs to estimate the @xmath5-value .
* logarithmic binning . *
the statistical nature of sampling will lead to the increasing noise in the tails of empirical power - law - type distributions .
applying the procedure of logarithmic binning @xcite can smooth the noisy tail .
logarithmic binning is a procedure of averaging the data that fall in the specific bins whose size increases exponentially . for each bin
the observed value are normalized by dividing by the bin width and the total number of observations ( see fig .
this work was partially supported by the national natural science foundation of china ( nnsfc ) under grant nos . 11222543 , 11205040 , 11275186 and 91024026 , program for new century excellent talents in university under grant no .
ncet-11 - 0070 , and huawei university - enterprise cooperation project under grant no .
ybcb2011057 . 99 gonzlez , m. c. , hidalgo , c. a. & barabsi , a. l. understanding individual human mobility patterns . _ nature _ * 453 * , 779 - 782 ( 2008 ) .
lu , x. , bengtsson , l. & holme , p. predictability of population displacement after the 2010 haiti earthquake .
_ * 109 * , 11576 - 11581 ( 2012 ) .
jiang , b. , yin , j. & zhao , s. characterizing the human mobility pattern in a large street network .
e _ * 80 * , 021136 ( 2009 ) .
yoon , j. , noble , b. d. , liu , m. & kim , m. building realistic mobility models from coarse - grained traces . in .
_ proc . of the acm mobisys06 _ , ( uppsala , sweden ) , pp 177 - 190 ( 2006 ) .
vespignani , a. predicting the behavior of techno - social systems .
_ science _ * 325 * , 425 - 428 ( 2009 ) .
barthlemy , m. spatial networks .
* 499 * , 1 - 101 ( 2011 ) .
balcan , d. & vespignani , a. phase transitions in contagion processes mediated by recurrent mobility patterns .
_ * 7 * , 581 - 586 ( 2011 ) .
belik , v. , geisel , t. & brockmann , d. natural human mobility patterns and spatial spread of infectious diseases .
x _ * 1 * , 011001 ( 2011 ) .
ni , s. & weng , w. impact of travel patterns on epidemic dynamics in heterogeneous spatial metapopulation networks .
e _ * 79 * , 016111 ( 2009 ) .
zhao , z .- d . ,
liu , y. & tang , m. epidemic variability in hierarchical geographical networks with human activity patterns . _
chaos _ * 22 * , 023150 ( 2012 ) .
horner , m. w. & okelly , m. e. s. embedding economies of scale concepts for hub networks design . _
j. transp .
geogr . _ * 9 * , 255 - 265 ( 2001 ) .
um , j. , son , s .- w . , lee , s .-
i . , jeong , h. & kim , b. j. scaling laws between population and facility densities .
_ * 106 * , 14236 - 14240 ( 2009 ) .
zheng , v. m. , zheng , y. , xie , x. & yang , q. collaborative location and activity recommendations with gps history data . in _ proceedings of the 19th international conference on world wide web _ , ( new york , acm press ) , pp 1029 - 1038 ( 2010 ) .
clements , m. , serdyukov , p. , de vries a. p. & reinders , m. j. t. personalised travel recommendation based on location co - occurrence .
scellato , s. , noulas , a. & mascolo c. exploiting place features in link prediction on location - based social networks . in _ proc . of the acm kdd11 _ , ( new york , acm press ) , pp 1046 - 1054 ( 2011 ) .
brockmann , d. , hufnagel , l. & geisel , t. the scaling laws of human travel .
_ nature _ * 439 * , 462 - 465 ( 2006 ) .
, hao , q. , wang , b .- h . & zhou , t. origin of the scaling law in human mobility : hierarchy of traffic systems . _ phys .
e _ * 83 * , 036117 ( 2011 ) .
song , c. , koren , t. , wang , p. & barabsi , a. l. modelling the scaling properties of human mobility .
phys . _ * 6 * , 818 - 823 ( 2010 ) .
petrovskii , s. , mashanova , a. & jansen , v. a. a. variation in individual walking behavior creates the impression of a lvy flight . _
proc . natl .
_ * 108 * , 8704 - 8707 ( 2011 ) .
chalasani , v. s. , engebretsen , .
denstadli , j. m. & axhausen , k.w .
precision of geocoded locations and network distance estimates .
_ j. transport .
* 8 * , 1 - 15 ( 2005 ) .
clauset , a. , shalizi , c. r. & newman , m. e. j. power - law distributions in empirical data .
_ siam rev . _ * 51 * , 661 - 703 ( 2009 ) .
song , c. , qu , z. , blumm , n. & barabsi , a. l. limits of predictability in human mobility .
_ science _ * 327 * , 1018 - 1021 ( 2010 ) .
balescu , r. _ equilibrium and nonequilibrium statistical mechanics_. ( new york : john wiley ) , ( 1975 ) .
willumsen , l. g. travel networks . in _ handbook of transport modelling _
, eds hensher , d. a. & button , k. j. ( new york : pergamon ) , pp 165 - 180 ( 2000 ) .
rietveld , p. , zwart , b. , van wee , b. & van den hoorn , t. on the relationship between travel time and travel distance of commuters .
sci . _ * 33 * , 269 - 287 ( 1999 ) .
li , s. , wang , h. & wang , z. a study on tour time planning of domestic sightseeing travel itineraries .
_ * 20 * , 51 - 56 ( 2005 ) .
oosterhaven , j. a. & rietveld , p. transport costs , location and the economy . in . _ location and competition _
, eds brakman , s. & garretsen , h. ( new york : routledge ) , pp 32 - 60 ( 2005 ) .
bazzani , a. , giorgini , b. , rambaldi , s. , gallotti , r. & giovannini , l. statistical laws in urban mobility from microscopic gps data in the area of florence , _ j. stat .
_ p05001 ( 2010 ) .
roth , c. , kang , s. m. , batty , m. & barthlemy , m. structure of urban movements : polycentric activity and entangled hierarchical flows .
_ plos one _ * 6 * , e15923 ( 2011 ) .
jiang , b. & jia , t. exploring human mobility patterns based on location information of us flights .
arxiv:1104.4578v2 liang , x. , zheng , x. , l , w. , zhu , t. & xu , k. the scaling of human mobility by taxis is exponential .
_ physica a _ * 391 * , 2135 - 2144 ( 2012 ) .
gallotti , r. , bazzani , a. & rambaldi , s. towards a statistical physics of human mobility .
c _ * 23 * , 1250061 ( 2012 ) .
peng , c. , jin , x. , wong , k. c. , shi , m. & li , p. collective human mobility pattern from taxi trips in urban area .
_ plos one _ * 7 * , e34487 ( 2012 ) .
simini , f. , gonzlez , m. c. , maritan , a. & barabsi , a .-
l . a universal model for mobility and migration patterns .
_ nature _ * 484 * , 96 - 100 ( 2012 ) .
barabsi , a. l. the origin of bursts and heavy tails in human dynamics .
_ nature _ * 435 * , 207 - 211 ( 2005 ) .
malmgren , r. d. , stouffer , d. b. , motter , a. e. , & amaral l. a. n. a poissonian explanation for heavy tails in e - mail communication .
_ * 105 * , 18153 - 18158 ( 2008 ) .
hidalgo , c. a. conditions for the emergence of scaling in the inter - event time of uncorrelated and seasonal systems .
_ physica a _ * 369 * , 877 - 883 ( 2006 ) .
wu , y. , zhou , c. , xiao , j. , kurths , j. & schellnhuber , h. j. evidence for a bimodal distribution in human communication .
_ * 107 * , 18803 - 18808 ( 2010 ) .
proekt , a. , banavar , j. r. , maritan , a. & pfaff , d. w. , scale invariance in the dynamics of spontaneous behavior , _ proc .
_ * 109 * , 10564 - 10569 ( 2012 ) .
mitzenmacher , m. a brief history of generative models for power law and lognormal distributions . _ internet math . _ * 1 * , 226 - 251 ( 2004 ) .
newman , m. e. j. power laws , pareto distributions and zipf s law .
_ contemp .
phys . _ * 46 * , 323 - 351 ( 2005 ) .
simkin , m. v. & roychowdhury , v. p. re - inventing willis .
_ * 502 * , 1 - 35 ( 2011 ) .
simon , h. a. on a class of skew distribution functions .
_ biometrika _ * 42 * , 425 - 440 ( 1955 ) .
barabsi , a. l. & albert , r. emergence of scaling in random networks .
_ science _ * 286 * , 509 - 512 ( 1999 ) .
l , l. , zhang , z .- k .
& zhou , t. deviation of zipf s and heaps laws in human languages with limited dictionary sizes .
* 3 * , 1082 ( 2013 ) .
garlaschelli , d. , capocci , a. & caldarelli , g. self - organized network evolution coupled to extremal dynamics .
phys . _ * 3 * , 813 - 817 ( 2007 ) .
zhou , t. , medo , m. , cimini , g. , zhang , z .- k . & zhang , y .- c .
emergence of scale - free leadership structure in social recommender systems .
_ plos one _ * 6 * , e20648 ( 2011 ) .
kim , b. j. , trusina , a. , minnhagen , p. & sneppen , k. self organized scale - free networks from merging and regeneration .
j. b _ * 43 * , 369 - 372 ( 2005 ) .
valverde , s. , cancho , f. & sol , r. v. scale - free networks from optimal design .
lett . _ * 43 * , 369 - 372 ( 2002 ) .
bartumeus , f. , da luz , m. g. e. , viswanathan , g. m. & catalan , j. animal search strategies : a quantitative random - walk analysis
. _ ecology _ * 86 * , 3078 - 3087 ( 2005 ) .
baiesi , m. & manna , s. scale - free networks from a hamiltonian dynamics .
e _ * 68 * , 047103 ( 2003 ) .
perotti , j. i. , billoni , o. v. , tamarit , f. a. , chialvo , d. r. & cannas , s. a. emergent self - organized complex network topology out of stability constraints .
lett . _ * 103 * , 108701 ( 2009 ) .
szell , m. , sinatra , r. , petri , g. , thurner , s. & latora , v. understanding mobility in a social petri dish .
* 2 * , 457 ( 2012 ) .
milojevi , s. power law distributions in information science : making the case for logarithmic binning .
tec . _ * 61 * , 2417 - 2425 ( 2010 ) . | uncovering human mobility patterns is of fundamental importance to the understanding of epidemic spreading , urban transportation and other socioeconomic dynamics embodying spatiality and human travel . according to the direct travel diaries of volunteers ,
we show the absence of scaling properties in the displacement distribution at the individual level , while the aggregated displacement distribution follows a power law with an exponential cutoff .
given the constraint on total travelling cost , this aggregated scaling law can be analytically predicted by the mixture nature of human travel under the principle of maximum entropy .
a direct corollary of such theory is that the displacement distribution of a single mode of transportation should follow an exponential law , which also gets supportive evidences in known data .
we thus conclude that the travelling cost shapes the displacement distribution at the aggregated level . positioning systems in mobile phones and vehicles and wi - fi devices in laptop computers and personal digital assistants
have made quantitative analyses of human mobility patterns possible @xcite .
these analyses have a significant potential to reveal novel statistical regularities of human behavior , refine our understanding of the socioeconomic dynamics embodying spatiality and human mobility @xcite , and eventually contribute to controlling disease @xcite , designing transportation systems @xcite , locating facilities @xcite , providing location - based services @xcite , and so on .
aggregated data from bank notes @xcite , mobile phones @xcite and onboard gps measurements @xcite showed that the displacement distribution of human mobility , for both long - range travel and daily movements , approximately follows a power law .
the scaling laws in long - range travel may result from the hierarchical organization of transportation systems @xcite , while the scaling laws in daily movements have recently been explained by the _ exploration and preferential return _
mechanism @xcite .
thus far , we still lack solid results about human mobility patterns at the individual level .
inferring individual features from the aggregated data is very risky because the scaling law for the population could be a mixture of many individuals with different statistics @xcite .
in addition , the aforementioned data are not sufficient to draw conclusions at the individual level .
first , data such as gps records from taxis and the trajectories of bank notes consist of many individual movements , but these individuals are not easy to be distinguished from each other .
second , data such as gps records from mobile phones and the trajectories of bank notes could not accurately capture purposeful travels with explicit origins and destinations .
in fact , the displacement between two activations of a mobile phone may be just a tiny portion of a purposeful trip or a combination of several sequential trips , while the displacement between two registrations of a bank note could be the result of a number of sequential trips made by different people . instead of using proxy data
, we analyze the travel diaries of hundreds of volunteers .
though the data set is small , it contains personal profiles and explicit positions of origins and destinations , allowing quantitative and authentic analyses at the individual level .
in contrast to the scaling laws in aggregated data , individuals show diverse mobility patterns , and few of them display the scaling property .
in fact , the trajectories of students and employees are dominated by trips connecting homes with schools and workplaces , respectively , while trips are distributed more homogeneously among different locations for others such as retirees , homemakers and unemployed people .
the aggregated displacement distribution follows a power law with an exponential cutoff , which can be analytically explained by the mixed nature of human travel under the principle of maximum entropy . in addition
, this theory predicts that the displacements using a single mode of transportation will follow an exponential distribution , which is also supported by the empirical data on taxi trips , car trips , bus trips and air flights . |
in the minimal supersymmetric standard model ( mssm ) the electroweak symmetry is broken via the higgs mechanism , leading to five physical scalars : three neutral ( h , h , a ) and two charged ( h@xmath0 , h@xmath1 ) higgs bosons . at tree level , two parameters , usually chosen to be the mass of the pseudo scalar higgs boson ( @xmath2 ) and the ratio of the vacuum expectation values of the two higgs fields ( @xmath3 ) , determine the masses and couplings in the higgs sector .
radiative corrections , chiefly coming from the top - stop and , at large @xmath3 , from the bottom - sbottom sector , introduce further model parameters . in the frequently studied decoupling limit ( @xmath4 ) ,
the lightest neutral higgs boson is standard model - like , while h and a are close in mass and both couple with @xmath3 to down - type fermions and with @xmath5 to up - type fermions .
the dominant production mechanism for heavy neutral mssm higgs bosons at large @xmath3 is gg @xmath6 bbh / a with negligible contributions from other processes .
as the main decay mode h / a @xmath6 bb suffers from high qcd background , the decay h / a @xmath6 @xmath7 , reaching about 10% rate for @xmath810 , promises the most sensitivity .
charged higgs bosons couple to fermions as h@xmath0ud @xmath9 @xmath10 .
they are produced in top quark decay , if @xmath11 , or in association with a top quark , if @xmath12 .
light charged higgs bosons decay almost exclusively to a tau lepton and a neutrino for @xmath13 . for @xmath14 ,
the decay h@xmath15 @xmath6 tb dominates with an important contribution of h@xmath15 @xmath6 @xmath16 at large @xmath3 .
the following processes are considered here : * associated bbh / a production followed by h / a @xmath6 @xmath7 with all possible @xmath7 final states ( jet jet , e jet , @xmath17 jet and e @xmath17 ) @xcite ; * production of a light charged higgs boson in tt @xmath6 h@xmath15bwb with subsequent h@xmath15 @xmath6 @xmath18 , @xmath19 @xmath6 jet and w @xmath6 @xmath20 decays @xcite ; * associated h@xmath15t(b ) production of heavy charged higgs boson with h@xmath15 @xmath6 @xmath21 , @xmath19 @xmath6 jet and hadronic top quark decays @xcite ; * associated h@xmath15t(b ) production of heavy charged higgs boson with h@xmath15 @xmath6 tb and one of the top quarks decaying leptonically @xcite .
the final states in the search for heavy mssm higgs bosons are complex : they typically contain leptons , b- and light - flavoured hadronic jets and missing transverse energy .
the selections rely on @xmath19-jet or lepton ( e , @xmath17 ) triggers .
@xmath19-identification based on vertex reconstruction and impact parameter measurements is powerful against hadronic jets .
a single b - tag is sufficient to suppress events from drell - yan , qcd multi - jet and w+jets processes . against the difficult tt
background a central jet veto is applied .
missing transverse energy ( @xmath22 ) reconstruction is important to account for the neutrinos from the higgs or the @xmath19 decays .
top quark and w boson mass reconstruction provides a further handle on the background .
for example , in the bbh / a , h / a @xmath6 @xmath7 @xmath6 @xmath23 + jet searches , it is instrumental in vetoing leptons from w decay by reconstructing the transverse mass of the [ lepton , @xmath22 ] system .
the main background comes from tt events , with significant contribution from w+jets events in the h@xmath15 @xmath6 @xmath21 searches and drell - yan processes in the h / a @xmath6 @xmath7 @xmath6 @xmath23 + jet searches .
a major exception is the h / a @xmath6 @xmath7 @xmath6 jet jet analysis , where qcd events dominate the background . the typical selection efficiency is below 1% .
_ higgs boson mass reconstruction _ does not only increases the sensitivity of the searches , but it will also play a crucial role in constraining the mssm model parameters . in the h / a @xmath6 @xmath7 selection
the higgs mass can be measured by the di - tau mass assuming that the @xmath19 decay products are collinearly emitted .
while the procedure has large inefficiency , it achieves a mass resolution of about 20% . in the heavy
charged higgs search : h@xmath15 t @xmath6 @xmath21 bqq , the reconstruction of the top quark mass is instrumental in fully reconstructing the signal .
the charged higgs mass is then estimated by the transverse mass of the [ @xmath19 , @xmath22 ] system ( @xmath25 ) . by requiring @xmath26100 gev ,
an almost background - free selection is achieved .
the results presented here are based on the full simulation of the cms detector at low luminosity ( @xmath27 @xmath28s@xmath29 ) , assuming a total integrated luminosity of 30@xmath3060 fb@xmath29 .
pythia is used as the main monte carlo generator with notable exceptions ( e.g. matrix element generators for multi - parton or toprex for tt final states in some cases ) .
background cross - sections are normalized to nlo calculations where available .
tau decays are modeled by tauola . for the numerical results ,
the minimal supergravity inspired @xmath31-max benchmark scenario is used , which maximises the theoretical upper bound on @xmath31 for a given @xmath3 , @xmath32 and @xmath33 , the susy mass scale .
higgs boson masses , cross - sections and branching ratios are calculated by feynhiggs 2.3.2 . the estimated 5-sigma discovery reach is shown on figure [ fig : h](a ) for the pp @xmath6 bbh / a searches and on figure [ fig : h](b ) for the h@xmath15 @xmath6 @xmath21 searches @xcite . in all cases the systematic uncertainties
are included in the calculation of signal significance .
5-sigma discovery reach of searches for ( a ) bbh / a production with h / a @xmath6 @xmath7 , ( b ) h@xmath15 @xmath6 @xmath21 and ( c ) inclusive h @xmath6 @xmath34 and vector boson fusion qqh / h with h / h @xmath6 @xmath7 .
, title="fig : " ] 5-sigma discovery reach of searches for ( a ) bbh / a production with h / a @xmath6 @xmath7 , ( b ) h@xmath15 @xmath6 @xmath21 and ( c ) inclusive h @xmath6 @xmath34 and vector boson fusion qqh / h with h / h @xmath6 @xmath7 .
, title="fig : " ] 5-sigma discovery reach of searches for ( a ) bbh / a production with h / a @xmath6 @xmath7 , ( b ) h@xmath15 @xmath6 @xmath21 and ( c ) inclusive h @xmath6 @xmath34 and vector boson fusion qqh / h with h / h @xmath6 @xmath7 .
, title="fig : " ] \(a ) ( b ) ( c ) we also considered the extended @xmath31-max scenario , where the supersymmetric higgs mass parameter can take different values : @xmath35 .
the cross - section is enhanced for large negative and reduced for large positive values of @xmath17 , but the change in the higgs branching ratios due to decay modes to supersymmetric particles partially compensates this effect . in general , the default value of @xmath36 gev gives the most conservative discovery reach .
for a detailed discussion see @xcite . _ systematic uncertainties _
dominate these searches .
therefore , several methods are explored to measure them from collision data @xcite .
the main sources of uncertainty come from b - tagging , @xmath19- and lepton identification , missing energy measurement and energy scale calibration . together with the theoretical uncertainties on the background production cross - sections , we estimate typically 12% error on the main tt background , 9% on z/@xmath37 , 16% on bbz/@xmath37 , 10@xmath3014% on w+jets and 15% on tw processes .
the qcd background will be measured from data with a 5@xmath3020% statistical error .
the inclusion of systematic uncertainties has a significant impact on the expected sensitivities .
for example , in the charged higgs boson h@xmath15 @xmath6 @xmath16 search , the 5-sigma reach is decreased from about 125 to 110 gev at the most difficult value of @xmath38 . at values of @xmath39 ,
the downward shift is even larger ranging from about 30 to 80 gev . in the search for gg @xmath6 h@xmath15 tb with h@xmath15 @xmath6 tb and tt
@xmath6 @xmath20b qqb requiring four b - tags , the main background comes from ttbb and mistagged tt + jets processes . with the simulation of the background processes by comphep ( with the cross - section calculated by alpgen ) and the inclusion of realistic experimental ( about 5@xmath305% on b - tagging efficiency and the mistag rate ) and theory uncertainties ,
no discovery potential remains for this channel .
the searches presented above for heavy higgs bosons loose their sensitivity for small and intermediate @xmath3 and leave open the so - called lhc wedge region , where only a light neutral higgs boson can be discovered .
this is illustrated on figure [ fig : h](c ) , where the sm higgs boson searches @xcite ( for inclusive pp @xmath6 h+x production with h @xmath6 @xmath34 and for the vector boson fusion process qq @xmath6 qqh / h with h / h @xmath6 @xmath7 @xmath6 @xmath23 + jet ) are reinterpreted in mssm @xcite .
the author is partially supported by otka grant nk67974 . | the discovery potential of the cms detector for the mssm neutral and charged higgs bosons at the lhc is presented based on studies with full detector simulation and event reconstruction of the principal discovery channels . |
it has been recently demonstrated that , in addition to energy spectral localization ( the fermi - pasta - ulam recurrence phenomenon @xcite ) and to energy spatial localization ( soliton generation from initial data @xcite ) , a nonlinear chain of oscillators possess another striking fundamental property called _ nonlinear supratransmission _ @xcite .
this phenomenon was shown to occur in the nonlinear sine - gordon chain , which possess a natural forbidden band gap , when it is submitted to irradiation at a frequency in the stop gap .
while in a linear chain the signal would exponentially vanish in the medium , it does not do so in the nonlinear case if its amplitude exceeds a threshold value .
in @xcite the irradiation of the medium was modelized by prescribing the boundary value at one end of the chain .
it is worth mentioning that nonlinear transmission does occur also in the case of a true wave scattering , namely when a monochromatic plane wave scatters onto a nonlinear medium with a frequency in the gap @xcite . in the case of the sine - gordon model
, the threshold can be predicted exactly by invoking the static one - breather solution @xcite . in short
, energy penetrates the medium as soon as the amplitude @xmath0 of the harmonic driving , at a frequency @xmath1 in the gap , exceeds the maximum amplitude of the static breather of frequency @xmath1 .
this energy travels then through the medium by means of nonlinear localized excitations ( kinks , breathers , solitons ) .
the purpose of this paper is , after recalling results of @xcite , to display experimental results on the mechanical chain of pendula and to discuss the generality of our result by exploring extensions to different situations .
we shall in particular explore the robustness of nonlinear supratransmission when the medium experiences damping , different nonlinearities and different classes of boundary values .
the next section is devoted to a short reminder of the results published in @xcite and it is intended to settle formalism and basic facts about nonlinear supratransmission .
the model is the sine - gordon chain submitted to a dirichlet condition at the origin on a vanishing initial background ( pendula at rest ) .
then section [ sec : experiment ] relates the experiment made with a mechanical pendula chain , coupled by means of a coil spring , and which is forced at one end by a periodic torque .
a systematic exploration of the chain response in a frequency range within the gap shows spectacular agreement with the theory . in section [ sec : charact ] we numerically describe the characteristics of the breathers generated in the sine - gordon chain and discover some simple relations between the parameters of the emitted breather with respect to those of the boundary driving . the section [ sec : energy ] deals with the energy transmitted by the nonlinear medium as a function of the driving amplitude at given frequency in the gap .
particular emphasis will be put on the effectiveness of the effect in different , not necessarily integrable , cases .
the nonlinear instability which is the generating mechanism of nonlinear supratransmission is briefly discussed in section [ sec : instab ] .
although the mathematical analysis is still to be constructed we propose a quite simple illustration of the process by a perturbative analysis of the sine - gordon system driven close to a breather mode . as another domain of study , we consider in section [ sec : josephson ] the sine - gordon model for neumann conditions at the boundary ( the derivative at the origin is prescribed ) .
this is a model for a chain of short josephson junctions whose first one is submitted to an external ac current . by using numerical simulations ,
the process of nonlinear supratransmission is shown to hold , and the threshold of energy transmission to obey a similar simple rule .
remarkably , in this case the energy flows by means of kink ( or anti - kinks ) and not by breathers ( or kink anti - kink pairs ) .
this is an interesting issue in view of applications as kinks are the objects that have an experimental signature ( through the josephson current ) .
the following normalized discrete sine - gordon chain of @xmath2 locally damped coupled oscillators @xmath3 @xmath4 is submitted to the boundary value at the origin ( dirichlet condition ) @xmath5 acting on a medium initially at rest , namely @xmath6 in this section the damping coefficient @xmath7 is used to model a semi - infinite chain by an absorbing boundary .
more precisely we take @xmath8 with @xmath9,\end{aligned}\ ] ] where the parameter @xmath10 is adjusted to have a damping factor @xmath11 varying slowly from almost @xmath12 to almost @xmath13 on the last @xmath14 particles .
a typical experiment will have e.g. @xmath15 , @xmath16 , @xmath17 and @xmath18 .
note that in section [ sec : josephson ] we shall use instead the reflective boundary condition @xmath19 for an homogeneous damping @xmath20 for all @xmath21 .
the equation is considered as a second - order ordinary differential system for the @xmath2 coupled oscillators @xmath3 .
this system is then solved with the subroutine dsolve of maple software package which uses a fehlberg fourth - fifth order runge - kutta method .
the linear dispersion relation @xmath22 of the chain is @xmath23 for a driving boundary with frequency in the forbidden band gap ( fbg ) , namely @xmath24 , a linear chain would sustain the solution ( evanescent wave ) @xmath25.\ ] ] the parameter @xmath26 is given from the dispersion relation written for @xmath27 and @xmath28 , namely @xmath29 in the nonlinear case , in order to fit the boundary condition , the _ approximate _ solution ( exact in the continuous limit ) is given instead by the static breather @xmath30,\ ] ] where the breather center @xmath31 solves @xmath32.\ ] ] the above equation expresses that the static breather fits the driving field by adjusting its position such as to match the amplitude @xmath0 .
the spectrum of the breather signal does not match exactly the monochromatic forcing and it adapts by sending phonons at third , fifth , etc ... , harmonic frequencies .
the above assertion is checked by performing two simulations of at a given forcing frequency , say @xmath33 , and a given amplitude , say @xmath34 , in one case with the harmonic forcing , in the other case with the breather - like forcing @xmath35 where the value of @xmath31 is calculated from with @xmath34 . to avoid initial shock we also set initial velocities matching the ones of the static breather .
then we evaluate by fourier transform the spectra of one particle of the chain in both cases ( we have selected the particle 50 on a chain of 100 pendula ) .
the result is displayed on figure [ fig : spectra - comp ] where we see that phonons at frequency @xmath36 are indeed emitted for harmonic forcing while no phonon appear for a breather - like forcing .
the small contribution at frequency @xmath37 in figure ( a ) represents a collective motion of the pendula generated by the initial shock resulting from inadequation of the initial velocities with the boundary value .
the adjustment of the breather center @xmath31 provided by not always has a solution .
indeed it works up to the maximum value @xmath38 of the breather amplitude realized for @xmath39 . beyond this threshold , for a driving boundary with @xmath40 , we have shown in @xcite that nonlinear supratransmission occurs , i.e. the medium starts to transmit energy by means of nonlinear modes generation ( breathers and kink - antikink pairs ) . from , the threshold @xmath41 reads as the following function of the frequency @xmath1 @xmath42,\ ] ] which has the approximate value @xmath43.\ ] ] which would hold for the continuous sine - gordon equation obtained in the limit @xmath44 .
note that the maximum difference ( when @xmath1 varies in @xmath45 $ ] ) between the above two expressions is already @xmath46 for @xmath47 and goes down to @xmath48 for @xmath49 , the value we have used in most of the numerical simulations .
this qualitative definition of a bifurcation threshold can be checked on numerical simulations of with the boundary condition by varying , at given frequency @xmath1 , the amplitude @xmath0 around the above value @xmath50 ( for @xmath51 ) .
there are many means to determine appearance of nonlinear supratransmission , a simple one being the observation of the motion of one particle of the chain . as an example we display in figure [ fig : motions ] the motion of the particle @xmath52 of a chain of @xmath53 particles driven at frequency @xmath54 at amplitudes @xmath55 ( no supratransmission ) and @xmath56 ( supratransmission ) for a coupling factor @xmath57 .
each large oscillation in the second figure corresponds to a breather passing by .
two of them are generated and cross the site @xmath52 at times @xmath58 and @xmath59 .
the small oscillation seen between the humps are the harmonic phonons , mainly of frequency @xmath36 . for illustration
we show on figure [ fig : actual ] a picture of oscillators amplitude @xmath3 at given time ( here @xmath60 ) in the case of nonlinear supratransmission obtained for @xmath61 and @xmath62 .
a first breather propagates to the right while a second one is just being generated near the origin .
note the amplitude of the breather with respect to the one of the driving boundary . by a systematic exploration of the chain response
we draw figure [ fig : bif ] obtained for @xmath53 particles with a coupling @xmath63 ( some experiments have been actually made with smaller coupling and less number of points to shorten computation times ) for a typical time of @xmath53 ( for frequencies close to the gap value @xmath37 , time had to be increased up to @xmath64 ) .
the points on figure [ fig : bif ] are obtained with an absolute precision of @xmath65 for the amplitude @xmath0 .
they are compared to the theoretical threshold expression ( continuous curve ) .
the figure [ fig : bif ] shows excellent agreement to formula apart for some discrepancies starting below @xmath66 and @xmath67 .
this results from the driving which , thanks to the nonlinearity , generates phonons at multiple frequencies ( here third , fifth , etc ... ) .
if these frequencies lie in the phonon band , the phonons move away from the boundary and have no effect on the forcing .
if however they lie in the fbg , the related phonons do not propagate ( which we call phonon quenching ) and stick on the boundary where they add contribution to the driving .
this effect indeed disappears when driving the system with the exact breather boundary value @xmath35 for which we have checked that nonlinear supratransmission _ never occurs _ at an amplitude @xmath68 , while it occurs already at @xmath38 ( or for very small deformation of the perfect breather ) .
our purpose here is to show that the process of nonlinear supratransmission can be easily realized experimentally on a chain of coupled pendula as depicted in figure [ fig : chaine ] .
this chain has been built following m. remoissenet @xcite .
the pendula rotate freely around a piano wire stretched between two supports , they are coupled together by a coil spring tightened to each pendulum by a screw .
such a chain of 48 pendula is driven by an electrical engine steered by a generator of sinusoidal tension . upon varying the frequency , at low amplitude , in the phonon band
, we can determine the parameters of the chain ( coupling constant @xmath69 and angular eigenfrequency @xmath70 ) by measuring the wavelength @xmath71 of produced wave ( in units of pendula number ) .
by comparison to the dispersion relation of the linear chain @xmath72 we determined the following parameter values @xmath73 then in a time normalized to the eigenfrequency , we have the dispersion relation with @xmath74 which is the fundamental parameter of the model , entering in particular expression of the threshold amplitude for nonlinear supratransmission .
this is the formula we want here to confront to experiments .
we proceed with a systematic exploration of the chain response to a signal frequency in the gap ( of value less than @xmath70 ) .
the method consist in increasing slowly , at fixed frequency , the amplitude of the driving , up to the time when a nonlinear mode is seen to be generated ( a picture of a breather generated at a frequency of @xmath75 in normalized units is displayed in @xcite ) . repeating 3 times , for each driving frequency value , the measurements of the driving amplitude that generates a nonlinear mode
, we eventually obtain the figure [ fig : exp - res ] which displays the measured threshold amplitude in terms of the signal frequency ( normalized ) . there
, the full line curve is the function @xmath76 given in with @xmath77 . despite the small number of pendula ,
inducing reflection at the open end , the inherent damping , and other mechanical imperfections , the experiments provide a spectacular realization of the theoretical threshold prediction .
one important issue concerns the nature and the characteristics of the nonlinear modes which propagate in the medium . in the first instance
, we have to check that the emitted structures correctly match the moving breather , at velocity @xmath78 , which is convenient to write as @xmath79\ .\ ] ] by choosing correctly the two parameters @xmath80 and @xmath81 which represent respectively the amplitude and the group velocity of the breather , together with the space and time origins of the solution @xmath82 , one can match it to the asymptotic numerical solution with high precision .
this enables us to identify correctly the emitted breathers , to determine their characteristic with a great accuracy , and to observe their possible disintegration into kink - antikink pairs .
the results of such estimations are displayed in figure [ fig : ampl - breath ] obtained as follows . for each driving frequency @xmath1
, we have driven the system at threshold amplitude @xmath76 and then determined at the the parameters @xmath83 of the emitted breathers ( a given experiment at threshold driving produces repeatedly the same breather ) .
then the amplitude of the propagating breather is plotted as a function of the amplitude of excitation @xmath76 for various frequencies .
similarly , figure [ fig : vit - breath ] shows the group velocity of the breather as a function of @xmath76 .
both figures [ fig : ampl - breath ] and [ fig : vit - breath ] correspond to a region of the driving frequency close to the gap @xmath84 . for lower frequencies the emitted breathers ,
when they occur , are unstable and decay into kink - antikink pairs ( this will be discussed later ) .
the main conclusion here is the existence of a linear relation between the amplitude @xmath85 ( and the velocity @xmath81 ) of the generated breather and the driving amplitude at threshold @xmath41 .
so far we have no theoretical interpretation of these observations .
another characteristic of the breather is its proper frequency @xmath86 , related to the apparent pulsation period @xmath87 by the usual ( relativistic ) relation @xmath88 this frequency is plotted in figure [ fig : freq - breath ] as a function of the driving frequency @xmath1 and , again , we note a non trivial linear dependence .
we now turn to the stability of the emitted breathers which , for @xmath89 decay into kink - antikink pairs .
this can be qualitatively described by studying the binding energy @xmath90 , given by @xmath91 where @xmath92 is the single kink energy and @xmath93 the breather energy @xcite .
we have evaluated the binding energy @xmath90 of the breathers produced in the simulations ( by plugging in ( [ binding - e ] ) the measured values of @xmath80 and @xmath81 ) , the result of which is plotted in figure [ fig : energ - breath ] in terms of the frequency @xmath1 of the driving boundary .
this figure shows that the binding energy of the breather decreases with the driving frequency , and goes to zero for @xmath94 .
thus the breather decay in the frequency range @xmath95 $ ] is well understood from formula ( [ binding - e ] ) .
nonlinear supratransmission is a process where large amount of energy flows through the medium .
our purpose here is to evaluate numerically this energy for amplitudes around the threshold value .
the theoretical expression of the energy flow is calculated hereafter for a generic nonlinearity deriving from a potential energy @xmath96 ( the sine - gordon case corresponds to @xmath97 ) . from the energy density @xmath98 and the evolution equation
follow the conservation law @xmath99 with the current @xmath100 incorporating the potential energy resulting from the coupling of the first particle @xmath101 to the boundary @xmath102 , the _ total energy _ of the system reads @xmath103 in our case @xmath104 is the driving and the chain is supposed infinite with @xmath105 as @xmath106 . upon time derivation , with help of the conservation law , and using the assumed asymptotic @xmath107 for large @xmath21 ,
we arrive eventually at @xmath108 hence the total energy injected in the medium during time @xmath87 reads @xmath109.\ ] ] choosing for @xmath87 an integer multiple of the period of excitation makes this energy to vanish identically in the linear case if the driving frequency falls in the fbg . in the nonlinear case ,
expression is computed numerically on a chain of 60 particles with a coupling parameter @xmath57 and an absorbing end working over the last 40 particles . for a driving frequency @xmath54 and amplitudes running from @xmath110 to @xmath111
, we obtain the figure [ fig : sg-09 ] where the bifurcation is seen to occur for @xmath112 , the value predicted by formula .
this simulation has been run for frequencies in the range @xmath113 $ ] , with comparable expected results .
our approach stems from the existence of a breather solution of the model equation , allowing to determine the threshold amplitude .
it is then worth wondering if the process is robust _ against _ non - integrability . to give a partial still instructive answer to this question
, we have performed numerical simulations of two non - integrable evolution in the same class , i.e. with hamiltonian .
first the taylor truncated expansion of sine - gordon ( fifth order is kept to ensure a confining potential at large @xmath114 ) reads as the nonlinear klein - gordon chain : @xmath115 this system is solved with the boundary driving ( [ forcing ] ) and the energy ( [ energy ] ) is computed for the same parameter values as for figure [ fig : sg-09 ] .
the result is displayed on figure [ fig : kg-09 ] . by scanning the frequency range in the gap ,
we have obtained that the process occurs down to @xmath116 and then disappears .
we suspect that for such a polynomial potential energy , at high forcing amplitude , the _ incoming wave _ sees an almost parabolic potential , while for low amplitude driving the _ incoming wave _ does feel the actual structure of the local potential .
another interesting non integrable hamiltonian evolution where the local potential has a periodic structure is the double sine - gordon chain @xmath117=0 , \quad n=1,\cdots , n\ .\ ] ] once again we have solved this system in the same situation as before and obtained the brutal energy flow of figure [ fig : dsg-09 ] . in this case the process holds for any frequency ( tested from @xmath118 ) just as in the sine - gordon case .
thus we have seen that in those two cases nonlinear supratransmission does work which is a strong indication that it is a generic nonlinear process .
an interesting question is the mechanism that generates nonlinear supratransmission , which is discussed now .
the process at the origin of nonlinear supratransmission is a nonlinear instability .
although it is not yet fully understood , we discuss here some of its relevant aspects . to that end
we explore the properties of the sine - gordon chain submitted to initial - boundary value data that precisely mimic the breather @xmath119 centered at @xmath39 as defined in and which is convenient to write as @xmath120 note that this is an exact solution in the continuous limit @xmath121 and @xmath122 . in order to study the behavior at the threshold
, we impose the boundary value @xmath123 together with compatible initial data @xmath124 the boundary value at the other end of the chain ( @xmath125 ) can be taken as an absorbing end to simulate the infinite line or as the breather value @xmath126 , e.g. to check the accuracy of the solution .
the parameter @xmath127 measures the departure from the exact solution .
we are of course interested in what happens when @xmath127 is positive .
for @xmath128 we simply generate the approximate solution @xmath119 which in the numerical simulations is marginaly stable ( and would be indeed stable for a breather centered in @xmath129 ) . in the case
@xmath130 , the numerical simulations of the sine - gordon model immediately generate nonlinear supratransmission . from the initial - boundary value problem , it is natural to seek a solution as a perturbation of the breather under the form @xmath131 which by obeys at order 1 in @xmath127 @xmath132 the variable coefficients @xmath133 and @xmath134 of this equation are given by @xmath135 the initial - boundary value problem that goes with can be taken to results from , namely @xmath136 with the related initial data .
then we observe the following fundamental facts : * for @xmath137 the system is stable showing a long period oscillatory behavior , * for @xmath130 the system is unstable showing an exponential growth of the oscillations .
these observations are illustrated by figure [ fig : lin ] where we have computed the energy transmitted by the chain according to formula , as a function of time ( the points in time are chosen as integer multiples of the period @xmath138 ) .
the parameters related to this experiments are @xmath15 , @xmath49 and the energy has been calculated for 10 points ( up to @xmath139 ) . we have also displayed results for the _ linear version _ of system ( read with @xmath128 ) which is marginaly unstable showing a linear growth of the oscillations . the exponential growth in the case @xmath130 is the signature of the instability which is the mechanism of nonlinear supratransmission .
the mathematical approach of this instability for is reported to future studies .
a josephson junction behaves just like a single pendulum , the rotation amplitude being replaced by the phase difference between wave functions of cooper pairs @xcite . by connecting the junctions in parallel with super - conducting wires , the resulting model equation is just the sine - gordon discrete system submitted to damping and constant torque @xcite @xmath140 here above @xmath7 is the constant damping along the array , @xmath141 is the normalized intensity of the applied current and time has been normalized to the plasma frequency .
we consider here only the one - dimensional geometry but it is worth mentioning the coupled arrays ( ladders ) who revealed as a means to generate discrete breathers by convenient initial conditions @xcite with subsequent striking experimental realizations @xcite .
the applied current can have a dc component ( the so - called bias ) and an ac driving part .
the problem we are interested in is the behavior of the above chain initially at rest and whose first junction only is submitted to ac driving at a frequency @xmath1 in the gap and intensity @xmath142 .
then the model results as for @xmath143 and the bias @xmath144 together with the following relation for @xmath145 : @xmath146 it is important to remark that the above equation can be equivalently written as the _ discrete neumann condition _
@xmath147 when the system is assumed to hold also for @xmath145 .
this remark actually allows us to consider the continuous limit in order to determine the prediction of the threshold of nonlinear supratransmission .
the continuous version of and boundary value , for the variable @xmath148 and @xmath149 , reads as the system @xmath150 where we have defined @xmath151 .
this continous version corresponds to a long josephson junction whose extremity @xmath152 is submitted to external micro - wave irradiation , the amplitude @xmath153 being then related to the external magnetic field intensity @xcite .
this is a neumann boundary condition for the sine - gordon continuous equation .
the system will then adapt the breather _ derivative _ at the boundary centered in @xmath154 , namely @xmath155 where now the continuous version of the dispersion relation is @xmath156 for evanescent waves ( due to the change of space variable , we have @xmath157 where @xmath26 is defined in ) . upon varying the position @xmath158 , the above expression has a maximum value for @xmath159 given by @xmath160 and the related maximum amplitude of the derivative eventually results as the simple expression @xmath161 this is the threshold prediction for the neumann condition and we are going now to check it on numerical simulations . as in section
[ sec : sg ] , we solve the discrete system with no constant bias ( @xmath162 ) , without damping in a first stage ( @xmath163 ) , for a chain of @xmath52 particles with an absorbing boundary on the last @xmath164 , and with a coupling factor @xmath165 .
the result is displayed on figure [ fig : bif - jj ] which shows good agreement except maybe around the frequencies @xmath66 and @xmath67 as in figure [ fig : bif ] ( for the same reasons ) and close to the phonon band , which is due to the absorbing end .
here above the situation is that of a quasi - continuous undamped and semi - infinite sine - gordon chain submitted to neumann condition a the origin .
we turn back now to the discrete case , a one - dimensional finite - length ( open - ended ) array of coupled short josephson junctions , where the first pendulum is submitted to an external ac driving .
namely we consider the _ josephson transmission line _
system @xmath166 nonlinear supratransmission becomes here the property of the jtl to transmit energy under the form of kinks ( or anti - kinks ) as soon as the intensity @xmath142 of the ac driving of the first junction exceeds the threshold @xmath167 as given by the definition @xmath168 and from expression .
a new fact here is that , for neumann type boundary conditions as , only kinks ( or anti - kinks ) are produced , not breathers .
this is the result of the fact that producing a kink cost half of the energy used to produce a breather , and of the freedom left on the boundary value @xmath104 ( only the difference @xmath169 is prescribed ) allowing full @xmath170-rotations .
for dirichlet boundary condition as in preceding sections , the precription of @xmath104 , e.g. by prevents full rotation . to illustrate this property we draw in figure [ fig : out - kink ] a typical simulation with the following set of parameters @xmath171 constituting a reasonable choice for an experimental situation . at frequency @xmath116 ,
the threshold is @xmath172 and for an ac driving at amplitude @xmath173 the chain starts to _ rotate _ , which means that a number of kinks ( elementary @xmath170 rotation ) are generated by the ac driving . in the case of figure [ fig : out - kink ] about @xmath174 kinks have been generated during @xmath175 units of normalized time .
note that we have assumed a small constant bias dc current @xmath176 in order to select kinks instead of possible anti - kinks .
however the process does work without bias , though in some of the experiments the rotation would stop after some time ( for a reason that we do not understand ) .
finally we mention that we have selected compatible boundary condition , namely , first an ac forcing that starts from zero and slowly reaches the value @xmath142 within @xmath177 units of time , second the following initial positions and velocities @xmath178 then the nonlinear supratransmission does work in such jtl , despite the small number of oscillators , the presence of damping , the free - end boundary condition and the presence of constant bias . in order to compute the transmitted energy in the present situation , we start with expression but written for a _ finite number
_ @xmath2 of oscillators , i.e. @xmath179 as before , this expression is differentiated and , by use of the conservation law , the free boundary condition @xmath19 leads to the same result , namely @xmath108 now , to compute the total energy injected in the medium during time @xmath87 , it is necessary to use integration by parts together with the neumann condition and to chose for @xmath87 an integer multiple of the period of the driving .
we eventually obtain ( @xmath180 in the numerical simulations ) @xmath181 this is the quantity that we have computed , for each value of the ac - driving @xmath142 , in figure [ fig : jtl - ener ] .
the bifurcation of energy translission is now quite clear on the graph of the energy @xmath182 transmitted to the chain displayed in figure [ fig : jtl - ener ] obtained for @xmath116 and the parameters given in .
the ability of a nonlinear medium to transmit energy when submitted to periodic boundary driving at a frequency in a stop gap and at an amplitude beyond a threshold value , which we called nonlinear supratransmission , has been shown to have some universality ( so far in the domain of nonlinear wave type of equations ) .
the process does not rely on integrability and is robust against damping , discreteness ( in a reasonable range ) , finiteness and different classes of boundary values .
the mechanism at the origin of this process is a nonlinear instability which is still under study .
moreover , the nonlinear supratransmission has a simple experimental realization in the pendula chain which works surprisingly well .
some other experimental results are expected in josephson transmission lines . | a nonlinear system possessing a natural forbidden band gap can transmit energy of a signal with a frequency in the gap , as recently shown for the nonlinear chain of coupled pendula [ phys rev lett 89 ( 2002 ) 134102 ] . this process of _ nonlinear supratransmission _ , occurring at a threshold exactly predictable in many cases , is shown to have a simple experimental realization with a mechanical chain of pendula coupled by a coil spring .
it is then analyzed in more detail by first going to different ( non - integrable ) systems which do sustain nonlinear supratransmission .
then a josephson transmission line ( one dimensional array of short josephson junctions coupled through super - conducting wires ) , is shown to sustain also nonlinear supratransmission , though being related to a different class of boundary conditions , and despite the presence of damping , finiteness and discreteness .
finally the mechanism at the origin of nonlinear supratransmission is a nonlinear instability , and it is briefly discussed here . |
only a few detections of extrasolar planets atmospheric species are reported so far , but they have been recognized as important steps in our understanding of these objects . of particular interest are _ hubble space telescope _ ( hst ) observations of hd209458b during primary transit , which yielded the detection of nai ( @xcite ) , as well as hi , oi , and cii ( @xcite ) , hi ( @xcite ) , rayleigh scattering by h@xmath8 ( @xcite ) , along with upper limits on the presence of tio and vo ( @xcite ) and recently confirmation of cii and detection of siiii ( @xcite ) and even possibly siiv ( @xcite ) .
hd189733b is the second transiting planet for which hst observations yielded information on the atmospheric transmission spectrum .
hst / nicmos has been used to detect h@xmath8o and ch@xmath9 ( swain et al . 2008 ) , though the h@xmath8o feature has been challenged by a more sensitive search using filter photometry ( @xcite ) showing the presence of haze condensate at high altitude as previously observed in the optical using hst / acs ( @xcite ) .
in addition to hst , ground based , high resolution spectra have also allowed sodium detection in hd189733b ( @xcite ) , confirmed sodium in hd209458b ( @xcite ) , detected co and even winds in the planetary atmosphere ( @xcite ) , while the other important alkali metal potassium has been detected in xo2b using narrowband photometry ( @xcite ) .
sing et al .
( 2008a ) revisited all the existing hst / stis spectroscopic data of hd209458b transits gathered over the years .
their result is an estimate of the transit absorption depth ( ad ) as a function of wavelength from 3000 to 8000 .
the variation of absorption depth as function of wavelength has been interpreted by atmospheric signatures of na , h@xmath8 rayleigh scattering and tio vo molecules ( @xcite ) .
another important result obtained with the analysis of this transit absorption spectrum is the observation of a sharp drop in sodium abundance above the altitude corresponding to the absorption depth of 1.485% ( @xcite ) .
this drop is clearly seen in the profile of the sodium absorption line , which differs from the classical `` eiffel tower '' shape found when there is a constant sodium abundance ( _ e.g. _ , seager & sasselov 2000 ) .
the sodium line shows a plateau with an almost constant absorption depth around [email protected]% at several hundred angstroms around the line center .
closer to the line center in the line core , the absorption depth increases again with a shape more like the `` eiffel tower '' above the plateau , which shows that the sodium is still present above the altitude of the abundance drop . to match the width of the core of the sodium line , the drop in sodium abundance should be at least a factor of 10 ( @xcite ) .
moreover , if the interpretation of the ad rise at wavelengths shorter than @xmath5 5000 in terms of rayleigh scattering by h@xmath8 molecules is correct , then the pressure at the altitude of the abundance drop is constrained to be about 3 mbar ( @xcite ) .
/co equilibrium curve .
the dashed lines show the condensation curves for sodium , iron , and titanium . ]
the fit of the hd209458b absorption spectrum including profiles of sodium and rayleigh scattering by h@xmath8 yielded ( sing et al .
2008b ) an estimate of the temperature pressure ( t p ) profiles for the atmosphere from @xmath530 mbar to @xmath510@xmath10bar ( see also madhusudhan & seager ( 2009 ) for related issues when retrieving temperature and abundances from transit data ) .
this result has been obtained using a parametric profile to fit the data , assuming linear variations in temperature with altitude and two layers with constant slope of temperature versus altitude ( fig .
[ fig : tpprofile ] ) . nonetheless , with this method ,
two different profiles provide an equally good fit to the absorption spectrum : the first profile involves na condensation to explain the abundance drop at the absorption depth of 1.485% ; in the second profile the temperature is not low enough to allow the sodium condensation , and the decrease in sodium absorption at altitude above the plateau is interpreted by ionization . the two resulting t
p profiles are 1 . in the first profile the temperature decreases down to less than 900 k , causing a sodium abundance drop by _ condensation _ at about 3 mbar ( fig .
[ fig : tpprofile ] ) ; 2 . the second profile ( not shown in fig .
[ fig : tpprofile ] ) is nearly isothermal at about 2200 k , in which case the abundance drop is explained by partial _ ionization _ of atomic sodium above the 3 mbar level in a scenario proposed by fortney et al .
( 2003 ) .
both profiles do not present variations similar to model predictions as calculated in the literature specifically for hd209458b . the direct comparison between extracted observational t
p profiles and model predictions will be made in sect .
4 ( fig .
[ fig : t - p ] ) . to further reduce the impact of a priori hypothesis on the data interpretation , we propose here a new analysis of the absorption depth profile with particular attention to the inner part of the nai doublet .
this new analysis is done using a non
parametric approach , with no a priori assumption about any expected behaviour related to the t
p profile or to supposed nai abundance variations .
the hope is to be able to distinguish between the two scenarios described above , avoiding a priori assumptions . in sect .
[ sec : re ] we present the method for the data analysis . in sect .
[ sec : variations ] we compare the observations to the model calculations .
the results are discussed in sect .
[ sec : discuss ] where a new description of the upper atmospheric layers of hd209458b is proposed .
corresponding to a sharp decrease , [ ii ] to a plateau and [ iii ] to a more gentle decrease .
analysis of these three regions will be carried out starting from region [ iii ] inward.,scaledwidth=48.0% ] sing et al .
( 2008a ) evaluated the transit absorption depth around the center of the nai lines in variable bandwidths from 4.4 to 100 centered over both lines . at 12
bandwidth , the measured absorption depth agrees with the charbonneau et al .
( 2002 ) value . in the sing et al .
( 2008a ) approach , the absorption depth is evaluated _ relatively _ to a reference absorption depth in a fixed double spectral band ( noted _ ref _ ) on both sides of the nai lines : from 5818 to 5843 and from 5943 to 5968 , _
i.e. _ in a fixed 50 region .
the measurements are obtained by first integrating the raw data over the selected bandwidth before extraction of the absorption depth . because the measurements correspond to the total absorption depths within various bands , including the line center as done with filters , this absorption spectrum as a function of the bandwidth
can thus be regarded as a spectrum obtained with a `` photometric '' approach . the resulting `` photometric '' spectrum is plotted in fig . [ fig : obs1 ] . in another approach ,
sing et al .
( 2008a ) also extracted _ spectroscopically _ the absorption depth distribution over each pixel in the whole spectral range covered by the observations .
these spectroscopic absorption depths can also be averaged over different spectral bandwidths as in the photometric approach .
we calculated the difference between the spectrum averaged over various bands centered on the nai lines and the spectrum averaged in the reference band , _ ref _ , defined above .
the result obtained in this `` spectroscopic '' approach ( see figure [ fig : obs1 ] ) can be directly compared with `` photometric '' result .
small differences do show up between the two extraction methods , but they are well within the error bars .
the main differences probably arise from the limb darkening correction , which is applied at each pixel wavelength in the spectroscopic approach , while the correction is a weighted average limb darkening correction in the considered bands in the `` photometric '' approach .
therefore , the `` spectroscopic '' is expected to provide a better correction for the limb darkening variations within the stellar lines themselves and in particular within the two strong nai doublet lines . in the following ,
we use the spectroscopic absorption depths . in figure
[ fig : obs1 ] the measured absorption depths have been translated into planetary radii , i.e. , altitudes above the _ ref _ level , defined by the level of the absorption depth in the _ ref _ spectral domain . to translate ad into altitudes above the _ ref _ level , we used r@xmath11=1.32r@xmath12 ( @xcite ) for the planetary radius and an absorption depth of ad@xmath13 = 1.49145% at the _ ref _ level .
as expected , the observed altitudes ( _ i.e. _ ad ) vary as a function of the bandwidth with a slow rise toward smaller bandwidths ; however , three different regions could be seen in terms of ad variations : a region [ i ] ( see fig . [
fig : obs1 ] ) presenting a rapid decrease of the ad with increasing bandwidths followed by a kind of plateau ( region [ ii ] ) , a puzzling feature discussed in sect .
[ sec : discuss ] followed by a last region [ iii ] where a slow decrease in the ad could be seen .
these observations are not independent . in particular , the error bars are larger than the observations fluctuations .
indeed , all observations made in a given bandwidth are present in observation extracted in any larger bandwidth . for that reason
, we will complete our analysis of the observed behavior of the ad by starting from the larger bandwidths and moving inward .
this approach will give us first access to the atmospheric information at the lower altitudes and then , by eliminating this information in the narrower bandwidths , move progressively upwards in the atmosphere .
the variations in the absorption depth ( or variations of the corresponding altitude ) along the line profile are due to variations of the line opacity . at each wavelength , the absorption line probes different altitudes in the planetary atmosphere with a resolution on the order of the error bar on the absorption estimate , _
i.e. _ about 0.01% in absorption depth corresponding to about 315 km in altitude . because the absorption cross section increases toward wavelengths closer to the core of the lines , the absorption depth increases toward narrower bandwidths ( fig .
[ fig : obs1 ] ) . these variations in absorption altitude ( or absorption depth )
can be interpreted through comparison with model calculations . to interpret the measured variations in absorption with bandwidths
, we built the simplest possible model , _
i.e. _ an isothermal hydrostatic uniform abundance model ( _ ihua _ ) . the model with isothermal hydrostatic atmosphere at temperature @xmath14 with a single sodium abundance value ( [ na / h ] , relative to the solar abundance ) , and
a pressure @xmath15 at a reference level , is a three parameter model ( @xmath14,[na / h],@xmath15 ) .
this model is aimed at a better understanding of the observed absorption depth _
variations _ in terms of departures from the _ local _ _ ihua _ conditions , _
i.e. _ in terms of variations in temperature or sodium abundance , without any _ a priori _ hypothesis on the shape of the vertical temperature pressure profile as needed in sing et al .
( 2008a ) . despite its simplicity , assuming hydrostatic equilibrium , this ( _ ihua _ ) model still holds _ locally _ to interpret observational measurements because the atmospheric scale height is on the order of the observational accuracy , _
i.e. _ @xmath5 300 km for typical temperatures in the hd209458b atmosphere .
the exponential decrease in the volume density with altitude is very steep ; as a result , the variations in the measured absorption at a given altitude are directly related to the physical conditions ( temperature , abundance ) in the atmosphere at the corresponding altitude .
the model absorption depth is calculated in a straightforward manner following lecavelier des etangs et al .
( 2008a , b ) .
the total column density along the line of sight passing through the terminator at different altitudes is given by fortney et al .
( 2005 ) . the optical depth , @xmath16 , in a line of sight grazing the planetary limb at an altitude @xmath17 is given by @xmath18 where @xmath19 is the planetary radius , @xmath20 the atmosphere scale height , and @xmath21 the volume density at the altitude @xmath17 of the main absorbent with a cross section @xmath22 .
the scale height is given by the relation @xmath23 where @xmath24 is the boltzmann constant , @xmath14 the temperature , @xmath25 the mean mass of atmospheric particles taken to be 2.3 times the mass of the proton , and @xmath26 the gravity at the planetary radius , @xmath27 with @xmath28 equal to the planetary mass and g the gravitational constant .
the absorption at a given wavelength , @xmath29 is calculated by finding the altitude at which the optical thickness is @xmath30 , where @xmath31=0.56 ( @xcite ) . using the equations and quantities defined above , the effective altitude @xmath17 is given by @xmath32}\times p_{z=0}\times \sigma_{nai}(\lambda ) } { \tau_{eff}}\times\sqrt{2\pi r_p / kt \mu g}\right ) \label{z_lambda}\ ] ] where @xmath33 is the nai lines cross section
. equation [ z_lambda ] shows that the altitude ( or corresponding absorption depth ) depends on pressure and abundance , but the variation in the altitude as a function of wavelength does not depend upon pressure or abundance , but only upon the scale height , which is directly and only related to @xmath14 , the temperature at the corresponding altitude .
therefore , the interpretation of fig .
[ fig : obs1 ] ( in particular the slope of the absorption depth curve as a function of wavelength ) provides the temperature profile in the altitude range probed by the sodium line core : from 100 to 4000 km altitude .
this temperature profile can have a typical resolution close to the atmospheric scale height . as shown by eq .
[ z_lambda ] , the absolute values of the pressure and abundance impact only the absolute absorption altitudes and not their variations .
moreover , there is a degeneracy between these two quantities because only the product @xmath34\times p_{z=0}$ ] can be constrained by the measurement of the absorption altitude . in other words ,
a change in the abundance can be directly compensated by a similar change in the pressure .
that is the total quantity of gas in the line of sight .
one could assume that the pressure at the reference level , @xmath15 for our nai study in the core of the lines , _
i.e. _ as evaluated over the reference domain @xmath35 over which the ad@xmath13 evaluation is made and relative to which all our differential ad evaluations are made , is given by the h@xmath8 rayleigh scattering observed at wavelengths shorter than @xmath55000 ( @xcite ) .
this leads to a reference pressure @xmath363 mbar as shown in fig .
[ fig : tpprofile ] .
this pressure corresponds to the location where the nai abundance drops and above which only the core of the nai lines do give an absorption signature . to be more precise
, one could alternatively use the evaluated nai abundance , @xmath34\sim 2 $ ] found in the lower layers of the atmosphere by sing et al .
( 2008a ) , to evaluate the corresponding pressure at the reference altitude .
indeed with only the product @xmath34\times p_{z=0}$ ] being defined , by using the evaluated nai abundance in the lower layers one can deduce the pressure level sampled by the nai lines at any spectral region .
this is given by the following relation deduced from eqs .
[ tau_lambda ] and [ z_lambda ] .
@xmath37}\times\sigma_{nai}(\lambda ) \times \sqrt{2\pi r_p / kt \mu g}}\right ) .
\label{p_lambda}\ ] ] ( solid line ) .
twice the solar nai abundance is assumed . a temperature of 900k ( corresponding to the sodium condensation temperature ) is also assumed and the two spectral reference domains are indicated as vertical long dashed lines . in those regions the evaluated pressure ( thick solid lines ) is indeed on the order of 3 mbar as indicated by the horizontal dashed line.,scaledwidth=48.0% ] the result of eq .
[ p_lambda ] is shown in fig .
[ fig : ref_pressure ] for which the assumed abundance is twice the solar abundance , valid up to our reference level .
the pressure broadening in the nai line wings is larger than the natural width broadening at pressures above 29 mbar as it could be scaled from the corresponding values at 1 bar and 2000 k where the inverse lifetime related to pressure broadening is @xmath38@xmath39s@xmath40 to be compared to the natural inverse lifetime which is @xmath41s@xmath40 ( @xcite ) . because here we are analyzing the inner parts of the nai lines ( between the two reference domains , see fig .
[ fig : ref_pressure ] ) , we only investigate the upper part of the atmosphere where lower pressures are present .
we thus can ignore here the pressure broadening effects .
consequently , @xmath42 in eq .
[ p_lambda ] was evaluated by taking only into account the doppler core broadening and the natural wing broadening of both nai doublet lines .
we find that , using a temperature of 900 k and sodium abundance twice solar , absorption over the two spectral reference domains , on both sides of the nai lines from 5818 to 5843 and from 5943 to 5968 , is due to sodium located at a pressure of 3 mbar ( fig .
[ fig : ref_pressure ] ) .
if a different temperature is selected for the reference level ( as for instance 2000 k as suggested in the second possible scenario ) , then the evaluated pressure at the reference level would be slightly higher and on the order of 6 mbar , a value still within the pressure domain where the pressure broadening effect could be ignored . with this new reference pressure , the t
p profile obtained here would still hold with only a shift in the reference level from 3 to 6 mbar . once the reference pressure has been set
, our _ ihua _ model becomes a two parameter model ( @xmath14,[na / h ] ) and the selection of a different reference pressure only slightly shifts our temperature evaluations along the pressure scale , with the sodium abundance modified by the same ratio .
= 500 k and @xmath14=1000 k for various sodium abundances ranging from [ na / h]=10@xmath43 , 10@xmath44 , 10@xmath45 , 10@xmath46 , 10@xmath47
( top to bottom ) . ] in the framework of our study , we consider _ relative _ absorption depths , calculated by the difference in absorption depths between two spectral domains : the first one for a given bandwidth centered on the sodium lines , the second one in the constant _ ref _ domain . if the [ na / h ] abundance varies , the absolute absorption depth of all spectral domains changes by nearly the same amount ( see eq .
[ z_lambda ] ) . as a result ,
the _ relative _ absorption depth in the sodium line does not depend on the sodium abundance as shown in figure [ fig:1 ] .
indeed a change in the abundance produces about the same shift in the absorption depth at all wavelengths .
therefore , the difference between the absorption depth in a given bandwidth and the absorption depth in the reference wavelength domain ( ad ad@xmath13 ) barely depends on the sodium abundance . on the contrary
, the slope of the models clearly changes with temperature .
the sodium abundance thus can not be estimated by the transmission spectrum of _ relative _ absorption depth .
however , the variations in the _ relative _ absorption depth does allow estimating the local temperature @xmath14 at the absorption altitude . in effect , the variation in the absorption altitude with wavelength depends on the variation in the horizontal integrated density with altitude . for instance , if one considers two wavelengths @xmath48 and @xmath49 , with corresponding line cross section @xmath50 , a grazing line of sight will have the same optical thickness at these two wavelengths if they are at altitude @xmath51 at @xmath48 and @xmath52 at @xmath49 such that the variation in partial density of the absorbing element @xmath53 between @xmath51 and @xmath52 compensates for the different cross sections .
the optical thickness at these two wavelengths are equal if @xmath54 .
therefore , the variation in the observed absorption altitude as a function of wavelength ( equivalent to variation as a function of line cross section ) reveals the variation in the partial density at the considered altitude . at hydrostatic equilibrium ,
the density variation with altitude is determined by the temperature through the characteristic scale height .
as a consequence , the temperature constrains the slope of the absorption altitude as a function of wavelength .
the higher the temperature , the larger the slope . for a given absorption feature with a known cross section as a function of wavelength ,
the temperature @xmath14 can be derived from the measurements of the absorption altitude spectrum using ( @xcite ) : @xmath55 140k , 540k ( solid lines ) , shifted in ordinate toward the lower end of the first error bar at 105 in order to show the constraints on the model slopes ( not affected by the applied shift ) related to the observational error bars .
this leads to an estimate of the temperature at the lower levels to be 340@xmath56200k .
similar errors on the order of @xmath56200k are found for each observational region of about one scale height ( see text ) . finally , the 900k model ( dotted line ) corresponding to the sodium condensation temperature ( see fig . [
fig : tpprofile ] ) is also shown to illustrate its incompatibility with the observations.,scaledwidth=48.0% ] figure [ fig : compt ] shows the expected variations for fixed abundances of nai over the whole atmosphere , overplotted with isothermal _ ihua _ models at different temperatures .
as explained above , the higher the temperature , the higher the scale height in the atmosphere and thus the steeper the slope . here
, the na abundance corresponding to each temperature is assumed to be constant , while the model representations in figure [ fig : compt ] are arbitrarily vertically shifted ( thus keeping the same slope at each given wavelength ) in order to match the observations above the reference level . with _ ihua
_ models matching the observations up to 800 km in altitude . at the lowest altitude ,
the temperature is found to be @xmath5 340 k ( solid line ) . at altitude above @xmath5
200 km , we identified layers with different temperatures at about 550 k ( dashed line ) , 910 k ( dotted line ) , then back to 550 k ( dashed line ) .
a few error bars are indicated to underline that , in the lower altitude region [ iii ] from 0 to 800 km above the reference level , the observations are on the average incompatible with an _ ihua _ model slope corresponding to @xmath5900 k , the sodium condensation temperature , demonstrating that condensation has to take place in the planetary atmosphere.,scaledwidth=48.0% ]
using the method described above , we start by evaluating the temperature at altitudes just above the reference level ( altitude 0225 km , bandwidths 75110 ) , with the reference level corresponding to the level where the sodium abundance drops sharply ( @xmath53 mbar level ) . the temperature of the atmosphere just above the reference level ( see fig . [
fig : compt ] ) is found by the _ slope _ of the observed variations ( see eq .
[ t_dz ] ) . in this atmospheric layer ( region [ iii ] as shown in fig .
[ fig : obs1 ] ) , the absorption depth varies slowly as a function of the bandwidth . as a consequence ,
we find the lower temperature close to @xmath5 340 k. because this temperature appears to be extremely low , the assumption of uniform abundance is likely not to be valid for the scale height just above the reference level , a region probably still affected by the na depletion mechanism ( see sect . [ the na abundance profile ] ) . from eq .
[ z_lambda ] , the measured altitude is determined by the quantity of @xmath34\times\sigma_{nai}(\lambda)$ ] . if the na abundance drops with increasing altitude , for a given optical depth through the atmosphere , the na cross section needs to be larger to compensate , making the transmission spectrum optically thick closer to the na core . as a result , a depletion in sodium causes a shallower slope ( or even a plateau ) in the observed absorption depth versus bandwidth . in other words , if the temperature is measured by the slope of the absorption altitude as a function of wavelength assuming a constant abundance , a decrease in sodium abundance with altitude implies underestimation of the temperature ( eq . [ eq : t ] ) .
the temperature of 340 k is therefore a lower limit of the temperature in the layer located in the altitude range from 0 to 200 km above the reference level . in this layer ,
the temperature is likely higher and the abundance of sodium is very likely decreasing with altitude .
apart from the shallow slope of absorption as a function of the bandwidth described above , we observe two sodium absorption depth plateaus at 80 and 55 bandwidth .
this can also be interpreted as abundance drops at the corresponding altitudes .
we however ignore these two plateau regions because they are entirely immersed within the error bars domain ( see fig . [
fig : compt ] ) .
we thus assume locally constant abundance ( at least over one scale height ) to estimate the local temperature in the different layers of the atmosphere . for narrower bandpass , _
i.e. _ corresponding to regions [ i ] and [ ii ] at higher altitudes from 800 km to 4000 km.,scaledwidth=48.0% ] the temperature profile can thus be obtained step by step toward increasing altitude by analyzing the observed ad variations in narrower bandpasses . when starting from wavelength bandwidths of 110 , working toward shorter bandwidths , and assuming constant abundance , the ad variations can be interpreted in terms of temperature changes as a function of altitude with the following characteristics ( see fig .
[ fig : evaltnash ] ) : + i ) a lower limit of 340 k in temperature is obtained at the lowest altitudes , + ii ) between altitudes of 200 km and 400 km , the temperature is found to be about 550 k , + iii ) the ad slope constraints the temperature at about 910 k in the layer between 400 km and 600 km altitude , + iv ) the temperature seems to drop back close to 550 k between 600 km and 800 km altitude . + for each layer considered , the error bars on these temperature evaluations are of about @xmath57 k , similar to the error estimate made on the lowest temperature level at 340 k as shown in fig .
[ fig : compt ] .
therefore , the observed temperature variations correspond to a marginal rise in temperature with altitude .
most importantly , for about three scale heights from 200 to 800 km above the altitude where the nai abundance is seen to drop sharply , the atmospheric temperatures are found to be within the range @xmath5 550 910 k , _
i.e. _ below the sodium condensation temperature ( figure [ fig : tpprofile ] ) .
this result favors the sodium condensation scenario to explain the sodium abundance drop as described in ( @xcite ) .
recent global climate models of hd209458b also suggest cool temperatures of @xmath5 800 k ( at mbar pressures ) are possible on the night side of the planet and terminator ( @xcite ) , further supporting this sodium condensation model . for the narrowest bandpass ( region [ i ] ) .
near to the line center , the cross section is the highest , therefore probing the highest altitudes .
the ground based measurements obtained by snellen et al .
( 2008 ) are plotted with crossed squares together with the spectroscopic evaluations ( diamonds ) of sing et al .
( 2008a ) .
the variations of the absorption depth ( ad ) between 1.5 and 3 bandpass are consistent with a temperature from 1500 to 4000 k with the most plausible value on the order of 2500 k.,scaledwidth=48.0% ] the temperature at higher altitudes can be estimated by looking for the ad variations over narrowest bandpasses . there , an important and significant leveling off ( region [ ii ] ) of the ad variation at about 800 km altitude , between 32 and 20bandwidths , is seen ( note in fig . [ fig : evaltnashhigher ] its position relative to the observational error bars ) .
this plateau in absorption depth might be interpreted as a local drop in temperature but this would correspond to an extremely low temperature of less than 100 k !
this plateau is more likely related to a decrease in sodium abundance which will be further discussed in sect .
[ the na abundance profile ] . above that long
leveling off , _
i.e. _ above 800 km altitude ( region [ i ] ) , the ad increases again rapidly toward shorter bandwidths ( from 15 bandpass and below ) , revealing there a rise in temperature . using an _ ihua _ model fit to the ad profile
, we find a typical temperature of about 1500 k from about 800 km up to 4000 km altitude ( fig . [ fig : evaltnashhigher ] ) .
the measurement at 4000 km corresponds to the narrowest 3bandpass in the very center core of the line in hst observations .
it is noteworthy that this 1500 k atmospheric layer is observed to be isothermal over nearly 6 scale heights . over this wide altitude range ,
the error bar on the estimated temperature is rather small ( @xmath56100 k ) .
there the temperature is tightly constrained in the 1400k1600k range .
after publication of the hst data analysis by sing et al .
( 2008a ) , new measurements of the nai absorption depth have been obtained by snellen et al .
( 2008 ) using ground based observations .
these measurements have been obtained in narrower bandpasses , providing larger absorption signatures , therefore corresponding to absorptions at higher altitudes .
these new measurements are shown in fig .
[ fig : evaltnashhighest ] together with those of sing et al .
( 2008a ) .
although these various measurements have been obtained using very different techniques , and despite that ground based measurements are known to be extremely difficult because of the earth atmospheric perturbation , the very good agreement between the various data sets strengthens the confidence in those measurements permitting us to extend our analysis in altitude up to more than 6500 km . on the contrary , we decided not to use the recent ground based observations of langland shula et al .
( 2009 ) because their result suffers from uncertainties in the process of continuum fitting and subtraction of the earth atmosphere absorption .
we find that to match the variations of the absorption depth toward the narrowest 1.5 bandwidth as measured by snellen et al .
( 2008 ) , the temperature at altitudes above 4000 km needs to be on the order of 2500 k and within a 15004000 k range ( fig .
[ fig : evaltnashhighest ] ) .
an important comment here before presenting the result is to recall that the evaluated t
p profile is extracted from quantities `` integrated '' along the cords passing through portions of the limb presenting gradients , _ e.g. _ from the night to day sides , as well as `` averaged '' over the planetary limb , _
e.g. _ from equatorial to polar regions . in a first step , such an `` integrated averaged '' t
p profile could be compared to 1d model calculations , although it is clear that later steps , when observations permit it , would be to compare them to more sophisticated 3d models as illustrated in _
e.g. _ the iro et al .
( 2005 ) and the recent fortney et al .
( 2010 ) studies . from the temperature
altitude relation obtained in sects .
[ t at lower altitudes ] and [ t at higher altitudes ] , and assuming hydrostatic equilibrium , we can calculate the temperature pressure ( t p ) profile .
we start by using a reference pressure of 3 mbar at the reference 0 km altitude .
the pressure is calculated step by step , starting from the zero altitude level .
the result is shown in fig .
[ fig : t - p ] . acknowledging that the sodium condensation occurs at temperatures below 800900 k , the temperatures measured in the layer between 0 and 800 km above the reference level are found to be low enough to allow condensation .
this condensation seems to take place at pressures between 3 mbar and 10 @xmath25bar .
as already mentioned in the introduction , at pressures above 3 mbar , the evaluated observational profiles do not present variations similar to the model predictions as calculated in the literature specifically for hd209458b ( _ e.g. _ barman et al .
2002 ; burrows et al . 2003 ; iro et al . 2005 ; barman et al . 2005 ; fortney et al . 2008 ; showman et al .
2008 ; showman et al . 2009 ; guillot 2010 ; fortney et al . 2010 ) .
three of them ( barman et al . 2005 ; showman et al . 2008
; guillot 2010 ) are presented in the lower part of fig .
[ fig : t - p ] .
it is interesting to note that all temperature predictions at about 3 mbar are in good agreement with our evaluation simply made from the direct estimation of the scale height at these levels .
we also point out the large variations revealed by the two near terminator limb ( _ i.e. _ near the transit observed limb ) model calculations from guillot ( 2010 ; one for @xmath58 at the terminator limb and the other at @xmath59 close to the terminator limb ) as also shown in fig .
[ fig : t - p ] .
this may indicate that indeed such model predictions are difficult to compare to observations according to their extreme sensitivity to the precise limb geometry and , as already said , to the integrated nature of the observations along a grazing line of sight .
also , one could argue that at higher pressures , it is not surprising that observations ( and their interpretation ) differ from theoretical models , since for instance , a large heat advection could be critical in the troposphere leading to strong departures from equilibrium temperature profiles . above this condensation layer , a progressive rise in the temperature with altitude is needed to explain the profile of the sodium line toward narrower bandpasses .
the temperature in the upper atmosphere is found to reach @xmath52500 k at the very low pressures of @xmath53 to @xmath590 nbar .
this is likely the signature of the base of the thermosphere , which is the expected transition between the cool lower atmosphere and the hot upper atmosphere where hydrodynamic atmospheric escape takes place .
this scenario agrees with theoretical models of the atmospheric escape as published by , _
e.g. _ , lecavelier des etangs et al .
( 2004 ) , yelle ( 2004 ) overplotted in fig .
[ fig : t - p ] , tian et al .
( 2005 ) , yelle ( 2006 ) , muoz ( 2007 ) , and murray - clay et al .
( 2009 ) . a summary of our t
p profile is given in table [ table:1 ] .
the scale heights @xmath20 are given for each temperature layer .
this shows that the layers have thicknesses of at least one or more scale heights . with a pressure change by a factor of 2.7 over one scale height under hydrostatic equilibrium , fig .
[ fig : t - p ] shows that the bottom layer at the lowest temperature of @xmath60 340 k has a thickness of nearly 2 scale heights .
the three following layers at 550 k and 910 k are over about 1 scale height , while the atmospheric layer with a constant temperature at @xmath51500 k extends over nearly 6 scale heights .
finally the last upper layer at about 2500 k extends over more than 3 scale heights .
the calculated t
p profile is valid when assuming local hydrostatic equilibrium .
this assumption could be invalid because of the hydrodynamical escape of the upper atmosphere .
figure [ fig : t - p ] shows that the theoretical escape velocities triggered by the hydrodynamic flow as evaluated by yelle ( 2004 ) are negligible below the altitude corresponding to a pressure of @xmath510@xmath6 bar .
this shows that the local hydrostatic equilibrium assumption is valid over our studied range , even at the highest altitudes where we estimate a temperature of about 2500 k. .evaluated atmospheric parameters [ cols="^,^,^,^,^,^,^ " , ] the sodium abundance [ na / h ] can not be constrained in absolute terms according to the mentioned degeneracy between abundance and pressure in the absorption spectrum ( see lecavelier des etangs et al .
2008a and eq . [ z_lambda ] ) .
however , here the degeneracy can be solved by assuming a reference pressure at the reference level ( see section [ ref_pressure ] ) .
indeed we used a reference pressure of 3 mbar for the altitude corresponding to our reference level at 1.49145% absorption depth . using in addition the sodium abundance below the reference level as evaluated by sing et al .
( 2008b ) in the lower layers of the atmosphere , [ na / h]@xmath53@xmath710@xmath45 ( solar abundance equals 1.5@xmath710@xmath45 , asplund et al .
2006 ) , we can deduce the sodium abundance evolution in the upper layers . in agreement with the condensation scenario ,
the sodium abundance drops just above the reference level and then , following the upward evaluation of sing et al .
( 2008b ) , it should be reduced on the average over all the upper layers by about a factor 10 .
starting from the value at the reference level , the na abundance is evaluated at higher levels using a step by step procedure with _
local ihua _ models . a constant [ na / h ] abundance in the isothermal layers obtained in sects . [ t at lower altitudes ] and
[ t at higher altitudes ] , has to be present , in agreement with our _ ihua _ hypothesis .
a plateau in the measured absorption depth is observed at an altitude of about 800 km . this plateau is a kind of anomaly : with the increase of absorption cross section in the core of the line ,
the ad should rise toward smaller bandwidths .
if the plateau was extending to the smallest bandwidths at the center of the line , this would reveal the absence of the absorber at altitudes above this plateau .
however , here the plateau has a limited extension from @xmath61 to @xmath62 bandwidth .
the absorption depth rises clearly again at bandwidths smaller than about 15 ( see fig . [
fig : evaltnashhigher ] ) .
this reveals a drop in the sodium abundance at the altitude of the plateau , but the sodium is still present above , thereby producing the observed absorption rise toward the line center .
along the plateau , the variation on the line cross section with wavelength , from @xmath63 to @xmath64 , is compensated for by the variation in the sodium abundance , @xmath65}_1 $ ] and @xmath65}_2 $ ] above and below the plateau altitude . using eq .
[ z_lambda ] , the altitude of the plateau @xmath66 being constant from @xmath48 to @xmath49 , we find that @xmath67}_1 } { { \rm [ na / h]}_2}\right ) = \left(\frac{\sigma(\lambda_2 ) } { \sigma(\lambda_1)}\right).\ ] ] because here the absorption is dominated by the damping wings of the sodium line , we have @xmath68 with @xmath692 for the plateau at 800 km of altitude , we can determine that this plateau may be explained by a drop in sodium abundance by a factor [ na / h]@xmath8/[na / h]@xmath70 @xmath5 4 . according to our approach , which is essentially local
, we can conclude about this relative local drop , while as stated earlier , absolute values of the sodium abundances could not be assessed . to represent these absolute values
, we used the result of the global approach of sing et al .
( 2008a , b ) as shown in fig .
[ fig : t - max - nash - double ] .
bar ( dot - dashed line ) , although nlte effects could perturb that estimation ( see text ) .
the drop in abundance by about a factor of 10 showed at the condensation level ( at 3 mbar , tripled
dot dashed line ) is taken from sing et al .
( 2008a , b).,scaledwidth=48.0% ] considering the sodium abundance drop evaluated by sing et al .
( 2008a , b ) at the condensation level to be on the order of a factor of 10 , we assumed that [ na / h]@xmath8 should be on the order of 0.2 solar , while the na abundance above the altitude of the plateau should be [ na / h]@xmath70@xmath71 solar .
figure [ fig : t - max - nash - double ] shows the na abundance variations obtained from the observed ad over the whole domain sampled in fig . [ fig : t - p ] . to match the observed ad plateau
, one has to assume a local reduction in the na abundance by a factor of about 4 , which could indicate of a possible physical effect in the atmosphere , such as the ionization of nai into naii as predicted by several models , _ e.g. _ fortney et al .
the conclusion on the drop in the na abundance to interpret the ad plateau is expressed in terms of relative abundance , so this conclusion does not depend upon any assumption about pressure or absolute abundances .
we are evaluating variations in the na abundance in low volume density regions where nonlocal thermodynamic effects ( nlte ) should take place , _
i.e. _ below 1 mbar as evaluated by barman et al . ( 2002 ) .
there radiative processes dominate collisional ones and the nai population levels could be significantly away from the assumed local thermodynamic equilibrium ( lte ) ones .
as long as such effects do not vary significantly over one scale height , our temperature evaluations should not be affected by these nlte effects . on the contrary , such effects certainly should perturb abundance evaluations , and in particular , the previously mentioned factor 4 drop of abundance at @xmath72 bar could as well be the signature of such nlte effects at these altitudes , leading probably to an underestimation of the evaluated drop ( barman et al .
if the corresponding signature is indeed due to na ionization , it is in reasonably good agreement with the barman et al .
( 2002 ) evaluation who found that na would be mostly ionized in the limb for p @xmath73 10@xmath45 bar and only partially ionized ( @xmath5 3% ) below this point .
using the measured nai absorption depth as a function of the bandwidths , we estimated temperature variations as a function of altitude . our temperature profile is inconsistent with one of the two scenarios of sing et al .
( 2008b ) ( scenario with constant temperature and ionization at low altitude ) .
we found layers at temperatures clearly below the nai condensation temperature showing that sodium condensation has to take place at these levels , followed by a temperature rise at higher altitudes up to about 2500 k. this rise is likely the signature of the thermosphere needed to link the cool lower to the hot upper atmosphere , as suggested by the observations of atmospheric escape ( @xcite ) , observations of high temperatures ( @xcite ) , and models of the escape mechanism ( @xcite ) .
the results of our analysis favor the sodium condensation scenario to explain the deficiency of sodium observed in the line core , and the possible presence of an na ionization layer just below the base of the thermosphere .
our temperature profile presents the following patterns : a further temperature rise is observed at higher altitudes , up to 2500 k at the highest altitudes about 6500 km above the reference level , corresponding to a pressure of about @xmath76 bar .
this high temperature is likely related to the bottom part of thermosphere whose presence is predicted by theoretical models of the atmosphere . an nai abundance drop by a factor of 4 .
this is possibly due to ionization at the @xmath72 bar level , the strength of the observed drop depending upon the importance of the nlte effects at these altitudes .
d.e . acknowledges financial support from the centre national detudes spatiales ( cnes ) .
d.k.s . was supported by cnes at the start of this study .
this work is based on observations with the nasa / esa hubble space telescope , obtained at the space telescope science institute ( stsci ) operated by aura , inc . , d. k. , dsert , j.m .
, fortney , j. j. , lecavelier des tangs , a. , ballester , g. e. , cepa , j. , ehrenreich , d. , lopez morales , m. , pont , f. , shabram , m. , & vidal madjar , a. 2010 , , accepted ( arxiv:1008.4795 ) | a complete reassessment of the _ hubble space telescope _ ( hst ) observations of the transits of the extrasolar planet hd209458b has provided a transmission spectrum of the atmosphere over a wide range of wavelengths .
analysis of the nai absorption line profile has already shown that the sodium abundance has to drop by at least a factor of ten above a critical altitude . here
we analyze the profile in the deep core of the nai doublet line from hst and high resolution ground based spectra to further constrain the vertical structure of the hd209458b atmosphere . with a wavelength
dependent cross section that spans more than 5 orders of magnitude , we use the absorption signature of the nai doublet as an atmospheric probe .
the nai transmission features are shown to sample the atmosphere of hd209458b over an altitude range of more than 6500 km , corresponding to a pressure range of 14 scale heights spanning 1 millibar to @xmath0 bar pressures . by comparing the observations with a multi layer model in which temperature is a free parameter at the resolution of the atmospheric scale height , we constrain the temperature vertical profile and variations in the na abundance in the upper part of the atmosphere of hd209458b . we find a rise in temperature above the drop in sodium abundance at the 3 mbar level .
we also identify an isothermal atmospheric layer at @xmath1 k spanning almost 6 scale heights in altitude , from 10@xmath2 to 10@xmath3 bar . above this layer ,
the temperature rises again to @xmath4 k at @xmath510@xmath6 bar , indicating the presence of a thermosphere .
the resulting temperature pressure ( t p ) profile agrees with the na condensation scenario at the 3 mbar level , with a possible signature of sodium ionization at higher altitudes , near the 3@xmath710@xmath2 bar level .
our t
p profile is found to be in good agreement with the profiles obtained with aeronomical models including hydrodynamic escape . |
scanning tunneling microscope - style break junctions provide us with a way to sample a wide range of conductance values quickly .
however , when using a lock - in measurement of the conductance , the repeated change of the conductance over orders of magnitude also causes rapid changes in the phase of the conductive ( as opposed to displacement ) contribution to the current , and therefore the noise signal . as a practical matter it is not possible to adjust the phase of the lock - in measurement `` on the fly '' to separate out the conductive and displacement contributions to the current and the noise .
instead , in our measurement the lock - in was set to measure the rms magnitude @xmath67 of the output of the power detector .
unfortunately , this introduces an effective background . to see this ,
recall that @xmath68 , where @xmath38 , @xmath69 are the two orthogonal components of the measured signal . even without any input signal from the power detector @xmath38 and @xmath69
fluctuate about zero due to amplifier noise . as a result ,
@xmath67 averages to a non - zero positive value even in the absence of a real noise signal , and this value is not negligible in our measurement . during a real measurement ,
@xmath67 can be expressed as @xmath70 , where @xmath71 and @xmath72 represent the random errors ( amplifier noise , fluctuating about zero ) on the two components . to remove this background on average
, we have to perform one extra measurement at essentially zero bias , to characterize the background
. practically we still use a very low bias voltage , usually around 3 millivolts , which is small enough that no true excess noise is detectable at room temperature , but we are still able to measure conductance . in principle
the output of the noise measurement lock - in during this measurement gives @xmath73 .
after average @xmath74 , while @xmath75 , because the @xmath76 and @xmath77 terms average to zero
. we can then consider @xmath78 as the averaged mean square signal from the power detector , which is the desired quantity .
37ifxundefined [ 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 10.1103/physrev.32.97 [ * * , ( ) ] link:\doibase 10.1103/physrev.32.110 [ * * , ( ) ] link:\doibase 10.1016/s0370 - 1573(99)00123 - 4 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.75.3340 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.76.2778 [ * * , ( ) ] link:\doibase 10.1103/physrevb.49.14066 [ * * , ( ) ] link:\doibase 10.1103/physrevb.59.2871 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.82.1526 [ * * , ( ) ] link:\doibase 10.1021/nl060116e [ * * , ( ) ] .
link:\doibase 10.1103/physrevlett.79.2526 [ * * , ( ) ] @noop * * , ( ) * * , ( ) .
link:\doibase 10.1103/physrevb.46.1889 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.69.1403 [ * * , ( ) ] link:\doibase 10.1016/0375 - 9601(92)90814 - 3 [ * * , ( ) ] link:\doibase 10.1103/physrevb.51.16867 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.76.3806 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.78.3370 [ * * , ( ) ] link:\doibase 10.1103/physrevb.69.245302 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.94.206804 [ * * , ( ) ] link:\doibase 10.1103/physrevb.74.205438 [ * * , ( ) ] link:\doibase 10.1038/nphys1234 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.245427 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.146602 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.60.537 [ * * , ( ) ] link:\doibase 10.1103/physrevb.78.235421 [ * * , ( ) ] link:\doibase 10.1126/science.1087481 [ * * , ( ) ] .
link:\doibase 10.1016/s0370 - 1573(02)00633 - 6 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.95.256806 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.86.284 [ * * , ( ) ] link:\doibase 10.1103/physrevb.52.4740 [ * * , ( ) ] link:\doibase 10.1103/physrevb.61.2273 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1021/nn300440f [ * * , ( ) ] . | current noise in nanoscale systems provides additional information beyond the electronic conductance .
we report measurements at room temperature of the nonequilibrium `` excess '' noise in ensembles of atomic - scale gold junctions repeatedly formed and broken between a tip and a film , as a function of bias conditions . we observe suppression of the noise near conductances associated with conductance quantization in such junctions , as expected from the finite temperature theory of shot noise in the limit of few quantum channels . in higher conductance junctions , the fano factor of the noise approaches 1/3 the value seen in the low conductance tunneling limit , consistent with theoretical expectations for the approach to the diffusive regime . at conductance values where the shot noise is comparatively suppressed
, there is a residual contribution to the noise that scales quadratically with the applied bias , likely due to a flicker noise / conductance fluctuation mechanism .
shot noise , first discussed by schottky in 1918@xcite , comprises fluctuations in the steady - state , nonequilibrium current that originate from the discreteness of the electron charge .
this is `` excess '' noise in addition to the johnson - nyquist@xcite current fluctuations that are present at equilibrium in the absence of an applied bias current .
shot noise tends to be suppressed in macroscopic structures at finite temperatures due to electron - phonon interactions . in many mesoscopic systems small
compared to the inelastic scattering length for the electrons , shot noise survives and is strongly related to the quantum nature of transport@xcite .
many measurements have been performed in this regime on various devices in past two decades , including quantum point contacts@xcite , diffusive metal conductors@xcite , break junctions@xcite , and quantum hall systems@xcite .
most of these experiments are conducted at cryogenic temperatures to avoid thermal smearing of the noise , though shot noise measurements are possible at room temperature in sufficiently nanoscale structures@xcite .
the classical schottky shot noise power in the current is @xmath0 , where @xmath1 is the spectral density of shot noise , expressed as the mean squared variation in the current
@xmath2 per unit frequency . here
@xmath3 is the magnitude of the electron charge , and @xmath4 is the average dc bias current .
this expression is derived assuming the arrival of charge carriers is poisson distributed , with each electron unaffected by the arrival of a previous electron .
deviations from poissonian statistics may alter the noise , and these changes are usually expressed in terms of a fano factor , @xmath5 , such that the measured noise @xmath6 .
values of @xmath7 provide clues about the possible effects of interactions and underlying transport processes .
the shot noise of mesoscopic conductors at zero temperature is expressed@xcite in terms of quantum channels : @xmath8 where @xmath9 is the quantum of conductance , @xmath10 is the bias voltage across the junction , and @xmath11 is the transmission probability of the @xmath12th quantum channel . combining with the landauer formula of @xmath13 ,
the fano factor at zero temperature is : @xmath14 the fano factor carries extra information about the transmission probabilities that a conductance measurement alone can not provide . at nonzero temperature
( though assuming that energy is not exchanged between the charge carriers and other degrees of freedom such as phonons ) , the situation is more complex , as the thermal johnson - nyquist noise and shot noise are not readily separable .
the total current noise will be : @xmath15 \label{eq : finitet}\ ] ] @xmath16 is the boltzmann constant . in the equilibrium limit @xmath17 ,
both terms in this formula will survive and contribute to johnson - nyquist thermal noise of @xmath18 .
in the zero temperature limit , the total noise power will reduce to eq .
( [ eq : zerot ] ) .
temperature manifests itself through the smearing of the fermi - dirac distribution of the electrons . from these equations
it is clear that fully transmitting channels ( @xmath19 ) do not contribute to the shot noise .
this leads to a relative suppression of the noise in nanoscale systems when the conductance is largely from such open channels , as in semiconductor point contacts exhibiting quantized conductance@xcite , and in metal point contacts@xcite . in the few channel limit ,
the conductance combined with the noise allow the determination of the number of channels and their transmission probabilities@xcite . in the many - channel limit of diffusive conductors ,
random matrix theory has provided valuable insights , and the fano factor is expected to approach an average value between @xmath20 and @xmath21@xcite depending on bias conditions and sample geometry .
these predictions have been confirmed in the low temperature limit@xcite .
inelastic processes such as the excitation of local vibrational modes are predicted to alter @xmath5 as the bias voltage exceeds the energy scale of such excitations@xcite .
such effects have been observed at low temperatures in nanotubes@xcite , bilayer graphene@xcite , and very recently in atomic - scale au junctions@xcite , though the particular changes in @xmath5 depend in detail on channel transmission .
one motivation of this work is the need to perform experimental comparisons with eq .
( [ eq : finitet ] ) in a temperature regime where inelastic processes are favored by @xmath22 exceeding the characteristic energy scale for other degrees of freedom . for example ,
at 300 k , the lowest optical phonon mode in au ( @xmath23 17 mev@xcite ) should already be populated . in this paper , we consider the noise properties of ensembles of atomic - scale metal point contacts from tunneling to the multichannel ( @xmath24 ) regime , at room temperature , when inelastic processes involving phonons should be considerably more important than in the cryogenic limit .
of particular interest are the accuracy and utility of eq .
( [ eq : finitet ] ) under these conditions , over a broad range of applied bias , and the relative contributions of other noise mechanisms , such as conductance fluctuations@xcite . with biases ranging from @xmath25 , we find noise consistent with eq .
( [ eq : finitet ] ) , with clear relative suppression of the noise at conductance values corresponding to the quantized conductance peaks in the ensemble histograms . at still lower bias
, we can not resolve the excess noise , while conductance peaks remain still clear .
the fano factor in the high conductance , high bias regime is approximately a third of that in the @xmath26 tunneling regime . at the conductances where noise is relatively suppressed ,
the bias scaling of the averaged noise is consistent with conductance fluctuations , and the magnitude is not unreasonable considering previous experiments@xcite .
impedance except the sample .
, width=326 ] these experiments are performed using a scanning tunneling microscope ( stm)-style break junction , as has become very popular in the study of molecular conduction@xcite .
a junction is repeatedly made and broken in ambient conditions between a 50 nm - thick gold film evaporated on an oxidized silicon substrate , and a cut gold wire .
a computer - controlled piezo actuator is used to form and break the junction typically several times per second .
the noise measurement approach is similar that employed previously in a flexural mechanical break junction@xcite .
the desire to examine the ensemble - averaged noise leads to the choice of the stm break junction method ; the need for rapid measurement of the noise during the junction breaking process necessitates the use of a high bandwidth radio frequency ( rf ) technique . throughout the junction formation and breaking cycle , a `` dc '' bias square wave (
between 0 v and a desired voltage level , with a frequency of approximately 10 khz ) is applied across the series combination of the gold junction and a current - limiting 2 k@xmath27 resistance standard .
this low frequency square wave serves as the ( essentially ) dc bias that drives current through the junction .
the circuit , as shown in fig .
1 , employs bias - tees to separate the dc and rf signals coming from the junction .
a current preamplifier measures the current and is recorded electronically , giving a measure of the junction s conductance . at the same time , a lock - in amplifier synchronized to the square wave detects the difference between the rf power with and without bias applied to the junction ; this is the excess noise power .
the bandwidth of the noise measurement is roughly 250 - 580 mhz .
a detailed gain - bandwidth product measurement is employed . since both shot noise and johnson - nyquist noise
are expected to be white over this bandwidth , deviations from white noise arise from the impedance properties of the measurement circuit as a whole .
the details of the noise analysis and background subtraction are described in supplemental material . at every single bias ,
the stm style motion of the gold tip repeats hundreds times to generate a histogram of conductance , as well as a plot of the ensemble averaged excess noise power vs. conductance .
an example is provided in fig.2 . . ,
width=336 ] as has been seen in many previous experiments in atomic - scale metal contacts@xcite , peaks are observed in the conductance histograms , signifying preferred junction configurations with specific values of conductance .
peaks are observed at 1 @xmath28 , and near other integer multiples of @xmath29 , consistent with past results on au junctions at room temperature@xcite .
cryogenic experiments involving shot noise@xcite and subgap structure in superconducting contacts@xcite have demonstrated that the 1 @xmath28 peak in au junctions is dominated by configurations with a single highly transmitting channel ( @xmath30 ) . in our structures
usually the first three conductance peaks are readily resolvable in the histogram .
the related ensemble - averaged noise power measurement is also shown .
as is clear from the figure , the ensemble - averaged noise power is clearly suppressed near conductance values where the conductance histogram is peaked .
the transmission of rf signals always faces the problem of power reflection , which originates from impedance mismatch .
conversely , reflection itself carries information about impedance . in our measurement circuit , all the commercial rf electronic components are of 50@xmath27 impedance
we therefore expect significant impedance mismatch and reflections only between the stm - style gold junction and the transmission lines . as an added complication compared to a fixed device configuration , the tip s repeatedly vertical motions introduce the extra complexity of a strongly time - varying dc conductance into the junction s rf properties . in principle
a measurement should be performed to properly characterize the impedance mismatch between the junction and the rf measurement circuitry at each conductance value . ideally , knowing the rf properties of the nanoscale junction and the accompanying electronics , including the gain - bandwidth product of the amplifier chain , it should be possible to infer the actual current noise ( a@xmath31/hz across the junction ) from the measured rf power seen by the power meter . however , in the stm breakjunction setup , in which the dc conductance of the junction changes by orders of magnitude on millisecond timescales , with our equipment it is not possible to measure all of the relevant rf parameters in real time . as an approximation to this ,
we instead measure the reflection properties as a function of conductance _ averaged _ over the ensemble of junction configurations .
this should at least indicate whether there are gross variations in the efficiency of the junction s rf coupling to the rest of the circuit .
following on the approach reported previously in measurements of the impedance properties of a vacuum photodiode@xcite , we perform a reflectance measurement as shown in fig .
3 .
a commercial white noise source is used to provide wide - band white noise across the rf bandwidth of interest .
the amplitude of this noise is modulated by a subsequent rf switch that is turned on and off at the same ( acoustic ) frequency used for the square wave voltage bias we applied in excess noise measurement .
part of the white noise will be reflected at the boundary between the gold junction and the transmission line to the bias tee .
the reflected power goes through a directional coupler as well as the same amplifier chain used in the excess noise measurements , and is registered by the logarithmic power detector .
simultaneously , the junction s cyclical stm - style motion is executed , with an applied dc bias across the junction to allow the simultaneous acquisition of a conductance histogram .
this measurement gives a picture of the relation between ensemble averaged reflections and the conductance of the junction .
while the reflection measurement can not provide all the information about the junction s rf properties , at least it provides a rough check that nothing dramatic happens in terms of the ensemble - averaged impedance mismatch over the conductance range .
the reflection averaged over the ensemble of junction configurations is around 22% .
there is some systematic variation with conductance , but this is small , less than 2% over 10 @xmath29 .
since the impedance mismatch between the junction and the measurement circuit does not vary dramatically on average over the range of junction configurations , there should be a scale factor ( approximately constant across the conductance range ) between the measured power and the true current noise across the junction .
we attempt to find this factor by using the knowledge that the fano factor in the tunneling regime @xmath32 approaches one in the limit that a single poorly transmitting channel dominates the conductance , as apparent from eq .
( 1 ) .
we acquire conductance histograms and ensemble - averaged noise using a series of square wave bias voltages from a few millivolts to several hundreds of millivolts .
an example of such ensemble - averaged noise data for eight different bias levels is shown in fig.4 .
all the data shown in this figure were taken continuously in one day to ensure an identical experimental environment ( lab temperature , any stray rf background ) . in our system
no clear excess noise can be detected at bias voltages below about 25 mv , corresponding to @xmath22 at @xmath33 k. noise suppression at conductances corresponding to the first three peaks in the conductance histograms are very clear , and the magnitude of the detected noise power is monotonously increasing with bias as expected . note that the suppressions are not complete .
this indicates that at least some of the junction configurations corresponding to peak conductance values result from a mixture of multiple , partially transmitting quantum channels .
we are interested in the scaling of the noise with bias , as this reveals the fano factor .
working with the data sets in fig .
4 , we can specify a particular conductance value , and for each applied square wave bias voltage we can compute the actual voltage drop , @xmath10 , across the junction . starting from eq .
( [ eq : finitet ] ) , we can see that plotting the detected power spectral density as a function of @xmath34 $ ] should produce a linear graph , with the slope giving the ( zero temperature ) fano factor , @xmath5 . here
@xmath35 is the conductance of the junction and @xmath36 is assumed to be the ambient temperature , @xmath37 k. in fig .
5a , we plot data for three closely spaced values of conductance , all below @xmath28 .
the figure shows that the noise is in fact linear when plotted as a function of @xmath38 , and the fano factor is decreasing as the conductance approaches 1 @xmath28 .
these are consistent with the measured signal originating from shot noise , and the suppressions shown in fig .
4a originating from the saturated channel mechanism of eqs .
( [ eq : zerot ] , [ eq : finitet ] ) .
qualitatively similar declines in fano factor are also observed as the conductance approaches the values associated with the other two peaks in conductance histograms / suppressions in measured noise power .
we now consider the situation at higher conductances , when the number of channels involved in transport is large compared to one .
as discussed above , we assume that the fano factor in the low conductance regime is close to one , as expected if transport is dominated by a single poorly transmitting channel .
we perform a linear fit to the noise vs. @xmath38 data at @xmath39 , and assume that corresponds to a true fano factor close to one ( though a true single channel device would have @xmath40 near that conductance ) . in the particular data set shown in fig . 5a
, the lowest conductance data around @xmath41 are still well described by a similar fano factor , as shown by the blue line .
this suggests that , when looked over the full ensemble average , the conductance in this regime results from several poorly transmitting channels , rather than being dominated by a single channel with a transmittance of 0.45 .
this indicates that a large contribution to the conductance histogram results from junctions with comparatively blunt tips .
figure 5b shows , on the same plot as that low conductance data , the noise data vs. @xmath38 at around 4,5 and 6@xmath29 .
identically scaled linear fits show slopes of 0.39 , 0.34 , and 0.31 , respectively .
these relative slopes are comparatively insensitive to the choice of the low conductance fits used to find the scaling factor , provided @xmath42 . at even higher conductance values not shown in the figure , the inferred fano factors also do not vary by much , decreasing to around 0.25 near 10 @xmath28 .
these observations are roughly consistent with expectations for a crossover toward diffusive conduction . in diffusive conductors that are small
compared to the inelastic scattering length for both electron - electron ( @xmath43 ) and electron - phonon ( @xmath44 ) interactions , theoretical calculations@xcite and experiments on different samples at helium temperature@xcite agree that @xmath45 in this regime . in the limit that the constriction length @xmath46 exceeds @xmath43 , the expected reduction factor is predicted@xcite to be @xmath47 . since both longitudinal and transverse dimensions in our junctions are comparable to the atomic scale , it seems likely that @xmath48 , and our data show consistency with the prediction of 1/3 .
this consistency with a diffusive picture is a bit surprising , given that the length and transverse dimensions of the point contact should still be around 1 nm even when @xmath49 .
experiments involving pb junctions@xcite have shown a similar transition to the diffusive regime even at relatively low conductances , though in that work this is ascribed to the importance of multiple orbitals per atom contributing to the conductance , leading to enhanced channel mixing . while this multiorbital mechanism should not apply for au junctions , the lack of clear peaks in the conductance histograms ( and corresponding suppressions in the noise ) above 3 @xmath28 are consistent with enhanced channel mixing in our structures relative to that seen in other au experiments@xcite .
this is further supported by the observation mentioned in the previous paragraph that conduction in the tunneling regime in this data set appears to involve multiple poorly transmitting channels .
it has been pointed out before@xcite that the prediction of a suppression factor of about 1/3 is surprisingly robust .
investigations with different materials and over a broader range of temperature and bias should shed light on the extent of universality in this suppression . .
data points at around 0.45 @xmath29 , 0.78 @xmath29 and 0.85 @xmath29 are shown in the plot as well as linear fits to @xmath38 for those points .
bottom panel : data around 4 @xmath29 , 5 @xmath29 and 6 @xmath29 , with inferred fano factors ( relative to that at @xmath50 ) of 0.39 , 0.34 and 0.31 separately.,title="fig:",scaledwidth=50.0% ] .
data points at around 0.45 @xmath29 , 0.78 @xmath29 and 0.85 @xmath29 are shown in the plot as well as linear fits to @xmath38 for those points .
bottom panel : data around 4 @xmath29 , 5 @xmath29 and 6 @xmath29 , with inferred fano factors ( relative to that at @xmath50 ) of 0.39 , 0.34 and 0.31 separately.,title="fig:",scaledwidth=50.0% ] .
however , for the latter a deviation from a linear dependence is observed when the conductance is around 1 @xmath29 and the shot noise contribution is relatively suppressed .
the lowest ( gray ) points and linear fit correspond to the 1 @xmath28 data with a best - fit quadratic - in - voltage contribution removed .
panel ( b ) is a purely quadratic fit of the excess noise vs. bias voltage at the conductances corresponding to the first two suppressions .
the noise power here has been converted to mean squared voltage per frequency , for comparison with flicker noise .
both of these fitting procedures give flicker noise contributions of the same order of magnitude.,title="fig:",scaledwidth=50.0% ] .
however , for the latter a deviation from a linear dependence is observed when the conductance is around 1 @xmath29 and the shot noise contribution is relatively suppressed .
the lowest ( gray ) points and linear fit correspond to the 1 @xmath28 data with a best - fit quadratic - in - voltage contribution removed .
panel ( b ) is a purely quadratic fit of the excess noise vs. bias voltage at the conductances corresponding to the first two suppressions .
the noise power here has been converted to mean squared voltage per frequency , for comparison with flicker noise .
both of these fitting procedures give flicker noise contributions of the same order of magnitude.,title="fig:",scaledwidth=50.0% ] we also consider the scaling of the excess noise at conductance values where the shot noise contribution is expected to be maximally suppressed . we have taken data sets of the type shown in fig . 4 on multiple occasions with different tips , film samples , and cleaning / annealing procedures .
to best study the bias scaling of the remaining excess noise near 1 @xmath28 , we examine a data set when the suppression at 1 @xmath28 was particularly well defined .
the result is shown in fig .
6a . with the comparatively strong suppression of shot noise power in this data
set , we find that the measured noise vs. @xmath38 is nonlinear .
this nonlinearity is only clearly seen at conductances where the noise is relatively suppressed . a natural explanation for this is a contribution to the measured noise from conductance fluctuations , commonly termed `` flicker '' noise , which often has ( in macroscale systems ) a @xmath51 frequency dependence@xcite .
such flicker noise , originating with fluctua tions in the actual junction resistance , should scale quadratically with voltage across the junction at a given conductance , when recast as voltage fluctuations .
figure 6b plots the excess noise spectral density vs bias voltage at the first two suppressions , with the noise power converted to mean squared voltage per unit frequency .
quadratic fits of the form @xmath52 describe the data well , with a comparatively small linear term , @xmath53 , as well as an even more tiny residual intercepts , @xmath54 . with this fitting procedure
we find @xmath55 hz@xmath56 .
we note that such a fitting procedure overestimates the size of @xmath57 , since the hyperbolic cotangent term in eq .
( [ eq : finitet ] ) does contribute some nonlinearity in such a plot even when the only noise is finite temperature shot noise . instead if we fit a quadratic in voltage term in addition to the finite temperature shot noise expectation of eq .
[ eq : finitet ] , we find @xmath58 hz@xmath56 , smaller but of the same order of magnitude .
using the phenomenological hooge s law@xcite , @xmath59 with power @xmath60 ranging between @xmath61 , we compared our inferred magnitude of @xmath57 with values observed at lower frequencies and higher junction conductances in other gold point contacts by wu _ _ et al.__@xcite .
the result depends strongly on the assumed value of @xmath60 , which is expected to fall between 1 ( traditional @xmath51 noise ) and 2 ( expected for a single two - level fluctuator@xcite ) .
we find a flicker noise amplitude smaller than the wu _ et al . _
values extrapolated to 1 @xmath28 with the fitted @xmath60 value in their case ; however , if @xmath62 over frequency from their sub-100 khz measurements to our rf scale , our value for @xmath57 is compatible with their results . we can reasonably conclude that much of the nonlinearity in the noise vs. bias at conductances when shot noise is suppressed results from flicker noise / conductance fluctuations .
the relatively small magnitude of this noise suggests a comparatively rapid decay of flicker noise with frequency in the rf range in atomic - scale junctions
. it would be interesting to consider the temperature variation of the flicker noise magnitude over this frequency range , which would help clarify whether thermal activation of defect motion is relevant here . as a function of conductance .
, title="fig:",scaledwidth=50.0% ] as a function of conductance . , title="fig:",scaledwidth=50.0% ] it is worth considering what kind of mixing of quantum channels is required to produce the evolution of fano factor with conductance that we infer from our data , within the noninteracting picture of eq .
[ eq : zerot ] .
a simple `` toy model '' simulation of possible fano factor evolution is provided , which depends on the choice of how the channels @xmath63 open as conductance is increased .
the result is shown in fig.7 ( a ) .
we assume for simplicity a set of uniformly separated hyperbolic tangent functions for the form of @xmath63 as a function of @xmath35 , which is shown on the bottom .
each hyperbolic tangent is specified by a central conductance at which @xmath64 , and a width in @xmath35 over which the channel `` turns on '' .
the resulting fano factors have been plotted on the top .
multiple colors / linestyles represent different choices of the central conductances and turn - on widths . in top graph
, the uppermost line ( blue ) shows no clear suppression of noise vs. @xmath65 due to the heavy overlaps between channels .
in contrast , the lower most curve ( black ) is in the limit of well - separated channels that do not mix .
channels mixture tends to smear out suppressions and raise the fano factor , which is the qualitative reason that in a larger - sized diffusive conductor only a roughly constant fano factor results .
figure 7b shows an example ( nonunique ) , of a set of transmission channels that agree reasonably well with the measurements .
additional constraints ( such as those provided by subgap conductance in superconducting junctions@xcite ) are necessary to better constrain the specific channels and their evolution .
we note that we do not see clear signatures of bias - dependent changes in the fano factor in these ensemble averaged measurements . as the low temperature data in kumar _
et al . _
shows@xcite , changes in the apparent fano factor due to excitation of optical phonons can be of either sign in single - channel devices , with a crossover between enhancement and suppression of @xmath5 taking place at a particular transmittance , around 0.95 . with respect to the au optical phonon at @xmath23 17 mev ,
all of the data in this paper are in the high bias regime above the voltage threshold .
our measurements show that in au junctions at room temperature , there are no clear inelastic thresholds above this energy that survive ensemble averaging over junction configurations .
other analysis approaches that look at subensembles may be more revealing@xcite and are in progress .
here we have reported a calibrated measurement of shot noise and its bias dependence in stm - style gold junctions at room temperature .
we observe in general that the shot noise spectral density is proportional to @xmath66 $ ] , as expected for a nanoscale junction with a small number of channels .
the slope of such a plot is a means of extracting the zero - temperature fano factor of the noise .
the fano factor observed at higher conductances ( several conductance quanta ) is roughly 1/3 that seen in the tunneling regime , consistent with expectations as the diffusive limit is approached . at suppressions where shot noise is comparatively small , nonlinearities in the noise vs. @xmath38
are observed .
it is likely that the origin of this nonlinearity is conductance fluctuation / flicker noise ; this is supported by fits showing an approximate quadratic dependence of the noise on the bias across the junction , and the magnitude of this noise is roughly consistent with that extrapolated from other experiments@xcite .
the evolution of the fano factor with conductance can be described adequately within the noninteracting picture through reasonable choices of the evolution of channel numbers and transmittances with conductance .
these experiments highlight that the quantum nature of the electronic conduction remains detectable and important as electronic systems approach the atomic scale , even at room temperature .
the authors acknowledge the support of nsf award dmr-0855607 , and useful conversations with j. van ruitenbeek , c. schnenberger , and l. richardson . |
within the context of quantum information theory , quantum objects are considered as carriers of information @xcite .
a two - state system , such as single 2-level atom , called single qubit is one basic example of quantum information unit .
qubits represent a fundamental aspect of quantum computer @xcite .
dynamical properties of single qubits have been investigated in different directions . very recently , single control qubit shor algorithm for the case of static imperfections induced by residual couplings between qubits was formulated @xcite .
cryptographic applications of single - qubit rotations in quantum public - key cryptography has been discussed in references @xcite-@xcite .
the use of single photon qubit as a quantum encoder for single - photon qubits was reported in @xcite .
the authors in @xcite have used the charged qubit pair to perform quantum teleportation .
the speed of communication using single qubit has been treated in classical and quantum framework @xcite .
dynamical and spectral properties of a short pulsed -driven qubit in the absence of dissipation processes has been studied by many researchers(@xcite-@xcite and refs therein ) . in @xcite ,
the authors have investigated the fluorescence spectrum of a rectangular pulsed - driven single qubit within and outside the rotating wave approximation ( rwa , where fast oscillatory terms are dropped ) for different initial atomic states .
specifically , for initial atomic coherent state , the transient fluorescence spectrum exhibits asymmetric rabi splitting which turns to `` ringing '' for large pulse area .
the ringing behavior is attributed to the initial finite coherent dispersion and interference process between amplitude spectra of the atomic inversion and polarization variables .
properties of single qubit rotation operations using simple rf pulses have been investigated in @xcite . also ,
measure of the errors in single qubit rotations by pulsed electron paramagnetic resonance has been reported in @xcite . in the present work ,
we investigate the dynamics of the coded information in a single qubit subject to a rectangular pulse .
we quantify the exchange information between the qubit and the environment(i.e the pulse ) .
also , we quantify the speed of quantum communation by evaluating the speed of orthogonality of the density operator .
the paper is organized as follows . in sec.2 , we present the model and its exact operator solution within the rwa and its iterative solution outside rwa @xcite .
the fidelity and the exchange information is the subject of sec.3 . the speed of orthogonality and hence the speed of transfer information is discussed in sec.4 .
finally , a summary is given in sec.5 .
we consider a single 2-level atom ( qubit ) of transition frequency @xmath0 interacts with a rectangular pulse of circular frequency @xmath1 in the absence of any atomic dissipation .
the full hamiltonian of this system outside the rwa ( in units of @xmath2 ) is given by @xcite , @xmath3 where the spin-@xmath4 operators @xmath5 obey the @xmath6 algebra , @xmath7=2\hat{s_{z}},\quad [ \hat{s_{z}},\hat{s_{\pm}}]=\pm\hat{s_{\pm}}.\ ] ] for a rectangular laser pulse of duration @xmath8 , the rabi frequency @xmath9 with @xmath10 through the interval of time @xmath11 $ ] and @xmath12 otherwise , with @xmath13 taken real .
the pulse duration @xmath8 is much shorter than the lifetime of the atomic upper state , hence atomic damping can be discarded . introducing the rotating frame operators @xmath14 where the @xmath15 operators obey the same algebraic form of eq.([com ] ) ,
heisenberg equations for the atomic operators @xmath16 according to ( [ ham ] ) are of the form , @xmath17^\dag , \nonumber\\ \dot{\hat\sigma}_{z}&=&-i\frac{\omega(t)}{2 } \bigr[\hat{\sigma}_{+}(1+e^{2i\omega_l t})-\hat{\sigma}_{-}(1+e^{-2i\omega_l t})\bigl],\end{aligned}\ ] ] where @xmath18 , is the atomic detuning . not that the terms in @xmath19 represent the effect of the interaction of the atom with the pulse outside rwa (
note that , the issue of rwa is only associated with linearly polarized light ; cf@xcite ) . + * ( a ) exact solution within rwa .
* within rwa we discard the terms @xmath19 in ( [ motion ] ) and assume the qubit initially in the coherent state , @xmath20 where @xmath21 , @xmath22 and @xmath23 are the bare ground and excited atomic states respectively . using the notation , @xmath24 for the initial bloch vector where @xmath25 , the exact solution of eq.(4)in terms of the bloch vector @xmath26 within rwa are written in the following form @xcite , @xmath27 where @xmath28 , \nonumber\\ \alpha^{(0)}_y&=&-\left(\frac{\delta}{\omega_1}\right)\sin\omega_1 t , \nonumber\\ \alpha^{(0)}_z&=&\left(\frac{\delta\omega}{\omega_1 ^ 2}\right)(1-\cos\omega_1 t ) , \nonumber\\ \beta^{(0)}_x&=&-\alpha^{(0)}_y , \nonumber\\ \beta^{(0)}_y&= & \frac{1}{2}\bigl[\left(\frac{\omega}{\omega_1}\right)^2+\left(\frac{\delta^2+\omega_1 ^ 2}{\omega_1 ^ 2}\right)\cos\omega_1 t-\left(\frac{\omega}{\omega_1}\right)^2(1-\cos\omega_1 t)\bigr ] , \nonumber\\ \beta^{(0)}_z&=&-\left(\frac{\omega}{\omega_1}\right)\sin\omega_1 t , \nonumber\\ \gamma^{(0)}_x&=&\alpha^{(0)}_z , \quad \gamma^{(0)}_y=-\beta^{(0)}_z , \nonumber\\ \gamma^{(0)}_z&=&\left(\frac{\omega}{\omega_1}\right)^2\bigl[(\cos\omega_1 t+\left(\frac{\delta}{\omega}\right)^2\bigr],\end{aligned}\ ] ] with @xmath29 and we have used the notations @xmath30 are used for the exact rwa solution . + * ( b ) iterative solution outside rwa . * outside rwa ,
eq.([motion ] ) has an iterative solution to @xmath31 ; @xmath32 at exact resonance @xmath33 and is given by @xcite , @xmath34 where , @xmath35 , \nonumber\\ \alpha^{(1)}_y&=&-\frac{\lambda}{4}\bigl[(1-\cos2 \omega_l t)\sin\omega t-\sin 2\omega_l t\cos\omega t\bigr ] , \nonumber\\ \alpha^{(1)}_z&=&\frac{\lambda}{2}\left(\cos 2\omega_l t \cos\omega t-1\right ) , \nonumber\\ \beta^{(1)}_x&=&\frac{\lambda}{4}\bigl[(1-\cos2 \omega_l t)\sin\omega t+\sin 2\omega_l t\cos\omega t\bigr ] , \nonumber\\ \beta^{(1)}_y&=&\cos\omega t-\frac{\lambda}{4}\bigl[\sin 2\omega_l t\sin\omega t+(4-\cos2\omega_l t)\cos\omega t\bigr ] , \nonumber\\ \beta^{(1)}_z&=&-\sin\omega t-\frac{\lambda}{2}\sin2\omega_l t\cos\omega t , \nonumber\\ \gamma^{(1)}_x&=&\lambda(2-\cos2 \omega_l t-\cos\omega t ) , \nonumber\\ \gamma^{(1)}_y&=&\sin\omega t+\lambda\sin2\omega_l t , \nonumber\\ \gamma^{(1)}_z&=&\cos\omega t.\end{aligned}\ ] ] and have used the notations @xmath36 for the iterative solution outside rwa .
note , for @xmath37 , the solutios @xmath38 with @xmath39 reduce to those in @xmath40 ) with @xmath41 at @xmath42 .
let us assume that we have coded information in the initial state of the qubit .
the initial density operator takes the form , @xmath43 due to the interaction with the pulse the coded information may be exchanged between the pulse and the initial state in ( [ init ] ) .
our aim to is find how information evolves with time .
we measure the transfer information using the fidelity , @xmath44 @xcite , while the entropy exchange , @xmath45 is used as a measure of the information exchanged between the system and the environment @xcite . for this puropose
we plot some figures for different initial state settings within and outside rwa . + * ( a)within rwa*. in fig.(1 ) , we investigate the effect of the initial phase angle @xmath46 , on the fidelity of the transfer and exchange information .
fig.(1a ) shows the periodic dynamical behavior of the fidelity in the resonant case i.e @xmath42 .
the initial state of the coded information is chosen such that @xmath47 .
it is clear that , for @xmath48 the fidelity , @xmath49 decreases with time smoothly and completely vanishes for the first time round at @xmath50 .
as one decreases @xmath46 the fidelity decreases but does not completely vanish and the minimum values of the fidelity @xmath49 , decrease as @xmath46 decreases . and the exchange information @xmath51 , @xmath52 respectively against the scaled time @xmath53 for the system ( within rwa ) , @xmath37 with @xmath42 and @xmath54 .
the solid , dash and dot curves are for @xmath55 and @xmath56 respectively.,title="fig : " ] and the exchange information @xmath51 , @xmath52 respectively against the scaled time @xmath53 for the system ( within rwa ) , @xmath37 with @xmath42 and @xmath54 .
the solid , dash and dot curves are for @xmath55 and @xmath56 respectively.,title="fig : " ] ( -375,73)@xmath49 ( -190,75)@xmath51 ( -350,118)(a ) ( -160,118)(b ) ( -80,-10)@xmath57 ( -280,-10)@xmath57 in fig.(1b ) , we investigate the amount of information exchanged between the state and the environment for the same parameters as in fig.(1a ) .
as the fidelity decreases the exchange information increases but does not reach the maximum value i.e there is still coherent information in the sate which carries these coded information .
also , with the fidelity @xmath49 reaches to its minimum value there is no exchange information .
this means that , for some interval of time one can use the driven qubit safely in quantum communication . and different @xmath58 for the solid , dashed and dot curves respectively.,title="fig : " ] and different @xmath58 for the solid , dashed and dot curves respectively.,title="fig : " ] ( -375,73)@xmath49 ( -190,75)@xmath51 ( -350,118)(a ) ( -160,118)(b ) ( -80,-10)@xmath57 ( -280,-10)@xmath57 in fig.(2 )
, we consider different initial states setting for @xmath59 , where the phase angle is fixed at @xmath60 . in this case
the initial state of the driven qubit is defined by @xmath61 .
in general the behavior of the fidelity is the same as that depicted in fig.(1a ) , but the the fidelity does not completely vanish as in fig.(1a ) for some classes of initial states . as @xmath59 decreases , the minimum value of @xmath49 decreases .
also , the exchange information , fig.(2b ) , increases for large fidelity and vice - versa . at the minimum values of the fidelity the exchange almost vanishes .
, @xmath60 . the solid , dashed and dot curves for @xmath62 and @xmath63 respectively.,title="fig : " ] , @xmath60 . the solid , dashed and dot curves for @xmath62 and @xmath63 respectively.,title="fig : " ] ( -375,73)@xmath49 ( -190,75)@xmath51 ( -350,118)(a ) ( -160,118)(b ) ( -80,-10)@xmath57 ( -280,-10)@xmath57 and ( b ) the information exchange @xmath51 against @xmath57 for the system treated outside the rwa ( @xmath64 ) .
the initial state of the qubit is described by @xmath65 and @xmath66 .
the solid , dashed and dot curves for @xmath67 and @xmath68 respectively.,title="fig : " ] and ( b ) the information exchange @xmath51 against @xmath57 for the system treated outside the rwa ( @xmath64 ) .
the initial state of the qubit is described by @xmath65 and @xmath66 .
the solid , dashed and dot curves for @xmath67 and @xmath68 respectively.,title="fig : " ] ( -375,73)@xmath49 ( -190,75)@xmath51 ( -350,118)(a ) ( -160,118)(b ) ( -80,-10)@xmath57 ( -280,-10)@xmath57 the effect of the detuning parameter , @xmath69 is seen in fig.(3 ) , where we consider the choice for @xmath70 and @xmath60 which maximize the fidelity and minimize the change information . in this case the initial state in which we coded the information
is given by @xmath71 .
the general behavior of the fidelity is the same as shown in fig.(1 ) and fig.(2 ) . for a very small value of the detuning @xmath72 ,
the minimum value of @xmath49 is lesser than the resonance case : see the solid curve in fig.(2a ) and fig.(3a ) .
as one increases the value of @xmath69 , the curves shift to the left .
this means that the fidelity decreases faster at earlier time .
fig.(3b ) describes the dynamics of the exchange information , where for small values of the detuning the exchange operation starts with delay .
the maximum amount of information for different values of the detuning shows slight change . + * ( b ) outside the rwa . * the behavior of the fidelity @xmath49 and the information exchange @xmath73 for the system given by ( [ without ] ) where the system is treated outside rwa is shown in fig.(4 ) . in this figure
we set @xmath74 ( the case described by the solid curve in fig.(1a ) , within rwa ) .
the dynamics of the fidelity is displayed in fig.(4a ) for different values of the parameter @xmath75 .
it is clear that for small value of @xmath76 , the behavior is almost the same as that within rwa . however , for larger values of @xmath75 , the minimum value of the fidelity increases .
this means that the loss of the coded information decreases .
fig.(4b ) describes the dynamics of the exchange information between the travelling qubit and the environment for the same parameter as that used in fig.(1a ) .
this figure shows that the amount of information exchange between the qubit and the environment is much larger compared with that depicted in fig.(1b ) ( solid - curve within rwa ) . also , the minimum value of @xmath51 increases for larger values of @xmath75 . against @xmath57 outside rwa @xmath77 for fixed @xmath48 and different @xmath78 ( solid , dashed and dot curves respectively ) ( b ) the same as ( a ) but for fixed @xmath54 and different @xmath79 ( solid , dashed and dot curves respectively),title="fig : " ] against @xmath57 outside rwa @xmath77 for fixed @xmath48 and different @xmath78 ( solid , dashed and dot curves respectively ) ( b ) the same as ( a ) but for fixed @xmath54 and different @xmath79 ( solid , dashed and dot curves respectively),title="fig : " ] ( -375,70)@xmath49 ( -190,70)@xmath49 ( -350,118)(a ) ( -160,118)(b ) ( -80,-10)@xmath57 ( -280,-10)@xmath57 . the effect of the excitation angle @xmath59 and the phase @xmath46 is investigated in fig.(5 ) , where we set the iteration parameter @xmath80 .
the dynamics of the fidelity @xmath49 is shown in fig.(5a ) for different values of the excitation angle @xmath59 . at @xmath81 the fidelity @xmath82 , due to the effect of the @xmath75- parameter @xmath75 . however , as one decreases the value of @xmath59 , the upper and lower values of the fidelity show slight change .
fig.(5b ) , displays the behavior of @xmath49 for different phases .
as the phase decreases the minimum value of the fidelity is much larger and consequently the exchange information between the travelling qubit and the environment decreases .
so , by decreasing the value of the phase one can overcome the negative effect of the iteration parameter @xmath75 .
therefore , from our preceding results the fidelity of the transmitted information , which is coded in an initial state described by small excitation angle @xmath59 and phase @xmath46 decreases faster but the maximum value of the exchange information with the environment is smaller .
however , the minimum value of fidelity loss can be improved for larger values of the angle @xmath59 and the phase @xmath46 .
if we consider a small value of the detuning parameter , we can send information with high fidelity , while the exchanged information is always upper bounded . outside the rwa
the fidelity of the travelling qubit could improve , but at the the same time increases the possibility of information exchange with the environment .
the effect of the excitation angle @xmath59 outside rwa has a slight effect and the fidelity does nt reach its maximum value . however , for a smaller phase angle @xmath46 , the value of @xmath83 .
in the pervious section , we showed how the coded information evolves from one locations ( node ) to another @xcite . in a quantum computer there
it is important to know the speed of sending information from one node to another .
since the information is coded in a density operator , therefore we seek how fast the density operator will change its orthogonality .
in other words , we search for a minimum time needed for a quantum system to pass from one orthogonal state to another @xcite .
let us assume that the eigenvectors of the initial state are given by @xmath84 , @xmath85 and for the final state are given by @xmath86 .
the speed of quantum computation is defined by the maximum number of orthogonal states that the system can pass through per unit time .
the orthognathy is given by the scalar product of the vectors @xcite @xmath87 + * ( a ) within rwa .
