article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
frankl s ( union - closed sets ) conjecture is a celebrated unsolved problem in combinatorics that was recently brought to the attention of a wider audience as a polymath project led by timothy gowers @xcite .
a non - empty finite family of finite sets @xmath0 is union - closed ( uc ) if and only if for every @xmath1 it follows that @xmath2 .
frankl s conjecture states that for any uc family @xmath0 there exists an element in at least @xmath3 sets .
the problem appears to have little structure perhaps the very reason why a proof or disproof remains elusive . in this paper
we focus on a well - established method employed to attack the problem referred to as _ local configurations _ in bruhn and schaudt @xcite , namely uc families that imply the conjecture for all uc families which contain them . in other words
, these particular uc families always have an element that is frequent enough to imply the conjecture for all uc families that contain them . in this regard , given an uc family @xmath4 , poonen s theorem @xcite characterizes the necessity of the implication by the existence of weights on the elements of @xmath4 that obey certain inequalities .
define @xmath5:=\left\{1,2,\ldots , n\right\}$ ] .
a family of sets @xmath4 _ covers _
@xmath6 elements if and only if the union of all sets in @xmath4 is @xmath5 $ ] . following vaughan @xcite
, we say that an uc family of sets @xmath4 which covers @xmath6 elements is _ frankl - complete _ ( fc ) , if and only if for every uc family @xmath7 there exists @xmath8 $ ] frequent enough to satisfy frankl s conjecture .
an uc family @xmath4 which covers @xmath6 elements is _ nonfrankl - complete _ ( nonfc ) , if and only if there exists an uc family @xmath7 such that any @xmath8 $ ] is in less than half of the sets of @xmath0 .
nonfc - families are particularly useful in characterizing _
minimal _ fc - families , i.e. , fc - families that do not contain smaller fc - families , and also other objects of interests defined in morris @xcite , which help shed light into structural properties of the conjecture .
in addition , nonfc - families yield natural candidates for possible counterexamples .
however , on a more positive note , the pressing relevance of fc and nonfc - families is evident in existing literature : these objects are at the heart of arguments that yield improved bounds for the problem , as seen in poonen @xcite , gao and yu @xcite , morris @xcite , markovi @xcite , bonjak and markovi @xcite , and finally vukovi and ivkovi @xcite which features the current best bound on the size of the universe of elements at @xmath9 .
furthermore , fc - families are used in bruhn et al . @xcite to prove that frankl s conjecture holds for subcubic bipartite graphs . therefore characterizing a considerable number of previously unknown fc and nonfc - families the fundamental contribution of
this work which consequently helps settle several open questions of interest is a clear step toward a better understanding of frankl s conjecture . characterizing _ exactly _ which uc families are fc and nonfc is suprisingly difficult , as evinced by the relative dearth of known fc - families despite the past twenty - five years of research on the matter . indeed , as is typical of objects in mathematics that are not well - understood , efforts on the topic yield more questions than answers .
previous researchers use special structures and stronger than neccessary conditions to determine a number of fc - families .
in particular , poonen @xcite proves that any uc family which contains three 3-subsets of a 4-set satisfies the conjecture .
vaughan @xcite , @xcite , @xcite proves that the conjecture holds for any uc family which contains a 5-set and all of its 4-subsets , or ten of the 4-subsets of a 6-set and their unions , or three 3-subsets of a 7-set with a common element .
furthermore , using a heuristic precedure implemented in a computer algebra system , vaughan identifies potential weight systems for candidate fc - families and then proves through tedious and technical case analysis that a few more uc families are fc .
still , most fc - families vaughan discovers are not minimal , in the sense that they contain smaller fc - families as shown by subsequent research or results in this paper .
morris @xcite is able to characterize new fc - families on six elements and with the help of a computer program exactly characterizes all minimal fc - families on 5 elements .
given a family of sets @xmath10 , we say that @xmath10 _ generates _ ( or is a _ generator _ of ) @xmath0 , denoted by @xmath11 , if and only if @xmath0 is the smallest uc family that contains @xmath10 . in order to facilitate the combinatorial analysis of fc - families
, morris @xcite introduces the following notion .
let @xmath12 denote the smallest @xmath13 such that any @xmath13 of the k - sets in @xmath5 $ ] generate a fc - family .
as proven in gao and yu @xcite , @xmath12 is always defined for sufficiently large @xmath6 in relation to @xmath14 .
consequently , morris shows that @xmath15 , @xmath16 , @xmath17 , @xmath18 , @xmath19 and @xmath20 .
such characterizations further facilitate the search for better bounds ( or possible counterexamples ) .
finally , mari , ivkovi , and vukovi @xcite formalize a combinatorial search in the interactive theorem prover isabelle / hol and show that all families containing four 3-subsets of a 7-set are fc - families .
although not explicitely mentioned in their paper , their result implies that @xmath21 by the lower bound on the number of 3-sets of morris @xcite . in summary ,
previous research has yielded less than _ two dozen _
exact characterizations of _ minimal _ generators of fc - families , with roughly a dozen more characterizations of general fc - families . in light of the above , our main contributions in this paper
are the following : * in the appendix we feature roughly _ one hundred _
( with many more underway at the time of this writing ) previously unknown minimal non - isomorphic ( under some permutation of @xmath5 $ ] ) generators of fc - families , thus contributing to a significant increase in the state of knowledge , and improving on all previously known results of the kind .
furthermore , we give the _ first _ known exact characterizations of minimal generators of fc - families on @xmath22 .