* in fig.(6 ) , we have plotted @xmath88 , the amplitude values of @xmath89 against the scaled time to display its behavior for different initial atomic state settings . from our results in sec.@xmath90 , we know that for small values of @xmath59 and @xmath46 , one can transmit the coded information with high degree of fidelity . in fig.(6a ) , in the resonant case @xmath33 and @xmath54 and a very small value of @xmath91 , it is clear that @xmath88 vanishes periodically at some specific time , @xmath92 . at these times , the initial and final states are orthogonal and the information which is coded in the initial state is completely transfered to the final state . as one increases the phase angle @xmath93 , the speed of orthogonality decreases ( fig.(6b ) ) , where the number of vanishing @xmath88 is lesser than that for smaller phase ( see fig.(6a ) ) .
fig.(6c ) , is plotted for different class of initial state settings , where we consider @xmath94 .
it is seen that the number of orthogonality decreases and consequently the speed of computations is smaller than that depicted in fig.(6a ) .
the non - resonant case is shown in fig.(6d ) , for @xmath95 where @xmath88 does not vanish periodically as that predicted for the resonant case ( see fig.(6a ) ) .
as an example , around @xmath96 and @xmath97 , @xmath98 .
this means that the information has not completely transferred to the final state and consequently there are some information gained by the environment . +
* ( b ) outside rwa . * . ( a ) for @xmath99 , @xmath42 .
( b ) the same as ( a ) , but @xmath100 .
( c ) the same as(a ) , but @xmath101 .
( d ) the same as ( a ) , but for @xmath95.,title="fig : " ] . ( a ) for @xmath99 , @xmath42 .
( b ) the same as ( a ) , but @xmath100 .
( c ) the same as(a ) , but @xmath101 .
( d ) the same as ( a ) , but for @xmath95.,title="fig : " ] ( -350,118)(a ) ( -160,118)(b ) ( -390,70)@xmath88 ( -195,70)@xmath88 ( -80,-5)@xmath57 ( -280,-5)@xmath57 + . ( a ) for @xmath99 , @xmath42 .
( b ) the same as ( a ) , but @xmath100 .
( c ) the same as(a ) , but @xmath101 .
( d ) the same as ( a ) , but for @xmath95.,title="fig : " ] . ( a ) for @xmath99 , @xmath42 .
( b ) the same as ( a ) , but @xmath100 .
( c ) the same as(a ) , but @xmath101 .
( d ) the same as ( a ) , but for @xmath95.,title="fig : " ] ( -350,118)(c ) ( -160,118)(d ) ( -80,-5)@xmath57 ( -280,-5)@xmath57 ( -390,70)@xmath88 ( -195,70)@xmath88 ) vs time , @xmath57 . ( a ) for @xmath102 and @xmath103 .
( b ) the same as fig.(a ) , but @xmath104 .
( c ) the same as(a),but @xmath101 and @xmath105 .
( d ) the same as ( a ) , but @xmath106.,title="fig : " ] ) vs time , @xmath57 . ( a ) for @xmath102 and @xmath103 .
( b ) the same as fig.(a ) , but @xmath104 .
( c ) the same as(a),but @xmath101 and @xmath105 .
( d ) the same as ( a ) , but @xmath106.,title="fig : " ] ( -350,118)(a ) ( -160,118)(b ) ( -390,70)@xmath88 ( -195,70)@xmath88 ( -80,-5)@xmath57 ( -280,-5)@xmath57 + ) vs time , @xmath57 . ( a ) for @xmath102 and @xmath103 . ( b ) the same as fig.(a ) , but @xmath104 .
( c ) the same as(a),but @xmath101 and @xmath105 .
( d ) the same as ( a ) , but @xmath106.,title="fig : " ] ) vs time , @xmath57 . (
a ) for @xmath102 and @xmath103 .
( b ) the same as fig.(a ) , but @xmath104 .
( c ) the same as(a),but @xmath101 and @xmath105 .
( d ) the same as ( a ) , but @xmath106.,title="fig : " ] ( -350,118)(c ) ( -160,118)(d ) ( -80,-5)@xmath57 ( -280,-5)@xmath57 ( -390,70)@xmath88 ( -195,70)@xmath88 the dynamics of the orthogonality speed for a system treated outside rwa is displayed in fig.(7 ) . the effect of the iteration parameter @xmath75 is shown in figs.(7a@xmath1077b ) , where we set @xmath103 and @xmath108 respectively .
it is clear that , the orthogonality speed @xmath88 increases for small value of @xmath75 .
as one decreases the excitation angle @xmath109 , with other parameters kept the same as in fig.(7a ) , the speed of orthogonality decreases ( see fig.(7c ) ) .
however for large value of the phase angle , the orthogonality speed decreases- compare figs.(7a@xmath110 ) .
the dynamics of the coded information in a pulsed - driven qubit is investigated within and outside the rotating wave approximation ( rwa ) . within rwa ,
the initial atomic state setting plays the central role for the fidelity of sending information from one location to another .
the fidelity and the exchange information are increased by decreasing the phase @xmath46 and the polarized angle @xmath59 of the initial atomic coherent state @xmath111 .
the sensitivity of the fidelity of the transmitted information to the detuning parameter is discussed , where for large values of detuning , the fidelity decreases faster .
however the maximum value of the exchange information between the qubit and the environment is slightly affected as one increases the detuning parameter .
also , we have investigated the effect of the initial state setting and the detuning on the speed orthogonality where it is shown that , the driven qubit could be used to achieve quantum computations much faster .
outside rwa , we have examined the effect of the iteration parameter @xmath75 on the dynamics of the exchange information and the fidelity of the transmitted information . although the minimum value of the fidelity increases as one increases the value of @xmath75 , the exchange information between the travelling information and the environment increases . the excitation angle
( @xmath112 has a small effect on the dynamics of the fidelity of the travelling information .
however , for smaller value of the phase ( @xmath46)the minimum value of the travelling information s fidelity increases and its maximum value reaches one . the speed of orthogonality , for the system treated outside rwa is investigated for different values of @xmath75 , excitation and phase angles @xmath113 .
we show that , the orthogonality speed decreases for large values of @xmath75 and the initial phase @xmath46 .
however the speed of orthogonality decreases as one decreases the excitation angle ( @xmath59 ) .
n. margolus and l. b. levitin , physica d 120 , 188 ( 1998 ) ; j. batle , m. casas , a. plastino , a. r. plastino , phys .
a * 72 * 032337 ( 2005 ) .
a. borrs , m. casas , a. r. plastino and a. plastino , phys .
a * 74 * , 022326 ( 2006 ) .
p. a. rodgers and s. swain , opt , commu . *
81 * 2991 ( 1991 ) .
a. joshi and s. s. hassan , j. phys . b*35
* 1985 ( 2002 ) . s. s. hassan , a. joshi and n. m. m. al - madhari , j. phys .
b * 41 * 145503 ( 2008 ) ; and corrigendum : j. phys .
b * 42 * 089801 ( 2009 ) . | we investigate the transfer and exchange information between a single qubit system excited by a rectangular pulse .
the dynamics of the system is treated within and outside rotating wave approximation ( rwa ) .
the initial state of the qubit plays an important role for sending information with high fidelity . within rwa , and as the fidelity of the transformed information increases the exchange information with the environment increases . for increasing values of atomic detuning ,
the fidelity decreases faster and the exchange information has an upper limit .
outside rwa , the fidelity of the transformed information increases as one increases the perturbation parameter .
however the exchange information is very high compared with that within rwa .
the orthogonality speed of the travelling qubit is investigated for different classes of initial atomic state settings and field parameters .
information transfer and orthogonality speed via pulsed - driven qubit + n. metwally and s. s.hassan + department of mathematics , college of science , university of bahrain , + p. o. box 32038 kingdom of bahrain + [email protected] , [email protected] + * keywords : * driven qubit , exchange information , fidelity , orthogonality |
the primary goal of the wilkinson microwave anisotropy probe ( _ wmap _ ) mission @xcite is to study the cosmic microwave background ( cmb ) anisotropy , an undertaking that requires detailed knowledge of the instrumental angular response and noise , low systematic errors , and absolute calibration at sub - percent levels .
these attributes make _ wmap _ data of natural interest for ancillary studies of sky - based calibration objects .
this paper focuses on _ wmap _ observations of selected planets and fixed celestial objects which have been commonly used as calibrators in the microwave , either by ground- or space - based instruments .
we provide fundamental measurements for these objects which will be of use not only for calibration purposes , but also other studies , such as planetary radiative transfer . a brief discussion of _ wmap _ planetary radiometric data was made by @xcite , with an earlier discussion of jupiter temperatures by @xcite .
we extend these presentations to include _ wmap _ observations of mars , jupiter , saturn , uranus and neptune over a seven - year interval ( 2001 - 2008 ) , in five microwave passbands ranging from @xmath423 to 94 ghz .
we also select five bright celestial sources for more detailed study : cas a , cyg a , tau a , 3c58 and 3c274 .
the _ wmap _ project routinely compiles and distributes a point source catalog , which tabulates mean flux values in those same five passbands for bright sources off the galactic plane @xcite .
however , four of the five objects discussed here ( cas a , tau a , cyg a , 3c 58 ) have low galactic latitudes and are not included in the catalog .
an earlier analysis of three - year _ wmap _ tau a observations by @xcite included both intensity and linear polarization measurements .
this paper is one of a suite that details the analysis of _ wmap _ seven - year data .
an overview and description of data processing methods is provided by @xcite .
@xcite discuss galactic foregrounds and techniques for removing them ; an updated point source catalog is included .
analysis of cmb power spectra , best fit models and cosmological interpretations are presented by @xcite and @xcite .
@xcite assess potential cmb `` anomalies '' relative to the best - fit six parameter @xmath5cdm model .
we organize this paper as follows : a short discussion of the _ wmap _ instrument , plus data strengths and limitations for calibration source work , is provided in section [ sec : wmapintro ] .
the paper then divides into two major sections : analysis and results for planets ( section [ sec : planets ] ) , and then section [ sec : celsrc ] discusses analysis methods and results for the five bright celestial sources . a closing summary and conclusions follow in section [ sec : summary ] .
since its arrival at the earth - sun l2 point in 2001 august , _ wmap _ has produced a well - calibrated , multi - wavelength microwave survey of the entire sky with high sampling redundancy .
detailed descriptions of the mission profile and hardware can be found in @xcite , @xcite and @xcite . to assist with clarifying the terminology and analysis methods presented in this paper
, we describe the key features of the mission .
_ wmap _ observes the sky differentially : signal from the sky is simultaneously sampled in two different directions @xmath6 apart and guided by 10 feed horn pairs into 10 receivers , referred to as differencing assemblies ( das ) .
the receivers cover five frequency bands ( k , ka , q , v and w ) with approximate center frequencies of @xmath423 , 33 , 41 , 61 and 94 ghz , respectively .
there is one da each for k and ka bands ( k1 , ka1 ) , two das each for q and v bands ( q1 , q2 , v1 , v2 ) and four das ( w1 , w2 , w3 , w4 ) for w band .
the frequency passbands are broad ( @xmath7 fractional bandwidth ) , so that the exact frequency associated with an observation depends on the source spectrum .
an orthomode transducer at the throat of each feed separates the sky signal into two orthogonal polarizations , allowing for reconstruction of both intensity and linear polarization signals .
angular resolution is moderate , ranging from roughly @xmath8 fwhm in w band to @xmath9 fwhm in k band . as a survey mission , _ wmap _
is not commanded to point at specific targets , but executes a pole - to - pole sky - scanning strategy whose position on the ecliptic advances @xmath10 each day . to facilitate high redundancy in sky coverage
, the spacecraft rotates about a spin axis , which also precesses .
these combined motions produce a @xmath11 wide scan swath which looks like a `` belt '' projected onto the sky , passing through the ecliptic poles .
the scan swath returns to its original starting position on the ecliptic every year , after completing a full - sky survey with multiple observations per sky pixel . a primary strength of _ wmap _ data is the calibration accuracy , with an absolute calibration error of @xmath12 @xcite .
the absolute calibration is tied to knowledge of the cmb monopole and the spacecraft orbital velocity , and thus is independent of any reliance on previously established measurements of the sources under study . when coupled with _ wmap _ s temporal and spectral coverage , the accurate flux measurements can provide a useful measure of intrinsic or geometrically induced source variability .
there are some limiting factors , however .
temporal coverage of an object is constrained by the sky - scanning method .
for a fixed source on the ecliptic , the _ wmap _ scan strategy provides for two @xmath13-day windows a year in which to observe that source .
these `` observing seasons '' are spaced roughly six months apart .
while this geometry is modified somewhat for moving objects , only mars is near enough that the orbital motion can extend a single viewing window from 45 days to as much as 60 to 90 days . since _ wmap _
does not dwell on an object , but simply scans as it passes through the viewing swath , the number of observations for an individual object is also limited , which affects the signal - to - noise . finally , _
s moderate spatial resolution , which allows us to characterize the planets as calibrators , also places detailed knowledge of structures such as saturn s rings beyond instrument capabilities .
there are five planets with sufficient signal - to - noise for analysis . listed in order of decreasing signal - to - noise ,
these are jupiter , saturn , mars , uranus and neptune . since _ wmap _
points away from the sun , earth and moon in its l2 orbit , it can not observe mercury or venus .
planets are excluded from the standard map - based products used for _ wmap _ studies of the cmb and galactic foregrounds .
observations of the planets must be extracted from the calibrated time - ordered data ( tod ) , and then undergo quality assessment and background removal before they can be analyzed .
the extraction process gathers all observations from the tod taken within several degrees of each planet and associates ancillary information such as position in the focal plane , attitude , planet viewing aspect and data quality flags with each observation .
observations whose sky positions lie within a @xmath14 radius of other planets or which are of suspect quality are immediately rejected .
in addition , observations are excluded from use if either beam of a da feed - horn pair samples sky coordinates that fall within the boundaries of a spatial processing mask @xcite .
the purpose of this `` galactic masking '' is to mitigate background removal error by excluding observations located near strong - signal , high - gradient regions of the galaxy . for those observations which survive these quality control steps ,
galactic and cmb background signals are subtracted .
the background signal subtracted from each observation is computed from the da - appropriate seven - year mean map ( from which planetary data were excluded ) , using the value from the @xmath15 wide pixel which is closest in spatial coordinates to those of the time - ordered observation .
table [ tab : planetseasons ] lists the beginning and ending dates of the observing seasons for the five planets , i.e. , when they are directly in the _ wmap _ viewing swath .
the `` sky proximity '' column in the table indicate probable reasons for an unusually high rejection of observations within a season . as noted in @xcite , planets are `` point sources '' for _ wmap _ and thus the brightness temperature of a planet @xmath16 can be computed via the relation @xmath17 where @xmath18 is the peak response observed by _
wmap _ , @xmath19 is the main - beam solid angle , and @xmath20 is the solid angle subtended by the planet . changing distance between the planet and _ wmap _ , plus changes in viewing aspect of an oblate planet , will cause @xmath20 ( and also @xmath18 ) to vary with time
. we account for both geometrical effects and compute the brightness temperature relative to a fiducial solid angle @xmath21 via the use of time - dependent geometrical scaling factors @xmath22 and @xmath23 . the factors
@xmath22 and @xmath23 may be thought of as `` distance correction '' and `` disk oblateness correction '' factors : @xmath24 where @xmath25 and @xmath26 are a fixed fiducial disk area and distance , respectively .
@xmath27 is the projected area of the planetary disk , and @xmath28 the distance of the planet from _
, at the time of each observation .
the projected area of the oblate disk is computed from @xmath29 where @xmath30 is the planet s equatorial radius and @xmath31 is the projected polar radius given by @xmath32^{1/2 } , \label{eq : projrad}\ ] ] where @xmath33 is the planetary latitude of _ wmap _ ( `` sub-_wmap _ latitude '' ) .
the fiducial disk area @xmath25 is simply @xmath27 evaluated at @xmath34 of @xmath35 .
planetary radii and north pole directions used to compute @xmath33 are adopted from @xcite ; the polar and equatorial radii are listed in table [ tab : radii ] . because _ wmap _
is at l2 , @xmath33 is not very dissimilar to that of the sub - earth latitude , @xmath36 .
as the brightest `` point source '' visible to _ wmap _ , jupiter is used to characterize the main beam response of the instrument .
this process and analysis are described in detail in @xcite , @xcite , and @xcite .
the end result is a set of seven - year , azimuthally - symmetrized , one - dimensional radial beam profiles per da .
these radial beam profiles are then used to derive window functions used for cmb power spectrum deconvolution .
unlike the rest of the planetary temperatures in this section , the seven - year mean jupiter temperatures are derived from the @xmath37 value of the unnormalized window functions ( which measures @xmath38 ) , coupled with the fiducial solid angle for jupiter @xcite . for all other planetary analysis ,
brightness temperatures are derived using a template - fitting technique , which takes advantage of the high - quality beam information obtained from seven years of jupiter observations .
after the processing steps described above , radial response profiles are produced for each planet , per da , per season , by assigning observations to 0.25 arcmin bins increasing in angular distance from the planet s center .
each bin is assigned a mean value and an error based on instrument noise appropriate to the number of observations within the bin .
the known radial beam response , based on seven - year jupiter data , is used as a fitting template to the seasonal radial profile for each da .
the fit returns @xmath18 for the planet .
the brightness temperature is then derived from @xmath18 , the main - beam solid angle @xcite and the reference solid angle for the planet as per the above equations .
errors include the formal error in the template coefficient and the propagated error in the beam solid angle .
figure [ fig : radialprofiles ] illustrates sample w1 radial profiles for a single observing season for each of the planets .
the beam template is overplotted in red .
as well as illustrating the technique , the figure is also a useful comparator of signal quality for each planet at w - band .
beam dilution increases with decreasing _
frequency : @xmath39 is roughly 24.6 , 14.4 , 8.98 , 4.20 and 2.10 @xmath40 sr for the k - w bands respectively @xcite .
tabulated brightness temperatures are appropriate for the rayleigh - jeans ( rj ) approximation .
brightness temperatures provided in the tables do not include the cmb contribution blocked by the planet but included in the background .
as noted in the tables , absolute brightness temperature is obtained by adding 2.2 , 2.0 , 1.9 , 1.5 and 1.1 k in bands k , ka , q , v and w respectively @xcite , since in the rj approximation the cmb temperatures are frequency dependent .
_ wmap _ data points shown in comparison to measurements in the literature are converted to absolute brightness .
jupiter is a bright , accessible source whose whole - disk observations have long been used for calibration in the infrared through radio . @xcite
recommend the use of _ wmap _ radiometry of jupiter as the preferred method of transferring the _ wmap _ dipole calibration to another microwave instrument .
for some instruments , jupiter is unfortunately too bright , hence the interest in the use of fainter sources , and a set of observations which tie several common sources together on the same calibration scale .
we continue to recommend jupiter as the primary comparison calibrator , however .
seven - year mean brightness temperatures for jupiter , computed as described in section [ sec : methods ] , are presented in table [ tab : juptemp ] .
the seven - year values differ from those of the five - year analysis of @xcite by @xmath0 or less .
the largest changes , which hover near @xmath0 , are for k band and some of the w band das . a detailed explanation of these changes is given in @xcite .
briefly , the change in k band results from improved background subtraction , and the w band changes are driven by small revisions to the beam solid angles , based on modeling .
the seven - year temperatures for w band show a decreased scatter between the individual das compared to five - year , implying an improved calibration .
the jupiter temperature uncertainties listed in table [ tab : juptemp ] are derived from the quadrature sum of the solid angle error and absolute calibration error of @xmath12 @xcite , as they were in the five - year analysis .
thermal emission from jupiter s atmosphere is the primary signal source at wavelengths spanned by _
wmap_. as such , the observations are helpful for providing constraints on atmospheric models for jupiter , which in turn could be used for calibration purposes
. however , as noted by e.g. , @xcite , there is a fine point to consider : small contributions to the total observed intensity from a synchrotron component up to about 40 ghz are expected .
this non - thermal `` decimetric emission '' arises from charged particles trapped close to jupiter by its strong magnetic field , and dominates the spectrum at frequencies less than about 3 ghz .
estimates in the literature place the synchrotron contribution at 23 ghz near 1% of the total intensity @xcite , with a roughly @xmath41 frequency dependence @xcite . however , the synchrotron emission is known to be variable on a wide range of time scales @xcite , and accurate removal is difficult .
seven - year mean jupiter brightness temperatures are graphically compared against those in the literature in figure [ fig : jupspec ] .
ground - based observational values are taken from compilations of @xcite and @xcite , along with individual observations from @xcite , @xcite and @xcite .
estimates of the synchrotron emission have been removed from the observations in the figure by many , but not all , of the authors , and not necessarily to the same level .
observations are plotted `` as observed '' . as an estimate of the stability of jupiter s emission with orbital phase over the seven - year interval , we extend the five - year analysis of @xcite , in which temperatures for individual jupiter observing seasons are computed and compared to the multi - year mean values . individual seasonal temperatures
are derived using the same beam - profile fitting technique which is used for the other planets ( section [ sec : methods ] ) .
we compute @xmath42 as the mean deviation of all das from their seven - year mean values , and include a @xmath0 standard deviation as a measure of coherency .
these results are listed in table [ tab : juptime ] . using the _ wmap _ five - year data , @xcite quote an upper limit of @xmath43 based on the largest deviation and scatter seen in the first 10 observing seasons , which occurred in season four . a careful examination of the entries in table [ tab : juptime ] with the largest deviations and/or scatter ( observing seasons 4 , 11 , 12 ) shows that these particular instances have more to do with template fitting and background subtraction quality than with intrinsic source variability . the two seasons with high scatter ( 4 , 12 )
are those in which the number of available observations for fitting has been significantly reduced because of foreground masking ( `` % masked '' in the table ) , which increases the statistical error in the temperature determination .
season 11 , with the highest @xmath42 , is one in which jupiter is nearly aligned with the galactic center , but outside the processing mask .
this is a region in which background subtraction is subject to greater error because of higher temperature gradients across the pixels in the map used for background removal .
if one ignores these three entries , then an upper limit of @xmath1 is more appropriate .
we also searched for indications that k band might be more intrinsically variable than w band , because of the small synchrotron contribution .
there is no indication of such a frequency dependence in the data , but the upper envelope here is about 0.5% .
the _ wmap _ jupiter observations were analyzed to determine if any linear polarization signal could be detected .
a special version of the beam mapping software was run for jupiter , in which the angle between the polarization axis of each radiometer and the planet s magnetic dipole was computed for each observation . for each differencing assembly , all observations within a tight radius of beam center were accumulated , corrected to peak beam response , background subtracted , loss imbalance corrected , separated by radiometer , and then used to solve for stokes parameters i , q , u and their errors . for a more complete description of the relevant equations , see @xcite and @xcite .
normalized stokes parameters q / i and u / i are reported in table [ tab : juppoln ] , along with their propagated errors and estimated linear polarization fraction , @xmath44 .
no attempt has been made to correct for noise bias in either @xmath45 or its reported error , which is simply the propagated error .
the estimated signal - to - noise is low , with no significant detections in any differencing assembly .
linear polarization of the synchrotron emission at 15 and 22 ghz has been detected by @xcite using the very large array ( vla ) .
they were unable to obtain a synchrotron polarization fraction at 22 ghz , but estimated 20% polarization based on the 15 ghz observations , which implies roughly a 0.2% polarization fraction at k band using a combined thermal and non - thermal intensity .
_ results are consistent with this estimate , but do not provide any further constraint .
mars value as a calibrator depends upon the accuracy to which its whole - disk brightness variations ( both apparent and intrinsic ) can be modeled . much of the perceived variability from earth - orbit is attributable to geometrical factors : the apparent phase as seen by the observer and changes in solar irradiance as mars distance from the sun varies in its elliptical orbit @xcite .
there are a number of additional complexities , however .
the day / night cycle imparts a significant rotational modulation to the observed temperature as the surface heats and cools . in the infrared , variable atmospheric effects including molecular line absorption and dust storms
can be of importance . additionally , the martian surface is itself far from homogenous , with polar ice - caps of changing extent and variations in albedo and thermal inertia on relatively small scales @xcite .
planetary science studies notwithstanding , the astronomical community has most consistently referenced two basic models when calibrating to mars : the `` wright model '' ( @xcite , @xcite , @xcite ) and the `` rudy model '' ( @xcite , @xcite ) , both developed before the current level of detail available from orbiters such as mars global surveyor ( mgs ) . in the thermal emission regime ( @xmath46 ) , the planet s brightness temperature can be computed to first order as re - emission of absorbed solar energy by a surface layer with a characteristic albedo , emissivity and thermal inertia @xcite .
the wright model uses this approach to predict thermal emission in the 10 - 350 range , with no positional variations of parameters such as albedo and thermal inertia .
the rudy model originates from vla observations at 2 and 6 cm , which sample further into the martian subsurface .
concurrent measurements of the polarized flux are used to estimate radio reflectivities and absorption lengths as a function of planetary latitude , and brightness temperatures calculated via radiative transfer as a function of position on a longitude / latitude grid .
typical errors on @xmath47 quoted for the wright model are @xmath48 , or about @xmath49k , and those for the rudy model at 5%-10% .
we have chosen in this paper to compare the _ wmap _ observations against predictions from the wright model , primarily because of its long history of use and readily available model code @xcite , but also because the seasonal brightness temperatures quoted for _ wmap _ are the resultant average over several weeks of data , which means we are only sampling mean properties at best .
we have modified the 2007 wright model code in three ways : ( 1 ) jpl ephemeris positions are substituted for the lower precision formulae originally used ; ( 2 ) an option is added to use _ wmap _ spacecraft positions rather than those of the earth ; and ( 3 ) computations are extended out to _ wmap _ frequencies .
the first two changes are not critical to the analysis , as the combined modifications introduce at most at 0.6 k difference from the original results .
typically the model predicts brightness temperatures at 3 mm which are a few kelvins lower than those at 350 , and the temperature spread between the extrema of the 5 _ wmap _ frequencies is quite low , of order 0.5 k. @xcite presented the first five years of _ wmap _ mars observations at 94 ghz .
here we extend the w - band observations to seven years , and include the observations at the other four _ wmap _ frequencies .
mean brightness temperatures for mars at five frequencies and seven _ wmap _ observing seasons derived from the seven - year data are presented in table [ tab : mars ] .
after correcting to absolute brightness temperature and converting from rayleigh - jeans to planck brightness temperature , the seven seasonal w - band averages are shown in comparison to the wright model predictions at 3.2 mm in figure [ fig : marswband ] .
this plot is similar to figure 15 of @xcite , whose preliminary analysis found that the wright model was a good predictor of the mean temporal variation , but the model @xmath50 seemed high by roughly 10% .
our results generally agree , but we find that multiplying the wright model by a factor of @xmath51 fits the seven - year data best .
if one uses the wright model tabulated at 350 , the plot looks nearly identical , but the scaling factor is 0.941 , or about a 6% reduction .
a previous comparison by @xcite indicated that the 3 mm whole - disk observations of @xcite were within 7.5% of wright model predictions at 350 . the first _ wmap _
mars observing season occurred during the time of the 2001 july - october massive global dust storm which elevated atmospheric temperatures in the mid - ir by nearly 40 k @xcite .
the effect of dust storms on microwave data is expected to be less of a concern , but we did fit for the model scaling factor both with and without inclusion of the first observing season , with no significantly different results . the uncertainties on the derived _ wmap _ seasonal brightness temperatures increase with decreasing frequency , which follows as a result of decreasing signal - to - noise in the data .
however , it is possible to extract a trend in temperature with frequency by finding within each season the ratio of the brightness temperature at each frequency to that at w band , and then computing the mean over all seven observing seasons as a function of frequency .
these means are plotted in figure [ fig : marsfreqdep ] in black .
errors are standard deviations of the mean .
the weak frequency dependence can be characterized roughly as @xmath52 , indicated by the blue line .
the wright model , processed in the same fashion , predicts almost no variation with frequency ( red line ) . such a frequency dependence could be attributed to an emissivity change of @xmath53 over the wavelength interval . in a similar vein , a weak @xmath54 decline in surface emissivity with increasing wavelength was noted by @xcite in the 50 to 180 range , based on analysis of infrared space observatory ( iso ) observations of water vapor absorption lines .
some consideration was given to placing the _ wmap _ observations in a wider context , with the goal of gaining further insight into the model parameters .
a search was made for a dataset that would provide absolutely calibrated , whole - disk observations of mars which spanned both sides of the thermal energy distribution peak and were taken over a reasonably contiguous interval of time .
mars data taken by the cobe / dirbe instrument @xcite for a 146 day interval ( 1990 april 28 - september 20 ) most closely matched these requirements , although no analysis of the publicly available data appears to have been previously published .
there are six dirbe wavebands which sample the thermal emission portion of the mars spectrum , with effective wavelengths of 12 , 25 , 60 , 100 , 140 and 240 . of these
, the 25 observations can not be used as an independent check of the wright model , since the model itself was the ultimate source of the dirbe absolute calibration in this band ( s. moseley & w. glaccum 2009 , private communication , @xcite ) .
in addition , while the 12 , 25 , 140 and 240 detectors are well - behaved linear devices , the 60 and 100 detectors are known to suffer non - linear response , and mars is on the bright end of that response .
dirbe is a broad - band photometer , and reports monochromatic flux values at the effective wavelength of each band assuming a constant source spectrum in @xmath55 across the band .
one must apply a `` color correction factor '' to the reported fluxes if the source spectrum is different from that assumed .
the dirbe explanatory supplement provides such factors for gray - body and power - law spectra .
however , in the case of mars , some of the dirbe bands intercept known atmospheric absorption lines .
a significant co@xmath56 absorption centered at 15 lies well within the 12 passband , and grazes the blue edge of the 25 band .
numerous minor water vapor absorptions are known to occur at wavelengths sampled by the 60 and 100 bands @xcite .
we have used spectral data available from the mgs thermal emission spectrograph ( tes ) public archives and the iso spectrum published by @xcite to form approximate correction factors for the dirbe data at these bands : only the 12 band required significant correction .
for the 140 and 240 bands , color corrections were calculated using the spectrum predicted by the wright model .
a mean flux at each dirbe effective wavelength was evaluated over the 146 day interval , including application of color corrections and scaling to a fiducial solid angle at a distance of 1.5 au ( table [ tab : dirbemars ] ) .
although the epochs of _ wmap _ and dirbe observations are separated by more than 10 years , the viewing geometry at the time of the dirbe observations is most closely duplicated during _ wmap _ observing season four .
the _ wmap _
mean brightness temperatures for this season were also converted to fluxes using the same fiducial solid angle , and both sets of data plotted as a spectrum in figure [ fig : marsdirbecomp ] .
the spectra predicted by the wright model for the mean time of the two epochs are also shown on the plot .
for clarity , the quoted model errors of about @xmath48 are not shown on the plot .
on the logarithmic scale needed to show both sets of data together , the two model spectra overlay each other . within the quoted data and model errors , the dirbe measurements at 60 and 100 do not match well with the model .
the dirbe 12 data deviate by a little more than @xmath0 , accounting for both data errors and model accuracy . in the case of the 60 and 100 bands ,
we suspect a non - linear response to a bright object is the root cause .
the 12 band has a low absolute calibration error , but there is a significant and uncertain color correction due to the co@xmath56 atmospheric absorption
. ratios of the dirbe fluxes to wright model predictions are presented in table [ tab : dirbemars ] ; the major source of dirbe error is that of the absolute calibration , which is also listed .
dirbe observations at 140 and 240 agree within the 10% calibration errors , but can not be used to deduce any trend with frequency , such as seen by _
wmap_. we decided to forgo adjustments to the model at this point .
as currently implemented , the wright model may be used to predict microwave observations at _ wmap _ frequencies by applying a `` post - processing '' multiplicative scaling factor to the published values : @xmath57 . for both model and data evaluated at 3.2 mm ,
we have already noted a scaling factor of 0.953 .
we use the mean ratios of the remaining four frequencies to w band ( figure [ fig : marsfreqdep ] ) as scaling factors for these frequencies , which are in column 2 of table [ tab : marsmodelresid ] .
the w - band observations can be reproduced to @xmath58 , with confidence decreasing with frequency to @xmath59 at k band .
since the 3.2 mm model values we have computed are not published , we also provide scaling factors to the 350 temperatures , which are derived from the the scaling factor of 0.941 noted earlier and multiplied through by the 3.2 mm scaling ratios .
differences between measurements and the wright model at _ wmap _ wavelengths are not unexpected .
the wright model was designed to predict infrared emission arising from the top layer of the martian surface , whereas the microwave signal originates from subsurface layers of order several centimeters deep ( @xcite ) , with potentially different composition and compaction .
this was recognized by @xcite and @xcite , who recommended a logarithmic interpolation between the wright 350 prediction and the 3.3 mm brightness temperature versus mars - sun distance fit of @xcite , citing agreement with the rudy model within 1.2% for a dielectric constant revised upward from 2.25 to 2.8 . on average ,
the @xcite fit predicts brightness temperatures roughly 5% higher than the _ wmap _ mean seasonal values at 3.2 mm .
this suggests that the lower value for the dielectric constant is preferable .
we obtained disk - averaged rudy model predictions at _ wmap _ seasonal mean times of observation using a web - based , slightly revised version of the model maintained at nrao .
in general , rudy model predictions using the `` standard dielectric constant '' were slightly high compared to the _ wmap _ observations , by @xmath60% at 3.2 and 4.9 mm and roughly 3% at 7.3 , 9.1 and 13.1 mm .
it should be noted that the _ wmap _ observations provide model constraints only in the disk - averaged sense .
the post - processing scaling of the wright model suggested here simulates a frequency - dependent emissivity , which implies slow changes in mean subsurface properties with depth . in the rudy model , which already incorporates subsurface sampling , use of model parameters derived from the radio data
produces agreement with _
wmap _ within a few percent .
it is possible that this agreement could be improved either through the use of a revised dielectric function or updated thermal inertia and albedo maps .
the thermophysical parameters used by the rudy model are based on @xmath420 year old maps of thermal inertia and albedo from the infrared thermal mapper ( irtm ) aboard viking @xcite .
more recent mgs maps of these quantities obtained using the tes instrument are available with roughly 10 to 40 times better spatial resolution , full planet coverage and approximately twice the signal - to - noise ( @xcite ) .
the surface albedo has been observed to have some time dependence : a visual comparison between the viking irtm and tes albedo maps shows alterations in the spatial pattern have occurred since the viking mapping @xcite . as with mars ,
saturn s microwave brightness temperature varies considerably over time for near - earth observers , primarily as a result of geometrical effects . here
the dominant factors are the observer s changing viewing angle of the rings and projected disk of the oblate planetary spheroid .
the ring system presents the main difficulty in saturn s use as a calibrator in the microwave , although substantial progress has been made in understanding and modeling it . as evidenced by images from hubble space telescope ( hst ) and data from orbiter and flyby missions such as cassini and voyager ,
saturn s ring system is complex and rich in detail .
the brightest rings ( a , b , c ) are composed of objects typically less than 5 m in size @xcite and are composed primarily of water ices mixed with a small amount of impurities ( @xcite , and references therein ) .
those portions of the rings between the observer s line - of - sight and the planet serve to attenuate emission from saturn s disk .
the rings also contribute to the microwave signal via a mix of scattering and thermal re - emission of planetary radiation .
the crossover point between the two emission mechanisms occurs near 1 cm @xcite , with thermal emission dominating shortwards and scattering longwards .
optical depths have been measured at a variety of wavelengths : @xcite found that the optical depth of the combined abc rings is roughly gray over a wide frequency range .
voyager i radio occultation observations @xcite at 3.6 and 13 cm produced radial profiles of optical depth for the a and c rings and cassini division , but only approximate results for portions of the b ring with @xmath61 .
the most complex and recent modeling of microwave data is that of @xcite who use a monte carlo radiative transfer code to simulate the scattered and thermal emission of the rings based on vla and other high - resolution observations .
_ wmap _ does not spatially resolve the individual components of the saturn system , nor is the spectral coverage sufficient to definitively isolate ring scattering and thermal spectral regimes .
however , the changes in viewing geometry with orbital phase can be used to break degeneracy between ring and disk components , especially as the viewing aspect of the rings becomes more `` edge - on '' . over the currently available seven - year observing baseline ,
saturn s rings are seen at inclinations between @xmath62 and @xmath63 : the sign convention indicates that the planet s south pole is tilting slowly away from the observer as the @xmath35 edge - on equatorial configuration is approached .
figure [ fig : satincl ] illustrates the two extrema and table [ tab : saturn ] lists mean times and ring opening angles for each observing season .
@xcite briefly discussed _ wmap _ five - year w - band observations of saturn , which sampled ring opening angles @xmath64 .
they found that the brightness temperature could be fit remarkably well with a simple @xmath65 dependence .
this convenient relationship does not effectively characterize the behavior at lower inclinations , however , and so a more complex formulation is needed for the seven - year data .
we adopt a simple empirical model of saturn s microwave emission , variants of which have appeared in the literature for over 30 years ( e.g. , @xcite ) .
the goal is to provide a predictive formula for the unresolved observed brightness of saturn with as low an error as possible : however , given the _
data constraints , such a model will lack physical detail . at a given frequency @xmath66 , we assume a single temperature for the planetary disk , @xmath67 , and that all rings are characterized by the same temperature @xmath68 .
the model variant we adopt allows for seven radially concentric ring divisions as defined by @xcite , which are based on optical depth variations observed by voyager @xcite .
each of the seven ring sectors has its own ring - normal optical depth @xmath69 , with @xmath70 , but each @xmath69 is assumed to be both constant within its ring and frequency independent .
those portions of the planetary disk which are obscured by ring cusps will have their emission attenuated by a factor @xmath71 , where @xmath72 is the ring opening angle seen by the observer .
thus at a given frequency and ring opening angle @xmath72 : @xmath73 + t_{\mathrm{ring}}(\nu ) \sum_{i=1}^7 a_{r , i},\ ] ] where @xmath74 , @xmath75 and @xmath76 are the projected areas of the unobscured disk , the portion of the disk obscured by ring i , and @xmath77 ring , respectively .
these areas are normalized to the total ( obscured+unobscured ) disk area .
@xmath68 is an observed brightness temperature and hides such physical details as ring emissivity and the apportionment of the scattering versus thermal emission .
the assumption of a single mean ring temperature is a convenience rather than reality @xcite .
however , the apportionment between individual rings is not critical to the model fit , as it only depends on the summed total ring emission .
potential contributions to modeling error from nonuniform planetary disk emission are discussed later in this section .
the _ wmap _ brightness temperatures in table [ tab : saturn ] are sorted by frequency band and @xmath72 and then a single simultaneous fit for the model parameters is made .
two separate fits were tried , differentiated from one another in the handling of the ring optical depth . in one fit , the @xmath69
were held fixed at the values chosen by @xcite ( table [ tab : satmodel1 ] ) , and the ten disk and ring temperatures ( for each of 10 das ) were solved for . in an alternate fit , the relative ratios between the @xmath69 were fixed as per the table , but a single @xmath78 was solved for in addition to the ten free temperatures .
both fits returned similar results , with reduced @xmath79 of @xmath80 for @xmath81 degrees of freedom .
we show only the results for the fixed-@xmath82 fit .
the disk and ring brightness temperatures derived from this model fit are presented in table [ tab : satmodel2 ] .
formal errors for the ring and disk temperatures at lower frequencies are higher because the ring contribution is smaller and there is a larger covariance between the two components . the model fit and residuals are shown in figure [ fig : satmod ] .
the model reproduces the observations to within @xmath83 , but observations closer to @xmath84 would better constrain the disk temperatures and so reduce the covariance between the disk and rings .
the parameterization chosen for the ring optical depths , together with the incomplete sampling of the ring system orientation , allows for alternate models which would fit the data equally well but return somewhat different mean disk and ring temperatures .
this systematic is not reflected in the formal fitting errors in table [ tab : satmodel2 ] . in order to assess the magnitude of this error
, we explored a small number of model variants .
the model which fit the data equally well but returned temperatures with the largest differences compared to table [ tab : satmodel2 ] consisted of a combined abc ring represented by a single @xmath85 , which was a free parameter in the fit .
disk temperatures derived for this model were @xmath86 k higher , and ring temperatures @xmath60 k lower .
as a conservative approach , we additively combine this estimated systematic error with the formal fitting errors to produce the `` adopted error '' columns in table [ tab : satmodel2 ] .
the derived disk and ring temperatures match well with those already in the literature .
figure [ fig : satspec ] shows the _ wmap _ derived disk temperatures in context with the compendium of @xcite plus more recent interferometric observations .
the ring temperatures also compare well with those measured directly from high - resolution images of the ring ansae @xcite : see figure [ fig : satringspec ] .
figure [ fig : satringaffect ] illustrates the percentage contribution of the ring to our total model , which sums the ring emission and attenuation of disk radiation , as a function of frequency and ring inclination @xmath72 .
the empirical model described here assumes saturn s mean whole - disk temperature is time invariant at each frequency .
the presence of band - like structures ( which can persist for years ) and latitudinal brightness gradients at the @xmath87% level have been noted by several observers at wavelengths of 2 cm and longward @xcite . there is less discussion in the literature of disk temperature structure and/or variability at the frequencies observed by
there have been no reports of bands or localized structures , but some indication of north / south latitudinal brightness differences . within the seven - year epoch under consideration
, @xcite observed a latitudinal disk brightness gradient at 3 mm , such that the north pole was brighter than the south by @xmath87% .
they interpreted differences between their 1.3 mm and 3 mm data , taken roughly 4 months apart , as possible evidence for atmospheric changes on week to month timescales . in order to estimate potential model bias caused by the assumption of a uniform disk temperature ,
saturn seasonal brightness temperatures were simulated using geometrical models similar to those described in this section , with the addition of a disk latitudinal gradient as described by @xcite .
if one makes the assumption that such a temperature gradient is persistent over seven years and exists at all _ wmap _ frequencies , then such structure would produce a temperature behavior with inclination which is asymmetric about @xmath88 .
this is because the rings would obscure hotter regions in the northern portion of the disk but colder regions in the south , producing a maximum deviation from symmetry of @xmath60% at the largest ring opening angles . since the _ wmap _ seven - year observations sample slightly less than half of the full @xmath72 range ,
this asymmetry would not be evident , and a model fit which assumes a constant disk temperature would use the ring model component to compensate for the disk temperature gradient . for a simulated dataset matching the _ wmap _ seven - year observing seasons , the model fit returns a mean ring temperature biased by less than 1 k , with an unbiased recovery of the mean whole - disk temperature .
this estimate is of uncertain quality , however , because it is based on a snapshot of saturn at one wavelength and a narrow time window .
for this reason , the adopted errors in table [ tab : satmodel2 ] have not been adjusted to reflect this potential effect .
an extended , low optical depth dust ring associated with saturn s moon phoebe was recently discovered using spitzer imaging at 24 and 70 @xcite . with a quoted radial extent between at least 128 and 207 saturn radii , or roughly @xmath89 in apparent diameter
, it would be possible for _ wmap _ to resolve the phoebe ring , with w band as the best chance for doing so .
however , no signal from the ring was detected by _
wmap_. properties of the grains are not well specified , but naive estimates assuming typical dust emission and optical depth frequency dependencies would predict temperatures at 94 ghz of less than 1 @xmath90k , well below achievable noise levels of about 300 @xmath90k ( @xmath0 ) . with their lower brightness , lack of a dominant ring system , and small solid angle ,
uranus and neptune are used as primary calibrators in infrared through radio wavebands for both ground- and space - based instruments .
spectral coverage in the microwave is somewhat undersampled , however , leaving room for interpretation in atmospheric modeling efforts .
the microwave spectra of uranus and neptune lack the broad nh@xmath91 absorption centered near 24 ghz which is characteristic of jupiter and saturn . at millimeter wavelengths , collision - induced absorption by h@xmath92
is considered the dominant continuum opacity source .
co rotational absorptions have been observed for neptune in the sub - millimeter @xcite . at centimeter wavelengths , ammonia ( which is depleted relative to solar nitrogen levels ) and hydrogen sulfide
are the main opacity contributors @xcite .
of special note is uranus unique @xmath93 obliquity , which allows for a slowly changing pole - to - pole panorama of the planet as it moves in its 84 year orbit , and may also play a role in determining atmospheric conditions . for selected frequencies , observational databases spanning decades
have permitted characterization of the variability of these two planets .
uranus in particular has exhibited long - term whole - disk temperature changes since 1966 which correlate with the viewing aspect of the south pole .
after accounting for geometrical solid angle changes , @xcite reported a gradual temperature drop of order @xmath94 over 20 years ( 1985 - 2005 ) at 90 ghz .
this is interpreted as a true integrated - disk temperature change , as the 1985 face - on contribution from the bright south pole progressively morphs into a 2007 view dominated by colder equatorial regions .
similar findings , with similar phasing albeit different amplitudes , have been reported in the radio @xcite and visible @xcite .
there is weaker evidence that some portion of variability in the light curves is attributable to changes deep in the atmosphere @xcite ; @xcite used vla 2 and 6 cm `` snapshots '' of the disk over several years to argue for opacity changes in zonal bands .
neptune s microwave variability is less well documented , although there is ample evidence for variability in the visible and near - ir , as summarized by @xcite .
@xcite find the 90 ghz integrated disk temperature to be constant to within @xmath95 over 20 years .
peak _ wmap _ antenna temperatures for these distant `` ice giants '' range from roughly 1 mk at w band to @xmath40.2 mk at k band .
the signal - to - noise for individual observations of these objects is low ( @xmath40.3 at w band , @xmath40.14 at k band ) , resulting in large statistical errors in single - season disk temperature determinations .
tables [ tab : uran ] and [ tab : nept ] present the single - season brightness temperatures computed for uranus and neptune .
brightness temperatures are listed per frequency rather than per da as a means of boosting signal - to - noise for those frequencies with multiple das .
temperatures for uranus observing seasons four and five exhibit somewhat larger error bars than other seasons .
for these two seasons , there was a reduction in the number of observations available for analysis because data quality checks excluded observations in close proximity to mars sky coordinates .
sub-_wmap _ latitudes for uranus range between @xmath96 and @xmath97 over our seven - year baseline .
linear correlations of @xmath47 against both sub-_wmap
_ latitude and time produced no statistically significant trend other than a flat line .
this is not surprising given that the tightest seasonal temperature errors are of order 10% , which encompasses the entire twenty - year range of variation seen in the 90 ghz light curves of @xcite .
we performed similar correlations for neptune , with again the same null variability result .
note , however , that the sub-_wmap _ latitude for this planet is relatively unchanged over the epochs of observation . with no discernable variability over the observing baseline , we computed seven - year means of the disk brightness for each planet ; these are listed as the last line in tables [ tab : uran ] and [ tab : nept ] . these seven - year means in turn may be compared to observations in the literature and placed in context with the microwave spectra in general .
figure [ fig : uraspec ] overplots the seven - year mean _ wmap _ uranus temperatures at k , ka , q , v and w center frequencies on a composite spectrum derived from several sources in the literature .
points are color - coded to reflect the decade in which observations were taken .
data have been culled from @xcite , @xcite , @xcite , @xcite , and @xcite .
also included is the mean 3.5 cm value from the 1966 - 2002 compilation of @xcite , with a single large error bar indicating the @xmath3 k peak - to - peak variation in the light curve . within the quoted errors ,
there is general agreement between the _ wmap _ observations and those taken at previous epochs .
the _ wmap _
observations show a `` dip '' in ka band temperature compared to neighboring frequencies .
although not of high statistical significance , a feature of this nature is clearly of interest .
a literature search for corroborating observations near 30 ghz only returned those few acquired in the late 1960s , listed in table 1 of @xcite . for a variety of reasons ,
more recent observations would be preferable , but the older epoch does have the advantage of sharing a similar viewing geometry to that of the _ wmap _ data .
these older observations are plotted in gray in the figure , and agree well with the _ wmap _ measurements , but also do not define a `` dip '' at high statistical significance . although suggestive of an interesting atmospheric opacity constraint for future study , the possibility of such a feature has gone unremarked in the literature .
a composite microwave spectrum for neptune is shown in figure [ fig : nepspec ] .
the color coding again indicates observational epoch , with ground - based data taken from @xcite , @xcite , @xcite , @xcite , @xcite , and @xcite .
@xcite provide a summary of observations in the 1990s .
observations with error bars larger than about 30 k are not included in the plot .
_ wmap_-derived temperatures for q band agree well with those of @xcite , while the 94 ghz observations lie on the upper envelope of data taken in the 1970s and 1980s .
_ wmap _ observations occur when neptune s south pole is most fully viewable .
the observational error on the seven - year w - band mean is @xmath95 , which is equivalent to the level to which @xcite claimed 20 year stability .
reasonable temperature error bars for the _ wmap _ uranus and neptune observations are only achieved through averaging the seven - year data .
observers wishing to use these data for calibration of observations taken outside of the 2001 - 2008 epoch should keep in mind the variability and related geometric issues described above .
non - variable , spatially isolated fixed celestial calibrators for millimeter wavelengths are not common . at high galactic latitude ,
bright sources are predominantly identified as some form of agn / qso , which are prone to outbursts and variability on a wide range of timescales and frequencies .
brighter sources in the galactic plane tend to be hii regions ( which may not be `` point - like '' ) or supernova remnants ( snrs ) with potential variability . at moderate spatial resolution , confusion with neighboring diffuse and compact sources in the galactic plane can be an issue for background subtraction .
five sources were chosen for study out of an original list which included some of the brightest sources from @xcite and @xcite , in addition to the brightest , least variable objects from the seven - year _ wmap _ source catalog @xcite .
the five selected sources are listed in table [ tab : calobj ] .
some of the sources which were initially considered but ultimately rejected , primarily because of low background contrast , included 3c286 , ngc 7027 , 3c84 , 3c218 , 3c123 , and 3c147 .
there are unfortunately few suitable calibration sources in the southern hemisphere with a long - term history of observation .
flux densities of the selected sources are measured from the seven - year sky maps at healpix sr , with a mean spacing of 6.87 arcmin between pixel centers . ]
resolution 9 ( nside=512 ) . for each frequency band ,
the azimuthally symmetrized beam profile ( see section [ sec : methods ] ) is convolved with a sky map pixel to produce a map - based beam template .
the sum of the pixel - convolved beam template plus a sloping planar base level is fit to the stokes i , q , and u sky map data at each source position , using pixels within 3.5 times the beam width @xmath98 ( 1.5 times the fwhm ) in each band .
the peak source temperature from each fit is converted from thermodynamic temperature to rayleigh - jeans brightness temperature and then translated to a source flux density using a conversion factor @xmath99 that is a weak function of the source spectral index @xcite . for point sources ,
this method of flux measurement is more accurate than the aperture photometry method as used for example by @xcite for tau a. the _ wmap _ beam profiles have extended wings @xcite , so an aperture radius @xmath100 times the beam fwhm should be used and the results are more susceptible to error due to background confusion . the uncertainty in flux density is calculated as the quadrature sum of ( 1 ) map measurement uncertainty , ( 2 ) the uncertainty in @xmath99 ( 0.5% to 0.7% depending on the band ) , and ( 3 ) the 0.2% absolute calibration uncertainty .
for stokes i , map measurement uncertainty is estimated from the rms fit residual in the fit region .
the stokes i residuals generally appear to be dominated by beam asymmetry effects , but also include background confusion and noise .
( the source profiles in the maps are asymmetric due to asymmetry in the instantaneous beam profiles and the nonuniform distribution of scan angles over a year for sources away from the ecliptic poles . ) for stokes q and u , map measurement uncertainty is calculated either from the 1@xmath98 source peak uncertainty and base level uncertainty from the beam fitting or as the q or u flux times the fractional map measurement uncertainty for i , whichever is largest .
the latter method is the estimated uncertainty in q or u due to beam asymmetry effects , assuming the fractional uncertainty is the same in q or u as it is in i. the flux determination was tested on simulated sky maps containing a population of point sources with no other signals and no noise , generated using _ wmap _
beam window functions as described in @xcite . recovered flux densities were accurate to about 0.1% or better , except in w band where the recovered fluxes tended to be larger than the input fluxes by up to 1% .
this is allowed for by including an additional 1% uncertainty term in the quadrature sum for w band .
fractional year - to - year variability for the selected sources has been obtained using sky maps for individual years 1 - 7 . to remove confusion noise from the cmb and galactic foregrounds
, we subtract the seven - year average map from each individual year map for each band .
a pixel - convolved beam plus flat base level is fit to each difference map at each source position , giving a flux difference @xmath101 for the @xmath102th year for each source in each band .
uncertainty in @xmath101 is calculated from the quadrature sum of the source peak uncertainty and the base level uncertainty .
the flux difference is divided by the mean flux from the seven - year map to get the fractional flux variation @xmath103 . for k , ka , and q bands
, there is a small ( @xmath104 ) but significant year - to - year variation in the _ wmap _ calibration , which we have measured by correlating each yearly map against the seven - year map ( see figure 1 of @xcite ) .
the measured fractional flux variations in k - q bands are corrected for these calibration variations @xmath105 using @xmath106 .
source flux densities from the seven - year maps are presented in table [ tab : srcflux ] .
fractional uncertainties for the stokes i fluxes are typically 1 to 3% .
for some of the sources , the maximum source extent given in table [ tab : calobj ] is not entirely negligible relative to the _ wmap _ beam width in v or w band ( fwhm 19.6 in v , fwhm 12.2 in w , @xcite ) .
we have estimated the possible error due to source extent for tau a. the 1.4 ghz map of the tau a region from the nvss survey ( beam size 45 fwhm , @xcite ) was smoothed with the symmetrized _ wmap _ v or w - band beam and converted to a resolution 9 ( nside = 512 ) healpix map . the flux determined by our method was found to underestimate the true flux by 1.5% in v band and 3.7% in w band .
spatial variations of the spectral index over tau a are very small @xcite so the source extent at w band is probably similar to that at 1.4 ghz .
our v and w - band fluxes for cas a and 3c58 may also be underestimated by similar amounts . to allow for this ,
the uncertainty values listed for these three sources in table [ tab : srcflux ] include additional contributions of 1.5% in v and 3.7% in w , added in quadrature to the uncertainties calculated as described in section [ sec : calsrcmethods ] .
fractional year - to - year variability results are presented in figure [ fig : npovarplot ] and table [ tab : calvar ] .
significant secular decrease is seen for cas a and tau a. the results are consistent with a frequency independent decrease of about 0.53% per year for cas a and 0.22% per year for tau a. our results for cas a fall between the @xmath107 per year decrease found by @xcite near 15 ghz from 1965 to 1995 and the @xmath108 decrease found by @xcite at 33 ghz from 2001 to 2004 .
dependence of the cas a decrease on frequency and epoch has recently been discussed by @xcite and @xcite . for tau
a , we find a somewhat higher rate of decrease than that of @xmath109 per year found by @xcite at 927 mhz from 1962 to 1977 and @xmath110 per year found by @xcite at 8 ghz from 1968 to 1984 .
@xcite found a decrease of @xmath111 per year at 30 ghz from 2002 january to 2004 september .
no significant secular trends are seen for the other sources , with @xmath112 upper limits in k band of 0.12% per year for cyg a , 0.21% per year for 3c58 , and 0.54% per year for 3c274 .
the limit for cyg a is consistent with the limit of 0.10% per year found by @xcite at 33 ghz for the period 2001 march to 2004 may . for 3c58 , @xcite reported an increase of @xmath113 per year at 408 mhz from 1967 to 1986 .
significant yearly flux variation is not seen for cyg a and 3c58 ; the rms year - to - year variation is consistent with the uncertainties in each band .
the lowest rms variation is in k band , and is 0.27% for cyg a and 0.33% for 3c58 . @xcite
give an upper limit of 10% on cyg a core variability from observations at 5 , 15 , and 90 ghz over timescales from 10 months to 15 years .
the core contributes @xmath114% of the total flux at _ wmap _ frequencies ( e.g. , @xcite ) . for 3c274
there is evidence for year - to - year variability of about 2% in k , ka , and q bands .
the rms variation is 1.8 to 2.6 times the mean uncertainty in these bands , and the variations are correlated from band to band .
previous observations of 3c274 have shown greater variability . at 90 ghz
, @xcite reported a decrease of 1.4 jy for the core between 1985 june and 1986 august , which corresponds to @xmath417% of the total source flux . at 43 ghz
, vlba monitoring showed an increase of 0.57 jy for the core region between 2008 january and april @xcite , which corresponds to 4% of the total flux .
smaller core flux variations , less than 1% of the total flux , have been observed at lower frequencies @xcite .
spectra of the seven - year _ wmap _ fluxes together with previous measurements from the literature are shown in figures [ fig : casaspec ] - [ fig : spec3c274 ] .
for cas a , we have scaled the _ wmap _ fluxes , the archeops fluxes of @xcite , the scuba fluxes of @xcite , and the blast fluxes of @xcite to epoch 2000 using a secular variation of @xmath115 per year .
these are plotted with previous measurements that were scaled to epoch 2000 by @xcite using frequency - dependent scaling .
for tau a , we have scaled previous measurements from @xcite to the epoch of the _ wmap _ data using a secular variation of @xmath116 per year at all frequencies .
results are not significantly different if @xmath117 per year is used
. table [ tab : specfits ] presents parameters from fits to the spectra for _ wmap _ data alone and for the combined data .
the two fits are generally consistent within their uncertainties over the _ wmap_frequency range .
some notes on the individual sources follow .
_ cas a_ a slightly curved spectrum gives a better fit to the combined data than a power law ( chi - squared per degree of freedom @xmath118 compared to @xmath119 for a power law ) .
the flattening of the spectrum with increasing frequency was previously noted by @xcite .
this may be consistent with observations of spatial variations of the spectrum in cas a ( e.g. , @xcite , @xcite ) .
@xcite presented spectra of 26 brightness peaks from maps with @xmath120 resolution at 1.5 , 5 , 28 , and 83 ghz .
the data were mostly consistent with power - law spectra , with spectral indices ranging from -0.75 to -0.95 .
( for comparison , the overall spectral index from a power - law fit to our integrated spectrum from 1.4 to 93 ghz is -0.73 ) .
such a variation will lead to curvature in the integrated spectrum .
@xcite also found curvature for some of the brightness peaks with the spectra progressively flattening at the higher frequencies , and noted that such curvature is expected from models of particle acceleration in cosmic ray modified shocks @xcite . within the _ wmap _ frequency range , all of the epoch 2000 scaled fluxes are consistent within the uncertainties with the fit to the combined data .
this includes the _ wmap _ fluxes , the absolutely calibrated fluxes from @xcite at 22.29 ghz and @xcite at 32 ghz , the 33 ghz flux from @xcite , which is calibrated using the five - year _
wmap_jupiter temperature , and the 86 ghz flux from @xcite , which is calibrated using the @xcite jupiter temperature . above 300 ghz
, there is excess emission above that expected for synchrotron emission , which has most recently been interpreted as emission from cool dust by @xcite .
the 353 and 545 ghz fluxes from archeops @xcite are much higher than the 600 ghz flux from blast @xcite and the 353 and 666 ghz fluxes from scuba @xcite .
the archeops measurements were made with a larger beam ( @xmath121 compared to @xmath122 or better for scuba and blast ) and appear to be affected by dust emission that is not associated with cas a. the _ wmap _ 23 ghz ( k band ) polarization map for cas a exhibits unexpected structure .
figure [ fig : ipforcelobj ] shows seven - year mean intensity and polarization images centered on each source at each of the five _ wmap_frequency bands .
these images are @xmath123 on a side , have not been background subtracted , and are scaled such that brighter pixels are black .
the polarization ( p ) image for cas a shows an irregular @xmath124 mk ring at a radial distance roughly @xmath125 from the source position .
the angular extent of cas a is @xmath126 ( table [ tab : calobj ] ) , leaving the reality of the ring feature in question . we have attempted to simulate this feature under the hypothesis that it is an artifact introduced by a combination of beam and source spectrum characteristics .
cas a is a steep - spectrum source . as a result of effective frequency differences between the two k - band radiometers @xcite , the k11 radiometer ( fed by the axial omt port ) has an fwhm roughly 3% wider than that of k12 ( which is fed by the lateral omt port ) , and the peak observed signal in k11 is a few percent higher than that of k12 .
simulated beam maps for the two k - band radiometers were generated separately using jupiter data as a template , and then individually scaled to peak values representative of cas a. the difference between the two beams as a function of azimuthal angle and radial distance from beam center can be used to compute a rough estimate of induced cas a polarization signal , under the assumption of complete scan - angle coverage . figure [ fig : artifact ] shows the results of such a simulation , which produces a feature with an approximately correct peak position ( near @xmath127 ) , but slightly wider and 30%-40% brighter than that shown in the data image .
a more complete simulation would include scan - angle coverage effects , which in this scenario are presumed responsible for the gaps in the ring .
this is the only known instance of an apparent artifact in _ wmap _ polarization data .
_ cyg a_ the fluxes from @xcite at 22.29 ghz , @xcite at 89 ghz , and @xcite at 94 ghz are consistent with the _ wmap _ results .
the flux from @xcite at 33 ghz is lower than _ wmap _ ka - band flux by @xmath128 , taking both flux uncertainties into account .
we found that excluding the @xcite flux from the power - law fit to the combined data improved @xmath129 from 2.89 to 0.79 , so it was excluded for the fit plotted in figure [ fig : cygaspec ] and the fit parameters given in table [ tab : specfits ] .
most of the measurements above 100 ghz in figure [ fig : cygaspec ] are fluxes summed over the core and two hot spots in the radio lobes .
these are probably valid measurements of the total flux , since the contribution of extended emission from the steep - spectrum lobes appears to be small or negligible at these frequencies . at 230 ghz , @xcite found that the integrated flux over the entire source was only about 10% greater than the summed flux of the hot spots and core , which did not amount to a significant detection of emission from the lobes . _
tau a_ the _ wmap _ flux agrees with that of @xcite at 22.29 ghz and with that of @xcite at 33 ghz .
flux measurements from three - year _ wmap _ sky maps by @xcite are consistent with our results in k - v bands , but their w - band measurement is 2.8@xmath98 greater than ours .
the difference is probably due to difference in background subtraction error .
tau a has been recommended as a polarization calibrator for cmb experiments by @xcite , who report a polarized flux density of @xmath130 jy at 90 ghz .
the _ wmap _
value at 93 ghz is @xmath131 jy . for a @xmath132 beam , @xcite predict a position angle of @xmath133 ( statistical ) @xmath134 ( systematic ) in equatorial coordinates .
the _ wmap _
value at 93 ghz is @xmath135 ( statistical ) @xmath136 ( systematic ) . these measurements and previous position angle measurements integrated over the source
are shown as a function of wavelength squared in figure [ fig : polangles ] .
a linear fit to the data gives an intrinsic polarization position angle of @xmath137 and a rotation measure of @xmath138 rad m@xmath139 .
this mean rotation measure appears to be consistent with the rotation measure map obtained by @xcite from vla observations at @xmath140 resolution .
they found a large - scale rotation measure of @xmath141 rad m@xmath139 that they attribute to the interstellar medium , and from the observed depolarization they inferred absolute rotation measure values of a few hundred rad m@xmath139 in unresolved filaments .
@xcite report polarized flux densities they measured from 5-year _ wmap _ data that are smaller than the values in table [ tab : srcflux ] by 9% , 15% , 28% , and 67% at 23 , 33 , 41 , and 61 ghz , respectively . for cas
a , cyg a , and 3c274 , their polarized fluxes agree with those in table [ tab : srcflux ] within the uncertainties . _
3c58_ the _ wmap _
data show the spectral steepening first suggested by the 84 ghz measurement of @xcite and confirmed by the upper limits of @xcite in the infrared .
our adopted form for the spectrum transitions between a low - frequency power law and a high - frequency power law , and gives a good fit to the combined radio to infrared data .
our fit gives a change in spectral index of about 0.9 .
additional steepening is needed to explain the observed x - ray spectrum , and @xcite suggested there is a second spectral break somewhere in the infrared .
they presented model calculations in which the break at @xmath81 ghz is due to a break in the electron spectrum injected into the nebula and the higher frequency break is due to synchrotron losses . _
3c274_ the _ wmap _ k - band flux agrees with that of @xcite at 22.29 ghz .
the combined data are better fit by a quadratic ( @xmath142 ) than by a linear power law ( @xmath143 ) .
curvature in the overall spectrum might be expected since 3c274 has a steep spectrum halo . above 10 ghz ,
the contribution of the halo to the total flux is @xmath144 @xcite . over the _ wmap _ frequency range , the quadratic fit and the _ wmap _ only power - law fit are not significantly different .
the data of @xcite give a flatter 1.4 - 10.6 ghz spectral index than that of @xcite , and they suggested that this may be due to core activity that could cause variability at the 5 - 10% level .
as noted above , variability of about 4% and 17% has been observed at 43 ghz and 90 ghz , respectively .
_ wmap _ data provide well - calibrated radiometry of the outer planets and bright sources .
the seven - year data provide the longest _ wmap _ baseline to date from which to study temporal variability , plus an increase in signal - to - noise which allows us to include fainter objects such as uranus and neptune in the analysis .
jupiter temperatures derived from the seven - year data are within @xmath0 of the previously published five - year values of @xcite .
the disk - integrated temperatures derived for jupiter at _ wmap _ frequencies have uncertainties of less than 1% , with an apparent seasonal stability of @xmath145 . as a planet , jupiter has additional advantages in that it has a small apparent angular diameter and moves with respect to the fixed sky , allowing for well - characterized background subtraction .
of the ten objects studied , _ wmap _
uncertainties for jupiter are lowest , and we continue to recommend it as the best means for transferring the _ wmap _ dipole calibration to another microwave instrument . with the aid of models to predict geometrical and/or intrinsic variations ,
the _ wmap_-derived temperatures for mars and saturn are predictable to @xmath146 .
the recommendation for using the mars values needs to be tempered by the knowledge that these have been derived using mean seasonal properties only .
the saturn model is purely empirical , and can be improved with the addition of the 2009 and 2010 _ wmap _ observations when the rings are nearly edge - on .
uncertainties in the seven - year mean uranus temperatures range from 3% at w band to about 7% in k and ka band ; only the q , v and w measurements are suitable for calibration at the @xmath147 level .
some form of simple model including polar brightness effects would be required to extend the seven - year mean data here beyond the observing epoch , since long - term studies in the literature show clear seasonal temperature variations .
there is also the surprising hint of a dip in the microwave spectrum near 30 ghz , where none is expected based on discussions in the literature .
flux densities in stokes i for the five celestial calibrators have typical uncertainties of 1%-3% , and are in good agreement with previous measurements in the _ wmap _ frequency range . for four of the sources ,
the _ wmap _ data and other available data are consistent with a simple well - defined spectrum over two decades in frequency or more .
the _ wmap _
observations improve the accuracy of the spectral fit parameters and in some cases they extend the frequency range where the object may be used for calibration .
the uncertainties in the spectral fits are typically 1% or less in amplitude and 0.02 or less in spectral index .
we provide new estimates for the secular variation of cas a and tau a , and provide limits and an estimate for year - to - year variability of the other sources .
we present _ wmap _ polarization data with uncertainties of a few percent for tau a. the _ wmap _ mission is made possible by the support of the science mission directorate office at nasa headquarters .
this research has made use of nasa s astrophysics data system bibliographic services .
we acknowledge the use of the healpix package @xcite . ) to the mars model of wright ( 1976 , 2007 ) .
the _ wmap _
observations have been corrected to absolute brightness .
model values ( red line ) have been rescaled by a factor of 0.953 to bring them into overall agreement with the observations ; thick portions of the line indicate observing seasons .
data quality masking can skew the mean times of observations from the mean of the seasonal interval , as is evident in the second observing season . [
fig : marswband],height=432 ] at w band .
black diamonds are the _ wmap _ data ; error bars are obtained from the variation over the seven observing seasons within each frequency band . the same averaging procedure is carried out using the wright model predictions for the mean time of observation ( red line ) : small time - sampling and spectral variations between the frequencies cause a slightly jagged appearance .
the _ wmap _
data show a weak frequency dependence which can be approximated as ( @xmath66(ghz)/94)@xmath148 ( blue line ) .
this can be interpreted as a frequency - dependent emissivity .
[ fig : marsfreqdep],height=480 ] is the ring opening angle at mid - season .
the contours illustrate the components used in an empirical model consisting of the planetary disk and seven ring regions , each of a different fixed normal optical depth . only the a , b , c and cassini division rings are represented , and the radial divisions follow those used by @xcite .
[ fig : satincl],width=672 ] are shown in black for each _ wmap _ frequency band . where there are multiple differencing assemblies per frequency , multiple points are plotted at each inclination . an empirical model including both ring and disk components ( see the text ) is overplotted in red .
the temperature of the planetary disk predicted by the model occurs at @xmath72=0@xmath149 , when the rings are viewed edge - on .
the model is symmetric about @xmath72=0@xmath149 .
right : residuals ( data - model ) .
the model predicts the observed brightness temperature within @xmath1503% .
[ fig : satmod],height=624 ] in a side , and the vertical direction is aligned with galactic north .
sources are labeled along the top , and frequency bands are labeled along the left side .
bright pixels are displayed as black .
left : stokes @xmath153 logarithmically scaled ; peak pixel values in k band are 60.8 , 19.3 , 95.8 , 5.4 , and 7.2 mk thermodynamic , respectively , for the five sources , at a native resolution of healpix @xmath154 .
right : polarization @xmath155 , where @xmath156 and @xmath157 are stokes parameters , smoothed with a gaussian of @xmath158 .
scaling is linear between 0 and 0.25 mk thermodynamic , except for tau a , where scaling is linear between 0 and 1 mk thermodynamic .
green arrow : north in the equatorial coordinate system .
red bars : polarization direction in a square pixel @xmath159 wide , shown only where the s / n of @xmath160 is greater than @xmath42 . the broken ring around cas
a in k band is spurious and results from beam ellipticity coupled with effective frequency differences between the two orthogonally polarized k - band radiometers .
[ fig : ipforcelobj],width=624 ] .
the black line is the radial profile of an azimuthally symmetric polarization artifact centered on the source , resulting from a combination of beam , detector and source properties .
the simulation approximately reproduces the observed ring structure surrounding cas a in figure [ fig : ipforcelobj ] , but is brighter by @xmath161 , and does not include details such as the asymmetric scan - angle distribution in the data .
[ fig : artifact],width=624 ] are shown in red , with error bars that include a systematic uncertainty of @xmath162 .
results shown in black are from references in @xcite and from @xcite at 90 ghz for a 10 arcmin beam .
the linear fit gives an intrinsic polarization position angle of @xmath137 and a rotation measure of @xmath138 rad m@xmath139 .
[ fig : polangles],width=624 ] jupiter & 1 & 2001 oct 8 & 2001 nov 22 & + & 2 & 2002 feb 11 & 2002 mar 27 & g + & 3 & 2002 nov 10 & 2002 dec 24 & + & 4 & 2003 mar 15 & 2003 apr 29 & g + & 5 & 2003 dec 11 & 2004 jan 23 & + & 6 & 2004 apr 15 & 2004 may 30 & + & 7 & 2005 jan 9 & 2005 feb 21 & + & 8 & 2005 may 16 & 2005 jul 1 & + & 9 & 2006 feb 7 & 2006 mar 24 & + & 10 & 2006 jun 16 & 2006 aug 2 & + & 11 & 2007 mar 10 & 2007 apr 24 & + & 12 & 2007 jul 19 & 2007 sep 3 & + & 13 & 2008 apr 11 & 2008 may 27 & + saturn & 1 & 2001 sep 7 & 2001 oct 22 & + & 2 & 2002 jan 14 & 2002 feb 26 & + & 3 & 2002 sep 22 & 2002 nov 5 & + & 4 & 2003 jan 28 & 2003 mar 13 & g + & 5 & 2003 oct 7 & 2003 nov 20 & + & 6 & 2004 feb 12 & 2004 mar 26 & g + & 7 & 2004 oct 20 & 2004 dec 3 & + & 8 & 2005 feb 25 & 2005 apr 10 & + & 9 & 2005 nov 4 & 2005 dec 17 & + & 10 & 2006 mar 11 & 2006 apr 24 & + & 11 & 2006 nov 18 & 2006 dec 31 & + & 12 & 2007 mar 25 & 2007 may 9 & + & 13 & 2007 dec 1 & 2008 jan 13 & + & 14 & 2008 apr 7 & 2008 may 22 & + mars & 1 & 2001 aug 10 & 2001 oct 14 & + & 2 & 2003 apr 19 & 2003 jul 17 & u , n + & 3 & 2003 oct 10 & 2003 dec 30 & u + & 4 & 2005 jul 13 & 2005 oct 1 & + & 5 & 2005 dec 16 & 2006 feb 19 & + & 6 & 2007 sep 18 & 2007 nov 20 & + & 7 & 2008 jan 30 & 2008 mar 30 & + uranus & 1 & 2001 sep 30 & 2001 nov 13 & + & 2 & 2002 may 20 & 2002 jul 5 & + & 3 & 2002 oct 4 & 2002 nov 17 & + & 4 & 2003 may 25 & 2003 jul 9 & m + & 5 & 2003 oct 8 & 2003 nov 21 & m + & 6 & 2004 may 28 & 2004 jul 13 & + & 7 & 2004 oct 12 & 2004 nov 24 & + & 8 & 2005 jun 2 & 2005 jul 17 & + & 9 & 2005 oct 16 & 2005 nov 28 & + & 10 & 2006 jun 6 & 2006 jul 22 & + & 11 & 2006 oct 20 & 2006 dec 3 & + & 12 & 2007 jun 10 & 2007 jul 26 & + & 13 & 2007 oct 24 & 2007 dec 7 & + & 14 & 2008 jun 14 & 2008 jul 29 & + neptune & 1 & 2001 sep 15 & 2001 oct 29 & + & 2 & 2002 may 2 & 2002 jun 16 & + & 3 & 2002 sep 17 & 2002 oct 31 & + & 4 & 2003 may 4 & 2003 jun 19 & m + & 5 & 2003 sep 20 & 2003 nov 3 & + & 6 & 2004 may 6 & 2004 jun 21 & + & 7 & 2004 sep 21 & 2004 nov 4 & + & 8 & 2005 may 8 & 2005 jun 23 & + & 9 & 2005 sep 23 & 2005 nov 7 & + & 10 & 2006 may 11 & 2006 jun 26 & + & 11 & 2006 sep 26 & 2006 nov 9 & + & 12 & 2007 may 14 & 2007 jun 28 & + & 13 & 2007 sep 28 & 2007 nov 11 & + & 14 & 2008 may 15 & 2008 jun 30 & + ccccc k1 & 22.85 & 13.1 & 136.2 & 0.85 + ka1 & 33.11 & 9.1 & 147.2 & 0.77 + q1 & 40.92 & 7.3 & 154.4 & 0.91 + q2 & 40.71 & 7.4 & 155.2 & 0.87 + v1 & 60.41 & 5.0 & 165.0 & 0.78 + v2 & 61.28 & 4.9 & 166.1 & 0.74 + w1 & 93.25 & 3.2 & 173.0 & 0.84 + w2 & 93.73 & 3.2 & 173.3 & 0.91 + w3 & 93.72 & 3.2 & 173.7 & 0.94 + w4 & 93.57 & 3.2 & 173.9 & 0.92 + k & 22.85 & 13.1 & 136.2 & 0.85 + ka & 33.11 & 9.1 & 147.2 & 0.77 + q & 40.82 & 7.3 & 154.8 & 0.67 + v & 60.85 & 4.9 & 165.6 & 0.59 + w & 93.32 & 3.2 & 173.5 & 0.54 + cccrcccc 1 & 2001 oct 08 & 2001 nov 22 & @xmath168 & 0.28 & 0 & 194 & 13 + 2 & 2002 feb 11 & 2002 mar 27 & & & 95 & 189 & 5 + 3 & 2002 nov 10 & 2002 dec 24 & @xmath169 & 0.28 & 0 & 214 & 40 + 4 & 2003 mar 15 & 2003 apr 29 & @xmath170 & 0.48 & 60 & 206 & 32 + 5 & 2003 dec 11 & 2004 jan 23 & @xmath171 & 0.22 & 0 & 254 & 59 + 6 & 2004 apr 15 & 2004 may 30 & @xmath172 & 0.32 & 0 & 237 & 55 + 7 & 2005 jan 09 & 2005 feb 21 & @xmath173 & 0.27 & 12 & 311 & 56 + 8 & 2005 may 16 & 2005 jul 01 & @xmath174 & 0.39 & 0 & 294 & 60 + 9 & 2006 feb 07 & 2006 mar 24 & @xmath175 & 0.29 & 18 & 344 & 35 + 10 & 2006 jun 16 & 2006 aug 02 & @xmath176 & 0.45 & 0 & 335 & 43 + 11 & 2007 mar 10 & 2007 apr 24 & @xmath177 & 0.31 & 0 & 2 & 9 + 12 & 2007 jul 19 & 2007 sep 03 & @xmath178 & 0.49 & 42 & 356 & 17 + 13 & 2008 apr 11 & 2008 may 27 & @xmath179 & 0.37 & 0 & 17 & -19 + cccccccc k1 & -0.306 & 0.601 & 0.658 & 0.428 & 0.726 & 0.464 & 1.62 + ka1 & -0.181 & 0.408 & -0.191 & 0.293 & 0.263 & 0.352 & 0.79 + q1 & -0.168 & 0.430 & -0.244 & 0.304 & 0.296 & 0.349 & 0.90 + q2 & -0.151 & 0.428 & -0.062 & 0.297 & 0.163 & 0.411 & 0.41 + v1 & -0.149 & 0.358 & 0.462 & 0.252 & 0.486 & 0.263 & 1.88 + v2 & 0.143 & 0.305 & 0.067 & 0.223 & 0.158 & 0.292 & 0.56 + w1 & 0.040 & 0.319 & 0.393 & 0.233 & 0.395 & 0.234 & 1.69 + w2 & 0.156 & 0.380 & 0.188 & 0.264 & 0.245 & 0.317 & 0.82
+ w3 & 0.006 & 0.397 & -0.377 & 0.277 & 0.377 & 0.277 & 1.36 + w4 & -0.058 & 0.371 & 0.311 & 0.273 & 0.316 & 0.277 & 1.15
+ ccccccc 1 & @xmath180 & @xmath181 & @xmath182 & @xmath183 & @xmath184 & @xmath185 + 2 & @xmath186 & @xmath187 & @xmath188 & @xmath189 & @xmath190 & @xmath191 + 3 & @xmath192 & @xmath189 & @xmath193 & @xmath194 & @xmath195 & @xmath196 + 4 & @xmath197 & @xmath198 & @xmath199 & @xmath200 & @xmath201 & @xmath202 + 5 & @xmath203 & @xmath204 & @xmath205 & @xmath206 & @xmath207 & @xmath208 + 6 & @xmath209 & @xmath183 & @xmath188 & @xmath210 & @xmath190 & @xmath211 + 7 & @xmath212 & @xmath213 & @xmath214 & @xmath215 & @xmath216 & @xmath217 + crcc @xmath218 & 148594 & 3.1 & 0.89 + @xmath219 & 186136 & 14.6 & 0.92 + @xmath220 & 50606 & 10.0 & 0.66 + @xmath221 & 24477 & 12.8 & 0.73 + @xmath222 & 18991 & 10.1 & 1.02 + @xmath223 & 7217 & 10.1 & 1.05 + 1 & 2175.58 & -26 & @xmath238 & @xmath239 & @xmath240 & @xmath241 & @xmath242 & @xmath243 & @xmath244 & @xmath244 & @xmath245 & @xmath246 + 2 & 2305.74 & -25 & @xmath247 & @xmath248 & @xmath240 & @xmath241 & @xmath249 & @xmath243 & @xmath250 & @xmath251 & @xmath244 & @xmath251 + 3 & 2554.16 & -26 & @xmath252 & @xmath239 & @xmath253 & @xmath254 & @xmath255 & @xmath242 & @xmath245 & @xmath245 & @xmath244 & @xmath245 + 5 & 2931.78 & -24 & @xmath256 & @xmath257 & @xmath258 & @xmath240 & @xmath259 & @xmath259 & @xmath250 & @xmath250 & @xmath260 & @xmath250 + 7 & 3310.06 & -21 & @xmath261 & @xmath262 & @xmath263 & @xmath263 & @xmath264 & @xmath241 & @xmath265 & @xmath265 & @xmath265 & @xmath265 + 8 & 3441.94 & -23 & @xmath266 & @xmath257 & @xmath267 & @xmath239 & @xmath253 & @xmath254 & @xmath249 & @xmath268 & @xmath268 & @xmath255 + 9 & 3690.17 & -17 & @xmath269 & @xmath270 & @xmath262 & @xmath271 & @xmath239 & @xmath239 & @xmath240 & @xmath240 & @xmath264 & @xmath264 + 10 & 3813.21 & -20 & @xmath272 & @xmath256 & @xmath273 & @xmath274 & @xmath275 & @xmath267 & @xmath254 & @xmath276 & @xmath277 & @xmath278 + 11 & 4068.06 & -12 & @xmath279 & @xmath270 & @xmath280 & @xmath281 & @xmath282 & @xmath263 & @xmath267 & @xmath258 & @xmath248 & @xmath258 + 12 & 4196.39 & -15 & @xmath283 & @xmath252 & @xmath280 & @xmath284 & @xmath282 & @xmath239 & @xmath248 & @xmath248 & @xmath248 & @xmath285 + 13 & 4443.26 & -6 & @xmath286 & @xmath256 & @xmath262 & @xmath257 & @xmath263 & @xmath282 & @xmath248 & @xmath240 & @xmath267 & @xmath240 + 14 & 4577.96 & -9 & @xmath287 & @xmath270 & @xmath280 & @xmath238 & @xmath263 & @xmath239 & @xmath263 & @xmath239 & @xmath239 & @xmath239 + cccc a & @xmath288 & @xmath289 & @xmath290 + cassini division & @xmath291 & @xmath288 & @xmath292 + outer b & @xmath293 & @xmath291 & @xmath294 + inner b & @xmath295 & @xmath293 & @xmath296 + outer c & @xmath297 & @xmath295 & @xmath292 + middle c & @xmath298 & @xmath297 & @xmath299 + inner c & @xmath300 & @xmath298 & @xmath301 + ccccccc k & 133.8 & 1.5 & 4.0 & 6.7 & 1.1 & 2.1 + ka & 138.8 & 1.2 & 3.7 & 9.6 & 0.8 & 1.8 + q & 142.1 & 0.9 & 3.4 & 11.3 & 0.6 & 1.6 + v & 146.1 & 0.7 & 3.2 & 14.5 & 0.5 & 1.5 + w & 146.0 & 0.6 & 3.1 & 19.8 & 0.4 & 1.4 + cccccccc 1 & 2204 & -26.4 & @xmath302 & @xmath303 & @xmath304 & @xmath305 & @xmath306 + 2 & 2437 & -19.0 & @xmath307 & @xmath308 & @xmath309 & @xmath310 & @xmath311 + 3 & 2573 & -22.4 & @xmath312 & @xmath313 & @xmath314 & @xmath315 & @xmath316 + 4 & 2807 & -15.2 & @xmath317 & @xmath318 & @xmath319 & @xmath320 & @xmath321 + 5 & 2942 & -18.5 & @xmath322 & @xmath323 & @xmath324 & @xmath325 & @xmath326 + 6 & 3176 & -11.2 & @xmath307 & @xmath327 & @xmath328 & @xmath329 & @xmath330 + 7 & 3312 & -14.5 & @xmath331 & @xmath332 & @xmath333 & @xmath334 & @xmath335 + 8 & 3546 & -7.3 & @xmath336 & @xmath337 & @xmath338 & @xmath339 & @xmath340 + 9 & 3681 & -10.7 & @xmath341 & @xmath342 & @xmath343 & @xmath344 & @xmath345 + 10 & 3915 & -3.4 & @xmath346 & @xmath347 & @xmath348 & @xmath349 & @xmath350 + 11 & 4050 & -6.7 & @xmath351 & @xmath352 & @xmath353 & @xmath354 & @xmath355 + 12 & 4284 & 0.5 & @xmath356 & @xmath357 & @xmath358 & @xmath359 & @xmath360 + 13 & 4419 & -2.8 & @xmath361 & @xmath362 & @xmath363 & @xmath364 & @xmath365 + 14 & 4654 & 4.4 & @xmath366 & @xmath367 & @xmath368 & @xmath369 & @xmath370 + mean & & & @xmath371 & @xmath372 & @xmath373 & @xmath374 & @xmath375 + cccccccc 1 & 2185 & -27.4 & @xmath376 & @xmath377 & @xmath378 & @xmath379 & @xmath380 + 2 & 2420 & -27.9 & @xmath381 & @xmath382 & @xmath383 & @xmath384 & @xmath385 + 3 & 2555 & -27.6 & @xmath386 & @xmath387 & @xmath388 & @xmath389 & @xmath390 + 4 & 2799 & -27.9 & @xmath391 & @xmath392 & @xmath393 & @xmath394 & @xmath395 + 5 & 2923 & -27.6 & @xmath396 & @xmath397 & @xmath398 & @xmath399 & @xmath400 + 6 & 3155 & -28.0 & @xmath401 & @xmath402 & @xmath403 & @xmath404 & @xmath405 + 7 & 3290 & -27.7 & @xmath406 & @xmath407 & @xmath408 & @xmath409 & @xmath410 + 8 & 3522 & -28.0 & @xmath411 & @xmath412 & @xmath413 & @xmath414 & @xmath415 + 9 & 3657 & -27.9 & @xmath416 & @xmath417 & @xmath418 & @xmath419 & @xmath420 + 10 & 3890 & -27.9 & @xmath421 & @xmath422 & @xmath423 & @xmath424 & @xmath425 + 11 & 4025 & -27.9 & @xmath426 & @xmath427 & @xmath428 & @xmath429 & @xmath430 + 12 & 4257 & -27.7 & @xmath431 & @xmath432 & @xmath433 & @xmath434 & @xmath435 + 13 & 4391 & -27.6 & @xmath436 & @xmath437 & @xmath438 & @xmath439 & @xmath440 + 14 & 4625 & -27.6 & @xmath441 & @xmath442 & @xmath443 & @xmath444 & @xmath445 + mean & & & @xmath446 & @xmath447 & @xmath448 & @xmath449 & @xmath450 + lrrrrlcl cas a & 23 23 24 & + 58 48 54 & 111.74 & -2.13 & shell type snr & @xmath451 & @xcite + cyg a & 19 59 28 & + 40 44 02 & 76.19 & 5.76 & radio galaxy & @xmath452 & @xcite + tau a ( crab ) & 05 34 32 & + 22 00 52 & 184.56 & -5.78 & filled - center snr & @xmath453 & @xcite + 3c58 & 02 05 38 & + 64 49 42 & 130.72 & 3.08 & filled - center snr & @xmath454 & @xcite + 3c274 ( m87 , vir a ) & 12 30 49 & + 12 23 28 & 283.78 & 74.49 & radio galaxy & @xmath455 & @xcite + lccccccccc cas a & 22.68 & 0.2497 & 232.1@xmath456 2.9 & [email protected] & [email protected] & 0.93 @xmath456 0.09 & [email protected] & 85.5@xmath456 2.7 ( 66.0 ) & + cas a & 32.94 & 0.2041 & 179.8@xmath456 1.5 & [email protected] & [email protected] & 0.59 @xmath456 0.08 & [email protected] & 80.6@xmath456 4.2 ( 61.1 ) & + cas a & 40.62 & 0.2162 & 153.7@xmath456 1.5 & [email protected] & [email protected] & 0.53 @xmath456 0.09 & [email protected] & 80.6@xmath456 5.3 ( 61.1 ) & + cas a & 60.48 & 0.2084 & 116.3@xmath456 2.0@xmath457 & [email protected]@xmath457 & [email protected]@xmath457 & 0.28 @xmath456 0.16 & [email protected] & [email protected] ( 82.9 ) & + cas a & 92.90 & 0.1800 & 84.0@xmath456 3.4@xmath458 & [email protected]@xmath458 & [email protected]@xmath458 & 0.13 @xmath456 0.28 & [email protected] & [email protected] ( 104.4 ) & + & & & & & & & & + cyg a & 22.65 & 0.2493 & 57.7@xmath456 1.3 & [email protected] & [email protected] & 0.48 @xmath456 0.06 & [email protected] & 43.6@xmath456 2.9 ( 164.8 ) & + cyg a & 32.91 & 0.2039 & 38.2@xmath456 0.6 & [email protected] & [email protected] & 0.45 @xmath456 0.08 & [email protected] & -13.5@xmath456 6.0 ( 107.7 ) & + cyg
a & 40.58 & 0.2159 & 29.8@xmath456 0.4 & [email protected] & [email protected] & 0.61 @xmath456 0.09 & [email protected] & -8.4@xmath456 4.8 ( 112.8 ) & + cyg a & 60.42 & 0.2083 & 18.6@xmath456 0.2 & [email protected] & [email protected] & 0.46 @xmath456 0.17 & [email protected] & [email protected] ( 104.3 ) & + cyg a & 92.83 & 0.1802 & 11.1@xmath456 0.3 & [email protected] & [email protected] & 0.28 @xmath456 0.27 & [email protected] & [email protected] ( 109.7 ) & + & & & & & & & & + tau a & 22.70 & 0.2501 & 383.8@xmath456 9.6 & [email protected] & [email protected] & 27.17 @xmath456 0.68 & [email protected] & -88.5@xmath456 0.1 ( 149.1 ) & + tau a & 32.96 & 0.2043 & 342.8@xmath456 6.4 & [email protected] & [email protected] & 23.80 @xmath456 0.44 & [email protected] & -87.7@xmath456 0.1 ( 149.9 ) & + tau a & 40.64 & 0.2163 & 317.7@xmath456 8.6 & [email protected] & [email protected] & 22.12 @xmath456 0.60 & [email protected] & -87.3@xmath456 0.2 ( 150.3 ) & + tau a & 60.53 & 0.2085 & 276.0@xmath456 5.2@xmath457 & [email protected]@xmath457 & [email protected]@xmath457 & 19.31 @xmath456 0.36 & [email protected] & -87.7@xmath456 0.4 ( 149.9 ) & + tau a & 92.95 & 0.1798 & 232.8@xmath456 9.7@xmath458 & [email protected]@xmath458 & [email protected]@xmath458 & 16.60 @xmath456 0.73 & [email protected] & -88.7@xmath456 0.7 ( 148.9 ) & + & & & & & & & & + 3c58 & 22.70 & 0.2500 & 20.8@xmath456 0.4 & [email protected] & [email protected] & 1.20 @xmath456 0.05 & [email protected] & -11.0@xmath456 1.4 ( 5.4 ) & + 3c58 & 32.96 & 0.2043 & 19.0@xmath456 0.3 & [email protected] & [email protected] & 1.23 @xmath456 0.07 & [email protected] & -5.8@xmath456 2.3 ( 10.7 ) & + 3c58 & 40.64 & 0.2163 & 17.1@xmath456 0.3 & [email protected] & [email protected] & 1.09 @xmath456 0.08 & [email protected] & -12.6@xmath456 2.3 ( 3.9 ) & + 3c58 & 60.52 & 0.2085 & 14.1@xmath456 0.3@xmath457 & [email protected]@xmath457 & [email protected]@xmath457 & 0.77 @xmath456 0.12 & [email protected] & -2.7@xmath456 6.3 ( 13.7 ) & + 3c58 & 92.94 & 0.1798 & 11.4@xmath456 0.6@xmath458 & [email protected]@xmath458 & [email protected]@xmath458 & 0.67 @xmath456 0.26 & [email protected] & [email protected] ( -2.9 ) & + & & & & & & & & + 3c274 & 22.68 & 0.2497 & 21.6@xmath456 0.6 & [email protected] & [email protected] & 0.88 @xmath456 0.06 & [email protected] & 28.1@xmath456 2.3 ( 45.5 ) & + 3c274 & 32.94 & 0.2041 & 16.8@xmath456 0.4 & [email protected] & [email protected] & 0.81 @xmath456 0.10 & [email protected] & 28.9@xmath456 4.0 ( 46.3 ) & + 3c274 & 40.61 & 0.2161 & 14.3@xmath456 0.5 & [email protected] & [email protected] & 0.68 @xmath456 0.12 & [email protected] & 26.9@xmath456 5.5 ( 44.3 ) & + 3c274 & 60.48 & 0.2084 & 10.6@xmath456 0.3 & [email protected] & [email protected] & 0.20 @xmath456 0.23 & [email protected] & [email protected] ( 34.6 ) & + 3c274 & 92.89 & 0.1800 & 7.8@xmath456 0.3 & [email protected] & [email protected] & 0.59 @xmath456 0.37 & [email protected] & [email protected] ( 40.4 ) & + crrrrr cas a & @xmath459 & @xmath460 & @xmath461 & @xmath462 & @xmath463 + cyg a & @xmath464 & @xmath465 & @xmath466 & @xmath467 & @xmath468 + tau a & @xmath469 & @xmath470 & @xmath471 & @xmath472 & @xmath473 + 3c58 & @xmath474 & @xmath475 & @xmath476 & @xmath477 & @xmath478 + 3c274 & @xmath479 & @xmath480 & @xmath481 & @xmath482 & @xmath483 + lccccccccc cas a & wmap only & 22 - 93 & power law & [email protected] & [email protected] & & + cas a & combined@xmath484 & 1.4 - 250 & curved & [email protected] & [email protected] & 0.038 @xmath456 0.008 & + & & & & & & & + cyg a & wmap only & 22 - 93 & power law & [email protected] & [email protected] & & + cyg a & combined & 2 - 94 & power law & [email protected] & [email protected] & & + & & & & & & & + tau a & wmap only & 22 - 93 & power law & [email protected] & [email protected] & & + tau a & combined@xmath485 & 1 - 353 & power law & [email protected] & [email protected] & & + & & & & & & & + 3c58 & combined & 0.04 - 83000 & rolloff & [email protected] & [email protected] & [email protected] & [email protected] & + & & & & & & & + 3c274 & wmap only & 22 - 93 & power law & [email protected] & [email protected] & & + 3c274 & combined & 0.4 - 93 & curved & [email protected] & [email protected] & 0.050 @xmath456 0.008 & + | we present _ wmap _ seven - year observations of bright sources which are often used as calibrators at microwave frequencies .
ten objects are studied in five frequency bands ( 23 - 94 ghz ) : the outer planets ( mars , jupiter , saturn , uranus and neptune ) and five fixed celestial sources ( cas a , tau a , cyg a , 3c274 and 3c58 ) .
the seven - year analysis of jupiter provides temperatures which are within @xmath0 of the previously published _ wmap _ five - year values , with slightly tighter constraints on variability with orbital phase ( @xmath1 ) , and limits ( but no detections ) on linear polarization .
observed temperatures for both mars and saturn vary significantly with viewing geometry .
scaling factors are provided which , when multiplied by the wright mars thermal model predictions at 350 , reproduce _ wmap _ seasonally averaged observations of mars within @xmath2 .
an empirical model is described which fits brightness variations of saturn due to geometrical effects and can be used to predict the _ wmap _
observations to within 3% .
seven - year mean temperatures for uranus and neptune are also tabulated .
uncertainties in uranus temperatures are 3%-4% in the 41 , 61 and 94 ghz bands ; the smallest uncertainty for neptune is 8% for the 94 ghz band .
intriguingly , the spectrum of uranus appears to show a dip at @xmath3 ghz of unidentified origin , although the feature is not of high statistical significance .
flux densities for the five selected fixed celestial sources are derived from the seven - year _ wmap _ sky maps , and are tabulated for stokes i , q and u , along with polarization fraction and position angle .
fractional uncertainties for the stokes i fluxes are typically 1% to 3% .
source variability over the seven - year baseline is also estimated .
significant secular decrease is seen for cas a and tau a : our results are consistent with a frequency independent decrease of about 0.53% per year for cas a and 0.22% per year for tau a. we present _ wmap _ polarization data with uncertainties of a few percent for tau a. where appropriate , _ wmap _ results are compared against previous findings in the literature . with an absolute calibration uncertainty of 0.2% , _ wmap _
data are a valuable asset for calibration work . |
in a series of instructive papers , froeschl et al .
( 2000 , 2005 ) , lega et al .
( 2003 ) , and guzzo et al .
( 2005 ) presented the results of a detailed numerical investigation of the phenomenon of _ arnold diffusion _ ( arnold 1964 ) in a hamiltonian system of three degrees of freedom that satisfies sufficient conditions for the holding of the _ nekhoroshev theorem _
( nekhoroshev 1977 , benettin et al .
1985 , lochak 1992 , pshel 1993 ) .
the aim of these studies was to establish quantitative estimates as regards a ) the critical value of the small parameter @xmath5 below which the system enters into the so - called ` nekhoroshev regime ' , and b ) the dependence of the local diffusion coefficient @xmath6 , along a particular resonance , on @xmath1 . in guzzo
( 2005 ) the authors reported to have also numerically observed global diffusion over an extended domain of the arnold web .
other numerical studies over the years related to the same problem are : kaneko and konishi ( 1989 ) , wood et al .
( 1990 ) , dumas and laskar ( 1993 ) , laskar ( 1993 ) , skokos et al .
( 1997 ) , giordano and cincotta ( 2004 ) . some early results of the literature are briefly commented in the discussion . in what follows we focus on an analysis of point ( b ) above . according to lega et al .
( 2003 ) , a tedious numerical calculation yields that the local diffusion coefficient @xmath6 along a resonance depends monotonically on @xmath1 .
while the limited range of the numerical data did not allow for a precise fitting , the authors presented evidence that the numerical function @xmath7 was decreasing faster than a power - law @xmath8 , while it was consistent with an exponential law @xmath9 for some unspecified constant @xmath10 . a subsequent study in a mapping model analogous to the above hamiltonian model yielded the concrete estimate @xmath11 ( froeschl et al .
2005 ) . in the present paper
our aim is to provide a deeper understanding of these numerical results by having recourse , precisely , to the analytical apparatus of the nekhoroshev theory , that is , the construction of a _ resonant birkhoff normal form _ and of estimates on the size of the remainder @xmath12 of the normal form at the optimal order of normalization .
the optimal remainder function @xmath13 is a crucial quantity for the dynamics because it represents the true size of the perturbation of the system that corresponds to its deviation from an integrable system .
indeed , in nekhoroshev theory all bounds on the variations of the actions follow from estimates on the size of @xmath13 . in order , therefore , to establish the link between the nekhoroshev theory , on the one hand , and the problem of the speed of diffusion , on the other hand , in the present paper we seek to determine : \a ) the dependence of the local diffusion coefficient @xmath6 on the size of the remainder @xmath0 at the optimal order of normalization , and \b ) the precise dependence of @xmath0 on @xmath1 . in particular , we seek to calculate @xmath0 when @xmath1 becomes small enough so that the exponentially small scaling of @xmath0 with @xmath1 clearly shows up .
this required to consider values of @xmath1 one order of magnitude smaller than the values considered in the numerical experiments of lega et al .
( 2003 ) .
we can immediately state a brief summary of our main results .
we find that : \a ) @xmath6 scales with @xmath0 as a power law @xmath14 , with @xmath15 . to determine this relation , use was made of the data of lega et al .
( 2003 ) for the numerical values of @xmath6 , and a program was written by the author in order to compute the birkhoff normal form and thereby the remainder @xmath13 .
\b ) in the range @xmath16 , @xmath0 scales as @xmath17 with @xmath18 and @xmath19 .
this also yields the exponential scaling @xmath20 .
note that the exponentially small scaling shows up clearly only for @xmath21 , while in the range @xmath22 a power - law fitting @xmath23 yields marginally better results than the exponential fitting .
in fact , we find that although the onset of the ` nekhoroshev regime ' can be placed at a threshold value @xmath24 , the deviation of the remainder from a power law becomes clear only for @xmath1 one order of magnitude smaller than this value .
the computational algorithm used to obtain the above results is described in section 2 .
it follows essentially the normalization scheme described in the lecture notes on exponential stability of giorgilli ( 2002 , pp.86 - 87 ) .
this scheme is different from the traditional scheme based on normalization by successive orders of @xmath1 .
we further modify it to avoid using @xmath25truncated resonant modules ( see section 2 ) .
people working on mathematical aspects of the nekhoroshev theory are definitely familiar with all related notions .
however , a somewhat extended description is given in section 2 , to help rendering the subject more accessible also to people oriented towards the applications .
section 3 then focuses on the implementation of the algorithm in the particular model of froeschl et al .
( 2000 ) . by constructing the normal form and identifying the optimal order of normalization
, the size of @xmath13 is evaluated for various values of @xmath1 .
this yields the exponential scaling of @xmath0 on @xmath1 . using also the data of lega et al .
( 2003 ) yields the power - law relation of @xmath6 with @xmath0 .
section 4 contains a brief discussion of the results .
the analytical part of the nekhoroshev theorem is based on the construction of a normal form for a @xmath26 degrees of freedom hamiltonian system of the form @xmath27 where @xmath28 are action - angle variables , @xmath29 satisfies appropriate non - degeneracy and convexity conditions , and @xmath30 is analytic in a complexified domain of the actions and the angles .
we shall be concerned with a _ local _ construction of a normal form for the hamiltonian ( [ ham ] ) , valid within some ( small ) preselected open domain @xmath31 of the action space ( the index @xmath10 means action space ) .
the analyticity condition implies that if @xmath32 is fourier expanded @xmath33 there are positive constants @xmath34 such that for all values of @xmath35 the coefficients @xmath36 are bounded by @xmath37 , where @xmath38 .
furthermore , bounds on the function @xmath32 can be found in domains @xmath39 , for any positive @xmath40 .
now , the purpose of the normal form construction is to perform a series of canonical transformations @xmath41 , such that , after @xmath42 normalization steps , the hamiltonian is expressed as @xmath43 in the last expression , the first term @xmath44 , called the _ normal form _ ,
is constructed in such a way as to yield a simple dynamics in the variables @xmath45 ( e.g. in the non - resonant or simply resonant case the hamiltonian @xmath44 is integrable ) . on the other hand ,
the second term @xmath46 , called the _ remainder _ , represents a perturbation to the dynamics induced by the normal form .
the goal of the theory is to proceed with the normalization as far as necessary in order to minimize the size of the remainder . in exceptional cases of _ integrable _ hamiltonian systems one may proceed by infinitely many steps , within some domain of convergence , and
show that the remainder tends to zero in the limit @xmath47 . in the generic case of non - integrable systems , however , one can only reduce @xmath46 to a finite minimum size .
this is found at a specific order of normalization which is hereafter called the optimal order of normalization @xmath48 .
the size of the remainder @xmath49 for any other order @xmath50 is bigger than the size of @xmath51 .
standard theory yields estimates @xmath52 and @xmath53 for some positive integers @xmath54 depending on the number of degrees of freedom . the classical construction leading to expressions of the form ( [ hamnorm ] ) is nowadays analyzed in many books ( see e.g. boccaletti and pucacco ( 1996 ) , contopoulos ( 2002 ) , morbidelli ( 2002 ) , or ferraz - mello ( 2007 ) for instructive introductions ) .
the main steps can be summarized as follows : \a ) one makes a choice as regards which fourier terms @xmath55 in @xmath32 will be retained and which terms will be eliminated in the normalized hamiltonian .
the terms to be retained are specified by their integer vector @xmath56 belonging to a particular subset of @xmath57 called _ resonant module _ ( denoted hereafter by @xmath58 ) ) .
the choice of resonant module is dictated by the location of the domain @xmath31 in the action space , i.e. , by whether this domain is close to or far from particular resonances ( see subsection 2.3 ) .
\b ) the fourier terms not belonging to @xmath59 are the ones to be eliminated by a canonical transformation .
let us consider the first normalization step .
we shall consider canonical transformations of old variables @xmath60 to new variables @xmath61 obtained via a lie generating function @xmath62 .
the function @xmath62 will be selected so as to be of the ` first order of smallness ' ( the precise meaning of this sentence is analyzed below ) .
the canonical transformation is defined by @xmath63 where @xmath64 is the poisson bracket operator , and @xmath65 is the formal exponential development of @xmath66 .
a well known property of lie transformations is that for any function of the canonical variables @xmath67 one has @xmath68 as a result , the transformed hamiltonian , after the action of the generating function @xmath69 is given by @xmath70 in order to specify the function @xmath62 , we select from @xmath32 all the terms satisfying the following two conditions : a ) being of the ` first order of smallness ' , and b ) not belonging to @xmath58 . denoting by @xmath71 the sum of these terms , the terms are eliminated by selecting @xmath69 so as to satisfy : @xmath72 equation ( [ homo1 ] ) is called homological .
it is reduced to a diagonal system of linear algebraic equations .
namely , writing @xmath71 as @xmath73 and setting @xmath74 we find , through ( [ homo1 ] ) , the solution for the coefficients @xmath75 given by : @xmath76 where @xmath77 is the frequency vector of the unperturbed hamiltonian at the point @xmath78 of the action space .
\c ) in the same way we eliminate the terms of the second order of smallness in @xmath79 or , in general , the terms of the r - th order of smallness in @xmath80 . the recursive formulae for the r - th canonical transformation @xmath81
, @xmath82 are : @xmath83 following iteratively the above procedure , the function @xmath84 takes the form ( [ hamnorm ] ) .
the remainder consists of all the terms of @xmath84 of ` order of smallness ' larger than @xmath42 .
it can be written as @xmath85 this series is an analytic function in a restriction of the domain of analyticity of the original hamiltonian provided that the lie transformation in ( [ nfrec ] ) is convergent at every step ( which is ensured for ` small enough ' generating functions @xmath86 ) .
a clarification of the meaning of ` order of smallness ' is now in order .
when one implements the recursive scheme of eqs.([lietran ] ) and ( [ nfrec ] ) , one needs to decide how should @xmath84 at every step be split into terms of different orders .
such a splitting should be based on the size of each term relative to the size of the other terms . in a computer program ,
it is customary to introduce a so - called ` book - keeping ' parameter @xmath87 .
this is a parameter the powers of which appear as coefficients in front of the various terms in the expansion . a term with coefficient @xmath88
is said to be ` of order of smallness ' @xmath89 .
this helps separating the terms in the program by , say , recalling the @xmath90 in an algebraic program like mathematica in order to get the term @xmath91 in eq.([nfrec ] ) . as the program carries on the normalization , the book - keeping coefficients change
according to the rule that the coefficient of the poisson bracket of two terms with coefficients @xmath88 and @xmath92 is @xmath93 . in the end
we set @xmath94 to recover the correct values of all the coefficients .
this way of organizing the series is practical , but it also reflects a fundamental decision on what quantity we consider to represent ` smallness ' in the perturbation series .
if the overall size of @xmath32 is an @xmath95 quantity , then it is customary to identify @xmath1 itself as the small parameter .
in that case the normalization scheme proceeds by ascending powers of @xmath1 , and the generating functions @xmath86 are of order @xmath96 .
we shall see in subsection 2.4 how to modify the ` book - keeping ' so as to separate terms of different smallness which co - exist within @xmath32 , and then within @xmath84 , provided that the function @xmath32 satisfies analyticity conditions of the form ( [ expbd ] ) .
the choice of resonant module depends on the location of the domain @xmath31 in the action space and on the properties of the unperturbed hamiltonian @xmath29 .
let us consider the homological equation of the first order under the choice of book - keeping @xmath97 .
this reads : @xmath98 where @xmath99 is the part of @xmath32 containing terms to be eliminated .
the solution of ( [ homo2 ] ) is @xmath100 provided that the resonant manifolds @xmath101 are dense in the action space , for any small open domain @xmath31 there are infinitely many resonant manifolds crossing @xmath31 .
this implies that there is a dense set of values of the actions @xmath78 within @xmath31 for which a divisor in ( [ homosol2 ] ) becomes exactly equal to zero . in practice
, this means that a generating function like @xmath69 in ( [ homosol2 ] ) can not in fact be determined in any open domain @xmath31 . in nekhoroshev theory , an often stated remedy to this problem consists of splitting the set of fourier harmonics @xmath102 into low order and high order harmonics according to whether @xmath103 is smaller or larger than some positive integer @xmath104 ( see , for example , morbidelli and guzzo 1997 ) .
the choice of value of @xmath104 is arbitrary , but the choice @xmath105 , for some constant @xmath10 , is suggested by the request that there be no overlapping of the resonant domains ( of width @xmath106 ) formed around each of the resonant manifolds @xmath101 , with @xmath107 , other than the overlapping near the loci at which the manifolds intersect each other ( see morbidelli 2002 , p.119 ) . fixing @xmath104 and requesting that only the harmonics with @xmath107 be normalized ensures that divisors @xmath108 with @xmath109 will not appear in the series .
furthermore , the domain @xmath31 is crossed by only a _
finite _ number of resonant manifolds with @xmath107 .
more precisely , denoting by @xmath110 the subset of integer vectors @xmath56 for which the resonant manifolds @xmath111 intersect the domain @xmath112 , and by @xmath113 a subset of @xmath110 such that any vector of @xmath110 can be produced by a linear combination of the vectors of @xmath113 with integer coefficients , the relevant property is that @xmath113 contains only a finite number of vectors @xmath56 with @xmath107 and an infinite number of vectors with @xmath114 . denoting by @xmath115 the subset of vectors @xmath116 with @xmath107 , the resonant module @xmath117 is determined by :
@xmath118 let @xmath119 be the complement of @xmath120 with respect to @xmath110 . under the ` book - keeping ' choice @xmath97 ,
one then has the following iterative algorithm of determination of the normal form locally , i.e. , within the domain @xmath112 : \a ) assuming that @xmath121 normalization steps were completed , split @xmath122 as @xmath123 , where @xmath124 contains all the terms with @xmath107 not belonging to @xmath125 .
\b ) determine the generating function @xmath86 by solving @xmath126 \c ) find the new hamiltonian @xmath127 in which @xmath128 , @xmath129 are normal form terms belonging to the resonant module @xmath125 , @xmath130 contains all the fourier terms belonging to @xmath119 , and @xmath131 contains the remaining fourier terms that are not in any of the previously determined sets .
this completes one step of the iteration algorithm . as shown in morbidelli and giorgilli ( 1997 ) , while all the terms of order higher than @xmath42 in these two functions are bounded by a quantity @xmath132 , i.e. , they are both exponentially small in @xmath104 , the terms of @xmath130 are the ones making the leading contribution to the remainder .
two problems arise in the implementation of the above algorithm : \a ) eqs.([homosol1 ] ) or ( [ homosol2 ] ) imply that the denominators of all the series coefficients are functions of the actions , in the form of products of divisors @xmath133 .
this implies that even if @xmath134 is polynomial in the actions , the fourier coefficients of the terms @xmath102 in @xmath84 or @xmath135 are in general rational in the actions .
the extra computational load of dealing with rational expressions ( and their derivatives entering into poisson brackets ) makes the whole algorithm hardly tractable in this form .
we defer the solution of this problem after the analysis of point ( b ) below , to which it is linked .
\b ) one needs to store many more intermediate coefficients than those eventually needed in order to have a complete knowledge of the transformed hamiltonian or of the generating function within any specified domain of interest .
this is a rather technical problem , the solution of which is , nevertheless , crucial in constructing an efficient algorithm for the computation of the normal form . a detailed quantitative discussion of this problem is deferred to the appendix .
here we state the main result : under the book - keeping scheme @xmath97 , assume we are interested in specifying the transformed hamiltonian in a ` domain of interest ' defined by the maximum orders @xmath136 , where @xmath137 denotes the maximum order in the book - keeping variable @xmath1 and @xmath138 denotes the maximum order in fourier space .
then , in order to be able to specify all the terms belonging to the ` domain of interest ' , we must store , at intermediate normalization orders @xmath139 , all the hamiltonian terms of fourier order @xmath140 .
as shown in the appendix these intermediate terms outnumber by far the terms finally stored within the ` domain of interest ' . issue ( b ) can be resolved by introducing a different book - keeping of the terms , on the basis of their _ fourier order _ rather than @xmath141order .
this organization of the series is analyzed in detail in giorgilli ( 2002 , pp.86 - 87 ) and we present the main points of this analysis in the appendix . essentially , it reflects the fact that for any order @xmath139 , many fourier terms with a coefficient @xmath142 in front , which would thus be formally stored as of order @xmath143 under the book - keeping @xmath144 , are actually of much smaller size than @xmath142 , because the size of any @xmath102 term scales as @xmath145 by virtue of the analyticity condition .
this scaling is already present in the original hamiltonian and it propagates at all subsequent normalization steps .
precisely , this fact is recognized by ` book - keeping ' the terms on the basis of their fourier rather than @xmath1-order .
the exact algorithm , which replaces the algorithm of subsection ( 2.3 ) , reads as follows : \1 ) define @xmath146,1\}$ ] .
\2 ) place a book - keeping coefficient @xmath88 in front of each fourier term of the form @xmath102 in @xmath32 , where @xmath147 + 1 $ ] .
\3 ) carry on the normalization ( eqs.([nfrec ] ) ) by successive powers of @xmath87 .
as shown in the appendix , with such an algorithm there are no extra terms , outside the domain of interest , that need to be computed .
furthermore , it is possible to incorporate a solution to issue ( a ) in the same algorithm .
first , we notice that , as defined in step ( 1 ) above , the constant @xmath148 does not pose an upper bound ( like @xmath104 ) in the order of the fourier harmonics being normalized at successive steps , since the algorithm requests that terms of increasing order @xmath149 be normalized at the r - th normalization step . in principle , this would cause a problem when @xmath42 becomes larger than @xmath150 .
but in practice , one wishes to avoid the problem of the appearance of the actions in the denominators of the fourier coefficients for both @xmath107 or @xmath109 , that is for both @xmath151 and @xmath152 .
it can be shown ( section 3 ) that if the domain @xmath31 is resonant , that is @xmath153 , the width of @xmath31 scales with @xmath1 as @xmath106 .
non - resonant domains can also be covered by subdomains of size @xmath106 .
choosing a point @xmath154 in the interior of the domain @xmath155 , we then develop the hamiltonian locally , within @xmath155 , by expanding the actions as @xmath156 the transformation @xmath157 is not canonical under the usual poisson structure , but the dynamics remains unaltered if a new hamiltonian @xmath158 substitutes @xmath30 as generator of the equations of motion .
for example , in a thirring - type model @xmath159 we get , introducing also the ` book - keeping ' parameter @xmath87 : @xmath160 + 1}c_{k_1,k_2 } \exp(i(k_1\phi_1+k_2\phi_2))\bigg)~~.\end{aligned}\ ] ] consequently , the perturbation has been rescaled to @xmath161 , i.e. , it follows the size @xmath106 of the domain @xmath31 .
nevertheless , the unperturbed frequencies are now @xmath162 , thus they do not depend on the actions , since the terms quadratic in the actions are now formally of order @xmath87 .
this implies that the divisors are of the form @xmath163 , i.e. , independent of the actions .
consequently , the algorithm proceeds with polynomial rather than rational ( in the actions ) fourier coefficients at every step .
this also means that there is no longer need for introducing a k - truncated resonant module .
thus , in all calculations the resonant module is specified only by the resonances between the frequencies @xmath164 .
if @xmath164 are incommensurable , one may specify a module corresponding either to no resonance or to an approximate resonance between the frequencies @xmath164 . as discussed by morbidelli ( 2002 , p.48 - 49 ) ,
such a construction is still consistent with the appearance of an exponentially small remainder at the optimal order of normalization .
this is because , as demonstrated in the appendix , the order of normalization has a linear relation with the maximum fourier order of the terms contained in the normal form .
furthermore , we find that @xmath165 which also implies @xmath166 .
hence the dominant terms in the remainder are of fourier order @xmath167 , i.e. , they have a size @xmath168 .
this means that @xmath169 plays now a role similar to @xmath104 in subsection 2.3 .
it is interesting to note , however , that such a result is obtained here _ a posteriori _ ,
i.e. , after the construction of the series .
that is , with the present algorithm one may proceed by calculating the series up to any desired value of @xmath42 , and then by checking whether the optimal order @xmath48 was reached .
this will also specify automatically the optimal order @xmath169 . on the contrary , following the algorithm of subsection ( 2.3 ) , one would have to fix a value of @xmath104 in advance and then check whether such a value yields the optimal estimate for the remainder .
otherwise , the calculation should be repeated from the start , by making a different choice of @xmath104 .
the hamiltonian model of froeschl et al .
( 2000 ) reads : @xmath170 where @xmath171 , @xmath172 are action - angle variables and @xmath1 is a perturbation parameter ( note the change of notation for the actions , @xmath157 , with respect to the previous section , for consistency with the notation of froeschl et al .
( 2000 ) ) . in the action space for the value of the energy @xmath173 .
the three gray - shaded planes correspond to the resonances @xmath174 ( parallel to the double arrow ) , @xmath175 , ( to the left of the double arrow ) , and @xmath176 ( to the right of the double arrow ) .
the intersection of all the resonant planes with the paraboloid of constant energy produces a set of parabolic curves which is the web of resonances. ( b ) the projection of the web of resonances @xmath177 for @xmath178 on the @xmath179 plane .
whenever the coefficient of the term ( for one particular resonant vector @xmath56 ) of the fourier development of @xmath32 is positive , we only plot a single line corresponding to the projection of the associated parabolic curve on the plane @xmath179 .
if this coefficient is negative , we plot the central line and two other parallel lines marking the borders of the resonance as obtained by an analytic estimate of the associated separatrix width when @xmath180 .
these rules correspond to a poincar surface of section @xmath181 , where @xmath182 is the resonant angle associated with each resonance ( see text for details ) , and due to this reason fig.1b is directly comparable to fig.1 of lega et al .
the analysis in that paper and below refers to orbits exhibiting chaotic diffusion along the direction indicated by the double arrows at the borders of the resonance @xmath174.,title="fig : " ] in the action space for the value of the energy @xmath173 .
the three gray - shaded planes correspond to the resonances @xmath174 ( parallel to the double arrow ) , @xmath175 , ( to the left of the double arrow ) , and @xmath176 ( to the right of the double arrow ) .
the intersection of all the resonant planes with the paraboloid of constant energy produces a set of parabolic curves which is the web of resonances. ( b ) the projection of the web of resonances @xmath177 for @xmath178 on the @xmath179 plane .
whenever the coefficient of the term ( for one particular resonant vector @xmath56 ) of the fourier development of @xmath32 is positive , we only plot a single line corresponding to the projection of the associated parabolic curve on the plane @xmath179 .
if this coefficient is negative , we plot the central line and two other parallel lines marking the borders of the resonance as obtained by an analytic estimate of the associated separatrix width when @xmath180 .
these rules correspond to a poincar surface of section @xmath181 , where @xmath182 is the resonant angle associated with each resonance ( see text for details ) , and due to this reason fig.1b is directly comparable to fig.1 of lega et al .
the analysis in that paper and below refers to orbits exhibiting chaotic diffusion along the direction indicated by the double arrows at the borders of the resonance @xmath174.,title="fig : " ] the integrable part @xmath29 produces a simple dynamics : @xmath183 and @xmath184 , @xmath185 , @xmath186 .
thus all three actions are integrals of motion and all three angles circulate with constant frequencies .
the surface of constant energy @xmath187 is a paraboloid in the action space given by @xmath188 ( fig .
1a ) . on the other hand , the resonances @xmath189 @xmath190 , with @xmath191 , @xmath192 , define planes in the action space which are always normal to the plane @xmath179 . the intersections of the resonant planes with the surface of constant energy define the web of resonances on this surface which is a set of parabolic curves ( fig .
when viewed from the top of fig.1a , the projection of these curves on the plane @xmath179 defines a set of straight lines ( fig .
1b ) . when @xmath193 all the points on the surface of constant energy of fig.1a are neutral equilibria , which correspond to invariant 3-tori in the six - dimensional phase space .
if , however , @xmath194 , an @xmath195 volume of invariant tori of the phase space are destroyed , according to the kam theorem , and replaced by a chaotic subset of orbits .
the constant energy condition foliates the phase space into 5-dimensional hypersurfaces . for any fixed value of the angles ,
the projection of one hypersurface on the space of actions defines a manifold resembling to a paraboloid , like in fig .
1a , which , however , has some deformation of order @xmath195 .
the chaotic orbits can drift on this manifold .
we shall focus on orbits sliding on the manifold of constant energy along the chaotic borders of resonances in the direction indicated by the arrows of figs.1a , b .
as shown below , the planes of resonances @xmath196 of the unperturbed system are transformed to resonant manifolds of the perturbed system that also resemble to planes , but with some thickness of order @xmath197 , where @xmath198 is the coefficient of the term @xmath102 in the fourier development of the perturbation @xmath32 .
thus the intersections of the resonant manifolds with the manifold of constant energy form resonant zones of thickness @xmath199 .
when projected on the plane @xmath179 the web of resonances looks like in fig.1b ( the reason why in some cases we show one single line for one resonance while in other cases we show a triple line is discussed below ) .
the chaotic orbits studied by lega et al .
( 2003 ) are calculated at a value of @xmath1 at which the system is in the so - called ` nekhoroshev regime ' , i.e. the width of the resonant zones is small and there is no significant resonance overlap .
the orbits drift slowly along the border of the resonance @xmath200 , starting with initial conditions close to , but outside the intersection of this resonance with a different resonance , namely @xmath201 .
the chaotic drift along the zone of the resonance @xmath202 produces a slow secular change of the value of the action @xmath203 ( see eqs.([genres ] [ hampen ] ) below ) , which is the action associated with the oscillation in the so - called `` direction of fast drift '' , i.e. , normal to the resonant plane .
the change of the value of @xmath204 with time was used by lega et al .
( 2003 ) in order to measure numerically the coefficient of the chaotic diffusion .
the latter causes also a slow change of the value of the action @xmath205 , due to the confinement of the orbits on the manifold of constant energy . of the fourier development ( [ fourier ] ) as a function of the modulus @xmath206 for all resonant terms up to @xmath207 .
the straight solid line is an exponential fit to the maxima of the @xmath208 values as a function of @xmath206 . ]
an estimate of the width of resonances can be made as follows .
if we complexify the angles in a domain of @xmath209 containing the real 3-torus , i.e. , if we make the replacement @xmath210 in @xmath32 , with @xmath211 real , then in any direction of the space @xmath212 the quantities @xmath213 can vary from zero up to a value at which @xmath32 becomes singular .
however , there is a lower bound on the distance of all the singularities of @xmath32 from the real 3-torus , i.e. , from the value @xmath214 .
thus , the values of @xmath213 at the point of the closest singularity can not be all infinitesimally small .
in fact , if we develop the denominator of @xmath32 up to terms of second degree in @xmath213 , we find an estimate of the position of the singularities of @xmath32 close to the real 3-torus given by the values @xmath215 satisfying @xmath216 and @xmath217 the first of the above equalities implies @xmath218 the interior of any parallelepiped in the u - space defined by the three inequalities @xmath219 , for some positive constant numbers @xmath220 , @xmath221 , constitutes a domain of analyticity of @xmath32 if it is contained within the sphere @xmath222 .
the optimal estimate for the domain of analyticity corresponds to the parallelepiped with maximum volume @xmath223 , that is to the choice @xmath224 . on the other hand ,
a more precise value of @xmath225 can be found by expressing @xmath32 as a triple fourier series @xmath226 where @xmath227 , @xmath228 . according to the fourier theorem on analytic functions , the coefficients @xmath208
are bounded from above by latexmath:[\[\label{expdecay } and @xmath34 positive constants .
thus eq.([expdecay ] ) predicts that the coefficients @xmath198 decay exponentially .
this is shown in fig
. 2 , where we plot the values @xmath208 , calculated by a series development of @xmath32 using mathematica up to the 40th order , against @xmath206 . a numerical fitting to the upper bound of this diagram ( solid line ) yields @xmath230 , and @xmath231 .
the latter value is close to the estimate @xmath232 found above heuristically .
having determined the size of the fourier coefficients @xmath198 , the normal form construction described in subsection ( 2.4 ) eliminates all harmonics in the hamiltonian perturbation @xmath32 ( written in the form ( [ fourier ] ) ) , except for the harmonic @xmath233 corresponding to the particular resonance into consideration . to the lowest order
, the resonant normal form reads : @xmath234 where we have used the property @xmath235 following from the reality and even parity with respect to the angles @xmath236 of the hamiltonian perturbation @xmath32 . introducing resonant action - angle variables @xmath237 , @xmath238 via the generating function : @xmath239 where the integers @xmath240 are any pair satisfying @xmath241 , the resulting canonical transformation is @xmath242 by virtue of ( [ restran ] ) , the resonant hamiltonian ( [ resnf0 ] ) takes the form ( apart from a constant ) : @xmath243 i.e. , it is split in two parts depending only on the actions @xmath204 , @xmath205 , and a third part which is the pendulum hamiltonian : @xmath244 the actions @xmath205 and @xmath245 are integrals of the hamiltonian @xmath246 ( in the case of the resonance @xmath200 we have @xmath247 and the precise definition of the fast action is @xmath248 which differs by a factor @xmath249 from the definition given at the beginning of this section for that particular resonance ) . on the other hand ,
the width of the resonance is determined by the separatrix half - width of the pendulum hamiltonian @xmath250 @xmath251 the phase portrait of @xmath250 is shown schematically in fig .
3 . in reality , the ideal separatrix given by ( [ hampen2 ] ) should be replaced by a thin chaotic layer produced by the weakly chaotic motion near the separatrix due to to higher order coupling terms of the hamiltonian .
when @xmath1 is small , however , the width of the chaotic zones is very small . in that case ,
eq.([sepwidth ] ) can be used to approximate the maximum normal distance to the resonance between the upper and lower branches of the separatrix - like chaotic zones .
the resonant action @xmath252 changes in time according to @xmath253 any variation @xmath254 of the resonant action results in variations @xmath255 , @xmath256 , such as to respect the integral @xmath204 , i.e. , @xmath257 .
eq.([restran ] ) then gives @xmath258 given by eq.([hampen2 ] ) ( arbitrary units ) .
the left and right dashed vertical lines indicate the section @xmath181 when @xmath259 and @xmath260 respectively ( i.e. @xmath261 and @xmath262 respectively ) . ] when the resonance web is projected on the @xmath179 plane , as in fig .
1b , the variations @xmath255 and @xmath256 of eq.([di ] ) correspond to motions _ across _ the resonance ,
i.e. , along a line normal to the resonance line @xmath190 ( this is called the `` direction of the fast drift '' by froeschl et al .
( 2000 ) ) on the other hand , when the web of resonances is visualized numerically in the action space , as , for example , by the method of the fast lyapunov indicator ( froeschl et al .
2000 ) , one has to make an appropriate choice of poincar surface of section in order to eliminate the angles from the calculation .
the choice @xmath263 , @xmath264 made by these authors corresponds essentially to setting the resonant angle @xmath265 to a value very close to zero , i.e. , @xmath266 .
we then see that this means to plot ( a ) two sets of points in the plane @xmath179 , corresponding to passing close to the maxima or minima of the theoretical separatrices , when @xmath260 , or ( b ) one set of points on the plane @xmath179 , corresponding to passing close to the x - point of each separatrix , when @xmath259 . consequently , the thin chaotic borders of the resonances will appear as follows on the surface of section : in case ( a ) ( @xmath260 ) we see two lines parallel to the resonant line @xmath190 .
these lines define a zone of width determined by eq.([di ] ) . in case ( b ) ( @xmath259 ) we see only one line coinciding with the resonance line itself .
these rules are followed in the plot of fig.1b , for all the resonances with @xmath267 .
this figure is to be compared with fig.1 of lega et al .
the two figures compare well not only qualitatively , but also quantitatively , i.e , , the theoretical resonance widths found above are very close to the widths determined numerically by the fli method . the resonance dealt with by lega et al .
( 2003 ) is @xmath174 .
the diffusion takes place on the thin chaotic border along this resonance ( fig.2 of lega et al .
( 2003 ) ) when the initial conditions are taken in a small region of the thin chaotic border . in order to perform a transformation of the type given by eqs.([newact ] ) and ( [ hamsc ] ) , we have to specify the central values @xmath268 with respect to which the hamiltonian is expanded . by visual inspection of fig.2 of lega et al.(2003 )
a convenient choice compromising the central values of @xmath269 in all the panels is @xmath270 , @xmath271 . renaming the variables of eq.([hamfr ] ) by the same symbols according to @xmath272 , @xmath273 , @xmath172 , and fourier - expanding up to order 40 , the hamiltonian reads ( apart from a constant ) : @xmath274 according to the definition of @xmath148 given in subsection 2.4 , the term @xmath275 should be considered of order of smallness @xmath276 .
then , it is convenient to slightly modify the definition of @xmath148 so as to render @xmath277 comparable to the factor @xmath161 .
we thus set : @xmath278 + 1~~.\ ] ] the range of values of @xmath1 for which lega et al .
( 2003 ) provide numerical data on the diffusion coefficient is @xmath279 . for most values within this range
@xmath148 is equal to a constant value @xmath280 .
we thus simply set @xmath280 in all the calculations .
this means that the fourier modes of @xmath281 with @xmath282 are book - kept as of order 1 , modes with @xmath283 as of order 2 , etc .
the terms of @xmath284 are of order zero , i.e. , they are the ones to be used in the kernel operator @xmath285 of the homological equation .
although the computer program allows the user to carry along the powers of @xmath161 in the expansion , the extra requirement of memory in order to store the associated exponents of @xmath161 is prohibitive .
we thus had to make separate runs of the computation of the normal form for various numerical values of @xmath1 .
that is , one numerical value of @xmath1 was supplied to the file storing the coefficients of ( [ hamfrexp ] ) for each run .
even so , a memory limit of 2 gb was reached by computing about @xmath286 complex coefficients of the hamiltonian and of the lie generating function per run .
such a number of terms corresponds to a truncation @xmath287 in the normalization order , or @xmath288 in fourier space . despite these limits , it was in the end possible to observe the asymptotic properties of the birkhoff series within the whole range of values of @xmath1 down to a value @xmath289 , which is one order of magnitude smaller than the value reached in the numerical experiments of lega et al .
( 2003 ) . as a function of the truncation order @xmath290 after one and after two normalization steps ( upper and lower points respectively ) when @xmath291 .
( b ) the size of the remainder as a function of the order of the normalization for three different values of @xmath1 .
@xmath0 as a function of @xmath1 is given by the size of the remainder at the minimum of each curve . ] after @xmath42 normalization steps , the hamiltonian reads : @xmath292 the normal form @xmath293 contains all the fourier terms up to @xmath294 belonging to the resonant module @xmath295 .
the normal form @xmath293 is an integrable hamiltonian , the integrals , besides the energy , being @xmath205 and @xmath296 .
the level lines @xmath297 are normal to the resonance line @xmath174 .
since we are interested in measuring the drift along the chaotic border of the resonance , we can estimate the values @xmath298 and @xmath299 at the chaotic border either numerically , i.e. , by measuring the half - width of the resonant zone in each panel of fig.2 of lega et al .
( 2003 ) , or by using eq.([sepwidth ] ) with @xmath300 .
the numerical calculation was obtained by measuring the half - width @xmath301 of the interval along the line @xmath302 passing through @xmath303 with endpoints at the two sides of the chaotic border marked yellow in lega et al .
( 2003 ) , and by dividing @xmath301 by @xmath161 according to the rescaling @xmath304 .
as expected , the estimates for three different values of @xmath1 turned to be approximately equal , yielding average values @xmath305 .
these values were inserted in ( [ nffr ] ) in order to calculate the norm of the remainder @xmath306 .
it should be noticed that @xmath205 does not appear in the remainder because it is a dummy action , i.e. , it only appears linearly in the unperturbed hamiltonian . according to standard theory ,
the remainder series @xmath307 should be convergent .
this is shown in fig.4a , in which the truncated sum @xmath308 for the remainders of the first two normalization steps @xmath309 and @xmath310 for @xmath291 ( the norm @xmath311 is taken as the sum of moduli of all the trigonometric coefficients of the involved function ) .
clearly , the remainder shows the tendency to converge to its final size after summing just its first two or three successive terms .
furthermore , the size of the remainder , estimated as @xmath312 exhibits the well known behavior expected for an asymptotic series ( fig.4b ) .
namely , the size of the remainder initially decreases , as @xmath42 increases , up to an optimal order @xmath313 corresponding to the minima of the curves of fig.4b for different values of @xmath1 . beyond the optimal order ,
however , the remainder becomes an increasing function of @xmath42 and one has @xmath314 , marking the eventual divergence of the normalization procedure . from fig.5b
we see clearly that the optimal order @xmath48 increases as @xmath1 decreases .
figures 5 and 6 show now the main result .
the abscissa in fig.5 shows the value of @xmath315 for ten different values of @xmath1 in the range @xmath316 .
the ordinate shows the value of the diffusion coefficient @xmath7 for the same values of @xmath1 as given by lega et al .
the straight line represents a power - law fitting of the relation between the optimal remainder and the diffusion coefficient .
the best fit law is @xmath317 .
thus , the scaling is essentially : @xmath318 on the other hand , fig.6 shows the scaling of @xmath0 with @xmath319 when @xmath1 reaches values as small as @xmath289 , i.e. , one order of magnitude smaller than the last point of fig.7 of lega et al .
( 2003 ) ( the right vertical dashed line shows the last point where the numerical calculation of lega et al . (
2003 ) was stopped ) .
the two solid lines passing through the data correspond to a power - law ( lower curve ) and an exponential law ( upper curve ) fitting the data . for small values of @xmath319 ( i.e. for large values of @xmath1 ) ,
the power - law fitting is better than the exponential fitting .
the limit beyond which the exponential fitting is acceptable is around @xmath320 ( indicated by the left vertical dashed line of fig.6 at @xmath321 ) .
we may identify this limit as roughly corresponding to the threshold of the nekhoroshev regime . on the other hand , in the interval @xmath322 , both the power law and the exponential law yield acceptable fittings .
nevertheless , beyond the value @xmath323 , the power - law clearly fails while the exponential fitting follows now narrowly the data .
the numerical best fit exponential law yields @xmath324 reported in lega et al .
( 2003 ) as a function of the size of the optimal remainder @xmath0 of the birkhoff normal form calculated for different values of @xmath1 . ] as a function of @xmath319 .
the data clearly deviate from a power law beyond @xmath325 , and they are well fitted by an exponential law , as predicted by the nekhoroshev theorem .
the left vertical dashed line gives a lower threshold of @xmath319 below which the exponential law is no longer valid .
the data are well fitted also by a power law in the range @xmath326 .
the right vertical dashed line shows the last point of the numerical calculation of the diffusion coefficient by lega et al .
( 2003 ) . ]
we note that froeschl et al .
( 2005 ) , making numerical experiments of the diffusion of orbits in a mapping model resembling to the hamiltonian ( [ hamfr ] ) , found an exponential scaling of the diffusion coefficient with @xmath1 in which the best - fit exponent is @xmath327 .
however , in that paper too the exponential scaling was unraveled by considering values of @xmath1 smaller by more than one orders of magnitude from a critical threshold value below which resonances do not significantly overlap .
we conclude that the exponential law predicted by the nekhoroshev theorem can be unambiguously seen in the real data by various methods when @xmath1 becomes at least one order of magnitude smaller than the critical value characterizing the onset of the ` nekhoroshev regime ' .
the present paper demonstrates the applicability of the analytical apparatus of the nekhoroshev theory in recovering the exponential scaling of the coefficient of arnold diffusion @xmath328 along a particular resonance in the model hamiltonian system of three degrees of freedom of froeschl et al .
in particular : \1 ) the normalization algorithm for the hamiltonian , based on an organization of the perturbation series in terms of the fourier order of the terms ( rather than on powers of @xmath1 ) , is presented , and its benefits are analyzed in detail .
\4 ) the size of the optimal remainder is found to scale exponentially with the inverse of the perturbation parameter , namely @xmath330 , with @xmath18 .
the exponential scaling clearly shows when one considers values of @xmath1 as small as @xmath331 , i.e. , one order of magnitude smaller than in the numerical experiments of lega et al .
( 2003 ) .
estimates on the speed of arnold diffusion based on numerical experiments have been given in the literature by a number of authors .
we note in particular the early work of kaneko and konishi ( 1989 ) , in which an exponential scaling law is found with @xmath10 between @xmath332 and @xmath333 , i.e. , consistent with the present results . on the other hand ,
wood et al .
( 1990 ) provided estimates of the diffusion coefficient in standard - like multidimensional symplectic maps on the basis of the arnold - melnikov method , i.e. , the exponentially small splitting of separatrices due to the presence of higher order coupling terms in the resonant hamiltonian .
an exponential scaling of @xmath6 with @xmath319 was also found in that case , favoring though a value of @xmath10 rather close to @xmath334 .
thus , a detailed comparison of the results by the nekhoroshev and arnold - melnikov theories is in order . to our knowledge ,
the only hint in that direction is a paper by morbidelli and giorgilli ( 1997 ) .
nevertheless , more work is necessary in order to clarify this connection by specific quantitative studies , as well as to compare the predictions of the two theories for arnold diffusion in various types of multidimensional systems .
finally , the fact that the scaling of @xmath0 with @xmath1 appears also as a power law in a transient interval of values of @xmath1 , before the onset of the exponential regime , is consistent with some past theoretical work ( chirikov et al .
1985 ) calling such a transient regime ` modulational diffusion ' .
according to this theory , there is an intermediate interval of values of @xmath1 within which many high order resonances , located in the chaotic border of the low - order resonance along which the diffusion takes place , still overlap .
we conjecture that these are resonances of some order @xmath206 which must be above , but close to a truncation order @xmath104 estimated as @xmath105 ( according to the analysis of subsection 2.3 ) . at any rate ,
the results of the present paper provide a clear relation between the local value of the diffusion coefficient @xmath6 along a resonance , on the one hand , and the size of the optimal remainder @xmath0 of the normal form for the same resonance , on the other hand .
furthermore , for @xmath1 sufficiently small @xmath0 is found to scale with @xmath1 precisely as predicted by the nekhoroshev theory .
thus , the final conclusion of our study that _ the analytical techniques of the nekhoroshev theory can be used with much profit , in order to construct precise quantitative estimates of the speed of arnold diffusion in hamiltonian systems_. + + * acknowledgements : * i would like to thank professors a. giorgilli , for explaining to me the benefits of normalizing a series according to the fourier order of the terms , c. froeschl , for an invitation to nice where this project started , and z. kneevi and g. contopoulos , for many useful comments on the manuscript . c. lhotka made a careful reading and prepared fig.1 .
arnold , v.i . ,
1964 : _ sov .
dokl . _ * 6 * , 581
. benettin , g. , galgani , l. , and giorgilli , a. : 1985 , _ cel .
mech . _ * 37 * , 1 .
boccaletti , d. , and pucacco , g. : 1996,_theory of orbits _ , springer , berlin .
chirikov , b.v . ,
lieberman , m.a . ,
shepelyansky , d.l . , and
vivaldi , f.m . : 1985 , _ physica _ * 14d * , 289 .
contopoulos , g. : 2002 , _ order and chaos in dynamical astronomy _ , springer , berlin .
dumas , h.s . , and laskar , j. : 1993 , _ phys . rev .
lett . _ * 70 * , 2975 .
ferraz - mello , s : 2007 , _
canonical perturbation theories .
degenerate systems and resonance .
_ , springer , new york .
froeschl , c. , guzzo , m. , and lega , e. : 2000 , _ science _ * 289 * ( 5487 ) , 2108 .
froeschl , c. , guzzo , m. , and lega , e. : 2005 , _ cel .
* 92 * , 243 .
giordano , c.m . , and cincotta , p.m. : 2004 , _ astron .
astrophys . _ * 423 * , 745 .
giorgilli , a : 2002 , _
notes on exponential stability of hamiltonian systems _ , in dynamical systems .
part i : hamiltonian systems and celestial mechanics , pubblicazioni della classe di scienze , scuola normale superiore , pisa .
guzzo , m. , lega , e. , and froeschl , c. : 2005 , _ dis .
dyn . sys .
b _ * 5 * , 687 .
kaneko , k. , and konishi , t. : 1989 , _ phys .
a _ * 40 * , 6130 .
laskar , j. : 1993 , _ physica _ * d67 * , 257 .
lega , e. , guzzo , m. , and froeschl , c. : 2003 , _ physica d _ * 182 * , 179 .
lochak , p. : 1992
surv . _ * 47 * , 57 .
morbidelli , a. : 2002 , _ modern celestial mechanics .
aspects of solar system dynamics _ , taylor and francis , london .
morbidelli , a. , and guzzo , m. : 1997 , _ cel .
astron . _ * 65 * , 107 .
morbidelli , a. , and giorgilli , a. : 1997 , _ physica d _ * 102 * , 195 .
nekhoroshev , n.n . : 1977 , _ russ . math .
_ * 32*(6 ) , 1 .
pshel , j. : 1993 , _ math .
z. _ * 213 * , 187 .
skokos , c. , contopoulos , g. , and polymilis , c. : 1997 , _ cel .
astr . _ * 65 * , 223 .
wood , b.p . ,
lichtenberg , a.j . , and
lieberman , m.a . : 1990,_phys .
a _ * 42 * , 5885 .
let @xmath137 be the maximum order in the book - keeping variable @xmath97 , @xmath138 be the maximum order in fourier space specifying a domain in which we are interested in computing the hamiltonian , and @xmath104 the order of truncation in fourier space beyond which terms are not normalized ( in general we have to set @xmath335 so that some remainder terms of the hamiltonian be also stored ) .
figure 7a shows this domain as a gray - shaded parallelogram in the space @xmath336 .
now , it can be readily seen that many terms of the hamiltonian which lie _ outside _ the domain of interest at some particular order @xmath42 interact with the generating function in such a way as to produce , at some subsequent order , terms which lie _ inside _ the domain of interest .
indeed , consider a term of the form @xmath337 in the generating function , with @xmath338 and @xmath339 ( inside the domain ) , and a term @xmath340 in the hamiltonian , with @xmath341 and @xmath342 ( outside the domain ) .
assume that @xmath343 .
then , a poisson bracket operation between the two terms yields a term @xmath344\exp(i(k+k')\cdot\phi)$ ] . if @xmath345 ( this is possible because the components of @xmath56 and @xmath346 are added algebraically ) , the so produced term lies inside the domain of interest , despite the fact that the hamiltonian term from which it originated lies outside the domain of interest . ) of the terms that we wish to store in the transformed hamiltonian up to an order @xmath347 in the book - keeping parameter @xmath97 .
the inclined line , given by @xmath348 , represents the upper limit of a larger domain in fourier space within which intermediate terms must be stored at an order @xmath42 , yielding contributions in the gray - shaded domain at orders larger than @xmath42 .
( b ) under an appropriate book - keeping ( see subsection ( 2.4 ) ) the domain of interest can be made to coincide with the domain in which terms must be stored . ]
a careful examination of all the lie operations taking place up to the order @xmath137 shows that in order to obtain complete knowledge of the transformed hamiltonian in a domain @xmath136 , one must store , for every order @xmath139 , all the hamiltonian terms of fourier order @xmath349 .
this extended domain , shown by the triangular domain of figure 1a , contains many more terms than those of the domain of interest ( the ratio of the number of terms in the two domains is proportional to the ratio of the areas of the domains raised to a power @xmath350 , where @xmath26 is the number of degrees of freedom ) . at this point
we may argue that , for sufficiently large @xmath138 , the terms with @xmath351 have very small size because of the analyticity condition ( eq.([expbd ] ) ) , so that they can probably be ignored without great modification of the results .
however , precisely this type of argument brings about the real source of the above problem , which is the choice of ` book - keeping ' @xmath97 . the fact that the triangular domain of figure 1a contains many more terms than the gray parallelogram in the same figure simply depicts failure to recognize that for any order @xmath139 , many fourier terms formally stored as of order @xmath143 are actually of much smaller size than @xmath142 .
for example , while @xmath32 is an overall @xmath95 quantity , not all the fourier terms in @xmath32 have a similar size , because the size of an @xmath102 term in @xmath32 scales as @xmath145 .
this initial scaling propagates also at subsequent orders of normalization .
this fact suggests that the terms should be ` book - kept ' differently , i.e. , on the basis of their fourier order rather than their @xmath1-order . following the algorithm exposed in subsection 2.4 ( steps ( 1 ) to ( 3 ) )
, the domain of interest ( figure 1b ) is made to coincide with the domain in which terms should be stored , i.e. , terms within the gray - shaded area can only be produced by poisson bracket operations between terms which are also within the gray shaded area .
in particular , the meaning of step ( 2 ) is that in a hamiltonian like ( [ ham ] ) the parameter @xmath1 is not so relevant in characterizing the smallness of the various terms in @xmath32 ( @xmath1 can in fact be incorporated in the book - keeping scheme , as was done in section 3 ) .
a formal analogy can be made with a polynomial hamiltonian around an elliptic equilibrium : @xmath352 in the case of the hamiltonian ( [ hampol ] ) it is customary to proceed by normalizing iteratively by the degrees of the monomial terms in the canonical variables rather than by powers of @xmath1 .
such a choice underlines that the real smallness in the polynomial case is the distance @xmath353 from the equilibrium point .
in fact , it is often convenient to rescale the whole hamiltonian ( [ hampol ] ) so that @xmath1 disappears from the lowest order terms of the polynomial expansion in the r.h.s . of ( [ hampol ] ) .
the analogy with the generic case becomes clear by introducing action - angle variables @xmath354 , @xmath355 , @xmath356 .
we then readily find that a fourier term of order @xmath357 is always of fourier order @xmath358 ( and @xmath206 has the same parity as @xmath42 ) .
the relevant domain of interest is thus limited by the equation @xmath359 which yields a domain very similar to fig.7b ( the only difference is in the slope of the limiting upper line , which is @xmath360 $ ] in fig.7b and @xmath361 in the polynomial case ) . | the analytical techniques of the nekhoroshev theorem are used to provide estimates on the coefficient of arnold diffusion along a particular resonance in the hamiltonian model of froeschl et al .
( 2000 ) .
a resonant normal form is constructed by a computer program and the size of its remainder @xmath0 at the optimal order of normalization is calculated as a function of the small parameter @xmath1 .
we find that the diffusion coefficient scales as @xmath2 , while the size of the optimal remainder scales as @xmath3 in the range @xmath4 .
a comparison is made with the numerical results of lega et al .
( 2003 ) in the same model . |
based on their correlated x - ray spectral and rapid variability properties , the brightest persistent neutron star low - mass x - ray binaries ( nsxbs ) are often divided in two classes : the ` atoll ' and ` z ' sources @xcite , named after the shape of the tracks they trace out in x - ray color - color diagrams ( cds ) and hardness - intensity diagrams ( hids ) .
the z sources form the brightest of these two classes and are believed to accrete at near - eddington luminosities ( 0.51.0 @xmath2 , * ? ? ?
six galactic nsxbs have been classified as a z source : sco x-1 , gx 17 + 2 , gx 349 + 2 , gx 5 - 1 , gx 340 + 0 , and cyg x-2 .
they are characterized by three - branched tracks in their cds ( see figure [ fig : z_source ] for two examples ) , and hids , which in some cases resemble the character ` z ' . in other cases
they have a more ` @xmath3'-like shape , but we will refer to them as z tracks nevertheless . from top to bottom , the three branches of the z track are referred to as the horizontal branch ( hb ) , the normal branch ( nb ) , and flaring branch ( fb ) . only in cyg x-2 ,
gx 340 + 0 , gx 5 - 1 ( the so - called cyg - like z sources , * ? ? ?
* ) is the hb nearly horizontal , particularly in the hid . in gx 17 + 2 and
sco x-1 , the hb is nearly vertical , and in gx 349 + 2 no full - fledged hb has been observed so far .
on the other hand , only in the latter three sources ( the ` sco - like ' sources ) the fb can be identified with strong flaring , with count rates significantly higher than observed on the other two branches , whereas the cyg - like sources sometimes show count rate decreases on that branch . despite these ( and various other ) differences among the z sources
, it has often been assumed that in all of them the mass accretion rate ( @xmath0 ) increases monotonically from the hb , via the nb , to the fb .
there are several observational results that seem to support this : ( 1 ) the typical frequencies of most variability components ( which in many models scale with @xmath0 ) increase from the hb to nb , ( 2 ) the optical and uv flux , often thought to be the result of reprocessed x - rays , increase from the hb to fb ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ) , ( 3 ) jumps between branches are not observed , in accordance with the assumption that @xmath0 changes continuously and not abruptly .
however , this interpretation implies that x - ray luminosity does not track @xmath0 .
motion along the branches of the z usually takes place on time scales of hours to days . on longer time
scales it has been observed that the entire z track changes its location in the cd and hid , or even changes its morphology , as is best observed in cyg x-2 @xcite .
these longterm changes ( or ` secular changes ' as they are often referred to ) appear to be strongest in the cyg - like sources and it has been suggested that this is because the cyg - like sources might have a higher inclination @xcite .
more recently , @xcite suggested that the longterm changes of the z tracks might be the result of changes in @xmath0 , which would be in apparent conflict with the view that @xmath0 changes along the z track .
however , based on the observation that otherwise similar spectral / timing states can occur at different luminosity levels both across sources and in a given source , @xcite proposed a specific alternative quantitative measure to determine source state , namely @xmath0 through the disk , @xmath4 , normalized by its own longterm average .
changes in @xmath4 could then perhaps be responsible for motion along the z track _ and _ changes in the z track itself ( see 4.2 ) .
the power spectra of the z sources show several types of quasi - periodic oscillations ( qpos ) and noise components ( for reviews , see * ? ? ?
* ; * ? ? ?
their presence and properties are strongly correlated with the position of the source along the z track @xcite .
three types of low frequency ( @xmath5100 hz ) qpos are seen in the z sources : the horizontal branch ( hbos ) , normal branch ( nbos ) and flaring branch oscillations ( fbos ) .
their names derive from the branches on which they were originally found .
the hbo is found on the hb and nb with a frequency ( 1560 hz ) that gradually increases along the hb towards the nb .
the nbo and fbo are most likely part of one phenomenon .
they are found on the nb and fb ( near the nb / fb vertex ) but not on the hb . on the nb the qpo has a frequency of @xmath657 hz , which ( only in the sco - like sources ) rapidly increases to @xmath620 hz when the source moves across the nb / fb vertex @xcite .
twin khz qpos have been found in all z sources .
they are often observed simultaneously , with a frequency difference of @xmath6300 hz , and have frequencies between 215 hz and 1130 hz , which increase from the hb to the nb ( in sco x-1 also onto the fb , * ? ? ?
two types of noise are seen in the z sources .
these are the very low frequency noise ( vlfn ) and the low frequency noise ( lfn ) .
the vlfn , which is found at frequencies below 1 hz , can be described by a power law and is found on all branches , whereas the lfn , which is peaked or flat - topped noise component ( with cutoff frequencies of 210 hz ) , is only observed on the hb and nb .
in addition to the six sources that were classified as z sources by @xcite a few other persistent sources have shown properties that are characteristic of the z sources . in its high luminosity state cir x-1 has shown tracks in its hid that were identified with the hb , nb , and fb by @xcite .
these authors also detected qpos on the hb and nb that evolved in similar ways as the hbo and nbo in the z sources .
@xcite recently also reported the discovery of twin khz qpos in cir x-1 . at lower luminosities
, however , cir x-1 is similar to the atoll sources @xcite .
two other sources , both extra - galactic , were tentatively classified as z sources based on their tracks in a cd : lmc x-2 ( * ? ? ?
* changes in vlfn consistent with z sources ) and rx j0042.6 + 4115 in m31 ( * ? ? ?
* a very incomplete z track was observed and no timing information was available , so classification is uncertain ) .
the atoll sources distinguish themselves from the z sources not only by their lower luminosities ( @xmath60.010.5 @xmath2 ) and different tracks in cds , but in general they also have harder spectra and show stronger variability than the z sources . @xcite and
@xcite proposed to identify certain structures that atoll sources display in the cd , in luminosity ranges well below those of z sources , with hb , nb and fb . the rapid x - ray variability , and the order in , and timescales on , which sources trace out the branches are however not in accordance with an identification of this type @xcite .
most of the differences between the z and atoll sources are likely the result of a lower @xmath0 in the latter group .
however , it is still unclear if @xmath0 alone can explain all the observed differences .
in particular the four most luminous atoll sources gx 3 + 1 , gx 9 + 1 , gx 9 + 9 and gx 13 + 1 appear to have luminosities close to and perhaps overlapping those of z sources but quite different color and timing properties .
early on it was argued that a difference in strength of the neutron star magnetic field , possibly as the result of different evolutionary history , played a role in the observational appearance of the two classes @xcite ; the arguments for this were tied to specific models for the rapid variability @xcite and the spectra @xcite of these sources .
however , the discovery of khz qpos and millisecond x - ray pulsars have cast doubts on these models , and for certain qpo models the similarity in the range of khz qpo frequencies would require a fine - tuned balance between magnetic field and @xmath0@xcite .
the existence of cir x-1 , which can apparently switch between atoll and z - like behavior , suggests that @xmath0 alone might be enough to explain all differences between the two types of sources .
however , even when most similar to a z source cir x-1 is still peculiar , and most of the time the source displays behavior that is neither characteristically z nor atoll .
luminous transient nsxbs would be ideal sources to further test the relation of source types to @xmath0 , since they cover a large range in @xmath0 .
however , while many nsxb transients have shown atoll source characteristics , none has ever been observed to show the full range of properties associated with z sources , likely because most of them have peak luminosities well below @xmath2 . until recently ,
the source that has come closest to resembling a z source is xte j1806246 , a transient which during one observation near the peak of its 1998 outburst ( @xmath7 erg / s for 8 kpc , 230 kev ) showed an nbo - like 714 hz qpo that was found in a structure in the cd that resembled the z source nb / fb @xcite .
the source was poorly sampled near the peak of the outburst , so additional z - like properties may have been missed . at lower luminosities
xte j1806246 displayed atoll - like behavior .
xte j1701462 was first detected with the all - sky monitor ( asm ) on - board the _ rossi x - ray timing explorer _ ( _ rxte _ ) on 2006 january 18 @xcite .
a reanalysis of the asm data puts the start of the outburst between 2005 december 27 and 2006 january 4 ( ron remillard , private communication ) .
initial observations with the proportional counter array ( pca ) on - board _ rxte _ revealed qpos near 6 hz and 55 hz and large luminosity swings on time scales of a few minutes @xcite , but did not lead to a classification of the source .
follow - up observations @xcite , however , suggested that xte j1701462 is the first new z source in nearly 35 years , with the source exhibiting the typical z tracks in the cd , and showing the evolution of the timing properties along these tracks characteristic of z sources .
optical / near - ir @xcite and radio counterparts @xcite were suggested for xte j1701462 , which were later confirmed by a precise chandra localization of the source @xcite .
the source has also been detected with _ swift _
@xcite and _ integral _ @xcite . in this paper
we present an analysis of the first ten weeks of _
rxte_/pca observations of the outburst , which at the time of writing is still ongoing .
the main goal of this work is to show those results that firmly establish xte j1701462 as a z source , rather than present a complete analysis of the source .
an analysis of ( quasi-)simultaneous _ rxte _ and atca radio observations will be presented by fender et al .
( in prep . ) .
we have analyzed all 113 _
rxte_/pca @xcite observations of xte j1701462 made between 2006 january 19 and 2006 march 29 ( mjd 5375453824 ) .
four scanning observations , made on 2006 january 20 to improve the positional accuracy , were excluded from our analysis , since they did not yield useful data .
background subtracted light curves with a time resolution of 16 seconds were constructed from the ` standard 2'-mode data , covering the full effective pca energy range ( @xmath6260 kev , with 129 channels [ 1129 ] ) , using only data from proportional counter unit ( pcu ) number 2 ( the most reliable of the five pcus ) .
no dead - time corrections were applied , since these were relatively small : less than 5.5% for the highest count rates in our data set .
we also defined two x - ray colors , a soft color ( sc ) and a hard color ( hc ) , as the ratio of count rates ( extracted from ` standard 2 ' data ) in the @xmath64.07.3 kev ( channels 714 ) and @xmath62.44.0 kev ( ch.36 ) bands ( sc ) , and the @xmath69.818.2 kev ( ch .
2140 ) and @xmath67.39.8 kev ( ch .
1520 ) bands ( hc ) .
these colors were used to produce color curves , cds , hids , and soft color vs. intensity diagrams ( sids ) , with the intensity defined as the count rate in the full pca band ( ch .
1129 ) .
power density spectra were created from high time resolution data from all active pcus summed together , using standard fast fourier transform ( fft ) techniques @xcite .
the data were not background corrected and no dead - time corrections were applied prior to the ffts ; the effects of dead time were accounted for by our fit function .
we made power spectra from two energy bands , @xmath6260 kev and @xmath66.960 kev , with lengths of 16 seconds and nyquist frequencies of 8192 hz . the resulting power spectra were averaged , based on various properties ( e.g. time interval , count rate , color ) , and rms normalized @xcite .
these normalized power spectra were fitted with a combination of several lorentzians ( @xmath8^{-1}$ ] ) , a constant ( to represent the dead - time modified poisson level ) , and usually an additional power law ( @xmath9 ) . here
@xmath10 is the centroid frequency , @xmath11 the half - width - at - half - maximum , and @xmath12 the integrated fractional rms ( from @xmath13 to @xmath14 ) . instead of @xmath10 and @xmath11
we will quote the frequency @xmath15 at which the lorentzian attains its maximum in @xmath16 and the quality factor , q , where @xmath17 and @xmath18 @xcite .
the fractional rms amplitudes quoted in this paper were the integrated power between 0 and @xmath14 for the lorentzians , and between 0.1 hz and 100 hz for the power law component .
errors on fit parameters were determined using @xmath19 .
the low ( below a few hundred hz ) and high frequency parts of the power spectrum were usually fitted separately .
in figure [ fig : asm ] we show the _ rxte_/asm light curve of the ongoing outburst of xte j1701462 .
the source showed large flux variations during the first @xmath670 days of its outburst , which were followed by a long period in which the source flux appeared to be slowly declining . in this paper
we focus on the _
rxte_/pca data from the time interval marked by the two vertical dashed lines in figure [ fig : asm ] .
figure [ fig : curves ] shows the _
rxte_/pca light curve and color curves of xte j1701462 from this interval . immediately clear from the light curve
are the strong variations in the 260 kev count rate , by factors up to 2 , on time scales of hours to days .
these large count rate changes were accompanied by large color changes .
initially we created single cds , hids , and sids of all observations combined , but these diagrams revealed that on a time scale of days to weeks the shape of the tracks traced out by xte j1701462 changed considerably .
for this reason we decided to divide our data set into different intervals ; starting from the first observation we defined a new interval every time when in a new observation the pattern traced out in the cd had clearly shifted compared to the previous observations .
this method led to 10 intervals , labeled a to j , whose corresponding times are indicated in figure [ fig : curves]a .
the cds of each interval are shown in figure [ fig : ccd ] .
small shifts in count rate remained present in the hids / sids with the current choice of intervals .
however , these shifts did not show up in the cds .
the cds of the 10 intervals shown in figure [ fig : ccd ] reveal a large variety in shapes .
many of those shapes resemble the tracks traced out by the z sources and most of them are either similar to those of the cyg - like z sources gx 340 + 4 and gx 5 - 1 ( intervals a - d - see , e.g. , * ? ? ?
* ; * ? ? ?
* for comparison ) or the sco - like z sources ( intervals f , g , and j - see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* for comparison ) .
based on these similarities ( compare also with figure [ fig : z_source ] ) we can identify the branches of the tracks in figure [ fig : ccd ] with those seen in the z sources ( hb , nb , and fb ) , although support from the power spectra is required to confirm these identifications .
branch labels have been plotted in figure [ fig : ccd ] for the tracks of intervals b and g. the nb / fb vertex of the tracks seemed to move along a straight line in the cd , as the shape of the z track changed ( see fig .
[ fig : ccd ] ) .
the most striking difference between the cds of the cyg - like an sco - like intervals was the appearance of the fb .
in the cyg - like intervals the fb pointed downward - it was nearly vertical in interval c - and was relatively short . in the sco - like intervals the fb pointed upwards and was much longer .
intervals e , h , and i seemed to be intermediate between the cyg - like and sco - like intervals ; their fbs were short , like in the cyg - like intervals , but were nearly horizontal .
other differences between the intervals could be seen in the behavior of the hb , which was only clearly observed in intervals b , c , and f h . in intervals b and
c the hb pointed to the upper left and showed a prominent upturn to the upper right . in intervals
f h such an upturn was absent and the hb had become nearly vertical .
the orientation of the nb in the cds seemed to remain rather constant , but its length appeared to be shorter in the sco - like intervals .
as will be discussed in [ sec : compare ] , examples of these types of behavior have all been previously seen in other z sources .
changes in the shape and position of the z tracks were also seen in the sids and hids . in figure
[ fig : scd_hcd ] we show example sids and hids for intervals b and g. these two intervals showed nearly complete tracks in their cds and represent two opposites in terms of observed behavior in the sid and hid .
the hb and nb went from showing large count rate changes in intervals b and c to showing nearly constant count rates in intervals g and j. interval f , which appeared sco - like in the cd , still showed considerable changes in count rate on the hb , but nearly constant rates on the nb .
the ` intermediate ' interval h showed moderate variations in the count rate on the hb and nb .
large changes were also observed in the properties of the fb : in intervals a d the count rates decreased from the hb / nb vertex all the way to the end of the fb . in interval e complex behavior was observed , with decreases on the lower fb and increases on the upper fb ; the fb in interval f was similar , but the lower fb showed an additional loop , back through the nb , similar to what has been seen in cyg x-2 @xcite .
intervals g and j , on the other hand , showed count rate increases from the nb / fb vertex to the top of the fb .
note that the highest count rates in interval b were reached at the hb / nb vertex , whereas in interval g the count rates peaked at the top of the fb . in interval b ,
the top of the fb actually corresponded to the _ lowest _ count rates . despite minor differences , which will be discussed in [ sec : compare ] , most of the behavior seen in the hids of xte j1701462 has also been observed in the other z sources .
given the similarities with the z sources in the x - ray color diagrams , the results of our timing analysis of xte j1701462 will be discussed in terms of variability components commonly observed in the z sources .
like in other z sources we found that the power spectra of xte j1701462changed considerably as the source moved along its z track .
since interval b had the most complete track in the cd of all intervals , we chose to use that interval to study the evolution of the broadband power spectra in more detail .
similar evolution as described below was observed in the other intervals .
however , since our main aim is to establish xte j1701462 as a bona fide z source , a full analysis of all the other intervals is beyond the scope of this paper .
all values quoted for the noise and qpo properties were measured in the 6.960 kev band . in figure [
fig : scd_hcd ] we have marked ten locations along the track in the sid of interval b for which we have extracted power spectra .
the corresponding power spectra , which we will refer to , using roman numerals , as pds i - x , together with our best fits , are shown in figure [ fig : pds ] .
best fit parameters are given in table [ tab : fits ] .
the fits had reduced @xmath20 values between 0.81 and 1.25 .
figure [ fig : pds ] reveals clear changes in the shape of the pds and a decrease in the overall strength of the 1/16100 hz variability ( see also table [ fig : pds ] ) . on the hb ( pds i iv ) , where the power spectrum was dominated by a peaked / flat - topped noise component ,
the total strength of the 1/16100 hz variability smoothly decreased from @xmath622% rms to @xmath616% rms . a large drop in the strength of the variability occurred near the hb / nb vertex ( pds v ) where we measured @xmath67.7% rms . at that point
a power - law noise component became clearly visible in the power spectrum .
on the nb ( pds vi
viii ) the variability gradually weakened to @xmath63.5% rms near the nb / fb vertex ( ix ) , and on the fb ( pds x ) , where only power - law noise was detected , the strength of the variability was [email protected]% rms .
@xmath22 broad - band rms in the 1/16100 hz range + @xmath23 nbo parameters + @xmath24 parameter was fixed the peaked ( or sometimes flat - topped ) noise component that dominated on the hb was fitted with a lorentzian ( two were needed in pds i ) .
the q - value of these components was either fixed at 0 or less than 0.6 ( when not fixed ) . as the source moved along the hb onto the nb ( pds vi ) , the characteristic frequency @xmath15 of this component increased from @xmath63 hz to @xmath610 hz .
it is clear from figure [ fig : pds ] that the peaked noise became weaker in the direction of the nb , decreasing from @xmath618% rms in pds i to @xmath64% in pds vi . at some point near the middle of the nb ( pds vii )
the peaked noise was replaced by , or evolved into , a broad qpo , which we identified as the z source normal branch qpo ( nbo ) ; the nbo is discussed in more detail below . however , in addition to the nbo we detected a noise component with a frequency of 40 hz , which is probably not related to the peaked noise on the upper nb , but might be a ( broad ) remnant of the hbo .
on the lower nb ( pds viii ) we detected a broad component with properties similar to the peaked noise in pds vi .
no peaked or flat - topped noise was detected at the nb / fb vertex and on the fb , including the dipping part ( see pds ix and x ) .
the power - law noise component emerged near the middle of the hb . as the source moved along the hb to the lower nb , this component , which was fitted with a single power law , decreased in strength from @xmath67% rms ( 0.1100 hz ) to 1.4% rms . on the fb
its strength increased again to @xmath64% rms . moving from the hb to the fb the power - law index increased from @xmath60.8 to @xmath61.8 , with only pds ix departing from this trend ( with an index of [email protected] ) .
based on comparisons with other z sources , the peaked and power - law noise component can be identified as the z source low frequency noise ( lfn ) and very - low frequency noise ( vlfn ) . their respective parameters and their relative contribution to the power spectrum changed in ways comparable to the other z sources ( see , e.g. , fig . 5 in * ? ? ?
although the lfn was slightly stronger than in the other z sources , this is mainly the result of the slightly higher energy band we used in our analysis ( e.g. for pds i iii the rms of the lfn in the 260 kev band was a factor @xmath61.4 lower than in the 6.960 kev band ) and to a lesser extent due to a different lower frequency of the range in which we measured the rms ( 0 hz , instead of 0.1 or 1 hz ) .
in addition to the lfn and vlfn we also detected two types of low - frequency qpos , both of which are clearly visible in figure [ fig : pds ] , and khz qpos ( not seen in interval b ) . these qpos are described in more detail in the remainder of this section .
a variable qpo can be seen clearly in pds i vi .
it increased in frequency from @xmath69.5 hz on the hb upturn to @xmath653 hz on the upper nb as its strength decreased from @xmath611% rms to @xmath64% rms .
its q - value remained close to 6 in pds i iii after which it decreased to a value of @xmath61.7 in pds vi .
in pds i iii there were indications for broad features ( [email protected] ) at twice the qpo frequency , which were likely the qpo s second harmonic .
the frequency range of the main qpo , the presence of a second harmonic , and the relation of its frequency to position in the track was very similar to that of the hbo in the z sources .
the q - values and rms amplitudes were slightly higher than those of the hbo in the other z sources , but like in the case of the lfn this was the result of our choice of energy band .
we therefore identify this qpo as the z source hbo .
hbos were also detected on the hb of other intervals .
however , they were most prominent on the hb of the cyg - like intervals . in none of the other intervals
the hbo was observed at frequencies as low as in interval b , with interval c being the closest ( @xmath2511.5 hz ) .
the highest observed frequencies were near 59 hz , in interval h. another , broad ( [email protected] ) feature was detected at [email protected] hz on the middle / lower nb ( pds vii , see also figure [ fig : nbo ] ) .
depending on the type of interval , broad or narrow features around 7 hz were detected on the lower nb of almost every interval , except for intervals f ( which had poor coverage of the lower nb ) and j. in all cases these features were more prominent in the 6.960 kev band than in the 260 kev band .
the cyg - like intervals a d generally showed features with q - values less than 1.5 , with interval c showing the broadest feature ( [email protected] ) .
frequencies were around 6.3 hz ( a c ) or 8.1 hz ( d ) , with fractional rms amplitudes of 3.03.5% .
an example from interval b is shown in figure [ fig : nbo ] .
the ` intermediate ' intervals e , h , and i showed features with a higher coherence ( [email protected] ) that looked similar to the nbos seen in , e.g. , gx 340 + 0 , gx 17 + 2 , and sco x-1 ( see fig .
[ fig : nbo ] for an example from interval e ) .
the frequencies of these narrower qpos ranged between 6.9 hz and 7.3 hz , with fractional amplitudes between 2.2% and 3.6% rms .
based on the frequency range , q - values , and location on z track , we identify these narrow qpos , as well as the broader features around 7 hz in the cyg - like intervals , as the z source normal branch oscillations ( nbo ) . the sco - like intervals ( f , g , j ) provided only one nbo detection ( in interval g ) at [email protected] hz , with a q - value of [email protected] and an rms amplitude of [email protected]% . given the small number of detections in each interval , it was difficult to study the dependence of the nbo properties on the position along the z track . unlike what is seen in , e.g. ,
gx 17 + 2 and sco x-1 , the nbos in xte j1701462did not evolve into flaring branch oscillations .
no indications of such qpos were found on the flaring branches of any of the intervals .
we visually inspected all 260 kev and 6.960 kev power spectra for ( pairs of ) khz qpos , like those seen in the atoll and z sources @xcite .
in addition to the ( marginally significant ) pair of khz qpos reported by @xcite we found indications for three additional pairs and made one very significant ( 6.1@xmath26 ) detection of a single peak . the single peak ( see fig.[fig : khz ] ) was detected in observation 92405 - 01 - 03 - 05 ( march 7 ) at 760@xmath218 hz , with a q - value of [email protected] and a fractional amplitude of [email protected]% rms ( 6.960 kev ) .
an additional broad peak ( [email protected] ) was possibly present ( 2.7@xmath26 ) at 470@xmath2150 hz .
other twin khz qpos were only marginally detected , with single trial significances for the individual peaks ranging from only 2@xmath26 to 4@xmath26 .
the most significant pair ( 4@xmath26 and 3.8@xmath26 , see fig .
[ fig : khz ] ) was detected in observation 92405 - 01 - 03 - 05 ( march 21 ) ; the frequencies were 620@xmath2110 hz and 909@xmath218 hz , with q - values of 8@xmath213 and 16@xmath216 , and rms amplitudes of [email protected]% and [email protected]% .
the peak separation of this pair was 288@xmath2113 hz ; the other ( marginally detected ) pairs gave values of 293@xmath2156 hz , 291@xmath2125 hz , 249@xmath2118 hz , and 265@xmath2135hz . no significant dependence of peak separation on khz qpo frequency was found .
we note that all the pairs of khz qpos we have detected ( without reference to expected frequency - frequency relations ) , had frequencies matching the frequency - frequency relations of other z , and atoll , sources ( see , e.g. , * ? ? ?
all ( marginal ) detections of khz qpos were made in observations on the hb and only in intervals that were sco - like or intermediate between sco - like and cyg - like .
for each of our intervals we tried to increase our sensitivity to khz qpo by combining hb data from multiple observations ( selections were made in the sid ) .
this did not result in additional detections .
we determined upper limits ( 95% confidence ) on the strength of possible khz qpo in the hb power spectra ( pds i iv ) of interval b in the 1001500 hz range , with q values fixed to 2 , 5 , and 8 .
we found rms values ranging from 1.7% to 6.3% .
during the first ten weeks of our _ rxte_/pca observations of xte j1701462 , the source showed all the characteristic x - ray color and variability properties commonly associated with the z sources . in the cd three - branched tracks
were traced out , which , although variable in shape , often resembled those of the z sources ( [ sec : results_curves ] ) , and the power spectra showed various noise components and qpos that were similar , in terms of properties and evolution along the z track , to those seen in the z sources ( i.e. lfn , vlfn , hbo , nbo , and khz qpos ; [ sec : results_variability ] ) .
xte j1701462 displayed a surprisingly large variety in the shape of its z tracks in the cd . during the first part of the outburst ( intervals a d ) the z tracks showed similarities to those of the cyg - like z sources , in particular gx 5 - 1 @xcite and gx 340 + 0 @xcite .
these similarities included an upturn on the hb and a downturn on the fb , which has previously been referred to as the ` dipping ' fb ( * ? ? ?
* ; * ? ? ?
* no clear dips were seen in the light curves of xte j1701462 ) . in the second part of our data
set the z tracks were more similar to those of the sco - like z sources ( intervals f , g , and j ) with the fb being more prominent than in the first intervals . comparing the sid / hid tracks of xte j1701462 with those of the z sources revealed some minor differences .
for example , the hid tracks of the cyg - like intervals appear a bit different from those of gx 5 - 1 and gx 340 + 0 ; in xte j1701462 the count rate decreased from the hb / nb vertex to the end of the fb , whereas gx 5 - 1 and gx 340 + 0 showed a decrease from the hb / nb vertex to the nb / fb vertex , followed by an increase to the middle of the fb and dipping at the top of the fb .
cyg x-2 , on the other hand , does occasionally show similar count rate behavior as xte j1701462 did in its cyg - like intervals @xcite .
although changes near the middle of the fb were also seen in xte j1701462 ( interval f ) the observed behavior was the opposite of gx 5 - 1 and gx 340 + 0 ( decrease in count rate on the fb followed by an increase ) . among the sco - like intervals we observed some variety in the hid tracks .
intervals g and j were similar to the sco - like z sources , but interval f seemed to combine some properties of the sco - like and cyg - like sources .
all these differences show that similarities in the shapes of z tracks in the cd do not necessarily extend fully to the hid and sid .
the nb / fb vertex was the only vertex observed in all intervals and can serve as a reference point for following the evolution of the z track .
we estimated the count rate at the nb / fb vertex for each of the intervals , which led to the following ordering ( from high to low count rates ) : c , b / d , a , e , f / h , g / i , and j. following this sequence we observed that the z tracks shifted to harder colors as the source became weaker ( see fig .
[ fig : ccd ] ) , which has also been seen in cyg x-2 @xcite .
another , very clear change occurred in the orientation of fb in the cd , pointing downwards when the source was brightest and rotating counter - clockwise as the source became weaker . in the sid
the fb turned clockwise from the left ( c ) to upper left ( b / d , a ) , then partially overlapped with the nb ( e , f / h ) and finally pointed to upper right ( g , j ) . at the same time the range in count rates observed on the hb and nb became much smaller and the hb changed from pointing to the upper left to a nearly vertical position in the hid .
changes in the shape and/or position of the tracks in the cds , sids , and hids ( secular changes ) have been observed in gx 17 + 2 and the cyg - like sources , with cyg x-2 showing the most dramatic changes @xcite . in cyg x-2 it was found that the fb changed from pointing to the left in the hid to pointing to the upper left when the source changed between its high and medium intensity levels , i.e. , in the same sense relative to flux level as we observe in the sid of xte j1701462 . on one occasion
@xcite observed a flaring - like branch , with a shape similar to that of the fb observed in the sco - like source .
this occurred at the lowest intensity levels , also in line with what we observed for xte j1701462 .
a detailed timing analysis of all intervals needs to be done to compare the properties of individual variability components , but at first sight the evolution of the variability properties did not seem to be affected strongly by the changes in z track itself .
there were however changes in the presence of certain types of variability .
first , there was a clear evolution in the properties of the nbo , changing from a broad feature in the cyg - like intervals to a narrow feature in the intermediate intervals and perhaps disappearing in the sco - like intervals .
second , khz qpos were only detected in the intermediate and sco - like intervals , despite similar or even higher count rates in the cyg - like intervals and the fact that khz qpos have been found in all three cyg - like z sources @xcite .
however , it is interesting to note that the khz qpos of cyg x-2 where discovered when the source showed cd tracks similar to those of the sco - like intervals f and g @xcite , but they were not found when its cd tracks were more similar to those of the cyg - like intervals b and c @xcite .
as discussed earlier , xte j1701462 is not the first z source to show large secular changes , but it is the first in which these changes resulted in a switch between full - blown cyg - like and sco - like behavior ; cyg x-2 has also shown a sco - like flaring branch @xcite , but not a full sco - like z track .
observing both types of behavior in a single source puts a different perspective on two previously proposed explanations for the existence of two types of z sources and also on the origin of the secular changes .
@xcite suggested that the difference between the cyg - like and sco - like z sources is the result of a lower binary inclination in the latter group .
our observations of xte j1701462 would require a change in the binary inclination on a time scale of a few weeks , much faster than expected ( and requiring interaction with a third stellar component ) .
the accretion disk itself might be precessing on such a time scale , possibly leading to different viewing angles of the inner accretion flow ( as originally suggested by * ? ? ? * for cyg x-2 ) .
however , this would likely only have an observable effect in sources with a high binary inclination , for which we have seen no indications ( such as dips or eclipses ) .
moreover , at some point , when the inner disk returns to its original inclination , xte j1701462would have to switch back to cyg - like behavior , which has not yet been seen at the time of writing .
@xcite explained the difference between the two groups in terms of the neutron star magnetic field , which should be smaller in the sco - like group .
again , our observations of xte j1701462 require changes on a time scale of a few weeks , much faster than is expected for magnetic field decay .
another , somewhat related option might be magnetic screening , which causes the magnetic field to become buried by the freshly accreted material at high @xmath0 ( see , e.g. , * ? ? ? * ) .
however , at near - eddington @xmath0 the effects of the screening are much larger ( orders of magnitude decrease ) than the expected difference ( if any exists at all ) between the cyg - like and sco - like sources
. moreover , the fact that we see two types of persistent z sources , accreting at near - eddington rates , also rules magnetic screening out as a explanation , because the magnetic field should be buried in both types of sources .
since the above explanations do not work , we have to search for another parameter that might result in two types of z sources as well as the observed switch in xte j1701462 . given the transient nature of xte j1701462@xmath0 seems an obvious candidate , not only to explain the differences among the z sources , but also as the origin of the secular changes seen in individual sources ( as suggested by * ? ? ?
this option has long been disregarded as a viable alternative because it was generally assumed that @xmath0 is already responsible for changes along the z track .
however , the possibility of @xmath0 causing both changes along the z track _ and _ changes in the shape of the z track itself is not necessarily a problematic one . as a solution to the problem of the ` parallel tracks ' in nsxbs ( e.g.
* ) , @xcite suggested a model in which there exists both a prompt and a filtered response to changes in the @xmath0 through the disk , @xmath4 . suggested mechanisms responsible for such a filtered response are a radial inflow or nuclear burning . in the cases of the ` parallel tracks ' , the khz qpo frequency
is then determined by the ratio of the @xmath4 and filtered @xmath4 , @xmath5@xmath4@xmath1 , and the luminosity is determined by @xmath4 plus a contribution from @xmath5@xmath4@xmath1 . following @xcite
, we apply this model to xte j1701462 and suggest that the ratio @xmath4/@xmath5@xmath4@xmath1determines the position along the z ( which in z sources also means the khz qpo frequency ) .
motion along the z is then the result of spectral changes caused by variations in @xmath4/@xmath5@xmath4@xmath1 .
the details of how these spectral changes occur , which determine the shape of the z tracks , depend on the underlying physical components , whose properties vary with @xmath4 , @xmath5@xmath4@xmath1 , or a combination of the two . hence , as the luminosity changes ( due to changes in @xmath4 and/or @xmath5@xmath4@xmath1 ) , the shape of the z tracks is likely to change , as we observe xte j1701462 . as long as the total @xmath0 does not change too much , the overall shape of the z is expected to remain similar . assuming that the count rates at the nb / fb vertex in each interval are representative for the @xmath0 during those periods , we attempt to estimate the magnitude of the changes in @xmath0 that are required to change the appearance of the z track .
interval c probably had the highest @xmath0 ( see [ sec : compare ] ) , so cyg - like behavior can probably be associated with the highest @xmath0 range .
the count rates at the nb / fb vertex in the sco - like interval g were about a factor of 1.5 lower than those in the c , suggesting that changes in the count rate ( and perhaps @xmath0 ) by factors of 1.5 or less might already be enough to switch between the two types of z source behavior .
taking the peak count rates of these two intervals , instead of the nb / fb count rates , gave a similar factor ( 1.7 ) .
in the past the luminosities at the vertices of the z have been treated as standard candles ( possibly related to @xmath2 ) , to estimate the distances of the z sources ( see , e.g. , * ? ? ? * ) .
our xte j1701462 observations show that probably none of these vertices can be uniquely identified with @xmath2 or any other constant luminosity .
in fact , they suggest that _ the cyg - like z sources might be substantially more luminous ( @xmath150% ) then the sco - like z sources _ , which affects distance estimates that are based on the assumption that all z sources have similar luminosities . since sco x-1 , for which an accurate distance has been established through parallax measurements @xcite , is known to have super - eddington luminosities occasionally ( for a 1.4 @xmath27 ; * ? ? ? * ) , the cyg - like z sources are probably even further above the eddington limit . .
the arrows indicate how the different branches of the ( black ) tracks changes their orientation .
the grey area at the top of the atoll track represent the extreme island state that is only seen a few atoll sources and which was interpreted by @xcite and @xcite as the equivalent of the z source horizontal branch .
an analysis by @xcite has shown that the island state can connect to different locations along the extreme island state.,width=302 ] it is likely that xte j1701462 eventually decreases to count rates well below the range observed for the sco - like intervals and at that point it may turn into an atoll source .
based on the observed change between the two types of z behavior this will probably be a gradual process as well .
xte j1701462 has shown that the orientation and shape of the branches in the x - ray color diagrams change as a function of the overall count rate level . extrapolating this behavior to lower count rates
suggests that the branches of the z track could slowly transform into the branches seen in atoll sources . in figure
[ fig : evolution ] we have drawn a simple cartoon that depicts how the tracks in a cd might evolve as a function of decreasing @xmath0 .
the hb of the cyg - like z tracks rotates clockwise from the upper left to the upper right and together with the nb it transforms into the atoll island state .
the fb transforms from a short downward pointing branch into a long upward curved branch that becomes the atoll banana state .
of course , this proposed evolution is purely based on the shape of the tracks in the cds .
a detailed timing analysis is still necessary to check these claims .
moreover , sofar we have only described the observed spectral evolution in terms of phenomenological changes ( i.e. changing position and orientation of the different branches ) .
spectral fits are necessary to relate these phenomenological changes to physical changes in , e.g. , the boundary layer or comptonizing medium . as @xmath0 decreases
, xte j1701462 may first go through a phase in which it resembles the bright atoll sources gx 9 + 1 , gx 9 + 9 , gx 3 + 1 , and in particular gx 13 + 1 , which appears to be a hybrid atoll / z source @xcite , and then possibly starts showing the characteristic atoll island / banana states @xcite .
note that , although the cd track of interval j looks similar to the ones traced out by gx 13 + 1 @xcite , the timing properties in this interval are different from gx 13 + 1 .
if ideas attributing z behavior to a relatively stronger b field are correct then at lower luminosity , where scattering in the accretion flow preventing us from detecting the pulsations would be less , a ( possibly millisecond ) pulsar might be observed .
it is also quite possible that x - ray bursts will occur at some point .
burst and pulsations have not been seen in xte j1701462 at the time of writing .
based on the ( nearly ) horizontal orientation of the extreme island state in the cds of some atoll sources , @xcite and @xcite identified this state with the z source hb ( see also the cartoon in fig . [
fig : evolution ] ) .
in addition to the concerns raised by other authors @xcite , we note that at lower @xmath0 ( in the sco - like intervals ) the hb no longer seems to be horizontal .
so , unless the hb would return to its original horizontal position at even lower @xmath0 , for which we see no evidence in the more recent data of xte j1701462 , the extreme island state seen in some atolls is probably not related to the hb seen in the z sources .
perhaps the hb upturn observed in intervals b and c might related to the extreme island state , but this upturn was not observed often enough to test this
. why is xte j1701462 the only transient z source thus far ?
the persistent z sources differ from the persistent atoll sources in that they have a much higher @xmath0 .
it has been suggested that this could be the result of differences in the properties of the secondary star .
the long orbital periods found for sco x-1 ( 18.9 hr , * ? ? ? * ) and cyg x-2 ( 9.84 days , * ? ? ?
* ) suggest that the companion stars in z sources are evolved stars , resulting in higher mass transfer rates .
what matters for the properties of transient outbursts ( e.g. duration and perhaps peak luminosity ) is not the mass transfer rate from the secondary , but how much matter has been stored in the accretion disk before an outburst starts .
obviously , systems with longer orbital periods result in larger roche lobes for the neutron star , which allows for larger accretion disks that presumably can store more matter . calculating the exact properties of such a system
is beyond the scope of the current paper .
however , given the fact that at the time of writing xte j1701462 has been accreting at near eddington levels for more than 215 days , i.e. much brighter and for a longer period than most transient nsxbs , it is likely that , just like some of its persistent counterparts , the transient z source xte j1701462 has a long orbital period . at this point
no clear signatures of an orbital period have been found in x - ray or optical / ir light curves .
the best indications for possible modulations were observed in the _ rxte_/asm ( see fig . [
fig : asm ] ) and _ rxte_/pca light curves between mjd @xmath653820 and mjd @xmath653920 .
these modulations had a ( quasi-)period of @xmath625 days , but they are no longer present at the time of writing and their nature is unclear .
we probably have to wait until the source has reached quiescence to search for optical signatures of an orbital period . in any case , the fact that xte j1701462 is the first transient z source would suggest that the number of sources with long orbital periods is small , in line with small number of the persistent z sources . assuming that the shape of z tracks changes with luminosity in the same way in all z sources we can attempt to estimate the distance of xte j1701462 .
we do this by comparing the flux of xte j1701462 to that of sco x-1 , which is the only z source with a well determined distance ( [email protected] kpc , as determined by radio parallax measurements ; * ? ? ?
* ) . for our estimate
we use the track of interval g , which , despite a small difference in the orientation of the hb , resembled the track of sco x-1 most @xcite .
we choose the nb / fb vertex as a reference point . for sco x-1
we measured a 325 kev kev luminosity of @xmath28 ergs@xmath29 at that vertex ( obs - id : 40706 - 02 - 06 - 00 ) ; we find an unabsorbed 325 kev flux for xte j1701462 of @xmath30 ergs@xmath29@xmath31 ( obs - id : 92405 - 01 - 01 - 03 ) , which implies a source distance of 14.7 kpc , with an uncertainty of at least 20% . for that distance ,
the highest 218 kev luminosity measured in our data set ( i.e. near the hb / nb vertex in interval c , obs - id : 91106 - 02 - 03 - 14 ) would be @xmath6@xmath32 ergs@xmath29 , well above the eddington limit of a 1.4 @xmath27 neutron star . note , again , that both the distance and luminosity estimates rely on assumptions that can not be verified easily .
we observed xte j1701462 during the early phase of its first known outburst . based on its correlated x - ray color and timing properties we concluded that during this period the source could be classified as a z source , making it the first transient neutron star x - ray binary with such properties .
no x - ray bursts or pulsations have been observed at the time of writing .
a clear change was observed in the tracks traced out in x - ray color diagrams .
initially these looked very similar to those of the cyg - like z sources ( gx 5 - 1 , gx 340 + 0 , and cyg x-2 ) , but later they gradually transformed into tracks similar to those seen in the sco - like z sources ( gx 17 + 2 , gx 349 + 2 , and sco x-1 ) .
we observed noise components and qpos such as commonly also seen in other z sources and found that the properties and/or presence of the nbo and khz qpos depended on the shape of the z track .
however , the general evolution of the variability properties along the z track did not change much between the different tracks . the transient nature of xte j1701462 and the associated changes in @xmath0 are likely responsible for the large changes in the z tracks , which we interpret as extreme cases of the secular changes observed in the persistent z sources .
this interpretation requires both a prompt and a filtered response to changes in the @xmath0 if one assumes that @xmath0 also correlates with position along the z track .
as xte j1701462continues its decay , the branches of the z are likely to transform further .
xte j1701462 provides us with a unique opportunity to follow the behavior of accretion flows onto neutron stars from ( super-)eddington accretion rates all the way to quiescence .
the first ten weeks of our _ rxte _ observations already allowed us to construct a sequence of z tracks based on overall count rate levels .
hopefully , as the source decays , this sequence can be extended to include different types of phenomenology such as perhaps atoll - like tracks , x - ray bursts or even pulsations .
such a sequence would be of great importance for understanding the nature of the different types of persistent nsxbs .
the authors would like to thank jean swank and evan smith for scheduling the daily _ rxte _ observations of xte j1701462 , ron remillard for stimulating discussions , erik kuulkers and peter jonker for their feedback on an earlier version of this paper , and , finally , the referee for carefully reading our manuscript and providing helpful comments .
this research has made use of data obtained from the high energy astrophysics science archive research center ( heasarc ) , provided by nasa s goddard space flight center .
e.g is supported by nasa through _ chandra _ postdoctoral fellowship award pf5 - 60037 , issued by the _
chandra _ x - ray observatory center , which is operated by the smithsonian astrophysical observatory for and on behalf of nasa under contract nas8 - 39073 .
, m. 1989 , in proceedings of the nato advanced study institute on timing neutron stars , held in eme , .
izmir , turkey , april 415 , 1988 .
editors , h. gelman and e.p.j .
van den heuvel ; publisher , kluwer academic , dordrecht , the netherlands , boston , massachusetts , p. 27 , m. 1995 , in proceedings of the nato advanced study institute on the lives of the neutron stars , held in kemer , turkey , august 29-september 12 , 1993 .
editors , m.a .
alpar , u. kiziloglu , and j. van paradijs ; publisher , kluwer academic , dordrecht , the netherlands , boston , massachusetts , p. 301 | we report on the first ten weeks of _ rossi x - ray timing explorer _
observations of the x - ray transient xte j1701462 .
large count rate variations were accompanied by changes in the source s spectral and variability properties . based on comparisons with other sources
we conclude that xte j1701462 had all the characteristics of the neutron star z sources , i.e. , the class of the brightest persistent neutron star low - mass x - ray binaries .
these include the typical z - shaped tracks traced out in x - ray color diagrams on time scales of hours and the variability components detected in the power spectra , such as khz quasi - periodic oscillations ( qpos ) and normal and horizontal branch oscillations .
xte j1701462 is the first transient z source and provides unique insights into mass accretion rate ( @xmath0 ) and luminosity dependencies in neutron star x - ray binaries .
as its overall luminosity decreased , we observed that xte j1701462 switched between two types of z - source behavior never previously identified as clearly in the same z source , with most of the branches of the z - track changing their shape and/or orientation .
we interpret this switch as an extreme case of the more moderate longterm changes seen in the persistent z sources and suggest that they are the result of changes in @xmath0 .
we also suggest that the cyg - like z sources ( cyg x-2 , gx 5 - 1 , and gx 340 + 0 ) might be substantially more luminous ( @xmath150% ) than the sco - like z sources ( sco x-1 , gx 17 + 2 , and gx 349 + 2 ) . adopting a possible explanation for the behavior of khz qpos , which involves a prompt as well as a filtered response to changes in @xmath0
, we further propose that changes in @xmath0 can explain both movement along the z track and changes in the shape of the z track itself .
we discuss some consequences of this and consider the possibility that the branches of the z will smoothly evolve into the branches observed in the x - ray color diagrams of the less luminous atoll sources , although not in a way that was previously suggested . |
cloud radio access networks ( c - rans ) have received considerable research interest as one of the most promising solutions to mitigate interference , fulfil energy efficiency demands and support high - rate transmission in the fifth generation cellular network @xcite . in c - rans , a large number of remote radio heads ( rrhs ) are deployed , which operate as non - regenerative relays to forward received signals from mobile stations ( mss ) to the centralized base band unit ( bbu ) pool through wire / wireless backhaul links for uplink transmission @xcite . to suppress the inter - rrh interference by using cooperative processing techniques at the bbu pool ,
the channel state information ( csi ) of both the radio access links ( als ) and wireless backhaul links ( bls ) are required@xcite . in @xcite , though a segment - training scheme was proposed to estimate the individual channel coefficients for two - hop scenario under flat fading environments , this proposal results in high overhead consumption for backhaul transmission since the rrhs need to forward both al and bl training sequences to the bbu pool @xcite . on the other hand
, the superimposed - training scheme @xcite , where a training sequence is superimposed on the data signal , can significantly reduce the overhead and is valid to perform channel estimation for time - varying environments using complex - exponential basis expansion model ( ce - bem ) @xcite .
however , straightforward implementation of superimposed training in c - rans would degrade transmission quality due to the fact that superimposing both al and bl training sequences on the data signal declines the effective signal - to - noise ratio ( snr ) @xcite . motivated to reduce the training overhead and enhance the channel estimation performance at the bbu pool , a superimposed - segment training design is proposed in this letter , where superimposed - training is implemented for the radio al while the segment - training is applied for the wireless bl .
moreover , based on the training design , a ce - bem based maximum a posteriori probability ( map ) channel estimation algorithm is developed , where the bem coefficients of the time - varying radio al and the channel fading of the quasi - static wireless bl are first obtained , after which the time - domain channel samples of the radio al are restored in terms of maximizing the average effective snr .
_ notation _ : the transpose , hermitian and inverse of a matrix are denoted by @xmath0 , @xmath1 and @xmath2 , respectively ; @xmath3 represents the two - norm of a vector ; @xmath4 defines the magnitude of a complex argument ; @xmath5 is the kronecker product ; @xmath6 denotes the diagonal matrix with @xmath7 being the @xmath8th diagonal element ; @xmath9 represents the trace of a matrix ; @xmath10 and @xmath11 are the @xmath12 unit diagonal matrix and zero matrix , respectively ; @xmath13 stands for the @xmath14 vector with each entry being unit value ; @xmath15 denotes the expectation of a random variable , and @xmath16 represents its estimate .
consider a c - ran consisting of one bbu pool and multiple rrhs depicted in fig .
the rrhs operate in half - duplex modes , and different mss served by the same rru are allocated with a single subcarrier through the orthogonal frequency division multiplexing access ( ofdma ) technique .
it is assumed that mss move continuously , while rrus remain fixed .
thus , the channels of radio als would vary during one transmission block , while those of wireless bls undergo quasi - static flat fading . due to the orthogonality characteristics of ofdma for the accessing of multiple mss , we can focus on the transmission of only a single ms .
let @xmath17 and @xmath18 denote the data vector and cyclical training sequence transmitted from the ms , respectively . the training sequence from the rrh is denoted by @xmath19 .
the @xmath20-th channel sample of the time - varying radio al is denoted by @xmath21 with mean zero and variance @xmath22 , while the channel fading of the quasi - static flat bl is denoted by @xmath23 which has the complex gaussian distribution of mean zero and variance @xmath24 .
the transmit power of the ms and rrh are denoted by @xmath25 and @xmath26 , respectively .
the noise variance at the rrh and bbu pool is denoted by @xmath27 .
it is assumed that the bbu pool acquires the knowledge of @xmath18 , @xmath19 , @xmath22 , @xmath24 , @xmath25 , @xmath26 , and @xmath27 . during each transmission block ,
the ms transmits a signal @xmath28 of @xmath29 symbol length to the rru , in which the @xmath20-th entry of @xmath28 is given by @xmath30 where @xmath31 denotes the @xmath20-th entry of @xmath17 with m - ary phase shift keying ( mpsk ) modulation constrained by @xmath32 , and @xmath33 represents the @xmath20-th entry of @xmath18 with @xmath34 whose period is denoted by @xmath35 .
the @xmath36 is within @xmath37 . without loss of generality
, we further assume that @xmath38 is an integer@xcite .
the @xmath20-th observation at the rrh is written as @xmath39 where @xmath40 is additive white gaussian noise ( awgn ) at the rrh .
then the rrh scales the received signal by @xmath41 , and inserts @xmath19 prior to the received signal .
the sequence @xmath19 is of @xmath42 length and its @xmath20-th entry satisfies @xmath43 .
the bbu pool receives two separate signals as @xmath44 where @xmath45^{t}$ ] , @xmath46^{t}$ ] and @xmath47^{t}$ ] are awgn vectors with each entry having variance @xmath27 . in order to perform coherent reception and adopt cooperative processing techniques at the bbu pool ,
the knowledge of @xmath21 s and @xmath23 should be obtained , which is explained in the following section .
the ce - bem for time - selective but frequency - flat fading channels in @xcite@xcite is chosen to model the time - varying al as @xmath48 where @xmath49 s are the bem coefficients that remain invariant within one transmission block , and @xmath49 has the complex gaussian distribution with mean zero and variance @xmath50 constrained to @xmath51 . substituting into , @xmath52 can be rewritten as @xmath53\boldsymbol{\lambda}\!+\!\alpha g \mathbf{w}_{r}\!+\!\mathbf{w}_{ds},\end{aligned}\ ] ] where @xmath54 $ ] is an @xmath55 dimensional matrix with @xmath56 . by defining @xmath57 of @xmath58 dimension
, we can obtain _
q_s= _ s , & @xmath59 , + 0 , & @xmath60 , where @xmath61 is an @xmath62 dimensional vector . left multiplying @xmath63 by @xmath64 yields @xmath65 whose @xmath66$]-th to @xmath67$]-th entries denoted by @xmath68 can be expressed as @xmath69 it is testified that @xmath70 for @xmath71 with @xmath72 . although @xmath73 is not a gaussian vector , it is effective to use the gaussian distribution to model the noise behavior for estimation problems @xcite .
thus , we choose the following function to be the nominal likelihood function of @xmath74 as @xmath75 defining @xmath76^{t}$ ] , the map estimation for @xmath77 gives @xmath78 where @xmath79 , @xmath80 and @xmath81 are gaussian distribution functions . with a given @xmath23 , the estimate of @xmath49 can be obtained as @xmath82 substituting into , the estimate of @xmath23 can be obtained from @xmath83 note that , only the first term in @xmath84 relates to the phase of @xmath23 denoted by @xmath85 , and thus the estimate of @xmath85 can be directly estimated from minimizing @xmath86 as @xmath87 the estimate of @xmath88 must be either a local minima of @xmath89 or at the boundary @xmath90 , which can be obtained from solving @xmath91 .
unfortunately , a closed - form expression for @xmath92 is hard to derive since @xmath93 is an @xmath8(@xmath94)th - order polynomial of @xmath88 , and thus numerical methods such as one dimensional search are needed to compute the value of @xmath95 . to reduce the complexity of such approaches , an iterative approach is developed whereby @xmath96
is initialized from @xmath97 with @xmath96 obtained , @xmath98 s can be estimated according to with @xmath96 in place of @xmath23 .
then , @xmath96 can be further updated by substituting @xmath98 s into as @xmath99 with @xmath98 s obtained , @xmath100 is restored as @xmath101 where @xmath102 s are real factors .
the vector @xmath103 that maximizes the average effective snr ( aesnr ) @xcite denoted by @xmath104 is obtained from @xmath105 define @xmath106 , and denote @xmath107 and @xmath108 to be the estimation error of @xmath23 and @xmath49 , respectively .
it is obtained that @xmath109 note that , @xmath110 while the last term in has the order of @xmath111 $ ] , thus we remove the last term for high snr approximation , i.e. , @xmath112 similarly , @xmath113 due to the non - linearity of the map estimation , it is hard to derive the corresponding mse expressions in closed forms .
moreover , it is known that the map estimation mses converge to the cramr - rao bound ( crb ) @xcite when the training length is sufficiently large , and thus it is effective to use the crbs in the computation of the aesnr as @xmath114 substituting the above expressions for @xmath115 and @xmath116 into @xmath117 yields @xmath118 where @xmath119\!\frac{\sigma_{n}^{2}}{p_{s}},\end{aligned}\ ] ] and @xmath120 .
on setting @xmath121 , the optimization for @xmath103 transforms to @xmath122 clearly , the optimization problem described in constrained by is concave , and @xmath104 can be directly obtained from the lagrange dual function as @xmath123 substituting back to , the optimization problem becomes @xmath124 whose solution is @xmath125 leading to @xmath126 combining the estimation for @xmath77 and the restoration for @xmath21 s , the proposed channel estimation algorithm is summarized in table i. moreover , the following proposition is given to demonstrate the effectiveness of the proposed algorithm : the iterative channel estimation algorithm is convergent , and it achieves lower mse than that of the maximum likelihood ( ml ) method .
each iteration consists of @xmath127 steps .
denote the @xmath128th entry of @xmath77 by @xmath129 , and the updated estimate of @xmath129 denoted by @xmath130 satisfies @xmath131 .
this indicates that @xmath132 strictly increases after each step as well as after one round of iteration .
it is known that @xmath133 is continuous with respect to @xmath129 and @xmath134 .
thus , it is concluded that the iterative algorithm is convergent . from ,
the map estimate of @xmath23 with a given @xmath135 is @xmath136 whose mse is calculated as @xmath137 the ml estimate of @xmath23 gives @xmath138 whose mse is calculated as @xmath139 clearly , @xmath140 always holds , and it can be also testified that @xmath141 is satisfied similarly . _ remark _ : according to and , we see that @xmath142 and @xmath143 .
this is because the proposed map estimation algorithm is biased since @xmath144 .
p3.4in@ * * initialize * @xmath96 in accordance with ; i_index@xmath145i_times(@xmath146 ) * * repeat * * * for each @xmath147 , update @xmath98 by substituting @xmath96 into . *
* update @xmath96 by substituting @xmath98 s into . * * i_index@xmath145i_index@xmath148 * * until * i_index@xmath149 is satisfied . *
* calculate * the optimal @xmath104 according to . * * restore * @xmath100 s according to by using @xmath98 s and @xmath104 .
* * return * @xmath100 s and @xmath96 .
numerical results are provided to evaluate the performance of the ce - bem based map ( c - map ) channel estimation algorithm . the al channel @xmath150 and bl channel @xmath23 are generated from the spatial channel model ( csm ) in _ 3gpp
tr 25.996 _ @xcite .
the parameters are set as @xmath151 , @xmath152 and @xmath153 .
we assume binary - phase - shift - keying ( bpsk ) modulation for @xmath154 , while @xmath61 is selected as the @xmath155nd column of the @xmath156 discrete fourier transform ( dft ) matrix and @xmath19 is as the @xmath157rd column of the @xmath158 dft matrix .
the transmit power @xmath25 and @xmath26 are set to be equal , and the noise variance is set to be of unit value ; thus the snr is equal to @xmath25 . in fig .
[ fig2 ] , the average mses of both al and bl channels for c - map estimation are compared with that for ml estimation . it is observed that the proposed c - map estimation outperforms the traditional ml method since c - map achieves lower mses of both al and bl channels than ml . moreover , it is seen that the mse of the al channel for the ml method is not convergent .
this is because random generation of @xmath23 would result in singularity , e.g. , @xmath159 , leading to @xmath160 for the ml method , while the proposed c - map algorithm is robust at the singularity . in fig.[fig3 ] , we evaluate the aesnr performance for the optimal weighted approach ( owa ) to channel restoration .
it is seen that owa obtains higher aesnrs than the baseline ( restoring according to ce - bem ) , especially in the low snr region , and it draws near to the baseline as snr increases .
a superimposed - segment training design has been proposed to decrease the training overhead and enhance the channel estimation accuracy in uplink c - rans . based on the training design ,
a ce - bem based map channel estimation algorithm has been developed , where the bem coefficients of the time - varying radio al and the channel fading of the quasi - static wireless bl are first obtained , then the time - domain channel samples of al are restored from maximizing aesnr .
simulation results have demonstrated that the proposed algorithm declines the estimation mse and increases the aesnr .
s. zhang , f. gao , c. pei and x. he , segment training based individual channel estimation in one - way relay network with power allocation , _ ieee trans .
wireless commun .
3 , pp . 13001309 , mar .
2013 .
g. dou , c. he , c. li and j. gao , a weighted first - order statistical method for time - varying channel and dc - offset estimation using superimposed training , _ ieee commun .
5 , pp . 852855 , may 2013 . | to decrease the training overhead and improve the channel estimation accuracy under time - varying environments in uplink cloud radio access networks ( c - rans ) , a superimposed - segment training design is proposed whose core idea is that each mobile station puts a periodic training sequence on the top of the data signal , and remote radio heads ( rrhs ) insert a separate pilot prior to the received signal before forwarding to the centralized base band unit ( bbu ) pool . moreover , a complex - exponential basis - expansion - model ( ce - bem ) based maximum a posteriori probability ( map ) channel estimation algorithm is developed , where the bem coefficients of access links ( als ) and the channel fading of wireless backhaul links ( bls ) are first obtained , after which the time - domain channel samples of als are restored in terms of maximizing the average effective signal - to - noise ratio ( aesnr ) .
simulation results show that the proposed channel estimation algorithm can effectively decrease the estimation mean square error and increase the aesnr in c - rans , thus significantly outperforming the existing solutions .
channel estimation , cloud radio access networks , time - varying environment . |
confinement of water in nanopores affects many properties of water , such as freezing temperature , crystal structure @xcite , the glass transition temperature , and the position of the hypothesized liquid - liquid ( ll ) critical point @xcite . indeed ,
water confined in nanoscale geometries has received much recent attention , in part because of its importance in biology , engineering , geophysics and atmospheric sciences .
the effects of different kinds of confinement have been studied , both using experiments and simulations @xcite . bulk supercooled water
water cooled below the equilibrium freezing temperature shows many anomalous properties @xcite .
experiments find that at low temperatures , various response functions , such as isothermal compressibility and specific heat , increase sharply .
there has been comparatively less research on confined water .
using the st2 potential to model water confined between smooth plates @xcite , a ll phase transition has been proposed .
a liquid - to - amorphous transition is seen in simulations of water using the tip4p potential @xcite confined in carbon nanotubes @xcite .
recent theoretical work @xcite suggests that hydrophobic lennard - jones ( lj ) confinement shifts the ll transition to lower temperature and lower pressure compared to bulk water , a feature also found in simulations of water confined between hydrophobic plates @xcite . confinement is known to enhance solidification of molecules that are more or less spherical @xcite .
however , careful experiments on thin films of water show that water performs extremely well as a lubricant , suggesting that confined water may be more fluid than bulk water @xcite .
recent experiments show that water in hydrophilic confinement , when cooled to very low @xmath1 , does not freeze @xcite a phenomenon also supported by simulation studies @xcite . in contrast , simulations @xcite show that hydrophobically confined water does freeze into different crystalline structures , which do not have counterparts in bulk water .
indeed monolayer , bilayer and trilayer ice have all been found in simulations @xcite .
thus hydrophobic confinement seems to facilitate the freezing of water .
however the reason for this facilitation is not yet fully understood .
the hydrogen - bond interaction between water molecules is an order of magnitude stronger than van der waals attraction with the hydrophobic walls .
thus one may hypothesize that freezing in hydrophobic confinement depends critically on the separation between confining walls that may distort or facilitate a particular crystalline structure , rather than on the weak details of the water - wall interaction potential . to test this hypothesis
we perform molecular dynamics ( md ) simulations of water in two different forms of repusive confinement .
specifically , we study : * the @xmath0 repulsive part of the lj potential studied in ref .
@xcite . * the same potential used in ref .
@xcite but truncated and shifted such that there is no attractive part in the potential , analogous to the weeks - chandler - anderson ( wca ) potential @xcite .
this potential allows us to examine the role , if any , the attractive part of the water - wall lj potential plays in determining the thermodynamics and structure of confined water .
we compare the case when the water - wall interactions are purely repulsive ( `` repulsive confinement '' ) with the studied case of pure lj confinement @xcite .
we also compare the freezing in repulsive confinements with the freezing found when the water - wall interactions are represented by an lj interaction ( `` lj hydrophobic confinement '' ) @xcite .
this paper is organized as follows : in sec .
ii , we provide details of our simulations and analysis methods .
simulation results for the liquid state are provided in sec .
iii . in sec .
iv we discuss the freezing properties of our system .
we perform md simulations of a system composed of water - like molecules confined between two smooth walls .
the molecules interact via the tip5p pair potential @xcite which , like the st2 @xcite potential , treats each water molecule as a tetrahedral , rigid , and non - polarizable unit consisting of five point sites @xcite .
the tip5p potential predicts many of the anomalies of bulk water @xcite .
for example , tip5p reproduces the density anomaly at @xmath2 k and @xmath3 atm and its structural properties compare well with experiments @xcite .
tip5p is known to crystallize at high pressures @xcite within accessible computer simulation time scales , and shows a `` nose - shaped '' curve of temperature versus crystallization time @xcite , a feature found in experimental data on water solutions @xcite . in our simulations ,
@xmath4 water molecules are confined between two smooth planar walls , as shown schematically in fig .
[ fig : schematic ] .
the walls are located at @xmath5 nm ( wall - wall separation of @xmath6 nm ) , which results in @xmath7 layers of water molecules .
periodic boundary conditions are used in the @xmath8 and @xmath9 directions , parallel to the walls .
we study two different forms of purely repulsive water - wall interaction .
the first uses only the @xmath10 repulsive core , which we call the @xmath0 repulsive potential , @xmath11 .
\label{udz}\ ] ] here @xmath12 is the distance from the oxygen atom of a water molecule to the wall , while @xmath13 kj / mol and @xmath14 nm are potential parameters ( fig .
[ fig : potentials ] ) . similar but different parameter values were used in previous confined water simulations using the tip5p interaction potential @xcite .
specifically , in ref .
@xcite , the water - wall interaction was modelled using a 9 - 3 lj potential with @xmath15 kj / mol and @xmath16 nm . we choose a different @xmath17 in the case of repulsive confinement so that the repulsion between the water and wall decays to almost zero where the 9 - 3 lj -potential has a minimum .
the second purely repulsive potential uses both attractive and repulsive terms of the 9 - 3 lj potential , but truncates and shifts the potential at the position of the minimum to create a repulsive potential that exactly mimics the repulsion of ref .
@xcite , in analogy to the wca potential , @xmath18 + \frac{8\epsilon_{\rm ow}}{3^{3/2 } } & \textrm{if $ |z - z_w| < 3^{1/6}\sigma_{\rm ow}$}\\ 0 & \textrm{if $ |z - z_w| > 3^{1/6}\sigma_{\rm ow}$},\\ \end{array } \right.\ ] ] where @xmath15 kj / mol and @xmath16 nm . for the @xmath0 repulsive potential , we perform simulations for @xmath19 state points , corresponding to seven temperatures @xmath20 k , @xmath21 k , @xmath22 k , @xmath23 k , @xmath24 k , @xmath25 k , and @xmath26 k , and eight `` geometric densities '' @xmath27 g/@xmath28 , @xmath29 g/@xmath28 , @xmath30 g/@xmath28 , @xmath31 g/@xmath28 , @xmath32 g/@xmath28 , @xmath33 g/@xmath28 , @xmath34 g/@xmath28 , and @xmath35 g/@xmath28 the same as studied in ref .
the geometric values of density do not take into account the fact that the repulsive interactions of molecules with the walls increases the overall amount of available space , since the @xmath36 parameter of the @xmath0 repulsive potential is smaller than the @xmath36 used for the lj confined system [ fig : potentials ] . for systems confined by lj interactions , there is a well - defined preferred distance from the wall , making it relatively straightforward to evaluate the `` effective '' density of molecules confined by the attractive wall . in our system with only repulsive interactions , there is no such preferred distance , as emphasized by fig .
[ fig : potentials ] .
we can approximate the effective density by examining the local density @xmath37 ( fig .
[ fig : rhoz ] ) .
we utilize the fact that @xmath37 has an inflection , and estimate the effective @xmath38 by the location where the second derivative of @xmath39 , or when first derivative of @xmath37 has a maximum .
we must also add to this value of @xmath38 the molecular diameter of water(0.278 nm ) to calculate the real space available along @xmath40-direction .
the resulting `` effective densities '' are @xmath41 g/@xmath28 , @xmath42 g/@xmath28 , @xmath43 g/@xmath28 , @xmath44 g/@xmath28 , @xmath45 g/@xmath28 , @xmath46 g/@xmath28 , @xmath47 g/@xmath28 , and @xmath48 g/@xmath28 .
we will use these effective densities throughout the paper , since they will be most comparable to the effective densities with lj confinement .
for the wca potential , we perform simulations for 32 state points , corresponding to 8 different temperatures and 4 different `` geometric densities '' , @xmath49 g/@xmath28 , @xmath29 g/@xmath28,@xmath30 g/@xmath28,@xmath31 g/@xmath28 respectively .
these geometric densities correspond to `` effective densities '' @xmath50 g/@xmath28 , @xmath51g/@xmath28 , @xmath52g/@xmath28 , @xmath53g/@xmath28 respectively .
note that these effective densities were calculated using the method described in @xcite .
we control the temperature using the berendsen thermostat with a time constant of @xmath54 ps @xcite and use a simulation time step of @xmath55 fs , just as in the bulk system @xcite .
water - water interactions are truncated at a distance @xmath56 nm as discussed in ref .
one of the defining characteristics of water is the existence of a temperature of maximum density ( tmd ) . relative to bulk water lj confinement shifts the locus of the tmd to lower @xmath1 by @xmath57 k @xcite .
additionally , the sharpness of the density maximum is markedly decreased in comparison to the bulk .
[ fig : pxy - comp ] shows isochores of @xmath58 for lj confinement , @xmath0 repulsive confinement and wca confinement , for similar densities .
a tmd in this plot is coincident with the minimum in the isochore . for @xmath0 repulsive confinement
, the minimum is very weak , but the location of the flatness in the isochore is near to that of the system with lj confinement .
this result suggests the the @xmath0 repulsive confinement further suppresses the structural ordering of the molecules that is known to be responsible to the presence of a density maximum .
the tmd for the case of wca confinement again appears at the same @xmath1 as the @xmath0 repulsive confinement and lj confinement cases but the isochore in the tmd region is flatter than for case the of lj confinement confinemen .
hence both kinds of the repulsive confinement supprsess the structural ordering in lateral directions compared to the case of bulk and lj confinement .
we further notice that the value of the lateral pressure @xmath59 in the case of the lj confinement approaches the value of @xmath59 in case of wca confinement at high temperatures .
this behavior of @xmath59 for lj confinement should be expected since at very high temperatures the molecules will not feel the potential minimum of the water - wall interaction . in order to compare the structural properties of repulsive confinement with those of lj confinement , we calculate the lateral oxygen - oxygen radial distribution function ( rdf ) defined by @xmath60.\ ] ] here @xmath61 is the volume , @xmath62 is the distance parallel to the walls between molecules @xmath63 and @xmath64 , @xmath65 is the @xmath40-coordinate of the oxygen atom of molecule @xmath63 , and @xmath66 is the dirac delta function . the heaviside functions , @xmath67 , restrict the sum to a pair of oxygen atoms of molecules located in the same slab of thickness @xmath68 nm .
the physical interpretation of @xmath69 is that @xmath70 is proportional to the probability of finding an oxygen atom in a slab of thickness @xmath71 at a distance @xmath72 ( parallel to the walls ) from a randomly chosen oxygen atom . in a bulk liquid
, this would be identical to @xmath73 , the standard rdf .
figure [ grxy ] shows the temperature and density dependence of the lateral oxygen - oxygen pair correlation function for both @xmath0 repulsive ( fig .
[ grxy]a , fig .
[ grxy]b ) and wca ( fig .
[ grxy]c , fig . [ grxy]d ) confinements . for both repulsive confinements ,
the qualitative behavior of the dependence of @xmath74 is the same . at low temperature and low density , the first two peaks in @xmath75
appear at @xmath76 and @xmath77 , but at high densities the second peak moves to a larger distance .
this behavior is nearly identical to that observed for water confined between lj surfaces , and is discussed in detail in ref .
@xcite .
we also confirm the structural similarity with lj confinement by calculating the lateral static structure factor @xmath78 , defined as the fourier transform of the lateral rdf @xmath69 , @xmath79 here the @xmath80-vector is the corresponding wave vector in the @xmath81 plane and @xmath72 is the projection of the position vector on the @xmath81 plane . in fig .
[ fig : sqxy ] , we show the temperature and pressure dependence of lateral structure factors for both repulsive confinements . for both forms of repulsive confinement ,
the temperature and density dependence of @xmath82 is similar .
we find that confined water has a weaker pre - peak at @xmath83nm@xmath84 compared to bulk water ( fig .
[ fig : sqxy - comp ] ) , consistent with the possibility that the local tetrahedrality is weakened by repulsive confinements . of the three forms of confinement ,
the @xmath82 for lj confinement is most like bulk water ( fig .
[ fig : sqxy - comp ] ) .
local tetrahedrality becomes weaker in case of repulisive confinements compared to lj confinement .
further , we see that the water in @xmath0 repulsive confinement is less structured in lateral directions compared to water in wca confinement , indicated by a less sharp and broad pre - peak at @xmath83 nm@xmath84 in @xmath78 ( fig .
[ fig : sqxy - comp ] ) .
bulk tip5p water crystallizes within the simulation time for @xmath85 g/@xmath28 at low temperatures @xcite .
crystallization of confined water is seen in some simulations @xcite .
a similar crystallization appears in simulations when an electric field is applied in lateral directions to a system of water confined between silica walls @xcite . at plate separation of 1.1 nm with hydrophobic lj confinement
, water crystallizes to trilayer ice @xcite . from our simulations of tip5p water in repulsive confinements with the same plate separation of 1.1 nm
, we find that the system does not freeze within accessible simulation time scales for @xmath0 repulsive confinement ; however the system freezes for wca confinement . as a more stringent confirmation of this fact
, we also use a starting ice configuration obtained from simulations with lj confinement for the same thickness , and confirm that the ice melts to a liquid with @xmath0 repulsive confinement . in fig .
[ fig : melting - repulsive ] , we show the evolution of potential energy and lateral structure factor with time , when the crystal formed in lj confinement @xcite is kept between the @xmath0 repulsive walls .
the potential energy first increases and then reaches its equilibrium value of the liquid potential energy accompanied by a structural change from a crystal ( presence of sharp bragg peaks ) to a liquid ( absence of bragg peaks ) . based on this observation , it is tempting to claim that repulsion inhibits crystallization , and that a preferable distance from the wall determined by the attractive portion of the lj potential is necessary to induce crystallization
. however , as discussed above for the same plate separation @xmath6 nm , @xmath0 repulsive confinement with the chosen parameters increases the available space for molecules relative to lj confinement . hence to properly compare the crystallization behavior
, we must adjust the separation of the wall so that the available space for the water molecules is the same in both systems .
we can make the available space the same by tuning the separation of the plates or by tuning the potential . by tuning the parameters ( see fig .
[ fig : potentials ] , where the values of parameter @xmath17 and @xmath86 for the modified @xmath0-repulsive potential are @xmath87 kj / mol and @xmath88 nm respectively ) of the @xmath0repulsive potential in the repulsive system such that the density profile along the @xmath40-axis becomes similar , we have identical values of the available space between the plates ( see fig . [
fig : rhoz - comp ] ) , and we find that an initial crystal configuration does _ not _ melt , emphasizing that the presence of the crystal is very sensitive to density _ and _ to plate separation since the separation determines the accessible packing arrangements between the plates . similar sensitivity to plate separation for monolayer ice was seen in ref .
@xcite .
in addition to examining the stability of initially crystalline structures , we also consider whether freezing from the liquid state occurs when we have the same effective plate separation .
we find that for the repulsively confined systems , the crystal will also spontaneously form if the available space between the plates is the same as that for which initially crystalline configurations are stable .
hence plate separation appears to be the dominant cause in determining whether or not a crystal will form .
we thank f. sciortino for helpful discussions , and the nsf for support under grants che 0096892 , che 0404699 , and dmr-0427239 .
we also thank the boston university computational facility for computational time and the office of academic affairs at yeshiva university for support .
two positive point charges of charge @xmath89 ( where @xmath90 is the fundamental unit of charge ) are located on each hydrogen atom at a distance @xmath91 nm from the oxygen atom ; together they form an @xmath92 angle of @xmath93 .
two negative point charges ( @xmath94 ) representing the lone pair of electrons ( @xmath95 ) are located at a distance @xmath96 nm from the oxygen atom .
these negative point charges are located in a plane perpendicular to the @xmath92 plane and form an @xmath97 angle of @xmath98 , the tetrahedral angle . to prevent overlap of molecules ,
a fifth interaction site is located on the oxygen atom , and is represented by a lj potential with parameters @xmath99 nm and @xmath100 kj / mol .
p. kumar , s. v. buldyrev , and h. e. stanley , in _ soft matter under extreme pressures : fundamentals and emerging technologies _ editors : sylwester j. rzoska , victor mazur [ proc .
nato arw , odessa , oct 2005 ] ( springer , berlin , 2006 ) . along z - direction for four different bulk densities at t=@xmath23 k for @xmath0 repulsive confinement .
( b ) density profile @xmath102 along z - direction for four different bulk densities at t=@xmath103 k for the case of wca confinement .
both the repulsive confinements show similar layering of water molecules as seen in the case of lj confinement @xcite.,title="fig:",width=377 ] along z - direction for four different bulk densities at t=@xmath23 k for @xmath0 repulsive confinement .
( b ) density profile @xmath102 along z - direction for four different bulk densities at t=@xmath103 k for the case of wca confinement .
both the repulsive confinements show similar layering of water molecules as seen in the case of lj confinement @xcite.,title="fig:",width=377 ] for one isochore for the purely repulsive , lj confinement and wca cases . here
the effective density is @xmath104 g/@xmath28 for the @xmath0 repulsive potential and @xmath105 g/@xmath28 for lj confinement .
these effective densities for both systems correspond to the same geometric density of @xmath106 g/@xmath28 .
all forms of confinement show a tmd , indicated by the minimum of the pressure ; however the tmd is very `` flat '' for @xmath0 repulsive and wca confinement .
as expected the value of @xmath59 approaches the value of @xmath59 for the case of lj confinement at high temperatures.,width=377 ] for the case of @xmath0 repusive and wca confinements . shown
are four different densities at two fixed temperatures ( a ) @xmath20 k and ( b ) @xmath107 k for @xmath0 repulsive confinement and ( c ) and ( d ) for wca confinement .
note that with increasing density , the second neighbor peak at @xmath108 nm becomes less pronounced and at high temperatures moves to a larger distance seen in lj confinement @xcite .
, title="fig:",width=302 ] for the case of @xmath0 repusive and wca confinements . shown
are four different densities at two fixed temperatures ( a ) @xmath20 k and ( b ) @xmath107 k for @xmath0 repulsive confinement and ( c ) and ( d ) for wca confinement .
note that with increasing density , the second neighbor peak at @xmath108 nm becomes less pronounced and at high temperatures moves to a larger distance seen in lj confinement @xcite .
, title="fig:",width=302 ] for the case of @xmath0 repusive and wca confinements . shown
are four different densities at two fixed temperatures ( a ) @xmath20 k and ( b ) @xmath107 k for @xmath0 repulsive confinement and ( c ) and
( d ) for wca confinement . note that with increasing density , the second neighbor peak at @xmath108 nm becomes less pronounced and at high temperatures moves to a larger distance seen in lj confinement @xcite .
, title="fig:",width=302 ] for the case of @xmath0 repusive and wca confinements . shown
are four different densities at two fixed temperatures ( a ) @xmath20 k and ( b ) @xmath107 k for @xmath0 repulsive confinement and ( c ) and ( d ) for wca confinement . note that with increasing density , the second neighbor peak at @xmath108 nm becomes less pronounced and at high temperatures moves to a larger distance seen in lj confinement @xcite .
, title="fig:",width=302 ] for the case of @xmath0 repulsive and wca confinements .
shown are four different densities at two fixed temperatures ( a ) @xmath20 k and ( b ) @xmath109 k for @xmath0 repulsive confinement and ( c ) and ( d ) for wca confinement for the same temperatures .
the first peak of @xmath78 corresponding to the hydrogen - bonds weakens as density is increased and is absent at high densities and high temperatures .
the first peak of @xmath110 in the case of @xmath0 repulsive confinement weaker than the lj confinement.,title="fig:",width=302 ] for the case of @xmath0 repulsive and wca confinements . shown
are four different densities at two fixed temperatures ( a ) @xmath20 k and ( b ) @xmath109 k for @xmath0 repulsive confinement and ( c ) and ( d ) for wca confinement for the same temperatures .
the first peak of @xmath78 corresponding to the hydrogen - bonds weakens as density is increased and is absent at high densities and high temperatures .
the first peak of @xmath110 in the case of @xmath0 repulsive confinement weaker than the lj confinement.,title="fig:",width=302 ] for the case of @xmath0 repulsive and wca confinements . shown are four different densities at two fixed temperatures ( a ) @xmath20 k and ( b ) @xmath109 k for @xmath0 repulsive confinement and ( c ) and ( d ) for wca confinement for the same temperatures .
the first peak of @xmath78 corresponding to the hydrogen - bonds weakens as density is increased and is absent at high densities and high temperatures .
the first peak of @xmath110 in the case of @xmath0 repulsive confinement weaker than the lj confinement.,title="fig:",width=302 ] for the case of @xmath0 repulsive and wca confinements .
shown are four different densities at two fixed temperatures ( a ) @xmath20 k and ( b ) @xmath109 k for @xmath0 repulsive confinement and ( c ) and ( d ) for wca confinement for the same temperatures .
the first peak of @xmath78 corresponding to the hydrogen - bonds weakens as density is increased and is absent at high densities and high temperatures .
the first peak of @xmath110 in the case of @xmath0 repulsive confinement weaker than the lj confinement.,title="fig:",width=302 ] repulsive confinement , lj confinement , wca - confinement are chosen to be 0.950 g/@xmath28 .
we choose the structure factor for bulk water at density 1.00 g/@xmath28 .
a diminished peak at @xmath83 nm@xmath84 shows that the local tetrahedral structure is weakened in case of all forms of confinement .
a further comparison of lj confinement at 0.950 g/@xmath28 with @xmath0 repulsive confinement shows that water in @xmath0 repulsive confinement is less tetrahedral , as the first peak of @xmath78 is much weaker than the first peak of @xmath78 for lj confinement .
water in wca confinement is more structured than @xmath0 repulsive confinement , but is less structured than lj confinement.,width=377 ] , when the crystal formed in lj confinement @xcite is kept between the @xmath0 repulsive wall .
( b ) the crystal structure indicated by the sharp bragg peaks melts , and ( c ) turns into a liquid indicated by the absence of bragg peaks.,width=453 ] of water along z - direction for different potentials at the same geometric density .
the repulsive confinement system freezes spontaneously when the parameters of the potential are modified such that the effective @xmath111 calculated from the @xmath102 ( red dotted line ) for the repulsive system is same as that for the lj system ( blue - dashed line ) ( see fig .
[ fig : potentials]).,width=453 ] , for @xmath0 repulsive confinement when the effective @xmath111 is same as the effective @xmath111 of lj confinement at @xmath112 k and geometric density @xmath113 g/@xmath28 .
the confined water spontaneously freezes , indicated by the drop in potential energy .
( b ) the structure factor of the ice such formed resembles the trilayer ice seen in case of lj confinement @xcite.,width=453 ] | much of the understanding of bulk liquids has progressed through study of the limiting case in which molecules interact via purely repulsive forces , such as a hard - core potential . in the same spirit , we report progress on the understanding of confined water by examining the behavior of water - like molecules interacting with planar walls via purely repulsive forces and compare our results with those obtained for lennard - jones ( lj ) interactions between the molecules and the walls . specifically , we perform molecular dynamics simulations of 512 water - like molecules which are confined between two smooth planar walls that are separated by 1.1 nm . at this separation
, there are either two or three molecular layers of water , depending on density .
we study two different forms of repulsive confinements , when the interaction potential between water - wall is ( i ) @xmath0 and ( ii ) wca - like repulsive potential .
we find that the thermodynamic , dynamic and structural properties of the liquid in purely repulsive confinements qualitatively match those for a system with a pure lj attraction to the wall . in previous studies that include attractions , freezing into monolayer or trilayer ice was seen for this wall separation .
using the same separation as these previous studies , we find that the crystal state is not stable with @xmath0 repulsive walls but is stable with wca - like repulsive confinement . however , by carefully adjusting the separation of the plates with @xmath0 repulsive interactions so that the effective space available to the molecules is the same as that for lj confinement , we find that the same crystal phases are stable
. this result emphasizes the importance of comparing systems only using the same effective confinement , which may differ from the geometric separation of the confining surfaces . |
galaxies are traditionally divided in active and inactive , according to whether their nuclei exhibit the telltale signs ( broad emission lines , flux variability , radio compactness , strong emission at high energies ) of a central engine " or not . however , the distinction is not sharp and there is a continuum of properties between these two extremes , with a wide range of nuclear activity levels , that accordingly define different types of active galactic nuclei ( agns ) .
the most powerful among these are luminous emitters over the whole electromagnetic spectrum up to the tev energies , highly variable and highly polarized ( bolometric luminosities can reach @xmath0 erg s@xmath1 and are often dominated , especially during outbursts , by the gamma - ray emission ) . these are called blazars " .
there is unanimous consensus on the fact that blazars owe their extreme physical behavior to the presence of a jet that is closely aligned with the observer s direction ( estimated viewing angles @xmath2 deg ) where the plasma moves with a lorentz is defined as @xmath3 , where @xmath4 , with @xmath5 representing the plasma velocity . ]
factor @xmath6 of the order of @xmath7 , and occasionally as high as @xmath850 @xcite .
the highly relativistic kinematic regime and the small viewing angle produce a doppler aberration ( quantified by the factor @xmath9^{-1}$ ] ) , that foreshortens the observed time - scales , blue - shifts the observed spectrum and magnifies , where @xmath10 , depending on the geometry of the emitting region and on spectral index . ]
the luminosities at all wavelengths ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
direct evidence for this interpretation comes from the detection of superluminal motion of plasma blobs along the jet , the apparent velocity of which , as measured with radio interferometry , can be up to @xmath11 @xcite . when corrected for doppler aberration ,
these velocities are obviously lower than the speed of light , and compatible with a lorentz factor of the order of 10 .
indirect arguments for relativistic beaming include : i ) the super - eddington accretion regimes for putative central black hole masses of @xmath12 m@xmath13 ; ii ) the transparency of the emitting regions to electron - positron pair - production at gamma - ray energies higher than @xmath81 mev @xcite ; iii ) the huge radio brightness temperatures .
if no beaming is invoked then one would expect self - absorbed radio spectra over wide radio wavelength ranges and the compton catastrophe in the x - rays .
none of these effects are observed @xcite .
the interpretation of the blazar class as relativistically beamed objects implies the existence of a parent population of misaligned sources .
based on optical and radio properties and population density , the most favored candidates are radiogalaxies erg s@xmath1 hz@xmath1sr@xmath1 .
their radio morphologies are also different , being core - dominated or lobe - dominated , respectively .
the comparison of radio power and morphology and host galaxy luminosity of blls and radiogalaxies suggests that the dominant parent population consists of fr i radiogalaxies . ]
that would be detected and classified as blazars when viewed at small angles @xcite .
the term _ blazar _ encompasses bright radio - loud agns with compact radio cores , high amplitude multi - wavelength variability and large radio and optical polarization .
historically , those with weak or absent optical spectral lines were named bl lac objects ( see also section [ sec : specscopy ] ) , and the others received definitions that reflected their main properties : optically violently variables , or highly polarized quasars ( e.g. , * ? ? ?
more recently , the distinction within the blazar class was drawn between objects with luminous broad emission lines , often accompanied also by prominent ultraviolet - optical continuum emission ( _ blue bump _ ) of thermal origin as normally seen in quasars , that are called flat spectrum radio quasars ( fsrq ) and blls , where these broad emission lines are weak or absent @xcite .
this distinction is reflected also in the radio polarization properties : at cm wavelengths blls tend to have polarization vectors nearly parallel to the jet , while in fsrqs the vectors cover a wider range of directions , favoring directions perpendicular to the jet @xcite . owing to these different properties , and in particular to the lack of a substantial thermal component ( accretion disk and/or dusty torus ) , blls offer a more direct view into the primary energy - production mechanism , with respect to fsrqs .
this review is organized as follows : in section [ sec : mwlref ] we present the basic multi - wavelength properties and the reference model for blls ; section [ sec : demograph ] reviews the methods of detection and selection of blls through surveys and the luminosity functions ; section [ sec : opticalvar ] reports on the optical , near - infrared and ultraviolet variability of blls ; section [ sec : polarim ] summarizes the optical polarization issues ; the optical spectral properties , both from broad - band photometry and spectroscopy , are described in section [ sec : specscopy ] ; the observations of host galaxies are reviewed in section [ sec : hostgals ] ; a few bll sources exhibited optical counterparts to their radio jets , as detailed in section [ sec : mwljets ] ; the close environments and clustering properties of blls are reported in section [ sec : environs ] ; section [ sec : bhmass ] focuses on the central supermassive black hole , on its influence on the observed properties and on its mass measurement .
section [ sec : conclrem ] contains a summary and future perspectives .
blls are characterized by strong non - thermal emission at all frequencies . in the radio band
, they exhibit bright ( up to @xmath14 erg s@xmath1 hz@xmath1 , * ? ? ?
* ) , compact cores with power - law spectral continua with very flat shapes ( @xmath15 with @xmath16 ) , mostly produced by synchrotron radiation , often partially self - absorbed at the ghz frequencies
. at infrared - optical wavelengths the spectrum of blls also follows a power - law , when corrected for galactic extinction and host galaxy contribution .
the optical / near - infrared spectral indices are in the range @xmath17 , i.e. they are steeper than in the radio @xcite . if the radio - to - optical spectrum is produced by the same optically thin synchrotron radiation component , the spectral steepening between these bands is naturally expected by increasing synchrotron losses at higher frequencies and adiabatic expansion of the jet plasma . the precise spectral slopes and their variability are also regulated by acceleration mechanisms and time - scales @xcite . in the x - rays blls
are powerful sources ( @xmath18 erg s@xmath1 hz@xmath1 ) , and in fact they were detected by many satellites in the last 40 years @xcite .
the x - ray spectra are generally well described by single power - laws or curved slopes , depending on the emission mechanism .
if this is synchrotron radiation , x - rays are generated by the highest energy electrons , so that the spectrum has a relatively steep spectral index ( @xmath19 ) .
if the cooling frequency ( or break frequency ) of the synchrotron spectrum is located in the x - ray range , the spectrum has a convex shape in this band . on the other hand , in a leptonic jet scenario ( which we favour with respect to the hadronic one ,
see below ) , x - rays may be due to compton up - scattering of lower energy photons by the same electrons that produce the synchrotron spectrum ( synchrotron self - compton ) , so that the x - ray spectral shape is flat ( @xmath20 ) . in blls where the synchrotron component peaks below the x - ray range , the x - ray spectrum is occasionally concave and joins smoothly with the high energy tail of the observed synchrotron spectrum .
the observations conducted in the mev - gev range from satellites ( first the _ compton _ gamma - ray observatory in 1991 - 2000 and recently the orbiting _
fermi _ gamma - ray satellite and _ agile _ ) detected large numbers ( @xmath81000 to date ) of extra - galactic sources that turned out to be blazars , with very few exceptions @xcite .
a large subset of them are blls .
the integrated luminosities in the range 0.1 - 10 gev can reach @xmath21 erg s@xmath1 .
very high energy photons ( @xmath22 100 gev ) from blazars were detected from the ground by atmospheric cherenkov radiation telescopes starting about 20 years ago @xcite .
about 50 tev blazars are now known ( see e.g. * ? ? ?
* ; * ? ? ?
? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , many of which have redshift higher than @xmath8 0.2 . a substantial fraction of these are blls , especially of high - frequency - peaked type , in which tev detection is favoured by the broad - band spectral shape ( see below ) . at @xmath23
the extragalactic infrared background suppresses the tev flux via pair production , so that detection beyond this redshift is virtually impossible .
the highest observed luminosities at energies larger than 100 gev are @xmath24 erg s@xmath1 @xcite .
the gamma - ray emission dominates the spectral energy distribution of blazars in general , and bll in particular , especially during flaring states , and is highly variable ( down to sub - hour time - scales , * ? ? ?
* ) . in the leptonic scenario
, this is ascribed to inverse compton scattering of lower energy photons off the most relativistic electrons .
therefore the gamma - ray spectral shape is determined by the high energy tail of the electron distribution and by the shape of the spectrum of the scattered photons , which can be either optical - to - soft - x - ray synchrotron photons ( synchrotron self - compton ) or photons external to the jet , like those from the accretion disk , broad - line region , dust torus ( external compton ) . ) .
panel ( a ) : spectral flux distribution from radio to tev frequencies ; panel ( b ) : @xmath25 representation of the spectrum , that emphasizes the frequencies at which most of the power is emitted . adapted from @xcite . ] in figure [ sed21552006 ] we report an example of an overall spectrum of bll .
the spectral energy distribution has two humps , one peaking at the characteristic synchrotron cooling frequency ( typically between the far - infrared and soft x - rays ) , and a second one attributed to inverse compton cooling in the mev - gev range .
these frequencies vary from object to object in a correlated way : to a larger frequency of the first hump corresponds a larger frequency of the second hump .
accordingly , blls depending on their spectral energy distribution were traditionally divided in low - frequency - peaked objects ( lbls ) and high - frequency - peaked objects ( hbls , * ? ? ?
objects with intermediate properties are called intermediate blls ( ibls ) .
this distinction reflects broadly the previous concept of blazar classification according to radio or x - ray selection .
the characteristic spectral maxima in fact favor naturally the detection of a given blazar in a specific band .
more extended multi - wavelength observations on large datasets led to a more complex picture ( see section [ sec : demograph ] ) .
in figure [ figblazarseqprops ] we show spectral energy distributions of representative blazar sources . while the double - humped shape is present in all of them ,
the frequencies of the spectral peaks and their relative intensities differ significantly .
@xcite found that the spectral energy distribution shapes define a continuum of properties , whereby ( i ) the first peak occurs in different frequency ranges for different samples / luminosity classes , with most luminous sources peaking at lower frequencies ; ( ii ) the peak frequency of the gamma - ray component correlates with the peak frequency of the lower energy one ; ( iii ) the luminosity ratio between the high and low frequency components increases with bolometric luminosity ( fig .
[ figblazarseqprops ] ) . in this picture , called _ blazar sequence _ , blls are the sources with lower luminosities and higher characteristic synchrotron and inverse compton frequencies .
the interpretation of this sequence is complex and somewhat controversial .
the parameters that govern it are the cooling efficiency of the relativistic particles , the accretion efficiency and ultimately the mass of the central black hole , blls being those with lower efficiencies and lower masses @xcite .
a number of objects deviate at first glance from this scheme , having comparatively high total luminosities and high synchrotron peak frequencies @xcite . the central engine that powers the relativistic jet
is assumed to be a rotating accreting black hole of about @xmath26 .
the mechanism of energy extraction and particle acceleration along the jet is however not completely understood ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the dynamics of the jet may be rather complex and it has in some cases been interpreted in terms of helical structures tightly related to the black hole spin or to the orbital motion of a binary black hole system @xcite . on sub - parsec scales ,
jets are pervaded by magnetic fields of the order of 0.1 - 10 gauss that may have tangled geometries and cause particles leptons or hadrons to radiate through the synchrotron mechanism .
as seen above , in the leptonic scenario the main emission mechanisms are synchrotron radiation of relativistic electrons at the radio to x - ray wavelengths , which is responsible for the first hump of the spectral energy distribution , and compton up - scattering of synchrotron photons ( self - compton ) or photon fields external to the jet at the higher energies , responsible for the second hump @xcite . for self - compton scattering ( that is the most common case in blls )
, klein - nishina suppression @xcite may reduce the luminosity of the second hump and decrease its peak frequency . in hadronic models , both primary electrons and protons
are accelerated to ultrarelativistic energies , with protons exceeding the threshold for photo - pion production on the soft photon field in the emission region . while the low - frequency emission is still due to synchrotron emission from primary electrons , the high - energy emission is dominated by proton synchrotron emission , neutral pion decay photons , synchrotron and compton emission from secondary decay products of charged pions , and the output from pair cascades initiated by these high - energy emissions intrinsically absorbed by photon - photon pair production @xcite .
in order to understand the cosmological evolution of blls in comparison with other agn and derive the properties of the unbeamed parent population it is mandatory to construct statistically significant samples of sources with measured redshift .
statistical samples are ultimately extracted from unbiased surveys carried out at various wavelengths .
depending on the adopted band , objects identified as bll candidates , and then confirmed through the essential criterion of optical spectroscopy , are reported usually as radio , x - ray , optically , and also gamma - ray - selected objects ( see section [ sec : mwlref ] ) .
the characteristic double - peaked spectral energy distribution of blls ( see fig .
[ figblazarseqprops ] ) emphasizes the effects of single band flux - limited surveys . as an example
, radio surveys preferentially select the sources with a low - frequency - peaked synchrotron component . on the other hand , in gev surveys
the objects with high - frequency inverse compton component are favoured .
the first complete flux - limited radio - selected sample of blls was constructed from the 1 jy catalogue of radio sources of @xcite containing 518 sources with s@xmath27 1 jy at 5 ghz .
the selection of blls from this catalogue was performed according to various properties ( flat or inverted radio spectrum , optical magnitude and presence of weak spectral lines ) .
this selection produced a list of 34 radio - selected blls @xcite .
the number counts derived from this sample ( see fig .
[ fig : ncounts ] , left ) showed a roughly euclidean behavior ( no evolution ) , a result that was then later confirmed by more extended and complete studies ( e.g. * ? ? ? * and references therein ) .
deep and systematic surveys undertaken in the 80 s with the then orbiting x - ray satellites discovered blls and thus contributed to the construction of complete samples of blls selected on the basis of their x - ray emission . from the extended medium sensitivity survey of the _ einstein _ x - ray satellite ( 0.33.5 kev , sensitivity @xmath28 erg s@xmath1 @xmath29 ,
* ) @xcite extracted a sample of 22 x - ray - selected blls .
the _ einstein _
slew survey @xcite covered a much larger fraction of the sky with shallower sensitivity ( @xmath30 erg s@xmath1 @xmath29 ) .
a result of this survey is a sample of 48 blls @xcite .
further x - ray surveys with _
rosat _ and _ swift _ allowed the discovery of about 100 new blls @xcite . in the optical band
the selection of blls is more tricky .
the first blls discovered in the optical were derived from the search of ultraviolet excess stellar objects in the palomar plates @xcite .
this resulted in the discovery of 7 objects @xcite .
most fruitful recent optical searches of blls resulted from the analysis of pure optical spectra of sdss sources .
@xcite , starting from the 7th sdss spectroscopic data base @xcite present a sample of @xmath8700 bll optically selected candidates .
a large fraction of the objects in the survey are of unknown redshift .
search for radio counterparts of these candidates showed that @xmath890% of these are radio - loud .
owing to the blazar spectral energy distribution , that extends to very high energies , surveys in the gev band turned out to be extremely efficient in finding blls , thanks to their regular and uniform sky coverage . in the 2nd fermi lat agn catalog @xcite , which contains @xmath81000 sources , 80% are blazars , of which about half are blls
the majority of the remaining 20% are blazar candidates . in the selection of candidates , near -
infrared observations from the ground have been used in the past ( see e.g. * ? ? ?
* ; * ? ? ?
* ) , but particularly effective were proven to be the spectra obtained from satellites ( e.g. _ spitzer , _ * ? ? ?
* ) , and especially the infrared colors from _ wise _ observations , because gamma - ray blazars detected by _ fermi _ occupy a distinctive strip in the 3.4 - 4.6 , 4.6 - 12 @xmath31 m color - color diagram @xcite .
these candidates must be eventually confirmed via optical spectroscopy . at variance with most agns ,
since blls are strong emitters at any wavelength , they can be detected in surveys performed in various spectral bands . moreover , owing to the high - amplitude flux variability they may appear in a survey depending on their activity state .
modern bll catalogs are based on multi - wavelength data and reflect more accurately the full range of their spectral and timing properties @xcite
. the knowledge of the luminosity function of blls is instrumental to quantifying the beaming and thus to identifying the parent population .
its construction and evolution ( redshift dependence ) are more complicated than for other classes of agn , because of the difficulty of distance determination , and of the consequences of the radiation beaming @xcite .
the luminosity functions of blls at radio and x - ray frequencies had been evaluated , often with the purpose to compare them with those of radiogalaxies , in search of similarities that could prove the direct link between the two classes , under the assumption that they only differ in the viewing angle .
a common result of these efforts is that blls and fsrqs can be plausibly connected to fr i and fr ii radiogalaxies , respectively @xcite , a conclusion that was independently reached also based on their similar host galaxies and environments ( see also sections [ sec : hostgals ] and [ sec : environs ] ) , emission line intensities , and radio morphologies ( for a review see * ? ? ?
the detailed comparison of radio and optical properties of blls and of their alleged parent population ( radiogalaxies ) might require more complex jet geometries ( e.g. structured jets , * ? ? ?
* ; * ? ? ?
the partial and sometime contradictory results obtained from the analysis of luminosity functions are plagued by the limited size of the original radio and x - ray samples ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the current surveys in gamma - rays yield much more numerous samples . in their compilation of blls based on the first year of _ fermi
_ operations @xcite @xcite attempt to overcome the previous limitations by including blls with spectroscopic redshifts and blls with matching redshift estimates from broad - band photometry and intervening absorption systems spectroscopy ( fig . [
fig : ncounts ] , right ) .
using this complete sample of 211 blls , they construct accurate gamma - ray luminosity functions of these sources .
their tentative outcome is that positive evolution is suggested for the majority of sources at luminosities larger than @xmath32 erg s@xmath1 , while those at low luminosities may behave differently .
their comparison with fsrqs @xcite suggests that the local ( @xmath33 ) luminosity function of blls overlaps and connects smoothly to that of fsrqs , highlighting the similarity between the two classes , with blls having on average lower luminosity ( and thus very likely lower lorentz factors ) than fsrqs .
@xcite also propose that blls correspond to the final ( gas - starved , inefficiently accreting ) and long - lasting phase of an earlier , short - lived , merger - driven , gas - rich epoch , represented by the fsrqs .
a confirmation of this transition scenario can only rely on a more robust beaming correction and knowledge of the black hole masses and host galaxy environments , which are at present not well constrained .
one of the distinctive properties of blls is high amplitude flux variability ( up to 5 magnitudes ) over a range of time - scales from months down to minutes @xcite .
variability is present over the whole near - infrared - to - ultraviolet range and does not exhibit periodic behavior ; occasional outbursts of few magnitudes occurring in tens of days are observed .
it is generally achromatic , in particular for sources with their synchrotron peak located at frequencies higher than the optical ( fig .
[ phvar_bllac ] ) .
30 years since 1976 .
flux variations of up to @xmath83 magnitudes are observed , while the color varies within @xmath81 magnitude . ( b ) expanded view of variability during the 3-weeks period indicated as a vertical dashed line in panel ( a ) .
variations of the order of @xmath81 magnitude are seen in a time - scale of few days , while the color remains almost constant .
the light curves were produced using data from the webt archive ( courtesy of c.m .
raiteri and m. villata ) . ]
while in the majority of agns the variability time - scale can be almost directly linked to the size of the central supermassive black hole and to the accretion rate @xcite , in blazars this correlation does not hold because the luminosity is due to the highly anisotropic and relativistic jet component , rather than to the accretion disk ( see section [ sec : intro ] ) .
the relativistic aberration magnifies the variation amplitudes and foreshortens the time - scales , so that small ( 10 - 20% ) monthly or yearly flux changes , that are typical in radio - quiet quasars , become inter- or intra - day variations with amplitudes of a factor of up to 10 or more . owing to the distance and to the extreme compactness of the regions at the base of the jet , where most of the radiation is emitted ( the light - crossing times inferred from tev observations
are less than an hour ) , direct imaging of the nuclear zones , aimed at mapping the structure of the jet and the plasma outflow is not possible .
rapid variability and correlated variations of the emission at different wavelengths represent the only effective investigation tool of the jet geometry and dynamics .
ideally , these should be coupled with high angular resolution radio observations ( vlbi ) in order to correlate the flux variations at optical , x- and gamma - rays with the appearance and motion of new radio components moving across and away from the nucleus .
multi - wavelength energy distributions and their variability provide some fundamental physical parameters , most of which are not model - dependent and can thus be considered solid , although affected by the measurement uncertainties and limitations . as an example , the cooling phase of an outburst ( flux decrease ) yields the average magnetic field , and the break frequencies of the synchrotron and inverse compton spectral components give the electron energy .
their values constrain the jet models .
therefore , multi - wavelength observing campaigns of blls are organized in conjunction with outbursts , to study the jet behavior in these sources where its phenomenology is most pure and undiluted .
observations in the optical domain are crucial , because the emission at these wavelengths is pristine " , i.e. not scattered or reprocessed , and rich , i.e. not photon - starved " as it is often at x- or gamma - rays , where the photons are normally much fewer .
the first optical studies , started in the 1970 s , were based on measurements taken on photographic plates and with photomultipliers @xcite .
more recently , ccd photometry has replaced the previous types of measurements almost entirely .
since blls are very bright ( most of those at @xmath340.5 have magnitudes in the range @xmath35 ) , accurate studies of variability are possible even with small telescopes ( 0.5 - 1 m aperture ) and indeed , the advent of robotic optical and infrared facilities , primarily oriented to the follow - up and monitoring of rapid transients ( e.g. gamma - ray bursts ) , has greatly benefitted the bll field .
these small telescopes ( e.g. , rem , rotse , smarts ) have flexible schedules and pointing constraints , therefore they can be devoted to long looks of variable sources like blls @xcite . in few cases , optical orbiting telescopes designed for long , uninterrupted monitorings , like kepler , have been used for bll variability studies @xcite .
photometry from a satellite also presents the advantage of no atmospheric contamination ( seeing " ) , which may be an issue at small variability amplitudes , where atmosphere instability may mimic intrinsic flickering . in the ultraviolet domain ( 1200 - 3000 ) , an important role for bll monitoring
was played in the past by the hst high speed photometer @xcite and especially by the iue satellite , whose flexibility of scheduling and photometric stability made it apt for ultraviolet variability studies @xcite . currently , _
swift_/uvot , which couples good flexibility and sensitivity , is better suited for this class of studies than other ultraviolet orbiting facilities @xcite .
often , optical and near - infrared monitorings of blls are organized as part of multi - wavelength campaigns , that have an emphasis on the x- and gamma - ray emission , and for which the information in the optical domain is crucial to reconstruct the broad - band spectrum for comparison with time - dependent modelling .
this approach has reached a high degree of sophistication , thanks to the formation of consortia of optical observers ( often including groups of amateur astronomers , * ? ? ?
* ) , who have made their expertise and resources available to this particular purpose .
an instance of this is the _ fermi _ - _ agile _ support program , organized within the whole earth blazar telescope .
while optical observations coordinated with wide multi - wavelength efforts cover generally outbursts and high activity states of blls @xcite , standalone optical programs may take place independently of the source flux state and thus offer an unbiased view of the source behavior in different states @xcite .
as already mentioned , variations of bll emission is seen on time - scales from minutes to many months and years , with amplitudes from a fraction of a magnitude to a factor of 100 or more ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
modern observing techniques make it possible to correct reliably for spurious ( non - astrophysical ) causes of flux variations ( e.g. atmospheric instabilities ) , even when these are small ( micro - variability : @xmath36 mag , time - scales of less than one hour , * ? ? ?
the vast majority of observed optical variations at all time - scales are intrinsic , i.e. not related to interstellar scintillation ( irrelevant at optical wavelengths ) or lensing by intervening sources , the role of which was proven to be inconclusive @xcite .
often , the rapid variability events violate the limits on efficiency of conversion of kinetic energy into luminosity @xcite , and/or are inconsistent , based on light - crossing times considerations , with independent estimates of the linear sizes of the inner emitting regions ( e.g. * ? ? ?
this suggests lower limits on the doppler boosting factors of the order of @xmath37 , similar to those derived based on observed radio brightness temperatures and pair - production compactness at gamma - rays ( see section [ sec : intro ] ) .
the correlated optical / near - infrared flux and spectral variability in blls can be complex , depending on the source state and on its broad - band spectrum @xcite .
@xcite and @xcite report that the color - magnitude diagrams exhibit hysteresis cycles ( as seen occasionally in x - rays , e.g. * ? ? ?
* ; * ? ? ?
the former authors also found somewhat larger amplitude variability at the optical than near - infrared wavelengths in blls , while the latter authors observed the opposite : this divergence is possibly related to the different time - scales probed . in low flux states ,
the optical / near - infrared energy distribution may deviate from a power - law shape and show an excess toward the short wavelengths @xcite , suggesting the presence of a thermal component , that is normally better detected in fsrqs .
the intra - day variations between different bands in bl lacertae , the namesake of the class and a typical low - frequency - peaked object ( according to the definition of * ? ? ?
* ) , are correlated without measurable time lags most of the time , and suggest that its optical variability properties remarkably resemble the x - ray variability properties of high - frequency - peaked blls .
the similarities imply a common origin of the variations , plausibly the most energetic tails of the synchrotron emission produced by the relativistic electrons in the jets , for both the optical emission of low - frequency - peaked blls and the x - ray emission of high - frequency - peaked blls @xcite .
time algorithms have been devised and applied to bll light curves in search of characteristic time - scales , recurrences and to measure variability indices @xcite .
these parameters give in turn insight into the behavior of the jet at different scales and they can be an effective tool to map its geometry , especially when coupled with information at other wavelengths .
however , caution must be used in the interpretation of the outputs of correlation and auto - correlation functions and time structure functions in that often their features do not trace necessarily an authentic and intrinsic characteristic time - scale @xcite .
unlike other fronts of research on blls that have progressed very quickly in the last two decades , the field of time analysis still presents facets that are not completely understood , so that despite the availability of advanced observing facilities that allow sophisticated monitorings in excellent conditions , the results are still inconclusive .
periodicity has been searched for in the light curves of blls at all wavelengths on a wide range of frequencies and never been detected in a consistent way @xcite .
the only exception is oj 287 ( @xmath38 ) , for which a regular flux brightening recurrence every @xmath812 years was reported @xcite and confirmed by the detection of 9 equally spaced maxima - of which the latest 4 are very well sampled - in the optical light curve so far ( see fig . [ oj287lc ] ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the period is ascribed to the presence of a binary system of massive black holes at the center of the nucleus , that produce outbursts , often accompanied by precursors , when approaching pericenter @xcite . in these occasions
the flux may brighten by more than a factor of @xmath8100 with respect to quiescence .
shorter - term variations with periodic and quasi - periodic character were also detected , that may be related to the matter dynamics in the close vicinity of the binary black hole system @xcite . on the other hand , @xcite
suggest maxima at times that are not foreseen by the model , making the interpretative picture not completely satisfactory .
blazars are the astrophysical sources with the highest percentage of optical linear polarization in the radio - to - ultraviolet range . unlike radio polarization ,
the optical polarized signal is immune from propagation effects like faraday rotation of the polarization plane ; moreover , optical polarimetry probes the central nuclear regions of blazar jets , where the radio emission is often self - absorbed . at optical and near - infrared wavelengths
, the linear polarization percentage can be as high as 40% and it is normally detected at the level of 5 - 10% @xcite . since blazar optical radiation is dominated by the synchrotron mechanism , this implies the presence of highly ordered large scale magnetic fields @xcite .
circular polarization could be expected with a degree proportional to @xmath39 , where @xmath40 indicates the individual lorentz factor of an emitting relativistic particle .
because of this it is very difficult to probe and detect @xcite .
high linear polarization was very soon recognized as a distinctive feature of blazars , to the point that it has been adopted as a blazar selection criterion @xcite .
low - frequency - peaked blls and fsrqs are indistinguishable with regard to linear polarization and variability thereof @xcite , while high - frequency - peaked blls are somewhat less polarized .
this partially confirms the pioneering study of @xcite .
since interstellar dust can spuriously polarize the optical light , the detection of linear optical polarization percentage must be accurately corrected for this effect by estimating the amount of intervening dust . on the other hand ,
the host galaxy stellar light may dilute the intrinsic polarization and lower it artificially , an effect that should also be corrected for @xcite
. however , variability of the polarized signal which is often observed in blls is a proof of its intrinsic origin . 2 .
( adapted from * ? ? ?
* ) .,title="fig : " ] 2 .
( adapted from * ? ? ?
* ) .,title="fig : " ] observations of optical polarization of blls have clearly shown that the degree of polarization varies on time - scales from intra - day to years ( fig .
[ figpolarim ] , left ) , with amplitudes from 20% to a factor of 2 @xcite . in some cases ,
even faster variations were claimed @xcite .
this variability is consistent with relativistically beamed synchrotron emission viewed at a very small angle to the line of sight .
although the polarized optical light variations of blls are occasionally well correlated with those of the total light @xcite , in general lack of correlation or anti - correlation is observed @xcite .
@xcite proposed that while the optical flux originates in the weakly polarized , stable jet component , the photopolarimetric variability results from the development and propagation of a shock in the jet .
as a consequence , the optical polarized emission is potentially a better tracer of the high - energy emission , revealing the importance of optical polarimetric monitoring in multi - wavelength campaigns . among these , some of the richest from the point of view of polarimetric coverage were presented by @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite .
these campaigns were often coordinated with vlbi monitoring of the radio components , and it was noted that the polarization variability events were often correlated with the crossing of the central core by a radio - emitting blob . in particular , the optical polarization angle settles on the direction parallel to newly born radio knot .
the observed multi - wavelength behavior of the outbursts can be explained within the framework of a shock wave propagating along a helical path in the blazar s jet @xcite . in few cases spectropolarimetry of bright blls
was obtained in optical and ultraviolet and suggested a monotonic increase of polarization percentage towards shorter wavelengths ( fig .
[ figpolarim ] , right ) . from these measurements
there is no evidence for the presence of an accretion disk or other thermal source contributing significantly to its ultraviolet continuum @xcite .
the ( quasi ) featureless optical spectra of blls are historically the main characteristic for this class of objects , that distinguishes them within the blazar family ( see section [ sec : intro ] ) .
this peculiarity made them rather elusive , because of the consequent difficulty to determine their redshift and thus the distance .
however , it was early recognized that for a number of low redshift blls it was possible to detect faint absorption and/or emission features ( see e.g. , * ? ? ?
* ; * ? ? ?