* we design the _
first general _ computational framework that is able to precisely characterize fc or nonfc - families by using exact integer programming and other redundant verification routines , thus providing an algorithmic road - map for settling open questions in morris @xcite and vaughan @xcite , @xcite . * given that our algorithm determines exactly whether a given uc family @xmath4 if fc or nonfc for @xmath23 , lower bounds for previously unknown @xmath12 in this range become trivial to obtain .
furthermore when coupled with a computer algebra system or graph isomorphism software to obtain the isomorphism types of generators , upper or exact bounds for previously unknown @xmath12 are obtained in the aforementioned range . *
in particular we construct an explicit counterexample to an eleven - year - old conjecture of morris @xcite about the structure of generators for nonfc - families .
furthermore we answer in the negative two related questions of vaughan @xcite and morris @xcite regarding a simplified method for proving the existence of weights that yield fc - families .
we believe these applications best illustrate the reach and potential of our method .
the connection between frankl s conjecture and mathematical programming is well - established in pulaj , raymond and theis @xcite , where the authors derive the equivalence of the problem with an integer program and investigate related conjectures .
furthermore , given an uc family @xmath4 , poonen s theorem yields a constructive proof to determine if @xmath4 is fc or nonfc in the form of a fractional polytope with a potentially exponential number of constraints .
in general , this makes it difficult to explicitly state the conditions which determine whether a given uc family is fc . to overcome this
, we design a cutting - plane method that computes the explicit weights which imply poonen s existence conditions .
in particular , this paves the way toward automated discovery of fc - families by computational integer programming , especially when coupled with an exact rational solver @xcite and other verification routines such as the recent work of cheung , gleixner and steffy @xcite .
our current implementation , we use cplex 12.6.3 @xcite for the seperation oracle , then recheck the results with the rest of the solvers .
in addition , the branch and bound tree of exact scip @xcite is verified with vipr @xcite . ] in scip 3.2.1 @xcite allows us to characterize _ any fc - family up to 10 elements _ tested so far .
as mentioned in section [ sec : introduction ] , poonen s seminal article @xcite precisely characterizes existence conditions for fc - families .
poonen s theorem is at the basis of all subsequent approaches for classifying fc - families , which in turn form an integral part of finding , as mention already , the best current bounds of @xmath9 .
let @xmath24}$ ] denote the power set on @xmath6 elements . for families of sets
@xmath4 and @xmath25 define @xmath26 . finally , for a family of sets @xmath0 define @xmath27 , and denote by @xmath28 the incidence vector of @xmath0 .
[ poonen ] let @xmath4 be an uc family that covers @xmath6 elements .
the following statements are equivalent : 1
. for every uc family @xmath29 , there exists @xmath8 $ ] such that @xmath30 .
+ 2 . there exist non - negative real numbers @xmath31 with @xmath32}c_i = 1 $ ] such that for every uc family @xmath33}$ ] with @xmath34 , the following inequality holds + @xmath35}c_i|\mathcal{b}_i| \geq |\mathcal{b}|/2 .
\label{poon}\ ] ] the proof of the theorem includes a beautiful application of the separating hyperplane theorem and points , at least algorithmically , in the right way .
indeed , for a fixed uc family @xmath4 that covers @xmath6 elements , it is easy to see that the second statement above describes a convex polyhedron @xmath36 defined by : @xmath37}c_i = 1;\\ \\ \sum_{i\in [ n]}c_i|\mathcal{b}_i| \geq |\mathcal{b}|/2 & \mbox { $ \forall \mathcal{b } \subseteq \mathcal{s}_{[n ] } : \mathcal{b } \uplus \mathcal{a } = \mathcal{b}$};\\ \\
c_i \geq 0 & \mbox{$\forall i \in [ n]$};\end{array } \right\}\ ] ] furthermore since the coefficients are all integral , if there exists a feasible point of @xmath38 , we can safely assume ( via fourier - motzkin elimination ) that the @xmath39 are rational .
thus , we can use the simplex or interior point methods to find a feasible point of @xmath38 , or show that one does not exist via farkas lemma .
furthermore let @xmath40 denote the following integer program : @xmath41 } \bar{c}_i & \\ \textrm{s.t . }
& \sum_{i\in [ n]}\bar{c}_i|\mathcal{b}_i| \geq \left(|\mathcal{b}|/2\right ) \sum_{i \in [ n ] } \bar{c}_i & \forall \mathcal{b } \subseteq \mathcal{s}_{[n ] } : \mathcal{b } \uplus \mathcal{a } = \mathcal{b } \\ & \sum_{i \in [ n ] } \bar{c}_i \geq 1 \\ & \bar{c}_i \in \mathbb{z}_{\geq 0 } & \forall i \in [ n]\end{aligned}\ ] ] + [ equivalence ] let @xmath4 be an uc family that covers @xmath6 elements . then there exists a feasible point of @xmath38 if and only if there exists a feasible solution of @xmath40 .
we simply scale .
more precisely , let @xmath42 be a feasible point of @xmath38 .
we can safely assume that @xmath42 is a rational vector , i.e. , @xmath43 .
define @xmath44 , and @xmath45 for all @xmath8 $ ] .
it is easy to see that the defined @xmath46 yield a feasible solution of @xmath40 .
let @xmath47 be a feasible solution of @xmath40 , i.e. , @xmath48 .
define @xmath49 } \bar{c}_i)$ ] .
it is easy to see that the defined @xmath39 yield a feasible point of @xmath38 .
a _ separation oracle _ for a polyhedron @xmath50 is an algorithm that , queried on @xmath51 , either asserts that @xmath52 or returns @xmath53 such that @xmath54 for all @xmath55 . in order to design a separation oracle for @xmath38 ( or @xmath40 ) , we first need the following corollary of poonen s theorem , a version of which is already noted in morris @xcite .
we formalize it again here for clarity and reference .
[ weight ] let @xmath4 be an uc family that covers @xmath6 elements .
the following are equivalent : 1 .
frankl s conjecture holds for every uc family @xmath29 .
in particular , there exists @xmath8 $ ] such that @xmath56 .