* and also section [ sec : hostgals ] ) that allowed one to prove the extragalactic nature of the sources and assess their redshift . for low redshift targets the optical spectra ( see example in fig . [
fig : spec_mkn180 ] ) left no doubts that the bll phenomenon occurs in the nuclei of galaxies dominated by old stellar populations @xcite , a result that was also confirmed by first imaging studies ( see section [ sec : hostgals ] for details ) . in most cases , however , the high luminosity of the non - thermal emission from the nucleus with respect to the light from the underlying nebulosity yields spectra with very weak lines .
when observed with poor signal - to - noise ratio and resolution their optical spectra appear therefore featureless ( see example in fig .
[ fig : spec_1722 ] ) . and
na i @xmath41 atomic lines and dibs at 4428 , 4726 , 4882 , and 5772 @xcite .
see http://archive.oapd.inaf.it/zbllac/ for more examples . ]
the spectral features of blls can be grouped into three types : 1 ) spectral lines of stars from the host galaxy ; 2 ) weak emission lines characteristic of low density gas ; 3 ) intervening absorptions from cold gas .
while from the first two types a redshift can be derived , in the latter case only a lower limit on the redshift can be obtained .
the faintness of intrinsic absorption or emission lines strongly depends on the brightness level of the nuclear source that is dominated by the jet component .
given the significant flux variability of the nuclear source ( see section [ sec : opticalvar ] ) , the detection of features in the spectra depends on the brightness state of the bll .
faint states favour the detection of intrinsic emission lines while during high states intervening absorptions may be more easily detected because of the better signal - to - noise ratio .
although the optical spectrum of blls is characterized by a featureless continuum , in a number of cases ( see example in fig . [
fig : spec_0048 ] ) weak broad emission lines , similar to those observed in quasars but at lower intrinsic luminosities , are seen .
owing to the weakness of these lines they may be detectable depending on the state of the source and the quality of the observations .
such broad emission lines were occasionally found also in the prototype of the class , bl lacertae @xcite .
emission line at the redshift of the host galaxy of bl lacertae proved that the bll prototype is not a high - redshift micro - lensed quasar @xcite . ] comparing the spectra of blls and fsrqs it was found indeed that there is a continuity of emission line properties between the two sub - classes of blazars @xcite , depending on the relative strength of the various involved components ( jet emission , thermal disk or broad line region emission , host galaxy ) .
it turns out therefore that the classification of an object as bll or fsrq may depend on the spectroscopic adopted criteria ( see e.g. * ? ? ?
* for a detailed discussion on this issue ) .
the emission line photons may be comptonized by both cold and relativistic electrons in the jet ( external compton , * ? ? ?
while in fsrqs , owing to their conspicuous broad emission line luminosity , this is the dominant mechanism for production of gamma - rays , in blls it never exceeds significantly self - comptonization of synchrotron radiation . as discussed in section
[ sec : hostgals ] virtually all blls reside in massive elliptical galaxies ( or galaxies with a prominente spheroidal component ) .
this means that their stellar population is dominated by old stars and the main absorption features are ca ii 3934,3967 , g band 4304 , h@xmath42 4861 , mg i 5175 , nai 5875 .
these lines can be detected over the non - thermal component depending on the relative power of the two components ( non - thermal and host galaxy ) and on the signal - to - noise ratio and spectral resolution of the observations .
early spectroscopic studies of blls ( see e.g. , * ? ? ?
* ) were limited by relatively modest spectroscopic capabilities and many objects remained therefore classified as featureless . only for objects at low redshift and with faint nuclei these absorption lines
could be observed since the bright continuum swamps the thermal component thus reducing significantly their equivalent width .
high quality spectroscopic observations obtained with large telescopes @xcite improved significantly the ability to discover faint features from which the redshift is derived ( see example in fig .
[ fig : spec_0337 ] ) in very few cases high quality spectroscopy secured with high spatial resolution probed ongoing star formation in the inner region of the host galaxy through the detection of narrow emission lines .
a noticeable case is pks 2005 - 489 ( @xmath43 ) for which spatially resolved emissions were observed in an extended rotating ring , perpendicular to the position of the radio jet , at 4 kpc from the nucleus .
the presence of this rotating star - forming region is the signature of a minor accretion event that has funneled gas toward the central few kiloparsec region around the active nucleus @xcite . as in the case of high redshift quasars , gas along the line of sight to blls
may produce absorption systems at a number redshifts lower than the redshift of the object .
since blls exhibit very weak or no intrinsic absorption lines , they are ideal sources for studying intervening absorptions . in particular they may be due to the galactic interstellar medium . in the optical the ca ii and na i are noticeable , together with diffuse interstellar bands ( dib ) of molecular origin .
an example is given in figure [ fig : spec_1722 ] .
@xcite ] the bulk of baryonic matter ( @xmath44 ) probably resides around and between galaxies at low redshifts .
a large fraction of it ( @xmath45 ) is at high temperature ( @xmath46 ) .
it has been probed through the observation of o vii and o viii and other x - ray absorption lines .
in particular we recall here the observations of the bright bll pks 2155@xmath47304 ( e.g. * ? ? ?
* ; * ? ? ?
* ) . at lower temperatures ( @xmath48 )
the intervening gas can be traced by the ly@xmath49 forest absorption , while absorptions associated with galaxies halos are detected from metals ( mainly mg ii and c iv ) .
relevant results in connection with blls and this line of research are consequent to the installation into hst of the cosmic origin spectrograph ( cos ) , which is a medium - high ( @xmath50 ) resolution instrument of unprecedented sensitivity ( e.g. * ? ? ?
so far , spectra of four blls ( pg 1553 + 113 , s5 0716 + 714 , 3c 66a , and pg 1424 + 240 ) of previously unknown redshift were obtained in the wavelength range 1135 - 1795 .
the data show a smooth continuum , with narrow ( @xmath8100 km s@xmath1 ) absorption features arising mainly from the lyman series .
the redshift of each ly@xmath49 absorber can be measured in the @xmath51 interval .
the maximum redshift yields immediately a lower limit to the redshift of the bll , while an upper limit is derived from the expected distribution of the absorbers .
an example of a spectrum is given in figure [ danforth2010 ] .
an important aspect of these findings is that all sources are gev emitters ( three of them are also tev emitters ) . at tev energies and @xmath52 ,
the opacity of the extragalactic background light for pair production is @xmath53 ( see section [ sec : mwlref ] ) . using this technique a tev source ( pg 1424 + 240 ) was found at z @xmath22 0.6 .
the redshifts derived from cos data , together with the gamma - ray spectra are therefore an effective powerful probe of the extragalactic infrared background ( e.g. * ? ? ?
* and references therein ) . a similar technique can be applied using mg ii @xmath54 absorption doublet , which at @xmath55 can be observed from the ground . in figure [ sbarufatti2006 ]
we report the eso vlt x - shooter optical spectrum of pks 0048 - 09 where a mg ii absorption at @xmath56 is apparent .
soon after their discovery , it was noted that blls could be surrounded by a faint luminosity @xcite . in the late 1970s
the use of modern detectors ( ccd ) allowed observers to probe with better accuracy the nature of the nebulosity ( e.g. fig .
[ fig : bllac ] ) .
@xcite imaged the bll pks 0548 - 322 in various filters and found it to be composed by a giant elliptical galaxy ( m@xmath57 22 ) with a bright nucleus .
about 10 years later , at the second conference entirely dedicated to blls in como , m - h .
ulrich @xcite reported about 15 objects for which a host galaxy was measured and compared their properties with those of bright radio galaxies in order to test the consistency of the hypothesis that blls are indeed the parent population of fr - i radio galaxies with the jet pointing closely to the observer direction . with the caveat of the exiguity and non - homogeneity of the dataset
it was argued that the bll host population is very similar to that of radiogalaxies .
the following decade has seen a significant improvement in terms of quality and sample size of bll host galaxies .
various authors @xcite undertook ground based optical imaging studies of sizeable samples of objects including radio and x - ray - selected objects , according to the old definition .
all these observations led to the conclusion that blls are the active nuclei of giant / massive elliptical galaxies with average luminosity in the r band m@xmath58 @xmath8 23 .
only very few cases were reported to be different from this scheme suggesting that the host is a disk dominated galaxy ( example : 1415 + 255 ; 1413 + 135 ; * ? ? ?
* ; * ? ? ?
* ) but these cases were rather controversial and mostly due to the modest angular resolution of the images ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
optical observations were also complemented by near - infrared images , that would favour the detection of the host galaxy with respect to the nuclear emission owing to the difference of the ratio of the two components .
this is more relevant for relatively high resdshift sources since the emission from the host galaxy peaks in the near - infrared while in the optical band the light from the underlying stellar population diminishes significantly for objects with z @xmath22 0.4 - 0.5 .
near - infrared observations of blls were collected by @xcite and @xcite who extended the sample of resolved objects confirming previous results derived from optical studies . moreover in a number of cases both optical and near - infrared images were available thus allowing the estimate of the integrated broad band color ( r - h ) on blls hosts .
it turned out that the color is indistinguishable from that of inactive elliptical galaxies of similar luminosity @xcite . in spite of these imaging studies a significant fraction of observed objects remained unresolved and often these targets are also of unknown redshift .
this is because , in order to disentangle the light of the host galaxy from that of the bright nucleus it is mandatory to observe the nuclei with the best possible angular resolution ( i.e. narrow point spread function ) and also have adequate signal - to - noise ratio in order to characterise the faint extended emission around the nucleus .
this achievement became possible with the observations of the refurbished hst and wfpc2 .
the first images of blls @xcite clearly demonstrated the great advantage of the hst angular resolution in characterizing the properties of the galaxies ( see example in fig .
[ fig : hstima ] ) .
the success of the first images prompted the acquisition of a large dataset of images through an hst snapshot program . using the wfpc2 in the f702w filter short exposure ( 300 - 1000 s ) images of 110 blls were secured @xcite .
these form the largest and most homogeneous dataset of blls high resolution images . in this dataset it was possible to resolve all blls at redshift @xmath59 0.5 and many others at higher redshift or with unknown z.
all resolved host galaxies were found to be well fitted by an elliptical model ( or with a dominant bulge component ) and exhibited a relatively narrow range of absolute magnitude ( m@xmath58 = 23.0 @xmath60 0.6 ; * ? ? ?
* ) with no significant difference depending on the original selection procedure ( radio or x - ray ) .
comparison with the properties of nearby radiogalaxies strongly supports the unification picture with fr i galaxies as a favorite parent population @xcite . a detailed morphological analysis of the nearest ( z @xmath59 0.2 ) objects @xcite revealed that the host galaxies are mostly smooth and unperturbed ellipticals with no signature of dusty features or sub - structures , albeit with a high incidence of close companions ( see fig .
[ fig : hstima ] ) .
it suggests that strong gravitational interaction is not relevant for the ongoing activity .
moreover in all cases the nucleus was found very well centered on the main body of the galaxy arguing against a microlensing hypothesis for the interpretation of the class @xcite .
the characterisation of the host galaxies could also help to constrain the distance of the objects in the cases of pure featureless optical spectra .
to this aim it is mandatory to use high quality images of the targets . from the whole ensemble of imaging studies of blls , it turned out that the range of absolute magnitudes of the host galaxies of this class of agn is relatively narrow and therefore one can use their luminosity as a sort of _ standard candle_. under this assumption ( see fig . [
fig : hstmabs ] ) if the host is detected then one can derive the _ imaging redshift _ from the apparent magnitude of the host galaxy . the method , initially proposed and applied to ground based images @xcite was then expanded and perfected using hst images @xcite .
providing that the photometric and structural properties of the host galaxies are well determined ( by high quality images ) the method is relatively robust and can yield redshift estimate with an average accuracy of @xmath61 0.05 ( see figure 4 in * ? ? ?
also in the case of non detection of the host galaxy it is possible to set constraints on the minimum redshift of the object ( see e.g. * ? ? ?
radio - loud agns are often associated with jets .
the quasar prototype 3c 273 was very early recognized to exhibit an optical jet structure @xcite . in the case of blls
this association is even more cogent , since they are characterised by relativistic jets pointing in the observing direction ( see * ? ? ? * and section [ sec : intro ] ) .
nevertheless , the very fact of being pointed to the observer may make the study of the jet morphology rather elusive . in the radio band , thanks to the high angular resolution , jet detection is possible and in fact a large fraction of objects classified as blls show the signature of a jet , often with evidence of superluminal motion ( see e.g. * ? ? ?
the arcsec structures are well studied with vla or similar instruments @xcite .
radio jet sub structures are commonly detected with milliarsec resolution when observed with the vlbi ( * ? ? ?
* ; * ? ? ?
* and references therein ) . at z @xmath8 0.1
one can thus explore structures down to parsec scale .
some counterparts are found also in x - rays in chandra images ( angular resolution @xmath81 arcsec ) at relatively high redshift .
they are interpreted as a result of the compton scattering cosmic microwave background photons off the most energetic electrons @xcite .
jet morphology in x - ray was studied in some details for a number of close - by blls , like pks 0521365 @xcite , and pks 2201 + 044 and 3c 371 @xcite , which notably are systems where the beaming is supposed to be modest .
knot structures are observed . on the other hand , the jet is rarely detected in the optical and near - infrared bands .
this depends both on the more limited angular resolution with respect to the radio band and on the short lifetime of the high energy electrons producing the non - thermal emission of the jet .
because of the requirements on spatial resolution , in the optical and infrared the jet morphology was mainly studied with hst @xcite and more recently with adaptive optics methods @xcite .
a detailed study of the optical jet based on hst images was completed for pks 0521365 , pks 2201 + 04 and 3c 371 @xcite . in figure
[ fig : imajet0521 ] we show an example of optical jet of the bll pks 0521365 ( z = 0.0554 ) superposed onto the prominent massive host galaxy .
detailed analysis of the jet structure at radio , near - infrared and optical frequencies reveals a number of knots ( see fig .
[ fig : jet0521 ] ) that exhibit remarkable similarity at all frequencies @xcite .
after the recognition that blls are hosted in elliptical galaxies an obvious further step was to search for associated groups or clusters of galaxies .
early suggestions of such association in the case of pks 0548 - 322 @xcite turned out to be false ( but see below for surprise ) and few other cases were suggested for 3c66a , ap lib , pks 1400 + 164 @xcite but not properly confirmed .
for instance in the case of 1440 + 164 , for which an association with a group of galaxies was claimed , deep images obtained with modern instrumentation complemented by spectroscopy of a number of these galaxies showed that the group is not physically associated with the bll @xcite .
it was 25 years after the discovery of the bll class that a number of systematic studies of the cluster environments around blls were undertaken . using optical images
@xcite performed a study of the galaxy density for the 1 jy sample and found an increased galaxy density for low ( @xmath62 ) and intermediate ( @xmath63 ) redshift sources but not for higher redshift objects ( @xmath23 ) .
however , for the latter subsample their images were not deep enough to probe adequately the environments .
subsequent studies @xcite showed that on average blls inhabit poor galaxy environments ( abell richness class @xmath8 0 ) .
attempts to compare the environment properties with those of the alleged parent population ( see section [ sec : intro ] ) of fr
i radiogalaxies yielded somewhat controversial results @xcite .
all these studies were based only on photometry of the fields around the bll sources , therefore the association with group of galaxies can be assessed only basing on statistical considerations . for a number of selected targets
these imaging studies were complemented by spectroscopy of galaxies in the fields that demonstrated the physical association ( same redshift ) between the galaxies and the bll source ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) .
a interesting case of rich environment is found for the radio source pks 0548 - 441 .
this target was among the first blls to be pointed out as likely connected with a cluster of galaxies at @xmath64 @xcite .
the claim was , however , soon disproved as the redshift of the source ( z = 0.069 ) was derived @xcite .
this nearby object was deeply investigated by @xcite who obtained both high quality images and spectroscopy of galaxies in the field .
it turned out that the source is indeed located in a rich cluster of galaxies ( see fig .
[ fig : p0548 ] ) of which the bll host galaxy is clearly the dominant member with an absolute magnitude m@xmath58 = 23.5 .
the physical association of the cluster was proved by spectroscopy of many galaxies in the field that have the same redshift as the bll source .
this is a rare case of bll object associated with a relatively rich ( abell class @xmath8 2 ) cluster of galaxies while most of the blls reside in poorer environments . in spite of the relative bareness of bll galaxy environment ,
it is rather frequent to find close companions within 10 arcsec ( projected distance @xmath59 20 - 50 kpc for redshift in the range 0.10.4 ) from the target .
relevant examples are shown in figure [ fig : cc ] @xcite . in some cases
the companion was found to be physically associated with the active nucleus @xcite .
however , owing to the faintness of the companions and their proximity to the bright bll source , spectroscopy of the companions is lacking in many cases .
moreover in some cases also the redshift of the bll target is unknown , and it is not possible to prove the association .
in agns the mass of the black hole ( @xmath65 ) is a basic parameter characterizing the source .
the standard paradigm is that agn are supermassive accreting black holes @xcite .
the mass fixes an upper limit to the agn luminosity through the argument that the eddington luminosity can not be significantly overcome .
if the accretion occurs through a disk , its inner radius depends on the mass , because it is related to the gravitational radius , and in turn this constrains the internal temperature .
the disk inner dimension gives a number of natural time - scales , as the light crossing time , or the keplerian time .
the mass therefore enters as a basic parameter in the interpretation of the agn variability .
the most widely used method to estimate the mass of quasars or seyfert galaxies nuclei is based on the assumption that broad emission lines are produced by fluorescence of cold clouds irradiated by a central source , and orbiting the central black hole .
the line width is related to the velocity of the clouds while the distance of the cloud may be derived from _
reverberation mapping _
, that yields the time delay between a flare in the continuum , and the response of a line flux which is produced in the cloud @xcite . because of the expensiveness of this method ,
alternatively one can use an experimental relationship ( with significant dispersion ) between the size of the broad line region and the luminosity of the continuum close to the emission line .
this technique allowed the estimate of the mass of thousands of quasars with a precision of a factor @xmath8 3 @xcite . in the case of blls
this method can be applied only to few sources showing broad emission lines and with particular caution , since the observed emission is dominated by a relativistic jet .
therefore it is of obvious importance to adopt other methods to estimate the black hole mass of blls . in the local universe
it was demonstrated that most galaxies contain in their center a massive black hole and that its mass is correlated with some properties of its host galaxy ( luminosity , mass , velocity dispersion ) . as a rule of thumb the mass of the black hole
is @xmath8 0.1% of that of the spheroidal component .
the host galaxies of blls are supposedly giant ellipticals , i.e the whole galaxy coincides with spheroidal component .
the mass of the ellipticals scales as the luminosity .
therefore one can construct a black hole vs host galaxy absolute magnitude relationship , which can be used for establishing the mass once the host galaxy luminosity is measured . in the case of blls
the main difficulty is to measure the luminosity of the host galaxy ( see section [ sec : hostgals ] ) . from systematic measurements of the host galaxies luminosity of bll up to @xmath66 the black hole masses
have been calculated by @xcite and ( * ? ? ?
* see fig .
[ mbh_bll ] ) . beyond that redshift
limit the host galaxies become very difficult to detect , and evolutionary effects may become important , requiring corrections to the relations obtained locally .
relation @xcite . ] another parameter of the host galaxy , which is related to the mass of the black hole is the velocity dispersion @xmath67 of the central part of the galaxy .
it is measured through the width of the absorption lines ( typically @xmath68 ) of the stellar component . in terms of dispersion
the @xmath69 relation is somewhat preferable than that discussed in the previous paragraph .
the technique for mass determination of blls was introduced by @xcite , in this way it was determined the mass of a dozen objects @xcite and @xcite .
it was extended to @xmath860 objects by @xcite , who considered the sample of optically selected bll candidates from the sdss dr7 ( see section [ sec : demograph ] and fig .
[ mbh_bll ] ) .
also in this case there is a maximum value of @xmath70 , for the applicability of the method , related to the possibility of having reliable measurements of @xmath67 .
the mass distributions reported in fig .
[ mbh_bll ] are very similar to those of quasars , as deduced from the width of broad lines and the continuum intensity ( see e.g. * ? ? ?
finally another technique , mutuated from x - ray binaries , is based on shakura - sunyaev disk model @xcite .
the idea is to consider the spectral energy distribution of the source in a broad frequency range , exclude the contribution of the jet , and isolate the one from the accretion disk . in practice
one should be able to observe the region where @xmath71 has a maximum .
the fit with a shakura - sunyaev disk allows one to constrain the mass and the accretion rate .
if the redshift is known the latter quantity can be fixed and one obtains the mass . up to now , results have been obtained for high redshift blazars ( z @xmath72 ) , rather than blls @xcite .
about half a century after the discovery of the first blls the basic picture to interpret this enigmatic class of sources appears robust : blls are active nuclei of massive elliptical galaxies the emission of which is dominated by relativistic jets .
the bright compact radio cores , high luminosities and rapid , large amplitude flux variability at all frequencies and the strong polarization that characterise these objects are well explained in this scenario .
also the observed quasi featureless optical spectra of bll , that is one of the distinctive properties of the class , find a natural explanation in the above beaming model . at variance with other classes of agn , like quasars and seyfert galaxies ,
the lack of prominent spectral features was , and still in part is , a significant limiting factor for the determination of their redshift and ensuing evaluation of their physical properties .
the implication of this model is the existence of a large number of misaligned objects with the same intrinsic properties as blls .
the most obvious candidates are low - power radiogalaxies .
blls were mainly discovered as counterparts of radio and/or x - ray sources .
the surveys in these bands led to the compilation of the first , albeit relatively small , complete samples of objects from which it was possible to start the exploration of the cosmic evolution of the class as compared with other types of agn .
because of the limited number of known bll and the paucity of known objects at high redshift ( @xmath73 ) the evolution with cosmic time remains vague .
negative or positive evolution were suggested for different samples , although almost consistent with no evolution , yielding a picture substantially different from that of the majority of agns , where significant positive evolution is seen .
the investigation of the cosmic evolution of blls is an important issue that could offer relevant clues on the relationship between the formation of supermassive black holes and the development of relativistic jets . in order to achieve a significant progress in the understanding of the cosmic evolution of blls the extension of available samples to higher
redshift is mandatory .
the discovery and follow - up study of high - redshift blls will come within reach thanks to the new generation of observing facilities at all frequencies foreseen for the next decades ( e.g. , square kilometer array , extremely large telescopes , james webb space telescope , large synoptic survey telescope , advanced telescope for high energy astrophysics , cherenkov telescope array ) .
in particular , high signal - to - noise ratio spectroscopy of bll candidates obtained with adaptive optics instrumentation at extremely large telescopes in the near - infrared will allow one to observe the rest - frame optical spectrum of high - redshift sources and determine the redshift from absorption features of the host galaxies or faint emission lines .
this will produce the first sizable high - redshift sample of blls , that , compared with the lower redshift samples , will yield a firm conclusion on cosmic evolution of the class . this could be related to the role of the angular momentum of the central black hole in forming a relativistic jet .
although the observational knowledge of the bll phenomenon is relatively well established , on the theoretical front some fundamental aspects remain unsolved . specifically , a central problem of bll physics is the mechanism for the extraction and transfer of energy from the inner engine to the jets .
we trust that the observations with the above - mentioned future facilities will provide huge gain in terms of signal - to - noise ratio , spatial and time resolution for spectroscopy , photometry and polarimetry .
some examples of the expected capabilities for the determination of the redshift with future optical and near infrared facilities are given in @xcite and @xcite .
these crucial measurements will also contribute to the comprehension of the mechanisms of conversion of rotational energy into kinetic energy of the jet and into luminosity in various channels ( electromagnetic , neutrinos , gravitational radiation ) .
we are grateful to catherine boisson , alessandro caccianiga , stefano covino , luigi foschini , gabriele ghisellini , oliver king , alan marscher , claudia raiteri , kenji toma , massimo villata , andreas zech for inputs in the preparation of this work .
in particular , we thank gabriele giovannini and massimo persic for useful suggestions , comments and discussions on this review .
we are also very grateful to angela sandrinelli and marco landoni for a careful reading of the manuscript and for fruitful comments .
we acknowledge support from asi / inaf grant i/088/06/0 and prin miur 2010/2011 .
abazajian , k. n. , adelman - mccarthy , j. k. , ageros , m. a. , et al .
2009 , , 182 , 543 abdo , a. a. , ackermann , m. , ajello , m. , et al . 2010a , , 715 , 429 abdo , a. a. , ackermann , m. , agudo , i. , et al .
2010b , , 716 , 30 abraham , r. g. , crawford , c. s. , & mchardy , i. m. 1991 , , 252 , 482 acciari , v. a. , aliu , e. , arlen , t. , et al .
2011 , , 738 , 169 ackermann , m. , ajello , m. , allafort , a. , et al .
2011 , , 743 , 171 ackermann , m. , ajello , m. , ballet , j. , et al .
2012 , , 751 , 159 agudo , i. , marscher , a. p. , jorstad , s. g. , et al .
2011 , , 735 , l10 aharonian , f. a. 2000 , new astronomy , 5 , 377 aharonian , f. , akhperjanian , a. g. , anton , g. , et al .
2009 , , 502 , 749 ajello , m. , shaw , m. s.,romani , r. w. , et al .
2012 , , 751 , 108 ajello , m. , romani , r. w. , gasparrini , d. , et al .
2014 , , 780 , 73 aleksi , j. , antonelli , l. a. , antoranz , p. , et al .
2011 , , 530 , a4 aleksi , j. , alvarez , e. a. , antonelli , l. a. , et al .
2012 , , 542 , a100 aliu , e. , aune , t. , beilicke , m. , et al .
2011 , , 742 , 127 aliu , e. , archambault , s. , arlen , t. , et al .
2013 , , 775 , 3 allen , r. g. , smith , p. s. , angel , j. r. p. , et al .
1993 , , 403 , 610 andruchow , i. , cellone , s. a. , & romero , g. e. 2008 , , 388 , 1766 angel , j.r.p .
, & stockman , h.s .
1980 , , 18 , 321 asano , k. , takahara , f. , kusunose , m. , toma , k. , & kakuwa , j. 2014 , , 780 , 64 barres de almeida , u. , ward , m. j. , dominici , t. p. , et al .
2010 , , 408 , 1778 barth , a. j. , ho , l. c. , & sargent , w. l. w. 2003 , apj , 583 , 134 bauer , a. , baltay , c. , coppi , p. , et al .
2009 , , 699 , 1732 beckmann , v. , engels , d. , bade , n. , & wucknitz , o. 2003 , , 401 , 927 begelman , m. c. , blandford , r. d. , & rees , m. j. 1984 , reviews of modern physics , 56 , 255 begelman , m. c. , fabian , a. c. , & rees , m. j. 2008 , , 384 , l19 bersanelli , m. , bouchet , p. , falomo , r. , & tanzi , e. g. 1992 , , 104 , 28 birkinshaw , m. , worrall , d. m. , & hardcastle , m. j. 2002 , , 335 , 142 bjrnsson , c .- i . 1990 , , 242 , 158 blandford , r. d. , & znajek , r. l. 1977 , , 179 , 433 blandford , r. d. & rees , m. j. 1978 , in : pittsburgh conference on bl lac objects , pittsburgh , pa . , april 24 - 26 , 1978 , proceedings .
( a79 - 30026 11 - 90 ) pittsburgh , pa . , university of pittsburgh , 1978 , p. 328 - 341
baejowski , m. , sikora , m. , moderski , r. , & madejski , g. m. 2000 , , 545 , 107 bttcher , m. , reimer , a. , sweeney , k. , & prakash , a. 2013 , , 768 , 54 bonning , e. , urry , c. m. , bailyn , c. , et al .
2012 , , 756 , 13 bressan , a. , falomo , r. , valds , j. r. , & rampazzo , r. 2006 , , 645 , l101 butcher , h. r. , oemler , a. , jr . ,
tapia , s. , & tarenghi , m. 1976 , , 209 , l11 caccianiga , a. , maccacaro , t. , wolter , a. , della ceca , r. , & gioia , i. m. 2002 , , 566 , 181 caccianiga , a. , severgnini , p. , della ceca , r. , et al .
2008 , , 477 , 735 cagnoni , i. , nicastro , f. , maraschi , l. , treves , a. , & tavecchio , f. 2004 , , 603 , 449 camenzind , m. , & krockenberger , m. 1992 , , 255 , 59 capetti , a. , celotti , a. , chiaberge , m. , et al .
2002 , , 383 , 104 carini , m. t. , barnaby , d. , mattox , j. r. , et al .
2004 , astronomische nachrichten , 325 , 646 celotti , a. , ghisellini , g. , & chiaberge , m. 2001 , , 321 , l1 chen , p. s. , fu , h. w. , & gao , y. f. 2005 , , 11 , 27 chen , p. s. , & shan , h. g. 2011 , , 732 , 22 chiaberge , m. , celotti , a. , capetti , a. , & ghisellini , g. 2000 , , 358 , 104 chiaberge , m. , capetti , a. , & celotti , a. 2001 , , 324 , l33 ciprini , s. , tosti , g. , raiteri , c. m. , et al .
2003 , , 400 , 487 corbett , e. a. , robinson , a. , axon , d. j. , et al .
1996 , , 281 , 737 costamante , l. , ghisellini , g. , giommi , p. , et al .
2001 , , 371 , 512 costamante , l. 2013 , international journal of modern physics d , 22 , 30025 cusumano , g. , la parola , v. , segreto , a. , et al .
2010 , , 524 , a64 dabrusco , r. , massaro , f. , paggi , a. , et al .
2013 , , 206 , 12 danforth , c. w. , keeney , b. a. , stocke , j. t. , shull , j. m. , & yao , y. 2010 , , 720 , 976 disney , m. j. 1974 , , 193 , l103 disney , m. j. , peterson , b. a. , & rodgers , a. w. 1974 , , 194 , l79 dolan , j. f. , & clark , l. l. 2004 , , 607 , 84 dolcini , a. , farfanelli , f. , ciprini , s. , et al .
2007 , , 469 , 503 edelson , r. a. , saken , j. , pike , g. , et al .
1991 , , 372 , l9 edelson , r. , mushotzky , r. , vaughan , s. , et al .
2013 , , 766 , 16 elvis , m. , plummer , d. , schachter , j. , & fabbiano , g. 1992 , , 80 , 257 emmanoulopoulos , d. , mchardy , i. m. , & uttley , p. 2010
, , 404 , 931 falomo , r. , treves , a. , chiappetti , l. , et al .
1993 , , 402 , 532 falomo , r. , pesce , j. e. , & treves , a. 1993 , , 411 , l63 falomo , r. , pesce , j. e. , & treves , a. 1993 , , 105 , 2031 falomo , r. , bersanelli , m. , bouchet , p. , & tanzi , e. g. 1993 , , 106 , 11 falomo , r. , pesce , j. e. , & treves , a. 1995 , , 438 , l9 falomo , r. 1996 , , 283 , 241 falomo , r. , kotilainen , j. , pursimo , t. , et al .
1997 , , 321 , 374 falomo , r. , urry , c. m. , pesce , j. e. , et al .
1997 , , 476 , 113 falomo , r. , & kotilainen , j. k. 1999 , , 352 , 85 falomo , r. , scarpa , r. , treves , a. , & urry , c. m. 2000 , , 542 , 731 falomo , r. , kotilainen , j. k. , & treves , a. 2002 , , 569 , l35 falomo , r. , kotilainen , j. k. , carangelo , n. , & treves , a. 2003 , , 595 , 624 falomo , r. , carangelo , n. , & treves , a. 2003 , , 343 , 505 falomo , r. , pian , e. , treves , a. , et al .
2009 , , 501 , 907 fanaroff , b. l. , & riley , j. m. 1974 , , 167 , 31p finke , j. d. , dermer , c. d. , boettcher , m. 2008 , , 686 , 181 fleming , t. a. , green , r. f. , jannuzi , b. t. , et al .
1993 , , 106 , 1729 fosbury , r. a. e. , & disney , m. j. 1976 , , 207 , l75 foschini , l. , ghisellini , g. , tavecchio , f. , et al . 2007 , , 657 , l81 fossati , g. , maraschi , l. , celotti , a. , comastri , a. , & ghisellini , g. 1998 , , 299 , 433 fossati , g. , celotti , a. , chiaberge , m. , et al .
2000 , , 541 , 166 fried , j. w. , stickel , m. , & khr , h. 1993 , , 268 , 53 fujiwara , m. , matsuoka , y. , & ienaka , n. 2012 , , 144 , 112 gabuzda , d. c. , cawthorne , t. v. , roberts , d. h. , & wardle , j. f. c. 1989 , , 347 , 701 gaur , h. , gupta , a. c. , wiita , p. j. , et al .
2014 , , 781 , l4 ghisellini , g. , maraschi , l. , & treves , a. 1985 , , 146 , 204 ghisellini , g. , padovani , p. , celotti , a. , & maraschi , l. 1993 , , 407 , 65 ghisellini , g. , villata , m. , raiteri , c. m. , et al .
1997 , , 327 , 61 ghisellini , g. , celotti , a. , fossati , g. , maraschi , l. , & comastri , a. 1998 , , 301 , 451 ghisellini , g. , & tavecchio , f. 2008 , , 387 , 1669 ghisellini , g. , tavecchio , f. , foschini , l. , et al .
2010 , , 402 , 497 ghisellini , g. 2013 , lecture notes in physics , berlin springer verlag , 873 , gioia , i. m. , maccacaro , t. , schild , r. e. , et al . 1990 , , 72 , 567 giommi , p. , menna , m. t. , & padovani , p. 1999
, , 310 , 465 giommi , p. , piranomonte , s. , perri , m. , & padovani , p. 2005
, , 434 , 385 giovannini , g. , falomo , r. , scarpa , r. , treves , a. , & urry , c. m. 2004 , , 613 , 747 giroletti , m. , giovannini , g. , taylor , g. b. , & falomo , r. 2004 , , 613 , 752 gladders , m. d. , abraham , r. g. , mchardy , i. m. , et al .
1997 , , 284 , 27 green , r. f. , schmidt , m. , & liebert , j. 1986 , , 61 , 305 green , j. c. , froning , c. s. , osterman , s. , et al .
2012 , , 744 , 60 gu , m. f. , lee , c .- u . , pak , s. , yim , h. s. , & fletcher , a. b. 2006 , , 450 , 39 guilbert , p. w. , fabian , a. c. , & rees , m. j. 1983 , , 205 , 593 gupta , a. c. , krichbaum , t. p. , wiita , p. j. , et al .
2012 , , 425 , 1357 hagen - thorn , v. a. , larionov , v. m. , jorstad , s. g. , et al . 2008 , , 672 , 40 halpern , j. p. , impey , c. d. , bothun , g. d. , et al .
1986 , , 302 , 711 hanson , c. g. , & coe , m. j. 1985 , , 217 , 831 hartman , r. c. , bertsch , d. l. , bloom , s. d. , et al . 1999 , , 123 , 79 heidt , j. , wagner , s. j. , & wilhelm - erkens , u. 1997 , , 325 , 27 heidt , j. , & wagner , s. j. 1998 , , 329 , 853 heidt , j. , nilsson , k. , fried , j. w. , takalo , l. o. , & sillanp , a. 1999 , , 348 , 113 heidt , j. , & nilsson , k. 2011 , , 529 , a162 h.e.s.s .
collaboration , abramowski , a. , acero , f. , et al .
2012 , , 539 , a149 h.e.s.s .
collaboration , abramowski , a. , acero , f. , et al .
2013 , , 552 , a118 hoffmeister , c. 1929 , astronomische nachrichten , 236 , 233 hoyle , f. burbidge , g. r. , & sargent , w. l. w. 1966 , , 209 , 751 hudec , r. , bata , m. , pihajoki , p. , & valtonen , m. 2013 , a&a , 559 , a20 hutsemkers , d. , borguet , b. , sluse , d. , cabanac , r. , & lamy , h. 2010 , , 520 , l7 impey , c. d. , & tapia , s. 1988 , , 333 , 666 impey , c. d. , lawrence , c. r. , & tapia , s. 1991 , , 375 , 46 impey , c. d. , bychkov , v. , tapia , s. , gnedin , y. , & pustilnik , s. 2000 , , 119 , 1542 itoh , r. , fukazawa , y. , chiang , j. , et al .
2013 , pasj , 65 , 18 jannuzi , b. t. , yanny , b. , & impey , c. 1997 , , 491 , 146 jorstad , s. g. , marscher , a. p. , larionov , v. m. , et al . 2010 , , 715 , 362 kastendieck , m. a. , ashley , m. c. b. , & horns , d. 2011 , , 531 , a123 kato , t. , uemura , m. , ishioka , r. , et al .
2004 , , 56 , 1 kellermann , k. i. , & pauliny - toth , i. i. k. 1969 , , 155 , l71 kinman , t. d. 1975 , , 197 , l49 kirk , j. g. , rieger , f. m. , & mastichiadis , a. 1998 , , 333 , 452 kotilainen , j. k. , falomo , r. , & scarpa , r. 1998 , , 336 , 479 khr , h. , pauliny - toth , i. i. k. , witzel , a. , & schmidt , j. 1981 , , 86 , 854 lamer , g. , newsam , a. m. , & mchardy , i. m. 1999 , , 309 , 1085 landoni , m. , falomo , r. , treves , a. , et al .
2012 , , 543 , a116 landoni , m. , falomo , r. , treves , a. , et al .
2013 , , 145 , 114 landoni , m. , falomo , r. , treves , a. , and sbarufatti , b. , 2014 , a&a , submitted larionov , v. m. , villata , m. , & raiteri , c. m. 2010 , , 510 , a93 larionov , v. m. , jorstad , s. g. , marscher , a. p. , et al .
2013 , , 768 , 40 laurent - muehleisen , s. a. , kollgaard , r. i. , moellenbrock , g. a. , & feigelson , e. d. 1993 , , 106 , 875 laurent - muehleisen , s. a. , kollgaard , r. i. , ciardullo , r. , et al .
1998 , , 118 , 127 laurent - muehleisen , s. a. , kollgaard , r. i. , feigelson , e. d. , brinkmann , w. , & siebert , j. 1999 , , 525 , 127 maccacaro , t. , gioia , i. m. , maccagni , d. , & stocke , j. t. 1984 , , 284 , l23 maccacaro , t. , caccianiga , a. , della ceca , a. , wolter , a. , & gioia , i. m. 1998 , astronomische nachrichten , 319 , 15 macleod , j. m. , & andrew , b. h. 1968 , , 1 , 243 magic collaboration , albert , j. , aliu , e. , et al .
2008 , science , 320 , 1752 mannheim , k. , & biermann , p. l. 1992 , , 253 , l21 mannheim , k. 1998 , science , 279 , 684 maraschi , l. , ghisellini , g. , tanzi , e. g. , & treves , a. 1986 , , 310 , 325 maraschi , l. , ghisellini , g. , & celotti , a. 1992 , , 397 , l5 maraschi , l. , & rovetti , f. 1994 , , 436 , 79 maraschi , l. , & tavecchio , f. 2003 , , 593 , 667 march , m. j. m. , & caccianiga , a. 2013 , , 430 , 2464 marscher , a. p. , jorstad , s. g. , mattox , j. r. , & wehrle , a. e. 2002 , , 577 , 85 marscher , a. p. , jorstad , s. g. , darcangelo , f. d. , et al .
2008 , , 452 , 966 maselli , a. , massaro , e. , nesci , r. , et al .
2010 , , 512 , a74 massaro , e. , giommi , p. , leto , c. , et al .
2009 , , 495 , 691 massaro , e. , nesci , r. , & piranomonte , s. 2012a , , 422 , 2322 massaro , f. , dabrusco , r. , tosti , g. , et al . 2012b , , 752 , 61 massaro , f. , paggi , a. , errando , m. , et al .
2013 , , 207 , 16 mcbreen , b. 1979 , , 71 , l19 mchardy , i. m. , koerding , e. , knigge , c. , uttley , p. , & fender , r. p. 2006
, , 444 , 730 miller , j. s. , & hawley , s. a. 1977 , , 212 , l47 miller , j. s. , french , h. b. , & hawley , s. a. 1978a , bl lac objects , 176 miller , j. s. , french , h. b. , & hawley , s. a. 1978b , , 219 , l85 miller , h. r. , & mcgimsey , b. q. 1978 , , 220 , 19 miller , h. r. , carini , m. t. , & goodrich , b. d. 1989 , , 337 , 627 moderski , r. , sikora , m. , coppi , p. s. , & aharonian , f. 2005 , , 363 , 954 moore , r. l. , & stockman , h. s. 1981 , , 243 , 60 morris , s. l. , stocke , j. t. , gioia , i. m. , et al .
1991 , , 380 , 49 mcke , a. , & protheroe , r. j. 2001 , astroparticle physics , 15 , 121 mcke , a. , protheroe , r. j. , engel , r. , rachen , j. p. , & stanev , t. 2003 , astroparticle physics , 18 , 593 nesci , r. , tosti , g. , pursimo , t. , ojha , r. , & kadler , m. 2013 , , 555 , a2 nicastro , f. , zezas , a. , drake , j. , et al .
2002 , , 573 , 157 ojha , r. , kadler , m. , bck , m. , et al .
2010 , , 519 , a45 osone , s. 2006 , astroparticle physics , 26 , 209 ostorero , l. , villata , m. , & raiteri , c. m. 2004 , , 419 , 913 ostriker , j. p. , & vietri , m. 1985 , , 318 , 446 ostriker , j. p. , & vietri , m. 1990 , , 344 , 45 padovani , p. , & urry , c. m. 1990 , , 356 , 75 padovani , p. , & giommi , p. 1995
, , 444 , 567 padovani , p. , giommi , p. , landt , h. , & perlman , e. s. 2007 , , 662 , 182 padovani , p. , giommi , p. , & rau , a. 2012 , , 422 , l48 paggi , a. , massaro , f. , dabrusco , r. , smith , h. a. , masetti , n. , giroletti , m. , tosti , g. , & funk , s. 2013 , , 209 , 9 paltani , s. , & courvoisier , t. j .- l . 1994 , , 291 , 74 pavlidou , v. , angelakis , e. , myserlis , i. , et al .
2013 , arxiv:1311.3304 perlman , e. s. , stocke , j. t. , schachter , j. f. , et al .
1996 , , 104 , 251 perri , m. , maselli , a. , giommi , p. , et al .
2007 , , 462 , 889 persic , m. , & de angelis , a. 2008 , , 483 , 1 pesce , j. e. , falomo , r. , & treves , a. 1994 , , 107 , 494 pesce , j. e. , falomo , r. , & treves , a. 1995 , , 110 , 1554 peterson , b. m. 1997 , an introduction to active galactic nuclei , cambridge university press , new york , 1997 pian , e. , falomo , r. , scarpa , r. , & treves , a. 1994 , apj , 432 , 547 pian , e. , urry , c. m. , treves , a. , et al .
1997 , , 486 , 784 pian , e. , vacanti , g. , tagliaferri , g. , et al .
1998 , , 492 , l17 pian , e. , romano , p. , treves , a. , et al .
2007 , , 664 , 106 pietil , h. , takalo , l. o. , tosti , g. , et al . 1999 , , 345 , 760 pihajoki , p. , valtonen , m. , zola , s. , et al .
2013a , , 764 , 5 pihajoki , p. , valtonen , m. , & ciprini , s. 2013b , , 434 , 3122 piirola , v. 1973 , , 27 , 383 piner , b. g. , pant , n. , & edwards , p. g. 2010 , , 723 , 1150 plotkin , r. m. , anderson , s. f. , hall , p. b. , et al .
2008 , , 135 , 2453 plotkin , r. m. , anderson , s. f. , brandt , w. n. , et al .
2010 , , 139 , 390 plotkin , r. m. , markoff , s. , trager , s. c. , & anderson , s. f. 2011 , , 413 , 805 pollock , j. t. , webb , j. r. , & azarnia , g. 2007 , , 133 , 487 potter , w. j. , & cotter , g. 2013 , , 436 , 304 punch , m. , akerlof , c. w. , cawley , m. f. , et al .
1992 , , 358 , 477 raiteri , c. m. , villata , m. , capetti , a. , et al .
2007 , , 464 , 871 raiteri , c. m. , villata , m. , capetti , a. , et al .
2009 , , 507 , 769 raiteri , c. m. , villata , m. , bruschini , l. , et al .
2010 , , 524 , a43 raiteri , c. m. , villata , m. , dammando , f. , et al .
2013 , , 436 , 1530 rani , b. , krichbaum , t. p. , fuhrmann , l. , et al .
2013 , , 552 , a11 rector , t. a. , stocke , j. t. , perlman , e. s. , morris , s. l. , & gioia , i. m. 2000 , , 120 , 1626 rector , t.a . , &
stocke , j.t .
2003 , aj , 125 , 2447 rieger , f. m. 2004 , , 615 , l5 romanishin , w. 1987 , , 320 , 586 salpeter , e. e. 1964 , , 140 , 796 sambruna , r. m. , maraschi , l. , & urry , c. m. 1996 , , 463 , 444 sambruna , r. m. , donato , d. , tavecchio , f. , et al .
2007 , , 670 , 74 sandrinelli , a. , treves , a. , falomo , r. farina , e.p . ,
foschini , l. , landoni , m. , & sbarufatti , b. 2013 , , 146 , 163 sandrinelli , a. , covino , s. , & treves , a. 2014 , , 562 , a79 sasada , m. , uemura , m. , arai , a. , et al .
2008 , , 60 , l37 sbarrato , t. , ghisellini , g. , nardini , m. , et al .
2013 , , 433 , 2182 sbarufatti , b. , treves , a. , & falomo , r. 2005 , , 635 , 173 sbarufatti , b. , falomo , r. , treves , a. , & kotilainen , j. 2006a , , 457 , 35 sbarufatti , b. , treves , a. , falomo , r. , et al . 2006b , , 132 , 1 scarpa , r. , & falomo , r. 1997 , , 325 , 109 scarpa , r. , urry , c.m . , falomo , r. , & treves , a. 1999 , apj , 526 , 643 scarpa , r. , urry , c. m. , falomo , r. , pesce , j. e. , & treves , a. 2000 , , 532 , 740 schmidt , m. 1963 , , 197 , 1040 schmitt , j. l. 1968 , , 218 , 663 schwartz , d. a. , & ku , w. h .- m .
1983 , , 266 , 459 schwartz , d. a. , marshall , h. l. , lovell , j. e. j. , et al .
2000 , , 540 , l69 entrk , g. d. , errando , m. , bttcher , m. , & mukherjee , r. 2013 , , 764 , 119 shakura , n. i. , & sunyaev , r. a. 1973 , , 24 , 337 shaw , m. s. , romani , r. w. , cotter , g. , et al .
2013 , , 764 , 135 shen , y. , richards , g. t. , strauss , m. a. , et al .
2011 , , 194 , 45 sikora , m. , begelman , m. c. , & rees , m. j. 1994 , , 421 , 153 sillanp , a. , haarala , s. , & korhonen , t. 1988a , , 72 , 347 sillanp , a. , haarala , s. , valtonen , m. j. , sundelius , b. , & byrd , g. g. 1988b , , 325 , 628 sillanp , a. , takalo , l. o. , pursimo , t. , et al .
1996 , , 305 , l17 smith , p. s. , allen , r. g. , & angel , j. r. p. 1993
, , 415 , l83 smith , e. p. , odea , c. p. , & baum , s. a. 1995 , , 441 , 113 smith , p. s. , williams , g. g. , schmidt , g. d. , diamond - stanic , a. m. , & means , d. l. 2007 , , 663 , 118 sorcia , m. , bentez , e. , hiriart , d. , et al .
2013 , , 206 , 11 stalin , c. s. , gopal - krishna , sagar , r. , et al .
2006 , , 366 , 1337 stein , w. a. , odell , s. l. , & strittmatter , p. a. 1976 , , 14 , 173 stickel , m. , padovani , p. , urry , c. m. , fried , j. w. , & khr , h. 1991 , , 374 , 431 stickel , m. , fried , j. w. , & khr , h. 1993 , , 98 , 393 stocke , j. t. , morris , s. l. , gioia , i. , et al . 1990 , , 348 , 141 stocke , j. t. , wurtz , r. , wang , q. , elston , r. , & jannuzi , b. t. 1992 , , 400 , l17 strittmatter , p. a. , serkowski , k. , carswell , r. , et al .
1972 , , 175 , l7 takahashi , t. , tashiro , m. , madejski , g. , et al .
1996 , , 470 , l89 takalo , l. o. , sillanpaa , a. , nilsson , k. , et al .
1992 , , 94 , 37 takalo , l. o. , & sillanpaa , a. 1993 , ap&ss , 206 , 191 takalo , l. o. 1994 , vistas in astronomy , 38 , 77 tavecchio , f. , maraschi , l. , sambruna , r. m. , & urry , c. m. 2000 , , 544 , l23 tommasi , l. , palazzi , e. , pian , e. , et al . 2001a , , 376 , 51 tommasi , l. , daz , r. , palazzi , e. , et al .
2001b , , 132 , 73 tosti , g. , pascolini , s. , & fiorucci , m. 1996 , , 108 , 706 tosti , g. , fiorucci , m. , luciani , m. , et al .
1998 , , 339 , 41 tosti , g. , massaro , e. , nesci , r. , et al .
2002 , , 395 , 11 tramacere , a. , giommi , p. , massaro , e. , et al .
2007 , , 467 , 501 treves , a. , falomo , r. , & uslenghi , m. 2007 , , 473 , l17 trussoni , e. , capetti , a. , celotti , a. , chiaberge , m. , & feretti , l. 2003 , , 403 , 889 ulrich , m .- h .
1978 , , 222 , l3 ulrich , m. h. , hackney , k. r. h. , hackney , r. l. , & kondo , y. 1984 , , 276 , 466 ulrich , m. h. 1989 , bl lac objects , 334 , 45 ulvestad , j. s. , & antonucci , r. r. j. 1986 , , 92 , 6 urry , c. m. , & shafer , r. a. 1984 , , 280 , 569 urry , c. m. , & padovani , p. 1991
, , 371 , 60 urry , c. m. , maraschi , l. , edelson , r. , et al .
1993 , , 411 , 614 urry , c. m. , & padovani , p. 1995
, , 107 , 803 urry , c. m. , treves , a. , maraschi , l. , et al .
1997 , , 486 , 799 urry , c. m. , falomo , r. , scarpa , r. , et al . 1999 , , 512 , 88 urry , c. m. , scarpa , r. , odowd , m. , et al .
2000 , , 532 , 816 valtonen , m. j. , lehto , h. j. , nilsson , k. , et al .
2008 , , 452 , 851 vercellone , s. , striani , e. , vittorini , v. , et al .
2011 , , 736 , l38 vermeulen , r.c . ,
ogle , p.m. , tran , h.d . ,
browne , i.w.a . ,
cohen , m.h . ,
readhead , a.c.s . ,
taylor , g.b . , goodrich , r.w .
1995 , apj , 452 , l5 vron , p. , & vron , m. p. 1976 , , 47 , 319 vron - cetty , m .- p . , & vron , p. 2010
, , 518 , a10 villata , m. , raiteri , c. m. , kurtanidze , o. m. , et al .
2002 , , 390 , 407 villata , m. , raiteri , c. m. , kurtanidze , o. m. , et al . 2004a , , 421 , 103 villata , m. , raiteri , c. m. , aller , h. d. , et al .
2004b , , 424 , 497 visvanathan , n. , & griersmith , d. 1977 , , 215 , 759 visvanathan , n. , & wills , b. j. 1998 , , 116 , 2119 voges , w. , aschenbach , b. , boller , t. , et al .
1999 , , 349 , 389 wagner , s. j. , & witzel , a. 1995 , , 33 , 163 watson , d. , hanlon , l. , mcbreen , b. , et al . 1999 , , 345 , 414 weistrop , d. 1973 , nature , 241 , 157 weistrop , d. , smith , b. a. , & reitsema , h. j. 1979 , , 233 , 504 weistrop , d. , reitsema , h. j. , smith , b. a. , & shaffer , d. b. 1983 , , 271 , 471 westerlund , b. e. , wlerick , g. , & garnier , r. 1982 , , 105 , 284 westfold , k. c. 1959 , , 130 , 241 wolter , a. , comastri , a. , ghisellini , g. , et al .
1998 , , 335 , 899 woltjer , l. 1966 , , 146 , 597 woo , j .- h . , &
urry , c. m. 2002 , , 579 , 530 wright , s. c. , mchardy , i. m. , abraham , r. g. , & crawford , c. s. 1998 , , 296 , 961 wurtz , r. , ellingson , e. , stocke , j. t. , & yee , h. k. c. 1993 , , 106 , 869 wurtz , r. , stocke , j. t. , & yee , h. k. c. 1996 , , 103 , 109 wurtz , r. , stocke , j. t. , ellingson , e. , & yee , h. k. c. 1997 , , 480 , 547 zeldovich , y. b. 1964 , soviet physics doklady , 9 , 195 zhang , x. , zheng , y. g. , zhang , h. j. , et al .
2008 , , 136 , 1846 zhang , b .- k . ,
wang , s. , zhao , x .- y . ,
dai , b .- z . , &
zha , m. 2013a , , 428 , 3630 zhang , y .- h . , bian , f .- y . , li , j .- z . , & shang , r .- c . 2013b , , 432 , 1189 zhang , h. , chen , x. , & boettcher , m. 2014 , arxiv:1401.7138 | bl lac objects are active nuclei , hosted in massive elliptical galaxies , the emission of which is dominated by a relativistic jet closely aligned with the line of sight .
this implies the existence of a parent population of sources with a misaligned jet , that have been identified with low - power radiogalaxies .
the spectrum of bl lacs , dominated by non - thermal emission over the whole electromagnetic range , together with bright compact radio cores , high luminosities , rapid and large amplitude flux variability at all frequencies and strong polarization make these sources an optimal laboratory for high energy astrophysics .
a most distinctive characteristic of the class is the weakness or absence of spectral lines , that historically hindered the identification of their nature and ever thereafter proved to be a hurdle in the determination of their distance . in this paper
we review the main observational facts that contribute to the present basic interpretation of this class of active galaxies .
we overview the history of the bl lac objects research field and their population as it emerged from multi - wavelength surveys .
the properties of the flux variability and polarization , compared with those at radio , x - ray and gamma - ray frequencies , are summarized together with the present knowledge of the host galaxies , their environments , and central black hole masses .
we focus this review on the optical observations , that played a crucial role in the early phase of bl lacs studies , and , in spite of extensive radio , x - ray , and recently gamma - ray observations , could represent the future major contribution to the unveiling of the origin of these sources . in particular they could provide a firm conclusion on the long debated issue of the cosmic evolution of this class of active galactic nuclei and on the connection between formation of supermassive black holes and relativistic jets .
* preface * discovered and classified as an irregularly variable star by @xcite , bl lacertae was found to have a radio counterpart @xcite .
this , together with the high optical polarization , drew attention on this source and on others that exhibited similar properties , primarily a highly variable and featureless , power - law - shaped optical spectral continuum .
this class of objects was soon recognized to be extragalactic in nature and to share some of the characteristics of quasars @xcite . supporting evidence of this came from the detection of a surrounding nebulosity centered on the optical point - like source ( first noted by * ? ? ?
* ) , that hinted at the presence of a host galaxy @xcite .
the peculiarities that make bl lac objects ( blls ) so enigmatic can be composed in a coherent picture that foresees a relativistic jet viewed at a small angle as the main responsible of all multi - wavelength bll characteristics @xcite .
this explanation was later reinforced by the elaboration of a unifying scenario that envisages radiogalaxies , for which kiloparsec - scale jets are directly observed , and blls as members of the same population , with the viewing angle determining the different observed properties .
thanks to their widely extended spectral energy distribution , that covers sixteen decades in frequency , from the radio to tev , blls can be detected with observations at any frequency .
in fact , surveys at radio , optical , x - ray and gamma - ray frequencies in the last 40 years have contributed to select bll candidates . after identification via optical spectroscopy , samples of bona fide blls have been constructed at different frequencies and have made the statistical study of the population possible . a crucial parameter for assessing their cosmological role and evolution is their redshift , a quantity that is intrinsically difficult to determine in blls and for which accurate optical spectroscopy is critical . by 1976 there were 30 known blls @xcite ; almost forty years on , there are nearly 1400 catalogued blls @xcite . notwithstanding the importance of the multi - wavelength investigation , the observation and analysis of blls in the optical band remain a mainstay of research in this field , as witnessed by the advances in the last decade . in this review
we have therefore focused on the optical domain .
the first specific review on blls was that of @xcite .
blls were the subject of three dedicated conferences : at the university of pittsburgh in 1978 ( pittsburgh conference on bl lac objects , proceedings edited by a.m. wolfe , nsfp pittsburgh , 1978 ) , in como , italy , in 1988 ( bl lac objects , eds .
l. maraschi , t. maccacaro , m .- h .
ulrich , lecture notes in physics , springer - verlag , 1989 ) , and in turku , finland , in 1998 ( bl lac phenomenon , eds .
l. o. takalo and a. sillanp , astronomical society of the pacific , 1999 ) . |
dirichlet branes , i.e. hypersurfaces which the endpoints of strings are confined to , play a fundamental role in the recent developments of the string theory duality pattern @xcite and the connected genesis of a unique underlying theory .
such d - branes arise from open strings with free ends by t - dualizing type i string theory or have to be added to type ii theories in order to fulfil all the duality requirements . in @xmath0-model language
the t - duality transformation for open strings has been studied recently , both via formal functional integral manipulations as well as via canonical transformations @xcite .
the aim of this contribution is twofold .
at first we present in section 2 a streamlined discussion of t - duality for @xmath0-models on two - dimensional manifolds with boundary based on the functional integral formulation .
it is suited for a parallel study of both abelian and non - abelian gauge fields coupling to the boundary of the manifold @xcite .
the formalism can be used to give meaning to the notion of a matrix valued d - brane position .
secondly , we comment on a simple observation relevant for the issue of quantum equivalence of t - dual @xmath0-models on manifolds with boundary . to give to the results gained by formal manipulations with the functional integral or by classical canonical transformations an exact status
, one has to perform a careful analysis of the renormalization properties .
for @xmath0-models without boundary it has been shown that beyond 1-loop order the buscher rules @xcite require modifications @xcite .
concerning the gauge field sector common folklore seems to take the absence of such corrections for granted .
this is based on the remarkable duality property of born - infeld actions @xcite . in section 3
we will point out that in the d - brane case the relation between the functional derivatives of the born - infeld action and the @xmath1-functions is an inhomogeneous one .
this is in collision with @xcite and implies the need for a more detailed study of the problem in the gauge field sector , too .
our starting point is a generalized @xmath0-model describing an open string moving in a target space with fields @xmath5 and @xmath6 .
the fields @xmath7 correspond to the massless excitations of closed strings .
the gauge field @xmath6 taking its values in the lie algebra of some gauge group @xmath8 corresponds to the massless excitations of open strings . to identify the dual target space fields one can restrict oneself to the partition function ( a more complete study of the 2d field theory would require the inclusion of source terms for @xmath9 . ) z[,a]=dx^e^is[;x ] pe^i _ ma_dx^ , [ 1 ] with + [ 2 ] + & & \ { _ m x^ _ n x^ e^mn _ ( x(z ) ) + & & + ^ r^(2 ) ( x(z ) ) } , + & & e^mn_(x)=g^mng_(x)+ b_(x ) .
@xmath10 is the 2d curvature scalar on the manifold @xmath11 , @xmath12 the geodesic curvature of the boundary @xmath13 . in the case of a non - abelian gauge group @xmath8 , due to the path ordering implied in the wilson loop ,
the integrand of the functional integral appears not in the standard form as an exponential of a local action . to handle this complication
we introduce an auxiliary field @xmath14 living on the one - dimensional space of the variable @xmath15 parametrizing the boundary @xmath13 of the string world sheet @xmath11 .
we denote its free action by @xmath16 .
it has the propagator |_a ( s_1)_b ( s_2)_0=_ab ( s_2-s_1 ) [ 3 ] and couples to @xmath17 via the interaction term i|_a a_^ab(x(z(s)))_b ( s)_m x^(z(s))^m(s ) . [ 4 ] then we can write ( choosing @xmath18 ) & & + & & e^is_0[| , ] ( is[,|a ; x ] ) , [ 5 ] with s[,c;x]&=&s[;x][6 ] + & + & _ m c_(x(z(s)),s)x^ ds .
the field c_(x(z),s)=|(s)a_(x(z))(s ) [ 6a ] in contrast to @xmath19 is no longer matrix valued . the case of an abelian gauge field is included in the formalism . however
, then we can of course also drop the auxiliary field and put @xmath20 . in the following we interchange
the order of the @xmath9 and @xmath21 integration and adapt for our open string case the standard manipulations of the @xmath9-functional integral used to derive the t - duality rules in the closed string case @xcite .
apart from the final auxiliary field integration the only difference between abelian and non - abelian @xmath6 is the absence or presence of explicit @xmath15-dependence in @xmath22 .
we now assume the existence of one killing vector field @xmath23 and invariance of our partition function under a diffeomorphism in the direction of @xmath24 _ k = 0 , l_ka _
= d_a .
[ 7 ] choosing suitable adapted coordinates we have [ 8 ] + & & k^ = ( 1,0 ) , l_k = _ 0 . starting from an arbitrary gauge field fulfilling ( [ 7 ] ) we can arrive at @xmath25 by a gauge transformation .
then all our target space fields are independent of @xmath26 _ 0 = 0 , _ 0a _ = 0 . [ 9 ] in the rest of the paper we take flat 2d metrics and a flat boundary and disregard the dilaton field @xmath27 . its transformation under t - duality should arise from some functional jacobian as in the closed string case @xcite . to perform the dualization procedure we call the result of the @xmath9-integration in ( [ 5 ] ) @xmath28 $ ] and rewrite it with the help of a lagrange multiplier field @xmath29 as [ 10 ] + & & = dx^mdy_md^0 e^i|s[,c;x^m , y_m,^0 ] , with + & & \ { _ m x^m _ n x^n e^mn_mn+ 2_mx^m y_n e^mn_m0 + & & + y_m y_n e^mn_00 + 2 ^ 0^mn _ m y_n } + & & + _ m ( c_n _
n x^n+c_0y_n)^n ds . [ 11 ] the dualized version will be obtained by performing the @xmath30 functional integral . after a shift of the integration variable suited to decouple the integral in the closed string case we get
[ 12 ] + & + e^mn_mn_m x^m _ n x^n- k_m k_n } & + & _ m\{c_n _ n x^nz^n+v_m l^m - k_m l^m } ds , & with k^m = e^mn_0n_n x^n+^mn_n^0 , + l^m = c_0^m+x^0^m , + v_m = y_m+ .
[ 13 ] due to the dependence of @xmath31 on @xmath32 the @xmath33 integral even after the rescaling @xmath34 does not decouple .
since the @xmath33 integral is ultralocal we can use + & & _ mdv_me^ ( i_m v_ml^mds ) . [ 14 ] then the @xmath35 integral on the boundary @xmath13 produces a functional @xmath36 function which imposes for the remaining functional integral the constraints @xmath37 , i.e. 2^c_0 ( x^m(z(s)),s ) + x^0(z(s))=0 .
[ 15 ] including the remaining bulk @xmath35 integral into the overall normalization we finally arrive at ( the @xmath38 below the integration symbol indicates the boundary condition ( [ 15 ] ) . )
z[,c ] = _ d x^(is[,c ; x ] ) , [ 16 ] with ( note @xmath39 ) x^ = ( x^0,x^m ) , [ 17 ] a _
= ( 0,a_m ) , c _
= |a _ , [ 18 ] + [ 19 ] + [ 20 ] for later use we denote @xmath40 $ ] understood as a functional of @xmath41 and the function @xmath42 determining the boundary condition in ( [ 16 ] ) by @xmath43 z[,c ] = f[,c-2c_0 ] .
[ 21 ] the formulas ( [ 19 ] , [ 20 ] ) coincide with those in the closed string case @xcite .
the gauge field @xmath44 of the original model determines via ( [ 15 ] ) a dirichlet boundary condition for the functional integrand in the integration defining the dual model ( [ 16 ] ) .
what concerns the quantized theory , the dirichlet condition , therefore , has to be considered as an external constraint .
in contrast , on the classical level there are always equivalent formulations for a boundary condition , either as an external constraint or as part of the stationarity condition of the action , possibly after a suitable addition of some boundary interaction .
a discussion of t - duality and classical canonical transformations has been given in @xcite extending corresponding results for the closed string case @xcite .
in the abelian case we can drop the @xmath21 s and put @xmath20 , @xmath45= \hat z[\psi , a]$ ] . the boundary condition ( [ 15 ] )
has to be interpreted as a constraint confining the end points of the string to the hypersurface ( dirichlet brane ) @xmath46 with free movement on the brane . in the nonabelian case ( [ 16 ] ) still has a d - brane interpretation .
there is a one parameter family of d - branes @xmath47 the endpoint of the string at string world sheet boundary parameter value @xmath15 has to sit on the d - brane for parameter value @xmath15 . the final @xmath21-integration results in ordering the matrices sandwiched between @xmath48 and @xmath49 with respect to increasing @xmath15 .
but after performing the functional integral over the world surfaces @xmath50 there is no longer any correlation between a given target space point and @xmath15 .
the situation is more transparent if one treats ( [ 5 ] ) and ( [ 10 ] ) within the background field method ( @xmath51 ) .
then both in @xmath52 and @xmath53 all dependence on target space coordinates is realized via a classical string world sheet configuration @xmath54 .
the result of the @xmath21-integration is then @xcite + & & pf_bm [ , ; x_cl-2^a_0 ] .
[ 22 ] path ordering now refers to the classical path @xmath55 and involves both the matrices appearing in the second argument of @xmath52 as well as @xmath56 entering via the argument specifying the boundary condition .
the insertion of a matrix as boundary condition is performed _ after _
@xmath52 has been calculated with scalar ( not matrix valued ) @xmath15-dependent boundary condition . in this formalism
we can avoid wondering about the target space interpretation of matrix valued boundaries .
the situation is similar to dimensional regularization , the change from integer @xmath57 to complex @xmath57 is performed _ after _
@xmath58 has been calculated for integer @xmath57 .
the manipulations with the functional integral reviewed in the previous section are of pure formal nature .
no attention has been paid both to functional jacobians and renormalization issues . in the closed string case , studied in more detail so far ,
there are thorough arguments for the quantum equivalence of conformal theories t - dual to one another @xcite .
away from the points of conformal invariance the situation is less clear .
the buscher formulae ( [ 19 ] ) , ( [ 20 ] ) can be derived by pure classical considerations on the basis of canonical transformations @xcite .
the weyl anomaly coefficients , related closely to the renormalization group @xmath1-functions @xcite , fulfil at one - loop order the consistency equations derived from requiring ( [ 19 ] ) and ( [ 20 ] ) for the renormalized target space fields @xcite .
however , insisting on quantum equivalence in higher orders of renormalized perturbation theory requires a modification of the buscher rules starting at two loops @xcite.),([20 ] ) to hold for the bare target space fields . ]
we now want to comment on some simple observations concerning quantum equivalence of the models related by ( [ 17])-([20 ] ) in the case of open strings
. then all the problems present in the absence of a boundary of the world sheet are present , too .
in addition the dirichlet boundary condition should refer at first sight to the bare quantities . for shortness
we restrict ourselves to the abelian case and discuss the gauge field @xmath1-function only . in leading order of an expansion in the derivatives of the field strength @xmath59
this @xmath1-function is @xcite ( we absorb a factor @xmath60 into @xmath59 .
@xmath61 denotes the field strength for @xmath62 . )
^(a)_&=&(b+f)_^ _ [ 23 ] + & + & j^\ { d_(b+f ) _ + & & + ( b+f)_^ h _
( b+f)^ _ } with j^ = ( ( g-(b+f)^2)^-1)^ .
[ 24 ] the vanishing of this @xmath1-function is equivalent to the stationarity condition of the generalized born - infeld action @xcite & & [ 25 ] + & & d^dx e^- , = -e^- j^^(a ) _ .
[ 26 ] the action @xmath63 has a remarkable duality property @xcite . for target space fields of the type discussed in the previous section ( @xmath64 )
one finds by simple manipulations , based on the identity ( @xmath65 arbitrary matrix , @xmath66 ) @xmath67 the relation [ 27 ] + & & e^- . here
@xmath68 and @xmath69 are the metric and the antisymmetric tensor induced on the d - brane defined by @xmath70 i.e. g _ mn = g__mf^_nf^ [ 28 ] and similar for @xmath71 .
the fields @xmath72 are given by ( [ 18])-([20 ] ) , @xmath73 the @xmath1-functions of a @xmath0-model describing a string with its endpoints confined to some dirichlet brane have been calculated in ref.@xcite .
furthermore , in this paper it has been argued that the vanishing of the @xmath1-functions for the gauge field on the d - brane as well as that for the d - brane position is equivalent to the stationarity condition of the born - infeld action restricted to the d - brane @xmath74 $ ] . if this last statement was correct , due to ( [ 27 ] ) the gauge field part in the issue of quantum equivalence of t - dual models would be completely settled .
however , while we agree with the result of the @xmath1-function calculation of @xcite ( see also @xcite ) , the relation of these @xmath1-functions to the functional derivatives of the d - brane bi - action needs some more discussion . + the @xmath1-functions just mentioned are that for the gauge field @xmath75 living on the d - brane & & [ 29 ] + & & + ^nl\{d_l(+)_nm + & & + ( + ) _ m^ c_cne(+)^e _
l + & & + ( + ) _
m^ c_cf^k_nl } and that for the d - brane position @xmath76 ^(f)_&=&- _ [ 30 ] + & + & ^nl\{(+)_l^ c_nc - k_nl } , with @xmath77 related to @xmath78 like @xmath79 to @xmath80 in ( [ 24 ] ) and the curvature of the d - brane given by k^ _
nl = d_n_lf^ .
[ 31 ] the two @xmath1-functions fulfil the equation ^(a)_m&=&^nld_l(+)_nm[32 ] + & -&(+)_m^ n _
nf^^(f ) _ . for the functional derivatives of @xmath81 $ ] with respect to @xmath82 and @xmath83 ( the last one for variations orthogonal to the d - brane . )
we find & = & -e^ [ 33 ] + & & ( ^(a)_m-^nl(+)_m^ c_cf^k_nl ) , & = & e^^(f)_[34 ] + & -&2b__nf^ .the last two formulas disagree with ref.@xcite by the second terms on their r.h.s .
we want to stress that at least ( [ 33 ] ) is obvious without any calculation .
the subtraction of the curvature term in ( [ 33 ] ) ensures that the gauge field derivative has the same form as in ( [ 23 ] ) .
this has to be correct since the gauge field @xmath75 is defined on the d - brane only .
in contrast to @xmath84 and @xmath85 it is not induced from target space fields and does not `` feel '' the embedding .
now the problem connected with ( [ 33]),([34 ] ) is due to the presence of an inhomogeneity ( the curvature term ) in the otherwise linear relation between the @xmath1-functions and the functional derivatives of the bi - action .
we no longer can claim the equivalence of the vanishing of the @xmath1-functions and the stationarity of @xmath63 .
as far as one is interested in the equivalence of bi stationarity and the conformal invariance condition of the d - brane model there is still a way out .
the weyl anomaly coefficients are given by the @xmath1-functions plus some additive terms @xcite usually contributing in higher orders only . while the @xmath1 s are not diffeomorphism invariant the weyl anomaly coefficients are invariant under diffeomorphisms .
hence as long as these additive terms are not calculated one should state conformal invariance as the vanishing of the @xmath1-functions up to terms which can be generated by a diffeomorphism . the disturbing term in ( [ 33 ] ) has the form of a product of @xmath86 times a vector .
therefore , it can be generated by a combination of a diffeomorphism and a gauge transformation .
of course a final statement can be made only after a calculation of the difference between the weyl anomaly coefficients and the corresponding @xmath1-functions has been done .
+ * acknowledgement * + we would like to thank j. borlaf , s. frste , a. kehagias , l. palla , c. preitschopf , c. schmidhuber and a. tseytlin for useful discussions .
j. balog , p. forgacs , z. horvath , l. palla , ( 1996 ) 121 , hep - th/9606187 + j. balog , p. forgacs , z. horvath , l. palla , _ nucl .
b ( proc . suppl . ) _ * 49 * ( 1996 ) 16 , hep - th/9601091 + j. balog , p. forgacs , z. horvath , l. palla , _ contribution to these proceedings _ | this contribution gives in @xmath0-model language a short review of recent work on t - duality for open strings in the presence of abelian or non - abelian gauge fields .
furthermore , it adds a critical discussion of the relation between rg @xmath1-functions and the born - infeld action in the case of a string coupled to a d - brane . h february 1997 hub - ep-97/2 + hepth@xxx/9702018 * remarks on t - duality for open strings@xmath2 * + 7.mm h. dorn , h .- j .
otto + 0.1 cm _ institut fr physik _ + _ humboldt - universitt berlin _ + _ invalidenstr.110 , d-10115 berlin _ * abstract * _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ this contribution gives in @xmath0-model language a short review of recent work on t - duality for open strings in the presence of abelian or non - abelian gauge fields .
furthermore , it adds a critical discussion of the relation between rg @xmath1-functions and the born - infeld action in the case of a string coupled to a d - brane . ' '' '' width 5.cm 2.mm @xmath3talk given at the @xmath4 international symposium ahrenshoop on the theory of elementary particles , buckow , august 27 - 31 , 1996 ; to appear in nuclear physics b ( proc .
suppl . ) .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ |
studies of nanoscale electronic structures hold important promise both as laboratories for few - body , interacting quantum mechanical systems and as technological testbeds for future classical or quantum computing technologies .
novel probe technologies @xcite are very important for studying electronic properties in nanoscale systems because they are often beyond the resolution of conventional imaging techniques like optical microscopy , and traditional transport measurements can only measure spatially averaged physical properties such as the conductance or the current .
one scanning probe microscopy ( spm ) @xcite technique utilizes a charged metallic tip to perturb the local electronic density in a nanoelectronic structure while monitoring the resulting change in transport properties .
using this technique one can obtain spatially resolved measurement of the electronic properties , including the local electron density and , in principle , the wavefunction itself in the case of a one - electron system @xcite(see discussions below ) .
this imaging technique has been fruitfully applied to study the flow of ballistic electrons across a range of two dimensional heterostructures .
recently , a series of experiments @xcite applied spm techniques to study quantum wires . in these experiments
a segment of an inas nanowire lying on top of a two - dimensional sio@xmath2 layer was isolated from the rest of the wire by two short inp segments , forming a one - dimensional quantum dot with lithographically defined boundaries .
a negatively charged probe scanned controllably the two dimensional area around the wire and the conductance across the 1d quantum dot was measured as a function of the probe location .
both the voltage of the probe and its height above the surface can also be independently varied .
motivated by these new experimental possibilities , we turn to exact diagonalization techniques to study the conductance response of a few - electron quantum dot as a function of a spatially varied probe potential , in order to illustrate the kind of information that can be extracted in the case of a system of several electrons .
schematic geometry of coulomb blockade microscopy of a quantum wire containing four electrons .
when calculating the electron - electron interaction and the electron - tip interaction , we assume that the inp barriers have zero thickness , the inas wire is infinitely long , and the substrate layers extend to infinity in x and y directions.,scaledwidth=40.0% ] ( color online ) ( dotted)interaction potential @xmath3.(solid , long and short dashed ) tip @xmath4 potentials with a tip charge @xmath5 and locations @xmath6 where @xmath7.,scaledwidth=48.0% ] we consider a uniform inas ( dielectric constant @xmath8 , bohr radius @xmath9 ) nanowire of radius @xmath10 nm , which lies in vacuum atop a sio@xmath11 ( @xmath12 ) layer @xmath13 nm thick , separating it from conducting doped bulk silicon ( see fig . [
fig : geometry ] ) .
electrons are modelled as point charges traveling along the center axis of the wire , confined to interval @xmath14 by hard walls , representing the inp layers .
we consider length @xmath15 from @xmath16 nm to @xmath17 nm .
the electron - electron interaction @xmath18 was calculated using the commercial finite - element program comsolto solve the classical poisson equation for a point charge on the axis of an infinite wire above in a substrate with the geometry described in fig .
[ fig : geometry ] . at short distance @xmath19 ,
the potential was softened to account for the finite thickness of the electron wavefunction , by replacing @xmath20 with @xmath21^{-1/2}$ ] . following the approximation used by topinka @xcite
, we model the negatively charged probe as a fixed point charge of strength @xmath22 at a location @xmath23 relative to the center point of the wire .
this gives rise to a one - body potential @xmath24 for an electron on the wire axis at point @xmath25 , which we again obtain by solving the poisson equation ( results are shown in fig .
[ fig : tip ] ) . in this paper , we diagonalize the exact 1d many - body hamiltonian with the lanczos method @xcite for up to electron number @xmath26 : @xmath27 where @xmath28 is the full many - body wavefunction , depending on the position @xmath29 and spin @xmath30 of the electrons . to connect to the experimentally observable variables
, we consider the coulomb blockade peak @xmath31 of the transition from @xmath32 to @xmath33 electron ground states .
the conductance through the quantum wire is maximum when the chemical potential difference between the lead and the wire , controlled by the voltage @xmath34 on a back gate , is equal to the ground state energy difference between the two states in question .
we may write this condition as @xmath35 , where @xmath36 is a constant and @xmath37 is the proportionality constant between changes in the back gate voltage and the chemical potential in the quantum dot .
we probe the electronic states in the quantum wire through the dependence of @xmath38 on the tip position @xmath23 and potential strength @xmath22 .
an interesting set of spatially resolved information about the electrons in the wire can be extracted from this function , and we call this method `` coulomb blockade microscopy ''
. it is a special application of the `` scanning probe microscopy '' developed by the westervelt group @xcite . in calculations in this paper we focus on the transition from @xmath39 to @xmath26 electrons , but most of our conclusions are easily generalizable to other ground state transitions .
finally , we note that for four non - interacting electrons with spin in a wire of radius @xmath40 , when the dot length @xmath41 , the lowest four single particle energy levels are all longitudinal modes .
the shortest wire length we consider in this paper @xmath42 , we therefore expect the wires under consideration can be well approximated as strictly 1d under the assumption of weak interaction effects .
indeed , the gap between the transverse ground state and first excited state for our wire is @xmath43 .
the most confined geometry we discuss is shown in fig .
[ fig : length_strong_tip ] : a @xmath44 wire with four electrons under an extremely strong tips , squeezing them into the both ends of the wire . even in that case , the _ total _ energy per electron is less than @xmath45 , not enough to cause an excited transverse mode , with full interaction effects taken into account .
thus we expect the quantum wire to be well approximated by a 1d model for all the parameters we explored in this paper .
the existence of higher transverse modes will lead to a renormalization of the effective electron - electron interaction ( generally a softening of the potential at short distances ) even if the energies of the modes are well above the fermi energy .
this is because two electrons that come close together in the lowest transverse mode will mix , virtually , with states where the electrons are in excited modes .
as our calculations use , already , a crude phenomenological cut - off at short distances , we do not include explicitly effects of this renormalization .
technically , the existence of higher transverse modes can also lead to three - body and four - body effective interactions , arising from three - body and four - body collisions , but we do not expect such higher body terms to be important in the wires under consideration . we want to emphasize that although we choose a specific wire geometry and a specific form of interaction that model some characteristics of the setup of the westervelt group s on - going experiments , most features of the coulomb blockade micrographs we discuss below are applicable to any 1d quantum dot system under a mobile potential .
indeed , our discussions of the qualitative features of coulomb blockade micrographs under both the weak - tip and the strong - tip limits rely only on the general properties of 1d electronic systems , independent of the specific geometry and interaction we adopt .
our numerical results mostly serve for illustrative purposes .
( color online ) electronic densities for a @xmath44 wire and for a @xmath46 wire in absence of tip potential . only the right half is shown , as the plot is symmetric about @xmath47 .
we rescale the @xmath25 and @xmath48 with wire length @xmath15 to ease the comparison.,scaledwidth=40.0% ] in the absence of a probe potential @xmath49 , both the @xmath39 and @xmath26 wires the electronic density profile @xmath48 undergoes a crossover as a function of @xmath15 from a liquid state characterized by a @xmath50 friedel oscillations to a quasi - wigner crystal state characterized by a @xmath51 density oscillation .
such a crossover from a liquid state to quasi - wigner states with a decrease in density a very generic phenomenon for 1d interacting fermion system @xcite . indeed , for any interacting decaying no faster than @xmath52 at long distance a quasi - wigner crystal state is known to emerge at low density @xcite .
this requirement for interaction will hold for a system with long screening length as compared to mean inter - particle distance , as is the case for our geometry when the screening doped silicon layer is relatively far ( @xmath53 ) away . for our specific geometry and interaction
, the crossover happens at around density @xmath54 .
a wigner crystallized density variation is shown in the dashed curve of fig .
[ fig : length_density ] for a quantum dot of @xmath46 , whereas for @xmath55 the four electron density exhibits friedel oscillations .
( color online ) coulomb blockade micrographs for a 1d dot with @xmath56 nm and tip charge @xmath57 for three tip potential shown in fig.[fig : tip ] .
again , the right half is shown.,scaledwidth=40.0% ] ( color online ) @xmath57 weak tip limit for the @xmath39 to @xmath26 coulomb blockade transition for densities shown in fig .
[ fig : length_density ] .
the tip distance to the wire is @xmath58 . only the right half is shown.,scaledwidth=40.0% ] ( color online ) @xmath59 strong tip limit for the @xmath39 to @xmath26 coulomb blockade transition for densities shown in fig .
[ fig : length_density ] the tip to wire distance is @xmath60 . only the right half is scanned .
, scaledwidth=40.0% ] now we introduce a weak tip potential , corresponding to a negatively charged tip of strength @xmath57 , scanning above the center axis of the quantum wire along its direction @xmath61 , with the tip location vector @xmath62 . for a 1d quantum dot of length @xmath46 ,
which as shown in fig .
[ fig : length_density ] has @xmath51 wigner - crystal density variation , let us consider the three tip heights above the quantum wire , @xmath63 , corresponding to the three tip potential shown in fig .
[ fig : tip ] .
the resulting coulomb blockade peak position @xmath38 as a function of the tip coordinate @xmath64 along the wire , i.e. the coulomb blockade micrograph , is shown in fig .
[ fig : height_gap ] . clearly , in fig .
[ fig : tip ] the closer the tip approaches the wire , the more localized is the tip potential and a sharper tip potential make it easier to resolve the density variations , this is reflected in coulomb blockade micrograph scans in fig .
[ fig : height_gap ] . at @xmath65 from the tip to the center of the wire , the @xmath51 density oscillation of the quasi - wigner crystal state on the right can be detected in the coulomb blockade micrograph , whereas when @xmath66 away , the tip potential becomes much too broad to resolve the fine features of the density oscillations .
we note that although the resolution of the tip is largely determined by the distance @xmath67 , the contrast of a coulomb blockade micrograph , i.e. the magnitude of the @xmath51 variations in the micrographs , can be improved by modestly increasing the tip potential .
by contrast , in fig .
[ fig : length_gap ] the @xmath55 nm micrograph at @xmath58 does not show features of wigner crystal oscillations . however ,
this micrograph does not by itself give a clearcut indication of the absence of wigner crystal order for @xmath55 nm . with the current interaction and tip parameters
, one can not observe the crossover from the wigner crystal to the friedel oscillations because it happens at a inter - particle spacing @xmath68 , below the resolution of the micrograph even at @xmath69 nm . to be more specific
, we may define the onset of wigner - crystal order for our four - electron system as the length @xmath15 at which there first appears a local minimum of the mean density @xmath48 in the vicinity of @xmath70 . according to our calculations ,
this should occur at l=135 nm .
however , with the tip at height 30 nm , in the weak charge limit , the resolution of the micrograph is of the order of 60 nm , so we would not see a secondary minimum in the micrograph signal until @xmath71 to gain a more intuitive understanding of a weak tip coulomb blockade micrograph , we observe that a weak tip only slightly disturbs the electron density as it scans across the wire , thus , a simple first order perturbation theory should be a good approximation to compute the ground state energy in the presence of the tip potential : @xmath72 where @xmath48 is the non - interacting ground state density and @xmath73 is its energy .
we have checked that for tip charges up to @xmath74 the simple first order perturbation theory gives a decent fit to both the ground state energy and the coulomb blockade micrograph . since both the width and the center location of the tip potential @xmath75 can be adjusted experimentally , the coulomb blockade microscopy with a weak tip potential provides a flexible way to map the electronic densities in a quantum dot . indeed , as an example of such flexibility , we find that one can improve the `` contrast '' of a coulomb blockade micrograph ; i.e. , the prominence of the spatial variations in micrographs like fig .
[ fig : height_gap ] as compared with the total energy shift @xmath76 , can be improved by slightly increasing the tip potential while still staying within the weak tip perturbative approximation .
electronic densities in @xmath46 , @xmath26 quantum wire as a tip @xmath59 scan through @xmath77 .
the density throughout the entire wire @xmath78 is shown.,scaledwidth=48.0% ] in the opposite limit of strong tip , the coulomb blockade tip scans present a very different physical picture . in fig .
[ fig : length_strong_tip ] we observe that irrespective of whether the electronic state is liquid or wigner - crystal like as shown in fig .
[ fig : length_density ] , the coulomb blockade micrographs show similar behavior : in the case of @xmath39 to @xmath26 transition , both the @xmath55 nm and @xmath56 nm wire show two relatively sharp peaks for a large tip charge @xmath79 .
this is in contrast with the case of a weak tip fig .
[ fig : length_gap ] , where the coulomb blockade micrographs show smooth spatial dependence as well as sensitivity to the electronic states in the absence of the tip potential . to understand the physics of this strong tip limit
we note that the two sharp cusps in fig .
[ fig : length_strong_tip ] represents discontinuous slope changes in the @xmath26 electron ground state energy as a function of tip position @xmath64 .
similarly the deep valley in the figure corresponds to a cusp in @xmath39 ground state energy .
the origin of these three discontinuities in slopes can be seen in fig .
[ fig : strong_tip_density ] . in this limit ,
the negatively charged tip potential is so strong that it depletes the electronic density under it .
thus the tip creates an effective partition of the electrons in the wire into left and right sub - quantum dot . as shown in fig .
[ fig : strong_tip_density ] , as the tip move from the center to right of the wire with four electrons , the partitions of the electrons undergoes two abrupt transitions @xmath80 .
these two transitions correspond to the two upward cusps shown in the @xmath26 curve in fig .
[ fig : length_strong_tip ] .
similarly , the discontinuous slope change shown on the @xmath39 curve of the fig .
[ fig : length_strong_tip ] corresponds to the transition between the @xmath81 partition of the ground state .
thus the three discontinuities seen in the coulomb blockade micrographs in fig .
[ fig : length_strong_tip ] correspond to , alternately , the transitions between the integer partitioning of total electron numbers in the @xmath39 and @xmath26 system .
the upward slope of the curve near @xmath82 reflects an additional downward cusp at the origin , due to the transition ( 1,2)@xmath83(2,1 ) in the @xmath39 wire .
schematic illustrations of partition of the four - electron state by the potential barrier under a strong tip potential . in ( a ) and
( b ) , single - electron energy levels are shown for non - interacting electrons , for two positions of the tip .
because of spin degeneracy , we find either two electrons in each well , or all four in the same well , when the energy of the single - electron ground state on the right becomes higher than the first excited level on the left . in ( c)-(e ) , we `` incorporate '' interaction energy into `` single electron levels '' schematically by plotting the energy needed to add an additional electron . in drawing energy levels this way , it is clear that the spin degeneracy in the non - interacting case is lifted by electron interaction , and an additional ( 3,1 ) partition will appear for the `` energy level '' arrangement in ( d ) . filled circles show electrons in occupied levels.,scaledwidth=50.0% ] to better understand the transitions between different partitions , let us consider the transitions in a model of @xmath26 electrons , with spin , which have no coulomb repulsion between them but interact with a repulsive tip potential .
the scenario is illustrated in fig .
[ fig : strong_schematic ] , panels @xmath84 .
when tip is near the center of the wire , the electrons are partitioned @xmath85 and both electrons in each side reside in the single - particle ground state . as the tip moves rightward , the energy levels rise in the right partition and fall in the left .
when the first excited level on the left partition crosses the ground state on the right , @xmath86 electrons will move to the left partition . therefore ,
contrary to the interacting case , there is no energetically favorable state of @xmath87 partitioning in the non - interacting system . in the @xmath39 case ,
the @xmath88 partition is not affected by this , and for non - interacting system the transition @xmath81 will coincide with the transition in @xmath26 , so the coulomb blockade micrograph will show only a single peak .
this analysis can also be generalized to a wire containing multiple non - interacting electrons , such that all the @xmath89 partitionings of electron number will be missing . as schematically
illustrated in fig .
[ fig : strong_schematic ] , panels @xmath90 , when we take electron interaction into consideration , the one and two electron state would no longer be degenerate in either partition , so contrary to the non - interacting scenarios in panel @xmath84 , here a @xmath87 partition can survive as an intermediate stage between the @xmath85 and @xmath91 partitions . with the non - interacting case in mind , we postulate that the distance between two peaks in @xmath39 to @xmath26 micrographs , corresponding to the tip positions where @xmath87 partitioning in the @xmath26 wire is stable , can serve as an indicator of the relative importance of the interaction energy versus the sum of kinetic and single particle potential energies .
the less important interaction is compared to single particle energies , the less splitting would the one and two particle energies be , and the smaller is the region of stable @xmath87 partition .
this can be seen in fig .
[ fig : length_strong_tip ] .
the potential energy should have a larger share in the total energy in the longer wire with lower electronic density , and indeed we observe that the longer wire has a wider distance between the two peaks marking @xmath92 and @xmath93 transitions .
( color online ) . @xmath94 and @xmath95 intermediate tip potentials for @xmath39 to @xmath26 coulomb blockade transition in a wire of length @xmath96 .
the tip distance to the wire is @xmath58 .
only the right half is shown.,scaledwidth=40.0% ] ( color online ) electronic density for a @xmath96 wire under three tip potentials , with effective charges @xmath97 , @xmath94 and @xmath95 . only the right half is shown , as the plot is symmetric about @xmath47 .
we rescale the @xmath25 and @xmath48 with the wire length @xmath15.,scaledwidth=40.0% ] we have carried out calculations with various tip charges intermediate between the two limits discussed above . in general , the larger the charge on the tip , the more readily one sees the secondary minima in the micrograph signal , which are seen in fig .
[ fig : length_strong_tip ] for a charge @xmath59 , even at @xmath98 .
as one illustration , for a wire of length 180 nm , with a tip height of 30 nm , we find that the micrograph signal shows a secondary minimum at @xmath99 for @xmath95 , but shows no secondary minimum when @xmath94 , as seen in fig .
[ fig : micrograph_tip ] however , if we calculate the electron density in the wire when the tip is over the center of the wire ( @xmath100 ) , we find that a tip charge of @xmath94 is enough to substantially modify the density relative to the density in the absence of the tip .
as seen in fig .
[ fig : density_under_tip ] , the electron density below the tip , at @xmath47 is reduced by a factor of three relative to the density with no tip charge .
nevertheless , the oscillations seen in @xmath48 remain qualitatively similar to structure seen in the absence of the tip .
for example , the charged tip only pushes out the position of the secondary minimum in the density from @xmath101 to @xmath102 .
in general , when we increase the tip potential to intermediate values , the resolution of our micrographs improves compared with the weak tip limit .
this is evident in the fact that at @xmath95 we can already see signs of wigner crystallization at @xmath96 , in contrast with the weak tip case with @xmath57 where we can only detect quasi - wigner crystal at @xmath103 . on the other hand , at @xmath95
there is _ no _ signature of wigner crystallization when the electrons are in a liquid state in a @xmath44 wire .
thus we do not have `` false positive '' signature of quasi - wigner crystal , in contrast with the strong tip limit described in section[strong ] , where the micrographs show @xmath33 peaks for a @xmath33-electron wire regardless whether , in the wire in the absence of the tip , the electrons are in a liquid or a quasi - wigner crystal state .
because of these two characteristics , an intermediate tip potential may help an experimentalist to reliably detect the presence of a quasi - wigner crystal state in a shorter , higher density wire closer to the crossover from a liquid state .
beyond the system of interacting electrons discussed above , a possible further application of coulomb blockade microscopy is to experimentally `` map '' the rugged potential landscape produced by wire inhomogeneities and charged impurities in the substrate .
one would focus on the transition from @xmath0 to @xmath1 state , in which case the coulomb blockade micrograph would reveal information about the single particle density . by inverting the transformation in eq .
[ equ : perturbation ] , one may be able to approximately obtain the single particle ground state density @xmath48 . in the absence of an external magnetic field
, the ground state wavefunction @xmath104 has no nodes and can be chosen to be @xmath105 it is then straightforward to invert the schrdinger s equation to extract the potential landscape from the single particle wavefunction .
in summary , in this paper we show that tracking the peak position shift as a charged mobile tip moves above and across a nanowire , a technique we term coulomb blockade microscopy , can reveal spatially - resolved information about the electronic density and states of a quantum 1d dot .
a weak tip potential can serve as a probe with a tunable width , to reveal the spatial distribution of the electronic density in the wire .
a strong tip potential that depletes part of the wire can be used to manipulate individual electrons from one partition to the other , and the accompanying coulomb blockade micrograph can indicate the transitions between different partitionings . furthermore , a feature of the resulting micrograph , the distance between peaks marking the @xmath89 partitioning , can serve as an indicator of the relative strength of the interaction . in this paper
we have chosen extreme values of the tip charge @xmath22 to illustrate the physics in the two limits .
however , our calculations show that the discussions above hold true for a wider range of moderately small and large values of @xmath22 . to obtain a quantitative description of the energy shifts expected in coulomb blockade microscopy , particularly in the intermediate coupling regime ,
we see that it is necessary to perform a realistic calculation , which takes into account both the electron - electron interaction and the non - linear effects of the charged tip on the electronic state of the wire .
if one is prepared to carry out such a calculation , however , coulomb blockade microscopy can be a powerful probe of interaction effects in the wire .
we would like to thank erin boyd , halvar trodahl and jesse berezovsky and especially bob westervelt for helpful discussions . this work is supported in part by nsf grants phy-0646094 and dmr-0906475 .
numerical work was performed in part at the center for nanoscale systems , a member of the national nanotechnology infrastructure network ( nnin ) supported by nsf award ecs-0335765 .
jq is also supported in part by nim and dfg through sfb 631 . | motivated by recent experiments by the westervelt group , which used a mobile tip to probe the electronic state of a segmented nanowire , we calculate shifts in coulomb blockade peak positions , as a function of tip location , which we term `` coulomb blockade microscopy '' .
we show that if the tip can be brought sufficiently close to the nanowire , one can distinguish a high density electronic liquid state from a wigner crystal state by microscopy with a weak tip potential . in the opposite limit of a strongly negative tip potential
, the potential depletes the electronic density under it and divides the quantum wire into two partitions .
there the tip can push individual electrons from one partition to the other , and the coulomb blockade micrograph can clearly track such transitions .
we show that this phenomenon can be used to qualitatively estimate the relative importance of the electron interaction compared to one particle potential and kinetic energies .
finally , we propose that a weak tip coulomb blockade micrograph focusing on the transition between electron number @xmath0 and @xmath1 states may be used to experimentally map the one - particle potential landscape produced by impurities and inhomogeneities . |
in this appendix , we present the details of our numerical simulations . our results mainly rely on a matrix - product state simulation based on a tebd code @xcite , so an instance of a tensor network state simulation . in order to corroborate the results ,
we have furthermore employed an exact diagonalisation code @xcite that uses the particle number symmetry and keeps track of the time evolution with a runge - kutta integration scheme . for the non - interacting case ,
further checks were performed by an explicit simulation of the dynamical evolution of the covariance matrix , which takes a particularly easy form in this case . for short times and the system sizes that can be achieved with exact diagonalisation , the codes agree up to a negligible error , thus also demonstrating that the chosen step size in the @xmath70-th order trotter decomposition used in tebd @xcite of @xmath71 does not produce significant errors .
this leaves only two potential error sources , the fact that numerics necessarily simulate a finite system and the possibility of discarded weights accumulating over time .
performing a finite size scaling , we find that comparably small systems are already indistinguishable from the thermodynamic limit for the quantities considered here , see fig .
this is in agreement with the very slow growth of lieb - robinson cones expected in these disordered systems . to be on the safe side
, we nevertheless carried out our numerical analysis on systems with @xmath72 sites and open boundary conditions .
having demonstrated that the considered system size is indistinguishable from the thermodynamic limit only leaves the discarded weight as potential error source , see fig .
the time evolution of this quantity , which is directly connected to spatial entanglement entropies , depends strongly on the chosen disorder realisation . in order to keep this discarded weight small enough ,
we increase the bond dimension in the simulation in a three - step procedure up to @xmath73 , which is sufficient to guarantee a discarded weight smaller than @xmath74 for all disorder realisations . | the phenomenon of many - body localised ( mbl ) systems has attracted significant interest in recent years , for its intriguing implications from a perspective of both condensed - matter and statistical physics : they are insulators even at non - zero temperature and fail to thermalise , violating expectations from quantum statistical mechanics . what is more , recent seminal experimental developments with ultra - cold atoms in optical lattices constituting analog quantum simulators have pushed many - body localised systems into the realm of physical systems that can be measured with high accuracy . in this work ,
we introduce experimentally accessible witnesses that directly probe distinct features of mbl , distinguishing it from its anderson counterpart .
we insist on building our toolbox from techniques available in the laboratory , including on - site addressing , super - lattices , and time - of - flight measurements , identifying witnesses based on fluctuations , density - density correlators , densities , and entanglement .
we build upon the theory of out of equilibrium quantum systems , in conjunction with tensor network and exact simulations , showing the effectiveness of the tools for realistic models .
many - body localisation provides a puzzling and exciting paradigm within quantum many - body physics and is for good reasons attracting significant attention in recent years .
influential theoretical work @xcite building upon the seminal insights by anderson on disordered models @xcite suggested that localisation would survive the presence of interactions .
such many - body localised models , as they were dubbed , would be insulators even at non - zero temperature and exhibit no particle transport .
maybe more strikingly from the perspective of statistical physics , these many - body localised models would fail to thermalise following out of equilibrium dynamics @xcite , challenging common expectations how systems `` form their own heat bath '' and hence tend to be locally well described by the familiar canonical gibbs ensemble @xcite .
following these fundamental observations , a `` gold rush '' of theoretical work followed , identifying a plethora of phenomenology of such many - body localised models .
they would exhibit a distinct and peculiar logarithmic scaling of entanglement in time @xcite , the total correlations of time averages have a distinct scaling @xcite , many hamiltonian eigenstates fulfil area laws @xcite for the entanglement entropy @xcite and hence violate what is called the eigenstate thermalisation hypothesis @xcite . the precise connection and interrelation between these various aspects of many - body localisation is just beginning to be understood @xcite , giving rise to a vivid discussion in theoretical physics .
these theoretical studies have recently been complemented by seminal experimental activity , allowing to probe models that are expected to be many - body localised in the laboratory under remarkably controlled conditions @xcite .
this work goes much beyond earlier demonstrations of anderson localisation in a number of models @xcite , in that now actual interactions are expected to be relevant .
such ultra - cold atomic systems indeed provide a pivotal arena to probe the physics that is at stake here @xcite .
what is still missing , however , is a direct detection of the rich phenomenology of many - body localisation in the laboratory . rather than seeing localisation and taking the presence of interactions for granted
, it seems highly desirable to make use of these novel exciting possibilities to directly see the above features , distinctly separating the observations from those expected from non - interacting anderson insulators .
such a mindset is that of `` witnessing '' a property , somewhat inspired by how properties like entanglement are witnessed @xcite in quantum information . in this work
, we aim at capturing precisely those aspects of the rich phenomenology of many - body localisation that are directly accessible with present experimental tools .
we would like to provide a `` dictionary '' of possible tools , as a list or a classification of features that can be probed making use of only in situ site resolved measurements , including the measurement of density - density correlation and time of flight measurements , in conjunction with a variation of densities . in this way
, we aim at identifying a comprehensive list of features that `` could be held responsible '' for mbl , based on data alone .
while all we explicitly state is directly related to cold atoms in optical lattices , a similar approach is expected to be feasible in continuous cold bosonic atoms on atom chips @xcite , where correlation functions of all orders can readily be directly measured .
we leave this as an exciting perspective .
the setting we focus on is that of interacting ( spin - less ) fermions placed into a one - dimensional optical lattice , a setting that prominently allows to probe the physics under consideration @xcite .
such systems are well described by @xmath0 where @xmath1 denotes a fermionic annihilation operator on site @xmath2 and @xmath3 is the local particle number operator .
the disorder in the model is carried by the local potential - strength @xmath4 , which is drawn independently at each lattice site @xmath2 according to a suitable probability distribution .
experimentally , the disorder can either be realised by superposing the lattice with an incommensurate laser or by speckle patterns @xcite . from eq .
one obtains the disordered heisenberg chain @xcite by setting @xmath5 and scaling the disorder by a factor two . to keep the discussion conceptually clear , just as ref .
@xcite we make use of a uniform distribution on the interval @xmath6 $ ] , where we refer to @xmath7 as the disorder strength .
thus , for @xmath5 the ergodic to mbl phase transition is approximately at @xmath8 @xcite .
the phase diagramme is best known for @xmath9 corresponding to the non - interacting anderson insulator and for @xmath5 , the mbl phase . to add a flavour of usual phase transitions order parameters such as total correlations @xcite , fluctuations of local observables @xcite or the structure of the eigenstates @xcite have been suggested . while these quantites impressively signal the transition , it is not a priori clear whether they can be implemented in an actual experiment .
recent numerical studies@xcite show that pump - probe type setups and novel instances of spin noise spectroscopy @xcite as well as utilising mbl systems as a bath @xcite are indeed suited to distinguish the above phases , albeit experimental realisations of this endeavour appear to need substantial changes and innovations in realistic setups .
another possibility for the phase distinction , which has prominently been carried out experimentally @xcite , is given by observing the behaviour of quasi two - dimensional systems in comparison to their one dimensional counterparts .
while this impressively demonstrates the capabilities of optical lattices as platforms for quantum simulations , it does not test the properties of mbl in one dimension as such .
we set out to find comparably strong and direct signatures of one dimensional mbl , which however rely on simple established measurement operations .
hence , we start by summarising the measurements , which we conceive to be feasible in an optical lattice experiment .
we now turn to specifying what measurements we consider feasible in optical lattices with state of the art techniques . for this
, we focus on the following two types of measurements : * in - situ * : an in - situ measurement detects the occupation of individual lattice sites .
this technique only allows to resolve the parity of the particle number on each site , which for fermions constitutes no limitation , however . using the fact that single - shot measurements are performed , also higher moments like density - density correlators
can be extracted from this kind of measurements .
both ramifications will be used . * time - of - flight * : the time - of - flight ( tof ) measurement extracts position - averaged momentum information of the form @xmath10 where @xmath11 are the positions of the lattice sites , @xmath12 reflects the wannier functions in momentum space , and @xmath13 is a constant derived from the mass of the particles and the lattice constant . the main goal of this work is to identify key quantities that indicate that the system indeed is many - body localised based on measurement information extracted using these two techniques . here
, we want to show both that the system is localised and that it is interacting .
thus , we also want to convincingly detect the difference between an mbl system and a non - interacting anderson insulator . in order to approach this task
, we will look at the time evolution of an initial state that is particularly easy to prepare experimentally relying on optical super - lattices @xcite , namely an alternating pattern of the form @xmath14 this initial product state will , during time evolution , build up entanglement and become correlated @xcite .
naturally , this is far from being the only choice for an initial state and alterations in this pattern and correspondingly locally changing particle and hole densities would surely be insightful , specifically since a modulation of the density already points towards interactions in the mbl phase being significant .
in this work , we put emphasis on measurements although preparation procedures as the above mentioned density variations are an interesting problem in its own right .
however , as we will demonstrate , the above defined initial state already captures the colourful phenomenology of mbl in all of its salient aspects .
a fundamental characteristic of mbl is the presence of local constants of motion @xcite .
they are approximately local operators @xmath15 whose support is centered on lattice site @xmath2 , but which nevertheless commute with the hamiltonian , i.e. , @xmath16=0 $ ] .
these operators are mainly supported on a region with diameter @xmath17 , corresponding to the localisation length scale of the system .
in fact , in the mbl regime , the dynamics can be captured by a phenomenological model in terms of a set of mutually commuting quasi - local constants of motion , @xmath18 such a hamiltonian constitutes a second order approximation of what is known as the @xmath19-bit model @xcite to exemplify the dynamics . here
@xmath20 again denotes a quasi - local integral of motion centred on site @xmath21 , @xmath22 is a random onsite potential and the coupling strength @xmath23 between constants of motion is assumed to decay suitably fast in their distance @xmath24 . in particular , it is expected that the dynamics generated by the hamiltonian defined in eq . in the mbl regime corresponding to @xmath5 and @xmath25 can be well captured by the @xmath19-bit model .
this phenomenological model gives rise to a separation of time scales in the evolution into two regimes .
initially , there is a fast regime , where the evolution takes place mainly inside the support of each local constant of motion @xmath26 .
hence , for this time scale , transport is unconstrained and particles and energies can move freely over the localisation length .
correspondingly , information can spread ballistically . beyond the localisation length , the dynamics
is dominated by the coupling of the constants of motion , given by the second term in eq .
@xcite .
the intuition is that this evolution does not facilitate particle or energy propagation , leading to a complete break - down of thermal and electric conductivity .
nevertheless , the couplings between distant constants of motion allow for the creation of correlations over arbitrary length scales given sufficient time .
this dephasing mechanism in turn makes it possible to send information and yields an explanatory mechanism for the observed slow growth of entanglement @xcite , measured as the von neumann entropy of the half chain of an infinite system @xmath27 ( in landau notation ) .
mathematically , these two dynamical regimes are best distinguished by the effect of a local unitary excitation on distant measurements .
more precisely , given a local measurement @xmath28 supported in a spatial region @xmath29 and a unitary @xmath30 corresponding to a local excitation in a region @xmath31 , we wish to bound the change in expectation value of @xmath32 induced by the unitary excitation .
this can be cast into a lieb - robinson bound @xcite of the form @xmath33 where @xmath34 a constant depending on the support of @xmath29 . for the connection between different zero velocity lieb - robinson bounds and the necessity of a linear @xmath35-dependence in * ii * , see ref .
@xcite . here
@xmath36 corresponds to the ballistic regime and @xmath37 captures the slower dephasing . in the context of optical lattices ,
local excitations seem difficult to implement .
hence , in the following , we focus on the observation of indirect effects on the dynamical evolution in mbl systems . in the following , we demonstrate that local memory of initial conditions , slow spreading of correlations and equilibration of local densities provide clear measures to distinguish mbl systems from both the non - interacting anderson insulators and also from ergodic systems , i.e. , those where local measurements , after a short relaxation time , can be captured by thermal ensembles . in order to carry out our analysis
, we will complement the intuitive guideline provided by the phenomenological @xmath19-bit model by a numerical tensor network tebd simulation @xcite ( for details see appendix [ app : numerics ] ) .
the chosen parameters for the simulation are a disorder strength of @xmath38 and interaction strengths of @xmath5 or @xmath9 for the mbl and anderson case , respectively .
we begin by considering the influence of the suppression of particle propagation .
a defining feature of localised systems is that independent of the interaction strength , particles and energies do not spread over the entire system , but remain confined to local regions .
they merely redistribute inside the localisation length , which can be extracted from the constants of motion .
therefore , even for long times , the particle density profile of an mbl system will not move to its thermal form , but rather retain some memory of its initial configuration .
this gives rise to the following particle localisation measure .
we define the following measure @xmath39 , which probes particle propagation for a system of length @xmath40 @xmath41 on an intuitive level , this measure directly probes the spread of particles , including weights based on the distance to the initial position @xmath42 such that distant contributions are amplified . numerically , we find that @xmath43 initially shows a steep linear increase , indicative of the ergodic dynamics governed by the onsite term of eq .
( fig .
[ mainfig ] ) . in the second regime
it fluctuates without visible growth , indicating a break - down of particle transport on length scales beyond the localisation length .
thus , the length scale of the phase correlations established in the system can be bounded independent of time @xmath44 . for ergodic systems , where particles and energies spread ballistically
, the measure would grow in an unconstrained fashion over time .
based on this insight , we deduce that time - of - flight images , while clearly distinguishing between localised and ergodic phase , are not useful for the distinction between interacting and non - interacting localised systems .
again more formally , this measure can be understood by considering the time evolution of the correlation matrix given by the matrix elements @xmath45 where @xmath46 . for the non - interacting case of an anderson insulator ,
this evolution is unitary @xmath47 , where @xmath48 is the evolution of the fermionic mode operators . for an anderson insulator ,
dynamical localisation precisely corresponds to locality of the unitary evolution @xcite , meaning that the matrix elements of @xmath49 are expected to decay exponentially @xmath50 for some constant @xmath51 with high probability @xcite . in the case of interacting hamiltonians that conserve the particle number
, this time evolution can be captured in form of a quantum channel @xmath52 where the kraus operators @xmath53 depend on the full initial state . as particle propagation in an mbl system
is expected to also be localised , it is assumed that the individual kraus operators obey @xmath54 .
starting from an initial product state of the form in eq .
, we obtain @xmath55 this again results in a suppression with the distance between @xmath56 and @xmath57 , causing a saturation of the phase correlation measure @xmath58 independent of time .
while particles and energies remain confined in interacting localised systems , correlations are expected to show an unbounded increase over time @xcite , although slower than in the ergodic counterpart . in stark contrast , anderson localised many - body systems will not build up any correlations that go beyond the localisation length . in order to probe the spreading of correlations in the system , we focus on a quantity easily accessible in the context of optical lattices , using in - situ images for different evolution times . as it turns out ,
this kind of simple density - density correlator is already sufficient to separate anderson localisation from mbl systems . in order to examine the spatial spreading of density - density correlations , we define the quantity @xmath59 , @xmath60 @xmath61 is a direct indicator for the length scale over which
density - density correlations are established without having to resort to assuming an explicit form , such as a decay in terms of an exponential function .
comparable to the dynamics of the phase correlations , we numerically find a steep initial increase followed by a saturation for the non - interacting case ( fig .
[ mainfig ] ) .
the mbl system however , continues to build up density - density correlations for the times simulated .
there is a transition in propagation speed , which we ascribe to the two dynamical regimes discussed before .
hence we conclude , that density - density correlations can be used to discriminate mbl from its non - interacting counterpart .
an intuitive explanation for the spread of density - density correlations despite spatial localisation of particles is that after exploring the localisation length , the particles feel the presence of neighbouring particles .
mediated by this interaction , the local movement of particles , governed by the respective constant of motion , becomes correlated , even over large distances .
in contrast , in the anderson insulator where constants of motion are completely decoupled , this communication can not take place .
we can connect this intuitive explanation to the more rigorous setting of lieb - robinson - bounds . in the anderson insulator in one - dimension , there provably exists a zero - velocity lieb - robinson - bound , where the correlator on the left hand side of eq .
is bounded by a time independent factor @xmath62 .
this means that the detectability of an excitation created in region @xmath29 decreases exponentially with the distance to @xmath31 . on the contrary ,
in the mbl regime we expect a logarithmic lieb - robinson cone of the form of eq .
* ii*. hence , an unbounded growth of correlations between distant regions is in principle possible , given sufficient time .
furthermore , we have shown that this built - up of correlations does also happen on observable time scales as can be seen from the evolution of density - density correlations captured by * measure 2*. it is also instructive to study the differences between the anderson and mbl - regime with respect to their equilibration properties .
due to the interactions present , we expect equilibration of fluctuations to take place in mbl systems , whereas in anderson insulators the effective subspaces explored by single particles remain small for all times and hence fluctuations remain large
. this in turn implies that fluctuations of local expectation values die out in the interacting model , but persist in an anderson insulator .
this qualitative difference has already been identified as a signifier of interactions in a disordered system@xcite . here , we build upon this idea and propose to consider the average change rate of local expectation values in order to detect the decreasing fluctuations in the mbl phase .
[ def : derivative ] we consider the expectation value @xmath63 of a local density operator in the middle of the system . as a measure of local equilibration
, we introduce the averaged rate of change of this density as a function of time @xmath64 @xmath65 as laid out in fig .
[ mainfig ] again , this function over time indeed shows a remarkably smooth behaviour that allows for the clear distinction between an anderson localised system and its mbl counterpart in that after a mutual increase the anderson system saturates at a constant value , whereas in the mbl phase , @xmath66 shrinks successively .
if we again resort to the lieb - robinson bound picture , we find that in the anderson case a local excitation is confined to a distinct spatial region given by the zero - velocity lieb - robinson - bound introduced in the previous chapter .
this implies that the effective subspace explored is constant and specifically , the excitation can not build up long distance correlations and fluctuations remain large .
this can also be seen from the results of * measure 2*. if we now , however , turn to the interacting model , a local excitation will slowly explore larger and larger parts of the hilbert space , leading to a slow , but persistent decrease of the fluctuations . in this work
, we proposed an operational procedure for distinguishing mbl phases building upon realistic measurements , which can be performed in the realm of optical lattices with present technology . utilising a phenomenological model and the concept of lieb - robinson - bounds
, we explained the effects numerically investigated employing tensor network methods .
the equilibration of local observables allows for the distinction of anderson and mbl localised models .
density - density correlations allow for the same information bit extraction , while furthermore reproduce the expected phenomenology . further investigating
this quantity might yield information about the localisation length via the duration of the first evolution regime .
phase correlations , which are directly connected to tof imaging , can not detect interactions in a localised system due to their correspondence to particle transport .
there is yet other information the tof reveals : one can also lower bound the spatial entanglement of bosons in optical lattices @xcite , building upon the ideas of constructing quantitative entanglement witnesses @xcite , a notion of multi - partite entanglement @xmath67 detecting a deviation from a best separable approximation , as @xmath68 for all @xmath69 .
this quantity detects a reasonable notion of multi - particle entanglement , which is yet different from the bi - partite entanglement discussed above .
since this measure is only onsite local , we would expect that it can not distinguish the long - range correlations of an interacting disordered model from the dynamics inside the constant of motion .
this further motivates the quest to engineer appropriate entanglement witnesses both accessible in optical lattice architectures as well as probing key features of mbl , a quest that is in turn expected to contribute to our understanding of mbl as such .
we would like to thank the eu ( aqus , siqs , raquel ) , the erc ( taq ) , the bmbr ( q.com ) , and the dfg ( spp 1798 ) for support , and u. schneider , c. gross , a. scardicchio , and r. vasseur for discussions .
44ifxundefined [ 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 * * , ( ) @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 `` , '' link:\doibase 10.1103/physrevx.5.041047 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1038/nphys2232 [ * * , ( ) ] @noop `` , '' ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1007/bf01645779 [ * * , ( ) ] link:\doibase 10.1088/1742 - 5468/2004/04/p04005 [ * * , ( ) ] @noop `` , '' ( ) , @noop * * , ( ) @noop * * , ( ) @noop `` , '' ( ) |
single crystals of mnwo@xmath1 and cowo@xmath1 were grown from the melt . during the single crystal growth process of mnwo@xmath1
special care was taken to keep manganese in divalent state .
this was achieved by the use of high growth temperatures and the renouncement of melt solvents . using the top
seeded growth technique and low cooling rate ( 0.08 k / h ) , ruby - red transparent single crystals were obtained starting from a growth temperature of 1574 k.@xcite single crystals of cowo@xmath1 were obtained from na@xmath4w@xmath4o@xmath9 melt solution with a small surplus of wo@xmath3 ( molar ratio cowo@xmath1 : na@xmath4w@xmath4o@xmath9 : wo@xmath3 = 1 : 2 : 0.5 ) , starting at 1363 k and applying a cooling rate of 3 k / h .
the xas spectra were collected at the dragon beamline of the national synchrotron radiation research center in hsinchu , taiwan .
the samples were cleaved _ in - situ _ in an ultra - high vacuum chamber with pressures in the @xmath10 mbar range , guaranteeing the high surface quality required to perform bulk - representative xas studies in the total electron yield mode .
the degree of linear polarization was @xmath11% , with an energy resolution of approximately 0.3 ev .
the spectra were collected at room temperature in the paramagnetic phase of both compounds .
single crystals of mno and coo were measured simultaneously in a separate chamber as references for mnwo@xmath1 and cowo@xmath1 , respectively .
xas spectra of mnwo@xmath1 with the @xmath12 vector of the light parallel to the @xmath13 , @xmath14 , and @xmath15 crystallographic axes .
the spectrum of mno is included as reference .
bottom panels : a close - up revealing the polarization dependence of the spectra.,width=347 ] fig .
[ mnl ] shows the room temperature mn-@xmath0 xas spectra of mnwo@xmath1 taken with the @xmath12 vector of the light parallel to the @xmath13 , @xmath14 , and @xmath15 crystallographic axes .
the spectrum of a mno single crystal is also included for reference purposes .
the spectra are dominated by the mn @xmath16 core - hole spin - orbit coupling which splits the spectrum roughly in two parts , namely the @xmath17 ( @xmath18 638 - 645 ev ) and @xmath19 ( @xmath18 649 - 654 ev ) white line regions . the line shape strongly depends on the multiplet structure given by the mn 3@xmath20 - 3@xmath20 and 2@xmath21 - 3@xmath20 coulomb and exchange interactions , as well as by the local crystal fields and the hybridization with the o 2@xmath21 ligands .
unique to soft xas is that the dipole selection rules are very sensitive in determining which of the 2@xmath223@xmath23 final states can be reached and with what intensity , starting from a particular 2@xmath243@xmath25 initial state ( @xmath26 for mn@xmath2 ) @xcite .
this makes the technique extremely sensitive to the symmetry of the initial state , i.e. , the valence and the crystal field states of the ions . comparing the mnwo@xmath1 spectra with the mno
, one can immediately observe that the spectral features are similar with the @xmath27 main peaks at identical energies .
this indicates directly that the mn ions in mnwo@xmath1 are also in the high - spin mn@xmath2 electronic configuration as the mn ions in mno . yet , the clear differences between the spectra of mnwo@xmath1 and mno , as seen e.g. at the low and high energy shoulders of the @xmath27 white line , provide a hint to a crystal field level scheme different from @xmath28 for the mnwo@xmath1 system .
a small but clear polarization dependence for mnwo@xmath1 was found , as can be seen in the bottom panels of fig .
[ mnl ] , where a close - up is shown of the spectra taken with the three different polarizations .
the intensities as well as the energy positions of several peaks vary with the polarization of the incoming light .
this polarization dependence is yet another indication that the local symmetry of the mn ions in mnwo@xmath1 is lower than @xmath28 .
xas spectra of cowo@xmath1 .
the spectrum of multi - domain coo single crystal is included as reference.,width=302 ] fig .
[ col ] depicts the co-@xmath0 xas spectra of cowo@xmath1 taken with the @xmath12 vector of the light parallel to the @xmath13 , @xmath14 , and @xmath15 crystallographic axes , together with the spectrum of a multi - domain coo single crystal as reference . following the same argumentation
as given above for the mn @xmath0 edge , from the strong similarities between the cowo@xmath1 and the coo spectra we can conclude that cowo@xmath1 contains co@xmath2 ions with a @xmath29 high - spin configuration like in coo . in comparing the cowo@xmath1 with the mnwo@xmath1
, one observes that the polarization dependence is much larger . in the following
we will discuss the different origins of the linear polarization dependence for the two compounds and its implications for their very different magnetic properties .
to interpret and understand the spectral line shapes and their polarization dependence , we have performed simulations of the atomic - like @xmath30 ( @xmath26 for mn@xmath2 and @xmath31 for co@xmath2 ) transitions using the well - proven configuration - interaction cluster model.@xcite within this method we have treated the mn or co ion within an mno@xmath32 or coo@xmath32 cluster , respectively , which includes the full atomic multiplet theory and the local effects of the solid .
it accounts for the intra - atomic @xmath33@xmath33 and @xmath16@xmath33 coulomb interactions , the atomic @xmath16 and @xmath33 spin - orbit couplings , the local crystal field , and the o @xmath16mn @xmath33 or o @xmath16co @xmath33 hybridization .
this hybridization is taken into account by adding the @xmath34 and @xmath35 etc .
states to the starting @xmath36 configuration , where @xmath37 denotes a hole in the o @xmath21 ligands . and cowo@xmath1 .
for the calculations of the xas spectra , a local coordinate system ( @xmath38 , @xmath39 , @xmath40 ) was introduced , in which the axes point roughly along the mn / co - o bonds .
the global crystallographic coordinates system can be transformed into the local coordinate system by rotating along @xmath41 by @xmath42 and then along the @xmath43 axis by @xmath44 ( see text for more details).,width=302 ] to facilitate the setup and the interpretation of the calculations , a local coordinate system was introduced , with its axes ( @xmath38 , @xmath39 , @xmath40 ) pointing roughly along the mn - o and co - o bonds . to transform the global coordinates ( @xmath45 , @xmath41 , @xmath43 ) into the local system , first a rotation by 33.9@xmath46 along @xmath41 is applied , resulting in a new system with ( @xmath47 , @xmath48 , @xmath49 ) .
this system is then rotated by 45@xmath46 along the _ old _ axis @xmath43 , transforming ( @xmath47 , @xmath48 , @xmath49 ) into ( @xmath38 , @xmath39 , @xmath40 ) .
the local coordinate system is shown in fig .
[ structure ] .
the coordinates of the oxygen ions as well as the mn - o and co - o distances are summarized in table [ ltab ] .
the crystal structure has been taken from ref . .
there are two distorted octahedra in the unit cell and they are connected to each other by inversion , thereby contributing to the polarization dependence in the same manner .
.coordination of the mn@xmath2 and co@xmath2 ions in their distorted oxygen octahedra .
all oxygen positions are given in in the local coordinate system specified in fig .
[ structure ] , where the numbering of the atoms is also defined .
the hybridization strengths @xmath50 and @xmath51 are in units of ev . [ cols="^,^,^,^,^,^,^",options="header " , ] the mn - o and co - o bond lengths were used to estimate the hybridization strength using harrison s description,@xcite resulting in values for @xmath50 and @xmath51 as shown in table [ ltab ] .
the slater - koster formalism@xcite provides the angular dependence of the hybridization strength .
values for the crystal fields were tuned to find the best match to the experimental spectra .
parameters for the multipole part of the coulomb interactions were given by the hartree - fock values,@xcite while the monopole parts ( @xmath52 , @xmath53 ) were estimated from photoemission experiments on mno@xcite and previous work on coo.@xcite the simulations were carried out using the program xtls 8.3,@xcite and the parameters used are listed in refs . . the calculated polarization - dependent spectra for the mn @xmath0edge of mnwo@xmath1
are plotted in fig .
one can observe that the general line shape of the experimental spectra is very well reproduced .
equally important , also the small energy shifts in the peak positions as well as the small variations in the peak intensities as a function of polarization can all be simulated .
this indicates that we have been able to capture the local electronic structure of the mn ion in mnwo@xmath1 with great accuracy .
we now will look into this in more detail .
( @xmath54 ) and ( c , d ) co@xmath2 clusters , excluding and including the @xmath33 spin - orbit coupling @xmath55 .
only the states up to 1 ev are shown.,width=340 ] the ground state of a mn@xmath2 ion in @xmath28 symmetry is @xmath56 , and this will not split or change upon lowering the crystal symmetry .
the electronic charge distribution of this high - spin half - filled shell ion remains spherical .
the presence of lower symmetry crystal fields will then show up most clearly when an electron is added ( or removed ) from the valence shell . to illustrate such a situation , we depict in the left panel of fig .
[ elevel ] the total energy level diagram of a mn@xmath57 ( @xmath54 ) ion using the crystal and ligand field parameters which reproduce best the mn @xmath0 spectra of mnwo@xmath1 as described above .
one can readily observe a splitting of about 0.7 ev between states with the added electron in the @xmath58 vs. the @xmath59 orbitals .
upon switching off the @xmath33 spin - orbit interaction , see fig .
[ elevel ] ( a ) , one can also identify an intra-@xmath58splitting of about 65 and 95 mev as well as an intra-@xmath59splitting of 50 mev .
all of the orbitals are of mixed type as a result of the low local symmetry .
the two orbitals lowest in energy are approximately @xmath60 and @xmath61 , resembling the @xmath62 and @xmath63 orbitals , rotated by @xmath6445@xmath46 along @xmath40 .
this is a result of the local @xmath65 symmetry axis of the mno@xmath32 cluster , which lies in the @xmath66 plane with an angle of roughly 45@xmath46 to the oxygen bonds . also in the mn
@xmath0 xas process one extra electron is added into the valence shell .
the lowest lying peak at 639.6 ev , see fig .
[ mnl ] , involves the excitation of the @xmath16 core electron into the @xmath58 orbitals .
different polarizations access different orbitals with different probabilities . for the 640.6 ev peak ,
the polarization dependence is reflected in terms of shifts in its energy position .
if one could have measured the polarization dependence along the local coordinates of one mn ion , one may expect to observe peak shifts of 65 and 95 mev associated with the intra @xmath58 splittings shown in fig .
[ elevel ] . yet , having two mn ions in the unit cell , and measuring along the global @xmath45 , @xmath41 , and @xmath43 crystallographic directions , the shifts between the effectively composite peaks become reduced . in the experiments and in the simulations
we are left with shifts of about 20 and 30 mev as can be seen in the middle bottom panel of fig .
[ mnl ] . for the 639.6 ev peak ,
the shifts due to variation of the polarization are more difficult to quantify since there is also a variation in the intensity of the peaks .
the latter is apparently caused by the rather complicated multiplet effects involving the @xmath16 core hole , as demonstrated by the excellence of the match by the simulations , see the left bottom panel of fig .
these intensity variations are yet quite small , and _ integrated _ over the entire @xmath0 range , would have been even identical to zero if the non - cubic hybridization effects were absent . the simulations for the polarization dependent spectra for the co @xmath0 edge of cowo@xmath1 are shown in fig .
[ col ] . like in the mn case
, we have been able to achieve a satisfying fit to the experimental spectra : all features are well reproduced , including the strong polarization dependence .
we will now analyze the local electronic structure of the co ion on the basis of the parameters used in these simulations . for the co@xmath2 ions , the ground state of the @xmath29 configuration in @xmath28 symmetry is @xmath67 ( @xmath68 ) . unlike for the @xmath56 of the mn
, this @xmath67 state will be split upon going to lower crystal symmetry .
the degeneracy in the orbital part , i.e. in the @xmath58 sub - shell , will be lifted .
this can be seen from the total energy level diagram shown in the right panel of fig .
[ elevel ] .
the calculation with the @xmath33 spin - orbit interaction switched off , see fig .
[ elevel ] ( c ) , reveals the presence of three low - lying quartets within the first 0.1 ev . in an one - electron language
, each of them would correspond to the hole occupying one of the three @xmath58orbitals .
the intra-@xmath58 splittings are 40 mev and 80 mev .
these values are quite close to those of the mnwo@xmath1 case reflecting the similar crystal structure .
the importance of the lifted degeneracy is that the orbital occupation will no longer be isotropic , i.e. the co ion charge distribution will be highly non - spherical .
this then explains the strong polarization dependence in the absorption spectra .
it is also important to notice how the low - lying states of co@xmath2are influenced by the @xmath33 spin - orbit coupling .
the energy level diagram in fig .
[ elevel ] ( d ) reveals that there is indeed a large amount of mixing due to spin - orbit interaction .
the spin - orbit coupling constant of @xmath69 mev is of the same order of magnitude as the splittings in the @xmath58 subshell .
all this has direct consequences on the magnetism , as will be discussed in the next section . in cowo@xmath1 .
( a ) three - dimensional representation of the magnetic anisotropy in the unit cell .
the surface depicts the expectation value of the magnetic moment @xmath70 pointing parallel to a small exchange field applied to the ion .
( b ) two - dimensional cut of @xmath70 in the plane of easy and intermediate magnetic axis ( corresponding approximately to the @xmath66 plane in local coordinates ) .
( c ) cut through the plane of hard and intermediate magnetic axis.,width=226 ]
from the results of the simulation , we find that the @xmath33 spin - orbit interaction has a major influence on the energies and nature of the low lying states of the cowo@xmath1 system .
the magnetic moment will then not only have a spin contribution but also an appreciable orbital one .
this results in a strong coupling of the direction of magnetic moment to the crystal structure .
we can make an estimate of the single ion anisotropy of the co@xmath2 ion . for that we use the same parameters with which we have reproduced the xas spectra and the polarization dependence therein , and
we apply a small exchange field of @xmath71 ev ( @xmath72 oe ) to align the magnetic moment . here
we would like to note that the results hardly changes with increasing the strength of the field , i.e. they are robust even beyond @xmath73 ev . fig .