2 . for every @xmath57 with @xmath58
, there exist @xmath59 for all @xmath60 $ ] , with @xmath61}c_i = 1 $ ] , such that @xmath62 .
fix an uc family @xmath57 with @xmath58 .
then the following holds , _
s(_i sc_i - _ i sc_i ) & = 2_s_i sc_i -_s(_i sc_i + _ i sc_i ) + & = 2_s_i sc_i -_s_i c_i + & = 2_ic_i|_i| - ||_i c_i 0 + & _ ic_i|_i|||/2 .
+ + since the above holds for every uc family @xmath57 with @xmath58 , the desired result follows from poonen s theorem .
poonen @xcite shows that in order to prove that frankl s conjecture holds for uc families of all cardinalities it is sufficient to consider uc families that contain the empty set .
in the next proposition , we show that for fc or nonfc - families we can always assume ( when convenient ) that the empty set is present .
[ emptynot ] let @xmath4 be an uc family that covers @xmath6 elements such that @xmath63 .
then @xmath4 is a fc - family if and only if @xmath64 is a fc - family .
let @xmath4 be an uc family that covers @xmath6 elements such that @xmath63 .
suppose @xmath4 is a fc - family .
then for all uc families @xmath7 there exists @xmath8 $ ] such that @xmath56 .
hence @xmath65 also satisfies frankl s conjecture .
therefore @xmath64 is a fc - family .
suppose @xmath64 is a fc - family .
as before for all uc families @xmath66 frankl s conjecture is satisfied .
let @xmath67 be the collection of all such families .
then @xmath68 , where @xmath69 and @xmath70 .
define @xmath71 .
then it is clear that @xmath72 , which implies that @xmath4 is a fc - family .
corollary [ weight ] combined with an integer programming approach to uc families inspired by the work of pulaj , raymond and theis @xcite , provides the basis of our separation oracle .
fix an uc family @xmath73}$ ] that covers @xmath6 elements such that @xmath63 , and fix @xmath74 with @xmath61}c_i = 1 $ ] . furthermore , define @xmath44 , and @xmath45 for all @xmath8 $ ] as previously .
let @xmath75 denote the following polyhedron : @xmath76}$};\\ \\
\sum_{s\in \mathcal{s}_{n}}\left(\sum_{i \in s}\bar{c}_i - \sum_{i \notin s}\bar{c}_i \right)x_s + 1 \leq 0;\\ \\
x_t \leq x_u & \mbox { $ \forall s\cup t = u \in \mathcal{s}_{[n ] } : s \in \mathcal{a}$};\\
\\ x_s \in \{0,1\ } & \mbox{$\forall s\in \mathcal{s}_{[n]}$};\end{array } \right\}\ ] ] suppose @xmath75 is non - empty , and let @xmath77
. then @xmath78 where @xmath79}$ ] .
the first inequalities ensure that the chosen family @xmath25 is uc , and we denote them as uc inequalities .
the third class of inequalities ensures that @xmath80 , and we denote them as fixed - set ( fs ) inequalities .
we arrive at @xmath81 , an equivalent condition we find in vaughan @xcite , @xcite , @xcite .
] we explain the @xmath46 inequality in the next proposition .
[ franklintprop ] let @xmath4 be an uc family that covers @xmath6 elements , and let @xmath82 with @xmath61}c_i = 1 $ ] . if @xmath75 is unfeasible , then all uc families @xmath7 satisfy frankl s conjecture .
suppose that @xmath75 is unfeasible .
let @xmath83 be defined as @xmath75 without the @xmath46 inequality .
it is easy to see that @xmath83 is non - empty since the all zero vector is feasible .
suppose @xmath84 is an uc family .
then any @xmath78 with @xmath58 is feasible .
therefore if @xmath75 is unfeasible this implies there exists no uc family @xmath84 with @xmath58 such that : @xmath85 since @xmath86 for all @xmath87}$ ] , and @xmath88 for all @xmath8 $ ] , this implies that for all uc families @xmath84 with @xmath58 , the following inequality holds : @xmath89 it is easy to see that corollary [ weight ] still holds if we replace @xmath39 with @xmath46 .
therefore poonen s theorem implies that all uc families @xmath7 satisfy the frankl conjecture . a natural candidate for checking the feasibility of @xmath75 , for some @xmath4 and @xmath90 ,
is a standard branch and bound algorithm .
since we are mainly interested in proving that certain ( previously unknown ) uc families @xmath4 are fc or nonfc , we do not address questions of complexity , but simply optimize some linear objective function over @xmath75 in a general purpose integer programming solver as specified in section [ sec : introduction ] .
we do the same for optimizing over @xmath40 , and checking the feasibility of @xmath38 . for the remainder of the paper @xmath90 can refer either to a vector of non - negative rationals that equal one as above , or ( equivalently via scaling ) to a vector of non - negative integers which sum up to at least one .
[ oracle ] let @xmath4 be an uc family that covers @xmath6 elements . given @xmath82 with @xmath61}c_i = 1 $ ] as input , determining whether there exists a feasible point @xmath91 of @xmath75 is equivalent to a separation oracle for @xmath38 .
suppose @xmath78 is a feasible point of @xmath75 . as a result
, we arrive at the following valid inequality for @xmath38 ( where we distinguish the variables @xmath92 from the fixed @xmath39 , for all @xmath93 ) @xmath35}\tilde{c}_i|\mathcal{b}_i|\geq |\mathcal{b}|/2 \nonumber,\ ] ] which separates @xmath90 from @xmath38 since the following implications hold _
s(_i s|c_i - _ i s|c_i ) -1 & ||/2 - _ ic_i|_i| > 0 .
+ + otherwise , if @xmath75 is unfeasible , no such valid inequality exists . in this case , proposition [ franklintprop ] implies that @xmath94 .
furthermore , define as @xmath95 the binary program which maximizes the following linear objective function @xmath35}\bar{c}_i \left(\sum_{s \in \mathcal{s}_{[n]}}x_s - 2\sum_{s \in \mathcal{s}_{[n ] } : i \in s}x_s\right ) \nonumber,\ ] ] over @xmath75 .
let @xmath4 be an uc family that covers @xmath6 elements , and let @xmath82 with @xmath61}c_i = 1 $ ] .
an optimal solution of @xmath95 returns a maximally violated inequality for @xmath38 .
it is easy to see that in the following algorithm the tuples @xmath96 and @xmath97 may be used interchangeably with appropriate adjustments .
[ row generation ] @xmath98}c_i = 1 , c_i \geq 0 , \forall i \in [ n ] \right)$ ] , @xmath99 , @xmath100 ; let @xmath4 be an uc family that covers @xmath6 elements .
then algorithm [ row generation ] either outputs weights which prove that @xmath4 is fc , or an infeasible system of constraints which proves that @xmath4 is non - fc .
it is clear algorithm [ row generation ] finitely terminates .
let @xmath4 be an uc family that covers @xmath6 elements .
suppose @xmath4 is a fc - family .
by poonen s theorem there exist @xmath101 for all @xmath8 $ ] with @xmath32}c_i = 1 $ ] that satisfy all inequalities of type [ poon ] .
given @xmath4 , at some iteration of algorithm [ row generation ] , for some @xmath90 , by corollary [ oracle ] we arrive at @xmath102 , otherwise if the algorithm terminates without outputting some @xmath102 , it outputs an infeasible @xmath103 which implies that @xmath4 is not a fc - family and we arrive at a contradiction .
suppose @xmath4 is a nonfc - family .
this implies there exist no @xmath101 for all @xmath8 $ ] with @xmath32}c_i = 1 $ ] that satisfy all inequalities of type [ poon ] . by corollary [ oracle ] during all the iterations of algorithm [ row generation ]
we have that @xmath104 , otherwise we arrive at a contradiction . therefore algorithm [ row generation ] terminates when @xmath105 , which implies that @xmath103 is infeasible .
from the perspective of computational integer programming , valid inequalities are considered effective if they lead to a smaller branch and bound tree .
for all the results that we feature in this paper , adding a subset of the following inequalities to the root node significantly reduces the size of the branch and bound tree . for small @xmath6
this allows for manual inspection , as we illustrate in the appendix .
a family of sets @xmath10 _ generates _
@xmath0 _ with _
@xmath4 , denoted by @xmath106 , if and only if @xmath0 is the smallest uc family that contains @xmath10 such that @xmath107 .
[ fcineq ] let @xmath4 be an uc family that covers @xmath6 elements , and let @xmath82 with @xmath61}c_i = 1 $ ] .
suppose we have @xmath108 , @xmath109 , and @xmath110 .
then the following inequality @xmath111 is valid for @xmath75 .
suppose there exists a feasible solution of @xmath75 which yields an uc family @xmath0 such that the following inequality holds @xmath112 this implies that the number of variables which equal one with positive coefficients is greater than the number of variables with negative coefficients which equal one .
but if either @xmath113 or @xmath114 are one then @xmath115 is one ( if both are one then @xmath116 is one ) and we arrive at a contradiction . in the following definition
the role of a considered uc family @xmath4 is taken into account in the listed conditions . in the first condition
the role of @xmath4 is implicit in the existence of a fs inequality , whereas in the second condition the role of @xmath4 is implicit in _ generating _ the desired family , as discussed at the beginning of this section .
let @xmath4 be an uc family that covers @xmath6 elements , and let @xmath82 with @xmath61}c_i = 1 $ ] .
let @xmath117}$ ] , @xmath118 .
given @xmath119 , we say @xmath120 form a fc - chain which we denote by @xmath121 , if there exist tuples @xmath122 , @xmath123 } $ ] for all @xmath124 , such that for any tuple @xmath125 , at least one of the following conditions holds : 1 .
there exists @xmath126 such that @xmath127 is a valid fs inequality for @xmath75 .
there exists @xmath128 such that @xmath129 is a valid uc inequality for @xmath75 .
we denote an explicit fc - chain by @xmath130 , where @xmath131 satisfy either condition listed above . when needed we specify which type of inequalities form a fc - chain by @xmath132 for uc inequalities , and @xmath133 for fs inequalites .
[ fcineq ] let @xmath4 be an uc family that covers @xmath6 elements , and let @xmath82 with @xmath61}c_i = 1 $ ] .
let @xmath117}$ ] , @xmath134 .
for any @xmath135 define @xmath136 .
suppose that @xmath137 for all @xmath135 .
then the following inequality @xmath138 is valid for @xmath75 .
suppose there exists a feasible solution of @xmath75 which yields an uc family @xmath0 such that the following inequality holds @xmath139 it is clear that @xmath140 , otherwise we arrive at a contradiction .
therefore the inequality implies that the number of variables @xmath141 which equal one , for all @xmath142 is greater than the number of variables @xmath141 which equal one , for all @xmath143 .
let @xmath144 , and for all @xmath145 , let @xmath146
. then @xmath137 holds by hypothesis .
furthermore by the definition of a fc - chain , for all @xmath147 we conclude that @xmath148 .
thus we arrive at a contradiction .
in this section we exhibit a counterexample to a conjecture of morris @xcite about generators for non - fc - families .
a generator @xmath10 for a uc family @xmath0 is _ minimal _ if and only if there exists no @xmath149 such that @xmath150 is a generator for @xmath0 . in the following definition , it is sufficient to consider @xmath151 $ ] . let @xmath10 be a family of sets that covers @xmath6 elements .