[ dir ] shows the expectation value of the magnetization parallel to the applied exchange field @xmath70 as a three - dimensional representation and in a polar plot , the latter in local coordinates .
it can be seen that the single - ion anisotropy is of easy - axis type , with the axis lying along @xmath74 135@xmath46 at @xmath75 in the local coordinate system . in the direction of the easy axis , the moment is completely directed along the magnetic field , and the perpendicular component of the magnetization vanishes .
the easy axis is also perpendicular to the @xmath65 symmetry axis of the distorted octahedron , reflecting the orbital character of the two lower lying ligand field levels .
the total magnetic moments of the hard and easy axes are 1.2 @xmath76 and 3.0 @xmath76 , respectively , with an orbital contribution of 0.1 @xmath76 and 0.9 @xmath76 .
the easy axis of the total moment coincides directly with that of the orbital moment , which is also pointed out earlier by bruno.@xcite it has also been reported that for low symmetries , the orbital and total moment could differ in their behavior due to the magnetic dipole moment @xmath77.@xcite in the case of cowo@xmath1 , however , with the symmetry of the coo@xmath32 groups being not too far from cubic , this @xmath77 is around @xmath78 ev , and does not influence the anisotropy significantly . in cowo@xmath1 ,
the single - ion anisotropy is strong enough to lead to an effective pinning of the spins in the magnetically ordered phase .
isotropic superexchange will be the dominating magnetic interaction , although in addition anisotropic exchange is possible , as the bonds lack inversion centers .
but the dzyaloshinskii - moriya ( dm ) term @xmath79 , that favors canted magnetic moments , is small compared to the energy involved in the magnetic anisotropy of cowo@xmath1 , reflected by the large difference in moment of more than 1 @xmath76 .
the superexchange will then minimize the energy in the magnetically ordered phase by aligning the moments along the easy axis .
the easy axis is identical for both coo@xmath32 clusters in the unit cell , and thus the single - ion anisotropy itself will not lead to canting between the magnetic moments of the co@xmath2 ions .
the result is a collinear antiferromagnet with an easy axis approximately in the @xmath80 plane @xmath81 off from @xmath45 .
this direction is illustrated in fig .
[ dir ] and is close to the one found in the experiment by neutron diffraction .
@xcite returning to mnwo@xmath1 case : to the leading order , the local spherical @xmath56 ground state does not produce an orbital moment .
the spin is then free to point in any direction .
this is reproduced in the calculation , where the response of the magnetic moment to an exchange field is practically isotropic . yet ,
in second order , the spin - orbit coupling does affect the mn - o hopping .
this produces a non - vanishing moment perpendicular to the exchange field , albeit in the order of 10@xmath82 @xmath76 , being much lower than in the case of the spin - orbit active cowo@xmath1 .
consequently , the dm term is on the same energy scale as the anisotropy , and spin canting can easily occur .
the propagation direction of the spiral magnetic order is in the _ local _ @xmath66 plane , meaning that the second order process leads to a hard axis along @xmath40 .
this demonstrates the importance of the single - ion anisotropy in this system .
the form of the anisotropy , however , depends strongly on the magnetic exchange .
this explains that even three different magnetically ordered phases exist in the material at low temperatures.@xcite additionally to the hard axis , the easy plane also has two principal axes for the magnetic anisotropy with a small difference between the two .
this , in combination with the subtle dependency of the anisotropy on the exchange field , makes incommensurabilities likely to appear .
the single - ion anisotropies extracted from our soft x - ray absorption data and the corresponding full - multiplet calculations provide a natural explanation for the magnetic ordering phenomena occurring in cowo@xmath1 and mnwo@xmath1 . in cowo@xmath1 ,
the spin - orbit coupling and a non - vanishing orbital moment overwhelm the anisotropic dzyaloshinskii - moriya exchange and cause the formation of a collinear ordering of the moments .
contrary to that , although being isostructural , mnwo@xmath1 shows only a small magnetic anisotropy which is comparable in energy to the dzyaloshinskii - moriya interactions .
this will ca nt the spins in such a way that the spiral propagates perpendicular to the direction of smallest magnetic moment , allowing a modulated spin spiral which causes the ferroelectricity in the material .
we gratefully acknowledge the nsrrc staff for providing us with beamtime . the research in cologne is supported by the deutsche forschungsgemeinschaft through sfb 608 .
n. h. is also supported by the bonn - cologne graduate school of physics and astronomy .
we are grateful for discussions with m. w. haverkort and d. i. khomskii . | soft x - ray absorption experiments on the transition metal @xmath0 edge of multiferroic mnwo@xmath1 and antiferromagnetic cowo@xmath1 are presented .
the observed linear polarization dependence , analyzed by full - multiplet calculations , is used to determine the ground state wave function of the magnetic mn@xmath2 and co@xmath2 ions .
the impact of the local structure and the spin - orbit coupling on the orbital moment is discussed in terms of the single - ion anisotropy .
it is shown that the orbital moment in cowo@xmath1 is responsible for the collinear antiferromagnetism , while the small size of spin - orbit coupling effects make spiral magnetic order in mnwo@xmath1 possible , enabling the material to be multiferroic .
the combination of magnetic order and ferroelectricity in multiferroic compounds is a topic that has gained high interest in modern research.@xcite magnetic transitions directly linked to a spontaneous electrical polarization have been investigated in detail for materials like _ _ r__mno@xmath3 ( _ _ r__=y , tb , gd , dy ) @xcite and tbmn@xmath4o@xmath5 @xcite . among the multiferroic manganates ,
mnwo@xmath1 takes a special place as it does not contain a rare - earth ion , leaving mn formally as the only magnetic ion.@xcite multiferrocity in these materials is explained by the spin current model,@xcite where the electrical polarization is caused by a spiral order of the magnetic moments .
the spin rotation axis @xmath6 does not coincide with the magnetic propagation vector @xmath7 , leading to a non - vanishing spontaneous polarization @xmath8 . here
we investigate the conditions for the occurrence of spiral magnetic order responsible for the rich physics of mnwo@xmath1 .
we performed soft x - ray absorption spectroscopy ( xas ) on mnwo@xmath1 to determine the local electronic and magnetic properties of the mn ions .
the results are also compared to measurements on the isostructural compound cowo@xmath1 , which is a collinear antiferromagnet and not a multiferroic material .
we analyze the role of the spin - orbit coupling and its impact on the magnetic structure .
the magnitude of the single - ion anisotropy is calculated using the full - multiplet configuration - interaction approach .
this forms the key to understand why in mnwo@xmath1 spiral magnetic order is possible and why it is suppressed in cowo@xmath1 . |
this work was partially supported by dgi ( spain ) : bfm2002 - 03315 , fpa2002 - 04181-c04 - 04 and bfm2000 - 0600 and by the junta de andaluca . c.m .
acknowledges mec ( spain ) for a postdoctoral stay at university of sevilla ( sb2000 - 0427 ) .
m.c.m . acknowledges support from a fellowship from the fundacin cmara ( university of sevilla ) .
99 y. fukuda _ et al .
_ [ super - kamiokande collaboration ] , phys . rev . lett . * 81 * , 1562 ( 1998 ) ; c. k. jung , c. mcgrew , t. kajita and t. mann , ann .
nucl . part .
sci . * 51 * , 451 ( 2001 ) ; see for example a. de santo , int .
j. mod .
a * 16 * , 4085 ( 2001 ) ; `` neutrino nucleus interactions in the few gev region '' .
proceedings , 1st international workshop , nuint01 , tsukuba , japan , december 13 - 16 , 2001 .
t. k. gaisser and j. s. oconnell , phys .
d * 34 * , 822 ( 1986 ) ; h. c. kim , j. piekarewicz and c. j. horowitz , phys .
c * 51 * , 2739 ( 1995 ) ; h. c. kim , s. schramm and c. j. horowitz , phys .
c * 53 * , 2468 ( 1996 ) , _ ibid . _ * 53 * , 3131 ( 1996 ) ; s. k. singh and e. oset , phys . rev .
c * 48 * , 1246 ( 1993 ) ; j. engel , e. kolbe , k. langanke and p. vogel , phys .
d * 48 * , 3048 ( 1993 ) ; j. marteau , j. delorme and m. ericson , nucl .
instrum .
a * 451 * , 76 ( 2000 ) . c. bleve _ et al . _ ,
astropart .
* 16 * , 145 ( 2001 ) ; g. c , c. bleve , i. de mitri and d. martello , nucl .
suppl . * 112 * , 210 ( 2002 ) .
udas _ et al .
_ , phys .
c * 48 * , 2731 ( 1993 ) ; * 51 * , 3246 ( 1995 ) ; * 53 * , r1488 ( 1996 ) ; * 64 * , 024614 - 1 ( 2001 )
. w. m. alberico _ et al .
_ , nucl .
a * 623 * , 471 ( 1997 ) .
v. bernard , l. elouadrhiri and u. g. meissner , j. phys .
g * 28 * , r1 ( 2002 ) .
c.j . horowitz and b.d .
serot , nucl .
a * 368 * , 503 ( 1981 ) ; phys .
b * 86 * , 146 ( 1979 ) ; b.d .
serot and j.d .
walecka , adv .
* 16 * , 1 ( 1986 ) .
cooper , s. hama , b.c .
clark and r.l .
mercer , phys .
c * 47 * , 297 ( 1993 ) .
p. barreau et al . , nucl .
phys . * a 402 * , 515 ( 1983 ) .
m. albert _ et al .
_ [ lsnd collaboration ] , phys . rev .
c * 51 * , 1065 ( 1995 ) . c. athanassopoulos
_ et al . _
[ lsnd collaboration ] , phys .
c * 56 * , 2806 ( 1997 ) ; l. b. auerbach _ et al .
_ [ lsnd collaboration ] , phys . rev .
c * 66 * , 015501 ( 2002 ) .
see n. jachowicz , k. heyde , j. ryckebusch and s. rombouts , phys .
c * 65 * , 025501 ( 2002 ) ; c. volpe _ et al .
_ , phys .
c * 62 * , 015501 ( 2000 ) 015501 ; e. kolbe _ et al .
c * 52 * , 3437 ( 1995 ) ; e. kolbe _ et al .
a * 652 * , 91 ( 1999 ) ; y. umino , j. m. udas and p. j. mulders , phys . rev
. lett . * 74 * , 4993 ( 1995 ) ; y. umino and j. m. udas , phys .
c * 52 * , 3399 ( 1995 ) ; c. giusti and f. d. pacati , nucl . phys
a * 473 * , 717 ( 1987 ) and refs . therein | the quasielastic scattering of muon neutrinos on oxygen 16 is studied for neutrino energies between 200 mev and 1 gev using a relativistic shell model .
final state interactions are included within the distorted wave impulse approximation , by means of a relativistic optical potential , with and without imaginary part , and of a relativistic mean field potential . for comparison with experimental data the inclusive
charged current quasielastic cross section for @xmath0@xmath1 scattering in the kinematical conditions of the lsnd experiment at los alamos is also presented and briefly discussed . in the past few years the observation of neutrino oscillations at super
kamiokande @xcite and the subsequent proposal and realization of new experiments , aimed at determining neutrino properties with high accuracy @xcite , have renovated the interest towards neutrino scattering on complex nuclei .
in fact , neutrino detectors usually contain carbon or oxygen nuclei , and for a proper interpretation of the experimental results the description of the @xmath2nucleus interaction must be accurate @xcite . at intermediate neutrino energies , ranging from some hundreds mev to a few gev , @xmath2nucleus quasielastic scattering has been studied within several approaches @xcite .
relativistic and non relativistic studies of random phase approximation have shown nuclear structure effects to be relevant only at low momentum transfers , but indications have been found that for future and precise data analyses of , e.g. , atmospheric neutrino measurements , more accurate theoretical estimates may be needed . additionally , very recently new attention has been drawn towards final state interaction ( fsi ) effects , which , contrary to what is often assumed , may still be relevant even at the relatively high energy @xmath3 gev @xcite .
in this contribution we study charged current ( cc ) neutrino nucleus quasielastic scattering within the framework of a relativistic shell model ( rsm ) , already successfully employed to study exclusive electron scattering @xcite and neutral current neutrino scattering @xcite .
we compute inclusive @xmath0@xmath4 quasielastic cross sections for three values of the incident neutrino energy , namely @xmath5 mev , @xmath6 mev , and @xmath7 gev , which are representative of the kinematical range where quasielastic scattering gives the main contribution to the inclusive @xmath2nucleus process .
we describe the cc quasielastic scattering of neutrinos on a nuclear target within the impulse approximation ( ia ) , assuming that the incident neutrino exchanges one vector boson with only one nucleon , which is then emitted , while the remaining ( a-1 ) nucleons in the target are spectators . the nuclear current is assumed to be the sum of single nucleon currents , for which we employ the usual free nucleon expression ( see @xcite ) with the axial form factor parameterized as a dipole with cut off mass @xmath8 gev @xcite , and the states of the target and residual nuclei to be adequately described by an independent particle model wave function . to describe the bound nucleon states we use relativistic shell model wave functions , obtained as the self consistent ( hartree ) solutions of a dirac equation , derived , within a relativistic mean field approach , from a lagrangian containing @xmath9 , @xmath10 and @xmath11 mesons @xcite .
as the single - particle binding energies determine the threshold of the cross section for every shell , in the numerical calculations , we have used the experimental values corresponding to the binding energies of the different shells . for the outgoing nucleon
the simplest choice is to use plane wave spinors , i.e. , no interaction is considered between the ejected nucleon and the residual nucleus ( pwia ) . for a more realistic description ,
fsi effects should be taken into account . in our formalism
this is done by using distorted waves which are given as solutions of a dirac equation containing a phenomenological relativistic optical potential ( rop ) , consisting of a real part , which describes the rescattering of the ejected nucleon and of an imaginary part , that accounts for the absorption of it into unobserved channels . in this work
we use the rop corresponding to the edad-1 single - nucleon parameterization presented in ref .
@xcite .
the use of this phenomenological rop leads to an excellent agreement between theoretical calculations and data for exclusive @xmath12 observables @xcite ; however some caution should be taken in extending the conclusions drawn from the analysis of exclusive reactions to inclusive ones . in the latter , unless a selection of the single nucleon knockout contribution is experimentally feasible , all final channels are included and thus the imaginary term in the optical potential leads to an overestimation of fsi effects .
this is clearly illustrated in fig . 1 where we compare our theoretical results with the experimental cross section corresponding to inclusive quasielastic electron scattering on @xmath1 for a momentum transfer equal to 400 mev / c ( similar results are obtained for 300 and 500 mev / c ) . in our calculation
besides the complex rop , we also consider the potential obtained by setting the imaginary part of the rop to zero .
additionally one may also use distorted waves which are obtained as the solutions in the continuum of the same dirac equation used to describe the initial bound nucleon .
we refer to this approach , that should be adequate to describe fsi at moderate energy transfer , as relativistic mean field ( rmf ) . as shown in fig . 1 , the complex rop results clearly underestimate the data , while the reverse occurs for the pwia calculation . on the other hand ,
the rmf and purely real rop agree much better with experiment , particularly rmf for small transfer energy and real rop for higher @xmath10 .
we believe that the rmf and real rop results indicate a reasonable `` band '' where fsi effects should lie . versus the outgoing muon kinetic energy , for the quasielastic scattering of muon neutrinos on @xmath4 and for three choices of the incident neutrino energy : @xmath13 mev ( upper panel ) , @xmath6 mev ( middle ) and @xmath7 gev ( lower panel ) .
the solid curves correspond to the rsm with no final state interaction , while the remaining curves are calculated within the rfg , with @xmath14 mev and @xmath15 ( dot dashed ) and @xmath16 mev ( dotted).,scaledwidth=60.0% ] let us now present our results for neutrino scattering on @xmath4 .
to better illustrate our model , we start by neglecting fsi and comparing rsm pwia results with inclusive cross sections obtained within the relativistic fermi gas ( rfg ) @xcite .
this is is done in fig .
[ fig : o16_dsig_pwia ] , which shows the differential cross section @xmath17 as a function of the outgoing muon kinetic energy . with respect to the rfg curve calculated with no binding energy
we observe that the rsm cross section is reduced and shifted towards lower @xmath18 values , in a way which is similar to the effect of an average binding energy in the rfg .
in addition the rsm cross section has a different shape , due to the different momentum distributions of the single nucleon shells contributing to the process . since the various shells have different binding energies , the corresponding contributions to the cross section go to zero at different values of @xmath18 and this gives rise to the structure of @xmath19 observed at large @xmath18 .
fig .
[ fig : o16_dsig_pwia ] shows that nuclear model effects on the cross sections can be rather large at low neutrino energy , but become less relevant as @xmath20 increases , practically disappearing at @xmath21 gev . , but including fsi effects .
all curves are calculated within the rsm model in pwia ( solid ) , and within the rmf ( dashed ) , real rop ( dotted ) and complex rop ( dot dashed ) approaches .
, scaledwidth=60.0% ] on the other hand , the behavior of fsi effects is quite different , as illustrated in fig .
[ fig : o16_dsig_fsi ] . here the results obtained with the rsm in pwia
are compared with the cases where fsi are described within the rmf , the real rop and the complex rop approaches .
the use of real potentials ( rmf and real rop ) for describing the final nucleon states leads to the resonant structure observed for relatively high @xmath18 ( that is , small energy transfer @xmath10 ) .
note that in this work we only include single - particle excitations within a mean field picture . including residual interactions
would make the width and number of resonances to be considerably larger .
we observe that fsi effects produce a reduction of the cross section , particularly important in the case of the complex rop model due to the absorption introduced by the imaginary term : about 60% for @xmath13 mev and 50% for @xmath22 mev and @xmath3 gev in the region close to the maximum .
for the rmf and real rop , the reduction , similar in both cases , is about @xmath23 for @xmath24 mev and @xmath25 for the other energy values . for the quasielastic scattering of muon neutrinos on @xmath4 as a function of the incident neutrino energy .
the curves are calculated within the rfg model with @xmath26 mev and binding energy @xmath15 ( solid line ) and @xmath27 mev ( dashed ) .
the points correspond to rsm calculations without fsi ( stars ) and with fsi effects taken into account within the rmf ( empty squares ) , real rop ( full squares ) and complex rop ( circles ) approaches.,scaledwidth=50.0% ] nuclear model effects on integrated cross sections are studied in fig .
[ fig : o16_tot ] , where the cross section @xmath28 is plotted as a function of the incident neutrino energy . here
the contributions coming from the rmf and rop resonances have been included in the calculation , in order to respect the completeness of the set of final states predicted by the model .
these contributions are important at @xmath13 mev , where they amount to about 10% of the integrated cross section , while at higher energies these effects are about 2% ( 500 mev ) and 1% ( 1 gev ) .
again we see that within the pwia the discrepancy between different nuclear models is relatively small and decreases with increasing neutrino energy . on the contrary fsi effects
remain sizeable even at large @xmath20 . as in the previous figure ,
the imaginary term in the rop leads to a too large reduction ( @xmath29 ) of the integrated cross section .
the results for the rmf and real rop models , which are more reliable , show a smaller , but still sizeable ( @xmath30 at @xmath3 gev ) reduction .
the results shown in figs . [ fig : o16_dsig_fsi ] and [ fig : o16_tot ] lead us to conclude that fsi effects should be carefully considered in neutrino experiments which use oxygen based detectors .
analogous conclusions can be drawn for the case of a @xmath1 target .
for the purpose of comparison with the experiment , let us consider the inclusive @xmath31 cross section measured by the lsnd collaboration at los alamos , using a pion decay in flight @xmath0 beam , with energies ranging from muon threshold to 300 mev , and a large liquid scintillator detector @xcite .
this experiment has been compared to very different theoretical approaches @xcite , but an important discrepancy still remains .
although at these low energies processes different from the quasielastic nucleon knockout affect the inclusive cross section , we consider our fsi approach to be useful to estimate integrated cross sections . in fig .
[ fig : dsigexp ] we show the observed @xcite and calculated distribution of events , averaged over the 1994 los alamos neutrino spectrum @xmath32 within the energy range @xmath33 mev .
the shape and position of the maximum of the experimental distribution are approximately reproduced by the three calculations , but the results that include fsi with the rmf potential are clearly favored by the data .
this is consistent with the fact that including fsi with the mean field potential should be adequate at moderate kinetic energy of the ejected nucleon .
however the values we obtain for the the flux averaged integrated cross section overestimate the measured cross section by approximately 50% . more precisely , in the rsm , we obtain @xmath34 ( pwia ) , @xmath35 ( rmf ) and @xmath36 ( real rop ) @xmath37 @xmath38 .
additional corrections due to the outgoing muon coulomb distortion , evaluated within the effective momentum approximation @xcite , further increase these numbers by @xmath39 .
the corresponding final measured experimental value is @xmath40 @xmath37 @xmath38 @xcite . |
flavor changing neutral current ( fcnc ) @xmath5 decays do not occur at tree level in the standard model ( sm ) .
being loop induced , they feel scales of order @xmath6 and in principle much higher ones , making them important probes of the flavor sector of the sm and beyond .
rare radiative @xmath3 decays proceed via so - called electromagnetic penguins .
they have been measured in exclusive @xmath7 @xcite and inclusive @xmath8 @xcite decays . in dilepton channels
@xmath9 ( @xmath1 ) , we identify two additional structures in the feynman diagrams : boxes and @xmath4-penguins . none of the dilepton modes has been detected to date , but we expect large data samples from operating @xmath10-factories ( cleo , babar , belle ) , dedicated @xmath10-physics programmes at colliders ( tevatron run ii , hera - b ) and lhc - b in the long term . corresponding @xmath11 transition amplitudes are ckm suppressed @xmath12 .
the existing best bound in the dimuon channels for inclusive decays is @xmath13 at 90% c.l .
@xcite , which is one order of magnitude above the nlo sm expectation @xmath14 @xcite .
note that the nnlo calculation in @xmath9 is only partially available @xcite .
corresponding bounds for the exclusive channels are @xmath15 , @xmath16 at 90% c.l . @xcite and their respective sm predictions are @xmath17 , @xmath18 with the dominant theoretical uncertainty resulting from hadronic matrix elements , which are estimated here using light cone sum rules @xcite .
currently , the exclusive @xmath19 decay has the most interesting bound , which is only a factor of 2 away from the sm prediction . despite larger theoretical uncertainty than in the inclusive cases ,
rare exclusive decays are more accessible experimentally in the near future and have observables ( e.g. existence and position of the zero of the forward - backward asymmetry discussed below ) , which are as clean as the respective ones in the inclusive modes .
the calculational tool for the description of @xmath20 decays is the low energy effective hamiltonian @xmath21 @xcite .
this enables the analysis of relevant observables in a model independent way , with the goal being to extract the wilson coefficients @xmath22 from data @xcite .
the mayor player is the @xmath23 vertex @xmath24 .
its effective coupling strength @xmath25 is related to the branching ratio @xmath26 thus @xmath27 @xcite which is in good agreement with the sm value @xmath28 at @xmath29 at leading log .
we see that the @xmath3 data fix the modulus of @xmath25 , but not its sign ( phase )
. in @xmath9 decays , in addition to @xmath30 , also 4-fermi operators involving dileptons contribute , given by @xmath31 , @xmath32 . due to the charge assignments of lepton-@xmath4 couplings
the @xmath4-penguin contribution to @xmath33 is suppressed with respect to @xmath34 by @xmath35 .
we thus identify @xmath34 as a measure of the @xmath36 coupling modulo the box contribution @xcite .
decomposition of the @xmath37 branching ratio yields @xmath38 @xcite . using the cdf bound @xcite on this mode and allowing @xmath25 to have both sm - like and sm - opposite signs gives the present best bound on the strength of generic fcnc @xmath4-penguins of @xmath39 , which is a factor of 2 - 3 larger than the sm value @xmath40 @xcite .
scenarios with non - standard @xmath4-penguins arise in many extensions of the sm like such as supersymmetry , 4th generation and @xmath41 @xcite .
another interesting possibility to test the @xmath36 vertex arises in @xmath42 decays , since here no photon penguins contribute .
supersymmetric effects in inclusive @xmath9 decays have been studied in @xcite .
the reach of a new physics search in the dilepton invariant mass distribution in @xmath44 decays is exemplified in fig .
[ fig1 ] @xcite .
supergravity ( sugra ) ( dotted ) and a supersymmetric scenario with non - minimal sources of flavor violation in the mass insertion approximation ( dashed ) , can be well discriminated from the sm ( solid ) and its hadronic uncertainties ( shaded area ) ; the upper curves contain resonant @xmath45 background via @xmath46 , lower ones are pure short - distance contributions @xcite .
note that the dashed curve saturates the experimental bound in this channel .
similar findings are valid for @xmath47 decays , which however show less sensitivity to @xmath25 as the photon pole at @xmath48 is absent . in either case @xmath49 ( opposite - to - sm sign )
enhances the rates through constructive interference of @xmath25 with @xmath33 .
the forward - backward asymmetry in @xmath2 decays results from v / a interference in the lepton pair @xmath50 , shown in fig .
[ fig2 ] @xcite ; see @xcite for a discussion of the sign of @xmath51 , which is opposite to the forward - backward asymmetry @xmath52 of the cp conjugate channel in the cp conserving limit . in the sm ( @xmath53 , solid curve ) , @xmath51 has a zero around @xmath54 , which would disappear if @xmath25 would have the opposite sign ( long - short - dashed curve ) .
the existence of a zero in the forward - backward asymmetry in @xmath2 decays below the @xmath55 resonance is an important test of the sm and is independent of hadronic matrix elements @xcite .
positive @xmath25 occurs generically in supersymmetric theories @xcite , but only for large @xmath56 in relaxed and/or minimal sugra @xcite .
further , the position of the zero @xmath57 has very small hadronic uncertainties @xcite . in the limit where the final hadron has large energy , i.e. , small dilepton mass ,
the large energy effective theory @xcite is applicable and here all form factors cancel out in the ratios which determines @xmath57 @xcite .
a recently proposed observable , the forward - backward cp asymmetry @xmath58 in @xmath2 decays probes the phase of @xmath34 or of the @xmath36 vertex @xcite , respectively .
defined as @xmath59 , its magnitude scales with @xmath60 , where @xmath61 contains contributions from resonant and non - resonant @xmath62 states @xcite ; @xmath63 below threshold , so it is sizeable only in the high dilepton mass region .
integration over @xmath64 yields @xmath65 , which can be large in the case of @xmath66 phases .
hadronic uncertainties are not small , but the sm background is below @xmath67 and any effect above this would be due to a non - sm source of cp violation @xcite .
the sm branching ratio in @xmath8 decays is known at nlo with 10 % accuracy @xmath68 @xcite .
this is in good agreement with data from lep @xmath69 @xcite , and cleo , @xmath70 @xcite .
a promising observable in @xmath3 decays where dramatic signals of possible physics beyond the sm could show up is the cp asymmetry in the rate @xmath71 @xcite .
it is very tiny in the sm since @xmath72 , @xmath73 are wolfenstein parameters , thus @xmath74 @xcite
. however , large effects of ( 10 - 50 ) % are possible in scenarios with an enhanced chromo - magnetic dipole operator @xmath75 in the @xmath76 vertex .
cp asymmetry in exclusive @xmath7 decays has a less clean prediction due to strong phases , however , in the sm , @xmath77 holds @xcite .
any significant deviation from this would signal new physics . in both inclusive @xmath78 @xcite and exclusive cases @xmath79 @xcite ,
the measurements are not conclusive yet .
theoretically clear signatures of possible new physics can be experimentally isolated in @xmath3 and @xmath9 decays in the near future via the observables ( @xmath80 ) .
it is exciting to see whether the sm passes this next round of fcnc tests .
t.e.coan _ et al . _ , ( cleo collaboration ) , phys.rev.lett . 84 ( 2000 ) 5283 .
et al . _ , ( cleo collaboration ) , phys . rev . lett.74(1995)2885 ; + s. ahmed _ et.al . _ ( cleo collaboration ) , cleo conf 9910 ( hep ex/9908022 ) .
r. barate _
et al . _ , ( aleph collaboration ) phys
. lett . b429 ( 1998 ) 169 .
s. glenn _ et al .
_ , ( cleo collaboration ) , phys . rev . lett . 80 ( 1998 ) 2289 .
a. ali , l.t .
handoko , g. hiller and t. morozumi , phys .
d55 ( 1997 ) 4105 .
c.bobeth , m. misiak and j. urban , nucl .
b574 ( 2000 ) 291 . t. affolder _ et al .
_ , ( cdf collaboration ) , phys.rev.lett . 83 ( 1999 ) 3378 .
see , a. ali _ et al .
d61 ( 2000 ) 074024 , and references therein . a.j .
buras _ et al .
b424(1994)374 ; a.j .
buras and m. mnz , phys .
d52(1995)186 ; m. misiak , nucl .
b393(1993)23 ; err .
ibid . b439 ( 1995 ) 461 .
a. ali , g.f .
giudice and t. mannel , zeitschrift fr physik c67(1995)417 .
see , g. buchalla , g. hiller and g. isidori , hep - ph/0006136 , and references therein .
s. bertolini _ et al .
cho , m. misiak and d. wyler , phys .
d54(1996)3329;j.l . hewett and j.d .
wells , phys .
rev.d55(1997)5549 ; e. lunghi , a. masiero , i. scimemi and l. silvestrini , nucl .
b568 ( 2000 ) 120 . t. goto , y. okada and y. shimizu , phys .
a. ali , t. mannel and t. morozumi , phys .
b273(1991)505 ; f. krger and l.m .
sehgal , phys .
b380(1996)199 ; z. ligeti , i.w .
stewart and m.b .
wise , phys .
b420(1998)359 ; a. ali and g. hiller , phys .
d60 ( 1999 ) 034017 .
g. burdman , phys .
j. charles _ et al .
d60(1999)014001 . see , f. borzumati and c. greub , phys.rev .
d59 ( 1999 ) 057501 , and references therein .
a.l.kagan and m.neubert , phys.rev .
d58 ( 1998 ) 094012 . c. greub , h. simma and d. wyler , nucl .
b434 ( 1995 ) 39 . | we briefly review the status of rare radiative and semileptonic @xmath0 , ( @xmath1 ) decays .
we discuss possible signatures of new physics in these modes and emphazise the role of the exclusive channels .
in particular , measurements of the forward - backward asymmetry in @xmath2 decays and its zero provide a clean test of the standard model , complementary to studies in @xmath3 decays .
further , the forward - backward cp asymmetry in @xmath2 decays is sensitive to possible non - standard sources of cp violation mediated by flavor changing neutral current @xmath4-penguins .
slac - pub-8551 + august 2000 |
glasses are characterized by amorphous order and solid - like rigidies and have a wide range of applications in our day - to - day lives .
supercooled liquids and slightly polydispersed colloidal suspensions form glasses as the relevant control parameters , temperature ( @xmath4 ) and volume fraction ( @xmath5 ) respectively , are changed appropriately .
their dynamics slow down dramatically with decreasing @xmath4 or increasing @xmath6 and their viscosities ( @xmath7 ) or structural relaxation timescales ( @xmath8 ) grow by several orders of magnitude . below the glass transition temperature @xmath9 or above the glass transition volume fraction @xmath10 ( @xmath11 for hard - sphere colloids ) ,
@xmath12 becomes so large that no flow or structural relaxation can be detected in the system .
glass - forming liquids are called strong if @xmath13 ( or @xmath7 ) shows a nearly arrhenius growth as a function of the relevant control parameter , whereas they are termed fragile if a super - arrhenius trend is followed . in many theories and phenomenological models , the dramatic dynamical slowdown and the fragile behavior of glass - forming liquids
are understood by considering the cooperative movement of the constituent molecules or particles of the glass - forming liquid .
the distinct regions exhibiting correlated particle motion are called dynamical heterogeneities ( dhs ) .
the sizes of the dhs , or alternatively , the number of molecules or particles showing correlated movement ( @xmath14 ) , grow as the glass transition is approached .
the glass transition is manifested by a remarkable slowdown in the dynamics which results in the observed increase in @xmath15 .
using multi - point correlation function techniques , simulations and experiments have established the presence of dhs and the growth of @xmath16 in glass - forming liquids when their temperatures are quenched rapidly .
experimentally , such measurements in supercooled liquids are typically performed using non - linear dielectric susceptibility .
the present work characterises the dhs in a fragile colloidal glass .
colloidal clay suspensions are known to exhibit rich phase behavior , showing fluid , gel , ordered and disordered phases , and serve as excellent model systems to mimic the behaviour of atomic systems @xcite .
one ubiquitous example of a model colloidal system is laponite , a synthetic smectite clay which has been studied extensively for its soft glassy rheology @xcite and interesting aging properties as it spontaneously transforms from a liquid to a glass with increasing waiting time ( time since preparation ) , @xmath1 . in this article
, we characterize the growth of @xmath16 in synthetic laponite clay suspensions at several @xmath1 .
aqueous suspensions of laponite clay are known to form wigner repulsive glasses in the concentration range @xmath17 wt% due to the buildup of long range inter - particle electrostatic repulsive interactions with increasing @xmath1 @xcite .
the growth of its @xmath0 relaxation time in a super - arrhenius manner as @xmath2 increases has been established experimentally @xcite .
recent studies in our group have established that the slowdown in the dynamics of laponite suspensions resembles the slowdown reported in fragile glass - forming molecular liquids if @xmath2 of the former system is mapped to ( @xmath18 ) of the latter in the arrhenius and the vogel - fulcher - tammann equations , which represent , respectively , the secondary and primary relaxation modes . in an earlier study , s. jabbari - farouji _ et.al .
_ reported dhs in laponite suspensions by studying the particles rotational and translational diffusion coefficients as a function of @xmath3 .
it was observed that the rotational diffusion of the constituent particles slows down at a faster rate than their translational motion .
however , no report was found in the literature on estimates of the sizes of the dhs in aging laponite colloidal glasses . in this study
, we employ the three - point correlation function formalism to probe dhs and @xmath19 in laponite suspensions with changing @xmath2 .
the relaxation dynamics of laponite suspensions were studied by analysing the intensity autocorrelation functions obtained in dynamic light scattering experiments as a function of the age of the laponite suspension , @xmath1 .
the scattering function decay curves are fitted by parametrizing the two - step relaxation equation ( showing exponential and stretched exponential decays and representing , respectively , the secondary and primary relaxation processes of the laponite suspension ) with the control parameter @xmath2 .
the three - point correlation functions estimated using this analysis , which are the @xmath2 derivatives of the decaying scattering function curves , exhibit peaks whose heights increase with increasing @xmath1 .
@xmath16 is calculated from the peak height data and its evolution with @xmath20 , where @xmath21 is the non - ergodic transition time of the laponite suspension , bears a striking resemblance to exisiting calculations for the growth of correlations in rapidly quenched supercooled liquids .
we believe that our study , which demonstrates the existence of remarkable similarities between the kinetic arrest pheonemena in laponite clay suspensions and in supercooled liquids , provides valuable additional insight into the colloidal glass transition .
laponite rd@xmath22(byk , inc . )
powder was procured from southern clay products . as
clay particles are hygroscopic in nature , the powder was dried in a hot oven at 120@xmath23c for 16 hours
. 3.0 wt@xmath24 ( @xmath25 ) laponite concentration ( @xmath26 = 1.18@xmath2710@xmath28 ) suspension was prepared by adding the dried powder slowly to milli - q water ( resistivity 18.2 m@xmath29-cm ) .
@xmath30 was chosen to lie in a region of the phase space where the laponite suspension is expected to form a wigner glass .
the suspension was stirred vigorously for 1 hour using a magnetic stirrer .
the resulting optically clear and homogeneous suspension was filtered using a 0.45 @xmath31 m millipore millex - hv grade filter using a syringe pump at a constant flow rate of 3 ml / min . a very small volume fraction (
@xmath32 = 5.66@xmath2710@xmath33 ) of polystyrene ( ps ) probe particles ( 100 nm in size ) was added and mixed homogeneously .
these ps beads are expected to act as light scatterers in the otherwise highly transparent laponite suspensions .
the laponite - ps suspensions were subsequently sealed in a cuvette . the aging time or waiting time , @xmath1 , was calculated from the time when the stirring of the suspension was stopped and the cuvette was sealed .
auto - correlation functions of the intensity scattered by these suspensions were recorded in dynamic light scattering ( dls ) experiments @xcite using a brookhaven instruments corporation bi-200sm spectrometer and a bi-9000at digital autocorrelator .
a constant temperature of 25@xmath23c was maintained using a temperature controller ( polyscience digital ) attached to the dls system .
details of the set - up are given elsewhere @xcite .
the normalized intensity autocorrelation function of the scattered light , @xmath34 , was recorded as a function of delay time @xmath35 . here ,
@xmath36 and @xmath37 are the scattering wave vector and the intensity of the scattered light at a particular @xmath36 and @xmath35 respectively .
@xmath36 is related to the scattering angle @xmath38 by the equation @xmath39 , where @xmath40 and @xmath41 are the refractive index of the medium ( @xmath40 = 1.334 ) and the wavelength of the laser ( @xmath42 = 532 nm ) respectively .
the intensity autocorrelation data was recorded at @xmath38 = 90@xmath23 and @xmath38 = 60@xmath23 .
the three point correlation functions were computed by taking the derivatives of the time - dependent self - intermediate scattering function with respect to @xmath1 using mathematica .
decay curves vs. delay times recorded at various aging times ( @xmath1 ) at @xmath38=90@xmath23 .
curves are shifted for better visibility .
solid lines are fits to the two step relaxation function given in eq.3 and the fitting parameters obtained with increasing @xmath1 are given in the supporting information , table s1 .
( b ) mean structural relaxation time ( @xmath43 ) as a function of @xmath1 for a laponite - ps suspension .
the data plotted here is for a laponite suspension of concentration @xmath25 = 3 wt.% , and ps volume fraction @xmath32 = 5.66 @xmath27 10@xmath44 .
the solid line is a fit to eq.4.,width=316 ] the intensity auto correlation function @xmath45 in dls @xcite is related to the self - intermediate scattering function @xmath46 by the siegert relation , @xmath47 where @xmath48\}\right > \label{eq2}\ ] ] here , @xmath49 is the position of particle @xmath50 at time @xmath35 , @xmath51 is the spatial coherence factor and the brackets represent an ensemble average . the @xmath46 or two - point correlation function quantifies the relaxation dynamics between delay times of 0 and @xmath35 seconds . @xmath46
_ vs. _ delay time plots are obtained using dls experiments for aging laponite clay suspensions at several aging times @xmath1 .
representative plots are displayed in fig.1(a ) .
the normalized @xmath46 plots are seen to decay comparatively faster for smaller @xmath1 values .
as @xmath1 increases , the decay of @xmath46 slows down considerably .
furthermore , the decay of @xmath52 can be described as a two - step process @xcite . as reported earlier ,
the observed two - step relaxation function is a sum of a fast decaying exponential part and a comparatively slower stretched exponential part and can be expressed as @xmath53 here , @xmath54 is the fast or secondary relaxation time ( corresponding to the diffusion of a laponite particle inside the cage formed by the neighbors ) , @xmath55 is the structural or primary @xmath0-relaxation time ( representing its cooperative diffusion to a neighbouring position ) , @xmath56 is a stretching exponent , and @xmath57 is the weight factor for the faster secondary relaxation process .
the growth in the structural relaxation time @xmath55 arises from an evolution of the screened interparticle electrostatic repulsion due to a gradual process of tactoid exfoliation @xcite .
fits of the experimental data to eq.3 are shown in fig.1(a ) and the fitting parameters @xmath57 , @xmath55 and @xmath56 , obtained from the fits for several @xmath1 , are shown in table s1 in supporting information ( si ) .
@xmath58 is the fixed parameter in eq.3 which corresponds to the diffusion of a single laponite particle .
the details of the calculation is presented in si .
the dependence of the mean structural relaxation time @xmath59 on @xmath3 is obtained from the fits .
the @xmath43 is plotted in fig.1(b ) can be described by the vogel - fultcher - tammann ( vft ) law that governs the dynamical slowdown process in fragile supercooled liquids , but with @xmath60 is replaced with @xmath1 .
it is to be noted here that colloidal glasses have been studied successfully by invoking a vft growth of the primary relaxation timescale , but with @xmath18 replaced by @xmath6 .
similarly for aging laponite suspensions as investigated in earlier work @xcite , the vft equation is as follows : @xmath61 in this equation , @xmath62 , @xmath63 is the fragility parameter ( the inverse of @xmath63 quantifies the apparent deviation from the arrhenius trend ) , and @xmath64 is the waiting time at which the relaxation time diverges . for the data plotted in fig.1(b ) ,
@xmath65 = 1060@xmath66 65 @xmath31s , @xmath63 = 6.8 @xmath66 0.5 and @xmath67 = 31.3 @xmath66 1.3 hours .
such a vft like growth of @xmath43 with increasing @xmath1 as the laponite suspension approaches eventual kinetic arrest is strongly reminiscent of the observations reported in fragile molecular supercooled liquids with decreasing temperature @xmath4 and in many hard sphere colloidal glasses with increasing volume fraction @xmath6 .
typically , the fragile behavior of glass - forming liquids as they approach the glass transition is rationalized by enumerating the number of particles engaged in slow correlated motion , @xmath16 , in the dhs and the growing sizes of these regions of cooperative motion . moreover ,
the @xmath56 values , obtained from eq.3 are plotted as a function of @xmath1 and shown in fig.2(a ) .
the decrease in @xmath56 implies the width of time scale distributions characterising the dynamics of independently relaxing dhs increases monotonically with increasing @xmath1 .
the correlated dynamics can therefore be modelled by assuming linear superposition of exponential relaxation of dhs , with each dh having its own relaxation time @xmath68 .
distribution functions of relaxation times ( @xmath69 ) are calculated for the laponite system using the following kohlrausch - williams - watts equation @xcite as @xmath1 increases @xmath70 the distribution @xmath71 at four different @xmath72 and @xmath73 h are plotted in fig.2(b ) . a laponite suspension lying in the liquid - like regime ( _ i.e. _ at small @xmath2 values )
is characterised by a @xmath74 which shows a sharp peak at @xmath75 close to 1 .
in contrast , as the sample ages towards a non - ergodic state , a broad distribution of @xmath74 having peak position at @xmath76 is observed .
this suggests that at small sample ages , the relaxation timescales characterising the reorganization dynamics is more likely to have values within a narrow range .
the broad distribution at longer @xmath2 can be rationalized by considering the presence of simultaneous fast and slow moving cooperatively rearranging regions having their own independent relaxation times .
the broadening of @xmath74 with increase in the control parameter @xmath77 is reminiscent of observations in fragile supercooled liquids with decreasing @xmath4 and strongly indicates the presence of heterogeneous dynamics in aging colloidal laponite suspensions .
( b ) distribution of relaxation time scales @xmath69 plotted at various @xmath2 for the laponite - ps system of fig.1,width=316 ] characterizing @xmath16 of dhs in experiments is a challenging task .
theoretically , information about the sizes of the dhs and @xmath14 are embedded in the four point dynamic susceptibility which takes into account correlations in both space and time .
a detailed theoretical and mathematical treatment of the four point correlation function and its relation to @xmath16 can be found in the literature .
the four - point dynamic susceptibility @xmath78 is related to the fluctuating part of the self - intermediate scattering function @xmath79 by the following equation : @xmath80 experimentally therefore , @xmath81 can be obtained by resolving the dynamical behavior of the observable in both space and time .
in contrast to molecular systems , colloidal particles have comparatively larger sizes ( 10 @xmath40 m - 10 @xmath31 m ) .
this feature , and the easy tunability of inter - particle interactions , make colloidal systems robust model candidates for experimental and simulation studies of a variety of physical phenomena .
indeed , @xmath81 has been extracted using advanced microscopy techniques for colloidal glasses and in simulations . however , since computation of spatial correlations from experimental data on molecular and nanocolloidal suspensions , _ eg .
_ laponite , is extremely difficult , the four point dynamic susceptibility calculation method invariably fails in these cases . in order to study dhs in the aforementioned systems ,
therefore , the three - point correlation function , which is the lower bound of the four - point correlation function , has been introduced .
it can be accessed experimentally by probing the sensitivity of the two - point correlation function ( dielectric response or the scattering function ) to external control parameters such as @xmath4 in supercooled liquids and @xmath6 in colloidal systems . for hard sphere colloids ,
@xmath82 is written as @xmath83^{2 } \label{eq2}\ ] ] where @xmath84 is the fixed density value , @xmath85 is the particle number density , @xmath86 is the isothermal compressibility and @xmath87 is the derivative of @xmath88 with respect to @xmath6 .
the second term in eq.7 is the three - point dynamic susceptibility and can be accessed experimentally .
the derivative of @xmath88 with respect to @xmath6 , @xmath87 , has been obtained in dynamic light scattering experiments for colloidal systems , while in supercooled liquids , the thermal derivative of the dielectric response @xmath89 has been used to study @xmath16 @xcite . in these experiments ,
@xmath87 and @xmath89 were evaluated by probing , respectively , the scattering function and dielectric response at infinitesimal regular intervals of @xmath5 and @xmath4 .
the @xmath6-dependence of @xmath52 or the @xmath4 dependence of the dielectric spectrum are fitted with polynomial functions , with the derivatives of the fitted curve with respect to the control parameter yielding the three - point susceptibility .
the use of the three point correlation function formalism to calculate @xmath16 in a wide range of glass forming liquids using dielectric and light scattering data has been successfully demonstrated by dalle - ferrier _ et al . _ .
the authors showed that @xmath87 and @xmath90 show peaks similar to @xmath81 , with the peak height being proportional to @xmath16 .
decay curves at various @xmath2 for laponite - ps suspensions ( laponite concentration @xmath25 = 3 wt.% , and ps volume fraction @xmath32 = 5.66 @xmath27 10@xmath44 ) .
the solid lines are fits to eq.8 .
( b ) plots of @xmath91 _ ( q , t ) _
= @xmath92 as a function of delay time @xmath35 for the laponite - ps sample ( fig.1),width=316 ] calculated from the data in fig.3(b ) as @xmath1 increases towards the non - ergodic state .
the solid line in ( b ) is a power law fit of the form @xmath93 where @xmath94 = 1.17 @xmath66 0.02 and @xmath95 = 0.56 @xmath66 0.04.,width=316 ] ( black symbols ) and @xmath56 ( red symbols ) _ vs. _ @xmath96 .
@xmath16 shows a monotonic increase , with the initial rapid increase slowing down considerably at high @xmath96 while @xmath97 shows a monotonic decrease.,width=316 ] in our study , the procedure outlined above is implemented to obtain @xmath16 for aging laponite colloidal suspensions after replacing @xmath6 with @xmath1 .
the dependence of @xmath52 on @xmath1 is shown in fig.3(a ) .
it has been shown earlier that the decay curves in fig.3(a ) fit well to eq.3 @xcite . in order to obtain derivatives of @xmath52 with respect to @xmath1 , eq.3 is parametrized with eq.4 to obtain @xmath98 the @xmath99 decay curves , fitted to eq.8 , are depicted in fig.3(a ) using solid lines .
the fitting parameters thus obtained are in good agreement with the parameters extracted by fitting the data to eq.3 .
the fitting parameters are tabulated in table s2 of the si .
the three - point correlation function for aging laponite system , @xmath100 , is obtained by differentiating the fitted curves shown in fig.3(a ) with respect to @xmath1 , _
i.e. _ @xmath101 .
the calculated derivatives , shown in fig.3(b ) , exhibit peaks , with the peak heights increasing and the peak positions shifting to higher delay times @xmath35 with increasing @xmath1 .
the observed growth of the peak height of the @xmath100 , obtained using @xmath1 as the relevant control parameter , demonstrates the growing dynamical heterogeneities and the monotonically increasing @xmath16 during the spontaneous aging process of laponite colloidal suspensions .
the height of the peak , plotted in fig.4(a ) as a function of @xmath1 , is proportional to @xmath16 @xcite .
next , @xmath102 $ ] is calculated with the assumption that for the smallest waiting times probed in this work ( very low @xmath1 ) , correlated particle motion can be ruled out completely ( with particles moving independently ; @xmath103 1 ) .
the proportionality pre - factors in this relation for @xmath16 can be assumed to be independent of @xmath1 .
finally , @xmath104 $ ] is normalized with respect to the earliest @xmath1 value at all @xmath1 to obtain @xmath16 as a function of @xmath1 .
the corresponding @xmath14 plot is given in fig.4(b ) .
if @xmath105 s is defined as the non - ergodic transition point @xcite , a systematic growth of @xmath16 with increasing @xmath1 is observed .
the growth of @xmath16 shows a power law dependence on @xmath1 ( @xmath93 , where the exponent @xmath94 = 1.17 @xmath66 0.02 , @xmath95 = 0.56 @xmath66 0.04 ) and is shown by a solid line in fig.4(b ) .
similar power law growth of @xmath16 is observed for the measurements ( upto @xmath106 ) taken at a lower scattering angle @xmath38 = 60@xmath23 ( thereby probing longer length scales ) and the plot is given in the si in fig .
s1 . in fig.5
, @xmath16 is plotted as a function of @xmath15 normalized with @xmath107 s to observe its evolution .
@xmath56 values are also depicted in the same figure to correlate the evolution of the time scale distributions with growth of @xmath16 .
the decrease of @xmath56 ( or increase of the width of distribution of relaxation time scales , @xmath74 , as shown in fig.2(b ) ) as laponite suspensions age towards the glass transition indicates that the dynamics turn progressively heterogeneous.this implies growing correlations in the system _ i.e. _ the appearance of a large number of regions of correlated groups of particles with their own relaxation times , as the suspension ages .
this can be understood from the evolution of @xmath16 as a function of @xmath96 ( shown in fig.5 ) .
the appearance of dynamic heterogeneities of larger sizes ( characterised by @xmath16 ) results in a slowdown of the dynamics ( as reported in fig.1(b ) ) which manifests as a fragile supercooled liquid - like growth of @xmath43 ( vft - like plot in fig.1(b ) ) .
however , it is seen that @xmath16 grows rapidly at small @xmath15 , but slows down considerably close to the non - ergodic transition point .
remarkably , this trend is in close agreement with an observation in supercooled liquids , where @xmath108 is reported to show a power - law dependence at large @xmath4 ( _ i.e. _ in the liquid regime ) and a logarithmic growth at low @xmath4 ( _ i.e. _ near the glass transition ) .
this slow increase of @xmath16 close to the non - ergodic transition is attributed to activated dynamics . in this regime
, cooperative motion becomes increasingly difficult as the structural rearrangement time slows down dramatically as a consequence of the very high viscosity of the system .
this results in large @xmath43 values .
the increasing values of @xmath14 with @xmath1 , estimated from the three - point dynamic susceptibility in the present study , provides a quantitative measure of the fragile supercooled liquid - like heterogeneous dynamics in aging colloidal laponite suspensions .
dynamical heterogeneities in aging laponite colloidal suspensions are investigated by the three - point dynamic susceptibility formalism with respect to a new control parameter , _ viz . _
laponite colloidal suspensions exhibit fragile supercooled liquid - like dynamics , with the average primary relaxation time , @xmath43 , in its characteristic two - step relaxation process , exhibiting vft growth as a function of @xmath1 .
three - point dynamic susceptibilities are computed from dls experiments by taking the derivative of the two - point scattering decay function , @xmath52 , with respect to @xmath1 .
the three - point dynamic susceptibility exhibits a peak , with the peak height growing with increasing @xmath1 .
our calculations show that @xmath16 , the number of particles participating in correlated motion thereby setting the size of the dynamical heterogeneity , shows a power law increase with increase in the aging time @xmath1 of laponite suspensions .
furthermore , we show that the growth of @xmath16 is initially quite fast , before slowing down close to the glass transition .
we believe that our study provides valuable insight into the approach of laponite suspensions towards kinetic arrest by demonstrating that the growth of correlations in aging laponite suspensions closely resemble the observations reported for fragile supercooled liquids . unlike in fragile supercooled liquids , however , the dynamics of laponite suspensions are driven by athermal processes such as long - range screened inter - particle electrostatic repulsions and tactoid exfoliation .
given the similarities in the dynamical slowdown processes and the growth of correlations in laponite suspensions and fragile supercooled liquids , the ubiquitous kinetic slowdown in glass - forming liquids , driven by a growth in heterogeneous dynamics , could well be a universal feature of fragile glass - formers .
g. wolynes , v. lubchenko , _ structural glasses and supercooled liquids : theory , experiment , and applications _ , john wiley & sons , new jersey ( 2012 )
. m. d. ediger , c. a. angell , and s. r. nagel , j. phys . chem . * 100 * , 13200 - 13212 ( 1996 ) .
p. n. pusey and w. van megen , nature ( london ) * 320 * , 340 ( 1986 ) .
c. a. angell , science * 267 * , 1924 ( 1995 ) . c. a. angell , j. non - cryst .
solids * 131 - 133 * , 13 - 31 ( 1991 ) . _ dynamical heterogeneities in glasses , colloids and granular media _ , edited by l. berthier , g. biroli , j .-
bouchaud , l. cipelletti , and w. van saarloos ( oxford university press , new york , 2010 ) .
g. adam and j. h. gibbs , j. chem .
phys . * 43 * , 139 ( 1965 ) .
r. richert , j. phys . : condens .
matter * 14 * , r703 ( 2002 ) .
j. d. stevenson , j. schmalian , and p. g. wolynes , nature phys . 2 , 268 ( 2006 ) m. d. ediger , annu .
phys . chem . *
51 * , 99 ( 2000 ) .
z. zhang , p. j. yunker , p. habdas , and a. g. yodh , phys.rev.lett . * 107 * , 208303 ( 2011 ) .
p. g. debenedetti and f. h. stillinger , nature ( london ) * 410*,259 ( 2001 ) .
c. dalle - ferrier , c. thibierge , c. alba - simionesco , l. berthier , g. biroli , j .-
bouchaud , f. ladieu , d. lhte , and g. tarjus , phys . rev .
e * 76*,041510 ( 2007 ) . c. dalle - ferrier1 , s. eibl1 , c.pappas and c.alba-simionesco , j. phys . : condens . matter * * 20**494240(2008 ) .
g. brambilla , d. el masri , m. pierno , l. berthier , and l. cipelletti , phys.rev.lett .
* 102*,085703 ( 2009 ) .
l. berthier , g. biroli , j .-
bouchaud , l. cipeletti , d. el masri , d .
lhte , f. ladieu , and m. pierno , science * 310 * , 1797 ( 2005 ) . c. crauste - thibierge , c. brun , f. ladieu , d. lhte , g. biroli , and j .-
bouchaud , phys .
lett . * 104 * , 165703 ( 2010 ) . c. brun , f. ladieu , d. lhte , m. tarzia , g. biroli , and j .- p.bouchaud , phys .
b * 84 * , 104204 ( 2011 )
. t. bauer , p. lunkenheimer , and a. loidl , phys .
lett.*111 * , 225702 ( 2013 ) .
herman z. cummins , j. non-cryst.solids * 353 * , 3891 ( 2007 ) .
b.ruzicka and e.zaccarelli , soft matter , * 7 * , 1268 ( 2011 ) .
a. shahin and y. joshi , langmuir , * 26 * , 4219(2010 ) .
r. angelini , et.al .
. commun . * 5*,4049 ( 2014 ) . v. n. manoharan , science * 349
* , 1253751 ( 2015 ) .
p. sollich , f. lequeux , p. hbraud , and m. e. cates , phys .
lett . * * 7**8 , 2020 ( 1997 ) mourchid a , lecolier e , van damme h and levitz p , langmuir * 14 * , 4718 ( 1998 ) .
r. bandyopadhyay , d. liang , h. yardimci , m. a. borthwick , d. a. sessoms , s. g. j. mochrie , j. l. harden and r. l. leheny , phys . rev
. lett . * 93 * , 228302 , ( 2004 ) .
y. m. joshi , j. chem . phys . * 127 * , 081102 ( 2007 ) .
d. saha , y.m.joshi,and r. bandyopadhyay , europhys . lett . * 112 * , 46004 ( 2015 ) .
d. bonn , h.tanaka , g.wegdam,h .
kellay and j. meunier , europhys .
* 45 * , 52 - 57 , ( 1998 ) .
d. saha , r. bandyopadhyay and y.joshi , langmuir , * 31 * , 3012(2015 ) .
d. saha , y. m. joshi , and r. bandyopadhyay , soft matter * 10 * , 3292 ( 2014 ) .
b. ruzicka , l. zulian , and g. ruocco , j. phys . :
matter * 16 * , s4993-s5002 ( 2004 ) .
d. saha , y. m. joshi , and r. bandyopadhyay , j. chem . phys .
* 143 * , 214901 ( 2015 ) .
s. jabbari - farouji , g. h. wegdam , and d. bonn , phys.rev.e * 86 * , 041401 ( 2012 ) .
s. jabbari - farouji , r. zargar , g. h. wegdam and d. bonn , soft matter , * 8 * , 5507 ( 2012 ) .
l. cipelletti and l.ramos , j. phys . : condens
. matter , * 17 * , r253-r285 ( 2005 ) . b. j. berne and r. pecora , dynamic light scattering : with applications to chemistry , biology , and physics ( john wiley & sons , new york , 1975 ) .
p. n. pusey , in photon correlation spectroscopy and velocimetry , edited by h. z. cummins and e. r. pike ( plenum , new york , 1977 ) . w. van megen et.al . , phys.rev.e * 58 * , 6073 ( 1998 ) .
b. ruzicka , l. zulian , and g. ruocco , phys .
lett . * 93 * , 258301 ( 2004 ) .
s. ali and r. bandyopadhyay , langmuir , * 29 * , 12663 ( 2013 ) .
g. l. hunter and e. r. weeks , rep .
. phys . * 75 * , 066501(2012 ) . c. p. lindsey and g. d. patterson , j. chem
. phys . * 73 * , 3348 ( 1980 ) .
n. lacevic , f. w. starr , t. b. schroder , and s. c. glotzer , j. chem . phys . * 119 * , 7372 ( 2003 ) . c. toninelli , m. wyart , l. berthier , g. biroli , and j - p .
bouchaud , phys .
e * 71 * , 041505 ( 2005 ) .
k. n. nordstrom , j. p. gollub , and d. j. durian , phys .
e * 84 * , 021403 ( 2011 ) .
l. berthier , g. biroli , j - p .
bouchaud , w. kob , k. miyazaki and d. r. reichman , j. chem .
phys.*126 * , 184503(2007 ) .
l. berthier , g.biroli , j - p .
bouchaud , w. kob , k. miyazaki and d. r. reichman , j. chem . phys . * 126 * , 184504 ( 2007 ) . | an aqueous suspension of the synthetic clay laponite can undergo a transition from a liquid - like ergodic state to a glass - like nonergodic arrested state . in an observation that closely resembles the dynamical slowdown observed in supercooled liquids subjected to a rapid temperature quench , the phenomenon of kinetic arrest in laponite suspensions is accompanied by a growth in the @xmath0-relaxation or structural relaxation time with increasing sample aging time , @xmath1 .
the ubiquitous dynamic slowdown and fragile behavior observed in glass forming liquids approaching the glass transition is typically ascribed to the heterogeneous dynamics and the growth in the size of distinct dynamical heterogeneities . in this article , we present the characterization of the dynamical heterogeneities in aging colloidal laponite clay systems by invoking the three - point dynamic susceptibility formalism . the average time - dependent two - point intensity autocorrelation and its sensitivity to the control parameter @xmath2 are probed in dynamic light scattering experiments .
distributions of relaxation time scales deduced from kohlrausch - williams - watts equation widen with increasing @xmath2 signifying the heterogeneous dynamic slowdown . a suitable formalism to calculate three - point correlation function is employed for aging colloidal suspension where the main control parameter is @xmath1 .
the calculated three - point dynamic susceptibility exhibits a peak , with the peak height increasing with evolving @xmath3 .
the number of dynamically correlated particles , deduced from the peak - height , is seen to initially increase with increasing @xmath3 at a fast rate , before eventually slowing down close to the non - ergodic transition point .
this observation is in agreement with published reports on supercooled liquids .
our study confirms the presence and growth of dynamical heterogeneities in soft glassy suspensions of laponite , thereby shedding new light on the fragile supercooled liquid - like dynamics of aging suspensions of these anisotropic , charged , colloidal clay nanoparticles . |
regular and noninvasive measurement of vital signs such as pulse rate ( pr ) , breathing rate ( br ) , pulse rate variability ( prv ) , blood oxygen level ( spo2 ) and blood pressure ( bp ) are important both in - hospital and at - home due to their fundamental role in the diagnosis of health conditions and monitoring of well - being . currently , the gold standard techniques to measure the vital signs are based on contact sensors such as ecg probes , chest straps , pulse oximeters and blood pressure cuffs .
however , contact - based sensors are not convenient in all scenarios , e.g. contact sensors are known to cause skin damage in pre - mature babies during their treatment in a neonatal intensive care unit ( nicu ) .
non - contact methods for vital sign monitoring using a camera has been recently shown to be feasible @xcite .
being non - contact , camera - based vital sign monitoring have many applications from monitoring newborn babies in the nicu to in - situ continuous monitoring in everyday scenarios like working in front of a computer .
however , camera - based vital sign monitoring does not perform well for subjects having darker skin tones and/or under low lighting conditions as was highlighted in @xcite .
furthermore , current known algorithms require a person to be nearly at rest facing a camera to ensure reliable measurements . in this paper , we address the challenge of reliable vital sign estimation for people having darker skin tones , under low lighting conditions and under different natural motion scenarios to expand the scope of camera - based vital sign monitoring .
photoplethysmography ( ppg ) is an optical method to measure cardiac - synchronous blood volume change in body extremities such as the face , finger and earlobe . as the heart pumps blood , the volume of blood in the arteries and capillaries changes by a small amount in sync with the cardiac cycle .
the change in blood volume in the arteries and capillaries underneath the skin leads to small change in the skin color .
the goal of a camera - based vital sign monitoring system is to estimate the ppg waveform which is proportional to these small changes in skin color .
vital signs such as pr , prv , spo2 and br can be derived from a well - acquired ppg waveform . the two major challenges in estimating ppg using a camera are : ( i ) extremely low signal strength of the color - change signal , particularly for darker skin tones and/or under low lighting conditions , and ( ii ) motion artifacts due to an individual s movement in front of the camera .
our main contribution in this paper is a new algorithm , labeled as _ distanceppg _ , that improves the signal strength of camera - based ppg signal estimate , with following three key contributions . * a new method to improve the snr of camera - based ppg signal by combining the color - change signals obtained from different regions of the face using a weighted average . * a new automatic method for determining the weights based only on the video recording of the subject .
the weights capture the effect of incident light intensity and blood perfusion underneath the skin on the strength of color - change signal obtained from a region . * a method to track different regions of the face separately as the person moves in front of the camera using a combination of a deformable face tracker @xcite and klt ( kanade - lucas - tomasi ) feature tracker @xcite to extract the ppg waveform under motion . for different skin tones ( pale white to brown ) ,
the distanceppg algorithm improves the signal to noise ratio ( snr ) of the estimated ppg signal on an average by @xmath0 db compared to prior methods @xcite . particularly , the improvement in snr for non - white skin tone is @xmath1 db . we have evaluated ppg estimation under three natural motion scenarios : ( i ) reading content on screen , ( ii ) watching video , and ( iii ) talking .
distanceppg improves the snr of estimated camera ppg in these scenario by @xmath2 db on an average .
further , it improves the snr of camera based ppg by as much as @xmath3 db under low lighting condition when compared to prior methods @xcite .
the improvement in snr of camera - based ppg signal estimated using distanceppg reduces the error in pulse rate estimates in various scenarios . using distanceppg ,
the mean bias ( average difference between ground truth pulse oximeter derived pr and camera - based ppg derived pr ) @xmath4 is @xmath5 beats per minute ( bpm ) with @xmath6 limit of agreement ( mean bias @xmath7 standard deviation of the difference ) between @xmath8 to @xmath9 bpm for @xmath10 subjects having skin tones ranging from fair white to brown / black .
when using prior methods , the corresponding performance numbers are @xmath4 = @xmath11 with @xmath6 limit of agreement between @xmath12 to @xmath13 . under three motion scenario of reading , watching , and talking for @xmath14 subjects of varying skin tones , the mean deviation is @xmath4 = @xmath15 bpm with @xmath6 limit of agreement between @xmath16 to @xmath17 bpm using distanceppg . using prior methods , the corresponding performance numbers are @xmath18 bpm with @xmath6 limit of agreement between @xmath19 to @xmath20 bpm .
further , distanceppg reduces the root mean square error ( rmse ) in pulse rate variability estimation below @xmath21 ms for non - black skin tones using a @xmath22 fps camera . over the past decade , several researchers have worked on measuring vital signs such as pr , prv , br , and spo2 using a camera @xcite .
initially @xcite , external arrays of leds at @xmath23 nm and @xmath24 nm were used to illuminate a region of tissue for measuring pr using a monochrome cmos camera .
it was shown later @xcite that pr and br can also be determined using simply a color camera and ambient illumination .
the authors in @xcite found that the face is the best region to extract ppg signal because of better signal strength .
they also reported that the _ green channel _ of the rgb camera perform better than red and blue channel for detecting pr and br .
the fact that the green channel perform better is expected as the absorption spectra of hemoglobin ( hb ) and oxyhemoglobin ( hbo@xmath25 ) , two main constituent chromophores in blood , peaks in the region around @xmath26 nm , which is essentially the passband range of the green filters in color cameras .
further in @xcite , the authors used a color webcam under ambient illumination to detect simultaneous pr , prv and br of multiple people in a video by using automatic face detection to define the face region .
they used blind source separation ( bss ) to decompose the three camera channels ( red , green , blue ) into three independent source signals using independent component analysis ( ica ) algorithm , and extracted the ppg signal from one of the independent sources .
more recently , authors in @xcite demonstrated that cyan , orange and green ( cog ) channels work better than red , green , and blue ( rgb ) channels of a camera for camera - based vital sign estimation .
one possible explanation for better performance of cyan - orange - green ( cog ) channels could be the higher overlap between the passband of cyan ( @xmath27 nm ) , green ( @xmath28 nm ) , and orange ( @xmath29 nm ) with the peak in the absorption spectra of hb and hbo@xmath25 ( @xmath26 nm ) .
most of the past work , however , did not report how camera - based ppg performs on individuals with different skin tones , under low lighting conditions , and for different motion scenarios .
it is well - known that the higher amount of _ melanin _ present in darker skin tones absorbs a significant amount of incident light , and thus degrades the quality of the camera - based ppg signal , making the system ineffective for extracting vital signs in darker skin tones .
recently , a pilot study conducted in nicu for monitoring pulse rate using a camera - based method showed difficulty under low lighting conditions and/or under infant motion @xcite . to counter the motion challenge , the authors in @xcite used automatic face detection in consecutive frames to track the face .
but , they reported difficulties in continuously detecting faces under motion due to large false - negatives .
another method is to compute 2d shifts in face location between consecutive frames @xcite using image correlation to model the motion . by simply computing the global 2d shifts
, one can only capture the basic translation motion of face , and it is difficult to compensate for more natural motion like turning or tilting of the face , smiling or talking , generally found in in - situ scenarios .
for a camera - based ppg estimation we record the video of a person facing a camera , and the objective is to develop an algorithm to estimate the underlying ppg signal @xmath30 using the recorded video .
the recorded video is in the form of intensity signal @xmath31 comprising of sequences of frames @xmath32 . each frame of the video records the intensity level of the light reflected back from the face over a two dimensional grid of pixel @xmath33 in the camera sensor .
if the camera sensor has multiple color channels ( e.g. red , green , blue ) , one can get separate intensity signals corresponding to each channel ( e.g. @xmath34 ) .
in general , the measured intensity of any reflected light can be decomposed into two components : ( i ) intensity of illumination , and ( ii ) reflectance of the surface ( skin ) , i.e. @xmath35 the illumination intensity corresponds to the intensity of ambient or any dedicated light falling on the face . for ppg estimation , it is generally assumed that the light intensity remains same over the ppg estimation window ( typically @xmath36 sec in past works ) .
the skin reflectance @xmath37 is equal to the fraction of light reflected back from the skin and consists of two level of light reflectance : ( i ) surface reflection , and ( ii ) subsurface reflection or backscattering . a large part of the light incident on face gets reflected back from the surface of the skin , and is characterized by the skin s bidirectional reflectance distribution function ( brdf ) .
remaining part of the incident light goes underneath the skin surface and is absorbed by the tissue and the chromophores ( hb , hbo@xmath25 ) present in blood inside arteries and capillaries .
the volume of blood in the arteries and capillaries changes with each cardiac cycle and thus the level of absorption of light changes as well .
since ppg signal , by definition , is proportional to this cardio - synchronous pulsatile blood volume change in the tissue @xcite , one can estimate ppg signal @xmath30 by estimating these small changes in subsurface light absorption .
thus , the camera - based ppg signal is estimated by extracting small variations in the subsurface component of skin reflectance @xmath37 .
since the incident light intensity @xmath38 is assumed to be constant over ppg estimation time window , any temporal change in the intensity of the light reflected back from the face region will be proportional to the changes in the reflectance of the skin surface @xmath37 . generally , these temporal changes in recorded intensity will be dominated by changes in surface reflection component unrelated to the underlying ppg signal of interest . as a first step for camera - based ppg estimation , one can spatially average the recorded intensity level @xmath31 over all the pixels in the face region to yield a measurement point per frame in the video @xcite .
the basic idea is that by averaging the intensity signal , the incoherent changes in surface reflection component over all the pixels inside the face will cancel out , and the coherent changes in subsurface reflection component due to blood volume changes will add up to give an estimate @xmath39 .
the spatially averaged intensity signal @xmath39 would be proportional to the changes in the subsurface reflectance component , and thus to the underlying ppg signal of interest .
one generally filters @xmath39 signal between @xmath40 hz to @xmath14 hz ( frequency band of interest ) to extract the ppg signal .
* challenge 1 : very low signal strength * : ppg signal extracted from camera video have low signal strength .
this is because the skin vascular bed contains a very small amount of blood ( only @xmath41 of total blood in the body ) , and the blood volume itself experiences only a small ( @xmath42 ) change in sync with the cardiovascular pulse @xcite .
thus , the change in subsurface skin reflectance due to cardio - synchronous changes in blood volume is very small . the small change in subsurface reflectance results in very small change in light intensity recorded using a camera placed at a distance . on the other hand ,
the change in surface reflection component due to small movement of person is very large .
for example , see figure [ fig : roi ] where the top plot ( red ) shows recorded intensity changes in a single pixel marked on the forehead ( @xmath43 ) . when compared to the ground truth pulse oximeter signal ( bottom , @xmath44 ) , it is clear that the raw intensity change variations in @xmath43 are unrelated to the underlying ppg signal and is dominated by a significant amount of surface reflectance changes to estimate small changes in subsurface reflectance , a general idea used in most past work is to spatially average the recorded intensity level over the face region . for example , see the plot of @xmath39 in figure [ fig : roi ] . clearly , @xmath39 is a better estimate of the ppg signal as is evident by comparing it with the ground truth pulse oximeter signal ( bottom ) .
the amplitude of @xmath39 is within @xmath45 in camera intensity scale ( all these signals are filtered between @xmath46 $ ] and so they are centered around @xmath47 ) .
this is at the level of the quantization step of @xmath48 bit adc present in cameras and thus camera - based ppg estimate is corrupted with large quantization noise .
now , let us suppose we spatially average only the pixels within the square blocks @xmath49 marked on the face ( @xmath50 px in size ) in figure [ fig : roi ] . by comparing the plot of @xmath51 , @xmath52 , and @xmath53
, we can clearly see that there are differences in the quality of estimated color change ( or ppg ) signal from these regions ( @xmath54 is a better estimate than @xmath55 which is better than @xmath53 ) .
past work has not fully exploited the fact that the quality of ppg estimates obtained from different regions of the face varies significantly across the face .
strength of the ppg signal extracted from a patch of the imaged skin surface would depend on the intensity of light @xmath38 incident on that patch and on the total amount of blood volume change underneath the skin patch .
the total amount of blood volume change will depend on the blood perfusion in the region , which in turn will be determined by the density of blood carrying capillaries in that region .
when pixel intensity is averaged over the whole face region to get @xmath39 , one ends up also including skin patches that have very limited blood perfusion , and hence contribute more noise than signal to the overall estimate of the camera - based ppg signal .
thus , there is a need for an automated method to find out which regions in the face can be used to estimate better ppg signal and which regions should be rejected because they contribute more noise than signal to the overall ppg estimate .
this is a challenging task because we do not have access to the ground truth ppg signal , and thus can not determine the quality of ppg estimation from different regions in the face .
further , it will also be desirable to combine the ppg signal obtained from different regions of the face in a manner that the overall estimate has maximum signal to noise ratio ( snr ) .
0.35 ( img ) at ( 0,0 ) from different regions of the face using a weighted average to maximize the snr of the overall estimate of the ppg signal , @xmath43 shows the ppg signal obtained from a single pixel in the forehead ( marked in red ) for comparison .
, title="fig:",scaledwidth=100.0% ] ; ( text ) at ( @xmath56 ) ppg signal strength varies over different regions in the face and depends on intensity of incident light @xmath38 in the region and on the blood perfusion underneath the skin . ; [ fig : faceroi ] 0.65 table[x = time , y = raw]result / regionppg.txt ; ( raw_text ) at ( @xmath57 ) @xmath43 ; [ name = face , at=(@xmath58 ) , width=3 in , height=1 in , anchor = north , xmin=-1 , xmax=20 , xtick= , ymin=-.5 , ymax=.5 , ytick=-0.25,0.25 , axis x line = none , axis y line = left , y axis line style=- , ] table[x = time , y = face]result / regionppg.txt ; ( face_text ) at ( @xmath59 ) @xmath39 ; [ name = ra , at=(@xmath60 ) , width=3 in , height=1 in , anchor = north , xmin=-1 , xmax=20 , xtick= , ymin=-2 , ymax=2 , ytick=-1,1 , ylabel = camera intensity scale , axis x line = none , axis y line = left , y axis line style=- , ] table[x = time , y = r1]result / regionppg.txt ; ( ra_text ) at ( @xmath61 ) @xmath51 ; [ name = rb , at=(@xmath62 ) , width=3 in , height=1 in , anchor = north , xmin=-1 , xmax=20 , xtick= , ymin=-2 , ymax=2 , ytick=-1,1 , axis x line = none , axis y line = left , y axis line style=- , ] table[x = time , y = r2]result / regionppg.txt ; ( rb_text ) at ( @xmath63 ) @xmath52 ; [ name = rc , at=(@xmath64 ) , width=3 in , height=1 in , anchor = north , xmin=-1 , xmax=20 , xtick= , ymin=-2 , ymax=2 , ytick=-1,1 , axis x line = none , axis y line = left , y axis line style=- , ] table[x = time , y = r3]result / regionppg.txt ; ( rc_text ) at ( @xmath65 ) @xmath66 ; [ name = ground , at=(@xmath67 ) , width=3 in , height=1 in , anchor = north , xmin=-1 , xmax=20 , axis x line = bottom , axis y line = left , y axis line style=- , ymin=-0.5 , ymax=0.5 , ytick = , xlabel = time(s ) , ] table[x = time , y = ground]result / regionppg.txt ; ( ra_text ) at ( @xmath68 ) @xmath69 ; * challenge 2 : motion artifact * : until this point we have assumed that the person facing the camera is static ( we will assume that the camera is static throughout this paper ) .
if the person moves in front of camera , e.g. tilting , smiling , shifting , talking etc , one needs to at the very least track the whole face to extract the ppg signal . as the face is not a rigid body , in a sense that different regions within the face can move separately ( e.g. cheek muscles during talking , smiling etc )
, it would be ideal to track different non - rigid regions of the face independently . even if we could track different regions within the face faithfully ,
there are other challenges in faithful ppg estimation under motion .
first , most light sources have a spatial illumination variations @xmath38 , and when a tracked region moves in space , it leads to change in incident light intensity in that region over time .
such changes violate the assumption that incident light intensity is constant over ppg estimation window and corrupt the ppg estimate even when the tracking works perfectly .
second , even very small motion ( e.g. slight rotation ) of person s face relative to camera can lead to change in incident light direction ( from light source to skin surface ) and reflected light direction ( from skin surface to camera sensor ) .
such small changes in light direction can lead to large changes in skin surface reflectance which is characterized by the highly non - linear brdf of the skin surface .
this large change in surface reflectance can completely overwhelm the measurement of small changes in subsurface reflectance due to changes in blood volume in sync with the cardiac cycle ( our ppg signal ) .
the fact that blood volume change underneath the skin causes very small changes in the intensity of reflected light signal @xmath31 recorded by the camera can be used to our advantage . for any region ( skin patch ) inside the face to be useful in estimating the ppg waveform , the corresponding averaged pixel intensity signal from that region should not change by very large magnitude within the ppg estimation window . any large change in intensity is mostly due to changes in incident light intensity @xmath38 or surface reflectance , and would be much larger than the intensity changes we are interested in .
so , it would be wise to identify such bad regions , and reject them completely till the time they are contributing large artifacts .
we first propose a camera - based ppg signal acquisition model which captures the effect of incident light intensity , surface and subsurface skin reflectance , and camera noise on ppg signal estimation . based on the proposed signal acquisition model ,
we develop distanceppg estimation algorithm that comprises of two subparts : i ) _ mrc _ ( maximum ratio combining ) algorithm to combine the average pixel intensity signal extracted from different regions of the face to improve the overall signal to noise ratio ( snr ) of the final ppg estimate , and ( ii ) _ region based motion tracking _
algorithm that keeps track of different regions of the face as the person moves in front of the camera .
overall steps involved in distanceppg algorithm are shown in figure [ fig : blockdiag ] .
( step1 ) at ( 0,0 ) ; at ( @xmath70 ) * step 1 : * input : green channel video of a person , landmark points around eyes , nose , mouth in face detected ; ( step2 ) at ( 3.3,0 ) ; at ( @xmath71 ) * step 2 : * face is divided into seven regions , each region tracked using klt , motion modeled using rigid affine fit ; ( step3 ) at ( 6.51,0 ) ; at ( @xmath72 ) * step 3 : * each tracked region is divided into @xmath73x@xmath73 pixel block , avg .
pixel intensity @xmath74 computed from each roi ; ( step4 ) at ( 9.75,0 ) ; at ( @xmath75 ) * step 4 : * goodness metric @xmath76 computed for each roi , overall camera ppg estimated using weighted average ; ( step1.east ) ( step2.west ) ; ( step2.east ) ( step3.west ) ; ( step3.east ) ( step4.west ) ; ( @xmath77 ) ( @xmath78 ) ( @xmath79 ) ( @xmath80 ) ( @xmath77 ) ; ( @xmath81 ) ( @xmath82 ) ( @xmath83 ) ( @xmath84 ) ( @xmath81 ) ; at ( @xmath85 ) * region based motion tracking * ; at ( @xmath86 ) * mrc algorithm * ; given the intensity signal @xmath31 of a person facing a camera , we assign the set of pixels imaging the face into many small regions of interest ( rois ) denoted by the set @xmath87 .
these rois are small enough that the blood perfusion within them can be assumed to be constant .
let @xmath74 be the spatial average of the intensity of the pixel within the roi @xmath88 at time @xmath89 , @xmath90 is used to index the elements of the set @xmath87 .
also , the spatial illumination variation @xmath38 inside an roi @xmath88 can be assumed to be constant , and will be denoted as @xmath91 .
then , @xmath74 can be modeled as @xmath92 where @xmath91 is the incident light intensity in roi @xmath88 , @xmath93 is the strength of blood perfusion , @xmath94 is the surface reflectance from the skin in region @xmath88 , and @xmath95 is the camera quantization noise .
when incident illumination @xmath91 falls on skin roi @xmath88 , a large fraction of it ( @xmath94 ) is reflected back by the skin surface and does not contain any pulsatile component related to blood volume change .
some part of the incident light penetrates beneath the skin surface , and gets modulated by the pulsatile blood volume change waveform @xmath30 , due to light absorption , before reflecting back ( back - scattering ) . here , @xmath93 represent the strength of modulation of light backscattered from the subsurface due to the pulsatile blood volume change .
the parameter @xmath93 will primarily depends on the average blood perfusion in the selected roi , and thus varies over different regions of the face and is also different for different individuals .
the parameter @xmath93 will also depend on the wavelength of incident light since the absorption spectra of hb and hbo@xmath25 , two major chromophores present in blood , depends on wavelength of light . in this work ,
we only consider camera s green channel that coincides with the peak in the absorption spectra of hb and hbo@xmath25 , this choice provides best performance in ppg estimation as was also highlighted by other researchers @xcite .
so , we did not explicitly incorporate wavelength dependence of @xmath93 in the model equation .
our goal is to extract @xmath30 from the measurements @xmath74 in the presence of large surface reflectance @xmath94 which constitutes a dominant portion of the signal captured by the camera sensor .
quantization noise is also significant as our signal of interest @xmath30 varies by a small amount relative to the surface reflection .
if the person moves in front of the camera , then all the rois in set @xmath87 need to be tracked .
given the intensity signal @xmath31 of a person facing a camera , we first compute all the @xmath74 corresponding to the roi @xmath88 inside the face .
each @xmath74 contain different strength of the underlying ppg signal @xmath30 , and different surface reflection component @xmath96 .
then , we temporally filter all @xmath74 using a bandpass filter @xmath46 $ ] to reject the out of band component of skin surface reflection ( @xmath96 ) and other noise outside the band of interest to obtain @xmath97 . based on the camera - based ppg signal acquisition model in section [ sec :
ppg_acquisition_model ] , the filtered signals from different regions of the face can be written as @xmath98 where @xmath99 represents the corresponding roi number in @xmath87 . here ,
@xmath100 denote the strength of the underlying ppg signal in region @xmath88 and is determined both by the strength of modulation @xmath93 and the incident light illumination @xmath91 .
further , @xmath101 denote the noise component due to the camera quantization , unfiltered surface reflection and motion artifacts . here , @xmath102 can be considered as different channels that receive different strength of the same desired signal @xmath30 , and have different level of noise .
we can combine all these different channels using a weighted average @xmath103 the weights for each channel can be determined based on the the idea of maximum ratio diversity @xcite .
the maximum ratio diversity algorithm states that the assigned weights should be proportional to the root - mean - squared ( rms ) value of the signal component , and inversely proportional to the mean - squared noise in that channel , in order to maximize the signal - to - noise ratio of the overall camera - based ppg estimate @xmath104 ; mathematically , @xmath105 as both @xmath100 and @xmath106 are unknown in our case , we need to develop an alternative method to estimate these weights @xmath76 . for ease of reference , we label @xmath76 as _ goodness metric _ for region @xmath88 from hereon .
the maximum ratio diversity algorithm assumes that the signal component @xmath107 is locally coherent among all channel , i.e. there is no time delay between ppg signals extracted from different rois , and the noise component @xmath101 is uncorrelated
. generally speaking , ppg signal obtained from different regions of the skin would exhibit varying delays as the blood reaches these regions at different times .
for example , there is a time lag of @xmath108 ms in ppg recorded between finger and toe @xcite .
however , the time lag between the ppg signal obtained from close - by regions are small , e.g. all the regions inside the face .
our own measurements show that the delay is less than @xmath109 ms between farthest point in face . as this delay falls within one sample period of normal cameras having frame rate of @xmath110 hz
, we can neglect such small delays for all practical purposes , and thus the signal component @xmath107 can be considered as locally coherent . as the amplitude of the camera - based ppg signal of interest @xmath107 is generally very small ( within 1 - 2 bits of the camera adc )
, we also reject regions which have unusually large signals .
large variations are mostly due to illumination change or motion artifacts .
thus , we reject all regions having amplitude greater than a threshold @xmath111 . our final estimate of the ppg signal over a time window of @xmath112 sec is given by @xmath113 where @xmath114 is the indicator function , @xmath115 } \hat{y}_i(t)$ ] , @xmath116 } \hat{y}_i(t)$ ] is the maximum and minimum amplitude of the @xmath97 over a @xmath112 sec duration .
the ppg signal @xmath30 has a fundamental frequency of oscillation equal to the pulse rate .
thus , the spectral power of the ppg signal is concentrated in a small frequency band around the pr . moreover
, the spectral power of the noise @xmath101 present in @xmath97 will be distributed over the passband of the filter @xmath46 $ ] . based on this spectral structure of the signal
, we can estimate the goodness metric as a ratio of the power of recorded @xmath97 around the pulse rate ( pr ) to the power of the noise in the passband of the filter .
let @xmath117 be the _ power spectral density _
( psd ) of @xmath118 over time duration @xmath119 $ ] .
then goodness metric @xmath76 can be defined as @xmath120 where @xmath121 $ ] denote a small region around the pulse rate ( pr ) of the person , and @xmath122 $ ] is the passband of the bandpass filter ( @xmath46 $ ] ) .
the definition of @xmath76 depends on the pulse rate which is still unknown .
a two - step process can be followed to get a coarse estimate of pulse rate .
first , we assume that all the @xmath76 are equal to @xmath123 , and thus we compute a temporary coarse estimate of the ppg signal using equation .
second , we compute the frequency corresponding to the peak in the spectrum of this coarse estimate of ppg signal to get a coarse estimate of the pulse rate to be used in goodness metric definition .
figure [ fig : mrc_algorithm](a ) shows the psd of acquired ppg signal and highlights how the goodness metric is defined based on the area under the psd .
figure [ fig : mrc_algorithm](b ) shows a face with the goodness metric overlay .
the goodness metric is based only on the recorded video of a person s face , and thus adapts to changes in blood perfusion for different people or changes in lighting conditions .
regions shown as more red have higher goodness value , those shown in blue have lower goodness metric .
figure [ fig : mrc_algorithm](c ) shows the ppg signal extracted from four different regions marked on the face .
when we compare these ppg signals with the pulse oximeter based ground truth , it is evident that the goodness metric correctly predicts the strength of the ppg signal extracted from these regions relative to the noise power .
thus , forehead ( roi 1 , goodness metric = @xmath124 ) and cheek ( roi 2 , goodness metric = @xmath125 ) regions give higher strength ppg signal estimates , whereas the region around mouth ( roi 3 , goodness = @xmath126 ) and eyes ( roi 4 , goodness = @xmath127 ) do not yield good signal .
since the roi @xmath128 around eyes gives spiky signal due to eye movements , the signal amplitude in this region crossed the amplitude threshold @xmath111 .
thus , the weight given to roi @xmath128 is @xmath127 , or equivalently that region is rejected by the algorithm . in camera - based acquisition of ppg ,
even a slight motion of person in front of camera can change the relative position of selected camera s roi @xmath88 and imaged skin portion . as our mrc algorithm requires ppg extraction from all the different rois independently , we need to keep track of each selected roi on the skin surface as the person moves in front of the camera . as a first step in tracking ,
we extract landmark location from the face like location of eyes , nose , mouth and the outer boundary using a deformable face fitting algorithm described in @xcite .
relative to these landmark locations , we then define seven planar regions in the face denoted by set @xmath129 for tracking three on forehead , left and right cheek , and two on the chin .
particularly , we do not include regions around the mouth and eyes , as they exhibit non - rigid motion and are difficult to track and thus causes large motion artifacts . as a next step , to track these planar regions on the face across the video frames , we identify @xmath130 feature points inside each planar region .
these feature points are selected so that they could be tracked well , and are known as _ good features to track _
@xcite in computer vision literature .
we then use the kanade lucas tomasi ( klt ) feature tracker @xcite to track these features across the video .
afterwards , we compute a rigid affine fit for each planar region using the tracked feature points inside corresponding planar region to model its motion between consecutive frames we use more feature points than is minimally required for a rigid affine fit ( @xmath131 feature points ) as feature tracking is generally error prone due to feature points changing appearance or disappearing from camera view due to occlusion .
we use the method proposed in @xcite to automatically detect tracking failures based on the forward - backward error and do not consider erroneous points for affine model estimation . from the remaining well - tracked feature points , we use the random sample consensus ( ransac )
@xcite algorithm to compute a robust estimate of affine fit by considering only the inlier points for the model , and rejecting the outliers .
we then divide these seven planar regions @xmath129 on the face into many small ppg region of interest required by the mrc algorithm .
each ppg roi @xmath88 is tracked across the video frames using the estimated affine motion model of the encompassing planar region .
we then combine the ppg signal from all the tracked roi using the mrc algorithm .
after every @xmath112 seconds , we reset the klt tracker and reinitialize the deformable face model .
all the steps for region based motion tracking algorithm discussed above are summarized in algorithm [ motion ] .
@xmath132 lp @xmath133 deformableface(@xmath31 ) @xmath134 defineplanerregion(lp ) @xmath135 defineppgregion(@xmath136 ) @xmath137 goodfeatures(@xmath138 ) @xmath139 trackerklt(@xmath31,@xmath140,@xmath141 ) @xmath142 affineransac(@xmath143 ) @xmath144
@xmath145 @xmath146 @xmath147 * return * @xmath148 the rois inside the face region , @xmath88 , are chosen to be of the size of @xmath50 pixel block .
we have experimented with other sizes such as @xmath149 , @xmath150 px , but have found @xmath50 to be the best choice in terms of overall performance for our data - set .
note that the best choice of roi size may depend on camera resolution and distance of subject from the camera ( @xmath40 m for our data collection ) .
we reinitialize the deformable face model and restart the klt feature tracker every @xmath151 sec .
klt feature tracker is restarted because the local feature based motion tracker accumulates error over time .
we have found that restarting klt feature tracker every @xmath151 sec is sufficient for faithful facial region tracking during motions like reading , watching video or talking . selecting a larger time interval @xmath112 helps in improving accuracy of goodness
metric estimate , and smaller time interval helps in improving tracking performance .
we balance this tradeoff by selecting @xmath152 sec when the person is relatively static , and choosing @xmath153 sec when there is motion . at the end of every @xmath151 sec ( referred as an epoch from now on ) , we combine the ppg signal from all the tracked roi using the mrc algorithm . within each epoch , we drop those klt features that show more than @xmath154 px of forward - backward error .
further , we also reject those rigid regions @xmath129 where there are fewer than @xmath109 klt points left to track , as these regions might be undergoing excessive motion or large illumination changes , and including them will deteriorate the overall ppg estimate . for the ransac algorithm , we set the tolerance @xmath155 pixels , inlier fraction @xmath156 , and maximum number of iteration @xmath157 .
if a rigid affine model can not be found by ransac within this tolerance limit , we reject that region .
next , within each epoch , we filter the ppg signal obtained from well tracked roi ( @xmath88 ) using a zero - phase non - causal forward - backward bandpass filter , @xmath46 $ ] .
we then reject those roi where the signal amplitude crosses the amplitude threshold @xmath158 .
camera - based ppg signals are really weak , and hence they hardly cross @xmath159 value in the units of @xmath160-bit camera pixel intensity . any large change in intensity is mostly due to change in illumination or large motion artifact as discussed earlier .
we then compute a coarse estimate of the pulse rate @xmath161 , by combining the ppg signals from all the remaining rois .
this coarse pulse rate estimate can be erroneous at times , and so we keep track of the history of pulse rates over last @xmath128 epochs .
if the current estimate of @xmath161 is off by more than @xmath162bpm , then we replace the current estimate with the median of the last four estimate .
we then compute the goodness metric @xmath76 for all the remaining roi ( all regions rejected prior to this stage are given a weight @xmath163 for the current epoch ) and combine the ppg signals using the weighted average ( equation ) .
we recompute the goodness metric @xmath76 for each roi after every epoch .
for comparison , we implemented known past methods for camera - based ppg estimation .
the general steps taken are : ( i ) select the face regions using voila jones face detector as described in @xcite , ( ii ) extract ppg signal by first computing the spatial average of the pixel intensity within the selected face region , and then filtering ( detrending ) the estimate @xcite , ( iii ) to compensate for motion , compute the 2d shift in face between consecutive frames as described in @xcite and extracted the ppg from the tracked face region . in this sense
, we have used a combination of known methods for single channel ( green ) camera - based ppg estimation algorithm for comparison with our distanceppg algorithm .
we label this combination as _ face averaging method _ from now onwards .
another set of work in camera - based ppg estimation involves decomposing different camera channels ( e.g. red , green , blue ) into independent source signals using independent component analysis ( ica ) , and extracting the desired ppg signal from one of these independent sources @xcite .
our proposed distanceppg algorithm provides improvement in camera - based ppg estimate by spatially combining ppg signals from different regions of the face , and by improving upon the tracking algorithm , and uses only a single channel ( green ) of the camera . on the other hand ,
ica based methods utilizes multiple camera channels ( e.g. red , green , and blue @xcite or cyan , orange , and green @xcite ) to improve the performance of camera - based ppg estimate by separating independent sources .
thus , these two methods are characteristically different , and we will summarize the performance improvement provided by ica - based method and by distanceppg . for all single channel video recording in this study , we used flea 3^^ usb 3.0 fl - u3 - 13e4m - c monochrome camera operated at @xmath22 frames per second , with a resolution of @xmath164 , and @xmath160 bits per pixel .
we added a @xmath165 nm full - width half - max ( fwhm ) green filter ( fb550 - 40 from thor labs
^^ ) in front of the monochrome camera .
we selected green filter since the absorption spectra of hb and hbo2 peaks in this wavelength region .
moreover , all commercial color camera have highest number of pixel having green filter ( bayer pattern ) . for color video recording
, we used flea 3^^ usb 3.0 fl - u3 - 13e4c - c color camera operated at @xmath22 frames per second with rgb bayer pattern having a total resolution of @xmath164 , and @xmath160 bits per pixel .
we used texas instruments afe4490spo2evm pulse oximeter module to record contact - based ppg signal for comparison .
it operates at a sampling rate of @xmath166 hz .
the distance between camera and subject was @xmath40 m .
both the camera system and the pulse - oximeter were started simultaneously , and the data is recorded in a pc workstation . all processing is done using a custom built software written in matlab . all experiments ( except the lighting experiment ) is conducted under ambient ( fluorescent ) lighting at 500 lux brightness .
the main goal of the experiments reported here is to characterize and quantify the performance of the two components of our proposed distanceppg algorithm mrc algorithm and region based motion tracking algorithm , and compare them with face averaging method .
we evaluate the performance by varying three main parameters of interest : ( i ) skin tone of people , ( ii ) motion , ( iii ) ambient light intensity , and quantify their effect on ppg estimation accuracy .
we use the same raw video feed to evaluate our distanceppg algorithm in comparison to face averaging method , and so performance improvements reported here are due to the proposed algorithm , and is independent of any specific hardware or camera choices we made for the evaluation .
all the experiments done in this research were approved by the rice university institute review board ( protocol number : @xmath167e , approval date : @xmath168 ) .
for the first experiment , we collected single channel ( green ) video data from @xmath10 subjects ( @xmath169 male , @xmath14 female ) with different skin tones ( from light , pale white to dark brown / black ) . for this experiment ,
subjects were asked to face the camera and be static for a duration of @xmath165 seconds ( involuntary motions were not restricted ) . for the second set of experiments
, we collected ( single green channel ) video under three natural motion scenarios ( i ) reading on computer , ( ii ) watching video , ( iii ) talking .
the motion scenarios are representative of a general class of motion exhibit by users of tablet , phone , or laptops while facing the screen .
reading scenario involves lateral movement of the head while reading text on screen .
watching video also involves intermittent facial expression such as smiling , getting amazed , sad etc , apart from lateral movement of the head .
talking scenario involves a lot of non - rigid movement around the jaw and the cheeks , and thus are ideal to evaluate system perform in harsh scenarios . for each of the motion scenarios , we collected @xmath108 seconds video recordings for @xmath14 subjects having different skin tones , along with their stationary video recording for baseline comparison .
for the third set of experiments , we varied the illumination level from @xmath170 lux upto @xmath171 lux ( ambient light is around @xmath172 lux ) and recorded the single channel ( green ) video for a duration of @xmath165 seconds under each lighting conditions for two subjects having pale - white and brown skin tones . for comparison with ica - based method , we collected two sets of data : ( i ) static dataset comprising of color video ( red , green , blue ) of @xmath128 subjects of varying skin tones ( only non - caucasian ) for a duration of @xmath173 seconds at @xmath166 lux illumination , ( ii ) talking dataset comprising of color video of @xmath128 subjects of varying skin tone ( non - caucasian ) , with @xmath131 subjects having @xmath166 lux illumination , and @xmath174 subject having @xmath175 lux illumination .
this set is deliberately chosen to be extremely harsh ( lower light , non - caucasian skin tones , and large motion during talking ) to highlight scenario where current known algorithms ( including distanceppg ) fails , and provide dataset where future algorithms can improve .
the dataset will be released in public and other researchers can access it at http://www.ece.rice.edu/\texttildelow mk28/distanceppg / dataset/[http://www.ece.rice.edu/\texttildelow mk28/distanceppg / dataset/ ] . as a waveform estimation algorithm ,
distanceppg provides an estimate of the ground truth ppg waveform @xmath30 using video of a person s face .
thus , we use signal - to - noise ratio ( snr ) of the estimate @xmath104 to quantify performance for comparison .
the ground truth signal , @xmath30 , is recorded using a pulse oximeter attached to subject s ear .
we chose earlobe instead of finger probe because of its proximity to the face region .
further , we also evaluate the performance based on the accuracy of physiological parameter like pulse rate and pulse rate variability ( i.e. beat - to - beat changes in pulse interval ) that can be extracted from ppg waveform .
we report the mean error rates in estimating pr and prv using our distanceppg algorithm and using face averaging method under different experimental scenarios .
to define the snr of the estimated ppg waveform , we used standard ppg signal obtained from pulse oximeter connected to a person s earlobe as our ground truth signal .
the ppg waveform obtained from a contact pulse oximeter would also exhibit error due to motion and ambient light artifacts , and in that sense our choice of ground truth is a best effort choice .
nonetheless , the noise present in ppg waveform acquired using a contact pulse oximeter is orders of magnitude smaller than that obtained using a camera - based system , so our estimate of snr would still be reasonably accurate .
the amplitude of the ppg signal recorded by pulse oximeter is unrelated to the amplitude of the camera - based ppg signal as both systems have completely independent sensor architecture and analog gain stage .
so , here we have developed a definition of snr which is independent of exact amplitude of these waveform .
our snr metric captures how similar camera - based ppg signal is to the pulse oximeter based ground - truth ppg signal .
let @xmath176 denote the ppg acquired using pulse oximeter .
let @xmath177 denote the ppg estimated from a camera - based system ( @xmath178 from previous section ) .
@xmath179 here , @xmath180 is the noise in the ppg signal acquired from camera .
apart from the quantization noise , @xmath180 also includes uncompensated motion artifacts .
the noise present in the pulse oximeter system is denoted as @xmath181 .
let us assume that all signals are defined in the time interval @xmath119 $ ] and we use @xmath182 inner product and norms for all definitions . the noise @xmath180 would be uncorrelated to @xmath181 as both acquisition systems are unrelated , and both noise is uncorrelated to the underlying zero mean ppg signal @xmath30 .
if we assume that all signals follow ergodicity and consider integration over large time window ( @xmath112 is large ) , then @xmath183 and further , we can write @xmath184 since the signal quality of ppg derived from pulse oximeter is reasonably good , we can assume that noise power @xmath185 is much smaller than signal power @xmath186 in the denominator .
thus , @xmath187 can be approximated as @xmath188 further , if we assume that @xmath189 , we can estimate the noise present in @xmath177 as @xmath190 and , underlying signal @xmath191 can be approximated to @xmath192 which leads us to define signal to noise ration or snr simply as @xmath193 the snr measure defined in ( [ eq : snr ] ) will be used in the following sections to compare performance of distanceppg algorithm and face averaging method under various experimental scenarios . to estimate the pulse rate ,
we compute the spectrum of the ppg signal using fft algorithm over every @xmath109 sec window ( hamming window ) with @xmath14 sec overlap .
the pulse rate is estimated as the frequency that correspond to the highest power in the estimated spectrum ( pr @xmath194 bpm ) .
similarly , we obtain the reference pulse rate from synchronously acquired contact ppg .
to estimate the prv or inter - beat interval ( ibi ) , we first interpolate the camera - based ppg signal with a cubic spline function to a sampling rate of @xmath166 hz ( same as the sampling rate for pulse oximeter ) .
we then detect the peaks in the interpolated ppg signal using a custom algorithm for minima - detection . to avoid inclusion of artifacts due to noise or motion
, we use the estimated pulse rate to reject peaks within @xmath195 time difference . the time interval between consecutive peaks
are the prv ( or ibi ) .
same algorithm is used to find the reference prv from synchronously acquired contact ppg . due to the noisy estimate of ppg signal for darker skin tones and/or under large motion ( e.g. during talking ) ,
some pulse beats ( peaks ) can not be detected in camera ppg signal .
we determine the number of missing peaks by comparing peak location in camera ppg with the reference pulse oximeter ppg .
thus , we also report the percentage of missing peaks in the camera - based ppg signal as a performance metric associated with prv estimation , apart from the rmse of prv .
0.48 table[x = ids , y = old_snr ] result / snr_static.txt ; table[x = ids , y = mrc_snr ] result / snr_static.txt ; 0.48 table[x = time , y = old]result / raw_static_fair.txt ; [ name = mrc , at=(@xmath196 ) , width=2.5 in , height=1 in , anchor = north , xmin=-1 , xmax=40 , xtick= , ymin=-1.5 , ymax=1.5 , ytick=-1,1 , ylabel = camera ppg scale , axis x line = none , axis y line = left , y axis line style=- , ] table[x = time , y = mrc]result / raw_static_fair.txt ; [ name = signal , at=(@xmath197 ) , width=2.5 in , height=1 in , anchor = north , xmin=-1 , xmax=40 , axis x line = bottom , axis y line = left , y axis line style=- , ymin=-1.5 , ymax=1.5 , ytick=-1,1 , xlabel = time(s ) ] table[x = time , y = signal]result / raw_static_fair.txt ; 0.48 table[x = time , y = old]result / raw_static_medium.txt ; coordinates(-1,0 ) ; coordinates(-1,0 ) ; [ name = mrc , at=(@xmath196 ) , width=2.5 in , height=1 in , anchor = north , xmin=-1 , xmax=40 , xtick= , ymin=-1.5 , ymax=1.5 , ytick=-1,1 , ylabel = camera ppg scale , axis x line = none , axis y line = left , y axis line style=- , ] table[x = time , y = mrc]result / raw_static_medium.txt ; [ name = signal , at=(@xmath197 ) , width=2.5 in , height=1 in , anchor = north , xmin=-1 , xmax=40 , axis x line = bottom , axis y line = left , y axis line style=- , ymin=-1.5 , ymax=1.5 , ytick=-1,1 , xlabel = time(s ) , ] table[x = time , y = signal]result / raw_static_medium.txt ; 0.48 table[x = time , y = old]result / raw_static_brown.txt ; [ name = mrc , at=(@xmath196 ) , width=2.5 in , height=1 in , anchor = north , xmin=-1 , xmax=40 , xtick= , ymin=-1.5 , ymax=1.5 , ytick=-1,1 , axis x line = none , axis y line = left , y axis line style=- , ] table[x = time , y = mrc]result / raw_static_brown.txt ; [ name = signal , at=(@xmath197 ) , width=2.5 in , height=1 in , anchor = north , xmin=-1 , xmax=40 , axis x line = bottom , axis y line = left , y axis line style=- , ymin=-1.5 , ymax=1.5 , ytick=-1,1 , xlabel = time(s ) ] table[x = time , y = signal]result / raw_static_brown.txt ; figure [ fig : static_snr_waveform ] compares the snr and the waveform of ppg obtained using distanceppg and face averaging method for people having different skin tone categories ( i ) light / fair ( @xmath128 people ) , ( ii ) medium / olive ( @xmath128 people ) , and ( iii ) brown / dark ( @xmath128 people ) .
as seen in figure [ fig : static_snr ] , on an average distanceppg provides @xmath0 db of snr improvement for all skin tones compared to face averaging method . also , the gain in snr is more for darker skin tones ( around @xmath1 db ) .
further , figure [ fig : static_waveform_fair]-[fig : static_waveform_brown ] shows an example @xmath165 s plots of camera - based ppg estimate using both distanceppg and face averaging method for the three skin tone categories .
when compared with ground truth signal , it is evident that camera - based ppg signal estimated using distanceppg is more similar to the ground truth pulse oximeter signal .
0.48 table[x = mean , y = diff]result / bland_static_data_max_fair.txt ; ; table[x = mean , y = diff]result / bland_static_data_max_medium.txt ; ; table[x = mean , y = diff]result / bland_static_data_max_brown.txt ; ; coordinates ( 40,-0.02 ) ( 110,-0.02 ) ; coordinates ( 40,-0.75 ) ( 110,-0.75 ) ; at ( axis cs : 65,-1.5 ) mean @xmath198sd ; coordinates ( 40,0.72 ) ( 110,0.72 ) ; at ( axis cs : 65,+1.5 ) mean @xmath199sd ; 0.48 table[x = mean , y = diff]result / bland_static_data_old_fair.txt ; table[x = mean , y = diff]result / bland_static_data_old_medium.txt ; table[x = mean , y = diff]result / bland_static_data_old_brown.txt ; coordinates ( 40,-0.40 ) ( 110,-0.40 ) ; coordinates ( 40,-4.5 ) ( 110,-4.5 ) ; at ( axis cs : 65,-5.5 ) mean @xmath198sd ; coordinates ( 40,3.7 ) ( 110,3.7 ) ; at ( axis cs : 65,+5.5 ) mean @xmath199sd ; the improvement in snr for various skin tones reduces the error in pr estimation .
figure [ fig : stationary_bland ] shows the agreement between camera derived pr and ground truth pulse oximeter derived pr from @xmath10 subjects of different skin tones using bland - altman plot . for ppg
estimated using face averaging method , the mean bias ( average difference between ground truth pulse oximeter derived pr and camera - based ppg derived pr ) @xmath4 = @xmath11 bpm with @xmath6 limit of agreement ( mean bias @xmath7 standard deviation of the difference ) being @xmath12 to @xmath13 bpm .
using the distanceppg reduces the error in pr estimate to @xmath200 bpm with @xmath6 limit of agreement being @xmath8 to @xmath9 bpm . for computing these statistics ,
we have not included occasional extreme outliers ( error in pr @xmath201 bpm ) found in face averaging method ( @xmath128 outliers in @xmath202 sample points ) as these outliers significantly skew the performance of face averaging method and show it in bad light .
distanceppg algorithm did not exhibit any extreme outliers in the dataset for all skin tones and so we have reported statistics for all @xmath202 sample points .
0.48 table[x = ids , y = old ] result / ibi_rmse_static.txt ; table[x = ids , y = mrv ] result / ibi_rmse_static.txt ; 0.48 table[x = ids , y = mrv ] result / missing_static.txt ; table[x = ids , y = old ] result / missing_static.txt ; ppg signal estimated using distanceppg shows significant improvement in the root mean squared error ( rmse ) of prv estimate .
figure [ fig : stationary_ibi ] shows the rmse of prv estimation and the corresponding percentage of missing peaks in camera - based ppg for the three skin tone categories . using distanceppg , the rmse in prv for light and medium skin tone is less than @xmath21 ms . as @xmath22 fps camera is used for recording the video , one can not expect rmse to be much lower than half the sampling interval ( @xmath203 ms ) . 0.48 table[x = ids , y = old_snr ] result / snr_motion.txt ; table[x = ids , y = mrc_snr ] result / snr_motion.txt ; 0.48 table[x = time , y = old]result / raw_motion_read.txt ; [ name = mrc , at=(@xmath196 ) , width=2.5 in , height=1 in , anchor = north , xmin=-1 , xmax=40 , ylabel = camera ppg scale , xtick= , ymin=-1.5 , ymax=1.5 , ytick=-1,1 , axis x line = none , axis y line = left , y axis line style=- , ] table[x = time , y = mrc]result / raw_motion_read.txt ; [ name = signal , at=(@xmath197 ) , width=2.5 in , height=1 in , anchor = north , xmin=-1 , xmax=40 , axis x line = bottom , axis y line = left , y axis line style=- , xtick= , ymin=-1.5 , ymax=1.5 , ytick=-1,1 , ] table[x = time , y = signal]result / raw_motion_read.txt ; [ name = motion , at=(@xmath204 ) , width=2.5 in , height=1.0 in , anchor = north , xmin=-1 , xmax=40 , ylabel = motion ( px ) , axis x line = bottom , axis y line = left , y axis line style=- , xlabel = time ( sec ) , ymin=0 , ymax=10 , ytick=0,5 , ] + [ color = black , const plot mark mid , mark size=0.5pt ] table[x = time , y = motion]result / motion_read.txt ; 0.48 table[x = time , y = old]result / raw_motion_watch.txt ; coordinates(-1,0 ) ; coordinates(-1,0 ) ; + [ color = black , const plot mark mid , mark size=0.5pt ] coordinates(-1,0 ) ; [ name = mrc , at=(@xmath196 ) , width=2.5 in , height=1 in , anchor = north , xmin=-1 , xmax=40 , xtick= , ymin=-1.5 , ylabel = camera ppg scale , ymax=1.5 , ytick=-1,1 , axis x line = none , axis y line = left , y axis line style=- , ] table[x = time , y = mrc]result / raw_motion_watch.txt ; [ name = signal , at=(@xmath197 ) , width=2.5 in , height=1 in , anchor = north , xmin=-1 , xmax=40 , axis x line = bottom , axis y line = left , y axis line style=- , xtick= , ymin=-1.5 , ymax=1.5 , ytick=-1,1 , ] table[x = time , y = signal]result / raw_motion_watch.txt ; [ name = motion , at=(@xmath204 ) , width=2.5 in , height=1.0 in , anchor = north , xmin=-1 , xmax=40 , ylabel = motion ( px ) , axis x line = bottom , axis y line = left , y axis line style=- , xlabel = time ( sec ) , ymin=0 , ymax=12 , ytick=0,5 , ] + [ color = black , const plot mark mid , mark size=0.5pt ] table[x = time , y = motion]result / motion_watch.txt ; 0.48 table[x = time , y = old]result / raw_motion_talk.txt ; [ name = mrc , at=(@xmath196 ) , width=2.5 in , height=1 in , anchor = north , xmin=-1 , xmax=40 , xtick= , ymin=-1.5 , ymax=1.5 , ytick=-1,1 , axis x line = none , axis y line = left , y axis line style=- , ] table[x = time , y = mrc]result / raw_motion_talk.txt ; [ name = signal , at=(@xmath197 ) , width=2.5 in , height=1 in , anchor = north , xmin=-1 , xmax=40 , axis x line = bottom , axis y line = left , y axis line style=- , xtick= , ymin=-1.5 , ymax=1.5 , ytick=-1,1 , ] table[x = time , y = signal]result / raw_motion_talk.txt ; [ name = motion , at=(@xmath204 ) , width=2.5 in , height=1.0 in , anchor = north , xmin=-1 , xmax=40 , axis x line = bottom , axis y line = left , y axis line style=- , xlabel = time ( sec ) , ymin=0 , ymax=8 , ytick=0,5 , ] + [ color = black , const plot mark mid , mark size=0.5pt ] table[x = time , y = motion]result / motion_talk.txt ; [ fig : motion_improvement ] figure [ fig : motion_snr ] shows the snr improvements for four scenarios ( i ) stationary , ( ii ) reading , ( iii ) watching video , and ( iv ) talking for @xmath14 subjects of different skin tones .
distanceppg provides on an average @xmath2db of snr improvement under various motion scenario compared to face averaging method .
further , figure [ fig : motion_waveform_reading]-[fig : motion_waveform_talking ] shows typical @xmath165 s plots of camera - based ppg estimate using both distanceppg and face averaging method for the three motion scenarios . when compared with ground truth signal ,
it is evident that distanceppg improves the estimate of ppg during small - medium motion ( @xmath205 px per frame ) . under large motion
( @xmath206 px per frame ) , distanceppg also suffers due to motion artifacts . 0.48 table[x = mean , y = diff]result / bland_motion_data_max_read.txt
; ; table[x = mean , y = diff]result / bland_motion_data_max_watch.txt ; ; table[x = mean , y = diff]result / bland_motion_data_max_talk.txt ; ; coordinates ( 40,0.48 ) ( 110,0.48 ) ; coordinates ( 40,-5.73 ) ( 110,-5.73 ) ; at ( axis cs : 95,-20 ) mean @xmath198sd ; coordinates ( 40,6.70 ) ( 110,6.70 ) ; at ( axis cs : 95,20 ) mean @xmath199sd ; 0.48 table[x = mean , y = diff]result / bland_motion_data_old_read.txt ; table[x = mean , y = diff]result / bland_motion_data_old_watch.txt ; table[x = mean , y = diff]result / bland_motion_data_old_talk.txt ; coordinates ( 40,7.17 ) ( 110,7.17 ) ; coordinates ( 40,-18.70 ) ( 110,-18.70 ) ; at ( axis cs : 95,-25.5 ) mean @xmath198sd ; coordinates ( 40,33.04 ) ( 110,33.04 ) ; at ( axis cs : 95,+40 ) mean @xmath199sd ; the improvement in snr under various motion scenarios reduces the error in pr estimation .
figure [ fig : motion_bland ] shows the agreement between camera derived pr and ground truth pulse oximeter derived pr from @xmath14 subjects under the three motion scenario ( reading , watching , talking ) using bland - altman plot . for ppg
estimated using face averaging method , the mean bias @xmath18 bpm with @xmath6 limit of agreement between @xmath19 to @xmath20 bpm . using
the mrc reduces the error in pr estimate to @xmath207 bpm with @xmath6 limit of agreement between @xmath16 to @xmath17 bpm . note ,
distanceppg do not perform well for pr estimation under talking scenario possibly because of large non - rigid motion . for non - talking motion scenario ( reading+watching ) ,
the error in pr estimate is @xmath208 bpm with @xmath6 limit of agreement being @xmath209 to @xmath210 bpm which is much better .
thus , one can possibly use computer vision based activity recognition to not estimate pr during large non - rigid motion .
0.48 table[x = ids , y = old ] result / ibi_rmse_motion.txt ; table[x = ids , y = mrv ] result / ibi_rmse_motion.txt ; 0.48 table[x = ids , y = mrv ] result / missing_motion.txt ; table[x = ids , y = old ] result / missing_motion.txt ; figure [ fig : motion_ibi ] shows the rmse in prv estimation and the corresponding percentage of missing peaks in camera - based ppg for the three motion scenario along with stationary base case .
distanceppg provides reduction in rmse of prv ( by around 1.5x ) and the percentage of missing peaks ( by atleast by 2x ) when compared to face averaging method under motion scenarios .
still , there is scope for further improvement as the rmse in prv estimate is around @xmath211 ms , and one can achieve around @xmath203 ms rmse with the @xmath22 fps camera .
these errors are due to large non - rigid motion of the face during smiling , talking etc .
figure [ fig : snr_light ] shows the snr of camera - based ppg estimate under different lighting condition ( @xmath212 lux ) using both distanceppg ( marked as squares ) and face averaging method ( marked as circles ) for two subjects having pale white ( dashed line ) and brown skin ( solid line ) tones . on an average distanceppg provides an snr gain of @xmath213 db for pale white person , and an snr gain of @xmath3 db for a brown skin tone person . though we tried to keep all other experimental parameter ( particularly motion ) invariant ( static in this case ) while varying the light intensity , one should not read much from the ups and downs in the plot , as they mostly are due to uncontrolled experimental variations .
nonetheless , the general trend is that snr of camera - based ppg estimate increases as we increase the light intensity .
this trend is more evident for distanceppg compared to face averaging methodwhich shows that distanceppg is more immune to such experimental variations .
table [ x = light , y = snr_mrc_u1]result / snr_light.txt ; table [ x = light , y = snr_old_u1]result / snr_light.txt ; table [ x = light , y = snr_mrc_u2]result / snr_light.txt ; table [ x = light , y = snr_old_u2]result / snr_light.txt ; for static scenario ( @xmath128 subjects having non - caucasian skin tones ) , the mean snr of camera - based ppg using distanceppg is @xmath214 db , and using ica - based method is @xmath215 db . the mean error in pr estimation is @xmath216 bpm and rmse of prv estimation is @xmath217 ms for distanceppg , and corresponding errors for ica - based method is @xmath218 bpm and @xmath219 ms respectively .
thus , the small improvement in snr ( @xmath174 db ) by using distanceppg results in better performance in prv estimation accuracy . for talking scenario ( @xmath131 subjects under @xmath166 lux illumination , @xmath174 subject under @xmath175 lux ) , both distanceppg and ica - based method performs poorly , possibly because of large motion artifact which renders small ppg signal estimation difficult .
the mean snr of camera - based ppg are @xmath220 db and @xmath221 db for distanceppg and ica - based method respectively . due to such low snr ,
the mean error in pulse rate suffers for both methods , and is @xmath222 bpm and @xmath223 bpm for distanceppg and ica - based method respectively .
we used goodness metric as a substitute for the signal quality or snr of the ppg signal obtained from different regions of the face . to understand how good our goodness metric substitute is
, we compute scatter plot between goodness metric @xmath76 in db and snr of the ppg signal @xmath97 obtained from different roi @xmath88 inside the face .
figure [ fig : goodnesssnr ] shows this scatter plot for two subjects having white / fair and brown / dark skin tones .
0.45 table[x = snr , y = goodness]result / scattersnrgoodnessuid9.txt ; table[row sep = crcr ] -80 -3 + 15 -3 + ; 0.45 table[x = snr , y = goodness]result / scattersnrgoodnessuid12.txt ; table[row sep = crcr ] -80 -3 + 15 -3 + ; the scatter plot shows that goodness metric is a good substitute for snr in regions having goodness metric greater than @xmath224 db . for regions having lower
goodness metric , goodness metric overestimates the snr .
thus , we also reject regions having very low goodness metric score ( @xmath225 db or @xmath226 ) in the ppg estimation equation . completely rejecting regions having very small goodness metric value
@xmath226 slightly improve the overall snr of camera - based ppg signal ( by @xmath227 db ) .
our definition of goodness metric exploited a known prior structure in photoplethysmogram signals , i.e. ppg signals are periodic with dominant frequency equal to the pulse rate , and developed a signal quality index based on the spectrum of the ppg signal extracted from a region .
this signal quality index helped us propose weighted averaging algorithm to improve the snr of the overall ppg estimate .
thus , we broke down the problem of robust camera - based ppg waveform estimation into two stages : ( i ) estimation of dominant frequency ( pr ) , which is a relatively easy problem , and ( ii ) robust estimation of ppg waveform by defining a new signal quality index based on the pulse rate .
it is important to note that merely filtering the ppg signal in a narrow frequency band around the pulse rate ( @xmath121 $ ] hz ) , a method generally used for weak periodic signal detection , would not give a good estimate of the ppg waveform , as the spectral band of interest for ppg signal is much wider ( @xmath46 $ ] ) . in other words , we care about the exact shape of the ppg waveform ( e.g. inter - beat - interval information , systolic and diastolic peaks etc ) , and thus we should design an estimation algorithm which preserves the shape .
one may find similar opportunities and concerns in many other periodic biological signal estimation problems e.g. electrocardiogram ( ekg ) , event related potentials ( erp ) , etc .
goodness metric based signal quality index developed in this paper can be expanded for these scenarios as well , particularly when one records these biological signals using sensors distributed in space and wants to develop a technique to weigh the sensor based on their signal quality .
the goodness metric profile over a face depends on two factors - ( i ) spatial illumination profile over the face @xmath38 . and
( ii ) strength of modulation ( @xmath93 ) of reflected light which depends on the blood perfusion profile . as the illumination profile can change over time
, distanceppg re - estimates the goodness metric every @xmath112 seconds .
figure [ fig : mrc_algorithm ] of goodness metric overlay clearly shows that forehead region is best for extracting ppg signal from the face .
our finding is in agreement with the findings of other researchers that forehead represent a suitable site for camera - based ppg estimation .
this is because the forehead region have much better blood perfusion compared to other regions @xcite .
further , it might be feasible to develop a new metric for the blood perfusion profile over the face if one can separately estimate the spatial illumination profile and compensate for it .
blood flow in peripheral tissue is generally estimated using either laser doppler flowmetry or using laser speckle contrast imaging ( lsci ) techniques .
camera - based blood flow monitoring , if developed , will be highly useful .
the rmse of prv that we could achieve for fairer skin tones under stationary scenario is around @xmath228 ms .
it is difficult to achieve something significantly below this error value using a @xmath22 fps camera - based ppg signal .
this is because the time difference between two camera frames is @xmath229 ms , and thus there is an ambiguity of @xmath230 ms in peak detection .
no amount of interpolation ( we interpolated camera ppg to @xmath166 hz using spline interpolation ) can completely recover the higher frequency details lost because of the low sampling rate .
one solution to further reduce the rmse of prv is to use higher frame rate camera ( e.g. , @xmath231 fps ) . since maximum achievable exposure time of higher frame rate cameras
would be smaller ( e.g. @xmath109 ms ) , this can reduce the amount of light entering the camera sensor , and thus the snr of the estimated camera ppg signal as well .
this will make faithful pulse minima detection , an essential step for prv estimation , difficult .
the algorithmic snr gain achieved due to distanceppg can be useful here , as one can compensate the snr loss because of higher frame rate with the algorithmic snr gain due to distanceppg .
pulse rate variability ( prv ) is one of the most commonly used measure in psychophysiology research , and is employed broadly as a determinant of the status of the autonomic nervous system ( ans ) @xcite .
camera - based prv estimation can enable researchers to conduct large scale psychophysiological studies of statistical significance outside their lab settings .
one needs to evaluate and understand the impact of error in camera - based prv on various metric of interest in psychophysiology such as high frequency ( hf ) power , low frequency ( lf ) power , lf / hf power ratio of prv , standard deviation of all normal heart period ( sdnn ) etc .
this will be one of the future direction of our research .
in this paper , we dived deeper into understanding the challenges involved in estimating photoplethysmogram waveform using a camera - based system .
we developed a new method , distanceppg , that addresses these challenges .
we evaluated our algorithm on people having diverse skin tones ( pale / white to brown / dark ) , under various lighting conditions ( @xmath170 lux to @xmath171 lux ) and natural motion scenarios , and showed significant improvement in the accuracy of vital sign ( pulse rate , pulse rate variability ) estimation .
our major contribution is to develop a formal method to take into consideration differences in blood perfusion and incident light intensity in different regions of the face to improve the accuracy of vital sign estimation under difficult scenario ( e.g. dark skin tones and/or low lighting condition ) .
further , we proposed a new motion - tracking algorithm which tracks different regions of the face separately , to improve performance of vital sign estimation under natural motion scenarios like reading content on computer screen , watching video , and talking over skype .
we also highlighted limitations of distanceppg algorithm , e.g. during talking scenario , where vital sign estimation accuracy suffers .
the major challenge in such scenario is significant variations in skin surface reflectance component during large motion which falls in the frequency band of interest of ppg signal ( @xmath46 $ ] ) .
the small changes in skin subsurface reflectance , which encodes the ppg signal , is not recoverable in presence of large in - band surface reflectance changes .
thus , the general approach we adopted in distanceppg is to reject regions undergoing large surface reflectance changes .
consequently , during large motion we end up rejecting a majority of face region and thus our estimate of ppg signal was inaccurate .
two popular apps for measuring non - contact pulse rate using the color change ( or ppg ) signal from a person s face are philips vital sign camera @xcite and cardiio @xcite .
both these apps require users to be at rest facing the camera in a well lit environment to be effective @xcite .
the distanceppg algorithm discussed in this paper address these challenges and thus would extend the use case of mobile phone and computer apps for vital sign monitoring .
we are in the process of developing a realtime pc - based application to robustly estimate ppg signal using a webcam .
our future work includes porting our code to popular mobile platforms ( android / ios ) , and further improving the performance of distanceppg under motion scenarios .
this work was partially supported by nsf cns 1126478 , nsf iis-1116718 , rice university graduate fellowship , texas instruments fellowship , and texas higher education coordinating board : thecb - nharp 13308 | vital signs such as pulse rate and breathing rate are currently measured using contact probes .
but , non - contact methods for measuring vital signs are desirable both in hospital settings ( e.g. in nicu ) and for ubiquitous in - situ health tracking ( e.g. on mobile phone and computers with webcams ) .
recently , camera - based non - contact vital sign monitoring have been shown to be feasible .
however , camera - based vital sign monitoring is challenging for people with darker skin tone , under low lighting conditions , and/or during movement of an individual in front of the camera . in this paper , we propose distanceppg , a new camera - based vital sign estimation algorithm which addresses these challenges .
distanceppg proposes a new method of combining skin - color change signals from different tracked regions of the face using a weighted average , where the weights depend on the blood perfusion and incident light intensity in the region , to improve the signal - to - noise ratio ( snr ) of camera - based estimate .
one of our key contributions is a new automatic method for determining the weights based only on the video recording of the subject .
the gains in snr of camera - based ppg estimated using distanceppg translate into reduction of the error in vital sign estimation , and thus expand the scope of camera - based vital sign monitoring to potentially challenging scenarios .
further , a dataset will be released , comprising of synchronized video recordings of face and pulse oximeter based ground truth recordings from the earlobe for people with different skin tones , under different lighting conditions and for various motion scenarios .
10 w. verkruysse , l. o. svaasand , and j. s. nelson , `` remote plethysmographic imaging using ambient light , '' optics express * 16 * , 2143421445 ( 2008 ) .
m .- z .
poh , d. mcduff , and r. picard , `` advancements in noncontact , multiparameter physiological measurements using a webcam , '' ieee transactions on biomedical engineering * 58 * , 711 ( 2011 ) .
y. sun , s. hu , v. azorin - peris , s. greenwald , j. chambers , and y. zhu , `` motion - compensated noncontact imaging photoplethysmography to monitor cardiorespiratory status during exercise , '' journal of biomedical optics * 16 * , 077010 ( 2011 ) . l. a. m. aarts , v. jeanne , j. p. cleary , c. lieber , j. s. nelson , s. bambang oetomo , and w. verkruysse , `` non - contact heart rate monitoring utilizing camera photoplethysmography in the neonatal intensive care unit - a pilot study , '' early human development * 89 * , 943948 ( 2013 ) .
j. m. saragih , s. lucey , and j. f. cohn , `` deformable model fitting by regularized landmark mean - shift , '' international journal of computer vision * 91 * , 200215 ( 2011 ) .
b. d. lucas and t. kanade , `` an iterative image registration technique with an application to stereo vision , '' ( 1981 ) , pp .
674679 . c. tomasi and t. kanade ,
`` detection and tracking of point features , '' technical report mu - cs-91 - 132 , carnegie mellon university ( 1991 ) .
f. p. wieringa , f. mastik , and a. f. w. van der steen , `` contactless multiple wavelength photoplethysmographic imaging : a first step toward `` spo2 camera '' technology , '' annals of biomedical engineering * 33 * , 10341041 ( 2005 ) .
k. humphreys , t. ward , and c. markham , `` noncontact simultaneous dual wavelength photoplethysmography : a further step toward noncontact pulse oximetry , '' the review of scientific instruments * 78 * , 044304 ( 2007 ) .
m .- z .
poh , d. j. mcduff , and r. w. picard , `` non - contact , automated cardiac pulse measurements using video imaging and blind source separation , '' optics express * 18 * , 1076210774 ( 2010 ) .
b. d. holton , k. mannapperuma , p. j. lesniewski , and j. c. thomas , `` signal recovery in imaging photoplethysmography , '' physiological measurement * 34 * , 14991511 ( 2013 ) .
d. mcduff , s. gontarek , and r. picard , `` improvements in remote cardiopulmonary measurement using a five band digital camera , '' ieee transactions on biomedical engineering * 61 * , 25932601 ( 2014 ) .
j. allen , `` photoplethysmography and its application in clinical physiological measurement , '' physiological measurement * 28 * , r1 ( 2007 ) .
s. hu , v. azorin - peris , and j. zheng , `` opto - physiological modeling applied to photoplethysmographic cardiovascular assessment , '' journal of healthcare engineering * 4 * , 505528 ( 2013 ) .
d. brennan , `` linear diversity combining techniques , '' proceedings of the ieee * 91 * , 331356 ( 2003 ) .
m. nitzan , b. khanokh , and y. slovik , `` the difference in pulse transit time to the toe and finger measured by photoplethysmography , '' physiological measurement * 23 * , 8593 ( 2002 ) .
j. shi and c. tomasi , `` good features to track , '' in `` , 1994 ieee computer society conference on computer vision and pattern recognition , 1994 .
proceedings cvpr 94 , '' ( 1994 ) , pp .
593600 .
z. kalal , k. mikolajczyk , and j. matas , `` forward - backward error : automatic detection of tracking failures , '' in `` 2010 20th international conference on pattern recognition ( icpr ) , '' ( 2010 ) , pp .
27562759 .
m. a. fischler and r. c. bolles , `` random sample consensus : a paradigm for model fitting with applications to image analysis and automated cartography , '' commun .
acm * 24 * , 381395 ( 1981 ) .
m. fernandez , k. burns , b. calhoun , s. george , b. martin , and c. weaver , `` evaluation of a new pulse oximeter sensor , '' american journal of critical care : an official publication , american association of critical - care nurses * 16 * , 146152 ( 2007 ) .
j. a. j. heathers , `` smartphone - enabled pulse rate variability : an alternative methodology for the collection of heart rate variability in psychophysiological research , '' international journal of psychophysiology : official journal of the international organization of psychophysiology * 89 * , 297304 ( 2013 )
. philips , `` philips vital signs camera , '' http://www.vitalsignscamera.com/. cardiio , `` cardiio , '' http://www.cardiio.com/. |
the riemann - siegel @xmath7 function is defined by @xmath8 let @xmath9 be the approximate @xmath7 function where the @xmath10 function has been replaced with its stirling approximation . @xmath11 the @xmath6 function is not invertible but the inverse of its approximation @xmath12 is defined by a linear function of the lambert w function given by @xmath13 let @xmath14 be the function which gives the fractional part of a real number by subtracting either the floor @xmath15 or the ceiling @xmath16 of @xmath17 from @xmath17 , depending upon its sign .
furthermore , let @xmath18 be the argument of @xmath4 on the critical line .
[ ec]the for the imaginary part of the @xmath3-th zero of @xmath19 ( * ? ? ?
* equation 20 ) @xmath20 has a solution for each integer @xmath21 where @xmath22 enumerate the zeros of @xmath23 on the real line and the zeros of @xmath4 on the critical line @xmath24 where @xmath25 denotes the positive integers .
* equation 14 ) the @xmath3-th gram point is defined to be the solution of the equation @xmath26 a very accurate approximation @xmath27 to the gram points is @xmath28 is found by inverting @xmath12 to get the exact solution @xmath29 where @xmath30 is the lambert w function , and the approximation bounds @xmath31 when @xmath32 is @xmath33 , followed by @xmath34 and decreases monotincally with increasing @xmath3 , that is , @xmath35 .
the inverse of @xmath27 is given by @xmath36 now define the infinite sequence of functions indexed by @xmath37 @xmath38 near each `` bad '' gram point there will be a corresponding zero on the critical line which has an argument not on the principal branch .
the function @xmath39 determines how many multiples of @xmath40 to add or subtract to @xmath41 so that it agrees with the argument of @xmath4 at a zero on the critical line where it is discontinuous , having the value @xmath42 when @xmath43 .
let @xmath44 let @xmath45 be the sign of @xmath6 . [ c]the argument @xmath46 of @xmath19 at the n - th non - trivial zero @xmath47 on the critical strip is equal to @xmath48 , that is @xmath49 if conjecture [ c ] is true then conjecture [ ec ] is true and , due to @xmath50 [ fl ] , so is conjecture [ rh ] , the riemann hypothesis . if @xmath51=@xmath52 then @xmath52 is well - defined @xmath53 since @xmath51 is well - defined @xmath53 .
@xmath54 the argument of any positive number @xmath17 with @xmath55 is equal to @xmath56 and @xmath57 @xmath58 .
that is @xmath59
[ le]if the limit @xmath63 is exists and is well - defined @xmath64 then the left - hand side of equation ( [ ae ] ) is well - defined @xmath64 , and due to monotonicity , there must be a unique solution for every @xmath37 .
* ii.a ) the riemann - von - mangoldt formula makes use of cauchy s argument principle to count the number of zeros @xmath73 where @xmath74 with @xmath75 @xmath76 and this definition does not depend on the riemann hypothesis(conjecture [ rh ] ) .
this equation has exactly the same form as the asymptotic equation [ ae ] .
* equation 15 ) if the exact equation has a unique solution for each @xmath3 , then the zeros obtained from its solutions on the can be counted since they are enumerated by the integer @xmath3 , leading to the counting function @xmath78 in equation ( [ n0 ] )
. the number of solutions obtained on the would saturate counting function of the number of solutions on the so that @xmath79 and thus all of the non - trivial zeros of @xmath4 would be enumerated in this manner .
if there are zeros off of the critical line , or zeros with multiplicity @xmath80 , then the exact equation ( [ ee ] ) would fail to capture all the zeros on the critical strip which would mean @xmath81 .
* ix ) 1 guilherme frana and andr leclair .
transcendental equations satisfied by the individual zeros of riemann zeta , dirichlet and modular l - functions . _
communications in number theory and physics _ , 2015 . | the function @xmath0 is conjectured to be equal to @xmath1 when @xmath2 is the imaginary part of the @xmath3-th zero of @xmath4 on the critical line . if @xmath5 then the exact transcendental equation for the riemann zeros has a solution for each positive integer @xmath3 which proves that riemann s hypothesis is true since the counting function for zeros on the critical line is equal to the counting function for zeros on the critical strip if the transcendental equation has a solution for each @xmath3 . |
luminous sss are a class of x - ray object first observed with the _ einstein _ x - ray observatory in the late 1970s @xcite , of which cal 83 and cal 87 were the prototypes .
subsequently enhanced in number by the _
rosat _ soft x - ray surveys in the 1990s ( @xcite , for a thorough review , most recently updated by @xcite ) , their defining characteristic is their extreme luminosity ( up to @xmath2 ) at very soft ( @xmath3 kev ) x - ray energies . as such ,
the sss are rendered almost undetectable in the galactic plane due to interstellar extinction .
consequently , the known population of sss are predominantly in the magellanic clouds and m31 , although short sss phases have been observed in a number of classical and recurrent novae in the galaxy ( see e.g. bode ( 2011,2012 ) ) . while it was initially suggested that the sss contained black hole accretors , the combination of high luminosity and low temperature led ( * ? ? ?
* hereafter vdh92 ) to propose that sss were actually white dwarfs accreting at or close to the eddington limit .
such luminosities can be sustained via continuous nuclear burning of accreted material on the surface of the white dwarf .
but in order to achieve this a much higher mass transfer rate is required than is observed in most cataclysmic variables .
this in turn requires that the donor have a mass comparable to or greater than its accreting white dwarf , which will lead to thermally unstable mass transfer ( e.g. * ? ? ?
with sss phases now exhibited in many classical and recurrent nova outbursts @xcite , the vdh92 accreting white dwarf model has become a well - established paradigm . however , confirmation of the inverted mass ratio and very high mass transfer rates in order to explain the very highest luminosity sss rests on two independent , but still circumstantial , pieces of observational evidence : * @xcite observed bipolar emission line components in the transient sss , rx j0513.9 - 6951 , whose velocities were comparable to the escape velocity of a white dwarf .
* @xcite used _ xmm - newton _ to observe rx j0513.9 - 6951 through one of its transient sss events .
these revealed that , as the sss component fades , the effective radiating radius _ increases _ , shifting the peak of the emission to longer wavelengths and hence fading from view .
this supported the model of @xcite who had noted that the sss component occurs when the optical brightness _ decreases _ , and proposed that this is due to a _ contraction _ in the size of the white dwarf .
@xcite have discussed several mechanisms , such as photospheric adjustments , cessation of nuclear burning , absorption of x - ray emission by material above the white dwarf surface , amongst others , to explain the x - ray off - states seen in cal 83 .
they studied the early macho data of cal 83 , which suggested that there were three different optical states ( low , intermediate , and high ) .
unfortunately , only two x - ray off - states had been detected during that time , in 1996 @xcite and 1999 @xcite , and the variations in the macho light curve are complex compared to the ogle - iii lightcurve .
they concluded that the most likely explanation of the x - ray variations is the photospheric expansion / contraction model of @xcite . in an alternative model , @xcite ,
have used the more regular , recurrent optical high and low states of rx j0513.9 - 6951 to develop a `` limit cycle '' model , in which the sss phase drives a powerful wind from the disc .
this wind is capable of driving material from the donor so as to switch off effectively the mass transfer , thereby bringing the sss phase to an end .
the donor then recovers , mass transfer resumes , the sss switches on again , and the whole cycle continuously repeats .
further studies by @xcite showed that the duration of the optical high and low states , which they called `` cycle - length '' , varies for each cycle .
they suggested these variations to be caused by changes in the mass transfer rate ( by a factor of @xmath0 5 ) on timescales of years .
extensive x - ray observations of nearby galaxies , especially m31 , have been made with _
@xcite , _ chandra _ @xcite and _ xmm - newton _
@xcite , revealing a large number of extra - galactic sss .
these studies have been combined in a multi - wavelength ( x - ray , uv , optical ) census of sss in m31 by @xcite .
most of ssss discovered so far are in distant galaxies such as in m101 @xcite , m81 @xcite , ngc 4697 @xcite .
several of these ssss have been found to have very high luminosities ( @xmath4 ) , and it has been suggested that they may form a new class of sss called ultra - luminous supersoft x - ray sources " ( ulss ) , see e.g. @xcite .
their luminosities are extremely super - eddington for a normal white dwarf , moreover these sources exhibit state changes , they have a supersoft blackbody spectrum during the outburst and a hard power - law tail in the low state ( @xmath5 ) .
this make uls good candidates to be intermediate - mass black holes ( imbhs ) @xcite .
the prototype of this class is m101-ulx1 , which exhibited a supersoft spectrum with blackbody temperature of @xmath6 ev , and bolometric luminosity of about @xmath7 during its 2004 outburst @xcite .
however , this interpretation is still highly controversial , and the use of real wd atmosphere models ( rather than simple black - body spectra ) has already reduced the apparent luminosity in a number of cases @xcite . in this paper , in order to provide additional observational constraints on the various models proposed for sss properties , we have studied the long - term optical variations of the lmc sss cal 83 and rx j0513.9 - 6951 using their combined light curves from the macho and ogle project databases .
we compare their variability properties , with particular emphasis on the optical / x - ray correlated behaviour .
we accomplish this by using archival x - ray data of cal 83 obtained by _ xmm - newton_.
cal 83 has been monitored by the macho and ogle projects since 1993 , which now provide us with a combined light curve that spans 18 years .
the macho observations were made with the 1.27 m telescope located at mount stromlo observatory , australia @xcite .
this provides photometry in two passbands , a red band which we refer to as @xmath8-band ( @xmath0 6300 - 7600 , a slightly longer effective wavelength than johnson r ) and a blue band we refer to as @xmath9-band ( @xmath0 4500 - 6300 , slightly shorter effective wavelength than johnson v ) .
the data were taken during the interval 1993 to 2000 and are available through the macho website .
the data are in instrumental magnitudes .
the ogle project has used the 1.3 m warsaw telescope at las campanas observatory , chile , with phase iii running from 2001 until 3 may 2009 .
the photometric data for the ogle projects are taken in the i - band .
ogle - iii data for optical counterparts of x - ray sources located in the fields observed regularly by the ogle - iii survey are available through the x - ray variables ogle monitoring ( xrom ) website @xcite .
the ogle - iii data were calibrated using observations of @xcite standard stars which overlap with the ogle - ii fields @xcite .
these observations were then used to determine the offset between the ogle - iii instrumental magnitudes and the standard system , which we subsequently used to calibrate the remaining ogle - iii fields .
unfortunately , both cal 83 and rx j0513.9 - 6951 are not covered by the ogle - ii project , therefore we do not have any overlap between macho and ogle data ( usually from 1997 to 2000 ) .
the light curves were then analysed using the starlink period time - series analysis package , which is designed to search for periodicities in unevenly sampled data .
it has a variety of temporal analysis options , including the lomb - scargle periodogram ( ls periodogram , * ? ? ?
* ; * ? ? ?
* ) , as well as the capability to fold any data with a given period . the macho and ogle projects both included many regularly monitored fields within the magellanic clouds , thereby providing systematic monitoring of a significant number of luminous x - ray sources .
figure [ cal83long ] shows the combined macho and ogle light curve of cal 83 .
the macho light curve exhibits irregular variability , showing a long - lived intermediate state for the first @xmath0 700 days , and similar duration high state towards the end of the observations .
however , the ogle - iii light curve is rather different , exhibiting a regular succession of optical high and low states on timescales of 450 days with a difference in brightness of @xmath0 1 mag .
cal 83 spends roughly the same time ( @xmath0 200 days ) within each state , which is much longer than the duration of the transitions between states ( typically @xmath10 days ) .
it is instructive to compare the properties of the lmc supersoft source rx j0513.9 - 6951 with cal 83 .
its macho and ogle light curves are shown in figure [ rxj0513long ] .
rx j0513.9 - 6951 exhibits similar behaviour to cal 83 , but with a regular dimming on a shorter timescale ( @xmath0 168 days ) , and it also spends less time in the optical minimum state .
these timescales are nicely revealed in the lomb - scargle periodograms ( figure [ long_period_all ] ) of the combined macho and ogle light curves , which show broad peaks at periods of p = 450@xmath113 days and 168@xmath111 days for cal 83 and rx j0513.9 - 6951 , respectively .
we then folded the data on these long periods , with the results shown in figure [ long_period_all ] .
more importantly , for both systems , the regular dimming in the optical is associated with the turn - on of the supersoft source component .
this will be discussed further in the later sections .
we also exploit the ogle - iii data to search for short - term variations that could be related to the orbital properties of these systems .
the light curves were detrended by subtracting a linear fit to the optical low and high states .
the detrended light curves were then analysed using period .
the power spectrum of cal 83 ( figure [ cal83_per ] ) shows a significant peak at 1.047529 @xmath11 0.000001 days , which is very close to the modulation reported by @xcite of @xmath12 = 1.0436 days based on photometric and spectroscopic data .
the folded light curve on the presumed orbital period of @xmath13 = 1.047529 days is shown in figure [ cal83_per ] ( bottom ) .
it is sinusoidal , with a semi - amplitude of @xmath0 0.1 mag .
we performed the same analysis for rx j0513.9 - 6951 .
the periodogram of its detrended ogle - iii light curve reveals a significant peak at [email protected] days ( figure [ rxj_per ] ) , again very close to previously determined values found in the macho data alone by @xcite and @xcite , and spectroscopically by @xcite . the light curve ( folded on @xmath12 = 0.762956 days )
is shown in figure [ rxj_per ] , and is also well fitted by a sinusoidal modulation of semi - amplitude @xmath0 0.2 mag .
cal 83 was observed by _ xmm - newton _
( proposal i d 050086 and 050653 ) between may 2007 and may 2009 .
the observation logs of the x - ray measurements are summarised in table [ obslogs ] .
the observations were obtained with the epic - pn @xcite , epic - mos @xcite and rgs instruments .
the data reductions were performed using sas version 11.0.0 .
there are another four available _ chandra _ observations of cal 83 taken in 1999 and 2001 , but unfortunately three of them are during the x - ray off state , and the x - ray on state has already been reported by @xcite .
they are included in figure [ cal83long ] and are consistent with the anti - correlation .
we extracted the epic - pn and epic - mos light curves in the energy band @xmath14 kev .
the mean epic - pn and epic / mos1-mos2 count rates from each observation are plotted with the ogle - iii i - band light curve in figure [ opt_xray ] .
the x - ray mean count rates are highly variable and show two different x - ray states : a high state where the epic count rates are at a maximum ( @xmath15 for epic - pn , and @xmath16 for epic - mos ) , and a low state where the x - ray emission from the source is weak or completely off ( and which occur during optical maxima ) .
the anti - correlation between the optical and x - ray fluxes is clearly seen in figure [ opt_xray ] .
unfortunately , we have only a few ( 4 ) optical data points between mjd @xmath17 where most of the x - ray observations took place .
this makes it impossible for us to distinguish whether the x - ray variations lag or lead those in the optical .
the x - ray spectral analysis was done using the latest xspec version ( v 12.7.0 ) , and the spectra were extracted using only photons in the energy range @xmath18 kev .
we used a simple blackbody model ( _ bbody _ ) together with the tbingen - boulder ism absorption ( _ tbabs _ ) model , ism abundances from @xcite , and the photoelectric absorption cross - sections from @xcite .
the column density was fixed to the value derived by @xcite of @xmath19 = @xmath20 from _ hubble space telescope _
uv observations of cal 83 .
we add an additional absorption component with free column density and abundances set to 0.5 solar ( lmc abundances ) for elements heavier than helium .
the results from a simple blackbody model fit ( _ bbody _ ) to the epic - pn spectra are summarised in table [ res ] .
the values of the blackbody temperature , @xmath21 , inferred x - ray luminosity ( assuming an lmc distance of 50 kpc , @xcite ) , @xmath22 , and white dwarf radius , @xmath23 , vary significantly with the optical state of the source .
the blackbody temperature has a mean value of @xmath24 35 ev during optical low states when the x - ray flux is highest , and reduces to @xmath25 ev during optical high states .
figure [ opt_xray ] ( bottom ) shows the evolution of @xmath21 and @xmath26 with time , where they are clearly anti - correlated with its optical brightness .
the @xmath27 contour maps of the optical low state spectra ( obsid : 0500860201 ) and four merged optical high state spectra ( obsid : 0506530901 , 0506531001 , 0506531301 , 0506531401 ) are presented in figure [ chi_map ] , showing how the column density and the blackbody temperature correlate with each other .
.x - ray off - state observations [ cols="<,^,^ " , ] & & & & & & & & + ' '' '' 0500860201 & 2.0@xmath28 & 34.7@xmath29 & 4.0@xmath30 & 4.61 & 3.59 & 0.65 & 1.21 & 63 + ' '' '' 0500860301 & 0.8@xmath31 & 35.3@xmath32 & 2.7@xmath33 & 3.66 & 3.51 & 0.46 & 1.02 & 61 + ' '' '' 0500860401 & 2.6@xmath34 & 32.3@xmath35 & 4.9@xmath36 & 5.88 & 2.52 & 0.61 & 1.09 & 53 + ' '' '' 0500860501 & 1.4@xmath37 & 33.3@xmath38 & 3.3@xmath39 & 4.54 & 2.65 & 0.45 & 1.16 & 56 + ' '' '' 0500860601 & 1.4@xmath40 & 33.9@xmath40 & 3.8@xmath41 & 4.69 & 3.37 & 0.56 & 1.68 & 65 + ' '' '' 0500860901 & 3.6@xmath42 & 26.6@xmath43 & 0.9@xmath41 & 3.80 & 0.10 & 0.05 & 1.13 & 15 + ' '' '' 0506530201 & 0.0 & 29.6@xmath44 & 0.5@xmath45 & 2.33 & 0.32 & 0.05 & 0.65 & 23 + ' '' ''
0506530501 & 1.6@xmath46 & 30.3@xmath47 & 1.7@xmath48 & 3.96 & 0.77 & 0.17 & 0.75 & 32 + ' '' '' 0506530601 & 2.8@xmath49 & 28.1@xmath50 & 3.4@xmath51 & 6.54 & 0.67 & 0.24 & 0.64 & 36 + ' '' '' 0506530901 & 8.3@xmath52 & 23.5@xmath53 & 10.7@xmath29 & 16.45 & 0.12 & 0.30 & 0.72 & 22 + ' '' '' 0506531001 & 0.0 & 30.7@xmath54 & 0.3@xmath55 & 1.60 & 0.21 & 0.03 & 1.21 & 20 + ' '' '' 0506531301 & 8.9@xmath56 & 24.0@xmath57 & 10.5@xmath38 & 15.71 & 0.13 & 0.36 & 1.03 & 20 + ' '' '' 0506531401 & 0.1@xmath58 & 27.6@xmath59 & 0.4@xmath60 & 2.27 & 0.14 & 0.02 & 0.68 & 19 + ' '' '' 0506531501 & 0.4@xmath61 & 36.3@xmath62 & 2.4@xmath63 & 3.28 & 4.04 & 0.45 & 1.40 & 62 + ' '' '' 0506531701 & 1.7@xmath64 & 35.5@xmath65 & 3.3@xmath66 & 3.98 & 3.58 & 0.56 & 2.86 & 82 + [ res ]
in supersoft sources , the accretion disc is believed to be the main source of optical flux . however , these optical emissions arise in the accretion disc as a result of reprocessed supersoft x - rays from the hot white dwarf , and hence are not directly produced by accretion processes @xcite .
the reprocessed luminosity greatly exceeds that of the accretion luminosity , as the energy released by nuclear burning is significantly greater than that due to accretion onto a white dwarf , i.e. the disc luminosity will be comparable to that seen in normal galactic cataclysmic variables , at least one to two orders of magnitude below that seen here , see e.g. @xcite .
the ogle - iii light curve of cal 83 in figure [ cal83long ] shows a quasi - periodic variation on a timescale of @xmath0 450 days , with two well - defined optical states ( low and high ) .
the variation of 1 mag in optical brightness must be associated with the changes in the amount of irradiation of the accretion disk , which is most likely due to an expansion / contraction of the white dwarf photospheric radius @xcite , and/or the irradiated areas of the accretion disc itself @xcite .
the mass accretion rate required by the vdh92 model is very high , typically @xmath67 , and steady nuclear burning on the surface of the white dwarf can occur .
however , when the mass accretion rate exceeds the steady nuclear burning rate , the white dwarf photosphere will expand and its effective temperature will be reduced .
the duration of the optical low ( presumably , x - ray on ) and high states in cal 83 are @xmath0 200 days and @xmath0 250 days , respectively .
these durations are much longer than those observed in the recurrent supersoft source rx j0513.9 - 6951 ( @xmath0 40 days and @xmath0 140 days ) @xcite .
this implies that the mass accretion rate ( @xmath68 ) and the mass of the white dwarf ( @xmath69 ) in cal 83 are slightly lower than in rx j0513.9 - 6951 , as the duration of the x - ray on state is inversely proportion to the mass accretion rate @xcite . the higher mass accretion rate of rx j0513.9 - 6951 makes its optical luminosity ( i @xmath0 16.2 ) higher compared with that of cal 83 ( i @xmath0 16.7 ) , given that they are at the same distance and both are subject to very low extinction levels . nevertheless , the implied masses of the white dwarfs in both systems are @xmath70 .
the orbital periods of 1.05 days and 0.76 days that we have confirmed in cal 83 and rx j0513.9 - 6951 , respectively , ( on the basis of their stability over the last @xmath0 18 years ) are much longer than `` normal '' dwarf nova systems .
this also supports the vdh92 model of an evolved and more massive donor as required in order for the donor to be filling its roche lobe in such a system . in the vdh92 model , a steady nuclear burning of hydrogen near the white dwarf s surface is proposed to be the main source of supersoft x - rays .
this implies that variations in the x - ray flux may be caused by changes in burning rate , mass accretion rate or photospheric radius .
@xcite discussed these mechanisms in order to explain the observed x - ray off - states and optical variability in cal 83 .
they conclude that the model of @xcite ( expansion / contraction of wd s photosphere ) accounts for the x - ray and and optical variations . the x - ray off - state of cal 83 has been detected 8 times since its discovery ( see table [ offstate ] ) , and all of them occur during optical high states ( figure [ cal83long ] ) .
the observed x - ray spectrum slightly deviates from a blackbody fit ( especially at x - ray high state ) .
however , we can still see broad changes in its blackbody temperature and luminosity as a function of its optical brightness ( figure [ opt_xray ] ) .
the mass accretion rates in sss are comparable or slightly higher than the nuclear burning rate .
a modest changes in accretion rate will lead to an expansion of the white dwarf photosphere which enhances the irradiation of the accretion disc and the mass flow through the disc .
correspondingly , the effective temperature will drop , as the peak of the emission is shifted from supersoft x - rays towards longer wavelengths , principally into the extreme - ultraviolet .
however , the source will be brighter in the optical due to the increasing contribution from the expanding ( and still eddington - limited ) white dwarf emission , together with reprocessed radiation from the accretion disc / disc wind .
the results from our simple blackbody fits to the spectra show a higher effective temperature ( @xmath71 35 ev ) during optical low state , which then drops to @xmath71 30 ev during optical high ( if it is detectable at all ) .
this implies that the main source of the x - ray emission becomes cooler , and the peak flux will be shifted towards longer wavelengths .
this behaviour has already been seen ( and quantified ) in rx j0513.9 - 6951 by @xcite .
once the mass accretion rate drops , the white dwarf s photosphere will contract again , thereby raising its effective temperature and the source will re - enter the supersoft phase . on the other hand , the contraction of the white dwarf s envelope will reduce the amount of disc illumination .
this causes the optical brightness to reduce during the x - ray on - state .
the expanding / contracting photosphere model agrees very well with our optical and x - ray observations as well as the anti - correlation between x - ray and optical fluxes .
the most likely cause of the white dwarf s radius variations is the changes in mass accretion rate .
the physical mechanism underlying this process is still unclear , although several mechanisms have been suggested , such as the accretion wind evolution model @xcite where the strong wind from the white dwarf is colliding with and stripping off the outer layers of the slightly evolved secondary .
this will attenuate the mass outflow and therefore decrease the mass transfer rate .
some authors have suggested that the reduction of the mass accretion rate is connected to the magnetic activity of the secondary star , where mass loss from the secondary is attenuated when star spots pass over the inner lagrangian point @xcite .
this is similar to the proposed mechanism for the drops in accretion rate in vy scl stars @xcite , which has led to the suggestion that vy scl stars are low - mass extensions of supersoft sources @xcite . in the ogle - iii light curve ( figure [ cal83long ] ) , the times of optical low and high states in cal 83 are quite regular , apart for the last 300 days of the ogle - iii observation .
based on the expansion / contraction model the duration of the optical low state should be the duration where we expect an x - ray detection of the source . using the analytic model for recurrent supersoft sources by @xcite with recurrence time @xmath72 = 450 days and x - ray on time @xmath73 = 200 days ( the duration of optical low state ) , one would expect a white dwarf mass in the range of @xmath74 and mass accretion rate of @xmath75 .
we have studied the optical and x - ray variations of the prototypical sss , cal 83 .
the long - term optical light curve shows a quasi - periodic variation on a timescale of 450 days with two well defined optical states .
the duration of the optical low and high states ( 200 days and 250 days respectively ) are longer than those observed in rx j0513.9 - 6951 , which implies that the mass accretion in cal 83 is slightly lower than in rx j0513.9 - 6951 . using the combined macho and
ogle light curves , we have refined the short - term periods of cal 83 and rx j0513.9 - 6951 to values of 1.047529(1 ) d and 0.762956(5 ) d , respectively , which are consistent with previously reported values , and further strengthens their interpretation as being orbital in origin . the results from simple blackbody fits to the epic / pn spectra show a difference in blackbody temperature ( from @xmath76 30 ev to @xmath0 35 ev ) and luminosity between the optical high and low states .
the optical brightness and the x - ray luminosity in cal 83 clearly show an anti - correlation .
the non - detections of x - ray emission , in eight x - ray observations since its discovery , occur only during the optical high state .
this is in good agreement with the photospheric expansion / contraction model suggested by @xcite . | we have studied the long - term ( @xmath0 years ) temporal variability of the prototype supersoft x - ray source ( sss ) cal 83 in the lmc , using data from the macho and ogle projects .
the cal 83 light curve exhibits dramatic brightness changes of @xmath0 1 mag on timescales of @xmath0 450 days , and spends typically @xmath0 200 days in the optical low state . combined with archival _ xmm - newton _ x - ray observations
these represent the most extensive x - ray / optical study to date of this system , and reveal in much greater detail that the x - ray light curve is anti - correlated with the optical behaviour .
this is remarkably similar to the behaviour of the `` transient '' sss , rx j0513.9 - 6951 , where the sss outbursts recur on a timescale of @xmath0 168 days , and also anti - correlate with the optical flux .
we performed simple blackbody fits to both high and low state x - ray spectra , and find that the blackbody temperature and luminosity decrease when the optical counterpart brightens .
we interpret these long - term variations in terms of the limit - cycle model of @xcite , which provides further support for these systems containing massive ( @xmath1 ) white dwarfs .
in addition , we have refined their orbital periods in the macho and ogle - iii light curves to values of 1.047529(1 ) days and 0.762956(5 ) days for cal 83 and rx j0513.9 - 6951 , respectively .
[ firstpage ] stars : individual : cal 83 , stars : binaries : close , stars : white dwarfs , accretion , accretion disks , x - rays : stars , galaxies : individual : large , magellanic cloud |
recently , a unified theory of antiferromagnetism ( af ) and superconductivity ( sc ) has been proposed for the high @xmath1 cuprates@xcite .
this theory is based on the @xmath0 symmetry between af and sc , and offers a unified description of the global phase diagram for this class of materials .
while the theory was originally proposed as a effective field theory description , it was soon realized that the @xmath0 symmetry could be implemented exactly at a microscopic level@xcite , and it can also be checked numerically in common strongly correlated models such as the @xmath2 model@xcite . while the phase diagram@xcite and collective excitations@xcite in the sc state derived from these @xmath0 models bear strong resemblance with the high @xmath1 cuprates , and a number of novel experimental predictions have been made@xcite , the mott insulating behavior at half - filling is a puzzling aspect which challenges the fundamental validity of the @xmath0 models@xcite . to be more precise ,
the exact @xmath0 symmetry requires collective charge two excitation at half - filling to have the same mass as the collective spin wave excitations .
this condition is clearly violated in a mott insulating system where all charge excitations measured with respect to a particle hole symmetric point have a large energy gap of few @xmath3 , while the spin wave excitations are massless . in the original @xmath0 proposal ,
it was pointed out that this situation is analogous to a easy - axis antiferromagnet in a external uniform field , and a @xmath0 symmetry breaking term at half - filling was introduced in order to describe this asymmetric behavior between spin and charge .
the chemical potential also introduces a @xmath0 symmetry breaking term , however , it was shown that these two terms could compensate each other@xcite so that the _ static potential _ governing the @xmath0 superspin could still be @xmath0 symmetric . since the asymmetry between the charge and spin excitations at half - filling is of the order of the coulomb energy scale @xmath4 , the @xmath0 symmetry breaking terms must also be of that order .
since there are various types of symmetry breaking terms , one might hope that their effects could partially cancel each other to arrive at a qualitatively correct picture .
however , this type of cancellation is very delicate , and approximate calculations could easily lead to erroneous conclusions . in particular
, one is interested in which physical properties could exhibit @xmath0 symmetric properties in the limit when the coulomb gap is taken to infinity .
for example , one could ask the following questions : \1 ) one of the hallmarks of the @xmath0 symmetry is not only the degeneracy between the af and sc states at a given chemical potential , but the approximate degeneracy among all mix - states interpolating between af and sc , _ i.e. _ the independence of the ground state energy on the superspin angle . what is the potential barrier separating the af and sc states at their degeneracy point in the limit @xmath5 ? if there is a large energy barrier in this limit , one would argue that the concept of @xmath0 symmetry is not a useful one , at least not for quantitative calculations .
on the other hand , if the potential barrier is finite and small in the @xmath5 limit , the concept of a approximate @xmath0 symmetry would be a useful one .
\2 ) exact @xmath0 symmetry predicts four massless collective modes . in the half - filled af state , besides the two conventional massless spin wave modes , the exact @xmath0 symmetry predicts a massless doublet of @xmath6 modes , with charge @xmath7 .
however , a mott insulator has a large gap to all charge excitations .
therefore , it is clear that one of the @xmath6 has to be projected out of the spectrum in the limit @xmath5 , say the @xmath8 mode carrying charge @xmath9 .
what happens to the rest of the goldstone modes , the @xmath10 mode carrying charge @xmath11 and the @xmath12 triplet mode of the sc state ? in the @xmath5 limit , can they all be simultaneously massless at the transition point between af and sc ?
since the pure sc state can only be reached with a finite doping concentration , is it possible that the gutzwiller projection does not affect the @xmath12 triplet mode of the sc state ? in order to address these questions , it is desirable to construct a low energy effective theory without any parameters of the order of the coulomb scale @xmath4 . in this work
, we construct a class of projected @xmath0 models which treat the gutzwiller constraint exactly and locally on every site .
we use this model to answer the physical questions posed above and show that the answers are affirmative . in the @xmath5 limit , when the gutzwiller constraint is implemented exactly , the ground state energy can still be @xmath0 symmetric and independent of the superspin direction . after projecting out the @xmath8 mode , all other goldstone modes remain massless at the symmetric point .
the dispersion relation of the collective modes bear unique signature of the projected @xmath0 symmetry .
furthermore , the @xmath12 triplet modes of the pure sc states are unaffected by the gutzwiller projection .
these properties define the concept of a projected @xmath0 symmetry ( @xmath13 ) , whose properties and consequences we shall explore in this paper .
the fundamental quantity in the @xmath0 theory is the locally defined five component superspin vector @xmath14 describing the local af and sc order parameters respectively . in the nonlinear @xmath15 model formalism ,
these are treated as mutually commuting coordinates and their dynamics is given by their conjugate momenta @xmath16 .
the charge operator is the angular momentum in the @xmath17 plane : @xmath18 implementing the gutzwiller constraint corresponds to requiring @xmath19 for every local @xmath0 rotor . from equations ( [ charge ] ) and ( [ gutzwiller ] ) and subsequent discussions , we shall see that the gutzwiller projection in the @xmath0 formalism corresponds to going from a fully symmetric @xmath0 rotor model to a _ chiral _
@xmath0 rotor model , where both the static potential of the individual rotors and the coupling between the rotors are still @xmath0 symmetric , but the rotors are constrained to rotate only in one sense in the @xmath17 plane , consistent with ( [ gutzwiller ] ) .
this observation reveals a deep connection between the gutzwiller projection and the lowest - landau - level ( lll ) projection in the fractional quantum hall effect@xcite . to be more precise , the gutzwiller projection represented by equations ( [ charge ] ) and ( [ gutzwiller ] ) is analogous to the lll projection , where all states in the lll have a definite sign of angular momentum . the lll projection can be analytically implemented by separating the cyclotron degrees of freedom from the guiding center degrees of freedom , which amounts to changing the commuting property between the @xmath20 and @xmath21 coordinates to a canonically conjugate commutation relation : @xmath22 = i \l_0 \label{guiding - center}\ ] ] where @xmath23 is the landau length . exploiting this analogy , we find that the original @xmath0 model can be fully gutzwiller projected without changing its form , if one imposes the simple quantization condition between the superconducting components of the superspin vector : @xmath24 = i/2 \label{non - commute}\ ] ] in the symmetric @xmath0 model , the wave function of the @xmath0 rotors are functions of the local coordinates @xmath25 and @xmath26 , while the projected @xmath0 model only depends on their holomorphic combination @xmath27 and is independent of their anti - holomorphic combination @xmath28
this way , we arrive at a natural projection of the @xmath0 model where the local gutzwiller constraint is taken into account exactly , and the resulting model is free of the large coulomb @xmath4 parameter . because the functional form of the symmetric @xmath0 model remain the same and only the quantization condition is modified upon projection , many important properties associated with the @xmath0 symmetry remain .
the central hypothesis of the @xmath0 theory is that this projected model is quantitatively accurate in describing both the static and dynamic properties of the high @xmath1 cuprates , and we shall compare the properties of this model with the phenomenology of the high @xmath1 systems .
we begin with the symmetric @xmath0 hamiltonian defined on a lattice , @xmath29 where @xmath30 denotes the five component superspin vector on a given site , and @xmath31 is the @xmath0 symmetry generator , @xmath32 expressed here in terms of the superspin vector @xmath33 and its canonically conjugate momenta @xmath34 , @xmath35 = i \delta_{ab}\ ] ] this lattice quantum non - linear @xmath15 model can be rigorously derived as the low energy limit of a microscopic @xmath0 ladder model@xcite . on the ladder ,
the rung @xmath0 singlet state is the vacuum @xmath36 , from which the lowest so(5 ) multiplet @xmath37 is created by a quintet of bose creation operators , which satisfy here , @xmath39 denote the triplet ( magnon ) states , and @xmath40 are the hole and particle pair states .
( see fig .
( [ fig1 ] ) ) .
the superspin coordinates are microscopically constructed using these lattice bosons , @xmath41 = 4.0 cm due to their microscopic origin , these bosonic states are hard - core bosons , in the sense that one can not define two of them on the same rung .
the @xmath42 term in equation ( [ sigma ] ) describes the gap energy of the magnon and the pair states , the @xmath43 term stands for the hopping and the spontaneous creation / destruction process of these states , and the @xmath44 term describes their nearest - neighbor interaction .
this quantum nonlinear @xmath15 model can in principle also be derived in higher dimensions from a microscopic @xmath0 symmetric model@xcite , by introducing a superspin vector as a hubbard - stratonovich decoupling field , and integrate out the fermionic degrees of freedom in a gradient expansion .
however , we shall proceed more heuristically here@xcite . for a two dimensional system
, one can imagine that the quantum @xmath15 model hamiltonian is obtained from a
block spin " type of coarse - graining of the microscopic electron hamiltonian , and is defined on a lattice with twice the lattice constant compared to the microscopic electron model .
( this doubled unit cell is the minimal size needed to define the local af and d wave sc order parameters ) .
therefore , each site @xmath45 in the effective model correspond to a plaquette of the microscopic electron model .
( on a ladder , this corresponds to going from the lattice sites to ladder rungs ) .
the @xmath0 singlet state @xmath36 corresponds to a rvb " type of singlet state , while the five - fold states @xmath46 describe the triplet magnon states , and the @xmath47-wave hole and particle pair states on a plaquette . unlike the ladder case
, the magnon and the @xmath47- wave pair states could condense in the ground state to form af and sc broken symmetry states .
in fact , eder@xcite has recently shown that properties of the af states can be described by a coherent state of magnon condensation on top of a uniform spin liquid state .
our model therefore describes competition among the rvb " type of singlet vacuum and the two forms of broken symmetry order .
while it is reasonable to take @xmath43 and @xmath44 to be approximately equal for magnons and pairs , the gap energy @xmath42 for the neutral magnons and the charged pairs are very different in the insulating state at half - filling .
in fact , their difference is of the order of the insulating gap @xmath4 at half - filling . taking into account the hard - core condition and neglecting the nearest - neighbor interaction @xmath44 for now ( it has higher powers of time and space derivatives in the continuum limit ) , we can express the general anisotropic @xmath0 model as @xmath48 in this paper we shall use the convention where @xmath49 denote the superspin indices , @xmath50 denote the spin indices and @xmath51 denote the charge indices , and repeated indices are summed over .
the main focus of our paper is to consider the limit where @xmath52 .
let us define the charge eigen - operators @xmath53 and @xmath54 as @xmath55 from this definition , it is clear that @xmath56 is the creation operator for a hole pair and @xmath57 is the creation operator for a particle pair .
we can introduce a chemical potential term @xmath58 to describe the effects of doping . in the presence of this chemical potential term ,
the gap energy of the hole and particle pairs are @xmath59 and @xmath60 respectively .
a chemical potential of the order of the charge gap @xmath61 is needed to induce a metal - insulator transition in this system .
near such a transition point , the gap energy of the hole pair @xmath62 can be comparable to the spin gap @xmath63 , while the gap towards a particle pair excitation is of the order of twice the charge gap , and needs to be projected out of the spectrum in the low energy limit .
therefore , within this formalism , the gutzwiller projection is equivalent to restricting ourselves to the projected hilbert space where @xmath64 at every site @xmath45 . within this projected hilbert space
, the projected hamiltonian takes the form @xmath65 this hamiltonian has no parameters of the order of @xmath4 , and it is reasonable to expect @xmath66 and @xmath67 .
we see that the form of the hamiltonian hardly changes from the unprojected model , but the definition of @xmath25 and @xmath26 is changed from @xmath68 to @xmath69 from equation ( [ before ] ) , we see that @xmath25 and @xmath26 commute with each other before the projection .
however , after the projection , they acquire a nontrivial commutation relation , as can be seen from equation ( [ after ] ) : @xmath24 = i/2\ ] ] therefore , the gutzwiller projection can be analytically implemented in the @xmath0 theory by retaining the form of the hamiltonian and change only the quantization condition .
the discussions outlined above reveal a deep connection between the gutzwiller projection within the @xmath0 formalism and the projection onto the lowest landau level ( lll ) in the context of the fractional quantum hall effect .
consider the problem of a charged particle in a strong magnetic field @xmath70 and a rotationally symmetric potential @xmath71 . in the absence of a magnetic field , all eigenstates form irreducible representations of the two dimensional rotation group @xmath72 , characterized by integral eigenvalues of the angular momentum operator @xmath73 however , in the presence of a strong magnetic field and projected into the lll , only negative eigenvalues of @xmath74 are realized .
this is analogous to the situation encountered here .
the local charge operator in the @xmath0 theory takes the form of the angular momentum in the @xmath17 plane as given by equation ( [ charge ] ) .
when doubly occupied sites are locally projected out , the local charge operator , or the angular momentum in the @xmath17 plane , takes only negative values . since the chemical potential couples directly the angular momentum in the @xmath17 plane , it plays the role of a fictitious magnetic field threading every @xmath0 rotor in the @xmath17 plane .
the landau level spacing @xmath75 is analogous to the charge gap @xmath61 encountered here , and both are taken to be infinity in the projected models . after the projection ,
the hamiltonian in the landau level problem retains its @xmath72 symmetric form , @xmath76 although a new quantization condition is imposed between @xmath20 and @xmath21 , as given by equation ( [ guiding - center ] ) .
this is analogous to the observation we made here that the hamiltonian formally retains a @xmath0 symmetric form after the projection ( [ constraint1 ] ) , but the quantum dynamics is changed due to the non - trivial commutator between @xmath25 and @xmath26 . in both cases only a part of the full symmetry multiplets remain after the projection .
however , the formal symmetry of the hamiltonian has direct physical manifestations despite the projection .
for example , in the lll problems , semi - classical orbits of the guiding center coordinates are still @xmath72 symmetric . in our case
, we shall see that the static potential for the superspin vector can still be @xmath0 invariant despite the projection . perhaps the most explicit way to establish the precise connections between these two problems is to consider the constraints on the wave function . in the symmetric gauge of the lll problem , the annihilation operator for the cyclotron coordinates
takes the form@xcite @xmath77 where @xmath78 and @xmath79 .
projection onto lll requires @xmath80 which determines the form of the lll wave function to be @xmath81 where @xmath82 is a holomorphic function of @xmath83 only .
this holomorphic condition also places strong constraints in many - body systems and led to the celebrated laughlin s wave function .
our no - double - occupancy constraint ( [ constraint1 ] ) is analogous to the lll constraint ( [ constraint2 ] ) .
in fact from equations ( [ p - h ] ) and ( [ before ] ) , we obtain @xmath84 where @xmath85 and @xmath86 . for a single unprojected @xmath0 rotor ,
the wave function @xmath87 is a function of the superspin coordinates .
however , the gutzwiller projection ( [ constraint1 ] ) restricts the wave function to be @xmath88 where @xmath89 is a holomorphic function of @xmath83 . for a collection of @xmath0 rotors ,
the superspin coordinates are themselves functions of the lattice sites @xmath45 , and @xmath90 $ ] is a functional is the superspin coordinates at each site . for the projected @xmath0 models , this functional
is restricted to take the form @xmath91 = f(z(x),n_\alpha(x ) ) \prod_{x } e^{-z\bar z(x)/2 } \label{functional}\ ] ] where @xmath92 is a holomorphic functional of @xmath93 .
the formal but precise analogy between the two types of projection allows us to introduce the concept of a chiral @xmath0 rotor .
this is a system of rotors with @xmath0 invariant potential and coupling , however , the rotation within the @xmath17 plane is chiral , _
i.e. _ only one sense of the rotation is allowed .
such a system of chiral @xmath0 rotors is described by the wave functional in equation ( [ functional ] ) .
having discussed the general notions of the projected @xmath0 model , we are now in a position to explore the phase diagram of this model . as we commented earlier , the projected @xmath0 model describes the competition and unification of the spin liquid , af and the sc states . in the original unprojected @xmath0 symmetric model ,
not only are the af and sc states degenerate in energy , but they are also degenerate with all the intermediate coexistence states .
this points out a route from af to sc with no potential barrier , and introduces the concept that the metal - insulator transition in the high @xmath1 systems can be viewed as a smooth rotation of the @xmath0 superspin .
one of the key questions to be answered in this work is what happens to the picture in the case of projected @xmath0 symmetry . in anticipation of the competition of the states discussed above
, we construct a class of variational wave functions in the coherent state representation : @xmath94 here @xmath95 denotes a local singlet state defined by @xmath96 and @xmath36 is a product state of these local singlets , @xmath97 .
@xmath98 is a local variational parameter describing the competition between long range order and quantum disorder . for @xmath99 our variational wave function describe a spin singlet ground state , while a non - zero value of @xmath100 describes a coherent state formed by the local singlet , the magnon or the hole pair state
this wave function is a generalization of the coherent state description of a af state in terms of a magnon condensate@xcite . as we shall see from the following equation ( [ superspin_vector ] ) , @xmath101 stands for the length of the @xmath0 superspin vector . @xmath102 and @xmath103 are general complex variational parameters describing the local amplitude for magnons and hole pairs .
we notice that this wave function satisfies both the gutzwiller constraint ( [ constraint1 ] ) and the hard - core constraint for magnons and hole pairs exactly . it is easy to see that @xmath104 where @xmath105 and @xmath106 denote the real and imaginary parts of a complex number .
the coupling terms in the projected @xmath0 hamiltonian depend only on @xmath107 , @xmath108 and @xmath109 .
therefore , the coupling energy depends only on the real part of @xmath102 while it depends on both the real and imaginary parts of @xmath103 .
therefore , for discussing the ground state wave functions , we can assume without loss of generality that @xmath102 is real and @xmath110 .
the normalization condition @xmath111 can be implemented by the constraint that @xmath112 therefore , we see that although we have completely projected out the particle pair states , the local degrees of freedom can still be represented by a vector on a five dimensional sphere .
uniform states are obtained by taking all parameters to be constant . for @xmath113 and @xmath114 ,
our wave function @xmath115 describes a pure af state with the following properties : @xmath116 where @xmath117 is the number of lattice sites .
equation ( [ af - state ] ) describes a half - filled state with a macroscopic neel magnetization , and vanishing uniform magnetization .
furthermore , this state is composed as a linear superposition of eigenstates with different values of the total spin , and the fluctuation of the total spin scales like @xmath118 , just as one expects from a standard neel state .
on the other hand , for @xmath119 and @xmath114 , @xmath115 describes a pure sc state with the following properties : @xmath120 equation ( [ sc - state ] ) describes a state with a finite doping density and a finite sc order parameter .
just as in the standard bcs case , this state is composed as a linear superposition of eigenstates with different values of the total charge , and the fluctuation of the total charge scales like @xmath121 , just as one expects from a standard sc state .
however , besides these two _ pure _ states , there is a class of _ mixed _ states which interpolates between the pure af and sc states .
taking @xmath122 and @xmath123 , we see that the mixed states have the following property : @xmath124 therefore , we see that there is a continuous family of intermediate mixed states interpolating between the pure af state at half - filling and the pure sc state with finite doping density . as the @xmath0 angle
@xmath125 rotates continuously from a pure af state with @xmath126 to a pure sc state with @xmath127 , the hole density of the mixed state interpolates continuously between these two limits .
therefore , our wave function gives a unified description of af and sc and points out a precise route from af to sc as the doping level is varied . in order for this route , or small deviations from this route ,
to be physically realized in the high @xmath1 superconductors , we have to demonstrate that there is no large energy barrier for the intermediate mixed states , or that the ground state energy is approximately independent of the @xmath0 mixing angle @xmath125 .
in particular , we have to show that the energy barrier is independent of the hubbard energy @xmath4 , in the limit of large @xmath4 . in the following
, we shall investigate this question .
the energy functional @xmath128 describes the coupling between these five dimensional vectors @xmath102 , and it is given by @xmath129 this ground state energy functional describes a systems of coupled rotors satisfying the constraint ( [ normalization ] ) .
at the point @xmath130 and @xmath131 in parameter space , this rotor model is exactly @xmath0 symmetric .
this is a central observation of this work . from this consideration
we learn a very important lesson about the compatibility of the mott insulating gap and the idea of a smooth @xmath0 rotation from af to sc . as we have seen , the large asymmetry between the charge and spin gap at half - filling necessitates the removal of the @xmath132 part of the @xmath0 multiplets , therefore , the dynamics close to half - filling has to be modified .
but the static potential governing the transition from af to sc can remain @xmath0 symmetric , and in particular , the energy barrier separating these two states can remain small in the limit where the mott insulating gap tends to infinity .
we also observe a crucial difference between the projected and the unprojected @xmath0 models . in the unprojected @xmath0 model with the full @xmath0 symmetry , af and sc states are degenerate at half - filling , and the rotation between these two states
can be continuously performed without changing the density to going away from half - filling .
this case is similar to the well - known degeneracy between the cdw state and the @xmath133 wave sc state for the negative @xmath4 hubbard model at half - filling . in the projected @xmath0 model , where all particle pair states have been locally removed ,
the @xmath0 rotation from the af to sc states are accompanied by the continuous change of the hole density , and a pure sc state can only be reached at a finite critical hole doping density @xmath134 .
while the unprojected @xmath0 symmetry is only valid at half - filling , the projected @xmath0 symmetry can be valid for a range of doping concentration @xmath135 , since all these doping concentrations correspond to the same value of the chemical potential @xmath136 at which the ground state energy ( [ energy ] ) is @xmath0 symmetric .
however , it should be pointed out that the projected @xmath0 symmetry at @xmath136 has only been demonstrated within the variational mean field approximation .
this corresponds to the semiclassical limit , and becomes exact only in the large @xmath133 limit , where @xmath133 labels the representation of the local @xmath0 group at a given site .
quantum fluctuations can be systematically investigated as a @xmath137 expansion .
assuming uniform ground states , we have studied the effect of zero point fluctuations in sec .
[ fluctuations ] and found that at @xmath136 , the intermediate mixed states have slightly higher energy than the af and sc state .
therefore , quantum fluctuation leads to a slight breaking of the projected @xmath0 symmetry .
the important point here is that this symmetry breaking effect can be systematically controlled in the semi - classical @xmath137 expansion , and certainly is independent of the coulomb energy scale @xmath4 .
this fluctuation would induce a first order transition and predict phase separation of af and sc states at @xmath136 .
however , there are also other competing interactions such as nearest - neighbor and next - nearest - neighbor interactions which tend to reduce the barrier , and could also lead to non - uniform states like stripes . due to the complexity of the calculations
, we shall defer the detailed studies of these competing effects to future works .
in this section we investigate the phase diagram of the projected @xmath0 model within the framework of the variational wave function ( [ variational ] ) .
taking uniform values of the variational parameters @xmath100 , @xmath138 , @xmath122 and @xmath139 , the variational energy in ( [ energy ] ) reduces to @xmath140 in the following , we shall mainly study the @xmath0 symmetric case , and take @xmath141 . defining @xmath142 and @xmath143 , and the dimensionless coupling constants @xmath144 , @xmath145 and @xmath146
, we obtain @xmath147 we shall minimize ( [ mean_energy ] ) with respect to @xmath45 and @xmath148 , subject to the condition that @xmath149 .
the phase diagram can be plotted in the two dimensional parameter space of @xmath150 and @xmath151 .
we notice that @xmath152 depends linearly on @xmath148 , therefore , for @xmath153 we obtain @xmath154 and @xmath155 similarly , for @xmath156 we obtain @xmath157 and @xmath158 from these equation , we can determine the phase diagram as shown in fig .
( [ fig2 ] ) .
= 10.0 cm there are seven different phases on this phase diagram .
@xmath159 corresponds to a quantum disordered singlet state with no condensed bosons . @xmath160 and @xmath154 corresponds to a quantum disordered state with completely filled hole pairs . @xmath160 and @xmath157 corresponds to a quantum disordered state with completely filled magnons . @xmath161 and
@xmath157 describe a pure af phase , while @xmath161 and @xmath154 describes a pure sc phase .
when @xmath162 a continuous family of mixed af / sc states labeled by a free superspin angle @xmath163 is realized , while for @xmath164 a continuous family of quantum disordered states labeled by a free superspin angle @xmath163 is obtained .
the system traces out a one dimensional trajectory in this two dimensional phase diagram as the chemical potential is increased , as depicted in fig .
( [ fig2 ] ) .
increasing the chemical potential decreases the @xmath150 parameter while holding @xmath151 constant .
@xmath151 describes the degree of quantum spin fluctuations in the system , since in the af phase , the size of the neel moment @xmath165 decreases with increasing @xmath151 . for @xmath166 ,
the system goes through a phase transition from af to sc at @xmath167 . at this critical value of the chemical potential ,
the @xmath100 parameter remains fixed , but the @xmath125 parameter changes continuously from @xmath168 to @xmath169 , and correspondingly , the density changes from @xmath168 to @xmath170 this behavior gives a density versus chemical potential diagram as shown in fig . ( [ fig3 ] ) .
= 4.0 cm as we see , for densities in the range @xmath135 , the system is infinitely compressible since @xmath171 . for @xmath172 or @xmath173
, the system has a finite compressibility of @xmath174 .
it is interesting to plot both the af and the sc order parameters @xmath175 and @xmath176 as a function of the density for the whole range of @xmath177 .
we will restrict to the case of @xmath178 where the undoped state is a af state .
we obtain the following doping dependence of the sc order parameter : @xmath179 and the doping dependence of the of the af order parameter : @xmath180 these behaviors are depicted in fig .
( [ fig4 ] ) .
= 6.0 cm we note several interesting features of the phase diagram .
first of all , we can use the energy functional ( [ energy ] ) as a starting point for a finite temperature classical fluctuation analysis , and estimate the transition temperature due to the classical fluctuations .
within such a framework , the 3d af ( @xmath181 ) and sc ( @xmath1 ) transition temperatures are proportional to the stiffness of the spin and the phase fluctuations , which are in turn proportional to @xmath182 and @xmath183 respectively . therefore , fig .
( [ fig4 ] ) gives an approximate estimate of the transition temperatures .
for small @xmath184 , we see that there is a sharp drop of @xmath181 and a maximum of @xmath1 at @xmath185 .
the system is a af insulator at doping @xmath186 and a pure sc state for @xmath173 .
the reason for a maximum of @xmath1 at @xmath185 is due to the strong correlation of the charged bosons .
the charge bosons have a hard - core interaction , therefore , they are insulating at both @xmath186 and @xmath187 and have maximal charge stiffness at @xmath185 .
we can perform a rough translation of this optimal doping value in our effective model to the microscopic model . since
our effective model is defined on a unit cell with twice the lattice spacing of the microscopic model , @xmath185 therefore describes one hole pair per four sites in the microscopic model , or a doping of @xmath188 in the conventional language .
this crude argument tends to overestimate the value for optimal doping , since it neglects the effects of unpaired electrons .
however , considering the crudeness of the estimate , it is still reasonably close to the optimal doping @xmath189 observed in the lsco family of high @xmath1 superconductors . in the regime of @xmath135 ,
the system is a coherent mixture of af and sc order . for this entire range of densities , the system has a projected @xmath0 symmetry within the variational approximation discussed above .
the projected @xmath0 symmetry manifests itself in terms of a infinite compressibility in the region @xmath135 and , as we shall see in next section , a charge mode with a dispersion relation @xmath190 .
since such a state is rather unusual , and maybe highly susceptible to density fluctuations , we would like to discuss more detailed physical properties in this region . first let us comment on the fact that there are familiar physical systems whose uniform ground states are infinitely compressible .
the free boson model is certainly such an example , and the density mode also has a @xmath190 dispersion relation .
but the infinite compressibility is due to the absence of the interaction , which is not characteristic of the strongly interaction system considered here .
a less trivial example is the spin @xmath191 xxz ferromagnetic heisenberg model , given by the hamiltonian , @xmath192 where @xmath193 and the sum extends over nearest neighbor sites of a square lattice .
this model can be interpreted as quantum hard - core boson model , where the fully polarized spin down state could be identified with the vacuum of the bosons , the xy part of the hamiltonian describes the hopping of the bosons and the last term describes the nearest neighbor attraction between the bosons if @xmath194 .
when @xmath195 , the system is in the ising limit , and the spontaneous breaking of the @xmath196 symmetry implies phase separation of the bosons . on the other hand , when @xmath197 , the system is in the xy limit , and the ground state is a superfluid .
therefore , the anisotropy parameter describe the competition between superfluidity and phase separation . at @xmath198 ,
the system has a @xmath199 symmetry and the dispersion relation becomes quadratic .
different directions of the ferromagnetic polarizations are degenerate and can be changed without any energy cost .
since the @xmath83 component of the ferromagnetic polarization is identified with the total density of the hard - core bosons , the @xmath199 vacuum degeneracy implies infinite compressibility of the corresponding boson system for the entire range of boson densities @xmath177 .
these two examples illustrates that there is nothing intrinsically pathological about having a system with infinite compressibility .
the second example is more generic , and shows that uniform states with infinite compressibility can be obtained in systems on the verge of phase separation , and the infinite compressibility can be ensured by symmetry .
both of these properties are also shared by the projected @xmath0 model .
these models are on the verge of phase separation into af and sc phases , and the infinite compressibility is a result of the projected @xmath0 symmetry .
it indicates that small perturbations , such as next nearest neighbor interactions , quantum fluctuations , and quenched disorder will be very important to determine the true ground state . with such perturbations ,
the ground state is expected to be unstable toward the experimentally reported textures such as the spin glass , stripes and incommensurate spin density waves .
next let us investigate the phenomenological consequence of this remarkable property .
one of the most puzzling properties of the high @xmath1 superconductors is the constant chemical potential in the underdoped samples .
for lsco systems , where the doping level can be varied continuously by the @xmath200 concentration , this effect has been dramatically observed in the arpes experiments and the constant chemical potential persists from the weakly doped insulator to optimally doped superconductor@xcite .
in fact numerical calculations on the hubbard model also reveal similar divergent behavior of the compressibility as the metal - insulator transition is approached from the metallic side@xcite .
the simplest explanation of the small chemical potential shift is a two phase mixture with different densities at a first order phase transition .
if the system globally phase separates into two different spatial regions with different charge densities but the same free energy densities , the added charges only change the proportion of mixture of the two phases and do not change the energy , therefore , @xmath201 .
however , this situation of global phase separation can certainly not occur in a system with long ranged coulomb interaction and is ruled out in the real high @xmath1 system .
a phenomenon possibly related to the tendency of phase separation is the formation of stripes@xcite .
a stripe state can be viewed as microscopic phase separation of af and sc into alternating regions , where each region has different charge density and the same free energy density .
however , a crucial difference between the global phase separation and this picture of microscopic phase separation is that the stripe state has infinitely many surfaces between af and sc , and the surface energy makes a finite contribution to the total energy in the thermodynamic limit . in this picture
, doping can be accomplished by converting af stripes into sc stripes , thereby creating more surfaces separating af and sc regions .
finite doping density therefore leads to a finite density of surfaces and the accumulated surface energy would in general lead to a shift of the chemical potential _ under a generic situation_. additional physical conditions are needed to ensure the constant chemical potential in the stripe phase .
therefore , the absence of the chemical potential shift places a very strong constraint on possible theoretical explanations . the projected @xmath0 model proposed in this work offers a possible explanation for the absence of chemical potential shift . at a critical value of the chemical potential where
the af and sc states have degenerate energy density , we can have three situations , where the intermediate mixed states have higher , lower or degenerate energy compared to the af and sc states .
when the intermediate states have higher energy , the system will go through a first order phase transition at @xmath136 and this will lead to global phase separation into af and sc regions . on the other hand , if the intermediate states have lower energy , there exists a range of chemical potential @xmath202 where the mixed phase has a uniform and continuously varying density . in this case , @xmath203 is obtained .
when the region @xmath202 shrinks to zero , we obtain the limiting @xmath0 symmetric case where the system is on the boundary between a first order transition and two second order phase transitions . in this case , @xmath201 for a range of densities @xmath135 .
therefore , if we restrict to ground states where the density is not globally inhomogeneous , the absence of the chemical shift directly implies @xmath0 symmetry . within this model , we can therefore define a experimental procedure to measure one of the most crucial parameter of the theory , namely @xmath184 . using the experimental arpes data for lsco system we would identify @xmath184 to be approximately the same as the optimal doping density .
we recall that @xmath184 is also a measure of the degree of the quantum spin fluctuation in the system . for the bi - layer materials such as ybco and bisco superconductors , the quantum spin fluctuation are stronger due to the inter - layer spin exchange , and we would predict that @xmath184 should be less than the optimal doping value .
having discussed the ground state properties and the phase diagram of the model , we are now in a position to study the collective excitations of the model . we have argued that the ground state energy can remain @xmath0 symmetric despite the projection .
however , the projection does affect the collective excitation spectrum near half - filling .
nonetheless , as we shall see , there remains a unique signature of the projected @xmath0 symmetry in the collective excitation spectra . in principle
, the collective excitation spectra can be obtained straightforwardly by studying the quadratic fluctuations around the mean field minima .
the resulting quadratic boson hamiltonian can be simply diagonalized .
the main complication in the procedure is the hard - core boson constraint , which requires @xmath204 for every site .
there are several ways to implement this constraint rigorously .
one is to follow the mapping from the one - component hard - core boson model to the @xmath205 model and generalize it to a multi - component hard - core boson model .
one could also convert the above inequality constraint to an equality constraint by introducing a boson creation and annihilation operator for the singlet state .
this approach will be implemented in the appendix . for simplicity of presentation , here we shall adopt a less rigorous approach and introduce a on - site boson repulsion term @xmath206 to our hamiltonian ( [ hamiltonian ] ) and convert the hard - core constraint to a soft - core constraint .
we shall show later that all results can be expressed in terms of the order parameter , which is implicitly dependent on @xmath207 , but there is no explicit dependence on @xmath207 .
we have verified that all three methods give the same long wave length spectra for the collective modes in the limit of low boson density . in the next section
, we shall present another calculation based on the continuum effective lagrangian method , which also reproduces the same spectra . to simplify presentation
, we shall concentrate on the case where the ground state energy functional ( [ energy ] ) is @xmath0 symmetric , _
i.e. _ for coupling constants @xmath208 and @xmath209 .
we choose the direction of spontaneously broken symmetry to be @xmath210 and @xmath211 .
the extremal condition can be easily determined to be @xmath212 the combination @xmath213 expresses the fact that the classical minimum is @xmath0 symmetric .
we can therefore write @xmath214 and @xmath215 . expanding the boson operators as : @xmath216 we obtain the following quadratic hamiltonian @xmath217 we are in particular interested in the collective mode spectra for the af insulating state with @xmath126 , the mixed states with @xmath163 and the sc state with @xmath127 and how they connect to each other . from this quadratic hamiltonian we can learn a number of important features .
first we notice that the @xmath218 and @xmath219 modes are decoupled for all ranges of @xmath220 , but most importantly , their dispersion relations are independent of @xmath125 and given by @xmath221 where @xmath222 and @xmath223 is the lattice constant .
this is indeed a very remarkable property . at @xmath126 , @xmath218 and @xmath219
modes are nothing but the transverse af spin wave modes .
af spin waves are usually viewed as goldstone modes and their existence is due to the af long range order .
however , as @xmath125 changes continuously from @xmath168 to @xmath169 , the af long range order continuously diminishes until it vanishes at @xmath127 .
the reason that the properties of the @xmath218 and @xmath219 modes do not change at all is due to the @xmath0 symmetry of this model , since the diminishing af order is compensated by the increasing sc order as the superspin angle @xmath125 is varied
. as we shall see , the @xmath224 mode is the af spin amplitude mode at @xmath126 , but it becomes massless and degenerate with the @xmath218 and @xmath219 modes at @xmath127 .
these three modes form a massless @xmath225 triplet mode whose existence is purely a consequence of the sc order . therefore , as @xmath125 is continuously varied from @xmath168 to @xmath169
, the transverse af spin wave modes gradually change their character to become the @xmath225 triplet resonance of the sc state .
as we shall see , for @xmath172 , the @xmath225 triplet mode becomes massive . at the af point
@xmath126 , the spin amplitude mode @xmath224 is decoupled from the sc mode @xmath226 and can be diagonalized separately .
the dispersion for the spin amplitude mode has the conventional massive relativistic form for small @xmath227 @xmath228 on the other hand , we have a massless sc goldstone mode @xmath226 with the following dispersion , @xmath229 this mode is an important new prediction of the @xmath0 theory .
it is the counterpart of the @xmath225 resonance in the af state . in the unprojected @xmath0 model ,
there are two such modes , with charge @xmath7 , and they represent gapless fluctuations from af to sc at half - filling . in the projected @xmath0 model , the charge @xmath9 mode is projected out of the spectrum , however , the charge @xmath11 mode remain massless at @xmath136 .
it is also a manifestation of the gapless fluctuation from af to sc at half - filling , but the sc fluctuation is hole - like , rather than both hole and particle - like as in the unprojected case .
we see again that a large mott - hubbard gap is fully compatible with gapless sc fluctuation at half - filling .
experimental detection of this mode could provide a important test of the projected @xmath0 symmetry .
the @xmath224 and the @xmath226 modes also decouple in the pure sc state with @xmath127 .
however , their physical interpretation change .
the @xmath224 mode becomes gapless at this point with the same dispersion as in ( [ spin_wave ] ) .
therefore , the three modes @xmath224 , @xmath218 and @xmath219 form a gapless @xmath225 triplet mode of the pure sc state , and represent the gapless fluctuation from sc to af at @xmath136 , but with a finite hole density @xmath184 given in ( [ rho_c ] ) .
the dispersion for the @xmath226 mode is given by @xmath230 and has the natural interpretation of a linearly dispersing phase goldstone mode of the sc state .
in the intermediate mixed state with @xmath163 , the @xmath224 and the the @xmath226 modes are coupled .
diagonalization of these modes gives : @xmath231 the upper massive mode has predominantly spin amplitude character , and we see that the gap diminishes continuously until it reaches zero at @xmath127 to become the massless @xmath225 triplet mode .
the lower mode has predominantly sc fluctuation character , and has a gapless @xmath232 dispersion . in the mixed region where both the af and the sc order parameters are non - zero
, one would naturally expect a gapless phase mode corresponding to the sc order .
however , in a interacting boson system , the phase mode is expected to have linear dispersion on general ground .
therefore , what is surpassing and new here is not the gapless nature of the sc mode , but its _
quadratic dispersion_. in order to locate the origin of the quadratic dispersion , we have perturbed the model away from the projected @xmath0 symmetric point so that a uniform mixed state is stabilized as a classical minimum .
sc fluctuation around such a non-@xmath0 symmetric point is gapless and has linear dispersion .
a quadratic dispersion is only realized at the @xmath0 symmetric point .
therefore , the quadratic dispersion is a unique signature of the projected @xmath0 symmetry in the entire range of densities @xmath135 ! to understand the physical origin of this remarkable phenomenon , we notice that a boson system with gapless quadratic dispersion generally has infinite compressibility .
this can be directly seen from the compressibility sum rule @xmath233 where @xmath234 is the dynamical density correlation function .
because of the quadratic dispersion relation , we can see explicitly that @xmath235 for small @xmath227 , therefore , a infinite compressibility is obtained . on the other hand ,
a infinite compressibility implies that @xmath236 , _
i.e. _ the chemical potential is independent of doping .
but this is exactly the prediction of the projected @xmath0 model ! for @xmath135 , the chemical potential is pinned at the @xmath0 symmetric point @xmath136 , where the superspin vector can point in any direction . to accommodate a long wave length fluctuation of the hole density
, the system rotates into another degenerate minimum with a different superspin angle @xmath125 and a different hole density . for this reason ,
the chemical potential does not change and the system is infinitely compressible .
at the mean - field level , the ground - state energy of the hamiltonian ( [ hamiltonian ] ) with ( [ soft - core ] ) depends on the af and sc order parameters @xmath45 and @xmath148 only via their combination @xmath237 reflecting the so(5 ) invariance of the mean - field result .
however , the zero - point energy of the bosons in the quadratic hamiltonian gives a correction to the ground - state energy due to quantum fluctuations .
( calculations of the quantum fluctuation effects at the @xmath238 spin - flop transition have been studied in ref .
as we will show below , this correction turns out to depend on @xmath239 and @xmath240 separately . for the soft - constraint case
, @xmath241 can be expressed as @xmath242 where @xmath243 are the four collective modes described in sec .
[ modes ] ( although extended to all values of @xmath227 in the bz ) . in a systematic @xmath137 expansion , this correction to the ground state energy scales like @xmath137 , and
is therefore small in the semi - classical limit .
alternatively , @xmath241 can be seen as a small correction to the mean - field energy @xmath244 for small values of the parameter @xmath245 . however , contrary to @xmath244 , @xmath241 also depends on the superspin angle @xmath125 and thus produces a small so(5)-symmetry breaking .
we now evaluate the @xmath125-dependent part of eq .
( [ egs ] ) in the small-@xmath246 limit .
we thus parametrize @xmath247 ( @xmath248 ) , differentiate the expression ( [ egs ] ) with respect to @xmath249 , and expand it up to second order in @xmath246 . by further transforming @xmath250 with @xmath251 , the derivative of @xmath241
can be expressed as the sum of two terms @xmath252 where @xmath253 with @xmath254 @xmath255 and the density of states @xmath256 is defined as @xmath257 \;.\ ] ] unfortunately , the terms of the @xmath246 expansion in @xmath258 diverge when integrated over @xmath259 .
it is thus convenient to carry out the transformation @xmath260 in @xmath258 and _ then _ expand in powers of @xmath246 .
we obtain @xmath261 for small @xmath246 , the integral in @xmath83 can be extended to @xmath262 and we obtain from eq .
( [ intg ] ) @xmath263 while in evaluating eq .
( [ e1bf ] ) we only need the density of states at @xmath264 , @xmath265 , for the first term @xmath266 one needs @xmath256 in the whole domain @xmath267 however , @xmath266 is independent of @xmath249 and thus it merely fixes the value of the critical chemical potential . a numerical integration yields @xmath268 by integrating over @xmath249 eqs .
( [ e1af ] ) and ( [ e1bf ] ) , we finally obtain the total contribution to the ground - state energy correction .
@xmath269 @xmath241 thus lifts the degeneracy as a function of @xmath249 and initially favors the pure superconducting phase ( @xmath270 ) .
a small chemical potential term @xmath271 with @xmath272 restores the degeneracy between the pure superconducting ( @xmath270 ) and the pure antiferromagnetic ( @xmath273 ) phases .
however , due to the convexity of @xmath241 as a function of @xmath249 there is always a barrier ( @xmath274 ) between the two phases since the mixed phase always has a higher energy .
this means that at @xmath275 and for intermediate densities the system prefers to phase separate between the two pure phases rather than choosing the mixed phase .
however , the important point is that this barrier remains small in the limit @xmath276 . in view of the symmetry breaking effects of the quantum fluctuations
, it would be interesting to see whether there is a limit when the the wave function ( [ variational ] ) and the projected @xmath0 symmetry becomes exact .
rokhsar and kotliar@xcite have shown that this type of wave functions are actually exact in the limit of infinite dimensions . therefore , besides the @xmath137 expansion , we could also use a @xmath277 expansion ( where @xmath47 is the space dimension ) to systematically control the @xmath0 symmetry breaking quantum effects . besides quantum fluctuations , there are also other symmetry breaking terms .
nearest - neighbor and next - nearest - neighbor coulomb interactions also break the @xmath0 symmetry , however , their corrections to the ground state is concave , _
i.e. _ the energy of the intermediate states are lowered .
therefore , they can also lead to uniform mixed states in some region of the phase diagram .
the detailed study of all these competing effects will be carried out in subsequent works .
while the projected @xmath0 model defined on a lattice enables us to make some connection to the underlying microscopic physics , for most discussions concerning the long wave length and low energy degrees of freedom , it is desirable to have a effective continuum lagrangian .
such a formulation can be directly obtained by taking the long wave length limit of the projected bosonic model discussed previously . however , in order to make the connection to the unprojected @xmath0 model clearer , we shall motivate our discussion from the original @xmath0 effective model .
the effective lagrangian for a fully @xmath0 symmetric model takes the form of @xmath278 where @xmath249 measures the superspin susceptibility , @xmath279 measures the superspin stiffness , @xmath280 denotes the spatial directions and @xmath281 is a scalar function of the superspin magnitude @xmath282 only .
there are three important symmetry breaking effects connected with the presence of a large mott - hubbard gap .
first is an asymmetry in the scalar potential , which can be described by a additional term @xmath283 which for positive @xmath284 favors af at half - filling .
second is the asymmetry between the spin ( @xmath285 ) and the charge ( @xmath286 ) susceptibilities , which modifies the kinetic energy to @xmath287 the last symmetry breaking effect is due to the chemical potential @xmath288 , which enters the lagrangian as a gauge coupling in the time direction , and modifies the charge part of the kinetic energy to @xmath289 combining these three symmetry breaking terms , we obtain @xmath290 in the presence of a large mott - hubbard gap , all these three symmetry breaking terms are of the order of @xmath4 , _
i.e. _ @xmath291 .
therefore , this lagrangian contains high energy degrees of freedom of the order of @xmath4 .
however , as already observed in @xcite , at @xmath292 , their effects cancel completely in the time - independent part of the lagrangian , and the static potential is @xmath0 symmetric just as in the original unprojected model .
we also observe that near the af / sc transition point where @xmath293 , the first order time derivative term is of the order of one .
furthermore , in the spirit of the low frequency and wave vector expansion , we only need to retain the first order time derivative term in the charge sector and can drop the second term in the above lagrangian . combining all these considerations , we obtain the following low energy effective lagrangian near the af / sc transition region , which is free of any parameters of the order of @xmath4 : @xmath294 this is exactly the lagrangian counterpart of the @xmath0 projection procedure discussed previously in the hamiltonian language . dropping the second order time derivative terms
removes half of the ( high energy ) degrees of freedom , and redefines the canonical conjugacy of the dynamical variables .
in particular , the conjugate variable of @xmath25 is nothing but @xmath26 itself , since @xmath295 .
standard quantization procedure requires the canonical commutation relation @xmath296=i$ ] , which in this case just reproduces equation ( [ non - commute ] ) .
this confirms the fact that the @xmath0 projection does not change the form of the interaction potential , only the commutation relation between @xmath25 and @xmath26 .
it is easy to see that the low energy effective lagrangian ( [ projected_l ] ) produces exactly the same long wave length collective mode spectrum as the projected @xmath0 hamiltonian ( [ hamiltonian ] ) defined on a lattice . to facilitate the comparison , we take the @xmath0 potential to be @xmath297 assuming broken symmetry in the @xmath25 and @xmath298 directions , we find that the @xmath299 and @xmath300 modes always decouple , and they have a linear spin wave dispersion relation with @xmath301 . the euler - lagrangian equations of motion gives the following dispersion relation for the @xmath25 and @xmath298 modes : @xmath302 for the af state with @xmath303 , @xmath304 for the sc state with @xmath305 , and @xmath306 for the mixed state with @xmath307 and @xmath308 .
these dispersion relations agree exactly with the lattice model results at the projected @xmath0 symmetric point if we make the following identification @xmath309 the effective lagrangian can be easily used to discuss effects of @xmath0 symmetry breaking .
the simplest form of symmetry breaking is increasing the chemical potential beyond the critical value @xmath310 , so that a pure sc state is realized .
the chemical potential enters the effective lagrangian through the gauge coupling in the time direction via the following substitution @xmath311 where @xmath312 is the deviation of the chemical potential away from the critical value . in this case , the spin triplet excitations acquire a finite mass gap , with the following dispersion relation : @xmath313 and the mass gap increases with increasing doping in the sc state .
we summarize the behavior of the collective modes obtained in the previous two sections in fig .
( [ fig5 ] ) .
= 6.0 cm we see that while there are significant modifications of the collective mode spectra in the density regime @xmath135 from the unprojected @xmath0 symmetry , the spectra beyond @xmath184 is essentially identical to the behavior expected from the unprojected @xmath0 symmetry .
this should be expected from our general considerations about the gutzwiller projection without much detailed calculations .
we argued that the only effect of the gutzwiller projection is to change the quantum commutation relation between the sc components of the superspin @xmath25 and @xmath26 .
however , for @xmath173 , the system is in a pure sc phase where these components acquire classical expectation values . in this case
, the modification of the quantum commutation relation does not have any significant effect .
this argument can also be illustrated by a simple picture of a _ chiral _ @xmath0 sphere , as depicted in fig .
( [ fig6 ] ) .
= 8.0 cm in this picture , the north and south poles represent the three af directions , and the equatorial plane represent the sc directions .
the sphere is perfectly @xmath0 symmetric .
however , the chemical potential along the pole direction acts like a fictitious magnetic field which restricts the sense of the rotation in the sc plane .
small oscillations of a vector pointing close to the north pole enclose the fictitious magnetic flux , and can only execute chiral rotations .
this amounts to the projection of the particle - pair states at half - filling . on the other hand , small oscillations of a vector
pointing anywhere along the equator does not enclose the fictitious magnetic flux , and their dynamics is therefore unaffected by the projection .
dynamics of a vector pointing anywhere between the north pole and the equator is also partially affected by the projection , but the symmetry of the static potential bears a unique signature .
the main purpose of this paper is to introduce the concept of projected @xmath0 models and discuss the properties of this model in connection with high @xmath1 superconductivity .
the projected @xmath0 model describes the low energy and long distance bosonic degrees of the freedom near the af / sc transition .
we showed that the gutzwiller projection can be implemented analytically on every site in the @xmath0 theory . in the presence of a infinite mott - hubbard gap ,
we show that static properties of the model can remain @xmath0 symmetric , while the modification of the dynamics can be completely cast into a non - trivial commutation relation between the two sc components of the @xmath0 superspin . unlike the unprojected @xmath0 models which can only have the full dynamic @xmath0 symmetry at half - filling , the projected @xmath0 model can have static @xmath0 symmetry at a critical value of the chemical potential @xmath310 and for a finite range of doping @xmath314 . at @xmath186 ,
the system has a af ground state and zero compressibility . in the intermediate regime @xmath135 ,
the system has mixed af and sc order and infinite compressibility . for @xmath173 ,
the system has a pure sc ground state , the sc order parameter rises to a maximal value before it decreases with doping . at the projected @xmath0 symmetric point
, we can understand precisely the evolution of the collective modes . on the af side
, we have two gapless spin wave modes and a gapless charge mode describing the gapless fluctuation from af to sc . in the intermediate density regime @xmath135 ,
the physical properties of the spin waves remain unchanged , while the massive spin amplitude mode gradually decreases its energy and merges with the two spin wave modes at @xmath315 .
the charge mode in the intermediate density regime is gapless , but has quadratic dispersion relation , which is a unique signature of the projected @xmath0 symmetry . for @xmath173 ,
the charge mode is gapless with linear dispersion relation , and the @xmath225 triplet spin mode becomes massive , and gradually increases its energy with increasing doping .
in this regime , the behavior of the collective modes are identical to the unprojected @xmath0 model .
this very simple model can form the basis to understand many novel and puzzling properties of the high @xmath1 superconductors in a unified framework .
it points out a route from af to sc through a gradual rotation of the superspin angle . at the projected @xmath0
symmetry point the mean field energy is independent of the superspin angle , and therefore it offers an explanation of the absence of the chemical potential shift in the underdoped regime without global phase separation .
it predicts a phase diagram which is qualitatively consistent with the observed phase diagram in the high @xmath1 materials . in the underdoped regime of the phase diagram ,
the system have large af and sc fluctuations , and these fluctuations can be responsible for the pseudogap physics observed in these materials .
there are many possible directions to carry out this line of research in the future .
the most important issue is to understand the precise nature of the intermediate state in the regime @xmath135 .
since the system has infinite compressibility in this regime , different small perturbation may select different ground states .
such perturbing effects might include quantum fluctuations and longer ranged interactions .
in particular , we would like to investigate the possibility that these perturbations might lead to the formation of incommensurate order or stripes .
in this work , we have discussed extensively the collective fluctuations in the long wave length limit . due to the definitions of our effective lattice model
, the @xmath316 limit corresponds to the @xmath316 limit in the sc correlation functions and the @xmath317 limit of the af spin correlation functions . within the @xmath0 theory ,
the @xmath225 resonance in the sc state is viewed as the @xmath0 symmetry partner of the @xmath316 goldstone mode of the sc phase fluctuation .
while the commensurate neutron resonance mode is observed in both ybco and bisco superconductors , all high @xmath1 systems also have incommensurate spin fluctuations .
how can these features be explained within the current theoretical model ?
the fact that lsco and ybco have very different fermi surface shapes and yet have similar incommensurate magnetic peaks strongly suggests that the incommensurate peaks are not sensitive to fermi surface effects and should be explainable within an effective bosonic model .
let us recall that the collective mode of a superfluid boson system consists of linearly dispersing phonon branch and another roton branch with a minimum located at the inverse inter - particle spacing .
so far , we have only studied the phonon branch of the charged bosons .
by analogy , the roton branch should also exist , with a wave vector determined by the density of the charged bosons or doping . within the @xmath0 theory , while the commensurate neutron resonance can be viewed as the @xmath0 partner of the sc phase mode , the incommensurate magnetic peaks can be viewed as the @xmath0 partner of the roton minimum of the charged bosons .
a detailed quantitative analysis of this picture will be carried out in the future .
however , while the ground state in the doping range @xmath135 may depend sensitively on small perturbation effects , at finite temperature , these perturbation effects should be small and the system should display more universal properties .
we have shown that the projected @xmath0 symmetry should be valid for the entire doping range @xmath135 , and we shall quantitatively study the manifestation of this symmetry at finite temperature , and see if the projected @xmath0 symmetry can give a universal explanation of the pseudogap physics . _
note added : _ after completing this work , we received a very interesting paper by coen van duin@xcite , in which he also observed the remnant @xmath0 behavior in the large @xmath4 limit " .
we would like to acknowledge useful discussions with d. arovas , j. berlinsky , e. demler , r. eder , c. kallin , s. kivelson and d. scalapino .
scz and jph are supported by the nsf under grant numbers dmr-9814289 .
wh and ea are supported by forsupra ii , bmbf ( 05 605 wwa 6 ) and by the deutsche forschungsgemeinschaft ( ar 324/1 - 1 ) and ( ha 1537/17 - 1 ) .
aa is supported by the israel science foundation .
aa , ea and wh would like to acknowledge the support and hospitality of the stanford physics department , where most of this work was carried out .
alternatively , one can enforce the hard - core constraint ( [ hard - core ] ) by introducing an additional `` slave '' boson for each lattice site .
the presence of this boson ( @xmath318 ) indicates that the lattice site @xmath45 is in the singlet state . the `` less or equal '' hard - core condition ( [ hard - core ] )
is replaced with the equality constraint @xmath319 with @xmath320 . since in physical states
one always has one and only one boson per lattice sites , destruction ( creation ) of a boson @xmath321 ( @xmath322 ) must always be accompanied by creation ( destruction ) of the empty boson @xmath323 ( @xmath324 ) . in this way
, the _ physical _ operators for creating ( destroying ) a triplet ( @xmath325 ) or a hole pair ( @xmath326 ) acquire the form @xmath327 ( @xmath328 ) .
the advantage of this method is that the constraint can be enforced exactly ( at least in principle ) by introducing an additional time - independent field @xmath329 at each lattice site , by adding to the hamiltonian a term @xmath330 and by integrating over the @xmath329 on the imaginary axis .
the whole hamiltonian ( [ hamiltonian ] ) thus takes the form ( apart for a constant ) @xmath331 in practice , one starts with a mean - field approximation and expands the boson operators around their mean - field values as in sec .
[ modes ] .
this expansion can be rigorously controlled by generalizing the constraint ( [ constraint ] ) to large values of @xmath332 , whereby one scales @xmath333 ( cf .
@xcite ) . physically
, this corresponds to allow for a large number of bosons to be present at each site and thus to have a large value for the total spin , or more precisely for the so(5 ) quantum number @xmath133 , at each site .
the mean - field result thus corresponds to the @xmath334 limit , while the quadratic expansion corresponds to the first @xmath335 correction . at the mean - field level ,
the constraint is fulfilled exactly and indeed one obtains the same result and the same phase diagram as the variational ansatz ( [ variational ] ) discussed in sec .
[ phase ] . by expanding the bosons quadratically around the mean - field
one obtains the same modes as for the soft - core hamiltonian with a similar dispersion .
( here , we restrict again to the so(5)-symmetric case ) .
specifically , in the mixed phase one obtains two spin - wave modes with dispersion @xmath336 one massive spin - amplitude mode @xmath337 and a quadratic mode @xmath338 in the pure superconducting phase , one has a @xmath225-triplet with dispersion @xmath339 and the sc goldstone mode with dispersion @xmath340 these results coincide with the ones of the soft - constraint approximation eqs .
( [ spin_wave ] ) to ( [ mixed_modes ] ) in the limit of small @xmath341 , i. e. , for low boson density . | we construct a class of projected @xmath0 models where the gutzwiller constraint of no - double - occupancy is implemented exactly .
we introduce the concept of projected @xmath0 symmetry where all static correlation functions are exactly @xmath0 symmetric and discuss the signature of the projected @xmath0 symmetry in dynamical correlation functions .
we show that this class of projected @xmath0 models can give a realistic description of the global phase diagram of the high @xmath1 superconductors and account for many of their physical properties . |
circa april 2010 , more than 70 exoplanets had been discovered transiting their primary stars .
transits are useful in the characterization of exoplanets because they break the the mass / orbital - inclination degeneracy of radial - velocity measurements .
moreover , a planet s radius can be directly measured , yielding mass ( ) and radius ( ) correlations with which theorists can extract useful information about bulk composition and structure and can attempt to fit radius evolution models ( guillot et al .
1996 ; burrows et al . 2000 ; brown et al . 2001 ; hubbard et al . 2001 ; baraffe et al .
2003 ; chabrier et al . 2004
; charbonneau et al . 2007 ) . however , since atmospheric opacity is a function of wavelength , a planet s transit radius is also a function of wavelength ( seager & sasselov 2000 ; brown 2001 ; spiegel et al . 2007 ;
fortney et al .
2003,2010 ) .
the variation with wavelength in the measured radius ( actually an impact parameter ) provides an ersatz spectrum directly related to the planet s atmospheric composition near the terminator .
this radius spectrum , unlike the planet s spectrum at secondary eclipse , is more dependent upon composition than upon the temperature profile , and so provides complementary information to a planet s own direct emissions .
the latter can be measured during secondary eclipse , but also during the traverse by the planet of its orbit as it traces out its phase light curve .
the transit radius spectrum and the planet emission spectrum as a function of phase can , therefore , together help constrain planet properties ( fortney et al .
2006,2010 ) .
however , the stellar irradiation of such an exoplanet severely breaks what would otherwise be quasi - spherical symmetry , producing large day - night contrasts in thermal structure , zonal flows and banding , and violent atmospheric dynamics ( showman & guillot 2002 ; guillot & showman 2002 ; cho et al .
2003,2008 ; burkert et al .
2005 ; cooper & showman 2005,2006 ; showman et al . 2008,2009 ; showman , cho , & menou 2010 ; dobbs - dixon & lin 2008 ; dobbs - dixon , cumming , & lin 2010 ; rauscher et al . 2007,2008
; menou & rauscher 2009 ; rauscher & menou 2010 ) . importantly , such behavior may have photometric and spectral signatures .
it is the exploration of such signatures that motivates this paper .
the state - of - the - art in the 3d modeling of exoplanet atmospheres is still evolving and has not yet reached a level of maturity where detailed predictions for each known exoplanet are robust .
hence , in this paper we focus on a few generic ideas and conjectures that emerge from our modeling efforts .
the procedure we have pursued is the following : first , using a 3d general circulation model ( gcm ) we derive dynamic model atmospheres at various epochs after they have achieved a steady state .
second , we post - process these 3d model atmospheres with a spectral atmosphere code to obtain transit radius spectra at various epochs during ingress , egress , and total transit .
we assume equilibrium molecular compositions and solar metallicity .
third , using the same 3d model atmosphere , we integrate over the visible " disk and calculate phase light curves as a function of wavelength and for various standard photometric bands .
the results are generic predictions which , though not expected to be quantitatively precise and constraining on any particular giant exoplanet , nevertheless contain qualitatively interesting features that should inform future measurements . for specificity ,
we focus on two models of hd 209458b ( charbonneau et al . 2000 ; henry et al . 2000 ) , one with and one without an extra absorber " and a thermal inversion ( hubeny , burrows , & sudarsky 2003 ; burrows et al . 2007 ; fortney et sl .
2008 ; spiegel et al . 2009 ; knutson et al .
2008 ) , but suggest our qualitative results are generic beyond this planet and this modeling paradigm ( see also fortney et al .
2006,2010 ) . in [ techniques
] , we summarize our methodology and techniques . then , in [ describe ] we describe the 3d models and various of their salient characteristics .
we go on in [ wave_rad ] to present our results for the wavelength - dependent transit radius and ingress - egress asymmetries and in [ phase ] we turn to a full discussion of the derived light curves .
this section contains not only wavelength - dependent planet - star flux ratios as a function of phase , but the phase variation of several photometric planet band fluxes . in [ conclusions ] , we summarize our general conclusions .
we use the same modified version of the university of reading s intermediate general circulation model ( hoskins & simmons 1975 ) as presented in menou & rauscher ( 2009 ) and rauscher & menou ( 2010 ) , where a detailed description of the numerical implementation can be found .
we use a horizontal spectral resolution of t31 , roughly equivalent to a 4@xmath4 resolution on the sphere , and 45 vertical levels which are logarithmically spaced in pressure from 100 bar to 10 microbar .
we employ a simplified newtonian relaxation scheme for the radiative forcing , so that everywhere the atmosphere heats or cools toward a prescribed three - dimensional equilibrium temperature profile , on some representative radiative timescale ( iro , bzard , & guillot 2005 ) .
the equilibrium temperature profiles are chosen to have a horizontal dependence such that on the dayside the fourth power of the equilibrium temperature goes linearly with the cosine of the angle from the substellar point and on the nightside it is constant . the amplitude of the temperature difference between the substellar point and the nightside is taken from 1d radiative transfer models ( burrows et al .
2008 ; see fig .
[ fig1 ] ) , as are the dependences with pressure and depth of the equilibrium temperatures needed in the newtonian scheme .
our choice for the angular dependence of the dayside equilibrium temperatures is loosely motivated by the fact that the equilibrium black body temperature ( @xmath5 ) of an irradiated surface at a slant angle is determined by setting the incident flux on the angled surface equal to the emitted flux and by the @xmath6 dependence of the latter .
we scale the entire vertical profile of internal equilibrium temperatures to which the newtonian cooling scheme is dynamically driving the actual temperature by this same angular factor . in this way , the expectation that the annuli on the planet away from the substellar point will be heated less by the star than the gas at the substellar point is realized . for the night side , which by definition is not irradiated by the star , we employ the simplifying assumption that in equilibrium without night - side dynamics the planet would resemble more an isolated brown dwarf with uniform emission and profiles .
note that these assumptions about the solid - angle distribution of the equilibrium temperature profiles still allow the 3d gcm dynamics to redistribute heat in all regions of the planet s atmosphere and that the actual temperatures perforce deviate from these equilibrium temperatures due to such advection , sometimes to a significant degree .
however , clearly our ansatz could be improved by doing the 3d radiative transfer that the problem will eventually demand . for
now , we believe our approach allows us to capture the essence , if not the detail , of the thermal character of the dynamical atmospheres we are studying . thus , the only difference between the inputs for our two 3d gcm models of hd 209458b is the different 1d profiles of the substellar equilibrium temperatures , where one was calculated including an extra absorber high in the atmosphere and one was calculated without .
the radiative timescales are constant on a given pressure level and they increase with pressure at depth . they are taken from the profile in iro et al .
( 2005 ) , except for the deepest model levels .
figure [ fig1 ] provides a detailed profile of the radiative times adopted .
all other planetary and numerical parameters match those described in rauscher & menou ( 2010 ) .
each model was run for 500 planetary days ( @xmath7 orbits ) , after which a statistically steady state has been reached for all levels at pressures less than 1 bar
. levels deeper than this are not directly probed by the types of observations considered here . to summarize , the key differences between the present models and those in rauscher & menou ( 2010 ) are ( i ) the different relaxation temperature profiles used and ( ii ) the greater vertical extent of pressures modeled , up to 10 microbar in the present study .
to calculate the spectrum of the transit radius , we create a data cube in latitude , longitude , and altitude of the temperatures , densities , and pressures of one of the 3d gcm models described in
[ gcm ] . at a given wavelength ,
using the snapshot at day 500 in a 500-day model , we cast 200 rays through an annulus at a given impact parameter and calculate the optical depth ( @xmath8 ) along the chord .
the miniscule index of refraction effects ( fortney et al . 2003 ) are ignored .
we use the opacity database described in sharp & burrows ( 2007 ) and equilibrium chemical abundances taken from burrows & sharp ( 1999 ) and burrows et al .
solar abundances were assumed , and we have used the inclination angle of 86.6@xmath4 suitable for hd 209458b .
the differenital area of the annulus associated with a given ray is then weighted by the quantity @xmath9 .
a new impact parameter is then chosen and the process is repeated . by this means , the total effective area of the atmosphere at a given wavelength is obtained .
it is this effective area that determines the magnitude of the occultation by the planet of the star due to the atmosphere .
since there is no core nor structural information in the atmospheric gcm , we assume that the baseline planet radius is the radius measured for hd 209458b in the optical by knutson et al .
2007 ( 1.32 ) and shift the differential area for a given wavelength calculated as described above by the average of the corresponding calculation in the 0.5 to 0.7 wavelength region , rendering the average differential area in the 0.5 to 0.7 wavelength region zero .
other procedures can be followed , but all of them are equivalent to first order in the ( small ) differential area , where only the first - order term is relevant .
when the transit is partial ( during ingress and egress ) , we use the formalism of mandel & agol ( 2002 ) to determine planet - star intercept angles as a function of orbital phase , between which the calculations of the total relevant area of the atmosphere are performed . since the planet is assumed to be in synchronous rotation
, it will rotate with the orbit during transit .
this effect , which can amount to a 10 - 20@xmath4 turn and which slightly changes the regions of the atmosphere intercepted by the cast rays included in the optical depth calculation , is incorporated in the formalism .
we perform such calculations for 2000 frequencies from 0.4 to 30 and 200 phase angles from just before ingress to just after egress .
though we performed some transit calculations at other days ( e.g. , days 451 , 476 , 496 ) in the 500-day runs , we found little variation ( less than 0.5% ) in the results .
hence , we focused on only the day-500 3d gcm models for the two input models , one with an extra absorber and one without , as described in [ gcm ] .
note that such a small temporal variation justifies the use of a single snapshot for both these trasnit radius calculations and the phase light curves calculations described below . to obtain the total planetary photon fluxes as a function of wavelength
, we use the code cooltlusty ( hubeny 1988 ; hubeny & lanz 1995 ; burrows , sudarksy , & hubeny 2003 ; burrows et al . 2008 ) .
this code incorporates the same composition and opacity algorithms we employ to perform the transit spectrum calculations described in [ transit_method ] .
first , we divide the surface seen by the observer at a given phase angle into about 2000 patches .
then , we cast a ray from the observer to the patch and into the planet at the associated slant angle .
the densities , temperatures , and pressures at that slant angle are then used to calculate the instantaneous radiation field emerging from the patch ( without allowing the code to seek radiative equilibrium and ignoring the time - dependent term in the transport equation ) . that intensity is then multiplied by the projected area of the patch and all the patches seen by the observer at that phase angle
were added .
this is done for 1000 frequency points from 0.4 to 30 .
we also calculated the associated photometric colors in 24 bands , 9 of which we report here .
the stellar flux is multiplied by the angle of incidence on the patch with respect to the star and a kurucz ( 1994 ) model for hd 209458 is employed .
the stellar flux is necessary to handle the rayleigh scattered component and to obtain the planet / star flux ratio .
we remind the reader that the 3d gcm model ( rauscher & menou 2010 ) is calculated using newtonian cooling and not with radiative transfer , so there are inconsistencies in our general approach . however
, our focus here is on the character of generic results ( for instance for different photometric bands and given general characteristics of the 3d flow ) , and not on specific predictions for hd 209458b .
although the relaxation temperature profiles are very different for our two atmospheric models with and without a high - altitude absorber ( fig . [ fig1 ] ) , we nevertheless find that these models have rather similar flow patterns .
this is perhaps even more striking since these circulation patterns are also qualitatively similar to that described in rauscher & menou ( 2010 ) , which has a rather different temperature relaxation profile .
this suggests that the circulation pattern studied here is more strongly governed by the choice of the @xmath10-behaves - linearly - with-@xmath11 horizontal dependence than by the amplitude or the detailed vertical structure of the relaxation temperature profile adopted .
this supports the notion that the main qualitative results shown in the present study may be extended to other close - in giant planets .
while details differ , key flow features are common to both model atmospheres : transonic wind speeds , with peak wind speeds of 10 and 20 km s@xmath12 high up in the models with and without an extra absorber , respectively ; a super - rotating ( eastward ) equatorial jet that extends across many pressure scale heights ( see fig .
[ fig2 ] ) ; a pronounced chevron - shaped shock - like feature at a longitude of @xmath13135@xmath4 at intermediate pressure levels ; and large stationary vortices in each hemisphere that extend almost from the equator to the poles and are centered at a longitude of @xmath13 - 120@xmath4 ( rauscher & menou 2010 ) .
high in the atmosphere , where heat advection is not dominant , the day - night temperature difference is greater in the model run with an extra absorber because of the stronger forcing profile adopted . in both models ,
the horizontal temperature field is close to equilibrium , with a hot dayside and a cold nightside at pressures less than 1 mbar .
the hottest atmospheric region is advected eastward by @xmath1345@xmath4 at the 100 mbar level and by @xmath1390@xmath4 at the 1 bar level , while the temperature field is fairly well longitudinally homogenized at deeper pressure levels .
the general trend in these models is to have the hottest atmospheric regions advected farther away from the substellar point at deeper pressure levels , with potentially observable consequences as discussed in
[ func_band ] . for the model with the extra absorber
, however , there is an additional complication .
although both models show localized heating where a shock - like feature develops in the flow , this heating is significantly more pronounced in the model with absorber .
we attribute this difference to faster winds driven by a stronger day - night forcing in the model with absorber .
this , rather than direct heat advection from the dayside , determines the location of the hottest atmospheric region at pressure levels 15 - 45 mbar in the model with absorber . as a result
, the peak temperature region at these levels does not simply transition away from the substellar point gradually with depth , but it is significantly affected by a systematic shift to @xmath13135@xmath4 , where the shock - like feature is located .. figure [ fig3 ] depicts temperature maps with and without dynamical redistribution and with and without an extra optical absorber . to emphasize the contrasts ,
the atmospheric scale heights are exaggerated by a factor of ten .
the longitudinal flows distribute heat to the nightside , and thereby partially even out the scale height bulge differential between the dayside and nightside .
additionally , figs .
[ fig4 ] depict the distribution of methane ( ch@xmath14 ) at a pressure level of 5.7 millibars , if in chemical equilibrium , over the surface of our 3d gcm models of hd 209458b , with and without an extra high - altitude optical absorber .
if not in methane , carbon would reside in carbon monoxide in these models , which predominates on the day sides ( particularly , in this model set , for the model with the extra optical absorber ) . in the model without the extra optical absorber , we note the presence of the tongue of methane - depleted ( co - rich ) material dragged from the dayside to the nightside in the equatorial belt ( see also cooper & showman 2006 ) .
the degree of such advection is a function of vertical height / pressure level . in principle
this feature could be detected in a precision comparison of the ingress and egress transit spectra ( fortney et al .
2010 ) . however , we find , as did fortney et al .
( 2010 ) , that this differential effect can be quite small , less than one percent of the already small variation with wavelength expected for the transit radius .
be that as it may , these two figures partially represent the model context in which we have performed our post - processing exercise .
the top panel of fig . [ fig5 ] shows the normalized fractional area of the atmosphere occulting the star as a function of wavelength as the planet enters and leaves transit for the model without an extra optical absorber at altitude .
the fractions are all normalized to the instantaneous average value in the optical wavelength range from 0.5 to 0.7 microns .
hence , the average value of this fraction in that wavelength range ( optical " ) is zero . for this model ,
the base position in the atmosphere where the chord optical depth is of order unity for optical photons ( which defines the transit radius ) is deep inside the atmosphere ( near 0.1 bars ) .
the bottom panel of fig .
[ fig5 ] is the same as the top panel , but for the model with the extra absorber at altitude .
the action of the extra absorber is clearly seen in the optical region , and the normalization is as performed for the top panel .
however , for this model the chord optical depth is of order unity for a base pressure near a millibar , much higher in the atmosphere than for the model without the extra absorber .
this difference results in perceptible differences between the two classes of models .
water is ubiquitous in our planet model atmospheres and its features dominate these spectra . for our models ,
the egress and ingress are rather symmetrical , differing by no more than @xmath130.8% at any wavelength , due predominantly to slight differences in the scale heights near the respective terminators .
these small ingress / egress differences are most prominent in the water bands , but there are slight differential signatures in the methane and carbon monoxide bands at @xmath131.6 and @xmath133.3 and at @xmath132.3 and @xmath134.6 , respectively , as might be expected from the advection of gas across the terminator depicted in figs .
[ fig4 ] .
figure [ fig6 ] compares the transit radii during full transit as a function of wavelength for both models depicted in fig .
the normalization for both models is to the measured radius in the optical ( @xmath131.32 ; knutson et al . 2007 ) .
we believe this to be the most physically and observationally sensible normalization .
note that when so normalized the two models , which differ only in whether there is an extra optical absorber at altitude , predict very different radii , though the relative differences when comparing radii at different wavelengths outside the optical are similar .
therefore , comparing the radius as inferred from optical observations to that inferred from infrared observations is diagnostic of the presence of an extra absorber .
unfortunately , the infrared transit radius data are not consistent with either the extra - absorber model or the no - extra - absorber model . a comparison with the _ spitzer_/irac data from beaulieu et al .
( 2010 ) and the 24-micron mips data from richardson et al .
( 2006 ) suggests that neither model fits both sets of data simultaneously , but a major discrepancy is also the wide spread between irac 3/4 and irac 1/2 . the same discrepancy between model and data
was noted by fortney et al .
( 2010 ) for their models and is as yet unexplained .
our implementation of the extra absorber puts it throughout the upper atmosphere , at the lowest pressures . given this , and as fig .
[ fig6 ] demonstrates , the corresponding model does not evince much contrast in and out of the na - d doublet .
this is in contradiction with the observations of charbonneau et al .
( 2002 ) and may hint at an upper boundary to the extent of the optical absorber ( if it exists ) if the na - d data are to be accommodated .
the planet fluxes themselves , and their phase variation , should speak volumes about the atmospheres of hot jupiters , in particular their thermal and compositional profiles and the longitudinal dependence of their circulation regimes ( cowan & agol 2008 ) . using the methodologies described in [ spectral_method ] , we now present results for the planet - star flux ratios and photometric variations for our two models of hd 209458b as it traverses its orbit .
figure [ fig7 ] portrays the planet - star flux ratios versus wavelength , for the models with and without the extra optical absorber and as a function of phase in the planet s ( hd 209458b s ) orbit . here
, zero phase is at secondary eclipse .
superposed are various data sets and the corresponding
1d " spectral models using the method of burrows et al .
( 2007,2008 ) .
the slight differences between the zero - phase 3d calculations and the
1d " predictions are due to the different methods of determining the average effect of emission from the day - side hemisphere and to the dynamics inherent in 3d models .
we note many things about these curves .
first , the data points at secondary eclipse are much more consistent with the zero - phase prediction(s ) for the model with an extra absorber .
this was pointed out in burrows et al .
( 2007 ) and knutson et al .
second , the variation with phase from the dayside to the nightside is much larger for the model with an extra absorber and inversion .
this is a generic feature of predictions with strong optical absorption at altitude on the dayside .
third , there are interesting differences in the phase - dependence at different wavelengths .
the variations longward of @xmath1315 microns are generally largest .
the variation at @xmath1310 microns is muted .
the water feature near @xmath136.2 microns becomes relatively more prominent on the nightside , while modest on the dayside .
the phase variation near @xmath134 microns is slight .
the phase variation from @xmath138 to @xmath1310 microns is much larger for the hot day - side model with the extra absorber . in sum
, there are important variations in infrared colors with orbital phase that may be diagnostic of the atmospheere models .
the planet - star flux ratios depicted in figs .
[ fig7 ] are integrals over the surface of the planet of brightness maps such as are shown in figs .
the associated panels in fig [ fig8 ] are representative planet brightness maps in the irac 3 band , the @xmath15 band , and the @xmath1 band , the latter for the models with and without the extra absorber .
all the maps are at full phase ( secondary eclipse ) . in principle , these are what the planet hd 209458b would look like in glasses " for the various bands , given the various relative color maps .
the color mappings depict the magnitude difference between the brightest point on the planet and are different for each given rendering and band .
note that the hot spot in each band is shifted by a different angle relative to the substellar point , reflecting the different approximate
photospheric " pressures of the various bands and the variation in the degree and strength of zonal heat advection with depth ( showman et al .
in particular , the irac band fluxes , formed as they are at altitude , are shifted the least , by no more than @xmath1320@xmath4 but mostly close to @xmath130@xmath4 .
these differential phase shifts are useful diagnostics of the zonal flow regimes and models ( cowan & agol 2008 ) .
this is particularly clear from fig .
[ fig9 ] , which depicts for the two different atmosphere models and for our fiducial hd 209458b assumptions integrated relative light curves versus orbital phase in the @xmath16 , @xmath15 , @xmath1 , @xmath2 , @xmath3 , irac 1 , irac 2 , irac 3 , and irac 4 bands .
these data are derived by integrating maps such as are shown in figs .
the different phase shifts of the hot spots suggested in figs .
[ fig8 ] are manifest in these plots . while the shifts in the irac bands are smallest ( generally being formed at altitude at lower pressures ) , the shift in irac 1 being the largest among these ( @xmath1320@xmath4 for the no - absorber model and @xmath1310@xmath4 for the absorber model ) , for the model with no extra absorber the shifts in the @xmath15 and @xmath1 bands can be @xmath1340@xmath4 and @xmath1345@xmath4 , respectively .
for the absorber model , the shift in the @xmath1 band is comparable , while the shift in the @xmath15 band is only @xmath1310@xmath4 .
the fact that many of the curves are not flat on the far right of these plots near 180@xmath4 is another indication of the skews in the phase light curves and of the ( differential ) dynamical advection of heat by planetary zonal winds . ignoring the irac bands , the contrasts are generally largest for bands at the shortest wavelengths . in fig .
[ fig9 ] , the differences between the models with and without the high - altitude absorber are clear , with the brightness variations with phase being larger for the absorber model in the irac , @xmath16 , and @xmath15 bands .
the two models are comparable in the near - infrared bands .
the specific numbers derived are not as important as the general trends revealed by this representative modeling exercise . in principle , by measuring such phase curves in different bands , one could back out information on the thermal and wind profiles as a function of depth and longitude .
this is in part because the different bands are formed and have approximate photospheric positions at different pressures in the atmosphere .
such remote sensing of a hot jupiter atmosphere would enable a new level of scrutiny for exoplanets and may be possible using jwst .
using a 3d gcm with newtonian cooling and day- and nightside equilibrium temperature profiles based on two 1d spectral models ( with and without a thermal inversion and a hot upper atmosphere ) , we created dynamical models of the transiting giant exoplanet hd 209458b .
we post - processed these 3d model atmospheres with a detailed opacity code to obtain transit radius spectra at various epochs during the primary transit . then , using a spectral atmosphere code , we integrated over the face of the planet seen by an observer at various orbital phases and calculated light curves as a function of wavelength and for nine different photometric bands from @xmath16 ( @xmath130.6 ) through irac 4 ( @xmath138 ) .
the products of this study are generic predictions for the character of the expected phase variations both of a giant planet s transit spectrum and of its light curves . since our gcm employed newtonian cooling ( and not radiative transfer using opacities that corresponded with those used in the post - processing ) ,
the calculations are slightly inconsistent .
nevertheless , the results have interesting features that should inform future measurements .
we found that the temporal variations in the derived integral quantities for this model suite are small ( less than 1% ) and that the ingress / egress contrasts due to zonal flows , while also small ( @xmath17 % ) , are most manifest due to the scale - height asymmetries near the terminators and in the water bands .
while models we have generated with and without the methane and carbon monoxide bands indicate that ingress / egress differences in the co and ch@xmath14 bands might be diagnostic of both carbon chemistry and abundance asymmetries due to zonal flows near the terminators , the differential effects due to different carbon - species abundances at the different terminators are small ( less than a few tenths of a percent ) .
a similar conclusion was reached by fortney et al .
( 2010 ) , who nevertheless found a slightly larger quantitative effect .
since the transit radius of hd 209458b has been measured in the optical , it should not be allowed to differ when comparing models with and without an upper - atmosphere absorber and the predicted radii at other wavelengths must be determined relative to it . when this is done , the predicted radii in the near- and mid - infrared are very different ( by as much as 0.05 ) , though our current models do a poor job of fitting all the transit radius data for hd 209458b simultaneously ( e.g. , knutson et al . 2007 ; beaulieu et al . 2010
; richardson et al . 2006 ) .
the reason for this discrepancy is currently unknown .
we determined that the angular phase shifts of the brightness peaks relative to the orbital ephemeris are functions of photometric band .
the @xmath1 , @xmath2 , and @xmath3 bands are shifted most , by as much as @xmath1345@xmath4 , while the irac bands are shifted least .
this is because different bands are spectrally formed at different pressure depths and the zonal winds that advect heat are depth - dependent .
the irac band photospheres are generally at altitude , while those in the near infrared are at deeper pressures near and above @xmath130.1 bars .
therefore , the question of the magnitude of the downwind shift in the planetary hot spot " due to equatorial winds is nuanced , depending upon the waveband in which measurements are made .
this wavelength dependence of the phase shift in the brightness light curve , and the associated day / night contrasts , can in principle be used to constrain the circulation regimes of irradiated giant planets and to probe different pressure levels of a hot jupiter atmosphere . though our calculations focused on models of hd 209458b , similar calculations for other transiting hot jupiters in low - eccentricity orbits should yield transit spectra and light curves of a similar character . as the subject of comparative exoplanetology matures , and jwst comes online , such observational manifestations of global circulation
may well become possible for a wide range of irradiated planets .
our models have been constructed to help prepare the way , however imperfectly , for that era .
we acknowledge useful conversations with ian dobbs - dixon and brad hansen and ivan hubeny for his general support of the cooltlusty code .
ab wa supported by nasa grant nnx07ag80 g and under jpl / spitzer agreements 1328092 , 1348668 , and 1312647 . er was supported by a nasa graduate student research program fellowship , contract nnx08at35h , and km was supported by the spitzer space telescope program under contract jplcit 1366188 .
the authors are pleased to note that part of this work was performed while in residence at the kavli institute for theoretical physics , funded by the nsf through grant no .
phy05 - 51164 .
burrows , a. , sudarsky , d. & hubeny , i. 2006 , , 650 , 1140 burrows , a. , hubeny , i. , budaj , j. , knutson , h.a . , & charbonneau , d. 2007 , , 668 , l171 burrows , a. , budaj , j. , & hubeny , i. 2008 , , 678 , 1436 burrows , a. & orton , g. 2010 , in exoplanets , edited by s. seager , to be published in the spring of 2010 in the space science series of the university of arizona press ( tucson , az ) ( refereed ) , ( arxiv:0910.0248 ) charbonneau , d. , brown , t. m. , noyes , r. w. , & gilliland , r. l. 2002 , , 568 , 377 charbonneau , d. , brown , t.m . , burrows , a. , & laughlin , g. 2007 , in protostars and planets v , " ed . b. reipurth and d. jewitt , p. 701
( university of arizona press ) , astro - ph/0603376 knutson , h. , charbonneau , d. , noyes , r.w . ,
brown , t.m . , & gilliland , r.l .
2007 , , 655 , 564 knutson , h.a .
, charbonneau , d. , allen , l.e . , burrows , a. , & megeath , s.t .
2008 , , 673 , 526 kurucz , r. 1994 , _ kurucz cd - rom no .
19 _ , ( cambridge : smithsonian astrophysical observatory ) rauscher , e. , menou , k. , cho , j.y .- k . ,
seager , s. , & hansen , b.m.s .
2007 , , 662 , l115 , arxiv:0712.2242 rauscher , e. , menou , k. , cho , j.y .- k . , seager , s. , & hansen , b.m.s .
2008 , , 681 , 1646 rauscher , e. & menou , k. 2010 , , 714 , 1334 ( arxiv:0907.2692 ) richardson , l.j . , harrington , j. , seager , s. , & deming , d. 2006 , .
649 , 1043 seager , s. & sasselov , d. 2000 , , 537 , 916 sharp , c.m . & burrows , a. 2007 , , 168 , 140 showman , a.p . ,
cho , j. y .- k . , & menou , k. 2010 , in exoplanets , edited by s. seager , to be published in the spring of 2010 in the space science series of the university of arizona press ( tucson , az ) ( arxiv:0911.3170 ) showman , a.p . ,
cooper , c.s . ,
fortney , j.j .
, & marley , m.s .
2008 , , 682 , 559 ( arxiv:0802.0327 ) showman , a.p . ,
fortney , j.j . ,
lian , y. , marley , m.s . ,
freedman , r.s . ,
knutson , h.a . ,
charbonneau , d. 2009 , , 699 , 564 showman , a. p. & guillot , t. 2002 , , 385 , 166 , as a function of the latitude and pressure level in the atmosphere , for the model with an absorber ( left ) and the one without ( right ) . the solid yellow lines separate regions of postitive ( eastward ) flow from negative pattern is qualitatively similar in both models .
( also compare to figure 3 of rauscher & menou 2010.),title="fig : " ] , as a function of the latitude and pressure level in the atmosphere , for the model with an absorber ( left ) and the one without ( right ) .
the solid yellow lines separate regions of postitive ( eastward ) flow from negative pattern is qualitatively similar in both models .
( also compare to figure 3 of rauscher & menou 2010.),title="fig : " ] , @xmath18 , @xmath19 , and 3.6 bars , and the colors depict the local temperatures on each level .
red is hot and blue is cold . ,
title="fig:",width=336 ] , @xmath18 , @xmath19 , and 3.6 bars , and the colors depict the local temperatures on each level .
red is hot and blue is cold .
, title="fig:",width=384 ] , @xmath18 , @xmath19 , and 3.6 bars , and the colors depict the local temperatures on each level .
red is hot and blue is cold .
, title="fig:",width=336 ] , @xmath18 , @xmath19 , and 3.6 bars , and the colors depict the local temperatures on each level .
red is hot and blue is cold .
, title="fig:",width=384 ] ) over the surface of the 3d gcm models of hd 209458b we are using for this study at a pressure level of 5.7 millibars .
chemical equilibrium is assumed .
the left model is with and the right model is without the extra absorber .
the nightsides for both are on the right .
reds and yellows indicate low methane abundances and blues and greens represent high methane abundances . if not in methane
, carbon would reside in carbon monoxide in these models , which predominates on the daysides ( particularly , in this model set , for the model with the extra optical absorber ) .
color bars indicate the logarithm ( base ten ) of the methane mixing ratio .
the position of the terminator is clearly indicated with a black line .
note on the model without the extra optical absorber ( right ) the tongue of methane - depleted ( co - rich ) material dragged from the dayside to the nightside in the equatorial belt .
the degree of such advection is a function of vertical height / pressure level and is greater for the model without the extra absorber.,title="fig : " ] ) over the surface of the 3d gcm models of hd 209458b we are using for this study at a pressure level of 5.7 millibars .
chemical equilibrium is assumed .
the left model is with and the right model is without the extra absorber .
the nightsides for both are on the right .
reds and yellows indicate low methane abundances and blues and greens represent high methane abundances .
if not in methane , carbon would reside in carbon monoxide in these models , which predominates on the daysides ( particularly , in this model set , for the model with the extra optical absorber ) .
color bars indicate the logarithm ( base ten ) of the methane mixing ratio .
the position of the terminator is clearly indicated with a black line .
note on the model without the extra optical absorber ( right ) the tongue of methane - depleted ( co - rich ) material dragged from the dayside to the nightside in the equatorial belt .
the degree of such advection is a function of vertical height / pressure level and is greater for the model without the extra absorber.,title="fig : " ] band map for the same ; the bottom left panel is the @xmath1 band map for the model with the extra absorber ; and the bottom right panel is the corresponding @xmath1 band map for the model without the extra absorber .
all the maps are at full phase ( secondary eclipse ) .
the colors and color bars are in relative magnitudes , with the brightest regions rendered in yellow and the dimmest regions in blue .
see text for details.,title="fig : " ] band map for the same ; the bottom left panel is the @xmath1 band map for the model with the extra absorber ; and the bottom right panel is the corresponding @xmath1 band map for the model without the extra absorber .
all the maps are at full phase ( secondary eclipse ) .
the colors and color bars are in relative magnitudes , with the brightest regions rendered in yellow and the dimmest regions in blue .
see text for details.,title="fig : " ] band map for the same ; the bottom left panel is the @xmath1 band map for the model with the extra absorber ; and the bottom right panel is the corresponding @xmath1 band map for the model without the extra absorber .
all the maps are at full phase ( secondary eclipse ) .
the colors and color bars are in relative magnitudes , with the brightest regions rendered in yellow and the dimmest regions in blue .
see text for details.,title="fig : " ] band map for the same ; the bottom left panel is the @xmath1 band map for the model with the extra absorber ; and the bottom right panel is the corresponding @xmath1 band map for the model without the extra absorber .
all the maps are at full phase ( secondary eclipse ) . the colors and color bars are in relative magnitudes , with the brightest regions rendered in yellow and the dimmest regions in blue .
see text for details.,title="fig : " ] | using a 3d gcm , we create dynamical model atmospheres of a representative transiting giant exoplanet , hd 209458b .
we post - process these atmospheres with an opacity code to obtain transit radius spectra during the primary transit .
using a spectral atmosphere code , we integrate over the face of the planet seen by an observer at various orbital phases and calculate light curves as a function of wavelength and for different photometric bands .
the products of this study are generic predictions for the phase variations of a zero - eccentricity giant planet s transit spectrum and of its light curves .
we find that for these models the temporal variations in all quantities and the ingress / egress contrasts in the transit radii are small ( @xmath0% ) .
moreover , we determine that the day / night contrasts and phase shifts of the brightness peaks relative to the ephemeris are functions of photometric band .
the @xmath1 , @xmath2 , and @xmath3 bands are shifted most , while the irac bands are shifted least .
therefore , we verify that the magnitude of the downwind shift in the planetary hot spot " due to equatorial winds is strongly wavelength - dependent .
the phase and wavelength dependence of light curves , and the associated day / night contrasts , can be used to constrain the circulation regime of irradiated giant planets and to probe different pressure levels of a hot jupiter atmosphere .
we posit that though our calculations focus on models of hd 209458b similar calculations for other transiting hot jupiters in low - eccentricity orbits should yield transit spectra and light curves of a similar character . |
the decay mode @xmath3 is analogous to the mode @xmath4 , which has been studied extensively in view of cp violation measurements . in these two decay modes ,
the cp asymmetry predicted by the standard model is very small , and the observation of a sizeable effect would be a clear signal of physics beyond the standard model . the decay channel @xmath3 can be used to measure various parameters in the @xmath5-meson system . in the atlas experiment , assuming an integrated luminosity of 30 fb@xmath6 , roughly 10 000 decays can be reconstructed with a signal - to - background ratio of about 1:1 , as will be shown later in this paper .
it can be thus foreseen that the measurement of the branching fraction of this mode , the measurement of the @xmath5 lifetime , as well as other measurements will be feasible , which constitues an important cross - check of other measurements performed using other decay channels .
moreover , the decay mode @xmath3 , when combined with its u - spin symmetric channel , @xmath7 , provides a new strategy for the extraction of the @xmath8 angle of the unitarity triangle @xcite .
indeed , while the phase @xmath9 is ckm suppressed for the decay mode @xmath3 , this is not the case for the decay mode @xmath7 .
cp violation effects could thus potentially be more easily accessible through the latter mode by measuring the time - dependent asymmetry of @xmath10 .
the overall normalization can be fixed by measuring cp averaged rates of @xmath7 and @xmath3 , which assumes the validity of the u - spin symmetry ( the approximate symmetry of u , d and s quarks ) .
the measurements seem to be , however , out of reach due to the small expected branching ratio of @xmath7 , and the overlapping mass peaks of the @xmath5 and @xmath11 @xcite .
the @xmath12 system is characterized by two eigenstates with different masses and decay rates .
the experimental determination of the mass difference @xmath13 and the rate difference @xmath14 will be valuable input for flavour dynamics in both the standard model and its possible extensions . while the measurement of @xmath13 has been proven to be accessible through the decays @xmath15 and @xmath16 , @xmath14 is expected to be precisely measured in atlas through the decay channel @xmath4 .
signal samples of @xmath17 were generated using the monte carlo program pythia 5.7 @xcite .
the j/@xmath18 meson was forced to decay into @xmath19 , and only events passing the atlas level-1 trigger requirements for b hadrons @xcite ( a muon with a @xmath20 6 gev and @xmath21 ) is used for the transverse momentum with respect to the beam direction , and @xmath22 for the pseudorapidity . ] and the @xmath22 meson decaying into @xmath23 were retained .
the branching ratio for the decay @xmath24 is ( @xmath25 ) % @xcite .
events were further selected to satisfy the atlas second level trigger : the presence of a second muon with @xmath26 gev and @xmath27 .
the resulting signal samples were consisting of about 17,000 @xmath28 events and 15,000 @xmath29 events .
a full geant - based simulation was used to simulate the response of the atlas inner detector and the electromagnetic calorimeter .
the muon chambers were not included in the simulation , but only real muons were included in the analysis .
the muon identification efficiencies were assumed to be 85% for the first muon ( the muon with @xmath20 6 gev ) and 78% for the second muon ( the muon with @xmath20 3 gev ) . the 85% efficiency for the muon with the higher transverse momentum includes both reconstruction and level-1 trigger efficiencies .
the j/@xmath18 was reconstructed by fitting pairs of opposite - charge muons passing the level-2 selection ( @xmath30 ) ( @xmath31 ) is used to indicate a muon with a @xmath20 6 ( 3 ) gev . ] to a common vertex and by calculating their invariant mass . the @xmath32/d.o.f . of the vertex fit
was required to be less than 6.0 , and the decay vertex of the j/@xmath18 was required to be detached from the primary vertex by at least 250 @xmath33 m in the transverse plane .
the j/@xmath18 mass resolution was found to be 39 mev .
the resolution of the j/@xmath18 decay vertex in the transverse plane was 64 @xmath33 m @xcite .
the efficiency of the reconstruction within a three - standard - deviation mass window was 79% . if a j/@xmath18 was found , the analysis proceeded with @xmath22-reconstruction in the electromagnetic calorimeter .
the @xmath22 reconstruction is described in more detail in ref .
@xcite . the mass distribution , fig . 1 ,
was fitted with two gaussians , and the obtained mass resolution was @xmath34 .
the total reconstruction efficiecny was 2.3@xmath35 within two standard deviations from the nominal mass . to the j/@xmath18 and @xmath22 candidates were combined to reconstruct both the @xmath36 and @xmath37
. a mass resolution of around 67 mev was found for both channels , with an overall signal efficiency of @xmath38 1@xmath35 within three standard deviations from the nominal mass . to the background considered was a sample of @xmath39 decays , processed and selected in the same way as the signal events .
the resulting efficiency was less than 0.009@xmath35 .
the table below summarizes the event rates for both signals and background . .assumed branching ratios for @xmath17 , and the estimated numbers of signal and background events for an integrated luminosity of 30 fb@xmath6 .
@xmath40 is the @xmath41 mixing angle ( see ref .
@xcite ) . [ cols="<,^,^",options="header " , ] for the measurement of the cp asymmetry in the @xmath42-system , the so - called same - side jet - charge tagging method was used for distinguishing production of @xmath42 and @xmath43 @xcite .
the jet charge was defined as : @xmath44 where @xmath45 is the charge of the @xmath46 particle in the jet , and @xmath47 is the momentum .
all charged particles with @xmath48 gev , @xmath49 around the reconstructed b - meson were included in the jet , if the distance between the particle and the b - meson , @xmath50 , was less than 0.8 .
particles not originating from near the primary vertex were excluded by requiring that the transverse impact parameter @xmath51 was less than 1 cm and the @xmath52-coordinate of the particle at the point of closest approach was within 5 cm of the @xmath52 of the primary vertex .
the particles from the b - meson decay itself were excluded as well . in the analysis ,
the reconstructed b - meson was defined as @xmath42 ( @xmath43 ) if the jet - charge had @xmath53 ( @xmath54 ) , where @xmath55 is a tunable cut .
the exponent @xmath56 was optimized to maximize the tagging quality factor @xmath57 , where @xmath58 is the tagging dilution factor , originating from wrong - sign tags ( @xmath59 , where @xmath60 is the fraction of wrong - sign tags ) , and @xmath61 is the tagging efficiency .
the optimum parameters were found to be @xmath62 , @xmath63 , resulting in a quality factor of 3.85% with @xmath64 , @xmath65 .
the observable asymmetry is : @xmath66 where @xmath67 combines the experimental dilution factors due to mistagging and background .
@xmath68 is the ratio of the number of observed signal events to the total number of observed events . assuming that the decay time resolution is @xmath69 = 0.073 ps , the error on the cp asymmetry was estimated to be : @xmath70 where @xmath71
where @xmath72 , @xmath73 is the cabibbo angle , @xmath22 is the height of the unitarity triangle and @xmath8 is one of the angles of the unitarity triangle .
the @xmath4 mode has a clean experimental signature . for its reconstruction ,
the same approach as for the @xmath17 was used .
details are given in @xcite the @xmath74 leads to three final state helicity configurations and their linear combinations are cp eigenstates with different cp parities @xcite . for the extraction of the cp - violating weak mixing phase @xmath75
the helicity amplitudes need to be separated .
experimentally , one can measure three independent angles and the @xmath5 proper decay time .
the initial @xmath5 flavour can be tagged in part of the events .
the atlas precision for these measurements was determined by detector response simulations and was used as input to angular analyses based on a maximum likelihood fit in repeated monte carlo experiments .
the likelihood function @xmath76 is defined as : @xmath77 with : @xmath78 where @xmath79=0.5 for untagged events , @xmath80=1-r , @xmath81=r if the @xmath5 is tagged as a particle and @xmath80=r , @xmath81=1-r if the @xmath5 tagged as an anti - particle , r is the wrong tag fraction , b is the level of background and @xmath82 is the average decay rate of the background .
the probability density @xmath83 is defined by : @xmath84 in which @xmath85 are expressed in terms of four parameters of @xmath5 weak decay and oscillations : @xmath86 , @xmath87 , @xmath88 , @xmath89 , and four independent parameters of helicity amplitudes : @xmath90 , @xmath91 , and the two strong phases @xmath92 and @xmath93 . the angular functions @xmath94 are given in @xcite , where more details about the analysis are also given . in the maximum likelihood fit the @xmath86 , the @xmath87 and the weak phase @xmath95 were simultaneously determined along with two helicity amplitude values and their strong phases .
the mixing parameter @xmath96 was assumed to have been measured in the @xmath15 and @xmath16 and was fixed . to
while all the eight parameters are independent in the theoretical models , the experimental resolution causes some of them become correlated , particularly the two strong phases , which prevented their simultaneous determination .
thus in the final analysis these two parametres were fixed . for an integral luminosity of 30 fb@xmath6
, @xmath86 can be detrmined with a relative error of 12@xmath35 , while the precision of @xmath95 depends on the value of @xmath96 and on the proper time resolution , as shown in fig . 3 , where the discovery lines for atlas and lhcb are displayed in the ( @xmath97 ) plane , together with regions allowed by the standard model and by one example of new physics known as the left - right symmetric model @xcite .
atlas is expected to measure several parameters in the bs meson system , which will be valuable input for the flavour dynamics in the standard model and its possible extensions .
a clear signal of the decay mode @xmath3 is expected to be reconstructed with the atlas detector , and several measurements are possible such as the measurement of the branching fraction and the bs meson lifetime .
the experimental resolution of the cp asymmetry measured shows some sensitivity to very large asymmetries as predicted by some models beyond the standard model .
the measurement of the @xmath8 angle seems , however to be compromised because the @xmath7 is not observable .
after 3 years at luminosity @xmath98 atlas is expected to determine @xmath86 with the relative errors of 12@xmath35 .
the atlas discovery line ( fig . 3 ) shows that the precision is high enough to be sensitive to new physics .
this work has been performed within the atlas collaboration , and the author would like to thank all members of the collaboration , particularly members of the atlas b - physics working group .
pz skands , jhep 0101:008,2001 c driouichi _
et al _ , observation of the decays @xmath99 in the atlas experiment at the lhc " .
phys . j. direct c4:n2 , 2002 .
t sjstrand , computer physics commun .