suppose @xmath10 is a minimal generator for an uc family @xmath0 , such that @xmath0 is a nonfc - family .
then @xmath10 is _ regular _ if and only if for any @xmath152 , @xmath153 , the uc family @xmath154 is nonfc . [ conjecture ] @xmath10 is regular for all @xmath155 .
morris @xcite checked the conjecture for all known families at the time , and therefore considered it plausible . in some sense , conjecture [ conjecture ] perfectly illustrates our general lack of knowledge about uc families since as a number of other related questions
it has eluded an answer for a relatively long time .
the obstacle in this case and others to follow is the lack of a method for exactly characterizing fc - families , a gap in knowlege which we correct with our framework .
this highlights the importance of appropriate computational techniques in areas where such applications are rarely encountered , but are otherwise quite effective tools .
our counterexample on six elements is minimal , in the sense that morris @xcite completely characterizes fc - families on 5 elements .
let @xmath156 .
furthermore , let @xmath157 .
hence it follows that @xmath158 .
it is easy to see that @xmath10 is a minimal generator for @xmath159 .
we will show that @xmath160 is a nonfc - family .
there is a stronger connection between the structure of inequalities featured in the proof below and questions of vaughan @xcite and morris @xcite we answer later in this work . in section [ sec : relaxation ]
we explicitely describe the structure of uc families from which the inequalities below are derived in relation to the questions of interest .
@xmath159 is a nonfc - family .
the proof is the output of algorithm [ row generation ] with @xmath159 as input , which is an infeasible system of constraints .
we display an irreducible infeasible subset of the given system .
we identify columns with zero one entries for each @xmath161 .
the six matrices featured below represent uc families .
the top row keeps track of the number of sets in each family .
in addition to rechecking with an exact rational solver @xcite and other solvers , we check that each matrix is uc via simple external subroutines and finally by hand .. ] furthermore , let @xmath0 be a family represented by one of the matrices below .
it easy to see that @xmath162 by inspection . in each matrix , we color columns which correspond to sets in @xmath10 , @xmath163 , red and blue , respectively .
each matrix yields an inequality of type [ poon ] ( multiplied by two ) featured below it .
the following system of constraints is infeasible in non - negative @xmath39 for all @xmath164 .
for each row we display the farkas dual values in square brackets .
this yields a certificate of infeasibility via a straightforward application of farkas lemma . for convenience
we state the lemma in the appendix .
+ + @xmath165:c_1 + c_2 + c_3 + c_4 + c_5 + c_6 = 1}$ ] .
+ @xmath166 \matrix ( m)[matrix of math nodes ] { \color{red}{0 } & \color{red}{1 } & \color{red}{2 } & \color{red}{3 } & \color{red}{4 } & \color{red}{5 } & \color{red}{6 } & \color{red}{7 } & \color{red}{8 } & \color{red}{9 } & \color{red}{10 } & \color{red}{11 } & \color{red}{12 } & \color{red}{13 } & \color{red}{14 } & \color{red}{15 } & \color{red}{16 } & \color{red}{17 } & \color{red}{18 } & \color{red}{19 } & \color{red}{20 } & \color{red}{21 } & \color{red}{22 } & \color{red}{23 } & \color{red}{24 } & \color{red}{25 } & \color{red}{26 } & \color{red}{27 } & \color{red}{28 } & \color{red}{29 } & \color{red}{30 } & \color{red}{31 } & \color{red}{32 } & \color{red}{33 } & \color{red}{34 } & \color{red}{35 } & \color{red}{36 } & \color{red}{37 } & \color{red}{38 } & \color{red}{39 } & \color{red}{40 } & \color{red}{41 } & \color{red}{42 } & \color{red}{43 }
\\ c_1 & \color{red}{0 } & 0 & 0 & 0 & 0 & 0 & 0 & \color{red}{0 } & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & \color{red}{1 } & 1 & 1 & \color{blue}{1 } & \color{red}{1 } & \color{blue}{1 } & 1 & \color{red}{1 } & 1 & 1 & \color{blue}{1}\\ c_2 & \color{red}{0 } & 0 & 0 & 0 & 0 & 0 & 0 & \color{red}{0 } & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & \color{red}{0 } & 0 & 0 & \color{blue}{0 } & \color{red}{1 } & \color{blue}{1 } & 1 & \color{red}{1 } & 1 & 1 & \color{blue}{1}\\ c_3 & \color{red}{0 } & 0 & 0 & 0 & 0 & 0 & 0 & \color{red}{0 } & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & \color{red}{1 } & 1 & 1 & \color{blue}{1 } & \color{red}{0 } & \color{blue}{0 } & 1 & \color{red}{1 } & 1 & 1 & \color{blue}{1}\\ c_4 & \color{red}{0 } & 0 & 0 & 0 & 1 & 1 & 1 & \color{red}{1 } & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & \color{red}{1 } & 1 & 1 & \color{blue}{1 } & \color{red}{0 } & \color{blue}{1 } & 0 & \color{red}{1 } & 1 & 1 & \color{blue}{1}\\ c_5 & \color{red}{0 } & 0 & 1 & 1 & 0 & 0 & 1 & \color{red}{1 } & 0 & 0 & 1 & 1 & 0 & 0 & 1 & 1 & 0 & 0 & 1 & 1 & 0 & 0 & 1 & 1 & 0 & 0 & 1 & 1 & 0 & 0 & 1 & 1 & \color{red}{0 } & 0 & 1 & \color{blue}{1 } & \color{red}{1 } & \color{blue}{1 } & 1 & \color{red}{0 } & 0 & 1 & \color{blue}{1}\\ c_6 & \color{red}{0 } & 1 & 0 & 1 & 0 & 1 & 0 & \color{red}{1 } & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & \color{red}{0 } & 1 & 0 & \color{blue}{1 } & \color{red}{1 } & \color{blue}{1 } & 1 & \color{red}{0 } & 1 & 0 & \color{blue}{1}\\ } ; \end{tikzpicture}\ ] ] + @xmath167:22c_1 + 46c_2 + 50c_3 + 50c_4 + 46c_5 + 46c_6 \geq 43}$ ] .