82 ( 1994 ) 74 .
atlas collaboration , atlas detector and physics performance technical design report vol i , cern / lhcc/9914 ( 1999 ) . de groom @xmath100 @xmath101 , eur .
j. c15 ( 2000 ) 1 and 2001 off - year partial update for the 2002 edition , available on the pdg www pages ( url : http://pdg.lbl.gov/ ) .
atlas collaboration , atlas detector and physics performance technical design report vol ii , cern / lhcc/9915 ( 1999 ) .
f ohlsson - malek and m melcher , ` atlas calorimeter performance for identification of @xmath8 , @xmath102 and @xmath22 particles at low transverse momentum ' , isn-01.75 and atlas internal note atl - com - phys-2001 - 021 ( 2001 ) .
t feldman , int .
a15 ( 2000 ) 159 .
a dighe , i dunietz , h lipkin , jl rosner , phys . lett .
b369 ( 1996 ) 144 - 150 .
m smizanska , ` determination of @xmath103 by atlas and the impact of non - zero @xmath103 on other @xmath5 studies ' , atl - phys-99 - 003 .
m smizanska , nucl .
a446 : 138 - 142 , 2000 .
p ball and r fleischer , phys .
b475 ( 2000 ) 111 . | the cp asymmetry predicted by the standard model in the decay modes @xmath0 and @xmath1 is very small , and the observation of a sizeable effect would be a clear indication of new physics beyond the standard model .
contrary to the final state @xmath2 , the former final state is a cp eigenstate , and therefore complex angular distribution analyses are not necessary for the asymmetry measurement . in this paper , the approaches for the study of these two decay channels are presented . |
in 1986 , a. r. kemer @xcite solved affirmatively the famous specht problem @xcite by proving that an arbitrary variety of associative algebras over a field of characteristic zero is finitely based .
it is also known @xcite@xcite that there are non - finitely based varieties of associative algebras over an arbitrary field of prime characteristic .
recall that a variety of algebras is said to be _ spechtian _ ( or to have the _
specht property _ ) if its every subvariety is finitely based .
the kemer s theorem has certain analogs in the cases of jordan , alternative , and lie algebras over a field of characteristic @xmath1 .
namely , a. ya .
vais , e. i. zelmanov @xcite proved the specht property of the variety generated by jordan @xmath2 on a finite set of generators .
a. v. iltyakov @xcite obtained the similar result for alternative @xmath3 and also proved in @xcite that the variety generated by a finite dimensional lie algebra is spechtian .
u. u. umirbaev @xcite proved the specht property of every variety of solvable alternative algebras over a field of characteristic distinct from two and three .
the questions about the specht property for the varieties of all alternative , lie , and jordan algebras over a field of characteristic zero are still open problems .
since 1976 , it is known @xcite that the variety of all right alternative metabelian algebras over an arbitrary field is not spechtian . in 1985 ,
i. m. isaev @xcite proved that non - finitely based varieties of right alternative metabelian algebras can even be generated by finite - dimensional algebras .
the specht property for certain varieties of right alternative algebras were also studied in @xcite@xcite . recall @xcite the notion of topological rank for spechtian varieties of algebras .
let @xmath4 be a spechtian variety of algebras and @xmath5 be a proper subvariety of @xmath4 . by a _ system distinguishing @xmath5 from @xmath4 _ we mean a set @xmath6 of nonzero homogeneous polynomials of the free @xmath4-algebra such that @xmath5 can be defined by the union of the identities @xmath7 with the defining identities of @xmath4 .
a _ degree of the system @xmath8 _ is the maximal degree of its polynomials .
dimension @xmath9 of @xmath5 with respect to @xmath4 _ is the minimal possible degree of a system distinguishing @xmath5 from @xmath4
. let @xmath10 be the _ set of all subvarieties of @xmath4_. for every @xmath11 and @xmath12 by @xmath13 we denote the _ set of all proper subvarieties @xmath14 of dimensions @xmath15 _ and put @xmath16 . by definition ,
it is clear that @xmath17 and , for every @xmath18 , we have @xmath19 . therefore considering a set @xmath20 as a base for the neighborhoods , we endow @xmath10 with a topology .
every subset @xmath10 gains a structure of the topological space with respect to the induced topology .
for every @xmath21 by @xmath22 denote a _ subspace of @xmath23 obtained by the exclusion of all its isolated points . _ note that by virtue of the specht property of @xmath4 , every descending chain of varieties in @xmath10 stabilizes and , consequently , every topological subspace of @xmath10 contains isolated points
. therefore one can consider the strictly descending chain @xmath24 the _ topological rank @xmath25 of the space @xmath23 _ is the minimal @xmath26 such that @xmath27 or @xmath28 if such an @xmath26 does nt exist .
the value @xmath29 is called the _ topological rank of the variety @xmath4 _ and is denoted shortly by @xmath30 . for example , if @xmath4 is nilpotent , then every subvariety of @xmath4 turns out to be an isolated point of @xmath31 and , consequently , @xmath32 . otherwise , @xmath22 consists of all non - nilpotent subvarieties of @xmath4 and its isolated points are the varieties that were limit points in @xmath23 for the sequences of nilpotent varieties only .
further , if @xmath33 is not empty , then its isolated points are the varieties that were limit points in @xmath22 for the sequences containing only isolated points of @xmath22 , etc .
the structures of a set of non - nilpotent subvarieties for various varieties of nearly associative metabelian algebras were studied in @xcite@xcite .
a. v. badeev @xcite provided a chain @xmath34 of varieties of commutative alternative nil - algebras over a field of characteristic three such that @xmath35 is a linear function on @xmath26 and @xmath36 . in 2007 , s. v. pchelintsev
@xcite constructed a variety @xmath5 of right alternative metabelian algebras of almost finite topological rank , i. e. a variety @xmath5 such that @xmath37 and @xmath38 is finite for every proper subvariety @xmath39 .
let @xmath40 be a _ field of characteristic @xmath41 _ and @xmath42 be the _ variety of right alternative metabelian algebras over @xmath40 _
defined by the identities @xmath43 where @xmath44 is the associator of the variables @xmath45 . by @xmath46
we denote the _ subvariety of @xmath42 distinguished by the identity _
@xmath47},\ldots , x_{s}\bigr],x_{s+1}\bigr]=0\ ] ] _ of lie - nilpotency of step @xmath0 _ , where @xmath48=xy - yx$ ] is the commutator of @xmath49 .
the specht property of @xmath46 is proved by the author in @xcite . by virtue of nilpotency of every commutative subvariety of @xmath42 ,
we have @xmath50 . s. v. pchelintsev established in @xcite that @xmath51 . in the present paper , we prove the following _ the topological rank of the variety @xmath46 is equal to @xmath0 for all natural @xmath0 .
_ the paper is organized as follows . in sec .
[ sec : basicoperatorrelations ] , we provide some preliminary results about the free @xmath52 @xmath53}$ ] on a countable set @xmath54 of generators over @xmath40 .
[ sec : relationsofthefreealgebra ] is devoted to the studying of relations of the free algebra @xmath55}$ ] . in sec .
[ sec : upperboundforthetopologicalrank ] , we construct a system of linear generators for the space of multilinear polynomials in @xmath55}$ ] of sufficiently high degree and obtain the upper bound @xmath56 by estimating the values of topological ranks of some subvarieties in @xmath46 of special type . finally , in sec .
[ sec : lowboundforthetopologicalrank ] , we construct an auxiliary @xmath57 and considering the identities of their grassmann envelopes obtain the low bound @xmath58 .
throughout the paper , @xmath40 is a field of characteristic @xmath41 ; all vector spaces ( algebras , superalgebras ) are considered over @xmath40 ; @xmath59 is a countable set ; @xmath60}$ ] is a free @xmath52 on the set @xmath54 of generators ; @xmath61 and @xmath62 are , respectively , the operators of right and left multiplication by the element @xmath63 ; @xmath64 ; @xmath65 is the associative algebra generated by all the operators @xmath61 and @xmath62 , for @xmath66 , acting on @xmath67 and by the identical mapping @xmath68 ; @xmath69 is the variety generated by an algebra @xmath70 .
recall @xcite that @xmath65 satisfies the relations @xmath71=0,\label{eq : rrl}\\ & \left[r_x , l_y\right]=-l_xl_y.\label{eq : rl - ll}\end{aligned}\ ] ] relations , imply immediately the following [ lm r2 centr ] the operator @xmath72 lies in the center of @xmath65 .
[ predl predvsoot ] the algebra @xmath65 satisfies the relation @xmath73+h_xr_y+h_yr_x.\ ] ] using , we have @xmath74-l_xr_y - l_yr_x-\left[r_x , l_y\right]=\\ = r_xr_y-2\left[r_x , l_y\right]-l_xr_y - l_yr_x.\end{gathered}\ ] ] combining the obtained relation with and , we get @xmath75-l_xr_y - l_yr_x=\\ = 2\left[r_x , r_y\right]-2\left[r_x , l_y\right]+\left(h_x - r_x\right)r_y+\left(h_y - r_y\right)r_x=\\ = 2\left[r_x , h_y\right]+h_xr_y+h_yr_x.\end{gathered}\ ] ] in what follows , we use the symbol @xmath76 as a common notation for the operator symbols @xmath77 and @xmath78 .
the notation @xmath79 means that each operator symbol of the word @xmath80 can be equal to @xmath77 or @xmath78 independently . in the case
when all operator symbols in some word are assumed to be equal to each other , we use the notation @xmath81 and set @xmath82 .
[ predl esph ] the algebra @xmath65 is spanned by the operators @xmath83 let @xmath84 be a linear span of all operators @xmath85 .
it suffices to prove the inclusions @xmath86 and @xmath87 .
note that yields @xmath88 .
hence the inclusion @xmath86 can be easily proved by induction on the length of the operator @xmath89 . at the same time
, lemma [ lm r2 centr ] implies @xmath87 .
let @xmath90 be a _ linear span in @xmath65 of all operators of the form _
@xmath91 by virtue of , @xmath90 forms an ideal of @xmath65 and by induction on @xmath26 one can prove the congruence @xmath92
let @xmath94}$ ] be the _
free @xmath93 on the set @xmath54 of generators .
_ lemma [ predl esph ] implies immediately the following [ lm esph ] the linear span of all operators of degree @xmath95 in @xmath96 is spanned by the operators @xmath97 in what follows , the term _ `` polynomial '' _ means a homogeneous polynomial of degree not less then two .
let @xmath98 be a symmetric relation on the set of polynomials of @xmath99 such that @xmath100 if @xmath101 , @xmath102 , and @xmath103 does nt depend on the variables @xmath104 .
by the same symbol @xmath98 we denote the induced relation on @xmath65 : @xmath105 for @xmath106 if @xmath107 and @xmath108 do not depend on @xmath109 . the algebra @xmath110 satisfies the relation @xmath111 using and , we have @xmath112 hence , @xmath113 .
therefore applying , for even @xmath114 , and taking into account , we obtain @xmath115 we say that _ almost all polynomials of @xmath110 ( operators of @xmath96 ) satisfy some condition @xmath116 _ if there is a natural @xmath26 such that @xmath116 holds for all polynomials ( operators ) of degree more then @xmath26 .
[ lm approx ] if @xmath117 for @xmath118 , then almost all operators of @xmath96 annihilate @xmath119 .
assume that @xmath120 , where @xmath119 does nt depend on @xmath104 .
in view of lemma [ lm esph ] , every operator word @xmath121 of the degree @xmath122 can be represented as @xmath123 hence by lemma [ lm r2 centr ] , we have @xmath124 [ lm opercosn ] almost all operators of @xmath96 are skew - symmetric with respect to all their variables .
we set @xmath125 . by virtue of and ,
the partial linearization ( see ( * ? ? ? * chap . 1 ) ) of has the form @xmath126 whence , @xmath127 .
thus in view of lemma [ lm esph ] , it remains to calculate @xmath128 the algebra @xmath110 satisfies the relation @xmath129 by virtue of lemmas [ lm r2 centr ] , [ lm opercosn ] , it suffices to verify that @xmath130 . using , , and lemma [ lm opercosn ] , we have @xmath131 [ predl soot rh2 ] the algebra @xmath96 satisfies the relations @xmath132+h_xh_y&\approx0,\label{soot 3r^2rn}\\ \left[r_x , h_yh_z\right]&\approx0.\label{soot rh2}\end{aligned}\ ] ] by lemma [ lm opercosn ]
, relation follows from . using and combining lemmas [ lm r2 centr ] , [ lm opercosn ] with the jacobian identity
, we obtain @xmath133\approx 2\bigl[r_x,\left[r_y , h_z\right]\bigr]\approx\\ \approx\bigl[r_x,\left[r_y , h_z\right]\bigr]-\bigl[r_y,\left[r_x , h_z\right]\bigr ] = \bigl[h_z,\left[r_y , r_x\right]\bigr]=0.\end{gathered}\ ] ] let @xmath84 be an ideal of @xmath96 . for @xmath134 we write
@xmath135 if there is a @xmath136 such that @xmath137 .
let @xmath138 @xmath139 be the _ ideal of @xmath140 generated by all the elements _ @xmath141 . the algebra @xmath96 satisfies the relation @xmath142 we set @xmath143 .
applying and , we have @xmath144
an _ @xmath145 variety _ ( @xmath146 ) is a subvariety @xmath4 of @xmath46 such that the free @xmath147 on the set @xmath54 of generators satisfies the relation @xmath148 where @xmath149},\ldots , x_{n}\bigr],x_{n+1}\bigr ] , & \enskip\text { if $ n$ is even},\\ \bigl[\bigl[\ldots{\left[x_1x_2,x_3\right]},\ldots , x_{n}\bigr],x_{n+1}\bigr ] , & \enskip\text { if $ n$ is odd}. \end{aligned}\right.\ ] ] by definition , every @xmath150 variety @xmath5 is right nilpotent .
moreover , applying lemma [ lm esph ] , it is not hard to prove that @xmath5 is nilpotent and , consequently , @xmath151 .
we also stress that the variety @xmath46 is @xmath152 : for even @xmath0 , it is clear by definition and , for odd @xmath0 , it follows from .
let @xmath4 be an _
@xmath145 variety _
@xmath153 , @xmath154 be the _
free @xmath147 on the set @xmath54 of generators _ , and @xmath155 @xmath156 be the _ subspace of multilinear polynomials in @xmath154 on the variables @xmath157 .
_ in order to avoid complicated formulas while writing down the polynomials of @xmath155 we omit the indices of variables at the operator symbols and assume them to be arranged at the ascending order .
for example , the notation @xmath158 means the monomial @xmath159 _ regular words _ are the polynomials of @xmath155 of the following types : @xmath160}h^{2j}r^{d-2j-2},\\ 3)&\enskip\,{\left[x_2,x_3\right]}h^{2j}r^{d-2j-2},\\ 4)&\enskip\,{\left[x_1,x_2\right]}h^{2k-1}r^{d-2k-1},\end{aligned}\ ] ] where @xmath161 @xmath162 @xmath163 @xmath164}{}{}{n}{2}}$ ] .
[ lm pravl slova ] almost all polynomials of @xmath165 are linear combinations of regular words . by lemma
[ lm opercosn ] , there is a degree @xmath166 such that every monomial @xmath167 is skew - symmetric w.r.t .
@xmath168 . consequently , in view of lemma [ lm esph ] and relation , @xmath155 can be spanned by the polynomials @xmath169}h^{k}r^{d - k-2},\ ] ] where @xmath170 , @xmath171 , and @xmath172 . linearizing and
, we have @xmath173}t_zt_t+{\left[x , t\right]}t_zt_y+{\left[z , y\right]}t_xt_t+{\left[z , t\right]}t_xt_y&\approx0.\end{aligned}\ ] ] applying these relations , it is not hard to prove that @xmath155 can be spanned by the polynomials : @xmath174}h^{k}r^{d - k-2},\\ 3')&\enskip{\left[x_2,x_3\right]}h^{k}r^{d - k-2},\end{aligned}\ ] ] where @xmath175 and @xmath172 . by @xmath176 denote the linear span of all regular words of types @xmath177@xmath178 .
note that the polynomials of types @xmath179@xmath180 lie in @xmath176 for even @xmath181 .
let us verify that the polynomials of types @xmath179@xmath180 for odd @xmath181 can be represented as linear combinations of regular words . by virtue of
, we have @xmath182 hence , in view of , every polynomial of type @xmath179 lie in @xmath176 .
further , using lemmas [ lm r2 centr ] , [ lm opercosn ] , the partial linearization @xmath183 of , identity and relation , we get @xmath184}h_y\approx{\left[x , y\right]}h_yr_zr_u \approx{\left[x , y\right]}r_zr_uh_y={\left[x , y\right]}r_yr_zh_u=\\ = \bigl({\left(xy\right)}y-{\left(yx\right)}y\bigr)r_zh_u \approx\bigl(2{\left(xy\right)}y+y^2x\bigr)r_zh_u = y^2\left(2l_x+r_x\right)r_zh_u=\\ = y^2\left(3r_x-2h_x\right)r_zh_u \approx y^2\left(2{\left[r_x , h_z\right]}-h_xh_z-2h_xr_z\right)h_u\approx\\ \approx y^2\left(2r_x - h_x\right)h_zh_u = y^2\left(r_x+l_x\right)h_zh_u = \left(y^2r_x+{\left(xy\right)}r_y\right)h_zh_u.\end{gathered}\ ] ] in view of the obtained relation implies that the polynomials of types @xmath185 for odd @xmath181 are skew - symmetric modulo @xmath176 with respect to all their variables .
therefore every polynomial of type @xmath185 is proportional modulo @xmath176 to a regular word of type @xmath186 .
[ lm 2n iz 2n+1 ] for every @xmath145 variety @xmath4 @xmath153 there is a punctured neighborhood @xmath187 such that every variety of @xmath187 is @xmath188 . by virtue of the restriction @xmath189 , lemma [ lm opercosn ] and relation
imply that in some punctured neighborhood of @xmath4 every variety can be defined by a system of identities where all polynomials starting from some sufficiently high degree are multilinear .
consequently by lemma [ lm pravl slova ] , we can choose a punctured neighborhood @xmath187 such that every variety @xmath190 satisfies an identity @xmath191 , where @xmath119 is a nontrivial linear combination of regular words of @xmath155 .
let @xmath154 be the free @xmath192 on the set @xmath54 of generators .
we write down relation shortly as @xmath193 if @xmath194 and as @xmath195 if @xmath196 .
first consider the case @xmath194 .
we prove that relation @xmath193 and identity @xmath191 imply @xmath195 . by lemma [ lm pravl slova ]
, @xmath119 can be presented in the form @xmath197}h^{2j}r^{d-2j-2 } \bigr)\pmod{\mathcal w_{3,4}},\ ] ] where @xmath198 is the linear span of regular words of types @xmath199 .
we fix @xmath200 and a minimal index @xmath201 such that @xmath202 .
then by the substitution @xmath203 , for @xmath204 , using the equality @xmath205 and relation , we obtain @xmath206 .
otherwise , we can rewrite @xmath119 in the form @xmath207 where @xmath208}x_3 + \beta_0{\left[x_3,x_1\right]}x_2 + \gamma_0{\left[x_2,x_3\right]}x_1\bigr)r^{d-3},\\ h_t&= \zeta_t{\left[x_1,x_2\right]}h^{2t-1}r^{d-2k-1},\\ \intertext{and } g_k&= \bigl ( \alpha_{k}\bigl[{\left[x_1,x_2\right]},x_3\bigr ] + \beta_{k}\bigl[{\left[x_3,x_1\right]},x_2\bigr ] + \gamma_{k}\bigl[{\left[x_2,x_3\right]},x_1\bigr]\bigr)h^{2k-1}r^{d-2k-2},\\ h_k&= \delta_{k}{\left(x_1\circ x_2\right)}h^{2k}r^{d-2k-2 } + \varepsilon_{k}{\left(x_1\circ x_3\right)}h^{2k}r^{d-2k-2}+ \zeta_k{\left[x_1,x_2\right]}h^{2k-1}r^{d-2k-1},\end{aligned}\ ] ] for @xmath209 .
if at least one of the coefficients @xmath210 is not zero , then by three successive substitutions @xmath211 @xmath212 , we have @xmath213 hence in view of the restriction @xmath189 , we obtain either @xmath206 or @xmath214 .
further , if @xmath215 , then by the substitution @xmath216 , we have @xmath206 . otherwise , if at least one of the coefficients @xmath217 or @xmath218 is not zero , by two successive substitutions @xmath219 @xmath220 , we obtain @xmath221 thus we have either @xmath206 or @xmath222 and , consequently , @xmath119 gets the form @xmath223 therefore by the same arguments as above , we obtain either @xmath206 or @xmath224 and @xmath225}h^{2t-1}r^{d-2t-1}.\ ] ] but in this case , the assumption of the lemma implies @xmath226 and , consequently , @xmath195 .
now consider the case @xmath196 .
we need to prove that relation @xmath195 and identity @xmath191 imply @xmath227 . unlike the case @xmath194 , the regular word of type @xmath186 corresponding to the index @xmath228 vanishes .
all the other regular words are the same . therefore by the similar arguments as above , reducing per unit the power @xmath229 for every substitution @xmath230 and assuming @xmath231
, one can prove that @xmath232 . by , the obtained relation yields @xmath227 . as it was stressed above
, every @xmath150 variety has the topological rank @xmath233 .
consequently lemma [ lm 2n iz 2n+1 ] implies that the topological rank of every @xmath145 variety is not more than @xmath26 .
in particular , @xmath56 .
let @xmath234 be a _ superalgebra _ ( @xmath235 ) with the _ even part @xmath236 _ and the _ odd part @xmath237 _ , i. e. @xmath238 for @xmath239 @xmath240 be the _ grassmann algebra _ on a countable set of anticommuting generators @xmath241 with the natural @xmath242 ( @xmath243 and @xmath244 are spanned by the words of even and , respectively , odd length on @xmath245 ) . the _ grassmann envelope _ @xmath246 of @xmath154 is the subalgebra @xmath247 of the tensor product @xmath248 .
it is well known that @xmath246 satisfies a multilinear identity @xmath191 iff @xmath154 satisfies the certain graded identity @xmath249 called the _ superization of @xmath191 .
_ here @xmath250 denotes the so - called _ superpolynomial corresponding to @xmath119 _ and we also say that @xmath154 satisfies the _ superidentity _ @xmath249 .
the descriptions of the process of constructing of superpolynomials ( the _ superizing process _ ) can be found in @xcite@xcite . for a given variety @xmath4 of algebras , @xmath154
is said to be a @xmath251 if @xmath252 , i. e. if @xmath154 satisfies all the superizations of the defining identities of @xmath4 .
let @xmath253 be one of the elements @xmath254 and @xmath255 be a superalgebra with the countable basis @xmath63 , @xmath256 @xmath257 such that @xmath258 and @xmath259 iff @xmath260 .
we introduce the multiplication of the basis elements of @xmath261 as follows : @xmath262 and all the other products are zero . by definition , it is not hard to see that @xmath261 is a metabelian algebra .
[ predl gapravh ] the algebra @xmath261 is an @xmath263 .
we check that @xmath261 satisfies the superization of : @xmath264 where @xmath265 are homogeneous basis elements of @xmath261 and @xmath266 denotes the parity of @xmath267 , i. e. @xmath268 for @xmath269 . in view of metability of @xmath261 and
the odd parity of @xmath63 , it suffices to verify the relation @xmath270}-{(-1)}^{i+j}l^2_x\bigr)=0.\ ] ] first for even @xmath271 , we calculate @xmath272 } = { ( -1)}^{i}\left ( a_{i,\,j+1}-a_{i+1,\,j}\right)r_x - a_{i,\,j+1}l_x=\\ = { ( -1)}^{i}\left(a_{i,\,j+2}-a_{i+1,\,j+1}\right ) -\frac{{(-1)}^{i}}{2}\left(a_{i,\,j+2}-2a_{i+1,\,j+1}+a_{i+2,\,j}\right)=\\ \shoveright{=\frac{{(-1)}^{i}}{2}\left ( a_{i,\,j+2}-a_{i+2,\,j } \right);}\\ \shoveleft{a_{i , j}\,l^2_x = { ( -1)}^{i}\left(a_{i,\,j+1}-a_{i+1,\,j}\right)l_x=}\\ = \frac{1}{2}\left(a_{i,\,j+2}-2a_{i+1,\,j+1}+a_{i+2,\,j}\right)+a_{i+1,\,j+1}-a_{i+2,\,j}=\\ = \frac{1}{2}\left ( a_{i,\,j+2}-a_{i+2,\,j } \right).\end{gathered}\ ] ] to conclude the proof it remains to make the similar calculations for odd @xmath271 : @xmath272 } = \frac{{(-1)}^{i}}{2}\left(a_{i,\,j+1}-2a_{i+1,\,j}+a_{i+2,\,j-1}\right)r_x - a_{i,\,j+1}l_x=\\ = \frac{{(-1)}^{i}}{2}\left(a_{i,\,j+2}-2a_{i+1,\,j+1}+a_{i+2,\,j}\right ) + { ( -1)}^{i+1}\left(a_{i,\,j+2}-a_{i+1,\,j+1}\right)=\\ = \frac{{(-1)}^{i+1}}{2}\left(a_{i,\,j+2}-a_{i+2,\,j}\right);\end{gathered}\ ] ] @xmath273 [ predl supercom ] the algebra @xmath261 satisfies the relations @xmath274}_{\mathrm s}&=a_{i+1,2j},\\ \label{soot supercom2 } { \bigl[{\left[a_{i , j},x\right]}_{\mathrm s},x\bigr]}_{\mathrm s}&=a_{i+2,\,j},\end{aligned}\ ] ] where @xmath275}_{\mathrm s}=ab-{(-1)}^{{\left|a\right|}{\left|b\right|}}ba$ ] is a supercommutator of the elements @xmath276 . first we calculate @xmath277}_{\mathrm s } = a_{i,\,2j}\,r_x-{(-1)}^{i } a_{i,\,2j}\,l_x = a_{i,\,2j+1}-\left(a_{i,\,2j+1}-a_{i+1,\,2j}\right)=a_{i+1,\,2j}.\ ] ] thus and , consequently , , for even @xmath271 , are proved .
further , for odd @xmath271 , we have @xmath278}_{\mathrm s } = a_{i , j}\,r_x+{(-1)}^{i } a_{i , j}\,l_x = a_{i,\,j+1}+\frac{1}{2}\left(a_{i,\,j+1}-2a_{i+1,\,j}+a_{i+2,\,j-1}\right)=\\ = \frac{3}{2}a_{i,\,j+1}-a_{i+1,\,j}+\frac{1}{2}a_{i+2,\,j-1}.\end{gathered}\ ] ] finally , combining the obtained relation with , we get @xmath279}_{\mathrm s},x\bigr]}_{\mathrm s } = \frac{1}{2}{\bigl[\left(3a_{i,\,j+1}-2a_{i+1,\,j}+a_{i+2,\,j-1}\right),x\bigr]}_{\mathrm s}=\\ = \frac{1}{2}\left(3a_{i+1,\,j+1}-3a_{i+1,\,j+1}+2a_{i+2,\,j}-a_{i+3,\,j-1}+a_{i+3,\,j-1}\right)=a_{i+2,\,j}.\end{gathered}\ ] ] let @xmath280 @xmath281 be the span of all elements @xmath282 such that @xmath283 . by the definition of multiplication in @xmath261 , every nonzero product @xmath284 is a linear combination of elements @xmath285 such that @xmath286 . consequently , @xmath280 is an ideal of @xmath261 . for every natural @xmath287
, we introduce the quotient superalgebra @xmath288 as follows : @xmath289 [ lm vgoban ] the algebra @xmath288 is an @xmath290 .
taking into account lemma [ predl gapravh ] , it suffices to prove that @xmath288 satisfies the superization of . by virtue of metability of @xmath288
, we need to verify the relation @xmath291}_{\mathrm
s},x\bigr]_{\mathrm s},\ldots , x\bigr]_{\mathrm s}=0.\ ] ] by the definition of @xmath288 , using , for @xmath292}{}{}{n}{2}}$ ] , we have @xmath293}_{\mathrm s},x\bigr]_{\mathrm s},\ldots , x\bigr]_{\mathrm s } = a_{i+2k,\,j}=0.\ ] ] thus the required relation is proved for even @xmath26 . in the case of odd @xmath26
, it remains to check the following : @xmath294}_{\mathrm s},x\bigr]_{\mathrm s},\ldots , x\bigr]_{\mathrm s } = \underbrace{\bigl[\ldots\bigl[[}_{2k}a_{0,0},{x]}_{\mathrm s},x\bigr]_{\mathrm s},\ldots , x\bigr]_{\mathrm s } = a_{i+2k,\,j}=0.\ ] ] [ lm vgoban-1 ] the variety @xmath295 is not @xmath188 . in the case
@xmath296 , we need to verify that @xmath297}_{\mathrm s},x\bigr]_{\mathrm s},\ldots , x\bigr]_{\mathrm s}r^j_x\neq0 , \quad j\in\mathbb n.\ ] ] applying , we calculate @xmath298}_{\mathrm s},x\bigr]_{\mathrm s},\ldots , x\bigr]_{\mathrm s}r^j_x = \underbrace{\bigl[\ldots\bigl[[}_{2k-1}a_{0,0},{x]}_{\mathrm s},x\bigr]_{\mathrm s},\ldots , x\bigr]_{\mathrm s}r^j_x = a_{2k-1,j}\neq0.\ ] ] for @xmath299 , we need to check that @xmath300}_{\mathrm s},x\bigr]_{\mathrm s},\ldots , x\bigr]_{\mathrm s}r^j_x\neq0 , \quad j\in\mathbb n.\ ] ] using , we get @xmath301}_{\mathrm s},x\bigr]_{\mathrm s},\ldots , x\bigr]_{\mathrm s}r^j_x = a_{2k-2,\,j+1}\neq0.\ ] ] in view of , lemma [ lm vgoban ] implies that the variety @xmath295 is @xmath145 . consequently by lemma [ lm vgoban-1 ]
, we have @xmath302 for @xmath303 .
finally , comparing this estimate with the result of sec .
[ sec : upperboundforthetopologicalrank ] , we obtain @xmath304 .
this article was carried out at the department of mathematics and statistics ( ime ) of the university of so paulo as a part of the author s post - doc project supported by the so paulo research foundation ( fapesp ) , grant 2010/518802 .
the author is very thankful to the ime for the kind hospitality and the creative atmosphere , to his supervisor prof .
i. p. shestakov for his attention to this article , and to prof .
s. v. pchelintsev for suggesting the problem and for the useful discussions on the obtained results .
a. v. iltyakov , finiteness of basis of identities of a finitely generated alternative pi - algebra over a field of characteristic zero , _ sib .
j. _ * 32 * ( 1991 ) 948961 ; translation from _ sib .
* 32 * ( 1991 ) 6176 .
a. medvedev , example of a variety of solvable alternative algebras over a field of characteristic 2 having no finite basis of identities , _ algebra logic _ * 19 * ( 1980 ) 191201 ; translation from _ algebra logika _ * 19 * ( 1980 ) 300313 .
e. i. zelmanov , i. p. shestakov , prime alternative superalgebras and nilpotentcy of the radical of a free alternative algebra , _ math .
ussr , izv .
_ * 37 * ( 1991 ) 1936 ; translation from _ izv .
nauk sssr , ser .
* 54 * ( 1990 ) 676693 . | in 1981 , s. v. pchelintsev introduced the notion of topological rank for spechtian varieties of algebras as a certain tool for studying the structure of non - nilpotent subvarieties in a given variety .
we provide a variety of right alternative algebras of arbitrary given finite topological rank .
namely , we prove that the topological rank of the variety of right alternative metabelian ( solvable of index two ) algebras that are lie - nilpotent of step not more than @xmath0 over a field of characteristic distinct from two and three is equal to @xmath0 . |
recently , considerable effort has been put into extending ads / cft correspondence beyond high - energy physics by constructing gravity models that are conjectured to be dual to various condensed matter systems .
this has lead to the discovery of holographic superconductors and holographic superfluids , describing condensed phases of strongly coupled , planar , gauge theories .
studying such models involves the construction of electrically charged black holes in an asymptotically ads spacetime , which , below a critical temperature become unstable to forming hair .
that is , a phase transition occurs to a superconductor / superfluid state , in which a sufficiently large @xmath3 charge density triggers the spontaneously breaking of the @xmath3 symmetry .
then an operator charged under the @xmath3 acquires a nonzero expectation value ( see @xmath6 @xcite for a review of these aspects ) . for @xmath5wave superconducting black holes ,
the condensing operator is a vector and hence rotational symmetry is broken .
such black hole solutions have been constructed using either charged non - abelian vector fields @xcite or , alternatively , charged two - forms @xcite .
however , most of the studies in the literature have assumed an @xmath7 construction of the lagrangean of the gravitational system , without a clear connection with a given supergravity model , which makes it rather difficult to describe precisely the application of the ads / cft dictionary . at the same time , the gauged supergravity models generically contain non - abelian vector fields , which may suggest the existence of @xmath5wave superconducting black hole solutions .
the case of @xmath8 , @xmath1 gauged supergravity @xcite , @xcite is of particular interest , given its connection with @xmath9 @xmath10 super - yang - mills theory in @xmath11 dimensions .
the bosonic sector of this theory consists of the metric , twenty scalars and fifteen @xmath12 yang - mills ( ym ) gauge fields gauged supergravity is richer .
however , a number of bosonic fields can be consistently set to zero @xcite . ] .
solutions of @xmath13 @xmath1 model have been considered by several authors for various consistent truncations , with subgroups of @xmath12 ( see @xmath6 @xcite and the references therein ) . however , to our knowledge , to date no attempt has been made to construct non - abelian superconducting black hole solutions in this context .
this paper is aimed as a first step in this direction , by taking a consistent truncation of the @xmath14 model corresponding to @xmath0 , @xmath15 romans gauged supergravity , with a single scalar field @xmath16 possessing a potential @xmath17 which is the sum of two liouville terms .
the scalar @xmath16 approaches asymptotically a constant value @xmath18 corresponding to an extremum of the potential , @xmath19 , which yields an effective cosmological constant @xmath20 .
it turns out that the basic properties of the @xmath0 solutions with non - abelian fields are rather similar to those found for pure @xmath1 einstein - ym-@xmath21 system @xcite , @xcite .
in particular , we find evidence for the existence , at low temperatures , of a superfluid state with a normalisable non - abelian condensate .
since romans theory arises as a consistent kaluza - klein truncation of the type iib supergravity on an @xmath22 @xcite and as a consistent compactification of @xmath23 supergravity @xcite , this shows the existence of holographic superfluids in @xmath24 supergravities .
the bosonic sector of the @xmath9 , @xmath1 romans gauged supergravity @xcite consists of gravity , a scalar @xmath16 , an @xmath2 ym potential @xmath25 ( with field strength @xmath26 and @xmath27 the su(2 ) gauge coupling constant ) , an abelian potential @xmath28 ( @xmath29 being the corresponding field strength ) , and a pair of two - form fields .
these two form fields can consistently be set to zero , which yields the bosonic part of the action @xmath30 where @xmath31 .
here @xmath32 is the dilaton potential , @xmath33 being the @xmath3 gauge coupling constant .
as discussed in @xcite , this theory has three canonical forms , corresponding to different choices of the gauge coupling constant @xmath33 .
the case of interest here corresponds to the @xmath0 version , in which @xmath34 and thus the dilaton potential is @xmath35 the field equations are obtained by varying the action ( [ action5 ] ) with respect to the field variables @xmath36 and @xmath16 @xmath37 where the energy - momentum tensor is defined by @xmath38 the scalar potential has exactly one extremum at @xmath39 , corresponding to an the effective cosmological constant @xmath40 then the effective ads length scale is fixed by the non - abelian gauge coupling constant , @xmath41 . as usual , one supplements ( [ action5 ] ) with a boundary term @xmath42 where
apart from the hawking - gibbons surface term we include also a counterterm part which is required to regularize the total action and the global charges . in the above relation
, @xmath43 is the ricci scalar for the induced metric @xmath44 of the boundary , @xmath45 is the trace of the extrinsic curvature , while @xmath46 ( this expression of the counterterm was derived in @xcite , in a more general context ) .
then , as in the well known pure - ads case @xcite , one can construct a divergence - free boundary stress tensor @xmath47 from the total action @xmath48 by defining @xmath49 where @xmath50 is the einstein tensor of the boundary metric , @xmath51 is the extrinsic curvature , with @xmath52 being an outward pointing normal vector to the boundary . thus , a conserved charge @xmath53can be associated with a surface @xmath54 ( with normal @xmath55 ) , provided the boundary geometry has an isometry generated by a killing vector @xmath56 .
for example , if @xmath57 is a timelike killing vector , then @xmath58 is the conserved mass @xmath59 .
we start with a discussion of the basic properties of the abelian black brane solutions of the @xmath0 romans model . they can be found as a particular limit of the black holes obtained in @xcite in the so - called stu model . in the general case
these black holes possess three different @xmath3 charges and two independent scalars . after setting one scalar to zero and taking two gauge fields to be equal
, one finds after a suitable field redefinition , the following black brane solution of the eqs .
( [ einstein - eqs])-([tij ] ) : @xmath60 with @xmath61 and the matter fields @xmath62 this solution is written in terms of three parameters @xmath63 , corresponding ( up to some factors ) to the global mass and two electric charges . in
what follows , to avoid cluttering our expressions with complicated factors of @xmath27 , we use the observation that the above solution is left invariant by the transformation @xmath64 , @xmath65 , @xmath66 and @xmath67 , and we set @xmath68 without any loss of generality . the horizon is located @xmath69 , with @xmath70 the largest positive root of the equation @xmath71 , which reduces to @xmath72 although one can write an expression for @xmath73 , it turns out to be more convenient to express @xmath74 in terms of @xmath75 : @xmath76 as usual , the constants @xmath77 in the expressions of @xmath78 are found by imposing the regularity of the one - forms @xmath79 on the horizon , which implies @xmath80 thus , physically they correspond to the two chemical potentials associated with the system .
a straightforward computation leads to the following expressions for the mass @xmath59 , electric charges @xmath81 and @xmath82 for the @xmath2 and @xmath3 fields , entropy @xmath83 and hawking temperature @xmath84 : @xmath85 with @xmath86 ; however , for the rest of this work , to simplify the expressions , we set @xmath87 , @xmath88 we shall work with mass , entropy and electric charge densities .
a straightforward computation shows that the solutions satisfy the first law of thermodynamics , @xmath89 and the smarr law , @xmath90 these @xmath4 solutions possess a relatively complicated thermodynamics . restricting for simplicity to a canonical ensemble
, we study black branes holding the temperature @xmath84 , and the charges @xmath91 fixed .
the associated thermodynamic potential is the helmholz free energy @xmath92={\cal m}-t_hs ~.$ ] thermodynamic stability requires the positivity of the specific heat at constant electric charges , @xmath93 .
a useful relation here is @xmath94 which defines @xmath95 , although an explicit formula can not be written in the general case .
analytic results are found only when discussing the limiting cases with a vanishing @xmath96 or @xmath74 .
the properties of the solutions with @xmath97 ( @xmath88 a consistent truncation of the model with a @xmath3 field only , @xmath98 ) are discussed at length in @xcite .
no extremal configurations are found in this case , since the temperature is bounded from below , @xmath99 . for any given @xmath100 ,
there are two branches of solutions , one of them being thermally stable . by constrast
, the solutions with @xmath101 ( @xmath2 gauge fields only , @xmath102 ) , admit an extremal limit which is approached for @xmath103 .
the entropy of the extremal solutions vanishes , a number of invariant quantities diverging in that limit . for nonextremal configurations
one finds a single branch of solutions , with @xmath104,\end{aligned}\ ] ] which possesses a positive specific heat .
the solutions with two @xmath3 charges exhibit a complicated picture , which is governed by the value of the relative ratio @xmath105 .
the picture in figure 1 appears to be generic : for any @xmath106 , the solutions with small to + enough @xmath74 are thermally stable , the entropy increasing with the temperature ( note that @xmath107 , the geometry remaining regular in this limit ) . however , when increasing @xmath74 we notice the occurance of three branches of solutions for some intermediate range of @xmath84 .
the physically relevant branch ( which has less free energy ) is the third one , which continues to @xmath108 ( the large temperature behaviour is @xmath109 ) . also , the second branch is unstable since it possess a negative specific heat . for a more systematic discussion of the properties of the generic abelian solutions , it turns out convenient to work with the following scaled quantities exact solution has an extra scaling symmetry @xmath110 , @xmath111 , @xmath112 , @xmath113 and @xmath114 , with @xmath115 and @xmath116 an arbitrary constant .
the quantities in ( [ metricu1xu1 - 13 ] ) are left invariant by this transformation . ]
@xmath117 with @xmath118 , @xmath119 , @xmath120 , @xmath121 , @xmath122 , and @xmath123 being constant factors which have been chosen such that the expressions below take a simpler form .
all the relevant quantities can then be expressed in terms of @xmath124 and @xmath125 only : @xmath126 it is clear that all solutions satisfy the condition @xmath127 the upper bound being approached for solutions with @xmath2 fields only . moreover ,
the condition @xmath128 imposses a lower bound for the reduced entropy : @xmath129 where @xmath130 one can also show that the scaled @xmath3 charge @xmath131 has a finite range , with @xmath132 solutions with a maximal value of @xmath131 correspond to extremal black holes , with @xmath133 and @xmath134 . from ( [ metricu1xu1 - 14 ] ) , the reduced entropy of the extremal solutions can also be written as @xmath135 which is a nonvanishing quantity for @xmath136 .
it is clear that the @xmath4 solutions should possess non - abelian generalizations .
these configurations are found when enlarging the @xmath2 ansatz to include a nonzero magnetic potential such that the gauge potential @xmath137 is approached only asymptotically . following previous work @xcite , @xcite on pure einstein - yang - mills ( eym ) solutions with vector hair , we choose an @xmath2 gauge fields ansatz possessing both electric and magnetic potentials , while the @xmath3 ansatz is still purely electric : @xmath138 also , as before , the dilaton field will depend on the @xmath139-coordinate only .
this leads to a diagonal energy - momentum tensor and thus it is consistent to choose a diagonal metric ansatz . before discussing the general solutions , it is instructive to consider the perturbative limit of the problem
. then @xmath140 is treated as a small perturbation around the @xmath4 solutions , @xmath141 . after substituting into the linearized ym equations
, one finds that @xmath142 solves @xmath143 for @xmath97 one finds the following exact solution of the above equation @xmath144 ( where @xmath145 are arbitrary constants ) .
as one can see , this solution posseses an essential logarithmic singularity at the horizon and thus can not be treated as a perturbation .
thus we conclude that only solutions which are charged with respect the @xmath2 fields may posses an instability .
although for @xmath106 the equation ( [ pert2 ] ) does not appear to be solvable in terms of known functions , one can construct approximate solutions near the horizon and at infinity .
as @xmath146 , the function @xmath142 behaves as @xmath147 while , for large @xmath139 , the approximate form of @xmath142 is @xmath148 with @xmath149 and @xmath150 free parameters .
solutions interpolating between these asymptotics are constructed numerically . ] .
the mechanism triggering the instability is similar to the pure eym-@xmath21 case @xcite , the magnetic gauge potential acquiring a tachyonic mass term near the horizon .
interestingly , the picture found for @xmath101 is rather similar to that valid for @xmath97 since in this case too no solutions of ( [ pert2 ] ) with correct asymptotics are found . we conclude that , somehow unexpectedly , both electric charges ( associated with the @xmath2 and @xmath3 fields ) should be nonvanishing for the existence of a normalizable zero mode . to some results of the numerical integration
are shown in figure 2 . there
, the part of the parameter space above the curves corresponds to the unbroken phase , where only abelian solutions exist . in figure 2 ( left )
we show the critical curve in the @xmath151 plane corresponding to configurations unstable with respect to non - abelian perturbations .
one can see that the reduced @xmath2 charge @xmath152 has a finite range , @xmath153 , an extremal configuration ( with @xmath154 ) being approached for the maximal value of @xmath124 and @xmath155 ( corresponding to @xmath156 ) . in figure 2 ( right ) we show the critical temperature @xmath157 as a function of the @xmath3 chemical potential @xmath158 ( both quantities are normalized @xmath159 the @xmath2 chemical potential @xmath160 ) ; note that the critical temperature is monotonically decreasing as we increase the ratio @xmath161 .
the instability of the @xmath4 solution pointed out in the previous section can be viewed as an indication of the existence of a branch of non - abelian solutions with nontrivial magnetic non - abelian fields outside the horizon .
in the numerical construction of these solutions , we adopt the following metric ansatz ) of the @xmath4 solution can also be written in the form ( [ metric - an ] ) ( with @xmath162 ) by defining a new radial coordinate . however , this results in rather complicated expressions . ] , which was first proposed in @xcite for the case of the pure eym-@xmath21 system : @xmath163 plugging ( [ metric - an ] ) and ( [ su2-an ] ) into ( [ einstein - eqs])-([tij ] ) results in the equations of motion ) does not contribute to the equations of motion so that the gauge fields do not interact directly . ] : @xmath164 @xmath165 , \\
\nonumber & & ( e^{2a\phi}r^3\frac{a_t'}{\sigma})'=\frac{e^{2a\phi}rf^4}{n\sigma}w^2a_t , ~~ ( e^{-4a\phi}r^3\frac{b_t'}{\sigma})'=0,~~ ( e^{2 a \phi } n f^4 r \sigma w ' ) ' = - \frac{e^{2
a \phi } r f^4}{n \sigma } a_t^2 w \ .\end{aligned}\ ] ] these equations possess the following scaling symmetries ( invariant functions are not shown ) @xmath166 with @xmath116 an arbitrary number .
the symmetries @xmath167 and @xmath168 are used to set @xmath169 , @xmath170 , while @xmath171 is used to set @xmath68 without any loss of generality , which fixes the ads length scale , @xmath172 .
note also that the last equation in ( [ eq - num ] ) implies the existence of the first integral @xmath173 with @xmath74 a constant fixing the @xmath3 electric charge .
we consider again black branes with an horizon at @xmath174 , where @xmath175 .
the non - extremal solutions have the following expansion as @xmath146 : @xmath176 with the independent parameters @xmath177 which fix the coefficients of all higher order terms .
we are interested in solutions approaching the @xmath4 configurations asymptotically .
we assume , @xmath178 which implies the existence of @xmath179 terms also in the expression of the metric functions , @xmath6 @xmath180 . ]
that , as @xmath181 , @xmath140 vanishes and @xmath182 decays as @xmath183 .
a systematic analysis then reveals the following expansion of the solutions at large @xmath139 : @xmath184 with @xmath185 arbitrary coefficients .
all physical quantities are fixed by the data at the horizon and at infinity . as in the abelian case , the global charges are the mass and the @xmath2 and @xmath3 electric charges , with case , the total mass is not given by the asymptotic value of @xmath186 , aquiring a contribution from the scalar field . ] @xmath187 while @xmath160 , @xmath158 are chemical potentials associated with the two gauge fields .
the entropy and hawking temperature of the solutions are given by @xmath188.\end{aligned}\ ] ] to for completness , we mention that the boundary stress - tensor @xmath189 as defined by ( [ s1 ] ) is diagonal , with the nonzero components : @xmath190 such that @xmath191
. eqs .
( [ eq - num ] ) with boundary conditions ( [ eh - an0 ] ) and ( [ bc23 ] ) , respectively , have been solved numerically using a standard shooting method .
in addition to using this algorithm , some solutions were also constructed by employing a collocation method for boundary - value ordinary differential equations equipped with an adaptive mesh selection procedure .
we have confirmed that there is good agreement between the results obtained with these two different methods .
as expected , some basic properties of these black branes are rather similar to those found in @xcite , @xcite in the case of the purely eym-@xmath21 model . however , the solutions in the present work feature a second control parameter , which is the @xmath3 electric charge @xmath131 ( or equivalently , the chemical potential @xmath158 ) . for all solutions , the functions @xmath192 and @xmath78 always increase monotonically with growing @xmath139 . however , @xmath193 and @xmath140 may feature a more complicated behaviour , with local extrema .
for sufficiently small @xmath194 , all field variables remain close to their values for the abelian configuration with the same @xmath195 .
significant differences occur for large enough values of @xmath194 and the effect of the magnetic fields on the geometry becomes increasingly more pronounced .
the profiles of a typical solution illustrating these features are presented in figure 3 . in the numerical approach
, we make use of the existence of the first integral ( [ fi ] ) to fix the value of the electric charge associated with the @xmath3 field , which implies @xmath196 in the near horizon expansion ( [ eh - an0 ] ) . the scaling symmetry @xmath197 in ( [ scal1 ] ) is used to set @xmath198 , such that the only remaining control parameters are @xmath199 and @xmath200 . we have studied in a systematic way families of solutions with fixed values of @xmath74 between 0.5 and 7 , the following picture being generic .
first , the behaviour of solutions for arbitrary data on the horizon is such that at large @xmath139 one finds @xmath201 and @xmath202 ( in which case the total mass as defined according to ( [ mcons ] ) diverges ) , or else there is a singularity at finite @xmath139 . given ( @xmath203 ) , solutions with the correct asymptotic behaviour .
it turns out that the configurations with nodes represent excited states whose energy is always greater than the energy of the corresponding nodeless configurations , and are therefore ignored in what follows . ]
are found only for a discrete set of values of @xmath204 , @xmath205 .
also , all solutions possess a + to + non - vanishing electric @xmath206 associated with the @xmath2 field .
moreover , for fixed @xmath207 , one finds a branch of non - abelian solutions for @xmath208 .
along this branch , the hawking temperature decreases , an extremal configuration being approached . ] for the maximal value of @xmath209 .
the numerical construction of the solutions with @xmath210 requires a different metric ansatz than ( [ metric - an ] ) and is beyond the scope of this work .
however , based on the results in the near - extremal case , we expect the extremal solutions to share the basic properties of their general abelian counterparts , possesing a regular horizon with non - vanishing entropy .
some results of the numerical integration are shown in figure 4 .
there we employ scale - free quantities defined in ( [ metricu1xu1 - 13 ] ) , which are invariant under the scaling transformation @xmath197 in ( [ scal1 ] ) ; also , we have found it convenient to define @xmath211 as a second scale - free control parameter .
one can see that , for all cases we considered , non - abelian solutions exist only for values of the hawking temperature smaller than a critical temperature @xmath157 .
this is the temperature at which the @xmath4 solution admits a static linearised perturbation , with nonvanishing but infinitesimally small @xmath212 .
moreover , the dependece of the order parameter @xmath150 on the hawking temperature is similar to that found in the literature for the @xmath213 case ( @xmath88 a eym-@xmath21 model ) .
also , as expected , we have found that the difference in the free energy density , @xmath214 , between a non - abelian solution and the @xmath4 solution with the same temperature and electric charges is negative , and thus the non - abelian solution is thermodynamically favoured .
in this paper we have studied electrically charged black branes of the @xmath0 @xmath15 gauged supergravity model with ads@xmath215 asymptotics .
apart from the abelian @xmath4 configurations , we have given numerical evidence that this model possesses also solutions with a non - vanishing magnetic @xmath2 fields .
remarkably , these emerge as perturbations of the abelian configurations at some finite temperature depending on the values of the electric charges , and can be viewed as @xmath5wave superfluids . moreover , by using the relations in @xcite , @xcite , one can uplift these configurations to ten dimensional type iib supergravity and @xmath23 supergravity .
this provides an explicit stringy construction of holographic superfluids .
our study should be viewed only as a preliminary investigation of the simplest non - abelian solutions of the @xmath0 model featuring superfluid properties .
various properties of these black branes remain to be investigated .
for example , it would be interesting to compute the conductivity of the solutions or to explore the connection with the unbalanced mixtures discussed recently in @xcite .
moreover , we expect the @xmath0 model to possess a variety of other electrically charged black brane solutions .
they would be found for a different ( typically more complicated ) matter field ansatz than ( [ su2-an ] ) . in this context , it is interesting to study solutions in which the chern - simons term enters the dynamics model @xcite , the properties of the spherically symmetric non - abelian solutions are very different once one switches on a chern - simons term in the action . ] . to this end , we have considered non - abelian black branes possesing a purely magnetic @xmath2 field , with @xmath216 and an electric @xmath3 field , @xmath217 .
this leads to an isotropic energy - momentum tensor , @xmath218 , in which case a suitable metric ansatz is given by ( [ metric - an ] ) with @xmath162 . then ( [ einstein - eqs])-([tij ] ) yield five equations of motion for @xmath219 , @xmath220 , @xmath212 , @xmath221 and @xmath16 which were solved numerically .
our results show that the properties of these solutions differ substantially from those found in the anisotropic case , discussed in section 4 .
first , when treating @xmath140 as a perturbation around the abelian solution ( which is ( [ metricu1xu1 - 1])-([metricu1xu1 - 5 ] ) with @xmath97 ) , the linearized equation can be solved in closed form . however , the solution looks very similar to ( [ pert1 ] ) ( with a @xmath222 term ) , with the result that no normalizable zero mode is found . also ,
different from the case of anisotropic non - abelian black branes , we could not find non - perturbative solutions with @xmath223 as @xmath181 . as a result ,
the mass of the solutions computed according to the counterterm prescription given in section 2 , diverges .
we therefore conclude that these isotropic black brane non - abelian solutions can not be interpreted as holographic superfluids .
however , the situation is likely to be different for a more general case featuring an anisotropic @xmath2 field and a purely magnetic @xmath3 field ( thus beyond the simple ansatz ( [ su2-an ] ) ) .
superconducting black brane solutions of this type have been studied recently in a truncation of @xmath0 romans model with a vanishing dilaton and an arbitrary chern - simons coupling constant @xcite .
we hope to return elsewhere with a systematic study of these aspects .
+ + * acknowledgements * + this work is carried out in the framework of science foundation ireland ( sfi ) project rfp07 - 330phy .
gratefully acknowledge support by the dfg , in particular , also within the dfg research training group 1620 models of gravity .
99 g. t. horowitz , arxiv:1002.1722 [ hep - th ] .
s. s. gubser , phys .
* 101 * ( 2008 ) 191601 [ arxiv:0803.3483 [ hep - th ] ] . f. aprile , d. rodriguez - gomez and j. g. russo , jhep * 1101 * ( 2011 ) 056 [ arxiv:1011.2172 [ hep - th ] ] .
m. gunaydin , l. j. romans and n. p. warner , nucl .
b * 272 * ( 1986 ) 598 .
m. cvetic , h. lu , c. n. pope , a. sadrzadeh and t. a. tran , nucl .
b * 586 * ( 2000 ) 275 [ arxiv : hep - th/0003103 ] .
m. cvetic , h. lu and c. n. pope , arxiv:0908.0131 [ hep - th ] . l. j. romans , nucl .
b * 267 * ( 1986 ) 433 .
r. manvelyan , e. radu and d. h. tchrakian , phys .
b * 677 * ( 2009 ) 79 [ arxiv:0812.3531 [ hep - th ] ] .
m. ammon , j. erdmenger , v. grass , p. kerner and a. obannon , phys .
b * 686 * ( 2010 ) 192 [ arxiv:0912.3515 [ hep - th ] ] .
h. lu , c. n. pope and t. a. tran , phys .
b * 475 * ( 2000 ) 261 [ hep - th/9909203 ] .
j. p. gauntlett and o. varela , jhep * 0802 * ( 2008 ) 083 [ arxiv:0712.3560 [ hep - th ] ] .
a. batrachenko , j. t. liu , r. mcnees , w. a. sabra and w. y. wen , jhep * 0505 * ( 2005 ) 034 [ hep - th/0408205 ] .
v. balasubramanian and p. kraus , commun .
phys . * 208 * ( 1999 ) 413 [ arxiv : hep - th/9902121 ] .
k. behrndt , m. cvetic and w. a. sabra , nucl .
b * 553 * ( 1999 ) 317 [ hep - th/9810227 ] .
j. erdmenger , v. grass , p. kerner and t. h. ngo , jhep * 1108 * ( 2011 ) 037 [ arxiv:1103.4145 [ hep - th ] ] .
l. a. pando zayas and d. reichmann , phys .
d * 85 * ( 2012 ) 106012 [ arxiv:1108.4022 [ hep - th ] ] .
y. brihaye , e. radu and d. h. tchrakian , phys . rev .
d * 81 * ( 2010 ) 064005 [ arxiv:0911.0153 [ hep - th ] ] ; + y. brihaye , e. radu and d. h. tchrakian , phys .
d * 85 * ( 2012 ) 044041 [ arxiv:1110.1816 [ gr - qc ] ] .
a. donos and j. p. gauntlett , jhep * 1112 * ( 2011 ) 091 [ arxiv:1109.3866 [ hep - th ] ] .
e. radu , class .
* 23 * ( 2006 ) 4369 [ hep - th/0601135 ] . | we analyze the properties of asymptotically ads electrically charged black brane solutions in a consistent truncation of the @xmath0 , @xmath1 romans gauged supergravity which contains gravity , @xmath2 and @xmath3 gauge fields , and a dilaton possessing a nontrivial potential approaching a constant negative value at infinity .
we find that the @xmath4 solutions become unstable to forming non - abelian hair .
these configurations emerge as zero modes of the abelian solutions at critical temperature and a critical ( nonvanishing ) ratio of the electric charges and can be viewed as holographic @xmath5wave superfluids . |
for diffuse , neutral gases in space , two fundamental processes influence the pattern of abundances of different elements in free atomic form : ( 1 ) the contributions from various nucleosynthetic sources and how the outcomes are influenced by mixing and transport processes , and ( 2 ) the removal of atoms as they are incorporated into dust grains .
there have been numerous studies of the gas - phase abundances of elements in the local region of our galaxy ( cowie & songaila 1986 ; savage & sembach 1996 ) , using information provided by uv interstellar absorption lines in the spectra of stars .
these investigations have generally focused on the second of the two processes ( jenkins , savage , & spitzer 1986 ; sofia , cardelli , & savage 1994 ) , using abundances in the sun or nearby , early - type stars as a comparison template for material that would be free of dust . by contrast , studies of very distant gas systems that produce absorption lines in quasars have concentrated on the history of element production ( pettini , boksenberg , & hunstead 1990 ; pettini et al .
1994 , 1997 ; timmes , lauroesch , & truran 1995 ; lu et al .
1996 ) , i.e. , its character and rate , and have regarded the depletions caused by dust formation as a possible complication that , if present , must be compensated for . it is generally recognized that the overall severity of a given element s depletion is correlated with the stability of chemical bonds of likely compounds in the grains .
this differential effect can be interpreted to result from either the relative completeness of the element s incorporation into grains in the outer envelopes of stellar atmospheres ( field 1974 ) , how rapidly the element contributes to the growth of grains in dense interstellar clouds ( snow 1975 ; tielens & allamandola 1987 ) , or the relative resistance of different grain components due to destruction by interstellar shocks ( seab 1987 ; jones et al .
1994 ; tielens et al . 1994 ) .
elements that have very strong bonds in dust compounds , such as al , ti and ca , are depleted by as much as 2 to 3 orders of magnitude ( jenkins 1987 ) , while more volatile elements , such as c , n and o , have depletions of only factors of 2 or 3 ( hibbert , dufton , & keenan 1985 ; meyer , et al .
1994 ; cardelli et al .
1996 ; meyer , cardelli , & sofia 1997 ; meyer , jura , & cardelli 1997 ) or perhaps even less see below .
noble gases represent the greatest extreme of volatility , since they generally do not form chemical bonds with solid compounds , and they can adhere to grains only through the very weak van der waals force ( watson 1976 ) .
this paper presents observations of the interstellar abundances of one such noble gas , the element argon in its neutral form .
one feature of this study is the fact that the lines of sight have a low average gas density , and this further reduces the likelihood that the abundances are modified by dust formation ( jenkins 1987 ) .
neutral argon has two strong resonance lines in the far ultraviolet .
their f - values differ by a factor of 4 ( federman et al . 1992 ) , which allows one to test for line saturation and apply corrective measures ( or realize that the saturation is hopelessly bad ) . while the interstellar ar features are easily recognizable in most spectra , their great strength turns out to be a liability because the lines become strongly saturated when dense clouds are being investigated .
for instance , for a velocity dispersion @xmath9 = @xmath10 and @xmath11 , the weaker of the two ar i lines at 1066.66 would have a central optical depth @xmath12 if the argon had a cosmic abundance ratio with respect to hydrogen .
the consequences of such saturation are lessened if the line s velocity structure is mostly resolved by the instrument , because one can interpret the apparent optical depth as a function of velocity ( savage & sembach 1991 ; jenkins 1996 ) instead of the much cruder approach of just measuring the equivalent widths and using the standard curve of growth to derive a total column density .
the observations reported here take advantage of an instrument that can , in most circumstances , resolve nearly all of the velocity detail of a feature , which puts us in a better position to derive accurate column densities than was possible with the _ copernicus _ satellite ( rogerson et al .
1973a ) .
the reference abundance of ar derived for the sun or b stars is now understood with reasonable accuracy ( see [ abundances ] ) , and the difference between the two is not very large .
the only major uncertainty is whether or not the solar or b - star abundances are indeed appropriate standards for comparison .
recently accumulated evidence ( cardelli et al .
1996 ; fitzpatrick 1996 ; meyer , cardelli , & sofia 1997 ; meyer , jura , & cardelli 1997 ; sofia et al .
1997 ) indicates that some mildly depleted elements do not follow the usual trend of depletions becoming stronger in denser interstellar clouds ( spitzer 1985 ; jenkins , savage , & spitzer 1986 ; jenkins 1987 ) , as we would expect if these elements were major constituents in grains that are more abundant and larger in dense regions because they are protected from interstellar shocks and are allowed to grow rapidly ( tielens & allamandola 1987 ; jenkins 1989 ) .
this phenomenon has raised the possibility that the intrinsic abundances of heavier elements in the local interstellar medium ( in both solid and gaseous form ) may be lower than for the nearby stars , and that the depletions by grain formation for some of them are relatively small in comparison to this deficiency , we adopt the term @xmath13 to mean how much more of the neutral form of argon is reduced compared to that of hydrogen , in contrast to a straight reduction @xmath14 that may occur when the gases are in a fully neutral region but ar condenses out onto the grains . ] .
a noteworthy property of argon is that , while its neutral form has an ionization potential greater than that of hydrogen , its cross section for photoionization above this threshold is extraordinarily high . this is of no consequence for an interstellar cloud that is thick enough to have virtually all of its interior shielded from outside sources of ionizing radiation by a large column of neutral hydrogen .
however the situation for thinner clouds is different .
since the absorption cross section of hydrogen diminishes for energies well above the lyman limit , we naturally expect that relatively hard radiation must penetrate the cloud and partially ionize the hydrogen and argon atoms . in circumstances
where some or almost all of the hydrogen is ionized , the fraction of argon that remains in neutral form can be considerably lower than the corresponding fraction for hydrogen . as a result
, this element may deceptively appear to be `` depleted '' relative to hydrogen if one makes the simple - minded assumption that both elements are predominantly neutral .
we will explore this issue in more detail in [ ionization ] and present an assessment of the conditions where this effect is likely to be significant .
our discussion starts with the observations and data reduction in [ obs ] , followed by the resulting ar i column densities ( [ col_dens ] ) and how they relate to measurements of neutral hydrogen along the same lines of sight ( [ abundances ] ) . in [ previous ]
we briefly compare our results with those of previous observations at much lower resolution . on the theoretical side , in [ depletion ? ] we will present in greater depth the arguments that support our assertion that the apparent deficiencies of ar i relative to h i are unlikely to be caused by the depletion of ar onto dust grains .
we argue in [ ionization ] that for our observing circumstances the most likely reason for a reduction in ar i is because it is readily photoionized by energetic radiation that is able to penetrate the region .
we can use this effect to understand better the nature of the partially ionized clouds .
ar is a good tracer element to indicate how strongly the apparent abundances of other elements could be influenced by ionization , and this information is crucial for research on the chemical evolution of the gases in different contexts .
the ultraviolet absorption line data used for this study were obtained with the interstellar medium absorption profile spectrograph ( imaps ) during the first orfeus - spas space shuttle mission in september 1993 .
imaps is a simple , objective - grating echelle spectrograph that can record the spectrum of a bright , hot star in the wavelength interval 950 to 1150 , a spectral region that is densely populated with atomic and molecular transitions from the neutral and ionized interstellar medium ( ism ) ( morton 1975 ; spitzer & jenkins 1975 ; cowie & songaila 1986 ) .
the instrument was designed to achieve the high resolving power needed to show most of the detailed velocity structures in the ism .
jenkins , et al .
( 1996 ) give a comprehensive description of the imaps instrument , its flight on the first orfeus - spas mission , and the methods of data correction and analysis .
this paper includes observations in 9 sight lines for which the neutral argon resonance transitions at 1048.220 and 1066.660 were recorded .
these paths through the ism are toward nearby , early - type stars with low reddening .
table [ targets ] shows the list of target stars with their galactic coordinates , spectral types , visual magnitudes , distances , b@xmath15v color excesses , and line - of - sight fractions of molecular hydrogen .
lccccccc @xmath16 cas & 123.58 & @xmath152.15 & b0ive & 2.47 & 188 & 0.13 & @xmath155.50@xmath17 ori & 205.22 & @xmath1517.24 & b0iae & 1.70 & 412 & 0.05 & @xmath153.91@xmath1 ori & 206.46 & @xmath1516.59 & o9.5ibe & 2.05 & 251 & 0.05 & @xmath154.66@xmath18 ori & 214.52 & @xmath1518.50 & b0.5iav & 2.06 & 221 & 0.04 & @xmath154.92@xmath19 eri & 290.84 & @xmath1558.79 & b3vpe & 0.46 & 44 & 0.07 &
@xmath1 pup & 255.98 & @xmath154.71 & o5iaf & 2.25 & 429 & 0.04 & @xmath155.50 @xmath2 vel & 262.81 & @xmath157.70 & wc8+o7.5e & 1.78 & 258 & 0.05 & @xmath155.51 @xmath3 cru & 302.47 & 3.18 & b0.5iii & 1.25 & 108 & 0.03 &
@xmath3 cen a & 311.77 & 1.25 & b1iii & 0.61 & 161 & 0.02 & @xmath156.44 once the images of the echelle spectra have been reduced and appropriately coadded , the fluxes _ vs. _ wavelength must be extracted .
our basic extraction method was discussed in jenkins et al .
( 1996 ) , however we have altered some of the procedures since the time of that article .
jenkins et al . described a spectral extraction slit that had a uniform cross - dispersion shape and spacing between orders .
these parameters are important for the extraction because the low intensity wings of adjacent echelle orders overlap in the cross - dispersion direction .
we used extraction slits with predefined empirically determined cross - dispersion profile shapes which varied slightly as a function of spectral order .
we also allowed the profile to be broadened or narrowed , and the distance between the orders could vary slightly in order to better fit the data .
these modifications to the extraction slit were performed interactively during the extraction .
our background level determination for the extraction procedure also differed from that discussed in jenkins et al .
here we examined the regions which contain the least contribution from the spectral orders , halfway between the order being extracted and each of the adjacent orders . since the shape of the order s cross - dispersion profile is known , we are able to subtract the contributions of the 2 adjacent orders from each of these interorder regions .
the remaining light was assumed to be the smooth background which changes with position in the echelle dispersion direction .
the average of the two interorder points on either side of the order being extracted was used as the background level and subtracted . to test the reliability of this method , we inspected the cores of lines that , from their appearance , are likely to be fully saturated and hence have zero intensity in their cores .
we found that the rms dispersion of these zero levels was approximately 5% of the local continuum flux level .
figure [ intens_prof ] shows the extracted , normalized profiles of @xmath20 1048 and 1066toward the stars in our sample .
the sampling interval for the curves in figure [ intens_prof ] is one half - pixel on the ccd since the data are put into an oversampled format for reduction purposes ( jenkins et al .
1996 ) .
an interesting exercise is to compare the imaps absorption profiles with the corresponding ones observed in the 1970 s with _ copernicus _ ( rogerson et al .
although these two instruments had very different wavelength resolving powers ( _ copernicus _ could resolve a velocity difference of about @xmath21 compared to @xmath22 for imaps ) , the equivalent widths of the absorption lines should be the same if the correct continua and background corrections were assigned in each case .
table [ eqw ] shows the equivalent width measurements for the three transitions near @xmath23 1134and the two transitions at @xmath23 1048and @xmath23 1066for @xmath24 vel ( morton & bhavsar 1979 ) , @xmath1 pup ( morton 1978 ) and @xmath17 ori ( shull 1979 ) . in the cases where the absorption features reach near zero intensity ( all of the and @xmath23 1048features , as well as @xmath23 1066toward @xmath17 ori ) , the imaps equivalent widths agree well ( @xmath25 10% difference ) with those from _
copernicus_. however the lines which do not approach zero intensity , i.e. the weaker line of ( @xmath23 1066 ) toward @xmath1 pup and @xmath2 vel , show substantial differences ( @xmath26 10% ) when comparisons are made between the imaps and _ copernicus _ measurements .
background levels were often difficult to determine for _ copernicus _ data , and we have no reason to believe that the imaps background determinations for the line at 1066 are worse than for any other region of the spectrum . we therefore believe that our background determinations may be more accurate for these weak lines than those determined for _ copernicus _ in this spectral region .
llccccc @xmath17 ori & imaps & 112 & 76 & 128 & 137 & 141 & _ copernicus _ & 106 & 78 & 126 & 136 & 152 & & & & & & @xmath1 pup & imaps & 88 & 53 & 112 & 120 & 128 & _ copernicus _ & 89 & 65 & 114 & 123 & 131 & & & & & & @xmath2 vel & imaps & 78 & 41 & 92 & 118 & 118 & _ copernicus _ & 72 & 54 & 99 & 113 & 107 jenkins & peimbert ( 1996 ) analyzed the telluric excited lines that appeared in the imaps spectrum of @xmath1 ori a , in order to determine the width of the instrument s profile .
assuming that the lines are thermally broadened for @xmath27 = 1000k , they found that the instrumental smearing function for those observations was approximately a 3.5 pixel ( fwhm ) gaussian , a width that corresponds to a doppler shift of 4 km s@xmath28 .
it is reasonable to assume that the resolving power for the other stars observed during the orfeus - spas - i mission is approximately the same as that for @xmath1 ori a. it is lower than what should have been achievable in principle with imaps and the pointing stability of the spacecraft .
we attribute the degradation to the relief of mechanical stresses within a sticky bearing that supported the echelle grating .
this problem led to small motions of the spectrum while the exposures were underway .
the magnitude and character of this effect has been discussed in detail by jenkins , et al .
an important advantage of having observations that can resolve much of the velocity structure is the ability to more explicitly detect and possibly correct for line saturation . in addition
, sections of the profile at different velocities may exhibit interesting behavior that would otherwise be lost at low resolution .
while we could have used our observations to define some combination of separate components and then determined their amplitudes , velocity centroids and widths that are most consistent with the data , we instead have chosen a more generalized procedure .
we measured the intensity @xmath29 in terms of the continuum level @xmath30 and derived an apparent optical depth as a function of velocity , @xmath31 a function whose interpretation has been described in some detail by savage & sembach ( 1991 ) and jenkins ( 1996 ) . for the ideal cases where either the line is very weak or we are sure that the instrument has resolved the finest details in velocity , @xmath32 gives an accurate depiction of a differential column density per unit velocity through the relation @xmath33 where @xmath34 is the transition s @xmath34-value and @xmath23 is expressed in . in some circumstances , @xmath35 derived through the use of eqs .
[ tau_a ] and [ n_a ] can give an underestimate for the true column density per unit velocity @xmath36 .
if narrow , saturated structures within the velocity profile are smoothed over by the instrument , one will not fully recognize how closely the cores of the features approached the zero intensity level before the smoothing took place , thus leading to an incorrect representation of @xmath36 .
fortunately , we can sense when this is happening ( and over what velocity range ) if we are able to observe two or more lines with different @xmath34-values and note that the weaker line gives a larger @xmath35 than corresponding values for the stronger line .
jenkins ( 1996 ) has described how one can use this information to derive a good approximation to the true @xmath36 , aside from the fact that it has been smoothed .
this method relies on a point - by - point correction technique that mirrors the classical curve of growth analysis method that has been used for the equivalent widths of lines .
a proper application of the @xmath32 correction procedure requires that the velocity scales of the strong and weak lines are accurately matched . in their imaps spectrum that showed many features of h@xmath37 in the spectrum of @xmath1 ori
a , jenkins & peimbert ( 1996 ) found that the velocities were internally consistent to within an rms dispersion of only @xmath38 .
since our method of correcting for image distortions and deriving the wavelength scales is identical to that of jenkins & peimbert , we expect a similar accuracy in this study . beyond the considerations discussed above ,
other key factors that influence the accuracy of the column densities include a good knowledge of the background and continuum levels in the region of the absorption and , of course , accurate oscillator strengths for the transitions being measured .
we have already discussed in [ obs ] the probable accuracy of our background subtractions .
the stellar continua over the velocity breadth of the absorption features are quite smooth and well determined for all of the stars .
we found that a third - order polynomial fitted to the spectrum adjacent to the absorption gave a satisfactory definition of the continuum .
we estimate the continuum fitting error to be 0.35 times the rms deviation of the data about the fitted continuum in the spectral region adjacent to the absorption line of interest [ see savage , cardelli & sofia ( 1992 ) for a justification of this error estimate ] .
for the two ar transitions we used the oscillator strengths @xmath39 and @xmath40 reported by federman et al .
( 1992 ) .
the optical depth correction procedure discussed in [ method ] was applied to the ar lines toward the stars in our sample .
valid corrections can not be made for line pairs in regions where the weaker line has near zero intensity .
figure [ na_prof ] shows the logarithmic column densities as a function of velocity for the four sight lines with corrected optical depth profiles which we believe do not contain any uncertain points ( @xmath1 pup , @xmath2 vel , @xmath3 cen a and @xmath19 eri ) ; the remaining five sets of profiles could not be corrected across the entire velocity extent of the ar absorption . integrating eq .
[ n_a ] over the velocity extent of the ar absorption yields the column densities and limits listed in table [ abund ] .
the column densities were derived using the corrected optical depth profiles , and the limits are based on the uncorrected @xmath231066optical depths .
the reported uncertainties are based on the statistical errors in the data , the uncertainty in the oscillator strengths , the error in continuum definition , and a background uncertainty of 5% of the continuum flux level , all of which were added in quadrature . no uncertainty relating to the correction of the optical depth profiles
is included .
the differences between the column densities measured after and before the optical depth correction was applied to the weaker line are also shown in the third column of table [ abund ] for those lines which could be corrected .
the differences between the pre- and post - correction column densities are quite small for these sight lines , all within the 1-@xmath41 measurement errors .
this agreement gives us added confidence that our background determinations are probably reliable .
lcccc @xmath16 cas & @xmath2614.10 & & [email protected] & @xmath430.58@xmath17 ori & @xmath2614.39 & & [email protected] & @xmath430.59@xmath1 ori & @xmath2614.44 & & [email protected] & @xmath430.48@xmath18 ori & @xmath2614.18 & & [email protected] & @xmath430.88@xmath19 eri & 13.19 @xmath42 0.09 & 0.00 & @xmath2519.36 & @xmath430.69@xmath1 pup & 14.14 @xmath42 0.08 & 0.03 & 19.99 @xmath42 0.02 & @xmath150.37 @xmath42 0.09@xmath2 vel & 14.11 @xmath42 0.08 & 0.08 & 19.77 @xmath42 0.03 & @xmath150.18 @xmath42 0.10@xmath3 cru & @xmath2614.10 & & & @xmath3 cen a & 13.45 @xmath42 0.10 & 0.00 & 19.54 @xmath42 0.06 & @xmath150.61 @xmath42 0.12
in order to study the significance of our argon measurements , we must compare them to measurements of other elements in the same gas .
when possible , hydrogen is generally used as the comparison species .
the cosmic abundance ratio of ar to h is well determined .
anders & grevesse ( 1989 ) give a logarithmic abundance for ar / h of @xmath155.44 @xmath42 0.10 which is based on corrected solar measurements , and regions , and stars .
since our objective is to study the abundances of ar in the ism , it would not be appropriate to use a cosmic abundance which includes interstellar values .
keenan et al . ( 1990 ) and holmgren et al . ( 1990 ) find logarithmic ratios of @xmath155.50 and @xmath155.51 for ar / h in b star photospheres with uncertainties of @xmath42 0.05 and @xmath42 0.10 respectively .
solar photospheric abundances determined from correcting solar coronal observations have produced logarithmic ratios of @xmath155.46 @xmath42 0.1 ( anders & grevesse 1989 ) and @xmath155.45 @xmath42 0.08 ( meyer , 1989 ) .
we have taken the weighted average of these b star and solar ratios to get a logarithmic cosmic abundance ratio of @xmath155.48 @xmath42 0.04 for argon to hydrogen .
the expected dominant ionization state for argon in most neutral hydrogen regions is since the ionization potential for neutral ar is 15.76 ev compared to 13.59 ev for h. making the plausible initial assumption that all of the argon associated with neutral h is ( an assumption that must eventually be withdrawn , as discussed in
[ ionization ] ) , we determine the reduction of ar in the gas phase below its expected ratio to hydrogen through the relation @xmath44 table [ targets ] shows that the h@xmath45 column densities toward the stars in our sample are very small compared to and can therefore be neglected .
the interstellar neutral hydrogen column densities toward @xmath2 vel and @xmath1 pup are @xmath46 and @xmath47 , respectively , from a weighted average of the bohlin , savage & drake ( 1978 ) and diplas & savage ( 1994 ) values .
we adopt the determination @xmath48 toward @xmath3 cen a from rogerson & york ( 1973 ) .
a reliable neutral hydrogen column density is not available toward @xmath19 eri because the stellar h absorption badly contaminates the interstellar h spectral features . for this sight line
we estimate a total h ( i.e. 2h@xmath45 + + ) column density from the column density ; with an ionization potential of 23.33 ev , the ion resides in both the neutral and the ionized ism toward @xmath19 eri . assuming that s is undepleted ( fitzpatrick & spitzer 1997 ) and that the cosmic s / h abundance is represented by b stars [ log(s / h ) = @xmath154.91 ( snow & witt 1996 ) ] , we find that the total logarithmic interstellar hydrogen column density toward @xmath19 eri is log n(h ) = 19.36 for the s ii column density determined by rogerson , et al .
( 1973b ) .
based on the and abundances toward @xmath19 eri , it is possible that a substantial fraction of the h in the sight line is ionized .
for the three cases with good measurements of n ( ) we determined the apparent logarithmic deficiencies @xmath14 of argon with respect to hydrogen in the fifth column of table [ abund ] .
the accompanying errors include the uncertainty in the and abundance determinations , and the uncertainty in the cosmic ar - to - h abundance ratio added in quadrature .
the uncertain neutral h column density toward @xmath19 eri only allows a limit of @xmath49 toward this star .
lower limits for @xmath14 toward the other stars in our sample are given in table [ abund ] .
due to the low orbit ( 295 km ) of astro - spas , there is a possibility that telluric lines could contribute some additional absorption in our spectra .
neutral argon is an abundant species in the earth s atmosphere , so we must know the telluric absorptions effects on the interstellar profiles .
we calculated the column density of telluric neutral ar that imaps would observe through as a function of zenith angle using the global thermospheric model of hedin et al .
( 1974 ) with input from the measured solar parameters for the observation days ( k. schatten , private communication ) and the assumption of a spherical earth .
the expected ar absorption component in the atmosphere is generally small except for the largest zenith angles , i.e. , @xmath50 only when @xmath51 . with our computed time - averaged column densities for the target zenith angles of our exposures
, we estimate that the values @xmath52 , @xmath53 , @xmath54 and @xmath55 atoms @xmath56 apply to the telluric contamination of the 1066absorptions toward @xmath1 pup , @xmath2 vel , @xmath3 cen a and @xmath19 eri , respectively .
these numbers should serve only as an approximate indication of the possible telluric contamination , since the predicted column densities could vary by a factor of 2 in any given day .
the heliocentric velocity of the telluric absorption toward each of these stars is shown in figure [ na_prof ] .
any addition to ar column densities from telluric lines would tend to make the ar in a sight line appear less deficient in the ism than it actually is .
the _ copernics _ spectrometer was used to observe the neutral interstellar argon lines toward several stars .
these earlier results have been summarized by meyer ( 1989 ) . because of the great strength of the absorption features ( see [ intro ] ) and their tendency to saturate for moderately dense lines of sight
, these measurements often led to large uncertainties in the ar abundances .
for the very lightly reddened sight lines toward @xmath1 pup and @xmath2 vel , morton ( 1978 ) and morton & bhavsar ( 1979 ) have reported depletions of @xmath150.31 @xmath42 0.3 and @xmath150.20 @xmath42 0.15 , for the respective stars based on data from_copernicus_. these are very similar to the logarithmic deficiencies determined in this paper ( @xmath150.37 @xmath42 0.09 and @xmath150.18 @xmath42 0.10 ) even though the original @xmath34-values ( morton & smith 1973 ) and reference abundance ( withbroe 1971 ) used for the _ copernicus _ determinations have since been revised .
it is the accuracy of the imaps data , however , that allows us to confidently conclude that the ar is deficient with respect to our adopted reference abundance in our sample s sight lines , and that the relative reduction levels are not constant for this element in the ism .
federman et al . ( 1992 ) argue that the strength of the absorption transitions requires that sight lines with small columns of neutral gas be used to determine reliable interstellar column densities of ar .
using updated @xmath34-values and a new reference abundance , they reanalyzed the _ copernicus _ results from the low column density , low reddening sight lines toward @xmath19 vir ( york & kinahan 1979 ) and @xmath23 sco ( york 1983 ) and reaffirmed the previous conclusions that ar / h is not below the solar value for the interstellar medium near the sun .
this seems to conflict with our analysis of low column density , low reddening lines of sight which show that argon is indeed deficient in the neutral ism in the sun s vicinity .
the fact that we find varied levels of ar deficiencies , however , does not rule out such a difference as being valid .
in studies of lightly depleted elements such as o , n and c ( cardelli et al .
1996 ; meyer , cardelli , & sofia 1997 ; meyer , jura , & cardelli 1997 ) , the observed @xmath14 values ( about equal to @xmath150.37 , @xmath150.09 and @xmath150.39 , respectively as compared to solar values ) are uniform over vastly different interstellar conditions .
such findings have led to the interpretations that the abundances of these elements are not appreciably affected by the formation of dust grains , at least over the range of cloud thicknesses and internal densities surveyed .
instead , it has been argued that the total abundances of such elements in the local gas may be lower than those of the sun and possibly b stars ( fitzpatrick 1996 ) . while some of the deficiency of argon that we have detected may likewise be explained partly by this phenomenon
, the variation of @xmath14 shown for the entries for @xmath1 pup , @xmath57 vel and @xmath3 cen a in table [ abund ] indicates that some additional process must be at work .
we explore in this section the plausibility that argon could be depleted by interacting with other materials in the medium . in the interstellar medium ,
there are very few gas - phase compounds that incorporate the argon atom , since this element is chemically inert .
while ar can form a stable hydride arh@xmath58 , duley ( 1980 ) argues that the concentrations of arh@xmath58 in the interstellar medium should be insignificant . via the reaction
@xmath59 could be ignored . in
[ ionization ] we challenge the notion that ar is neutral in all h i regions . nevertheless , as indicated in table [ targets ] , there is practically no h@xmath37 present in our lines of sight .
] we are thus left to consider only the direct attachment of single argon atoms onto the surfaces of dust grains .
the characteristic e - folding time for the depletion of an element by condensing onto dust grains is given by @xmath60^{-1}~.\ ] ] the quantity @xmath61 ( spitzer 1978 , p. 162 ) is the cross sectional area of the dust grains normalized to the local density of atomic hydrogen @xmath62 .
the average thermal speed of the atoms is @xmath63 , and their sticking efficiency after impacting the grains is @xmath64 . for the impact of a gas atom that is heavier than typical atoms on the grain surfaces , an extrapolation of simple classical ( soft cube ) models for the interactions
indicate that @xmath64 should be of order unity ( burke & hollenbach 1983 ) . we can reason that for lines of sight with low average densities , such as those in this study , significant depletions must have had to occur at some time in the past when the material was in a dense cloud .
for instance , @xmath65 for ar at @xmath66k and @xmath67 is about @xmath68yr , a value that is considerably shorter than a usual cloud lifetime of @xmath69yr against collisions with other clouds ( spitzer 1978 , p. 231 ) or the average interval of @xmath70yr between the passages of shocks that can destroy the grains ( seab & shull 1986 ; jones et al .
1994 ) .
we note in passing that neutral argon has an ionization potential of 15.76 ev , i.e. , a value greater than that of hydrogen . as a consequence ,
practically all of the argon atoms are neutral within h i regions that have @xmath71 because these clouds are generally thick enough to have their internal portions shielded from external sources of ionizing radiation ( this does _ not _ apply to regions of lower density however , see [ ionization ] ) . because the atoms are neutral , the electrical charge of the grains is not a complicating factor that must be entered in calculations of an accretion rate for argon .
also , the interaction of ar as it comes in contact with a grain is not made more complicated by the energy liberated by the recombination of an ion or the creation of a chemical bond . for accretion to work
, it must not be overwhelmed by processes that tend to remove the atoms from the dust grain surfaces .
the ar atoms on such surfaces are bound by only a physical attachment caused by the van der waals interaction .
owing to the weakness of this binding , these atoms are susceptible to evaporation
. the residence of an atom on a surface can be terminated by the rare thermal fluctuations in the lattice at the dust temperature @xmath72 that can overcome the binding energy @xmath73 at the surface of the grain . the mean time interval for this escape to occur
is given by @xmath74 where @xmath75 is the vibration frequency of the grain lattice and @xmath76 is estimated to be about 900k for ar ( watson 1976 ) .
draine ( 1994 ) has estimated a characteristic grain temperature @xmath77k in the diffuse interstellar medium from the infrared emission spectrum observed by wright et al .
( 1991 ) [ see also hauser , et al . ( 1984 ) ] . substituting the values above for @xmath78 , @xmath76 , and @xmath72 into eq .
[ t_ev ] leads to the large inequality @xmath79 .
for very small grains , the situation is even worse : the discrete nature of the photoabsorptions that heat the grains leads to significant , transient excursions in @xmath72 in the positive direction .
for instance , graphite or silicate grains with diameters of order 50 or less spend a nonnegligible fraction of the time at temperatures @xmath80k ( guhathakurta & draine 1989 ) , where @xmath81 ! larger grains , however , should not have such large temperature fluctuations , and their equilibrium temperatures of around 15k [ for grains with a radius of around @xmath82 m ( draine & lee 1984 ) ] lead to @xmath83 which is much larger than the value @xmath84 given earlier .
another means for returning ar to the gas phase is photoejection .
the atoms on a grain surface can absorb the ambient stellar radiation at wavelengths very near the 1048 and 1066 resonance lines . for the excited atoms a new interaction potential
is created that ultimately results in a breaking of the bond with the grain , either because it is repulsive or causes an increase in the separation that can not be accommodated by the ground - state potential when the atom eventually decays ( watson & salpeter 1972 ) . for an average interstellar radiation flux @xmath85 at wavelengths near the ar transitions ( mathis , mezger , & panagia 1983 )
, we calculate that the photon excitation ( and ejection ) rate @xmath86 . the effect of this rapid ejection process can be overcome if , during the time they were being accreted , the ar atoms were smothered by molecules or other atoms in a time scale that was short compared with @xmath87 .
for instance , oxygen atoms ( 250 times more abundant than argon ) or oxygen - bearing compounds ( principally water ice ) could cover argon atoms after their initial attachment to the grains .
laboratory experiments have demonstrated that significant quantities of ar can be trapped during the accumulation of solid h@xmath37o on a cold surface ( bar - nun et al .
the time @xmath88 needed to cover an ar atom should be of order @xmath89 for @xmath66k ( watson & salpeter 1972 ) . to retain about one - half of the ar striking the grains
, it should be sufficient , for example , to have a cloud whose interior has a density @xmath90 that is shielded by about 0.5 magnitudes of visual extinction ( reducing @xmath3 to a value of about @xmath91 ) . in principle
, we could consider the possibility that argon atoms are locked inside coatings of water ice that accumulated at some previous time when the grains were inside a very dense cloud .
if this were true , the argon would not be quickly liberated by photodesorption after the cloud had dissipated .
however , observations of the @xmath92 m ice - band absorption feature indicate that ice - coated grains exist only within the densest portions of compact clouds . for such clouds , the approximate linear trends in @xmath93(3.05 ) vs. a@xmath94
extrapolate to zero ice absorption at a@xmath95 in different surveys ( whittet et al .
1988 ; eiroa & hodapp 1989 ; smith , sellgren , & brooke 1993 ) .
a star with very strong foreground dust absorption over a distance of 2.1kpc ( vi cygni no.12 with a@xmath96 ) shows no detectable ice band absorption ( gillett et al .
it therefore appears unlikely that our lines of sight hold grains that are coated with ice , and thus ar probably has little chance of being bound to the grains . from the above considerations , we conclude that ar should be able to experience some depletion well inside clouds that are optically thick to radiation at the wavelength of the ar transitions ( the same ones that we can observe ) , and that have large internal hydrogen densities and grains with large diameters
however , it appears likely that once the material is returned to the low density phase of the ism where we can observe the absorption lines , the argon atoms are rapidly returned to the free atomic form .
there is now good evidence for the existence of clouds that contain significant amounts of h i but with internal electron densities in the general range @xmath97 and temperatures @xmath98k ( spitzer & fitzpatrick 1993 , 1995 ; fitzpatrick & spitzer 1994 , 1997 ) . for these high - temperature clouds
it is clear that most of the electrons must come from the ionization of hydrogen ( rather than only from those elements with an ionization potential less than that of hydrogen ) .
sciama ( 1997 ) has presented some arguments that support the idea that the observations are indeed registering partially ionized regions , rather than fully ionized ones that are kinematically associated with neutral gases .
the presence of clouds that have partially ionized material seems to be plausible on theoretical grounds ( domgrgen & mathis 1994 ) , although the pervasiveness of such gas is constrained by an observed upper limit for the ratio of the emission lines [ o i ] @xmath236300/h@xmath19 ( reynolds 1989 ) .
probable external sources of ionizing photons for clouds with a hydrogen column density @xmath99 include recombination radiation from nearby , more fully ionized regions and the occasional early b - type stars that are unobscured by other clouds : a familiar example of the latter is the action of ionizing photons from @xmath17 cma that can penetrate and partially ionize the neutral cloud around the sun ( vallerga & welsh 1995 ) see [ lism ] . for thicker clouds , only the more energetic radiation from white dwarf stars ( dupree & raymond 1983 ; dupuis et al .
1995 ) , the hottest o - type stars , or emission from the surrounding hot gas ( bloch et al . 1986 ; cheng & bruhweiler 1990 ; snowden et al . 1997 ) or an evaporation interface ( slavin 1989 ) can overcome the opacity of the outer neutral material well enough to cause some ionization of hydrogen and helium in the interior .
we now examine what the relative ionization of argon should be in such regions , on the premise that it can markedly influence the ratio of observed h i to ar i. [ this phenomenon has also been investigated for elements that can be depleted onto dust by cardelli , sembach & savage ( 1995 ) , sembach & savage ( 1996 ) and welty , et al .
( 1997 ) . ]
the ionization potential of neutral argon at 15.76ev is slightly higher than that of hydrogen , which means that argon , like hydrogen , will have appreciable photoionization only for clouds that are thin enough to admit some photons from the outside .
the recombination coefficient of argon is about the same as that of hydrogen , but its cross section for photoionization is about one order of magnitude higher at most energies ( see fig .
[ pfig ] : the quantity @xmath100 is defined in eq .
[ peq ] ) . in our development of the equation that expresses the relative deficiencies of neutral hydrogen and argon caused by ionization
, we will begin with the simplest case where all of the electrons come from the ionization of hydrogen and there are no charge exchange reactions .
once the basic equation has been outlined , we will move on to consider complications that arise from charge exchange , doubly - charged ions , and the ionization of helium .
photoionization reduces the amount of neutral hydrogen by a factor @xmath101 where in this instance @xmath102 . for a hydrogen ionization rate @xmath103 and recombination rate , on the assumption that the region is optically thick to the recombination radiation at energies just above the lyman limit . for very thin regions , it would be more appropriate to include the recombination coefficient to all levels ( larger by 50% if @xmath104k ) , since a lyman limit photon will escape the region instead of re - ionizing another hydrogen atom . ]
@xmath105 the electron density is given by @xmath106 the recombination coefficient and ionization rate that apply to argon determine the outcome for the reduction in neutral argon , @xmath107 substituting eq . [ n(e)1 ] into eq .
[ delta_ar1 ] gives @xmath108^{-1}=\left [ 1+\left({\gamma({\rm ar})\alpha({\rm h})\over \gamma({\rm h})\alpha({\rm ar})}\right){n(e)\over n({\rm h})}\right]^{-1}~.\ ] ] we define a quantity @xmath109 such that @xmath110 so that the logarithmic difference in deficiencies is given by the simple expression @xmath111 to give a general impression of the expected values of @xmath100 under different circumstances for argon and other commonly observed elements , we show in fig .
[ pfig ] plots of @xmath112 as a function of the energy of the photoionizing radiation , assuming it is monoenergetic . for elements that are expected to be mostly singly ionized in h i regions ,
the @xmath100 values describe the magnitudes of the shift from singly to doubly ionized forms .
we calculated the photoionization cross sections from the analytic approximations of verner , et al .
( 1996 ) [ and verner & yakovlev ( 1995 ) for ni ] .
the recombination coefficients were evaluated from the parameters for the fitting equations given by shull & van steenberg ( 1982 ) [ and aldrovandi & pquignot ( 1974 ) for al ] . in real circumstances for the ism
, one must calculate explicitly the photoionization rates by integrating over energy the flux times the cross sections .
in deriving values of @xmath100 for fig .
[ pfig ] we assumed that @xmath113k .
the recombination coefficients for different elements change with temperature in a very similar way ( except for temperatures where dielectronic recombination becomes important ) , so @xmath100 will not change appreciably for temperatures that are somewhat different than 5000k .
[ deltad ] is an oversimplication for energies above the ionization potential of he , since an appreciable fraction of the electrons can come from the ionization of this element .
[ note that fig .
[ pfig ] shows that @xmath114 is about 10 , and this number times the cosmic abundance ratio ( he / h ) = 0.1 is 1 ] .
also , the earlier development neglected the influence of charge exchange reactions and photoionization to even higher stages . in this section
, we will evaluate more realistic ionization ratios that take these considerations into account , but we will express them by retaining the simple form of eq .
[ deltad ] and replacing the simple values of @xmath109 from eq .
[ peq ] with more correct ones ( to be denoted @xmath115 ) .
this representation is particularly useful since @xmath116 and @xmath117(h ) can often be determined by independent means , such as by observing the relative populations of atoms and ions in excited fine - structure levels ( bahcall & wolf 1968 ) . the ionization balance equations that yield the densities @xmath118 , @xmath119 and @xmath120 of the three lowest ionization stages of an element @xmath121 ( either he or ar ) in the presence of hydrogen , @xmath122n(x)=\big[\alpha(x)n(e)+c^\prime(x^+)n({\rm h})\big]n(x^+)\ ] ] @xmath123n(x^{++})\ ] ] may be combined to yield the ionization fractions @xmath124 \big\{\gamma(x^+)+\big[\alpha(x^+)n(e)+c^\prime(x^{++})n({\rm h})\big]\big\}\over \big[\alpha(x)n(e ) + c^\prime(x^+)n({\rm h})\big]\big[\alpha(x^+)n(e)+c^\prime(x^{++ } ) n({\rm h})\big]}\right)^{-1}~,\end{aligned}\ ] ] @xmath125/\gamma(x^+)}~,\end{aligned}\ ] ] and @xmath126 where the charge exchange rate constants @xmath127 and @xmath128 refer to the reactions @xmath129 and @xmath130 , respectively . [ for @xmath131k the rate constants @xmath132 and @xmath133 for the reverse reactions are very small for he and ar since their ionization potentials are greater than that of hydrogen by more than 1ev .
nevertheless we leave the @xmath132 term in eqs .
[ equi1 ] and [ n0 ] because it is very important for o and of some significance for n , both of which will be analyzed in [ lism ] .
] for regions that are partially ionized , the populations of atoms in high stages of ionization are severely limited by their large charge exchange reaction rates with h ( an effect that makes it safe for us to disregard these higher stages in the equilibrium equations ) .
[ n0]@xmath15[n+ ] with @xmath121=he must be solved simultaneously with the ionization balance for hydrogen , @xmath134 with @xmath135~,\ ] ] together with the constraints @xmath136\ ] ] and @xmath137 a satisfactory method for obtaining simultaneous solutions to eqs .
[ n0]@xmath15[n(e ) ] is to start with @xmath138 and @xmath139 equal to zero , and then iterate on the answers with gradual increases in the two helium ionization rates up to the final , correct values . after obtaining the final results for the coupled hydrogen and helium ionization balances ,
one can solve for the argon ionization balance with eqs .
[ n0]@xmath15[n+ ] .
these results can in turn be used to derive the more accurate versions of @xmath109 , @xmath140n({\rm h})\over f_0({\rm ar})[n({\rm h})+n({\rm h}^+)]}-1 \right\}\left[{n(e)\over n({\rm h})}\right]^{-1}~.\ ] ] table [ pprimes ] lists our calculations of @xmath141 ( and @xmath109 for comparison ) for monoenergetic ionizing fields at various energies . in many cases
the energies were chosen to be just below or above an ionization edge for he or ar .
sources for the photoionization cross sections and recombination coefficients were given in
[ simple ] ; the charge exchange rates were derived from the fits by kingdon & ferland ( 1996 ) to experimentally determined cross sections ( marr & west 1976 ) . for realistic cases in the ism , the radiation is usually distributed over a broad range of energy , and one must solve eqs .
[ n0]@xmath15[n(e ) ] with appropriate values of @xmath142 . weighted averages of the numbers in table [ pprimes ] for the respective contributions at various energies can give an approximate indication for @xmath141 in a complex radiation field .
rccccc h@xmath58&14&@xmath143&@xmath143&@xmath143&@xmath143h@xmath58 , ar@xmath58&16&0.764&0.761&0.732&0.518h@xmath58 , ar@xmath58&20&1.086&1.083&1.054&0.839h@xmath58 , ar@xmath58&24&1.310&1.306&1.278&1.063h@xmath58 , ar@xmath58 , he@xmath58&25&1.347&1.329&1.243&1.021h@xmath58 , ar@xmath58 , ar@xmath144 , he@xmath58&28&1.410&1.407&1.301&1.072h@xmath58 , ar@xmath58 , ar@xmath144 , he@xmath58&35&1.246&1.239&1.131&0.895h@xmath58 , ar@xmath58 , ar@xmath144 , he@xmath58&40&0.925&0.913&0.812&0.581h@xmath58 , ar@xmath58 , ar@xmath144 , he@xmath58&47&0.593&0.581&0.492&0.281h@xmath58 , ar@xmath58 , ar@xmath144 , he@xmath58&54&0.813&0.799&0.697&0.462h@xmath58 , ar@xmath58 , ar@xmath144 , he@xmath58 , he@xmath144&55&0.855&0.831&0.712&0.493h@xmath58 , ar@xmath58 , ar@xmath144 , he@xmath58 , he@xmath144&70&1.313&1.292&1.149&0.900h@xmath58 , ar@xmath58 , ar@xmath144 , he@xmath58 , he@xmath144&100&1.728&1.719&1.551&1.283 the largest deficiency of argon that we observed with imaps was for the line of sight toward @xmath3 cen a. if the true abundance of ar in the local ism is about equal to the weighted average for b stars and the sun that we derived in [ abundances ] , then the deficiency @xmath14 is equal to the value shown in table [ abund ] .
an alternative interpretation is that the abundance of ar in the ism is depressed below stellar abundances by an amount that could be as large as the value of @xmath14 that we observed for @xmath57 vel . in that case , the reduction of the neutral argon abundance toward @xmath3 cen a caused by ionization is equal to only @xmath150.43dex . in the discussion that follows
, we will adhere to this more conservative viewpoint . in order to achieve @xmath13 of @xmath150.43 , eq .
[ deltad ] tells us that @xmath145 must be about 0.23 if @xmath146 .
for @xmath113k we require that @xmath147 ^ 2\alpha({\rm h})n({\rm h})$ ] must be @xmath148 or @xmath149 for an assumed @xmath150 or @xmath151 , respectively .
for either value of @xmath62 , @xmath152 .
if all the material in front of @xmath3 cen a were inside a monolithic cloud with no internal sources of ionizing radiation , the opacity due to any substantial fraction of the hydrogen column density @xmath153 would make it difficult to achieve the required level of @xmath154 from any reasonable external source of ionizing photons .
we are therefore forced to conclude that the gas in front of @xmath3 cen a is broken into well separated fragments that each has a much lower column density for the radiation to penetrate .
just from the fact that we see two separate velocity components in the argon absorption lines ( see figs .
[ intens_prof ] and [ na_prof ] ) , we can conclude that the number of clouds is at least two .
another possibility , one that invokes a more specially contrived geometry , is that our line of sight skims through a long edge of a cloud , very close to a surface that is illuminated from the side . if we were to propose that the gas in front of @xmath3 cen a is comprised of about 5 separate clouds , each bathed in an ionizing field on all sides , the average atom will be shielded by approximately each cloud s column density @xmath155 divided by four , i.e. , @xmath156 . for an incident euv spectral energy distribution that is similar to a planck distribution with @xmath157k [ e.g. , the radiation from a star like @xmath17 cma see vallerga & welsh ( 1995 ) , or perhaps @xmath3 cen a itself ]
, the average energy of the photons that penetrate through the hydrogen is @xmath158ev .
the strength of this radiation would need to be equivalent to the flux from one such star at an unreasonably small distance of about 30pc to achieve the required @xmath154 for @xmath159 ( or 9pc for @xmath160 ) , but this requirement will be eased if there is a significant additional contribution from white dwarf stars or the soft x - ray radiation produced by hot gases in the intercloud medium . from table
[ pprimes ] we learn that @xmath161 when @xmath158ev and @xmath162 , and this value of @xmath141 is consistent with our original assumption that led to @xmath163 when @xmath164 .
additional sources of much harder radiation will raise @xmath165 which may in turn increase @xmath141 and make it easier to achieve our measured @xmath13 . we digress to consider the potential importance of ar i abundances for our understanding of the very local , partially ionized interstellar medium that surrounds our solar system , just outside the heliosphere .
the photoionization of hydrogen and helium in the local interstellar cloud ( lic ) is thought to be strongly dominated by radiation from @xmath17 cma ( vallerga & welsh 1995 ) , nearby white dwarf stars ( dupuis et al . 1995 ) , and possibly line emission from hot ( @xmath166k ) gas that surrounds the lic ( cheng & bruhweiler 1990 ) or radiation from the conductive interface at the boundary between the lic and this hot gas ( slavin 1989 ) .
as we summarize below , the simple picture of an equilibrium established by photoionization has problems that are difficult to reconcile with observations .
thus , it is worthwhile to investigate whether or not future measurements of @xmath167 to nearby stars can help us to understand better the real nature of the ionizing processes .
vallerga ( 1998 ) has combined _ euve _ measurements of 54 of the brightest stars that emit ionizing radiation in our vicinity to arrive at photoionization rates @xmath168 and @xmath169 .
the ratio of the former to the latter is not sufficient to explain the observations that indicate that the fractional ionization of helium seems to be as large as , or slightly greater than , that of hydrogen ( dupuis et al . 1995 ) , especially since @xmath170 over all temperatures of interest .
vallerga ( 1998 ) estimates that the radiation field arising from a large number of unobserved late - type stars could raise @xmath138 by approximately 14% , but clearly this increase is too small to explain the large ionization of he .
a reasonable proposal for overcoming the he ionization problem is one stated by cheng & bruhweiler ( 1990 ) , where euv radiation from hot gases known to be emitting soft x - rays provides the needed photons that are more energetic than the he ionization potential .
unfortunately , this interpretation is upset by the inability of jelinsky et al .
( 1995 ) to observe this diffuse radiation during long , dark - sky integrations with euve .
jelinsky et al .
concluded that their upper limits for the energetic line and continuum emission , if expanded to include the whole sky , translated to @xmath171 , an upper limit that is well below a value that would be sufficient to explain the ionization of he .
possible reasons for this discrepancy between the observations and simple theoretical predictions for the emission by a hot plasma may be attributable to non - equilibrium effects or a depletion of some heavy elements that are key contributors to the euv line emission .
lyu & bruhweiler ( 1996 ) proposed a possible solution to the helium ionization problem : the lic was ionized at some time in the past , possibly by the passage of a shock from a supernova , and the atoms within the lic have not yet reached their equilibrium states of lower ionization .
the time constant for such an equilibrium to be reached after the ionization is turned off is @xmath172 . to demonstrate how the abundance of ar i could give further insights on this ionization problem , we calculated photoionization equilibria using the methods outlined in [ complex ] , assuming that the radiation field of cheng & bruhweiler ( 1990 ) is in some sense correct but that , somehow , we are misinterpreting the restrictions on the steady - state flux imposed by the observations by jelinsky et al ( 1995 ) . in this exercise
, we took the local ionization rates arising from the local stellar radiation field calculated by vallerga ( 1998 ) and added them to those of an approximation - functions at 35 ev and 69 ev and made them match the specified @xmath138 and @xmath139 . to this we added a flat spectrum ( vs. @xmath173 ) between the h and he ionization edges to simulate the contribution from thermal bremstrahlung . ] to the radiation field from the hot gas calculated by cheng & bruhweiler ( 1990 ) after it has been absorbed by @xmath174 plus @xmath175 , giving the photoionization rates shown in table [ condx ] . from eqs .
[ n0]@xmath15[n+ ] with @xmath121=he , we found equilibrium conditions for the local gas that are shown in the top portion of table [ condx ] .
the effective path length of neutral gas that absorbs the ionizing radiation from the hot gas and the local electron density were both adjusted to give the best agreement with observed values of @xmath176 ( ajello et al .
1994 ; qumerais et al .
1994 ; qumerais , sandel , & de toma 1996 ) , @xmath177 ( vennes et al . 1993 ) and @xmath178 ( dupuis et al . 1995 ) .
however our value of @xmath179 that is consistent with these constraints turns out to be somewhat lower than the determination of @xmath180 by wood & linsky ( 1997 ) and considerably lower than determinations from various measurements of mg ion fractions ( frisch 1994 ; gry et al .
1995 ; lallement & ferlet 1997 ) .
it is possible that these ionization measurements are strongly influenced by material closer to the edge of the local cloud that is more ionized because there is has less shielding of the radiation ( vallerga 1998 ) . when we use eqs .
[ n0]@xmath15[n+ ] to solve for the argon ionization under conditions summarized in table [ condx ] , we find that @xmath181 , and @xmath182 ( with virtually no @xmath183 ) . substituting the values of @xmath184 , @xmath116 , @xmath185 and @xmath62 into eq .
[ pprime ] gives @xmath186 .
eq . [ deltad ] with @xmath141 in place of @xmath109 yields @xmath187
. it will be interesting to see if this prediction is confirmed by future observations of the ar i absorption lines in the spectra of very nearby stars by imaps or the far ultraviolet spectroscopic explorer ( fuse , to be launched in late 1998 ) .
if it is , one may find it difficult to abandon the notion that the material is close to ionization equilibrium .
if @xmath188 , then one could suppose that either ( 1 ) the gas was mostly ionized by a burst of photons at some time in the past , leaving the h and he to recombine at approximately equal rates , or ( 2 ) the gas was once collisionally ionized and has now cooled somewhat , but it has not yet had a chance to recombine and reach a new equilibrium ( see [ summary ] below ) .
rl @xmath27&7000k@xmath116&0.036@xmath189@xmath62&0.16@xmath189@xmath190&0.029@xmath189@xmath191&0.013@xmath189@xmath192&0.0060@xmath189@xmath193&0.00026@xmath189@xmath154&@xmath194@xmath138&@xmath195@xmath139&@xmath196@xmath197&@xmath198@xmath199&@xmath149 ccccc he&@xmath150.173&0.441&0.387&@xmath150.101n&@xmath150.122&0.906&0.221&@xmath150.049o&@xmath150.076&1.123&0.019&@xmath150.004ne&@xmath150.674&1.383&1.244&@xmath150.602ar&@xmath150.427&1.013&0.902&@xmath150.355 we have performed similar calculations for other abundant elements that are expected to be primarily neutral : he , n , o and ne . the results of those calculations ( plus a repeat of the information pertaining to he and ar ) are given in table [ other_elem ] . we draw attention to the fact that for n and o the values for @xmath200 are close to 1 and thus @xmath201 , making these two elements good substitutes for h if @xmath202 is not easy to measure . this result is a consequence of the very large charge transfer rates for these elements with hydrogen ( field & steigman 1971 ) . for n
, only the reaction @xmath203 goes easily and it is not as strong as that for o ( butler & dalgarno 1979 ) , so if @xmath204 , @xmath205 .
] , essentially locking their ionization ratios to nearly that of hydrogen .
it is clear that the simple @xmath100 values for these elements can be very misleading .
measurements of pick - up ions ( geiss et al .
1994 ) and anomalous cosmic rays50mev nucleon@xmath28 to become anomalous cosmic rays . ]
( fisk , kozlovsky , & ramaty 1974 ) have already given us some indications about the relative abundances of neutral species in the local cloud .
the pick - up ion ratios he / o , n / o and ne / o measured by geiss , et al .
( 1994 ) seem to be consistent , within the experimental errors , with the respective cosmic abundance ratios .
similarly , cummings & stone ( 1995 ) found that the abundances of the he , n , o , ne and ar components of anomalous cosmic rays seemed to reflect their cosmic abundances . to obtain these results , however , cummings & stone ( 1995 ) had to incorporate an empirical correction for the efficiency of the acceleration and propagation processes in the heliosphere .
this correction was simply a power - law in atomic mass , adjusted to make the abundances of he and ne agree with their respective cosmic ratios .
even with this adjustable parameter , it appears that the values predicted for photoionization equilibrium given in table [ other_elem ] are not easy to reconcile with the observations of anomalous cosmic rays , i.e. , he , ne and ar do not seem to be depressed below their cosmic abundances relative to o and n. it is possible that a modification of o and n abundances by the strong charge exchange reactions with solar wind protons in the heliospheric interface ( fahr , osterbart , & rucinski 1995 ) may effectively cancel the relative reductions in the noble gas abundances predicted in table [ other_elem ] .
we have measured the abundances of neutral argon relative to neutral hydrogen in the ism over 3 different lines of sight ( and obtained lower limits for 5 others ) .
these abundances are lower than those measured in b - type stars or the sun , repeating an effect seen for other elements that do not have strong depletions caused by dust formation .
in contrast to the relatively constant ism abundances of these other lightly depleted elements , our determinations indicate that argon s relative abundance seems to vary from one place to the next . we have stated theoretical arguments to suggest that , for our lines of sight , it is unlikely that an appreciable amount of argon is bound into the dust grains that accompany the gas . instead
, the missing ar can be explained by most of the atoms being ionized and thus becoming invisible to us in this survey .
the very large photoionization cross section of neutral argon , relative to that of hydrogen , can result in argon being ionized more thoroughly than hydrogen in partially ionized regions . for a line of sight that shows the largest reduction of ar (
that toward @xmath3 cen a ) we conclude that the cloud must be broken into at least about 5 subclouds to allow radiation from the outside to penetrate most of the material and cause the effect that we observe .
krypton is the only other noble gas that has been well studied in the neutral ism ( cardelli 1994 ; cardelli & meyer 1997 ) , and it too has a large cross section for ionization : @xmath206 above the ionization threshold ( marr & west 1976 ) . because kr has such a low cosmic abundance [ @xmath207 , ( anders & grevesse 1989 ) ] the uv absorption feature is very weak .
thus , the studies of kr / h have featured sight lines that have @xmath208 , and regions with this much column density are expected to be fully shielded and neutral , except for those elements with ionization potentials below that of h. hence it is no surprise that the kr investigations do not uncover the abundance fluctuations that we see here for ar .
the ratio of ar i to h i ( or some other element that is not usually depleted in the ism ) may offer a critical insight on the nature of the ionizing process that is responsible for the many neutral clouds that seem to be unexpectedly rich in free electrons .
if such clouds are ionized mainly by photons , we should expect to see a reduction in the apparent abundance of ar i. alternatively , if the ar i seems normal , partial recombination from a fully ionized condition or the effects of collisional ionization may be more plausible interpretations ( trapero et al . 1996 ) .
for instance , in collisional ionization equilibrium up to temperatures of around @xmath209k , the fractional abundances of argon in its neutral form are of order or slightly more than those of oxygen ( shull & van steenberg 1982 ) . above @xmath209k , @xmath210 , but both elements have less than 1% of their atoms in neutral form at these temperatures . for a gas that is cooling either isochorically or isobarically with some lag in the recombination , the fractional ionizations of o and ar are nearly identical for @xmath211 ( which means that @xmath212k ) ( benjamin & shapiro 1997 ) . in future studies of uv absorption lines , it will be interesting to confirm the validity of our proposal that partially ionized regions offer an explanation for the reduction in ar i. for instance , is there a correlation between @xmath213 and @xmath116 ( as indicated by the fine - structure excitations of c ii or si ii ) or the existence of atoms in moderately high stages of ionization , such as si iii or s iii ? if such correlations hold true , then we can use the ar i abundances as either a powerful diagnostic on the nature of the ionization or , alternatively , as an indicator of whether the higher ions arise from a completely ionized classical h ii region ( where there should be little ar i or h i ) or a partially ionized cloud of the type described here .
support for flying imaps on the orfeus - spas - i mission and the research carried out by ebj came from nasa grant nag5 - 616 to princeton university .
ujs performed most of his work on this project while he held a national research council - nasa / gsfc research associateship , and additional support was provided by the nasa - ltsarp grant nag5 - 3539 to villanova university .
most of the development of the data reduction software was done by ebj shortly after the mission , while he was supported by a research award for senior u.s .
scientists from the alexander von humboldt foundation and was a guest at the institut fr astronomie und astrophysik in tbingen , germany .
we thank drs .
b. t. draine , t. m. tripp and j. v. vallerga for their comments on an early draft of this paper .
the orfeus - spas project was a joint undertaking of the us and german space agencies , nasa and dara .
the successful execution of our observations was the product of efforts over many years by engineering teams at princeton university observatory , ball aerospace systems group ( the industrial subcontractor for the imaps instrument ) and daimler - benz aerospace ( the german firm that built the astro - spas spacecraft and conducted mission operations ) .
wright , e. l. , mather , j. c. , bennett , c. l. , cheng , e. s. , shafer , r. a. , fixsen , d. j. , eplee , r. e. , jr . , isaacman , r. b. , read , s. m. , boggess , n. w. , gulkis , s. , hauser , m. g. , janssen , m. , kelsall , t. , lubin , p. m. , meyer , s. s. , moseley , s. h. , jr . , murdock , t. l. , silverberg , r. f. , smoot , g. f. , weiss , r. , & wilkinson , d. t. 1991 , , 381 , 200 | we use the absorption features from the resonance transitions of neutral argon at 1048 and 1066 to determine interstellar argon abundances or their lower limits toward nine early - type stars .
these features were observed with the interstellar medium absorption profile spectrograph ( imaps ) along sight lines with low reddening and low fractional abundances of molecular hydrogen .
four of the sight lines in our sample have reliably measured interstellar ar abundances , of which three also have good determinations of @xmath0 .
we find that the interstellar ar is below its solar and b - star abundance with respect to hydrogen toward @xmath1 pup , @xmath2 vel and @xmath3 cen a with ( logarithmic ) reduction factors @xmath4 , @xmath5 , and @xmath6 , respectively .
while ar can condense onto the surfaces of dust grains in the interiors of dense clouds , it is unlikely that argon atoms are depleted by this process in the low - density lines of sight considered in this study .
instead , we propose that the relatively large photoionization cross section of ar makes it much easier to hide in its ionized form than h. in regions that are about half ionized , this effect can lower ar i / h i by @xmath7 to @xmath8 , depending on the energy of the photoionizing radiation and its intensity divided by the local electron density .
we apply this interpretation for the condition of the gas in front of @xmath3 cen a , which shows the largest deficiency of ar . also , we determine the expected magnitudes of the differential ionizations for he , n , o , ne and ar in the partly ionized , warm gas in the local cloud around our solar system . for the local cloud and
others that can be probed by uv studies , the observed ar i to h i ratio may be a good discriminant between two possible alternatives , collisional ionization or photoionization , for explaining the existence of partly ionized regions . |
the wavefunction approach has been very successfully applied to study single layer quantum hall ( slqh ) systems at laughlin series @xmath2 @xcite and jain s series at @xmath3 @xcite .
one of main reasons for the success of the wavefunction approach in slqh is that there is a gap in the bulk , suitable wavefunctions @xcite can describe both the groundstate and low energy excitations quite accurately .
its accuracy can be checked easily by exact diagonalization in a finite size system whose size need only go beyond a few magnetic length .
spherical geometry can be used to get rid of edge state effects quite efficiently . in general ,
trial wave function approach is a very robust approach to study slqh and multi - components systems as long as there is a gap in the bulk .
the gap protects the many properties of the system such as charge density distributions and energies from being sensitive to some subtle details of wavefunctions . however , the situation could be completely different in spin - polarized bilayer quantum hall system ( blqh ) at total filling factor @xmath4 .
this system has been under enormous experimental and theoretical investigations over the last decade @xcite . when the interlayer separation @xmath1 is sufficiently large , the bilayer system decouples into two separate compressible @xmath5 layers @xcite .
however , when @xmath1 is smaller than a critical distance @xmath6 , even in the absence of interlayer tunneling , the system undergoes a quantum phase transition into a novel spontaneous interlayer coherent incompressible phase which is an excitonic superfluid state ( esf ) in the pseudospin channel @xcite . in @xcite ,
halperin proposed the @xmath0 wavefunction to describe the groundstate of the esf state .
starting from the @xmath0 wavefunction , using various methods , several authors @xcite discovered a neutral gapless mode ( ngm ) with linear dispersion relation @xmath7 and that there is a finite temperature kosterlitz - thouless ( kt ) phase transition associated with this ngm . by treating the two layer indices as two pseudo - spin indices , girvin , macdonald and collaborators mapped the bilayer system into a easy plane quantum ferromagnet ( epqfm ) @xcite ( which is equivalent to the esf ) and explored many rich and interesting physical phenomena in this system . as first pointed out in @xcite
, the @xmath0 wavefunction may not be qualitatively good at finite @xmath1 , because @xmath0 is a broken symmetry state in a direction in @xmath8 plane of isotropic ferromagnet at @xmath9 instead of a easy - plane ferromagnet at finite @xmath1 .
the ngm is a hallmark of the interlayer coherent quantum hall state .
its existence is expected to dramatically alter the properties of the wavefunctions of the ground state , quasi - hole and quasi - particle . in @xcite ,
g. s. jeon and one of the authors studied properties of essentially all the known trial wavefunctions of ground state and excitations in bilayer quantum hall systems at the total filling factor @xmath10 . the results indicated that qualitatively good trial wave functions for the ground state and the excitations of the interlayer coherent bilayer quantum hall system at finite @xmath1 are still not available and searching for them remains an important open problem . specifically , they investigated the properties of the quasi - hole wave function , meron wavefunction and a pair of meron wavefunction built on @xmath0 state which have superscripts `` prime '' in this paper : @xmath11 where @xmath12 is the halperin s @xmath0 wavefunction : @xmath13 where @xmath14 in the balanced case and @xmath15 and @xmath16 are the coordinates in layer 1 and layer 2 respectively . in the following ,
we suppress the exponential factor .
these quasi - hole and meron wavefunctions differ only by `` normalization factors '' .
as shown in @xcite , the normalization factor @xmath17 is accurate only at long distance limit @xmath18 limit . near the origin , the `` meron '' and the " quasi - hole have similar behaviors .
normalization factors have been shown not to be important in single layer quantum hall ( slqh ) systems . however , as shown in @xcite , they make a dramatic difference in blqh .
although the smallest meron has a localized charge @xmath19 and logarithmically divergent energy , the charge of the quasi - hole excitation extends over the whole system and its energy also diverges linearly as the area of the system size .
this indicates that the quasi - hole wavefunction is not a good trial wavefunction for any low energy excitations .
the meron wavefunction is not a good trial wavefunction either , because it ignores the strong interlayer correlations @xcite .
it was found the energy of the possible wavefunction of a pair of merons in eqn.[prime ] increases quadratically @xmath20 instead of logarithmically as the separation of the pair increases .
all the results achieved in @xcite indicate that qualitatively good trial wave functions in the interlayer coherent bilayer quantum hall system at finite @xmath1 , both for ground state and excitations , are still unknown and searching for them remains an important open problem .
so the wavefunction approach to blqh is much more difficult and far less powerful in blqh than in slqh .
fortunately , effective theory approaches such as epqfm approach @xcite and composite boson theory approach @xcite circumvent this difficulty associated with the unknown wavefunction at any finite @xmath1 and are very effective to bring out most of the interesting phenomena in the pseudo - spin sector in this system .
in fact , all these effective theories start from the insights gained from halperin s @xmath0 wavefunction which is exact at @xmath9 . in a series of papers @xcite ,
one of the authors developed a systematic composite boson approach to study balanced and im - balanced bi - layer quantum hall systems in rather details .
the theory puts spin and charge degree freedoms in the same footing , explicitly bring out the spin - charge connection and classify all the possible excitations in a systematic way .
then he pushed the theory further to understand novel phases and quantum phase transitions as the distance between the two layers is changed .
he found that starting from the well studied excitonic superfluid ( esf ) state , as distance increases , the instability driven by magneto - roton minimum collapsing at a finite wavevector in the pseudo - spin channel leads to the formation of a pseudo - spin density wave ( psdw ) at some intermediate distances .
he constructed a quantum ginsburg - landau theory to study the transition from the excitonic superfluid ( esf ) to the psdw and analyze in detail the properties of the psdw .
he showed that a square lattice is the favorite lattice and the correlated hopping of vacancies in the active and passive layers in the psdw state leads to very large and temperature dependent drag observed in the experimental . in the presence of disorders ,
the properties of the psdw are consistent with all the experimental observations @xcite in the intermediate distances .
further experimental implications of the psdw are given .
then he extended the composite boson theory to study slightly im - balanced blqh . in the global @xmath21 symmetry breaking excitonic superfluid side , as the imbalance increases , the system supports continuously changing fractional charges . in the translational symmetry
breaking psdw side , there are two quantum phase transitions from the commensurate psdw to an in - commensurate psdw and then to the excitonic superfluid state .
these results explained the experimental observations in @xcite very nicely .
the author found that the theory can be easily extended to study some additional interesting phenomena in trilayer quantum hall systems @xcite .
it was concluded in @xcite that field theory approaches are much more powerful in blqh than in slqh .
obviously , the cb theory circumvent this difficulty associated with the unknown wavefunction at any finite @xmath1 and is used to achieve all these interesting and important results at two different distance regimes without knowing the precise wavefunctions for the ground state and excitations .
it would be interesting to use the cb theory to address the important and outstanding problem avoided in @xcite and in all the other pervious work that finding the good ground state and low energy excited wavefunction for blqh at any finite @xmath1 . in slqh , the cb theory developed in @xcite was used to re - derive the already well known laughlin s wavefunctions for ground state and quasihole at @xmath2 . as said previously
, the gap in the bulk protects the properties of the system such as charge density distributions and energies from being sensitive to some subtle details of wavefunctions .
here we are facing a more difficult and interesting task : to derive these unknown wavefunctions at finite @xmath1 .
the rest of the paper is organized as the following . in sec .
ii , in order to be self - contained , we review briefly the cb approach and its dual action developed in @xcite which are needed to derive the wavefunctions in the following sections . in sec .
iii , using the formalism presented in sec .
ii , we derive the ground state wavefunction which is different from the @xmath0 wavefunction at any finite @xmath22 . in sec.iv , using the dual action presented in sec .
ii , we derive the quasi - hole wavefunction and compare it with the `` quasi - hole '' and `` meron '' wavefunction built on @xmath0 wavefunction listed in eqn.[prime ] . in sec.v , we derive a pair of meron wavefunction with charge @xmath23 and compare it with the `` pair meron '' wavefunction built on @xmath0 listed in eqn .
[ prime ] . in sec.vi , we look at the instability in the ground state wavefunction as distance approaches @xmath6 . finally , we reach conclusions in sec .
some caveats related to the lowest landau level ( lll ) projection of the wavefunctions are also pointed out .
we note that there is also an alternative approach in @xcite .
in this section , we briefly review the formalism developed in @xcite which is needed to derive the wavefunctions in the following sections . consider a bi - layer system with @xmath24 ( @xmath25 ) electrons in left ( right ) layer and with interlayer distance @xmath1 in the presence of magnetic field @xmath26 ( fig.1 ) : @xmath27 where electrons have _ bare _ mass @xmath28 and carry charge @xmath29 , @xmath30 are electron operators in top and bottom layers , @xmath31 are normal ordered electron densities on each layer .
the intralayer interactions are @xmath32 , while interlayer interaction is @xmath33 where @xmath34 is the dielectric constant . performing a singular gauge transformation @xcite : @xmath35 where @xmath36 is the angle between the vector @xmath37 and the horizontal axis .
@xmath38 is the total density of the bi - layer system .
note that this transformation treats both @xmath39 and @xmath40 on the same footing .
this is reasonable only when the distance between the two layers is sufficiently small .
it can be shown that @xmath41 satisfies all the boson commutation relations .
we can transform the hamiltonian eqn.[first ] into the lagrangian in coulomb gauge : @xmath42 where @xmath43 and @xmath44 .
the chern - simon gauge field is @xmath45 . in coulomb gauge , integrating out @xmath46 leads to the constraint : @xmath47 .
note that if setting @xmath48 , then the above equation is identical to a single layer with spin in the absence of zeeman term , so the lagrangian has a @xmath49 pseudo - spin symmetry .
the @xmath50 term breaks the @xmath49 symmetry into @xmath21 symmetry . in the blqh at finite @xmath1 ,
@xmath51 , so the system is in the easy - plane limit .
we can write the two bosons in terms of magnitude and phase @xmath52 the boson commutation relations imply that @xmath53 = i \hbar \delta_{ab } \delta ( \vec{x}-\vec{x}^{\prime } ) $ ] . after absorbing the external gauge potential @xmath54 into @xmath55 , we get the lagrangian in the coulomb gauge : @xmath56^{2 } + \frac{1}{2 } \delta \rho^{+ } v_{+ } ( \vec{q } ) \delta \rho^{+ } - \frac { i } { 2 \pi } a_{0 } ( \nabla \times \vec{a } ) \nonumber \\ & + & \frac{i}{2 } \delta \rho^{- } \partial_{\tau } \theta^{- } + \frac { \bar{\rho } f } { 2 m } ( \frac{1}{2 } \nabla \theta_{- } ) ^{2 } + \frac{1}{2 } \delta \rho^{- } v_{- } ( \vec{q } ) \delta \rho^{- } - h_{z } \delta \rho^{- } \label{main}\end{aligned}\ ] ] where @xmath57 which is equal to 1 at the balanced case and @xmath58 plays a role like a zeeman field . performing the duality transformation on eqn.[main ]
leads to the dual action in terms of the vortex degree of freedoms @xmath59 and the corresponding dual gauge fields @xmath60 : @xmath61 where @xmath62 are the two source fields .
it is useful to stress that the dual cs term only appears in the charge sector . for simplicity
, we only consider the balanced case .
putting @xmath63 and @xmath64 into eqn.[main ] , we get the lagrangian in the balanced case where the symmetry is enlarged to @xmath65 .
in this section , we derive the ground state wavefunction from the formalism reviewed in the last section . from eqn.[main ] , we can find the corresponding hamiltonian in the charge sector : @xmath66 \label{ch}\ ] ] where @xmath67 and @xmath68 = i \hbar \delta ( \vec{q } + \vec{q}^{\prime } ) $ ] .
representing @xmath69 by @xmath70 , in the long wavelength limit , neglecting the coulomb interaction @xmath71 which is less singular than @xmath72 , we find the ground state wavefunction in the charge sector : @xmath73 = \exp [ - \frac{1}{2 } \sum_{q } \frac { 2 \pi } { q^{2 } } \delta \rho_{+ } ( - \vec{q } ) \delta \rho_{+ } ( \vec{q } ) ] \ ] ] in the following , we use @xmath74 to stand for the complex coordinate @xmath75 .
using @xmath76 and transforming to the position space , it is simply the modulus of the @xmath77 wavefunction in the balanced case @xmath78 .
this is the wavefunction in the bosonic picture . in order to get the wavefunction in the original fermionic picture
, we need to perform the inverse of the sgt in eqn.[singb ] on the bosonic wavefunction . in the first quantization form ,
the inverse is : @xmath79 performing the inverse transformation on the modulus leads to the @xmath77 wavefunction in the fermionic coordinates : @xmath80 in contrast to the slqh , there is also an additional pseudo - spin sector in the blqh which contains the most interesting physics . from eqn.[main ] , we can find the corresponding hamiltonian in the spin sector is : @xmath81 \label{spinh}\ ] ] where @xmath82 and @xmath83 = i \hbar \delta ( \vec{q } + \vec{q}^{\prime } ) $ ] ; @xmath84 is the spin stiffness @xcite . at small @xmath85 , @xmath86 @xcite where @xmath87 and @xmath88 remains a constant at small distances @xcite .
it is important to stress that this form of @xmath89 has the shape displayed in the fig.1 , it not only has a phonon part near @xmath90 , also has a roton part near @xmath91 where @xmath92 is the magnetic length .
representing @xmath93 by @xmath94 , we find the ground state wavefunction in the spin sector : @xmath95 = \exp [ - \frac{1}{2 } \sum_{q } \frac { \sqrt { v_{-}(q)/\rho_{e } } } { q } \delta \rho_{- } ( - \vec{q } ) \delta \rho_{- } ( \vec{q } ) ] \label{spinw}\ ] ] it is easy to see that the above equation make senses only when @xmath96 is positive for all @xmath85 . at @xmath9 ,
@xmath97 , @xmath98 , then eqn.[spinw ] becomes : @xmath99 = \exp [ - \frac{1}{2 } \sum_{q } \sqrt { c /\rho_{e } } \delta \rho_{- } ( - \vec{q } ) \delta \rho_{- } ( \vec{q } ) ] = const .
\label{spinw0}\ ] ] at any finite distance , as long as @xmath100 in the fig.1 , so the roton has a large gap , we can neglect the contributions from the roton part in fig.1 and only focus on the phonon contributions . in the long wavelength limit @xmath101 , @xmath102 . using @xmath103 and transforming to the coordinate space
, it is @xmath104 \label{spinw1}\ ] ] obviously , the above equation only holds at small distance @xmath105 and in the long distance limit @xmath106 @xcite .
the total wavefunction is : @xmath107 it is easy to see that the total wavefunction coincides with the @xmath0 wavefunction only in @xmath108 limit . at any finite @xmath1 , it has an extra factor from the gapless spin sector @xmath109 .
note that this extra spin factor eqn.[spinw1 ] is not in the lll , this should not be too worrisome , because similar to the meron wavefunction listed in eqn.[prime ] , modulus of the coordinates could appear in the long distance limit where eqn.[spinw1 ] hold .
by inserting one static vortex at the origin in layer 1 by setting @xmath110 or layer 2 by setting @xmath111 , from the dual action eqn.[dual ] , we will first try to derive the quasi - hole wavefunction and compare this quasi - hole wavefunction with the known `` quasi - hole '' wavefunction and `` meron '' wavefunction written down in @xcite . in order to derive the quasi - hole wavefunction
, we have to resort to the dual action eqn.[dual ] in the balanced case where the last term vanishes .
setting the two sources @xmath112 , in the coulomb gauge @xmath113 , eqn.[dual ] becomes : @xmath114 note the absence of cs term in the spin sector .
we only consider static vortices , so @xmath115 . integrating out @xmath116 and @xmath117 and transforming into the coordinate space lead to : @xmath118 the corresponding hamiltonian is : @xmath119 where @xmath120 = i \hbar \delta ( \vec{q } + \vec{q}^{\prime } ) $ ] . from the above hamiltonian in the bosonic representation ,
we can see the charge sector and spin sector remain `` decoupled '' @xcite .
due to the absence of the c - s term in the spin sector , the inserted vortex only shifts the total density variable in the charge sector , but does not couple to the relative density in the spin sector , so the hamiltonian in the spin sector remains the same as the ground state one eqn.[spinh ] , the corresponding wavefunction remains the same as the ground state one in the spin sector eqn.[spinw ] .
all the effects of the inserted vortex are encoded in the charge sector .
again , neglecting the coulomb interaction @xmath121 in the long wavelength limit , we find that the wavefunction in the charge sector is : @xmath122\ ] ] transforming to the coordinate space and setting @xmath123 @xcite lead to : @xmath124 \nonumber \\ & = & \prod^{n_{1}}_{i}|z_{i}|^{\frac{1}{2}}\prod^{n_{2}}_{i}|w_{i}|^{\frac{1}{2}}\prod^{n_{1}}_{i=1 } \prod^{n_{2}}_{j=1 } |z_{i}-w_{j } | \prod^{n_{1 } } _ { i < j } the sgt for the quasi - hole could be different from that for the groundstate . if one inserts a vortex at the origin at the layer 1 in the _ boson _ lagrangian eqn.[boson ] , in order to recover the original electronic hamiltonian eqn.[first ] , @xmath125 in eqn.[singg ] is needed to remove the c - s term , an additional sgt @xmath126 is needed to remove the effects of the inserted vortex . in the first quantization , it is easy to show that @xcite : @xmath127 the total sgt for the quasi - hole at the layer 1 is @xmath128 . performing the sgt on eqn.[qhm ] ,
we get the quasi - hole wavefunction , @xmath129 where @xmath130 is the ground state wavefunction eqn.[totalg ] and @xmath14 in the balanced case .
note that there is no singularity at the origin .
note that even in the @xmath108 limit , eqn.[qh ] differs from both the `` quasi - hole '' and the `` meron '' wavefunction listed in eqn.[prime ] .
this should not cause any problem .
it is known that @xmath0 wavefunction is the exact wave function in the @xmath131 limit . but
both both the `` quasi - hole '' and the `` meron '' wavefunction listed in eqn.[prime ] make sense only at finite distances .
it is known that the lowest energy excitation at @xmath9 is a skymion carrying charge 1 , while the `` quasi - hole '' and the `` meron '' carry total charge @xmath19 , so they are not valid excitations anymore at @xmath9 .
we conclude the quasi - hole eqn.[qh ] make sense only at finite @xmath1 .
it is not interesting to take @xmath108 limit to this equation @xcite .
compared to the `` quasi - hole '' and the `` meron '' wavefunction listed in eqn.[prime ] , we can see that there are two modifications in the eqn.[qh ] : ( 1 ) the ground state wavefunction is the correct one eqn.[totalg ] instead of the @xmath0 wavefunction . ( 2 ) the prefactor is different from both the `` quasi - hole '' and the `` meron '' .
we expect this prefactor takes care of the strong interlayer correlations .
all the wavefunctions in eqn.[prime ] are built upon the @xmath132 wavefunction .
as suggested in @xcite and explicitly shown in this paper , @xmath133 wavefunction is not even qualitatively correct at any finite @xmath1 . as shown in @xcite ,
although prefactors are not important in slqh due to the gap in the bulk , they maybe crucial in blqh due to the gapless mode in the interlayer correlations .
the two factors maybe responsible for the `` quasi - hole '' s charge distribution spreading over the whole system and its energy diverges linearly with the area of the system .
although , the `` meron '' wavefunction s energy is only logarithmically divergent , it ignores the strong interlayer correlations , so it is not a good trial wavefunction either .
we propose that the quasi - hole wave function eqn.[qh ] not only has logarithmically divergent energy , well localized charge distribution , but also correct interlayer correlations .
now we put two vortices into the blqh system .
one is on the top layer at @xmath134 and the other is on the bottom layer at @xmath135 .
the only change from the quasi - hole calculation in the last section is @xmath136 @xcite .
again , due to the lack of the cs term in the spin sector , the hamiltonian in the spin sector remains the same as the ground state one eqn.[spinh ] , so the wavefunction is not affected at all by the insertion of the two vortices and remains the same as the ground state one in the spin sector eqn.[spinw ] .
all the effects of the inserted two vortices are encoded in the charge sector .
the wavefunction in the charge sector in the bosonic picture is : @xmath137ln|x - y| \nonumber \\ & & [ ( \frac{1}{2}(\delta(\vec{y}-z_{0})+\delta(\vec{y}-w_{0}))+(\sum_{i}\delta({\vec{y}}-z_{i})+ \delta({\vec{y}}-w_{i})-\bar{\rho } ) ) ] \nonumber \\ & & = \prod^{n_{1}}_{i}|z_{i}-z_{0}|^{\frac{1}{2}}\prod^{n_{1}}_{i}|z_{i}-w_{0}|^{\frac{1}{2 } } \prod^{n_{2}}_{i}|w_{i}-z_{0}|^{\frac{1}{2}}\prod^{n_{2}}_{i}|w_{i}-w_{0}|^{\frac{1}{2 } } \label{pc}\end{aligned}\ ] ] it is easy to see that in the bosonic picture , the above wavefunction in the charge sector is symmetric under @xmath138 or @xmath139 separately .
this is under expectation , because the two layers are completely symmetric in the charge sector . just like deriving @xmath126 for the quasi - hole
, we can get @xmath140 for a pair of vortices inserted at @xmath141 in top layer and @xmath142 at the bottom layer : @xmath143 the total sgt for the meron pair is @xmath144 . performing the sgt on eqn.[pc ] ,
we get the wavefunction for a pair of quasihole : @xmath145 where @xmath130 is the ground state wavefunction eqn.[totalg ] and @xmath14 in the balanced case .
note that the pair wavefunction is not symmetric under @xmath138 or @xmath139 separately anymore .
this is because the sgt @xmath146 eqn.[v2 ] is not .
of course , it is still symmetric under @xmath147 simultaneously .
if we insert the two vortices at the same point , namely , putting @xmath148 in the above equation , as expected , we get : @xmath149 this corresponds to insert a single vortex through the two layers .
in contrast to the quasi - hole excitation eqn.[qh ] , eqn.[qh0 ] carries charge 1 and remains a valid wavefunction even at @xmath150 . indeed , in the @xmath108 limit
, it recovers the `` meron pair '' wavefunction listed in eqn.[prime ] .
if one splits the whole vortex , it will evolve into the pair wavefunction eqn.[pair ] .
the `` pair meron '' wavefunction built on @xmath0 is listed in [ prime ] ( essentially eqn.(110 ) in ref.10 ) .
it was shown that its energy @xmath151 instead of logarithmically as naively expected , because the charges are extended between @xmath141 and @xmath142 .
similar to the quasi - hole wavefunction eqn.[qh ] , there are two modifications in the eqn.[pair ] : ( 1 ) the ground state wavefunction is the correct one eqn.[totalg ] ( 2 ) the prefactor is different .
we expect this prefactor takes care of the strong interlayer correlations between the two vortices , the pair wave function not only has logarithmically divergent energy , well localized charge distribution , but also correct interlayer correlations .
when the distance is sufficiently small , the blqh is in the esf phase , we expect the ground state , quasi - hole and pair wavefunctions eqns.[totalg],[qh],[pair ] only hold in the esf phase . when the distance becomes sufficiently large ,
the two layers become two weakly coupled fermi liquid ( fl ) layers .
all these wavefunctions completely break down .
new set of wavefunctions are needed . although the esf phase and fl phase at the two extreme distances are well established , the picture of how the esf phase evolves into the two weakly - coupled fl states was not clear , namely , the nature of the intermediate phase at @xmath152 was still under debate .
recently , starting from the well studied excitonic superfluid ( esf ) state , as distance increases , one of the authors found @xcite that the instability driven by magneto - roton minimum collapsing at a finite wavevector in the pseudo - spin channel leads to the formation of a pseudo - spin density wave ( psdw ) at some intermediate distances .
he constructed a quantum ginsburg - landau theory to study the transition from the esf to the psdw and analyze in detail the properties of the psdw .
he showed that a square lattice is the favorite lattice .
as shown in @xcite , it is the original instability in @xmath153 which leads to the magneto - roton minimum in the fig.1a . by looking at the two conditions @xmath154 and @xmath155 ,
it is easy to see that @xmath96 indeed has the shape shown in fig.1b .
when @xmath156 , the minimum of @xmath96 at @xmath157 has a gap , the system is in the esf state , this is always the case when the distance @xmath1 is sufficiently small .
however , when @xmath158 , the minimum collapses and @xmath159 diverges at @xmath160 , which signifies the instability of the esf to the pseudo - spin density wave ( psdw ) formation .
when @xmath161 , the minimum drops to negative , the system gets to the psdw state , this is always the case when the distance @xmath1 is sufficiently large .
the zero temperature phase diagram in the balanced case as the distance between the two layers increases .
esf where @xmath162 stands for excitonic superfluid , psdw where @xmath163 stands for pseudo - spin density wave phase , fl stands for fermi liquid .
( a ) energy dispersion relation @xmath164 in these phases .
( b ) @xmath165 in these phases .
the cross in the psdw means the negative minimum value of @xmath166 is replaced by the psdw .
the two order parameters were defined in @xcite . in reality
, the instability happens _ before _ the minimum collapses .
we can easily see the instability from the ground state wavefunction eqn.[totalg ] .
the only distance dependence in the charge sector is encoded in the @xmath121 in the bosonic hamiltonian in the charge sector eqn.[ch ] , but this dependence is ignored in the @xmath0 wavefunction in the charge sector . the @xmath1 dependence in @xmath167 is smooth anyway .
essentially all the distance dependence is encoded in the ground state wavefunction in the spin sector eqn.[spinw ] . as can be seen from fig.1
, the instability happens at @xmath168 where @xmath50 becomes negative , but the spin stiffness @xmath169 remains _ non - critical _ through the esf to psdw transition . as @xmath170 , the sum over @xmath85 in eqn.[spinw ] becomes dominated by the regime @xmath171 .
when @xmath172 , eqn.[spinw ] breaks down .
a new wavefunction to describe the translational symmetry breaking psdw state is needed .
some trial wavefunctions are proposed in @xcite .
it would be interesting to derive the new wavefunction of the psdw state from the cb theory .
blqh differs from the slqh by its symmetry breaking ground state and associated neutral gapless mode in the pseudo - spin sector .
due to the gapless mode in the bulk , the groundstate wavefunctions could be considerably different from the well known @xmath0 wave function @xcite
. the low energy excited states could also be sensitive to details such as normalization factors .
one important problem is to find good trial wavefunctions for the ground state and low energy excited states .
we investigated this important open problem from the cb theory developed previously in @xcite to study blqh systematically .
we derived the ground state , quasi - hole and a pair of quasi - hole wavefunctions from cb theory and its dual action by the following procedures : we first performed the singular gauge transformation eqn.[singb ] to transform a fermionic problem into a bosonic problem , then found that all the wavefunctions in the bosonic picture are always the product of two parts , one part in the charge sector and the other in the spin sector .
all the distance dependence are encoded in the spin part , while all the excitations only happen in the charge sector . after transforming back to the original electron picture by proper inverse sgt s
, we get the final wavefunctions in the electron coordinates .
we found that the inverse sgt s are different in ground state , meron and a pair of merons . by considering the differences carefully , we derived all these wavefunctions in the original electronic picture a systematical way . at any finite @xmath1 ,
the ground state wavefunction in the charge sector is the same as the @xmath0 wavefunction , while that in the spin sector is highly nontrivial due to the gapless mode .
so the total groundstate wavefunction differs from the well known @xmath0 wavefunction at any finite @xmath1 . in the bosonic picture , when inserting vortices in the ground state , the spin part remains the same due to the lack of cs term in this sector , while the charge part changes accordingly .
however , due to the insertion of vortices , in order to recover the original electronic problem , the inverse sgt differs from that in the ground state . after transforming back to the original electron problem by the inverse sgt s , we showed that the quasi - hole and a pair of quasi - holes wavefunctions contain non - trivial normalization factors as shown in [ qh],[pair ] .
we expect that the quasi - hole and pair wave function not only has logarithmically divergent energy , well localized charge distribution , but also correct interlayer correlations .
it is important to test these trial wavefunctions by qmc simulations performed in @xcite for the states listed in eqn .
[ prime ] .
we also investigated the instability encoded in the spin sector which leads to the psdw solid formation proposed in @xcite . because the cb field theory has been used to describe the tri - layer quantum hall systems very successfully , the analysis in this paper can be easily extended to derive the wavefunctions in the tlqh @xcite .
it is well known that cb approach is not a lowest landau level ( lll ) approach @xcite , it is very difficult to incorporate the lll projection into the cb approach .
this may be partially responsible for the spin part of the ground state wavefunction eqn .
[ spinw1 ] not in the lll level .
but as explained below eqn.[totalg ] , this should not be too worrisome , because eqn.[spinw1 ] works only in long distance anyway .
as shown in @xcite and listed in eqn.[prime ] , in the long wavelength limit , the meron wavefunction s normalization factor contains modulus which is not in the lll either .
how to getting precise short distance behaviors of these wavefunctions from the cb theory remains an open problem .
j. ye thanks s. c. zhang for helpful discussions
. he also thanks yu lu and xiang tao for the hospitality during his visit to the institute for theoretical physics of china .
kun yang , k. moon , l. zheng , a. h. macdonald , s. m. girvin , d. yoshioka , and shou?cheng zhang , phys .
72 , 732?735 ( 1994 ) ; k. moon , h. mori , kun yang , s. m. girvin , and a. h. macdonald , l. zheng , d. yoshioka and shou?cheng zhang , phys . rev .
b 51 , 5138?5170 ( 1995 ) s. m. girvin and a. h. macdonald , phys .
58 , 1252 ( 1987 ) ; s. z. zhang , t. h. hansson and s. kivelson , phys .
62 , 82 ( 1989 ) .
s. z. zhang , int .
b 6 , 25 ( 1992 ) ; n. read , phys
62 , 86 ( 1989 ) .
recently , trial wavefunctions involving the coexistence of composite bosons and composite fermions at finite distance were proposed by s. h. simon , e. h. rezayi , and m. v. milovanovic , phys .
lett . * 91 * , 046803 ( 2003 ) .
m. v. milovanovic and stanic , phys . rev .
b * 72 * , 155306 ( 2005 ) .
we put `` quote '' here because eqn.[decouple ] is decoupled only in its _ face _ form .
as shown explicitly in @xcite , in the balanced case , there is a spin - charge separation _ only _ in the ground state .
however , in any excited states such as meron excitations , there are spin - charge connection which leads to the correct meron fractional charge and polarization listed in table 1 in @xcite .
it is this spin - charge connection which leads to the constraint of the vortices in the two layers @xmath173 used in this section and @xmath136 used in the next section . as shown in detail in @xcite , the functional form achieved from the composite boson theory
is the same at that achieved from the microscopic lll+hf approach .
but the coefficients should be taken from the lll+hf approach .
the above form of the phonon contributions to the ground state wavefunction , but with different dependence on the distance @xmath174 was already derived from composite fermion chern - simon theory in lopez and fradkin , phys . rev .
b 51 , 4347 ( 1995 ) .
the same transformation is needed to transform the modulus of the quasi - hole wavefunction in slqh at laughlin s series @xmath175 with @xmath176 odd @xmath177 . to the best of the authors knowledge ,
this very subtle and important point was not explicitly spelled out in any previous work .
it was shown in @xcite that although the `` quasi - hole '' and the `` meron '' wavefunction listed in eqn.[prime ] are completely different at any finite distance , they have finite comparable expectation energies at @xmath178 , so we expect the quasi - hole s expectation energy in eqn.[qh ] is comparable to both at @xmath9 . however , as explained in the main text , all these wavefunctions are not interesting anymore at @xmath9 .
l. brey , phys .
lett , 65 , 903 ( 1990 ) ; r. cote , l. brey and a. h. macdonald , phys . rev .
b 46 , 10239 ( 1992 ) ; x. m. chen and j. j. quinn , phys
lett , 67 , 895 ( 1991 ) , phys .
b 45 , 11054 ( 1992 ) , phys . rev .
b 47 , 3999 ( 1993 ) ; l. zheng and h. a. fertig , phys . rev .
b 52 , 12282 ( 1995 ) s. narasimhan and tin - lun ho , phys . rev .
b 52 , 12291 ( 1995 ) | bilayer quantum hall system ( blqh ) differ from its single layer counterparts ( slqh ) by its symmetry breaking ground state and associated neutral gapless mode in the pseudo - spin sector . due to the gapless mode , qualitatively good
groundstate and low energy excited state wavefunctions at any finite distance is still unknown .
we investigate this important open problem by the composite boson ( cb ) theory developed by one of the authors to study blqh systematically .
we derive the ground state , quasi - hole and a pair of quasihole wavefunctions from the cb theory and its dual action .
we find that the ground state wavefunction is the product of two parts , one in the charge sector which is the well known halperin s @xmath0 wavefunction and the other in the spin sector which is non - trivial at any finite @xmath1 due to the gapless mode .
so the total groundstate wavefunction differs from the well known @xmath0 wavefunction at any finite @xmath1 .
in addition to commonly known multiplicative factors , the quasi - hole and a pair of quasi - holes wavefunctions also contain non - trivial normalization factors multiplying the correct ground state wavefunction .
we expect that the quasi - hole and pair wave function not only has logarithmically divergent energy , well localized charge distribution , but also correct interlayer correlations .
all the distance dependencies in all the wavefunctions are encoded in the spin part of the ground state wavefunction .
the instability encoded in the spin part of the groundstate wavefunction leads to the pseudo - spin density wave formation proposed previously by one of the authors .
some subtleties related to the lowest landau level ( lll ) projection of the cb theory are also noted . |
anomalous enhancement of the sub - barrier nuclear fusion reaction d(d , p)t in a deuterated metallic matrix has been experimentally observed at energies of the incident beam lower than few kev @xcite .
electron screening is not sufficient to explain this enhancement , and other quantitative explanations are missing @xcite .
in fact , the calculated ( adiabatic limit ) electron screening potential energy @xmath0 is 28 ev , but experiments show an exponential enhancement of the cross section at very low energies that would correspond to a screening energy @xmath1 ev @xcite .
similar behaviors are found in gaseous targets .
these discrepancies between the calculated @xmath0 and the experimentally inferred @xmath2 have not yet been understood @xcite ; these puzzling results could have important consequences also for the study of nuclear fusion in astrophysical environments @xcite . in this paper
we discuss a possible explanation of this enhancement which is based on the quantum - uncertainty dispersive effect between energy and momentum that was proposed by galitskij and yakimets @xcite and recently discussed and applied by starostin et al .
experimental data for the d(d , p)t reaction in a deuterated tantalium target @xcite are reported for beam energies in the range 4 - 20 kev .
the target is cooled with nitrogen at a temperature of 10 @xmath3c , which corresponds to a thermal energy @xmath4 ev .
data from ref .
@xcite are plotted in fig .
[ fig : raiolasigmav ] . in panel ( a ) of fig .
[ fig : raiolasigmav ] the experimental values of @xmath5 are plotted versus the center of mass energy @xmath6 .
since the incoming particles are not relativistic and the target particles are practically at rest , these data are plotted using @xmath7 and @xmath8 , where @xmath9 is the reduced mass ) .
the unscreened cross section can be written as : @xmath10 with @xmath11 .
the astrophysical factor @xmath12 should vary slowly in this energy range and is linearly approximated as @xmath13 , where @xmath14 and @xmath15 are extrapolated from energies @xmath16 kev ( see panel ( b ) in fig .
[ fig : raiolasigmav ] ) ; the values reported in ref .
@xcite are @xmath17 kev b and @xmath18 b. the interacting nuclei `` feel '' a effective potential barrier lower by an amount equal the electron screening potential @xmath19 : the resulting screened cross section is : @xmath20 when @xmath21 a correction factor @xmath22 can be factorized : @xmath23 where @xmath24 the electron screening potential computed in the adiabatic limit @xmath0 constitutes a theoretical upper limit for @xmath19 : @xmath25 for d+d reactions , fig .
[ fig : raiolasigmav ] compares the `` bare curve '' of eq .
( [ eq : barecrosssec ] ) and the `` screened curve '' ( @xmath26 ev ) of eq .
( [ eq : screenedcrosssec ] ) with the experimental data .
the screened cross section with @xmath27 ev underestimates the experimental data by about an order of magnitude in the few kev energy range ; the data could be fitted only by using an unphysical electron potential much larger than the adiabatic upper limit : @xmath28 ev @xmath29 .
@xmath30 equation ( [ eq : screenedcrosssec ] ) implies @xmath31 and , once again , only the screened curve with @xmath28 ev fits the data . in summary , there is experimental evidence of anomalously high values of low - energy fusion cross sections that would require electron screening potentials @xmath19 one order of magnitude larger than their adiabatic limit , if explained in terms of screening . in deuterated metal targets ,
the effect depends strongly on the metal @xcite ) .
values of @xmath32 are required also for describing experiments with gas target , but violations of the requirement @xmath33 are less strong than in metal targets .
these anomalous values of @xmath19 are substantially unexplained ; the screened potential approach is probably trying to mimic important processes that have been disregarded .
one should attempt to find alternative mechanisms that could reproduce the enhancement of the cross section at low energy .
the parameters of the experiments ( temperature and beam energy ) are such that thermal corrections can be neglected .
in fact if we average @xmath34 over a maxwellian distribution of velocity , the temperature correction factor @xmath35 can be estimated @xcite : @xmath36 @xmath37 for d+d reactions ( @xmath38 kev ) even at room temperature ( @xmath39 kev ) and at the lowest energies of the beam ( @xmath40 kev ) the maxwellian thermal factor @xmath41 . as showed by galitskij and yakimets @xcite many - body collisions
broaden the relation between momentum and energy of the particles .
since momentum rather than energy determines the scattering amplitude , the reaction cross section must be averaged over a momentum distribution that may differ from the energy one .
the reaction rate for a mono - energetic beam is : @xmath42 where the center of mass energy @xmath6 is function of the beam - particle and target - particle momentum , and the momentum distribution function of the target particles @xmath43 depends on thermal distribution of the target - particle energies @xmath44 , which we take to be maxwellian , and on the probability that a target particle with energy @xmath45 has momentum @xmath46 ( dispersion relation ) , @xmath47 .
quantum effects are responsible for the dispersion relation between energy and momentum @xmath47 not being a @xmath48-function , but a broader distribution . according to galitskij and yakimets @xcite
the relation between energy and momentum is a lorentzian @xmath49 with @xmath50 , where @xmath51 and @xmath52 are the collisional frequency and cross section , while @xmath53 and @xmath54 are the colliding particles density and velocity . assuming that the collision cross section could be approximated by the bare coulomb cross section @xmath55
, the resulting @xmath56 is @xmath57 where the relations @xmath58 and @xmath59 have been used with @xmath60 the total density and @xmath61 the ( average ) atomic number . in the asymptotic regime @xmath62 ,
relevant for the particles that undergo fusion , the maxwellian contribution @xmath63 is negligible and the distribution of momenta becomes @xmath64 since only particles in the high - energy tail of the distribution @xmath62 contribute to the fusion rate , the quantum effect contribution @xmath65 is the only important contribution to the rate .
more in general , the fact that the relation between energy and momentum is not a @xmath48-function results in a distribution of momentum @xmath66 with a power - law asymptotic behavior in spite of the energy distribution @xmath67 being exponential .
this power - law tail becomes mostly important for reactions whose cross sections select high - momentum particles .
this quantum contribution can be calculated numerically and we are also developing useful analytical approximations : we shall report elsewhere the derivation of these approximations and their comparison with the exact numerical integration . in the following we give some preliminary results using a parameterization that is motivated by the form of eq .
( [ eq : qefftercorr ] ) and that can be used to qualitatively estimate the importance of such quantum broadening of the momentum distribution . if we use @xmath68 as an upper bound for @xmath69 in the integral in eq .
( [ eq : qefftercorr ] ) ( the precise value of the low bound is inconsequential ) and work in the relevant approximation that the target particles have energies lower than the beam particles , we obtain the following partial parameterization for the dominant contribution @xmath70 in fig . [
fig : enhancfactcompared ] we have plotted the ratio @xmath71 with @xmath72 given by eq .
( [ eq : qefflowenapprox ] ) for three different values of @xmath73 .
the function @xmath74 shows an evident enhancement at low energy starting from the region of 1 - 2 kev in qualitative agreement with experiments .
the threshold below which this enhancement becomes important depends on @xmath75 .
we are completing a more detailed calculation that does not require the introduction of the parameter @xmath73 and that , therefore , can better test the relevance of this quantum effect for the experimental results .
the theory of galitskij and yakimets predicts that quantum indeterminacy broadens the relation between energy and momentum .
we have performed a preliminary calculation to estimate the effect of this broadening on the momentum distribution of deuteron in metals and , therefore , on the cross section of the reaction d(d , p)t .
this calculation shows that this quantum effect should give an important enhancement of the cross section at low energies similar to the one observed in the bochum experiments .
a more quantitative comparison between theory and experiments requires the use of more sophisticated analytical or numerical analyses @xcite and the inclusion of the effects of screening both on the fusion cross section and on the ion - collision cross section .
this further work is being completed and will be published in the near future .
00 f. raiola _ et al .
_ , eur .
j. * a 13 * ( 2002 ) 377 .
f. raiola _ et al .
_ , phys .
* b 547 * ( 2002 ) 193 .
g. fiorentini , c. rolfs , f. l. villante and b. ricci , phys .
c * 67 * ( 2003 ) 014603 [ arxiv : astro - ph/0210537 ] .
m. coraddu , g. kaniadakis , a. lavagno , m. lissia , g. mezzorani and p. quarati , braz .
j. phys .
* 29 * ( 1999 ) 153 [ arxiv : nucl - th/9811081 ] .
m. coraddu , m. lissia , g. mezzorani and p. quarati , arxiv : nucl - th/0012002 . v. m. galitskii and v. v. yakimets , jept * 24 * ( 1967 ) 637 .
a. n. starostin , a. b. mironov , n. l. aleksandrov , j. n. fisch , and r. m. kulsrud , physica a * 305 * ( 2002 ) 287 .
a. n. starostin , v. i. savchenko , and n. j. fisch , phys .
a * 274 * ( 2000 ) 64 .
v. i. savchenko , arxiv : astro - ph/9904289 .
m. coraddu , g. mezzorani , yu .
v. petrushevich , p. quarati , and a. n. starostin , physica a , this issue . | the bochum experimental enhancement of the d+d fusion rate in a deuterated metal matrix at low incident energies is explained by the quantum broadening of the momentum - energy dispersion relation and consequent modification of the high - momentum tail of the distribution function from an exponential to a power - law . , , , , , . |
cosmic observations from type ia supernovae ( sne ia ) @xcite , large scale structure ( lss ) @xcite , baryon acoustic oscillations ( bao ) @xcite and cosmic microwave background ( cmb ) @xcite indicate that our universe is undergoing an accelerating expansion .
the reason for this acceleration , the so - called dark energy problem , remains a fascinating question today . the simplest model to explain this problem is the @xmath3cdm model , in which a time independent energy density is added to the universe .
however , the @xmath3cdm model suffers from both fine - tuning and coincidence problems @xcite . in general
, the ways to understand the cosmic acceleration can be separated into two branches .
one is to modify the matter by introducing some kind of dark energy .
the other one is to modify einsteins general relativity the modification of gravity . in modified gravity ,
one of the popular approaches is to promote the ricci scalar @xmath4 in the einstein - hibert action to a function , @xmath5 .
although there are several viable @xmath5 models , many of them are restricted to the regimes to be effectively identical to the @xmath3cdm by the observational constraints .
recently , linder @xcite has explored an @xmath5 theory named exponential gravity , which has also been discussed in refs .
@xcite . the exponential gravity has the feature that it allows the relaxation of fine - tuning and it has only one more parameter than the @xmath3cdm model .
in addition , the exponential gravity satisfies all conditions for the viability @xcite such as the local gravity constraint , stability of the late - time de sitter point , constraints from the violation of the equivalence principle , stability of cosmological perturbations , positivity of the effective gravitational coupling , and asymptotic behavior to the @xmath3cdm model in the high curvature regime . in this paper
, we will study the constraints given by latest observational data , reexamine the alleviation of the fine - tuning problem , and find the possibility of the derivation from @xmath3cdm .
we use units of @xmath6 and the gravitational constant is given by @xmath7 with the planck mass of @xmath8gev . the paper is organized as follows . in sec .
[ sec : exponential - gravity ] , we review equations of motion and the asymptotic behavior at the high redshift regime in the exponential gravity model . in sec . [
sec : observational - constraints ] , we discuss the observations and methods .
we show our results in sec .
[ sec : results ] .
finally , conclusions are given in sec .
[ sec : conclusion ] .
the action of @xmath5 gravity with matter is given by @xmath9+s_{m } , \label{action}\ ] ] where @xmath10 and @xmath5 is a function of the ricci scalar curvature @xmath4 .
in this paper , we focus on the exponential gravity model @xcite , given by @xmath11 where @xmath12 is related to the characteristic curvature modification scale .
since the product of @xmath1 and @xmath12 can be determined by the present matter density @xmath13 @xcite , we can choose @xmath1 and @xmath13 as the free parameters in the model .
we use the standard metric formalism . from the action ,
the modified friedmann equation of motion becomes @xcite @xmath14 where @xmath15 is the hubble parameter , a subscript r denotes the derivative with respect to r , a prime represents @xmath16 , and @xmath17 is the energy density of all perfect fluids of generic matter including ( non - relativistic ) matter , denoted by @xmath18 , and relativistic particles , denoted by @xmath19 . here , we only consider the matter density .
since the modification by the exponential gravity only happens at the low redshift , the contributions from relativistic particles are negligible . in a flat spacetime ,
the ricci scalar is given by @xmath20 following hu and sawicki s parameterization @xcite , we define @xmath21 where @xmath22 , @xmath23 is the effective dark energy density , and @xmath24 is the present matter density .
then , eqs . and
can be rewritten as two coupled differential equations , @xmath25 and @xmath26,\label{yr'}\ ] ] where @xmath4 and @xmath27 can be further replaced by @xmath28 and @xmath29 from equations in .
combining eqs . and , we obtain a second order differential equation of @xmath29 , @xmath30 where @xmath31 with @xmath32.\end{aligned}\ ] ] solving eq .
numerically , we can get the evolution of the hubble parameter in the low redshift regime ( @xmath33 ) .
the effective dark energy equation of state @xmath34 is given by @xmath35 in the high redshift regime ( @xmath36 ) , the exponential factor @xmath37 of @xmath5 in eq .
( [ fr ] ) becomes negligible ( @xmath38 ) .
the exponential gravity model behaves essentially like a cosmological constant model with the dark energy density parameter @xmath39 .
thus , the hubble parameter as a function of @xmath40 in this regime can be expressed as @xmath41 where @xmath42 is the density parameter of relativistic particles including photons and neutrinos , where @xmath43 is the present fractional photon energy density and @xmath44 is the effective number of neutrino species @xcite . ] .
the equation ( [ eqn2 - 11 ] ) will be used in the data fitting of cmb and the high redshift part of bao in section [ sec : observational - constraints ] .
to constrain the free parameters of @xmath1 and @xmath13 in the exponential gravity model , we use three kinds of the observational data including sne ia , bao and cmb .
the sne ia and cmb data lead to constraints at the low and high redshift regimes , respectively , while the bao data provide constraints at the both regimes .
the observations of sne ia , known as `` standard candles '' , give us the information about the luminosity distance @xmath45 as a function of the redshift @xmath40 .
the distance modulus @xmath46 is defined as @xmath47 where @xmath48 with @xmath49 is the present value of the hubble parameter .
the hubble - free luminosity distance for the flat universe is @xmath50 where @xmath51 .
the @xmath52 of the sne ia data is @xmath53^{2}}{\sigma_{i}^{2}},\end{aligned}\ ] ] where @xmath54 is the observed value of the distance modulus .
since the absolute magnitude of sne ia is unknown , we should minimize @xmath55 with respect to @xmath56 , which relates to the absolute magnitude , and expand it to be @xcite @xmath57 where @xmath58^{2}}{\sigma_{i}^{2 } } , \nonumber\\ % \begin{eqnarray } b&=&\sum_{i}\frac{\mu_{obs}(z_{i})-\mu_{th}(z_{i};\mu_{0}=0)}{\sigma_{i}^{2}},\quad c=\sum_{i}\frac{1}{\sigma_{i}^{2}}.\end{aligned}\ ] ] the minimum of @xmath55 with respect to @xmath56 is @xmath59 we adopt this @xmath60 for our later @xmath52 minimization .
we will use the data from the supernova cosmology project ( scp ) union2 compilation , which contains 557 supernovae @xcite , ranging from @xmath61 to @xmath62 .
the observation of bao measures the distance ratios of @xmath63 , where @xmath64 is the volume - averaged distance , @xmath65 is the comoving sound horizon and @xmath66 is the redshift at the drag epoch @xcite .
the volume - averaged distance @xmath67 is defined as @xcite @xmath68^{1/3},\end{aligned}\ ] ] where @xmath69 is the proper angular diameter distance : @xmath70 the comoving sound horizon @xmath71 is given by @xmath72 where @xmath73 and @xmath43 are the present values of baryon and photon density parameters , respectively .
we use @xmath74 and @xmath75 @xcite .
the fitting formula for @xmath66 is given by @xcite @xmath76,\end{aligned}\ ] ] where @xmath77 , \nonumber\\ b_{2}&=&0.238(\omega_{m}^{0}h^{2})^{0.223}.\end{aligned}\ ] ] the typical value of @xmath66 is about 1021 with @xmath78 and @xmath79 .
since @xmath66 is in the high redshift regime , we use eq .
( [ eqn2 - 11 ] ) to calculate @xmath80 . on the other hand
, @xmath67 is evaluated by the numerical result of eq .
( [ 2nd diff eq ] ) as it is in the low redshift regime .
the bao data from the two - degree field galaxy redshift survey ( 2dfgrs ) and the sloan digital sky survey data release 7 ( sdss dr7 ) @xcite measured the distance ratio @xmath81 at two redshifts @xmath82 and @xmath83 to be @xmath84 and @xmath85 with the inverse covariance matrix : @xmath86 the @xmath52 for the bao data is @xmath87 where @xmath88 .
the cmb is sensitive to the distance to the decoupling epoch @xmath89 @xcite .
it can give constraints on the model in the high redshift regime ( @xmath90 ) .
the cmb data are taken from wilkinson microwave anisotropy probe ( wmap ) observations @xcite . to use the wmap data , we compare three quantities : ( i ) the acoustic scale @xmath91 , @xmath92 ( ii ) the shift parameter @xmath4 @xcite , @xmath93 and ( iii ) the redshift of the decoupling epoch @xmath89 .
the fitting function of @xmath89 is given by @xcite @xmath94\left[1+g_{1}(\omega_{m}^{0}h^{2})^{g2}\right],\end{aligned}\ ] ] where @xmath95 the @xmath52 of the cmb data is @xmath96 where @xmath97 and @xmath98 is the inverse covariance matrix .
the data from seven - year wilkinson microwave anisotropy probe ( wmap7 ) observations @xcite lead to @xmath99 , @xmath100 and @xmath101 with the inverse covariance matrix : @xmath102 finally , the @xmath52 of all the observational data is @xmath103 in our fitting process , we did not use the markov chain monte carlo ( mcmc ) approach because the numerical calculation for each solution of @xmath5 theory is very time - consuming , and the necessary change to the code like cosmomc @xcite is very extensive with no obvious benefit in our study of the exponential gravity .
therefore , we take the simple @xmath52 method as our main fitting procedure .
the @xmath3cdm result obtained from sne ia , bao and cmb constraints with this @xmath52 method is @xmath104 , while that with the mcmc method is @xmath105 @xcite .
we note that the fitting in ref .
@xcite has also included the observational constraints from the radial bao and hubble parameter h(z ) . ]
.the best - fit values of the matter density parameter @xmath13 ( 68% cl ) and @xmath52 for the exponential gravity model with @xmath106 and the @xmath3cdm model .
note that the error for @xmath13 is obtained when @xmath1 is fixed.[tab : comparison ] [ cols="^,^,^,^",options="header " , ]
based on the methods described in sec .
[ sec : observational - constraints ] , we now examine the parameter space of the exponential gravity model . in fig .
[ fig : contours ] , we present likelihood contour plots at 68.3 , 95.4 and 99.7% confidence levels obtained from the sne ia , bao and cmb constraints .
the results show that the observational data give no upper bound on the model parameter @xmath1 , making it a free parameter .
hence , there is no fine - tuning problem .
however , a larger value of @xmath1 , which is closer to the @xmath3cdm model , is slightly preferred by the observational data as expected .
the lower bound on @xmath1 is @xmath107 ( 95% cl ) .
the present matter density parameter @xmath13 is constrained to @xmath2 ( 95% cl ) , which agrees with the current observations .
the best - fit value ( smallest @xmath52 ) in the parameter space between @xmath108 and 4 . for @xmath109
, it is almost the @xmath3cdm model . for @xmath110
, it is ruled out by the local gravity constraints and the stability of the de - sitter phase .
] is @xmath111 with @xmath112 and @xmath78 .
the comparison of the best - fit @xmath13 and @xmath52 for the model with @xmath106 and @xmath3cdm is shown in table [ tab : comparison ] . in fig .
[ fig : equation of state ] , we illustrate the evolution of the effective dark energy equation of state @xmath34 for @xmath106 with their best - fit @xmath13 , which is given in table [ tab : comparison ] .
we can see that , for every value of @xmath1 , the effective dark energy equation of state @xmath34 starts at the phase of a cosmological constant @xmath113 and evolves from the phantom phase ( @xmath114 ) to the non - phantom phase ( @xmath115 ) . and , for larger value of @xmath1 , the deviation from cosmological constant phase ( @xmath113 ) become smaller .
for @xmath116 , there is still another small oscillation after the main phantom phase crossing .
negative @xmath40 means the future evolution .
it is clear that the exponential gravity model has the feature of crossing the phantom phase in the past as well as the future @xcite . in fig .
[ fig : density parameter ] , we depict the effective dark energy density @xmath117 and non - relativistic matter density @xmath118 vs. the redshift @xmath40 .
corresponding to @xmath106 with their best - fit @xmath13 given in table [ tab : comparison ] .
[ fig : equation of state ] ] and non - relativistic matter density parameter @xmath118 as functions of @xmath40 , where the solid lines indicate the exponential gravity model with @xmath119 and the best - fit @xmath120 and the dashed lines represent the @xmath3cdm model with @xmath78 . for a higher value of @xmath1 ,
the evolution becomes closer to that in @xmath3cdm .
[ fig : density parameter ] ]
we have studied the exponential gravity model . in the low redshift regime , we follow hu and sawicki s parameterization to form the differential equation for the exponential gravity and solve it numerically . in the high redshift regime , we take advantage of the asymptotic behavior of the exponential gravity toward an effective cosmological constant . the analytical form of the hubble parameter @xmath121 as a function of the redshift @xmath40 can be expressed in the high redshift limit .
we have constrained the parameter space of the model by the sne ia , bao and cmb data .
we have found that there is a lower bound on the model parameter @xmath1 at 1.27 but no upper limit , and @xmath13 is constrained to the concordance value .
this means that the exponential gravity model shows no need of fine - tuning . nevertheless , the @xmath3cdm model is still included by the observational constraints since @xmath122 corresponds to the model .
current observational data still lack the ability to distinguish between the @xmath3cdm and exponential gravity models . finally , we remark that as seen from fig . [
fig : density parameter ] , the noticeable difference between the exponential gravity and @xmath3cdm models lies in the regime @xmath123 , and is maximized at @xmath124 if we compare their expected distance modulus .
an improvement on the bao observation may give a stronger constraint on this redshift regime or higher .
the ongoing and future dark energy survey projects which will observe bao include wigglez @xcite , boss ( baryon oscillation spectroscopic survay ) @xcite , hetdex ( hobby - eberly dark energy experiment ) @xcite , euclid @xcite , jdem ( joint dark energy mission)/omega with wide field infrared survey telescope ( wfirst ) @xcite , bigboss ( big baryon oscillation spec - troscopic survay ) @xcite , ska ( square kilometer array ) @xcite , lsst ( large synoptic survey telescope ) @xcite and des ( dark energy survey ) @xcite .
in addition , it is known that the measurement on the growth rate of @xmath125 has the potential to distinguish the models with the same expansion history but different physics . in the exponential gravity case ,
the growth index is @xmath126 for @xmath116 .
it is clear that if those surveys such as wigglez , euclid , bigboss and jdem / omega can measure the growth rate with a high accuracy , they will be able to discriminate the exponential gravity from the @xmath3cdm model .
we thank dr .
k. bamba for many helpful discussions and suggestions .
the work is supported in part by the national science council of r.o.c . under : grant # : nsc-98 - 2112-m-007 - 008-my3 and national tsing hua university under the boost program # : 97n2309f1 .
m. tegmark _
et al . _ , [ sdss collaboration ] , phys . rev .
d * 69 * , 103501 ( 2004 ) .
u. seljak _
[ sdss collaboration ] , phys .
d * 71 * , 103515 ( 2005 ) .
d. j. eisenstein _ et al .
_ [ sdss collaboration ] , astrophys .
j. * 633 * , 560 ( 2005 ) .
j. a. tyson , d. m. wittman , j. f. hennawi , and d. n. spergel , nucl .
. suppl . * 124 * , 21 ( 2002 ) , arxiv : astro - ph/0209632 ; j. a. tyson [ lsst collaboration ] , proc .
. eng . * 4836 * , 10 ( 2002 ) , arxiv : astro - ph/0302102 ; aip conf . proc . *
870 * , 44 ( 2006 ) , arxiv : astro - ph/0609516 ; see also http://www.lsst.org/. | we study the observational constraints on the exponential gravity model of @xmath0 .
we use the latest observational data including supernova cosmology project ( scp ) union2 compilation , two - degree field galaxy redshift survey ( 2dfgrs ) , sloan digital sky survey data release 7 ( sdss dr7 ) and seven - year wilkinson microwave anisotropy probe ( wmap7 ) in our analysis . from these observations ,
we obtain a lower bound on the model parameter @xmath1 at 1.27 ( 95% cl ) but no appreciable upper bound .
the constraint on the present matter density parameter is @xmath2 ( 95% cl ) .
we also find out the best - fit value of model parameters on several cases . |
blazars are observed to emit highly variable non - thermal radiation spanning the entire electromagnetic spectrum . the spectral energy distribution ( sed ) of blazars is assumed to be dominated by emission from a relativistic jet pointing close to our line of sight .
the jets in blazars consist of ultra - relativistic particles embedded in a magnetic field , with the entire plasma flowing outward from the central region with relativistic speed .
the characteristic sed of blazars shows two broad non - thermal well defined continuum peaks .
the first hump peaks somewhere between the infrared ( ir ) and x ray bands , whereas the second hump exhibits a maximum at @xmath0ray energies .
the origin of low energy peak is attributed to the synchrotron emission of relativistic electrons in the magnetic field of the jet . in leptonic models
@xcite , the @xmath0ray emission in high energy ( he : e@xmath9100 mev ) and very high energy ( vhe : e@xmath9100 gev ) regimes is attributed to the inverse compton ( ic ) scattering of low energy photons by the same electron population producing the synchrotron radiation ( ssc : synchrotron self compton ) or with the possible contribution from external photons ( ec : external compton ) . in hadronic models
@xcite , both electrons and protons are accelerated to ultra - relativistic energies , with protons exceeding the threshold for photo - pion production on the soft photon field in the jet . in these models ,
the he and vhe @xmath0ray emissions are dominated by proton synchrotron emission , neutral pion decay photons , synchrotron and compton emission from secondary decay products of charged pions .
blazars are broadly classified in two groups namely bl lacertae objects ( bl lacs ) and flat spectrum radio quasars ( fsrqs ) @xcite .
bl lacs are characterized by the weak or absence of thermal features like broad emission lines in their optical spectra .
on the other hand , fsrqs exhibit luminous broad emission lines in their optical spectra .
the peak frequency of the synchrotron component in the sed of blazars is usually used to subdivide them in 3-classes : low - synchrotron peaked ( lsp : @xmath10 hz ) , intermediate - synchrotron peaked ( isp : @xmath11 hz @xmath12
@xmath13 @xmath12 @xmath14 hz ) and high - synchrotron peaked ( hsp : @xmath15 hz ) blazars @xcite .
bl lacs are assumed to be good vhe candidates for ground based tev telescopes because their ic peak is in the tev regime . to date , bl lacs observed at vhe energies are predominantly high frequency peaked objects , which have been predicted as possible tev candidate blazars @xcite .
blazars often show violent flux variability from radio to vhe @xmath0rays at different time scales which may or may not be correlated . therefore , simultaneous multi - wavelength ( mwl ) observations are important to understand the underlying physics of blazars .
furthermore , vhe observations of blazars at cosmological distances would help in constraining the intensity and spectrum of extragalactic background light ( ebl ) , which is a very important physical quantity to understand the structure and star formation history @xcite .
thus , besides their importance for study of emission mechanisms and relativistic jet dynamics , growing interest for blazar study is also motivated by the use of vhe spectra as a probe for ebl @xcite .
1es 1218 + 304 ( ra:@xmath16 , dec : @xmath17 ) is an hsp bl lac object located at a redshift z = 0.182 @xcite .
the source was discovered as a candidate bl lac object on the basis of its x - ray emission and was identified as x - ray source 2a 1219 + 30.5 @xcite .
being an x ray bright source , it was predicted to be a tev candidate blazar from the position of the synchrotron peak in its sed @xcite .
the @xmath18 ( high energy gamma ray astronomy ) collaboration observed 1es 1218 + 304 during 1996 - 2002 for 3.9 h and reported an upper limit of 12@xmath2 of crab nebula flux above 750 gev @xcite .
the vhe emission from 1es 1218 + 304 was first detected by the @xmath7 ( major atmospheric gamma ray imaging cherenkov ) telescope in 2005 above 120 gev @xcite .
the differential energy spectrum ( d@xmath19/de = f@xmath20 e@xmath21 ) , with power law index ( @xmath22 ) of [email protected] , was also reported by the @xmath7 group . in may 2006 , 1es 1218 + 304
was the target of @xmath24 ( high energy stereoscopic system ) observation campaign and these observations did not yield any statistically significant signal from the source @xcite . the corresponding 99.9@xmath2 confidence level ( cl ) limit on integral flux above 1 tev was reported to be 17@xmath2 of the @xmath24 crab nebula flux .
the @xmath25 ( solar tower atmospheric cherenkov effect experiment ) detector monitored 1es 1218 + 304 during 2006 - 2007 above 160 gev but did not detect any significant @xmath0ray emission from the source @xcite . in 2007 ,
@xmath8 ( very energetic radiation imaging telescope array system ) telescope observed vhe emission above 160 gev from the source at a persistent level of 6@xmath2 of the crab nebula flux @xcite .
the differential energy spectrum of the source was found to be compatible with a power law of index [email protected] .
using the lower limit ebl from galaxy counts @xcite , the intrinsic power law index was found to be [email protected] . during the december 2008 - may 2009 monitoring campaign of 1es 1218 + 304
, @xmath8 telescope revealed a prominent flaring activity from the source at 20@xmath2 of the crab nebula flux above 200 gev @xcite .
the time averaged differential energy spectrum , in the energy range 0.2 - 1.8 tev , was found to be a power law with index of [email protected] .
this flaring activity was characterized by a variability time scale of days .
the corrected spectrum after accounting for absorption due to ebl suggests a very hard intrinsic source spectrum with index @xmath26 [email protected] @xcite in vhe regime based on lower limit ebl model @xcite . _
fermi_-lat ( large area telescope ) also detected significant emission from 1es 1218 + 304 during its first two years of operation @xcite .
the _ fermi_-lat spectrum of the source is described by a power law with index [email protected] , making it one of the hardest spectrum sources in mev - gev range .
due to relatively large redshift of the source , we expect a significant attenuation of tev photons above 1 tev due to ebl absorption .
this makes the source a good candidate for probing ebl using tev observations .
it is because of these characteristics of the source we were motivated to observe the source with @xmath1 ( tev atmospheric cherenkov telescope with imaging camera ) in the tev energy region .
the paper is organized as follows .
section 2 gives a brief description of the @xmath1 telescope and in section 3 we present details of observation and data analysis . in section 4
we present the mwl data analysis of the source from radio to he . in section 5
we discuss the results and the conclusions are presented in section 6 .
the @xmath1 @xmath0ray telescope is located at mount abu ( 1300 m asl , @xmath27n , @xmath28e ) , india @xcite . the telescope is equipped with a f/1-type tessellated light collector of approximately 9.5 @xmath29 area consisting of 34 front - face aluminium coated , spherical glass mirror facets of 60 cm diameter .
the point - spread function has a full width at half maxima ( fwhm ) of 0.185@xmath30 ( @xmath3112.5 mm ) and d@xmath32 @xmath330.34@xmath30 ( @xmath3122.8 mm ) . here
, d@xmath32 is defined as the diameter of a circle , concentric with the centroid of the image , within which 90@xmath2 of reflected rays lie .
the telescope deploys a 349-pixel imaging camera , with a uniform pixel resolution of @xmath34 and a @xmath35 field - of - view , to record atmospheric cherenkov events produced by an incoming cosmic - ray particle or a @xmath0-ray photon .
data used in this work have been collected with inner 225 pixels where the innermost 121 pixels ( 11 @xmath36 11 matrix ) are used for generating the event trigger .
the trigger scheme is based on a nearest neighbour non - collinear triplets trigger criterion . apart from generating the prompt trigger with a coincidence gate width of @xmath3718ns
, the trigger generator has a provision for producing a chance coincidence output based on @xmath38@xmath39 combinations from various groups of closely spaced 12 channels .
the data acquisition and control system of the telescope is designed around a network of pcs running the qnx ( version 4.25 ) real - time operating system .
the triggered events are digitized by camac based 12-bit charge to digital converters ( cdc ) , which have a full scale range of 600 pc .
the relative gain of the photomultiplier tubes is monitored regularly once in 15 minutes by flashing a blue led , placed at a distance of about 1.5 m from the camera .
other details regarding hardware and software features of the data acquisition and control system of the telescope are discussed in @xcite .
major upgrade , involving replacement of signal and high voltage cables and installation of new compound parabolic concentrators ( cpc ) was taken up in november - december 2011 for improving the sensitivity of the telescope . while detailed simulation and experimental results on the crab nebula , after upgrade will be presented elsewhere , we present here only a brief summary of the upgrade work . new cpcs with square entry and circular exit aperture were installed on the @xmath1 imaging camera in order to increase its photon collection efficiency .
the collection efficiency of the new cpcs was measured to be about 80@xmath2 in the wavelength range 400550 nm . apart from removing the dead space in between the photomultipliers substantially ,
the new cpcs has also helped us to improve the gamma / hadron segregation capability of the telescope .
in addition , the trigger criteria was also modified by including more nearest neighbor collinear triplet combinations .
a dedicated ccd camera was also installed for conducting detailed point run calibrations and data collected were successfully used for determining the position of the source in the image plane with an accuracy better than @xmath23 3 arc min .
the point run calibration data ( i.e. position of the star image in the camera ) were also incorporated in the analysis software so that appropriate corrections can be applied in an off line manner while calculating source position dependent image parameters .
the analysis procedure was improved by using the asymmetry parameter so that additional hadronic background can be further removed by identifying the `` head / tail '' feature of cherenkov images .
it is worth mentioning here that the @xmath0ray images have their head closer to the assumed source position in the imaging camera and thus can be selected preferentially by imposing asymmetry @xmath90 cut .
the upgrade of the telescope has led to an increase in the prompt coincidence rate from 2.33 hz to 4.70 hz close to the zenith .
furthermore , as a result of the upgrade the @xmath0ray rate from the crab nebula has also increased from 9.13 h@xmath5 to 15.30 h@xmath5 and this translates to the reduction in the telescope threshold energy from 1.2 tev to 0.87 tev .
the blazar 1es 1218 + 304 was monitored with the @xmath1 from march 1 , 2013 to april 15 , 2013 ( mjd 5635256397 ) at zenith angles between @xmath40 and @xmath41 .
all the data were collected in tracking mode , during which the telescope continuously monitored the source .
this mode of observation maximizes the source observation time and increases the possibility of detecting flaring activity from the source .
about 54 h of data were collected during 18 nights of observations as per the details summarized in table 1 .
.summary of @xmath1 observations of 1es 1218 + 304 [ cols="<,^,^,^,^",options="header " , ] the @xmath42 ( owens valley radio observatory ) is a fast cadence 15 ghz radio telescope with diameter of 40 m @xcite .
the telescope is a f/0.4 type parabolic reflector on an alt - azm mounting system and is equipped with dual beamed off axis optics and a cooled receiver installed at prime focus .
the source 1es 1218 + 304 was observed at 15 ghz using @xmath42 telescope for six days during @xmath1 observations as part of _ fermi _ mwl blazar monitoring program and data of these observations are used in the present work .
the mwl light curve of 1es 1218 + 304 observed by various instruments during march 1 , 2013 to april 15 , 2013 ( mjd 5635256397 ) is shown in the figure 2 . since
no significant @xmath0ray emission has been detected from the source with the @xmath1 , we have shown the 99@xmath2 cl upper limit on integral flux above 1.1 tev for @xmath1 observations in figure 2(a ) .
the five - day binned light curve of the source observed with _
fermi_-lat during the same period is presented in figure 2(b ) .
all the points reported in the figure 2(b ) correspond to ts @xmath9 4 . from the figure ,
it is evident that there is no statistically significant variation in the @xmath0ray activity in the energy range 100 mev-100 gev during @xmath1 observations and the average flux level during this period is found to be ( [email protected])@xmath43 photons @xmath4 s@xmath5 .
it is important to mention here that the source has been categorized as highly variable on monthly time scale in the second _ fermi _ catalog @xcite and thus some enhanced activity may be quite consistent with its past behavior .
nearly simultaneous x
ray light curve observed by _
swift_-xrt for six days of monitoring of the source is depicted in figure 2(c ) . from the figure
, we observe that the soft x ray emission from the source is consistent with the average flux level ( [email protected])@xmath44 erg @xmath4 s@xmath5 , except for some enhanced activity on march 15 , 2013 ( mjd 56366 ) . during the enhanced activity on march 15 in x - rays ,
the he activity is also observed to be slightly higher with respect to the average flux .
we have also looked into the archival x
ray data in the energy bands 2 - 20 kev and 15 - 50 kev from _ _ maxi _ _ and _ swift_-bat instruments respectively and these data do not show any detection above 3@xmath45 .
this indicates that during the period of @xmath1 observations , the source was not active in hard x - ray regime .
the simultaneous uv ( w1,m2,w2 filters ) and optical(v , b , u filters ) light curves of the source for six days of observations are presented in figure 2(d ) and 2(e ) respectively .
the uvot flux points included in the light curve have not been de - reddened .
no unusual activity is observed in the source with _
swift_-uvot instrument during the period of @xmath1 observations .
the radio observations at 15 ghz available for six days are shown in figure 2(f ) .
no significant variations are observed in radio emission from the source during this period . the spectral energy distribution of 1es1218 + 304 during march 1 , 2013 to april 15 , 2013 using broad band data discussed above is shown in figure 3 .
for vhe @xmath0rays we plot the data from the observations with @xmath7 and @xmath8 telescopes along with the 99@xmath2 confidence level upper limit on integral flux obtained from @xmath1 in the present work .
the flux points reported by @xmath7 group are based on the observations carried out during the period january 9 - 15 , 2005 @xcite . during six days of observations ,
no flux variability on timescales of days was found and the time averaged spectrum was described by a power law with @xmath46 .
the @xmath8 flux points taken from @xcite correspond to observations during january - march 2007 for a total observation time of 17.4 h. the time averaged differential spectrum was described by a power law with photon index [email protected] and the integral flux above 200 gev was 6@xmath2 of the crab nebula flux .
it is evident from the figure that the measured flux points are consistent with each other in the overlaping energy regime of the two telescopes .
the vhe flux points measured with @xmath7 and @xmath8 telescopes as well as upper limit from @xmath1 have been corrected for ebl absorption using the mean level density model proposed by franceschini et al .
( 2008 ) @xcite .
the lat data points have been obtained by dividing the energy range 0.1100 gev into four energy bands : 0.11 gev , 110 gev , 1020 gev and 20100 gev .
the time averaged gev spectrum measured by _
fermi_-lat during this period is described by a power law with normalization factor f@xmath20=([email protected])@xmath36 10@xmath47 @xmath4 s@xmath5 gev@xmath5 and photon index @xmath48 .
the he photon index @[email protected] obtained in the present study is consistent with the value reported from quiescent state monitoring of 1es 1218 + 304 by _
fermi_-lat @xcite .
the photon index observed by _
fermi_-lat during the first two years of observation of 1es 1218 + 304 is [email protected] @xcite .
the _ swift_-xrt flux points have been obtained from simultaneous fitting of the spectra of six observations in three energy bands:0.3 - 1 kev , 1 - 2.5 kev and 2.56 kev using cflux . beyond 6 kev , _
swift_-xrt observations are not statistically significant to perform spectral analysis .
the x - ray flux points measured with xrt as shown in figure 3 have been corrected for galactic absorption using a neutral hydrogen column density of 1.99@xmath49 @xmath50 obtained from ned .
the time averaged soft x ray spectrum measured with xrt during @xmath1 observations is described by a power law with photon index [email protected] .
the x ray emission level of ( [email protected])@xmath36 10@xmath51 erg @xmath4 s@xmath5 in the energy range 2 - 10 kev , observed in our present study is below the flux level ( [email protected])@xmath36 10@xmath51 erg @xmath4 s@xmath5 obtained from _ xmm - newton _ measurements during 2001 observations @xcite . the simultaneous optical and uv emission in the wavelength range 180 - 600 nm as measured by _ swift_/uvot instrument in all six filters ( v , b , u , w1,m2,w2 ) is shown in the figure 3 .
the flux densities in all filters are estimated from the dereddened magnitudes with galactic absorption @xmath52 and @xmath53 @xcite , using the methodology proposed in @xcite .
the uvot flux points have been obtained by multiplying the mean of flux densities with the bandpass ( fwhm ) of the corresponding filter @xcite .
error in the mean density is obtained through the standard error propagation method .
the radio flux point at 15 ghz obtained from ovro telescope data archive corresponds to the mean emission level from the source during @xmath1 observations .
the broadband data points presented in figure 3 indicate that the sed of the source can be described by two humps : first one peaking at x ray energies and second at gev energies .
this implies that the mwl emission from 1es 1218 + 304 can be possibly compared with the predictions of ssc model for blazar emission , but detailed sed modeling of the source is beyond the scope of this work .
tang et al .
( 2010 ) have reproduced the sed of 1es 1218 + 304 with the inhomogeneous jet model and the homogeneous ssc model @xcite .
they emphasize that the leptonic model is very successful in explaining multi - band emissions from the source and also point out that the vhe @xmath0ray data from the @xmath7 and @xmath8 telescopes can be fitted with the strict lower - limit ebl model . using the _ swift _ , @xmath7 and @xmath8 telescopes data the sed of the source has been modeled in @xcite by employing a time - dependent ssc code for obtaining the physical parameters of the emission region .
the short - time variability of the source has also been studied by weidinger & spanier ( 2010 ) @xcite for reproducing the light curve observed by @xmath8 telescope .
they suggest that the light curve can be reproduced by assuming a changing level of electron injection compared to the constant state .
our @xmath0ray observations of 1es 1218 + 304 ( z=0.182 ) with the @xmath1 from march 1 , 2013 to april 15 , 2013 ( mjd 56352 - 56397 ) for a total observation time of @xmath3339.62 h do not show any evidence of tev @xmath0ray signal from the source .
the mwl data in the x - ray and he bands , as measured by _ swift_-xrt ( 0.3 - 10 kev ) and _ fermi_/lat ( 0.1100 gev ) respectively , do not reveal any unusual activity from the source .
the optical and uv emissions observed with _
swift_-uvot instrument and radio observations at 15 ghz with 40 m @xmath42 telescope during @xmath1 observations also do not indicate any flaring activity from the source .
it is important to point out here that , because of the variable nature of blazars in general , the vhe emission from 1es1218 + 304 may increase significantly during future flaring episodes and may even easily exceed the limit reported in the present work .
hence the upper limit presented here only constrains the flux during our observation period .
our blazar observation program will continue to monitor this source .
we thank the anonymous reviewer for his / her suggestions which improved the quality of the paper .
the authors would like to convey their gratitude to all the concerned colleagues of the astrophysical sciences division for their contributions towards the instrumentation , observation and analysis aspects of the @xmath1 telescope .
we acknowledge the useful discussions held with n.g .
bhatt on various aspects of data analysis .
we acknowledge the use of public data obtained through _ fermi _ science support center ( fssc ) provided by nasa .
this work made use of data supplied by the uk _ swift _ science data centre at the university of leicester .
this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
l. maraschi et al .
1992 , apj , 397 , l5 c. dermer and r. schleickeiser , 1993 , apj , 416 , 458 m. sikora et al . , 1994 , apj , 421 , 153 m. blazejowski et al . , 2000 , apj , 545 , 107 k. mannheim and p. l. biermann , 1992 , a&a , 253 , l21 f. aharonian et al . , 2000 , new astron . ,
5 , 377 a. mucke and r. j. protheroe , 2001 , astropart .
15 , 121 a. mucke et al . , 2003 ,
phys . , 18 , 593 c. urry and p. padovani , 1995 , pasp , 107 , 803 a. a. abdo et al
. , 2010 , apj , 716 , 30 l. costamante and g. ghisellini , 2002 , a&a , 384 , 56 f. w. stecker et al . , 1992 , apj , 390 , l49 f. aharonian et al . , 2006 ,
nature , 440 , 1018 n. bade et al .
, 1998 , a&a , 334 , 459 a. s. wilson et al . , 1979 , mnras , 187 , 109 j. e. ledden et al . , 1981 , apj , 243 , 47 m. tluczykont et al . , 2003 , @xmath54 icrc ,
2547 j. albert et al . , 2006 ,
apj , 642 , l119 f. aharonian et al . , 2008 ,
a&a , 478 , 387 c. mueller et al . , 2011 ,
, 34 , 674 a. acciari et al . , 2009 ,
apj , 695 , 1370 p. madau and l. pozzetti , 2000 , mnras , 312 , l9 g. g. fazio et al . , 2004 ,
apjs , 154 , 39 a. acciari et al . , 2010 ,
apjl , 709 , l163 f. krennrich et al . , 2008 ,
apj , 689 , l23 l. levenson and e. wright , 2008 , apj , 683 , 585 p. l. nolan et al . , 2012 ,
apjs , 199 , 31 r. koul et al .
, 2007 , nim a , 578 , 548 k. k.yadav et al
. , 2004 , nim a , 527 , 411 a. m. hillas et al . , 1985 , proc .
19th icrc , 3 , 445 a. m. hillas et al . , 1998 ,
apj , 503 , 744 g. mohanty et al .
, 1998 , apj , 9 , 15 t. p. li and y. q. ma , 1983 , apj , 272 , 317 o. helene , 1983 , nim a , 212 , 319 f. a. aharonian et al . , 2000 , apj , 539 , 317 w. b. atwood et al . , 2009 ,
apj , 697 , 1071 j. r. mattox et al .
, 1996 , apj , 461 , 396 d. n. burrows et al . , 2005 , space sci .
120 , 165 p. w. a. roming et al .
, 2005 , space sci .
120 , 95 a. a. breeveld et al . , 2011 ,
aipc , 1358 , 373 t. s. poole et al . ,
2008 , mnras , 383 , 627 j. l. richards et al . , 2011 ,
apjs , 194 , 29 a. franceschini et al .
, 2008 , a&a , 487 , 837 a. a. abdo et al . , 2009 , apj , 700 , 597 a. a. abdo et al . , 2009 , apj , 707 , 1310 a. j. blustin , m. j. page and g. b. raymont , 2004 , a&a , 417 , 61 e. f. schlafly and d. p. finkbeiner , 2011 , apj , 737 , 103 y. tang , z. dai and l. zhang , 2010 , ra&a 10
, 415 m. ruger et al .
, 2010 , mnras , 401,973 m. weidinger and f. spanier , 2010 , a&a , 515 , a18 rays by @xmath1 , he @xmath0rays by _
fermi_-lat , x rays by _ swift_-xrt , uv / optical by _ swift_-uvot instruments and radio by @xmath42 telescope .
99@xmath2 confidence upper limit in vhe @xmath0rays by @xmath1 is indicated as downward arrow .
each point in the 5 day binned light curve of _ fermi_-lat in he @xmath0rays corresponds to ts @xmath58 .
x ray , uv / optical and radio light curves represent daily flux values for 6 days of observations available during the monitoring of the source with @xmath1 .
the horizontal dotted lines ( shown in ( b ) , ( c ) & ( f ) ) represent the average emission level in the respective energy bands.,scaledwidth=100.0% ] | in this paper , we present results of tev @xmath0ray observations of the high synchrotron peaked bl lac object 1es 1218 + 304 ( z=0.182 ) with the @xmath1 ( tev atmospheric cherenkov telescope with imaging camera ) .
the observations are primarily motivated by the unusually hard gev - tev spectrum of the source despite its relatively large redshift .
the source is observed in the tev energy range with the @xmath1 from march 1 , 2013 to april 15 , 2013 ( mjd 5635256397 ) for a total observation time of 39.62 h and no evidence of tev @xmath0ray activity is found from the source . the corresponding 99@xmath2 confidence level upper limit on the integral flux above a threshold energy of 1.1 tev is estimated to be 3.41 @xmath3 photons @xmath4 s@xmath5 ( i.e @xmath6 crab nebula flux ) assuming a power law differential energy spectrum with photon index 3.0 , as previously observed by the @xmath7 and @xmath8 telescopes . for the study of multi - wavelength emission from the source , we use nearly simultaneous optical , uv and and x ray data collected by the uvot and xrt instruments on board the _ swift _ satellite and high energy @xmath0ray data collected by the large area telescope on board the _ fermi _ satellite .
we also use radio data at 15 ghz from ovro 40 m telescope in the same period .
no significant increase of activity is detected from radio to tev @xmath0rays from 1es1218 + 304 during the period from march 1 , 2013 to april 15 , 2013 .
blazars : 1es 1218 + 304 , cherenkov imaging telescope : very high energy gamma - rays , multi - wavelength observations . |
an important and still open issue , at present under considerable discussion , of the high-@xmath3 cuprate superconductors ( htsc ) is the role that play in their superconducting phase transition the chemical and electronic disorder and inhomogeneities,@xcite and the loss of long - range phase coherence due to vortex fluctuations.@xcite the ongoing debate , whose elucidation may be an appreciable step toward the understanding of the pairing mechanism in htsc , is particularly well illustrated by the _ precursor _ diamagnetism observed above the meissner transition . in optimally
doped htsc , this effect was first studied quantitatively by lee , klemm and johnston @xcite and attributed , as in the case of the conventional ( bcs ) low-@xmath3 superconductors,@xcite to the presence of evanescent cooper pairs created by the unavoidable thermal agitation .
this last conclusion was supported by the fact that the measured diamagnetism could be accounted for in terms of the ginzburg - landau approach for multilayered superconductors with gaussian fluctuations of the superconducting order parameter ( ggl approach ) . since then
, the applicability of this _ conventional _ phenomenological ggl scenario , which was more recently extended empirically to high reduced temperatures by the introduction of a _ total energy _
cutoff,@xcite has been confirmed by different groups in many htsc,@xcite very in particular in underdoped compounds up to relatively high reduced field amplitudes , of the order of 0.2.@xcite further support was earlier provided by the consistency with other intrinsic rounding effects observed above @xmath3 , especially with the paraconductivity,@xcite and also with the precursor diamagnetism in dirty low-@xmath3 superconductors.@xcite in the last few years , however , the results summarized above are being strongly questioned , the claimed disagreements being with both the experimental and the theoretical aspects.@xcite among the first ones , is the anomalous precursor diamagnetism observed in various htsc with different dopings , including the optimal one.@xcite these anomalies are being proposed as a further confirmation of _ unconventional _ ( non - ggl ) scenarios , including those based on the presence of intrinsic @xmath3-inhomogeneities,@xcite on the breaking of the long - range phase coherence by vortex fluctuations,@xcite or on a 3d bose - einstein condensation.@xcite however , it has been shown recently that at least some of these anomalies may be easily explained in terms of @xmath3 inhomogeneities associated with the presence of chemical inhomogeneities in the samples.@xcite in fact , intrinsic - like chemical inhomogeneities and disorder , and then the corresponding @xmath3 inhomogeneities , may be always present in some cuprates ( in particular , in underdoped compounds ) , due to their non - stoichiometric nature.@xcite nevertheless , even when the measured precursor diamagnetism agrees with the ggl predictions , some authors contest the adequacy of these analysis invoking the apparently anomalous magnetization behavior below @xmath3 also in optimally doped cuprates.@xcite the challenge is , therefore , first to disentangle the diamagnetism due to superconducting fluctuations from the magnetization roundings associated with , intrinsic or not , chemical disorder and inhomogeneities , and then to attempt to analyze these data , consistently above and below @xmath3 , in terms of vortex and cooper pairs fluctuations .
this is the purpose of this paper , where we will first summarize the magnetization measurements in a tl@xmath0ba@xmath0ca@xmath0cu@xmath1o@xmath2 ( tl-2223 ) single crystal with very high chemical and structural quality .
this compound is stoichiometric , which minimizes the randomness of dopant atom distribution and then the corresponding disorder and inhomogeneities.@xcite these data will be then used to check if the existing gl approaches with fluctuations of vortices and of cooper pairs @xcite explain , at quantitatively and consistently above and below @xmath3 , the diamagnetism measured in homogeneous cuprate superconductors .
the tl-2223 sample used in this work , a @xmath4 mm@xmath5 single crystal , was chosen by its sharp low - field meissner transition from several high - quality single crystals of this nominal composition prepared by a. maignan and coworkers in caen .
details of the growth procedure of these crystals and of their structural characterization may be found in ref . .
the temperature and magnetic field dependence of the magnetization , @xmath6 , around @xmath7 were measured with a superconducting - quantum - interference - device ( squid ) magnetometer ( quantum design ) .
a crucial check of the quality of our crystal is provided by the temperature dependence of its low - field ( 1 mt ) magnetic susceptibility measured under field - cooled ( fc ) and zero field - cooled ( zfc ) conditions ( inset of fig . 1 ) .
under these very low fields , the superconducting fluctuation effects are expected to be negligible @xcite and , therefore , the observed susceptibility roundings above @xmath3 will be mainly due to inhomogeneities and disorder . as it may be appreciated , the meissner transition is very sharp , the transition temperature width , defined as usually,@xcite being around @xmath8 k , and the zero - field critical temperature @xmath9 k. these data also show an exceptionally high meissner fraction , @xmath10 .
dependence of the as - measured magnetization , for @xmath11 t perpendicular to the _ ab _ planes .
the line is the background contribution , obtained by fitting a curie - like function in the region 150 - 250 k. the shadow area represents the effect of the superconducting fluctuations .
inset : @xmath12 dependence of the low - field magnetic susceptibility , already corrected for demagnetizing effects , measured under fc and zfc conditions . ]
dependence of the excess magnetization around @xmath3 .
data in the critical region around @xmath13 are indicated in red .
upper inset : detail of the crossing point which reveals its splitting : the @xmath14 curves into the critical region cross at @xmath15 , while the ones outside cross at @xmath16 .
lower inset : @xmath17 phase diagram for the fluctuations .
the red curve is the critical region boundary as estimated from eq .
the data points represent the experimental window for the scaling approach in the critical region .
see the main text for details . ]
the _ in - plane _ diamagnetism ( for @xmath18 applied perpendicularly to the crystallographic _ ab _ planes ) may be parametrized through the so - called _ excess magnetization _ , @xmath19 , which as usual is defined as the difference between the as - measured magnetization and the normal - state or background contribution , @xmath20 .
this last was obtained by fitting to the @xmath21 curves well above @xmath3 ( between @xmath22150 k and @xmath22250 k ) a curie - like function ( see fig . 1 for an example corresponding to @xmath11 t ) .
an overview of the measured @xmath23 curves is presented in fig . 2 .
these data extend to both sides of @xmath3 and cover up to 2.5 orders of magnitude in @xmath18 , which allows a thorough check of the gl approaches in the different regions , where the influence of fluctuating cooper pairs and vortices are expected to be different.@xcite as shown in this figure , the @xmath23 curves cross at a temperature few degrees below @xmath24,@xcite which is a signature of the thermal fluctuations in highly anisotropic superconductors . when expanding these data ( upper inset of fig .
2 ) , the crossing point spans a temperature interval between @xmath25 k for high fields and @xmath26 k for low fields . a similar
splitting has been also observed in other highly anisotropic superconductors @xcite and will be analyzed below in detail .
above but not too close to @xmath3 , on the grounds of the ggl scenario in the 2d limit , well adapted to this highly anisotropic superconductor , @xmath27 is given by @xcite @xmath28 .
\label{prange}\end{aligned}\ ] ] here @xmath29 and @xmath30 are the gamma and digamma functions , @xmath31 the reduced magnetic field , @xmath32 the upper critical field extrapolated to @xmath33 k , @xmath34 the reduced temperature , @xmath35 the number of superconducting cuo@xmath0 layers in their periodicity length ( @xmath36 ) , @xmath37 the effective superconducting volume fraction ( approximated as the meissner fraction @xcite ) , @xmath38 the boltzmann constant , @xmath39 the flux quantum , and @xmath40 the total - energy cutoff constant.@xcite the solid lines in fig .
3(a ) are fits of eq .
( 1 ) in the @xmath41-region @xmath42 to some of the data above @xmath3 of fig . 2 , with @xmath32 as the only free parameter .
these few examples show an excellent agreement down to a few degrees above @xmath24 , where the gaussian approximation is no longer valid ( see below ) , and it leads to @xmath43 t , a value that is going to be used in the remaining analysis above and below @xmath3 . as illustrated by the examples shown in the inset of fig .
3(a ) , the agreement of eq .
( 1 ) with the @xmath18 dependence of @xmath27 measured above @xmath3 is also excellent . and vs. @xmath18 ( inset ) in the gaussian region above @xmath7 .
the lines are the ggl prediction [ eq .
( [ prange ] ) ] .
b ) scaling of the @xmath44 vs @xmath45 curves in the critical region around @xmath13 .
the line is the gl - lll scaling function [ eq .
( [ scalingfunction ] ) ] .
c ) @xmath27 vs @xmath18 in the gaussian region below @xmath13 .
the lines are fits to the blk theory [ eq . ( 6 ) ] .
the resulting @xmath46 is presented in the inset and close to @xmath24 it has the temperature behavior and amplitude predicted by the gl approach .
in contrast , if the conventional london term alone is used , @xmath46 diverges at @xmath47 . in all figures ,
data points in red are in the critical region . ]
near @xmath7 the fluctuations increase so much that the gaussian approximation breaks down . in a magnetic field
sufficiently strong that the fluctuations spectrum is restricted to the lowest landau - level ( lll ) , this critical region for two - dimensional systems is bounded by @xcite @xmath48 where @xmath49 is the specific heat jump at @xmath24 , and @xmath50 the in - plane superconducting coherence length amplitude . by using@xcite @xmath51
j / km@xmath5 and @xmath52 ( which corresponds to the @xmath32 value obtained above in the ggl region above @xmath3 ) , the limits of this _ critical _ region are represented in the lower inset of fig . 2 .
in this lll regime , the gl theory predicts that for two - dimensional systems @xmath27 follows a scaling behavior in the variables @xcite @xmath53 the scaling function being calculated by tesanovi and coworkers as @xcite @xmath54 where @xmath55^{1/2}$ ] , @xmath56 , and @xmath57 is close to the limit of the critical region below @xmath24 when @xmath58 .
equation ( 4 ) predicts a crossing of the @xmath59 curves at @xmath57 and @xmath60 , related through @xmath61 which allows a first direct comparison with the results of fig . 2 .
note first that the high - field crossing point observed at @xmath15 ( upper inset in fig .
2 ) falls into the critical region delimited by eq .
( [ criterio ] ) and , therefore , should be described by eq .
( [ crucealto ] ) . by using @xmath62
k , this equation leads to @xmath63 a / m , which is in good agreement with the experimental value ( @xmath64 a / m ) , taking into account the experimental uncertainties in @xmath65 and in the normal - state @xmath66 contribution . in fig .
3(b ) we represent the @xmath19 data in the critical region in terms of the @xmath45 and @xmath44 variables given by eq .
( 3 ) , calculated by using @xmath67 $ ] and @xmath43 t , the value obtained before in the ggl region above @xmath3 . as can be seen , the scaling of the @xmath68 curves is excellent and the scaling function , evaluated from eq .
( 4 ) by using again the same parameters , is also in agreement with the data at a quantitative level . a good agreement with the gl - lll scaling
was also observed by various groups in different optimally doped and intrinsically underdoped cuprates.@xcite for temperatures well below @xmath7 , outside the critical region , the fluctuation diamagnetism will be dominated by fluctuations of the two - dimensional vortex ( _ pancakes _ ) positions , which in terms of the gl approach may be seen as fluctuations of the order parameter phase .
this contribution has been calculated by bulaevskii , ledvig and kogan ( blk approach ) as,@xcite @xmath69 the first term on the right is the conventional london magnetization , whereas the second one is associated with fluctuations . in this equation
, @xmath70 is the magnetic penetration length in the @xmath71 planes , @xmath72 and @xmath73 are constants of the order of the unity related to the vortex structure , and @xmath74 .
this equation also predicts the crossing of the @xmath23 curves at a temperature @xmath16 a few degrees below @xmath24 , the corresponding excess magnetization being @xmath75 which differs from eq .
( 5 ) for the crossing point in the critical region , by a constant of the order of the unity . by using in eq .
( 7 ) the experimental @xmath16 and @xmath76 values ( upper inset of fig .
2 ) , which are well outside the critical region , one finds @xmath77 , in agreement with the blk approach . a thorough comparison of eq .
( [ blk ] ) with the experimental data is presented in fig .
for each isotherm , the only free parameters are @xmath78 and @xmath70 .
clearly , the fit quality is excellent for isotherms up to @xmath79 . as shown in the inset of fig .
3(c ) , close to @xmath24 the resulting @xmath78 follows a linear temperature dependence and tends to zero at @xmath24 , in full agreement with the conventional ( gl ) behavior for @xmath13 .
moreover , by using the previous @xmath43 t , one obtains , as expected , @xmath80 .
it is worth noting that the london theory alone ( without the term due to vortex fluctuations ) fits the data of fig .
3(c ) as well ( it follows the same @xmath18 dependence ) .
however , as earlier stressed by kogan and coworkers,@xcite it would lead to an anomalous @xmath13 close to @xmath24 , due to the presence of the crossing point .
this seemingly anomalous @xmath13 behavior has been recently presented as the hallmark of unconventional behavior for @xmath27.@xcite
we have presented detailled measurements of the diamagnetism around the meissner transition in a high quality tl-2223 single crystal .
these results were analyzed in terms of the ginzburg - landau approach with superconducting fluctuations in the three different fluctuation regions : the gaussian - ginzburg - landau regions above and below @xmath3 and the critical region closer to @xmath7 .
these analysis extend previous comparisons , up to now performed consistently only in one or two of these fluctuation regions.@xcite they provide , therefore , an important test of consistency for the applicability of the gl approaches to describe the meissner transition in an optimally doped cuprate .
when combined with previous measurements of different observables around @xmath3,@xcite in particular of @xmath27 in underdoped cuprates,@xcite our present results provide a strong experimental confirmation that the meissner transition in all cuprates is , independently of the doping , a conventional gl transition , although in many cases entangled with , intrinsic or not , disorder and inhomogeneities .
this work may then help to reach a general consensus on the phenomenological descriptions of the superconducting transition in cuprates , which in turn will provide a constraint for the explanations of the critical behavior of any observable around @xmath3 @xcite and , more important , for any proposal of pairing mechanism in these materials .
we thank a. maignan and a. wahl for providing us the tl-2223 crystals , the spanish ministerio de y ciencia ( grant no .
mat2004 - 04364 ) and the xunta de galicia ( grant no .
pgidit04tmt206002pr ) .
l.c . acknowledges support through a fpu grant .
j. mosqueira , c. carballeira , m.v .
ramallo , c. torrn , j.a .
veira and f. vidal , europhys . lett . *
53 * , 632 ( 2001 ) ; f. vidal , c. carballeira , s.r .
currs , j. mosqueira , m.v .
ramallo , j.a .
veira and j. via , _ ibid . _ * 59 * , 754 ( 2002 ) .
see e.g. , q. li , in _ physical properties of high temperature superconductors v _ , edited by d.m .
ginsberg ( world scientific , 1996 ) , chap . 4
; see also , a. a. varlamov , g. balestrino , e. milani , and d. v. livanov , adv . phys . *
48*(6 ) , 655 ( 1999 ) .
see e.g. , f. vidal and m.v .
ramallo in _ the gap symmetry and fluctuations in high - tc superconductors _ , vol .
371 of nato advanced study institute , series b : physics , edited by j. bok _ et al . _ ,
plenum , new york , 1998 , p.443 .
p. carretta , a. lascialfari , a. rigamonti , a. rosso , and a. varlamov , phys .
b * 61 * , 12420 ( 2000 ) ; a. lascialfari , a. rigamonti , l. romano , a.a . varlamov , and i. zucca , _ ibid . _ * 68 * , 100505(r ) ( 2003 ) , and references therein .
l. cabo , f. soto , m. ruibal , j. mosqueira , and f. vidal , phys .
b * 73 * , 184520 ( 2006 ) . for a review on the rounding effects around the superconducting transition associated with @xmath3 inhomogeneities at long length scales see f. vidal , j. a. veira , j. maza , j. mosqueira , and c. carballeira , _ materials science , fundamental properties and future electronic applications of high - tc superconductors _ , nato asi series , edited by s.l .
dreschler and t. mishonov ( kluwer - dordrech , amsterdam ) 2001 , p. 289 .
a recent version of this paper may be seen in arxiv : cond - mat/0510467 .
a. maignan , c. martin . v. hardy , ch .
simon , m. hervieu , and b. raveau , physica c * 219 * , 407 ( 1994 ) ; some earlier magnetization measurements may be seen in f. vidal et al .
, in _ superconducting and related oxides : physics and nanoengineering iii _ , edited by d. pavuna and i. bozovic , proceedings of spie volume 3481 ( 1998 ) , p. 32
just as an example , our present results do not support an anomalous ( non - gl ) superconducting transition in cuprates and , therefore , they favor the explanations of the high nernst signal observed above @xmath3 in htsc on the grounds of the ggl approach or as a normal state effect .
see e.g. , i. ussishkin , s.l .
sondhi , and d. huse , phys .
* 89 * , 287001 ( 2002 ) ; a.s .
alexandrov and v.n .
zavaritsky , _ ibid . _ * 93 * , 217002 ( 2004 ) ; _ ibid . _ * 95 * , 259704 ( 2005 ) . | the in - plane diamagnetism around the meissner transition was measured in a tl@xmath0ba@xmath0ca@xmath0cu@xmath1o@xmath2 single crystal of high chemical and structural quality , which minimizes the inhomogeneity and disorder rounding effects on the magnetization .
when analyzed quantitatively and consistently above and below the transition in terms of the ginzburg - landau ( gl ) approach with fluctuations of cooper pairs and vortices , these data provide a further confirmation that the observed meissner transition is a conventional gl superconducting transition in a homogeneous layered superconductor . |
quantum anomalous hall effect ( qahe ) is a quantum mechanical version of the hall effect in a ferromagnet in absence of an external magnetic field or landau levels .
different from the quantum hall effect in a strong magnetic field , it originates from the topological properties of band structure in solid .
usually the anomalous hall effect occurs in ferromagnetic metals due to the intrinsic spin - orbit coupling or extrinsic spin - orbit scattering.@xcite it has been realized that the hall conductance can be expressed as the summation of the berry curvature in momentum space over all occupied states of electrons .
this makes it possible to realize the quantum hall effect even without the landau levels.@xcite a simple picture for qahe was proposed in a two - dimensional electron gas with strong spin - orbit coupling.@xcite electron localization in disordered systems also provides an alternative approach to realize this effect.@xcite the discovery of the quantum spin hall effect ( qshe ) and topological insulators@xcite stimulated extensive interest to search qahe in realistic systems . as a result of time - reversal symmetry breaking in topological insulators ,
this effect was predicted in a magnetically doped hg@xmath0mn@xmath1te
quantum well@xcite and cr or fe doped topological insulators,@xcite in which the presence of magnetization suppresses one of helical edge states in qshe , and preserves a chiral edge state for the quantum hall effect .
very recently hgcr@xmath2se@xmath3 is proposed to be a chern semimetal and exhibits qahe in quantum - well structure.@xcite interaction - driven topologically non - trivial mott insulating phase displaying qahe or qshe has also attracted much research interest.@xcite while in search of ferromagnetism in diluted magnetic semiconductors , rigorous models of ferromagnetism@xcite or ferrimagnetism @xcite in strongly correlated electron systems provides an alternative context to realize ferromagnetism . a common feature of these examples is the flat or almost flat band of electrons . the coulomb interaction may remove high degeneracy of electrons in the band , leading to the ferromagnetism according to the stoner criteria .
recently it was predicted that the ferromagnetism may occur in the hubbard model with topological non - trivial bands.@xcite in this paper we propose a theory of qahe in a flat band ferromagnet with spin - orbit coupling .
we start with a two - dimensional decorated lattice model , which has a pair of flat bands in the middle of the energy spectrum .
the inclusion of spin - orbit coupling makes the energy bands topologically non - trivial , and gives rise to qshe .
based on the rigorous results for the hubbard model , the coulomb interaction may induce the ferrimagnetism in the ground state , when the middle flat bands are half filled .
it provides an ideal way to realize a magnetic staggered field .
the staggered field can modulate the topological numbers of electron bands by closing and reopening the energy gap.@xcite different configurations of topological invariants assigned to the electron bands can then be obtained and will give different topological phases . based on a self - consistent mean field calculation , we present the phase diagram of the ground state . we find that qahe with nonzero chern number can be realized in a ferromagnet due to the coulomb interaction .
we begin with the tight - binding hamiltonian on a two - dimensional decorated lattice [ see fig . [ fig01](a ) ] , @xmath4 where the spin - independent hopping term is given by @xmath5 @xmath6 and @xmath7 are the annihilation and creation operators of electron with spin @xmath8 on site @xmath9 .
@xmath10 means the summation over the nearest - neighbor sites .
@xmath11 is the hopping amplitude .
the spin - orbit coupling term has the form @xmath12c_{l},\ ] ] which gives a spin - dependent hopping between the next - nearest - neighbor sites with hopping amplitude @xmath13 [ shown by the dash lines in fig . [
fig01](a ) ] . here
the lattice is divided into two sublattices , @xmath14 and @xmath15 , shown by the light and dark dots in fig .
[ fig01](a ) , respectively . @xmath16 and @xmath17 denote the adjacent sites of site @xmath9 , and @xmath18 is the unit vector along the direction from site @xmath9 to site @xmath16 , and @xmath19 are the pauli matrices . and @xmath15 , respectively .
( b ) the energy dispersion for @xmath20 , which has three pairs of doubly degenerate bands , and the gaps open in the presence of spin - orbit coupling ( @xmath21).,scaledwidth=22.0% ] we choose the sites 1 , 2 , and 3 in fig .
[ fig01](a ) as the unit cell .
since the @xmath22 component of spin @xmath23 commutes with this hamiltonian , the hamiltonian has a block - diagonalized form in momentum space after the fourier transformation , @xmath24 where @xmath25 , @xmath26 denotes different spins , and @xmath27 @xmath28 is the time - reversal partner of @xmath29 .
the brillouin zone spans over @xmath30 and @xmath31 .
this hamiltonian preserves time - reversal symmetry , i.e. , @xmath32 , where @xmath33 and @xmath34 is the complex conjugate operator . in this spin-@xmath35 system ,
the time - reversal operator satisfies @xmath36 .
when @xmath21 , in each @xmath37 , all the three bands are well separated and can be characterized by the chern number@xcite @xmath38 _ { z},\ ] ] where @xmath39 and @xmath40 is the bloch function for the @xmath41th band of electrons with spin @xmath42 . since @xmath43 and @xmath44 are time - reversal partners , we have @xmath45 for each time - reversal pairs of bands , which are degenerate due to time - reversal symmetry .
therefore , the total chern number is zero . in the presence of spin - orbit coupling ,
the chern numbers of the three bands in @xmath43 ( @xmath46 ) are @xmath47 ( @xmath48 ) from top to bottom , with @xmath49 .
the nonzero difference between @xmath50 and @xmath51 is equivalent to a non - trivial z@xmath52 index , which can also be calculated explicitly.@xcite when the fermi level is located in the gap , the non - trivial @xmath53 index for the filled pairs of bands indicates qshe.@xcite
for an intuitive illustration on how qahe arises in this system , we first introduce a spin - dependent staggered potential term @xmath54 and a spin - independent staggered potential term @xmath55 respectively , where the summations run over the sublattice sites . in momentum space ,
the hamiltonian @xmath56 has a block - diagonalized form , @xmath57 with @xmath58 and @xmath59 , where @xmath60 and @xmath61 . when @xmath62 , the time - reversal symmetry is broken and the degeneracy of the time - reversal pair of bands is removed .
in the block - diagonalized form , we may say that @xmath29 feels a staggered potential of amplitude @xmath63 while @xmath44 feels @xmath64 .
these two parts of the hamiltonian can be investigated separately .
we notice that a staggered potential may change the chern numbers of the bands of @xmath65 and @xmath66 by closing and reopening the band gap in a band inversion . for @xmath65 ,
the bands cross only at the point @xmath67 when @xmath68 , @xmath21 .
the eigenvalues at this point are @xmath69 , respectively . as a result , with increasing @xmath70 from zero
, a band crossing happens at @xmath71 or @xmath72 .
for example near @xmath73 , we may obtain an effective hamiltonian near the point @xmath74 @xmath75 where @xmath76 and @xmath77 .
from this two - band massive dirac model it is known that the topological quantum phase transition occurs when the sign of @xmath78 or @xmath79 changes .
the chern number of the lower band is given by@xcite ( the upper one has a sign change ) @xmath80 .\ ] ] thus the sign change of @xmath78 indicates that the chern number changes from 0 to 1 or 1 to 0 .
[ fig02 ] depicts the band structure for @xmath65 and @xmath81 , with the chern numbers also denoted for each band .
there are three different cases : ( a ) when @xmath82 , the chern numbers of the bands in @xmath65 are @xmath83 from top to bottom ; ( b ) when @xmath84 , the two lower bands touch at * * k**@xmath85 and the chern numbers disappear as the two bands are not well separated ; ( c ) when @xmath86 the band gap reopens and the chern numbers become @xmath87 after the two inverted bands exchange their chern numbers . similarly , for @xmath81 ,
the chern numbers of the three bands are @xmath88 when @xmath89 and @xmath90 when @xmath91 .
it is noted that the band structure and chern numbers of @xmath81 with @xmath92 and @xmath13 are identical to those of @xmath65 with @xmath93 and @xmath94 .
and @xmath44 along @xmath95 $ ] , at @xmath96 .
the horizontal axis is @xmath97 and the vertical axis is @xmath98 .
we take the parameter @xmath99 . 0 and
@xmath100 indicate the chern number of each band . from the left to right , @xmath101 , @xmath102 , and @xmath103 .
@xmath44 with @xmath104 has the identical bands structure as @xmath29 with @xmath105 , but opposite chern numbers.,scaledwidth=50.0% ] in a strip geometry .
edge state spectra are shown by dark gray lines ( or red online ) .
( a ) @xmath106 .
( b ) @xmath107.,scaledwidth=50.0% ] , @xmath81 and @xmath108 , respectively .
the horizontal axis is @xmath109 $ ] and the vertical axis is @xmath98.,title="fig:",scaledwidth=50.0% ] , @xmath81 and @xmath108 , respectively .
the horizontal axis is @xmath109 $ ] and the vertical axis is @xmath98.,title="fig:",scaledwidth=50.0% ] , @xmath81 and @xmath108 , respectively .
the horizontal axis is @xmath109 $ ] and the vertical axis is @xmath98.,title="fig:",scaledwidth=50.0% ] according to the chern numbers , the edge - bulk correspondence tells that the edge states in a sample of strip geometry depend on parameters . in fig .
[ fig03 ] , we only present the edge state spectra of @xmath65 . when @xmath110 , the edge state spectra connect the upper and lower bands . since the chern number of the middle band is zero , this topologically trivial band only distorts the edge state spectra , but does not affect the existence of the edge states . the total chern number is still 1 and the system is topologically non - trivial when the middle and lower bands are fully filled .
when @xmath86 , the middle band becomes topologically non - trivial , and the lower band becomes trivial .
the edge state spectra only connect the middle and the upper bands .
now we can present the evolution of edge states in the total hamiltonian @xmath108 .
five cases are listed in fig .
[ fig04 ] . without loss of generality , we take @xmath111 and @xmath112 . in this case
we have @xmath113 .
when time - reversal symmetry is broken , an energy gap can open between a time - reversal pair of bands . a uniform magnetism term @xmath114 is introduced to shift the bands of @xmath115 and @xmath81 upward and downward , respectively , without changing the chern number of each band .
a gap is opened between two middle bands . at half filling , we assume that the fermi level is located in this gap . according to the chern numbers and relative positions of the energy bands , the system can be categorized into five cases .
case i : @xmath116 and @xmath117 .
the total chern numbers of three lower bands is @xmath118 .
however , the nonzero difference between the chern numbers of @xmath65 and @xmath119 indicates qshe .
we may have two counter - propagating edge states with different spins on each edge , although these two edge states do not form a time - reversal pair as time - reversal symmetry has already been broken .
case ii : @xmath120 and @xmath121 .
the chern numbers of both @xmath65 and @xmath66 change to zero due to the @xmath122 term . when the middle band of @xmath123 is higher than the middle band of @xmath124 , the system is in an insulating phase as shown in the figure .
when the middle band of @xmath65 is lower , the system exhibits qshe .
case iii and case iv : @xmath120 and @xmath117 .
case iii , when the middle band of @xmath65 is higher than the middle band of @xmath81 , the total chern number is @xmath125 .
this is a qahe phase , in which there exists one gapless spin - up chiral edge state .
case iv : when the middle band of @xmath65 is lower than the middle band of @xmath81 , the total chern number is @xmath126 , but it gives rise to qshe . to distinguish case iii and case iv , one can check the eigenvalues at @xmath127 , and we have case iii when @xmath128 .
case v : @xmath120 and @xmath129 . due to @xmath130 term , the chern numbers of both @xmath65 and @xmath81 change , and the total chern number is @xmath125 .
once again the system exhibits qahe as there is only one chiral edge state .
one of the prominent features of the model in eq . is the appearance of the flat bands due to the unequal numbers of sites of the two sublattices , even in the presence of spin - orbit coupling .
these flat bands give rise to the famous flat - band ferromagnetism when the coulomb interaction is turned on .
@xcite when the system is at half filling , the two lower bands are fully filled , and the total spin of these two bands is zero since electrons in these two bands have opposite spins .
the two middle bands are degenerate . in this case , if only one single middle band is fully filled , the expectation value of the coulomb interaction is minimized since the fully polarized electron spin in the middle band excludes the double occupancy completely at each site . in this way
, the ground state of the system is ferromagnetic .
the total spin is given by the degeneracy of the flat band , @xmath131.@xcite furthermore , since the antiferromagnetic correlation is dominant in the half - filled hubbard model , this ground state is actually ferrimagnetic in which ferromagnetic and antiferromagnetic long range orders coexist.@xcite when the spin - orbit coupling is present , the flat bands will be distorted by the ferrimagnetism .
when the coupling is strong , the ferrimagnetism would significantly distort the flat bands and is suppressed .
let us focus on the case of weak spin - orbit coupling , in which the band is expected to be almost flat .
it is still possible that the ferrimagnetism could survive if the coulomb interaction is strong enough over the band distortion.@xcite thus the combination of the flat band and the coulomb interaction provides a reliable mechanism to realize the spin - dependent staggered potential in this system .
the hamiltonian with the on - site and nearest - neighbor repulsive interactions has the form @xmath132 where @xmath133 is the on - site coulomb potential , @xmath134 is the nearest - neighbor repulsive potential , @xmath135 is the number operator for electron with spin @xmath136 on site @xmath9 , and @xmath137 .
the chemical potential @xmath138 determines the number of electrons in the system .
@xmath139 here is the number of unit cells , and the number of electrons is @xmath140 at half filling . in the mean field approximation ,
the on - site interaction is decoupled as@xcite @xmath141 where @xmath142 is the magnetization on site @xmath9 . due to the asymmetry of the two sublattices , the magnetization on site a and b
are different , saying , @xmath143 for the sublattice @xmath14 and @xmath144 for the sublattice @xmath15 . in this way the hubbard term is reduced to @xmath145 at half filling , since the antiferromagnetic correlation is dominant , @xmath143 and @xmath144 have opposite signs , based on the rigorous results for the hubbard model.@xcite this can also be illustrated from the calculation of the mean field theory .
the nearest - neighbor interaction may induce the instability of charge - density wave ( cdw ) . in the mean field approximation , we have @xmath146 where @xmath147 for the sublattice @xmath148 and @xmath149 for the sublattice @xmath150 .
the cdw order parameter @xmath151 is given by @xmath152 and @xmath153 due to the charge conservation .
phase diagram for @xmath154 at half filling .
the shadow marks the gapless regions at half filling , which should be metallic with strong anomalous hall effect and spin hall effect.,scaledwidth=45.0% ] after some tedious algebra we may have the zero - temperature mean field free energy at half filling @xmath155\left[- \mu+e_{i}(\mathbf{k})\right]+n(e_{0}+3\mu),\end{aligned}}\ ] ] where @xmath156 is the step function , @xmath157 is the eigenvalues of @xmath56 in eq . with @xmath158 and @xmath159 , where @xmath160 , @xmath161 , and @xmath162 .
the summation runs over the whole brillouin zone .
the order parameters @xmath143 , @xmath144 , @xmath164 , and @xmath138 can be determined self - consistently by minimizing the free energy .
the variational principle @xmath165 leads to a set of the mean field equation , @xmath166 we solve this set of equations numerically .
the calculated mean field results are consistent with the rigorous results for the hubbard model . @xmath167 and
@xmath168 have different signs , which demonstrates the existence of antiferromagnetic correlation .
@xmath169 demonstrates the ferromagnetic correlation . if @xmath170 , @xmath171 , which is one of the rigorous results for the hubbard model
. figure .
[ fig05 ] shows the mean field @xmath172 phase diagram for @xmath173 and @xmath174 .
the cdw order is zero when @xmath175 is small .
@xmath122 and @xmath176 would increase with @xmath177 and have the same sign .
the system transits from case i to case ii through a band inversion .
however we may see that the gap at half filling is not opened in the dashed area of fig .
[ fig05 ] .
when this gap opens , case i gives an afm quantum spin hall effect , and case ii is an afm insulating phase .
when @xmath175 is large , the cdw order may become nonzero and increase with @xmath175 .
case iii can be found near the transition point where @xmath130 is small , and case v can be found at larger @xmath175 where @xmath130 becomes large .
they both present qahe when the gap opens at half filling .
if large @xmath13 is chosen , the @xmath130 term would distort flat bands so much that the orders of @xmath122 and @xmath176 are suppressed .
thus to have a qahe , we need strong on - site interaction @xmath177 and a nonzero spin - independent staggered field .
we have found that the flat - band ferromagnet may exhibit qahe after the inclusion of spin - orbit coupling on a 2d decorated lattice .
the spin - orbit coupling can induce topologically non - trivial phase on this lattice , which exhibits qshe . in the present three - band system
, the existence of the topologically trivial flat band between the two non - trivial bands does not affect the formation of qshe .
the coulomb interaction may remove the degeneracy of electrons in the flat band and lead to spontaneous symmetry breaking , which gives rise to ferrimagnetism .
the coexistence of ferrimagnetism and cdw may break the balance between the helical edge states with spin up and spin down in qshe , and make it possible that one branch of edge states is suppressed completely , and the other survives . as a result
, it gives rise to qahe . | we proposed a theory of quantum anomalous hall effect in a flat - band ferromagnet on a two - dimensional ( 2d ) decorated lattice with spin - orbit coupling .
free electrons on the lattice have dispersionless flat bands , and the ground state is highly degenerate when each lattice site is occupied averagely by one electron , i.e. , the system is at half filling .
the on - site coulomb interaction can remove the degeneracy and give rise to the ferrimagnetism , which is the coexistence of the ferromagnetic and antiferromagnetic long - range orders . on the other hand
the spin - orbit coupling makes the band structure topologically non - trivial , and produces the quantum spin hall effect with a pair of helical edge states around the system boundary .
based on the rigorous results for the hubbard model , we found that the coulomb interaction can provide an effective staggered potential and turn the quantum spin hall phase into a quantum anomalous hall phase . |
the power systems are undergoing drastic changes on both the supply side and the demand side . on the supply side , an increasing amount of renewable energy ( e.g. wind energy , solar energy ) is penetrating the power systems .
the adoption of renewable energy reduces the environmental damage caused by conventional energy generation , but also introduces high fluctuation and uncertainty in energy generation on the supply side . to cope with this uncertainty in energy generation , the demand side is deploying various solutions ,
one of which is the use of energy storage @xcite . in this paper
, we study the optimal demand side management ( dsm ) strategy in the presence of energy storage , and the corresponding optimal economic dispatch strategy .
specifically , we consider a power system consisting of energy generators on the supply side , an independent system operator ( iso ) that operates the system , and multiple aggregators and their customers on the demand side . on the supply side ,
the iso receives energy purchase requests from the aggregators as well as reports of ( parameterized ) energy generation cost functions from the generators , and based on these , dispatches the energy generators and determines the unit energy prices . on the demand side ,
the aggregators are located in different geographical areas and provide energy for residential customers ( e.g. households ) or for commercial customers ( e.g. an office building ) in the neighborhood . in the literature , the term `` dsm '' has been broadly used for different decision problems on the demand side . for example , some papers ( see @xcite@xcite for representative papers ) focus on the interaction between one aggregator and its customers , and refer to dsm as determining the power consumption schedules of the users .
some papers @xcite@xcite focus on how multiple aggregators @xcite@xcite or a single aggregator @xcite@xcite purchases energy from the iso based on the energy consumption requests from their customers .
our paper pertains to the second category of research works .
the key feature that sets apart our paper from most existing works @xcite@xcite is that all the decision makers in the system are _
foresighted_. each aggregator seeks to minimize its _ long - term _ cost , consisting of its operational cost of energy storage and its payment for energy purchase .
in contrast , in most existing works @xcite@xcite , the aggregators are _ myopic _ and seek to minimizing their _ short - term _
( e.g. one - day or even hourly ) cost . in the presence of energy storage ,
foresighted dsm strategies can achieve much lower costs than myopic dsm strategies because the current decisions of the aggregators will affect their future costs .
for example , an aggregator can purchase more energy from the iso than that requested from its customers , and store the unused energy in the energy storage for future use , if it anticipates that the future energy price will be high .
hence , the current purchase from the aggregators will affect how much they will purchase in the future . in this case
, it is optimal for the entities to make _ foresighted _ decisions , taking into account the impact of their current decisions on the future .
since the aggregators deploy foresighted dsm strategies , it is also optimal for the iso to make foresighted economic dispatch , in order to minimize the _ long - term _ total cost of the system , consisting of the long - term cost of energy generation and the aggregators long - term operational cost .
note that although some works @xcite@xcite assume that the aggregator is foresighted , they study the decision problem of a _ single _ aggregator and do not consider the economic dispatch problem of the iso .
when there are multiple aggregators in the system ( which is the case in practice ) , this approach neglects the impact of aggregators decisions on each other , which leads to suboptimal solutions in terms of minimizing the total cost of the system . when the iso and _ multiple _ aggregators make _ foresighted _ decisions , it is difficult to obtain the optimal foresighted strategies for two reasons
. first , the information is decentralized .
the total cost depends on the generation cost functions ( e.g. the speed of wind for wind energy generation , the amount of sunshine for solar energy generation , and so on ) , the status of the transmission lines ( e.g. the flow capacity of the transmission lines ) , the amount of electricity in the energy storage , and the demand from the customers , all of which may change due to supply and demand uncertainty .
however , none of the entities knows all the above information : the iso knows only the generation cost functions and the status of the transmission lines , and each aggregator knows only the status of its own energy storage and the demand from its own customers .
hence , the dsm strategy needs to be decentralized , such that each entity can make decisions solely based on its locally - available information .
second , the aggregators are coupled in a complicated way that is unknown to them .
specifically , each aggregator s purchase affects the prices , and thus the payments of the other aggregators .
however , the price is determined by the iso based on the generation cost functions and the status of the transmission lines , neither of which is known to any aggregator .
hence , each aggregator does not know how its purchase will influence the price , which makes it difficult for the aggregator to make the optimal decision . to overcome the difficulty resulting from information decentralization and complicated coupling , we propose a decentralized dsm strategy based on conjectured prices . specifically , each aggregator makes decisions based on its conjectured price , and its local information on the status of its energy storage and the demand from its customers . in other words
, each aggregator summarizes all the unavailable information into its conjectured price .
note , however , that the price is determined based on the generation cost functions and the status of the transmission lines , which is only known to the iso .
hence , the aggregators conjectured prices are determined by the iso .
we propose a simple online algorithm for the iso to update the conjectured prices based on its local information , and prove that by using the algorithm , the iso obtains the optimal conjectured prices under which the aggregators ( foresighted ) best responses minimize the total cost of the system .
.comparisons with related works on demand - side management . [ cols="^,^,^,^,^,^,^",options="header " , ]
in this paper , we proposed a methodology to perform optimal foresighted dsm strategies that minimize the long - term total cost of the power system .
we overcame the hurdles of information decentralization and complicated coupling in the system , by decoupling the entities decision problems using conjectured prices .
we proposed an online algorithm for the iso to update the conjectured prices , such that the conjectured prices can converge to the optimal ones , based on which the entities make optimal decisions that minimize the long - term total cost .
we prove that the proposed method can achieve the social optimum , and demonstrate through simulations that the proposed foresighted dsm significantly reduces the total cost compared to the optimal myopic dsm ( up to 60% reduction ) , and the foresighted dsm based on the lyapunov optimization framework ( up to 30% reduction ) .
due to limited space , we give a detailed proof sketch .
the proof consists of three key steps .
first , we prove that by penalizing the constraints @xmath0 into the objective function , the decision problems of different entities can be decentralized .
hence , we can derive optimal decentralized strategies for different entities under given lagrangian multipliers .
then we prove that the update of lagrangian multipliers converges to the optimal ones under which there is no duality gap between the primal problem and the dual problem , due to the convexity assumptions made on the cost functions .
finally , we validate the calculation of the conjectured prices .
first , suppose that there is a central controller that knows everything about the system .
then the optimal strategy to the design problem should result in a value function @xmath1 that satisfies the following bellman equation : for all @xmath2 , we have @xmath3 defining a lagrangian multiplier @xmath4 associated with the constraints @xmath5 , and penalizing the constraints on the objective function , we get the following bellman equation : @xmath6 \right . \\ & & \left
. + \delta \cdot \sum_{s_0^\prime,\bm{s}^\prime } \rho(s_0^\prime,\bm{s}^\prime|s_0,\bm{s},a_0,\bm{a } ) v^{\bm{\lambda}}(s_0^\prime,\bm{s}^\prime ) \right\}. \nonumber\end{aligned}\ ] ] in the following lemma , we can prove that can be decomposed .
[ lemma : decompositionvaluefunction ] the optimal value function @xmath7 that solves can be decomposed as @xmath8 for all @xmath9 , where @xmath10 can be computed by entity @xmath11 locally by solving @xmath12 + \delta \cdot \sum_{s_i^\prime } \rho_i(s_i^\prime|s_i , a_i ) v_i^{\bm{\lambda}}(s_i^\prime ) \right\ } \end{array}.\end{aligned}\ ] ] this can be proved by the independence of different entities states and by the decomposition of the constraints @xmath0 .
specifically , in a dc power flow model , the constraints @xmath0 are linear with respect to the actions @xmath13 . as a result
, we can decompose the constraints as @xmath14 .
we have proved that by penalizing the constraints @xmath0 using lagrangian multiplier @xmath15 , different entities can compute the optimal value function @xmath16 distributively .
due to the convexity assumptions on the cost functions , we can show that the primal problem is convex .
hence , there is no duality gap . in other words , at the optimal lagrangian multipliers @xmath17 , the corresponding value function @xmath18 is equal to the optimal value function @xmath19 of the primal problem .
it is left to show that the update of lagrangian multipliers converge to the optimal ones .
it is a well - known result in dynamic programming that @xmath20 is convex and piecewise linear in @xmath15 , and that the subgradient is @xmath0 .
note that we use the sample mean of @xmath21 and @xmath22 , whose expectation is the true mean value of @xmath21 and @xmath22 . since @xmath0 is linear in @xmath21 and @xmath22 , the subgradient calculated based on the sample mean has the same mean value as the subgradient calculated based on the true mean values . in other words , the update is a stochastic subgradient descent method .
it is well - known that when the stepsize @xmath23 , the stochastic subgradient descent will converge to the optimal @xmath24 .
h. mohsenian - rad , v. w. s. wong , j. jatskevich , r. schober , and a. leon - garcia , `` autonomous demand - side management based on game - theoretic energy consumption scheduling for the future smart grid , '' _ ieee trans .
smart grid _ , vol . 1 , no
320331 , 2011 .
kim , s. ren , m. van der schaar , and j .- w .
lee , `` bidirectional energy trading and residential load scheduling with electric vehicles in the smart grid , '' _ ieee j. sel .
areas commun .
, special issue on smart grid communications series _ , vol .
31 , no . 7 , pp
. 12191234 , jul .
2013 .
italo atzeni , luis g. ordez , gesualdo scutari , daniel p. palomar , and javier r. fonollosa , `` noncooperative and cooperative optimization of distributed energy generation and storage in the demand - side of the smart grid , '' _ ieee trans . on signal process .
24542472 , may 2013 .
italo atzeni , luis g. ordez , gesualdo scutari , daniel p. palomar , and javier r. fonollosa , `` demand - side management via distributed energy generation and storage optimization , '' _ ieee trans . on smart grids _ ,
2 , pp . 866876 , june 2013 . e. altmam , k. avrachenkov , n. bonneau , m. debbah , r. el - azouzi , d. s. menasche , `` constrained cost - coupled stochastic games with independent state processes , '' _
technical report_. available : http://www-sop.inria.fr/members/konstantin.avratchenkov/pubs/constrgame.pdf j. hrner , t. sugaya , s. takahashi , and n. vielle , `` recursive methods in discounted stochastic games : an algorithm for @xmath27 and a folk theorem , '' _ econometrica _ , vol .
79 , no . 4 , pp .
12771318 , 2011 . | we consider a smart grid with an independent system operator ( iso ) , and distributed aggregators who have energy storage and purchase energy from the iso to serve its customers .
all the entities in the system are _ foresighted _ : each aggregator seeks to minimize its own _ long - term _
payments for energy purchase and operational costs of energy storage by deciding how much energy to buy from the iso , and the iso seeks to minimize the _ long - term _ total cost of the system ( e.g. energy generation costs and the aggregators costs ) by dispatching the energy production among the generators .
the decision making of the foresighted entities is complicated for two reasons .
first , the information is decentralized among the entities : the iso does not know the aggregators states ( i.e. their energy consumption requests from customers and the amount of energy in their storage ) , and each aggregator does not know the other aggregators states or the iso s state ( i.e. the energy generation costs and the status of the transmission lines ) .
second , the coupling among the aggregators is unknown to them due to their limited information .
specifically , each aggregator s energy purchase affects the price , and hence the payments of the other aggregators .
however , none of them knows how its decision influences the price because the price is determined by the iso based on its state .
we propose a design framework in which the iso provides each aggregator with a conjectured future price , and each aggregator distributively minimizes its own long - term cost based on its conjectured price as well as its locally - available information . the proposed framework can achieve the social optimum despite being decentralized and involving complex coupling among the various entities interacting in the system .
simulation results show that the proposed foresighted demand side management achieves significant reduction in the total cost , compared to the optimal myopic demand side management ( up to 60% reduction ) , and the foresighted demand side management based on the lyapunov optimization framework ( up to 30% reduction ) . |
the squeezed state is a nonclassical state , which has been discovered for a long time ( see , e.g. , a recent review @xcite and references therein ) . besides recognized potential applications in detecting extremely weak forces , squeezed states have recently attracted much more attention in the newly emerging field of quantum information and quantum computation . by means of squeezed states schemes for teleportation of continuous quantum variables @xcite , quantum cryptography @xcite , entanglement distribution @xcite , etc .
have been proposed and some of them have been realized .
the original squeezed state @xcite has been generalized to different types of higher - order .
one type of higher - order amplitude squeezing is suggested by hong and mandel @xcite .
another qualitatively different type is introduced by hillery @xcite which is then developed further by several authors ( see , e.g. , @xcite ) .
the hong - mandel type higher - order amplitude squeezing has been investigated in the fan - state @xmath12 which is introduced in @xcite as a linear superposition of @xmath8 @xmath8-quantum nonlinear coherent states @xmath13 @xmath14 @xcite in a phase - locked manner .
since there are states that exhibit hong - mandel type squeezing but not hillery - type one and vice versa @xcite , we would like to examine in this paper whether the hillery - type @xmath0-th power amplitude squeezing is possible and , if it is , how it behaves in the same fan - state .
the normalized fan - state is defined as @xmath15 where @xmath16 @xmath17 @xmath18 @xmath4 is in general an arbitrary real nonlinear operator - valued function ( in particular , it may be that @xmath9 ) of @xmath19 with @xmath20 @xmath21 the bosonic annihilation ( creation ) operator , @xmath22^{2 } } , \label{e1}\ ] ] with @xmath23 and the @xmath8-quantum nonlinear coherent state is given by @xmath24 with @xmath25 a fock state .
the state @xmath26 is a sub - state of the multi - quantum nonlinear coherent state which is by definition @xcite the eigenstate of the nonboson operator @xmath27 the notation @xmath28 appearing in eqs .
( [ e1 ] ) and ( [ e2 ] ) is understood as follows @xmath29 the hillery - type @xmath0-th power amplitude squeezing is associated with the operator @xmath30 of the form @xmath31 with @xmath32 an angle determining the direction of @xmath33 in the complex plane . according to @xcite , a state @xmath34 is said to be @xmath0-th power amplitude squeezed along the direction @xmath32 if there exists a value of @xmath32 such that the variance of @xmath35 satisfies the following inequality @xmath36 where @xmath37\rangle .
\label{rr3}\ ] ] it is easy to express the variance as @xmath38 where @xmath39}-2\left ( \re [ e^{-in\varphi } { \langle } a^{n}{\rangle ] } \right ) ^{2}\ } \label{r5}\ ] ] with @xmath40 denotes a normal ordering of the operators .
the @xmath41 can also be normally ordered and its explicit form reads @xcite @xmath42 with @xmath43 .
combining the inequality ( [ r3 ] ) and eq .
( [ r4 ] ) we recognize that the state becomes squeezed whenever @xmath44 gets negative . for convenience , we define a measure of squeezing degree by a quantity @xmath45 scaled as @xmath46 clearly that for a squeezed state @xmath47 and the ideal squeezing corresponds to @xmath48 without any loss of generality we choose the real axis to be along the direction of @xmath49 in the phase space .
this allows us to treat @xmath49 as a real number . in the fan - state
, we have @xmath50 where @xmath51 .
the function @xmath52 equals unity ( zero ) if @xmath53 is an integer ( non - integer ) and @xmath54 is the step function : @xmath55 @xmath56 for @xmath57 @xmath58 . putting eqs .
( [ r5 ] ) and ( [ r6 ] ) into eq .
( [ s ] ) and making a simple trigonometric transformation , we arrive explicitly at @xmath59 where @xmath60 @xmath61 @xmath62 and @xmath63 are determined by ( [ alm ] ) if @xmath64 is set to be @xmath65 @xmath66 @xmath67 and @xmath68 in what follows we distinguish two cases : @xmath9 which corresponds to light field and @xmath10 which may be associated with the vibrational motion of the center - of - mass of a trapped ion .
a qed - based scheme to generate the fan - state with @xmath9 which concerns radiation field in a cavity has been proposed in @xcite . for @xmath69 the expectation values @xmath70 and
@xmath71 simplify to @xmath72 @xmath73 using the property of the function @xmath74@xmath75 it follows from eqs .
( [ an ] ) and ( [ a2n ] ) that , for a given @xmath76 both @xmath77 and @xmath78 vanish if @xmath79 with @xmath80 on the other hand , @xmath81 is always positive .
hence , for @xmath82 the @xmath83-dependence in @xmath45 ( see eq . ( [ ss ] ) ) is killed and we always obtain @xmath84 , i.e. squeezing is impossible . as a consequence , squeezing turns out to be possible only for powers @xmath0 that are a multiple of @xmath85 a necessary condition for the hillery - type squeezing in the fan - state . yet ,
whether or not squeezing appears still depends on the value of @xmath86 to see this let us analytically calculate @xmath45 for a couple of concrete values of @xmath5 and @xmath87 for @xmath88 and @xmath89 we have obtained explicitly @xmath90}{2\xi ^{2}(\sinh ( \xi ^{2})-\sin ( \xi ^{2}))+d_{1}(\xi^2)}\ ] ] with @xmath91\ ] ] and squeezing occurs whenever @xmath92 the function @xmath93 equals zero at @xmath94 and decreases for increasing @xmath95 .
there is no squeezing for @xmath96 for which @xmath97 and no @xmath32 can be found to make @xmath98 negative .
this defines the range of @xmath11 within which squeezing takes place : @xmath99 figure 1 is a 3d plot of @xmath100 as a function of @xmath95 and @xmath32 .
interestingly to notice that a maximal squeezing occurs simultaneously along two directions characterized by @xmath101 and @xmath102 the two coexistent directions of squeezing can alternatively be seen by a polar plot of @xmath100 in fig .
2 with @xmath103 which looks like an eight - winged bow .
the shorter wings correspond to squeezing , while the longer ones to stretching .
to better understand fig .
2 let us also show in fig .
3 the variance @xmath104 itself as a function of @xmath32 for the same parameters as in fig .
2 , i.e. for @xmath105 . taking into account eqs .
( [ r4 ] ) and ( [ s ] ) yields @xmath106 in fig .
3 the variance @xmath107 is represented by the solid curve , while the dashed circle of radius @xmath108 determines the uncertainty region associated with the coherent state for which @xmath109 ( note , this corresponds to the center point in fig .
it is clear that , along the directions @xmath110 and @xmath111 we have @xmath112 signifying negativity of @xmath113 i.e. squeesing takes place . along the directions @xmath114 and @xmath115 however , the inequality @xmath116 holds .
that implies @xmath117 i.e. no squeesing occurs .
since the difference between the solid curve and the dashed circle along @xmath110 ( and @xmath118 ) is smaller than that along @xmath114 ( and @xmath119 ) , the squeezing ( stretching ) gives rise to a shorter ( longer ) wing in fig .
2 . for @xmath120 @xmath121 @xmath122 @xmath123
simultaneous squeezing along the same two directions may also result with appropriate values of @xmath49 but the squeezing degree is much lower than that in the case with @xmath124 for @xmath125 and @xmath126 we have obtained explicitly @xmath127+a(\xi
^{2})+2b(\xi ^{2})-d_{2}(\xi ^{2})},\ ] ] where @xmath128 @xmath129),\ ] ] @xmath130 + 6c(\xi ^{2})+2d_{2}(\xi ^{2})\ ] ] and @xmath131.\ ] ] we display in fig .
4 the polar plot of @xmath132 at @xmath133 @xmath134 @xmath135 and @xmath136 for @xmath137 @xmath138 @xmath139 @xmath123 simultaneous squeezing along the four above - mentioned directions may also result with appropriate values of @xmath49 but the squeezing is much less pronounced compared to the case with @xmath140
in general @xmath4 is a nonlinear function of the boson number of a field .
the concrete form of @xmath4 depends on the physical system to be considered . for example , for the so - called f - oscillator "
@xcite whose frequency is not a constant but depends on the number of the quanta of a field , @xmath4 can be given through @xmath141-deformed algebra @xcite as @xmath142 with @xmath143 a real parametr or , for the so - called photon - added coherent states @xcite one has @xmath144 with the parameter @xmath145 being a positive integer @xcite , etc . in this section , we deal with a realistic situation associated with the phonon field of the center - of - mass vibrational motion of a trapped ion . a laser - driven scheme to produce the fan - state in this context has been presented in @xcite .
in this specific physical system the function @xmath4 and the quantity @xmath49 are determined by @xmath146 @xmath147 where @xmath148 is the @xmath149-th generalized polynomial in @xmath53 for parameter @xmath145 , @xmath150 is the lamb - dicke parameter , @xmath151 with @xmath152 the phase of the driving laser which is resonant with ( detuned to the @xmath8-th red sideband of ) the electronic transition of the trapped ion , and @xmath153 are the rabi frequencies . as followed from eqs .
( [ f ] ) and ( [ xi ] ) , in this realistic system the function @xmath4 as well as the magnitude of @xmath49 can be controlled by adjusting the parameters of the driving lasers ( through @xmath154 , @xmath153 ) and/or the trapping potential ( through @xmath155 ) .
this provides an experimental means to tailor a fan - state on demand . as for
the question which power @xmath0 might give rise to squeezing for a given @xmath156 it is not affected by the concrete form of the function @xmath4 since in fact this question depends merely on the property ( [ j ] ) of the @xmath157 as argued in the preceding section .
this means that independent of @xmath4 the @xmath0-th power amplitude squeezing is possible for @xmath158 @xmath159 @xmath160 however , squeezing actually arises only in certain ranges of @xmath161 and @xmath162 for each chosen value of @xmath155 there exists a critical @xmath163 such that squeezing takes place when @xmath164 .
this is illustrated in figure 5 which plots @xmath165 for @xmath166 and @xmath167 as a function of @xmath168 : @xmath169 .
the number of squeezing directions is determined in the same way as for @xmath9 , i.e. squeezing if it exists is observed simultaneously and equally maximal at @xmath8 directions .
it is worthy to emphasize that the number of simultaneous squeezing directions is dictated only by @xmath5 but not by @xmath170 an interesting fact that can be justified from the symmetry property of the fan - state .
the figure of merit of the case @xmath10 in comparison with @xmath9 is that we can manage the squeezing directions by controlling the physical parameters of the system under consideration .
namely , we could control @xmath150 and @xmath49 so that @xmath171 to have @xmath8 squeezing directions determined by @xmath172 or , if needed , we could chose @xmath150 and @xmath49 so that @xmath173 to make squeezing along the other directions determined by @xmath174 transparently , the squeezing directions determined by eqs .
( [ phi1 ] ) and ( [ phi2 ] ) are rotated " by @xmath175 relative to each other . in other words , by adjusting the system driving parameters we are able to interchange the direction of squeezing and stretching if we like .
we have dealt with the hillery - type @xmath0-power amplitude squeezing in the fan - state @xmath176 eq .
( [ fan ] ) , for both @xmath9 and @xmath177 independent of @xmath4 we find out that for a given @xmath5 squeezing may occur within a certain range of @xmath11 for powers @xmath0 that are a multiple of @xmath178 the remarkable feature in the fan - state is that whenever squeezing exists it arises simultaneously and equally in @xmath8 directions , as opposed to conventional states in which squeezing can be observed only along one direction .
while the number of squeezing directions is fixed to @xmath8 for a given @xmath76 these directions themselves can be rotated by adjusting the physical system parameters when @xmath10 as demonstrated in this work in the ion trap context .
though it can be hoped , it seems premature to see that the multidirectional character of squeezing might find an actual application in practice in general or in quantum information processing in particular .
for example , it is known that for quantum continuous variables teleportation @xcite the necessary resource , i.e. the two - mode squeezed entangled state , is produced by superimposing on a beamsplitter two single - mode squeezed states whose squeezing directions must be perpendicular to each other . yet , as at present it remains unknown what happens and what are the benefits if one or both of the superimposed states would exhibit squeezing in more than one direction .
such kinds of work are exciting and do deserve further efforts . | we study the hillery - type , i.e. @xmath0-th power , amplitude squeezing in the fan - state @xmath1 characterized by @xmath2 , @xmath3 and @xmath4 a nonlinear operator - valued function .
we show that for a given @xmath5 there exists a critical @xmath6 such that for @xmath7 squeezing occurs simultaneously in @xmath8 directions for the powers @xmath0 which are a multiple of @xmath8 .
this result does not depend on the concrete form of @xmath4 , i.e. it holds true for both @xmath9 and @xmath10 .
however , for @xmath10 , which is realized here in the ion trap context , the squeezing directions as well as the magnitude of @xmath11 can be controlled by adjusting the system driving parameters . |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.