+ + @xmath168 \matrix ( m)[matrix of math nodes ] { \color{red}{0 } & \color{red}{1 } & \color{red}{2 } & \color{red}{3 } & \color{red}{4 } & \color{red}{5 } & \color{red}{6 } & \color{red}{7 } & \color{red}{8 } & \color{red}{9 } & \color{red}{10 } & \color{red}{11 } & \color{red}{12 } & \color{red}{13 } & \color{red}{14 } & \color{red}{15 } & \color{red}{16 } & \color{red}{17 } & \color{red}{18 } & \color{red}{19 } & \color{red}{20 } & \color{red}{21 } & \color{red}{22 } & \color{red}{23 } & \color{red}{24 } & \color{red}{25 } & \color{red}{26 } & \color{red}{27 } & \color{red}{28 } & \color{red}{29 } & \color{red}{30 } & \color{red}{31 } & \color{red}{32 } & \color{red}{33 } & \color{red}{34 } & \color{red}{35 } & \color{red}{36 } & \color{red}{37 } & \color{red}{38 } & \color{red}{39 } \\
c_1 & \color{red}{0 } & 0 & 0 & 0 & 0 & 0 & 0 & \color{red}{0 } & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & \color{red}{1 } & 1 & 1 & \color{blue}{1 } & \color{red}{1 } & \color{blue}{1 } & 1 & \color{red}{1 } & 1 & 1 & \color{blue}{1}\\ c_2 & \color{red}{0 } & 0 & 0 & 0 & 0 & 0 & 0 & \color{red}{0 } & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & \color{red}{0 } & 0 & 0 & \color{blue}{0 } & \color{red}{1 } & \color{blue}{1 } & 1 & \color{red}{1 } & 1 & 1 & \color{blue}{1}\\ c_3 & \color{red}{0 } & 0 & 0 & 0 & 0 & 0 & 0 & \color{red}{0 } & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & \color{red}{1 } & 1 & 1 & \color{blue}{1 } & \color{red}{0 } & \color{blue}{0 } & 1 & \color{red}{1 } & 1 & 1 & \color{blue}{1}\\ c_4 & \color{red}{0 } & 0 & 0 & 0 & 1 & 1 & 1 & \color{red}{1 } & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & 0 & 0 & 0 & 0 & \color{red}{1 } & 1 & 1 & \color{blue}{1 } & \color{red}{0 } & \color{blue}{1 } & 0 & \color{red}{1 } & 1 & 1 & \color{blue}{1}\\ c_5 & \color{red}{0 } & 0 & 1 & 1 & 0 & 0 & 1 & \color{red}{1 } & 0 & 0 & 1 & 1 & 0 & 0 & 1 & 1 & 0 & 0 & 1 & 1 & 0 & 0 & 1 & 1 & 0 & 0 & 1 & 1 & \color{red}{0 } & 0 & 1 & \color{blue}{1 } & \color{red}{1 } & \color{blue}{1 } & 1 & \color{red}{0 } & 0 & 1 & \color{blue}{1}\\ c_6 & \color{red}{0 } & 1 & 0 & 1 & 0 & 1 & 0 & \color{red}{1 } & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & \color{red}{0 } & 1 & 0 & \color{blue}{1 } & \color{red}{1 } & \color{blue}{1 } & 1 & \color{red}{0 } & 1 & 0 & \color{blue}{1}\\ } ; \end{tikzpicture}\ ] ] + + @xmath169:46c_1 + 14c_2 + 42c_3 + 42c_4 + 42c_5 + 42c_6 \geq 39}$ ] .
+ + @xmath170 \matrix ( m)[matrix of math nodes ] { \color{red}{0 } & \color{red}{1 } & \color{red}{2 } & \color{red}{3 } & \color{red}{4 } & \color{red}{5 } & \color{red}{6 } & \color{red}{7 } & \color{red}{8 } & \color{red}{9 } & \color{red}{10 } & \color{red}{11 } & \color{red}{12 } & \color{red}{13 } & \color{red}{14 } & \color{red}{15 } & \color{red}{16 } & \color{red}{17 } & \color{red}{18 } & \color{red}{19 } & \color{red}{20 } & \color{red}{21 } & \color{red}{22 } & \color{red}{23 } & \color{red}{24 } & \color{red}{25 } & \color{red}{26 } & \color{red}{27 } & \color{red}{28 } & \color{red}{29 } & \color{red}{30 } & \color{red}{31 } & \color{red}{32 } & \color{red}{33 } & \color{red}{34 } & \color{red}{35 } & \color{red}{36 } & \color{red}{37 } & \color{red}{38 } & \color{red}{39 } & \color{red}{40 } & \color{red}{41 } & \color{red}{42 } & \color{red}{43 } & \color{red}{44 } & \color{red}{45 } & \color{red}{46 } \\ c_1 & \color{red}{0 } & 0 & 0 & 0 & \color{red}{0 } & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & \color{red}{1 } & 1 & 1 & \color{blue}{1 } & 1 & 1 & 1 & \color{red}{1 } & \color{blue}{1 } & 1 & 1 & 1 & 1 & \color{red}{1 } & 1 & 1 & \color{blue}{1}\\ c_2 & \color{red}{0 } & 0 & 0 & 0 & \color{red}{0 } & 0 & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & \color{red}{0 } & 0 & 0 & \color{blue}{0 } & 1 & 1 & 1 & \color{red}{1 } & \color{blue}{1 } & 1 & 1 & 1 & 1 & \color{red}{1 } & 1 & 1 & \color{blue}{1}\\ c_3 & \color{red}{0 } & 0 & 0 & 0 & \color{red}{0 } & 1 & 1 & 1 & 1 & 1 & 0 & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & 1 & 0 & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & \color{red}{1 } & 1 & 1 & \color{blue}{1 } & 0 & 0 & 0 & \color{red}{0 } & \color{blue}{0 } & 1 & 1 & 1 & 1 & \color{red}{1 } & 1 & 1 & \color{blue}{1}\\ c_4 & \color{red}{0 } & 0 & 0 & 0 & \color{red}{1 } & 0 & 0 & 0 & 0 & 1 & 0 & 0 & 0 & 0 & 1 & 0 & 0 & 0 & 0 & 1 & 0 & 0 & 0 & 0 & 1 & 0 & 0 & 0 & 0 & \color{red}{1 } & 1 & 1 & \color{blue}{1 } & 0 & 0 & 0 & \color{red}{0 } & \color{blue}{1 } & 0 & 0 & 0 & 0 & \color{red}{1 } & 1 & 1 & \color{blue}{1}\\ c_5 & \color{red}{0 } & 0 & 1 & 1 & \color{red}{1 } & 0 & 0 & 1 & 1 & 1 & 0 & 0 & 1 & 1 & 1 & 0 & 0 & 1 & 1 & 1 & 0 & 0 & 1 & 1 & 1 & 0 & 0 & 1 & 1 & \color{red}{0 } & 0 & 1 & \color{blue}{1 } & 0 & 0 & 1 & \color{red}{1 } & \color{blue}{1 } & 0 & 0 & 1 & 1 & \color{red}{0 } & 0 & 1 & \color{blue}{1}\\ c_6 & \color{red}{0 } & 1 & 0 & 1 & \color{red}{1 } & 0 & 1 & 0 & 1 & 1 & 0 & 1 & 0 & 1 & 1 & 0 & 1 & 0 & 1 & 1 & 0 & 1 & 0 & 1 & 1 & 0 & 1 & 0 & 1 & \color{red}{0 } & 1 & 0 & \color{blue}{1 } & 0 & 1 & 0 & \color{red}{1 } & \color{blue}{1 } & 0 & 1 & 0 & 1 & \color{red}{0 } & 1 & 0 & \color{blue}{1}\\ } ; \end{tikzpicture}\ ] ] + + @xmath171:52c_1 + 46c_2 + 52c_3 + 28c_4 + 52c_5 + 52c_6 \geq 46}$ ] .
+ + @xmath172 \matrix ( m)[matrix of math nodes ] { \color{red}{0 } & \color{red}{1 } & \color{red}{2 } & \color{red}{3 } & \color{red}{4 } & \color{red}{5 } & \color{red}{6 } & \color{red}{7 } & \color{red}{8 } & \color{red}{9 } & \color{red}{10 } & \color{red}{11 } & \color{red}{12 } & \color{red}{13 } & \color{red}{14 } & \color{red}{15 } & \color{red}{16 } & \color{red}{17 } & \color{red}{18 } & \color{red}{19 } & \color{red}{20 } & \color{red}{21 } & \color{red}{22 } & \color{red}{23 } & \color{red}{24 } & \color{red}{25 } & \color{red}{26 } & \color{red}{27 } & \color{red}{28 } & \color{red}{29 } & \color{red}{30 } & \color{red}{31 } & \color{red}{32 } & \color{red}{33 } & \color{red}{34 } & \color{red}{35 } & \color{red}{36 } & \color{red}{37 } & \color{red}{38 } & \color{red}{39 } & \color{red}{40 } \\ c_1 & \color{red}{0 } & 0 & 0 & 0 & 0 & 0 & 0 & \color{red}{0 } & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & \color{red}{1 } & 1 & 1 & \color{blue}{1 } & 1 & 1 & 1 & \color{red}{1 } & 1 & 1 & 1 & \color{blue}{1 } & \color{red}{1 } & 1 & 1 & \color{blue}{1}\\ c_2 & \color{red}{0 } & 0 & 0 & 0 & 0 & 0 & 0 & \color{red}{0 } & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & \color{red}{0 } & 0 & 0 & \color{blue}{0 } & 1 & 1 & 1 & \color{red}{1 } & 1 & 1 & 1 & \color{blue}{1 } & \color{red}{1 } & 1 & 1 & \color{blue}{1}\\ c_3 & \color{red}{0 } & 0 & 0 & 0 & 0 & 0 & 0 & \color{red}{0 } & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & \color{red}{1 } & 1 & 1 & \color{blue}{1 } & 0 & 0 & 0 & \color{red}{0 } & 0 & 0 & 0 & \color{blue}{0 } & \color{red}{1 } & 1 & 1 & \color{blue}{1}\\ c_4 & \color{red}{0 } & 0 & 0 & 0 & 1 & 1 & 1 & \color{red}{1 } & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & \color{red}{1 } & 1 & 1 & \color{blue}{1 } & 0 & 0 & 0 & \color{red}{0 } & 1 & 1 & 1 & \color{blue}{1 } & \color{red}{1 } & 1 & 1 & \color{blue}{1}\\ c_5 & \color{red}{0 } & 0 & 1 & 1 & 0 & 0 & 1 & \color{red}{1 } & 0 & 0 & 1 & 1 & 0 & 0 & 1 & 1 & 0 & 0 & 1 & 1 & 0 & 0 & 1 & 1 & \color{red}{0 } & 0 & 1 & \color{blue}{1 } & 0 & 0 & 1 & \color{red}{1 } & 0 & 0 & 1 & \color{blue}{1 } & \color{red}{0 } & 0 & 1 & \color{blue}{1}\\ c_6 & \color{red}{0 } & 1 & 0 & 1 & 0 & 1 & 0 & \color{red}{1 } & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & \color{red}{0 } & 1 & 0 & \color{blue}{1 } & 0 & 1 & 0 & \color{red}{1 } & 0 & 1 & 0 & \color{blue}{1 } & \color{red}{0 } & 1 & 0 & \color{blue}{1}\\ } ; \end{tikzpicture}\ ] ] + + @xmath173:48c_1 + 40c_2 + 16c_3 + 48c_4 + 40c_5 + 40c_6 \geq 40}$ ]
. + + @xmath174 \matrix ( m)[matrix of math nodes ] { \color{red}{0 } & \color{red}{1 } & \color{red}{2 } & \color{red}{3 } & \color{red}{4 } & \color{red}{5 } & \color{red}{6 } & \color{red}{7 } & \color{red}{8 } & \color{red}{9 } & \color{red}{10 } & \color{red}{11 } & \color{red}{12 } & \color{red}{13 } & \color{red}{14 } & \color{red}{15 } & \color{red}{16 } & \color{red}{17 } & \color{red}{18 } & \color{red}{19 } & \color{red}{20 } & \color{red}{21 } & \color{red}{22 } & \color{red}{23 } & \color{red}{24 } & \color{red}{25 } & \color{red}{26 } & \color{red}{27 } & \color{red}{28 } & \color{red}{29 } & \color{red}{30 } & \color{red}{31 } & \color{red}{32 } & \color{red}{33 } & \color{red}{34 } & \color{red}{35 } & \color{red}{36 } & \color{red}{37 } & \color{red}{38 } & \color{red}{39 } & \color{red}{40 } & \color{red}{41 } & \color{red}{42 } \\ c_1 & \color{red}{0 } & 0 & 0 & 0 & \color{red}{0 } & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & \color{red}{1 } & 1 & \color{blue}{1 } & 1 & 1 & \color{red}{1 } & 1 & 1 & \color{blue}{1 } & 1 & 1 & 1 & \color{red}{1 } & 1 & \color{blue}{1}\\ c_2 & \color{red}{0 } & 0 & 0 & 0 & \color{red}{0 } & 0 & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & \color{red}{0 } & 0 & \color{blue}{0 } & 1 & 1 & \color{red}{1 } & 1 & 1 & \color{blue}{1 } & 1 & 1 & 1 & \color{red}{1 } & 1 & \color{blue}{1}\\ c_3 & \color{red}{0 } & 0 & 0 & 0 & \color{red}{0 } & 1 & 1 & 1 & 1 & 1 & 0 & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & 1 & 0 & 0 & 0 & 0 & 0 & 1 & 1 & \color{red}{1 } & 1 & \color{blue}{1 } & 0 & 0 & \color{red}{0 } & 0 & 0 & \color{blue}{0 } & 1 & 1 & 1 & \color{red}{1 } & 1 & \color{blue}{1}\\ c_4 & \color{red}{0 } & 0 & 1 & 1 & \color{red}{1 } & 0 & 0 & 1 & 1 & 1 & 0 & 0 & 1 & 1 & 1 & 0 & 0 & 1 & 1 & 1 & 0 & 0 & 1 & 1 & 1 & 0 & 0 & \color{red}{1 } & 1 & \color{blue}{1 } & 0 & 0 & \color{red}{0 } & 1 & 1 & \color{blue}{1 } & 0 & 0 & 0 & \color{red}{1 } & 1 & \color{blue}{1}\\ c_5 & \color{red}{0 } & 0 & 0 & 0 & \color{red}{1 } & 0 & 0 & 0 & 0 & 1 & 0 & 0 & 0 & 0 & 1 & 0 & 0 & 0 & 0 & 1 & 0 & 0 & 0 & 0 & 1 & 0 & 0 & \color{red}{0 } & 0 & \color{blue}{1 } & 0 & 0 & \color{red}{1 } & 0 & 0 & \color{blue}{1 } & 0 & 0 & 1 & \color{red}{0 } & 0 & \color{blue}{1}\\
c_6 & \color{red}{0 } & 1 & 0 & 1 & \color{red}{1 } & 0 & 1 & 0 & 1 & 1 & 0 & 1 & 0 & 1 & 1 & 0 & 1 & 0 & 1 & 1 & 0 & 1 & 0 & 1 & 1 & 0 & 1 & \color{red}{0 } & 1 & \color{blue}{1 } & 0 & 1 & \color{red}{1 } & 0 & 1 & \color{blue}{1 } & 0 & 1 & 1 & \color{red}{0 } & 1 & \color{blue}{1}\\ } ; \end{tikzpicture}\ ] ] + + @xmath175:44c_1 + 44c_2 + 42c_3 + 48c_4 + 20c_5 + 52c_6 \geq 42}$ ] . + + @xmath174 \matrix ( m)[matrix of math nodes ] { \color{red}{0 } & \color{red}{1 } & \color{red}{2 } & \color{red}{3 } & \color{red}{4 } & \color{red}{5 } & \color{red}{6 } & \color{red}{7 } & \color{red}{8 } & \color{red}{9 } & \color{red}{10 } & \color{red}{11 } & \color{red}{12 } & \color{red}{13 } & \color{red}{14 } & \color{red}{15 } & \color{red}{16 } & \color{red}{17 } & \color{red}{18 } & \color{red}{19 } & \color{red}{20 } & \color{red}{21 } & \color{red}{22 } & \color{red}{23 } & \color{red}{24 } & \color{red}{25 } & \color{red}{26 } & \color{red}{27 } & \color{red}{28 } & \color{red}{29 } & \color{red}{30 } & \color{red}{31 } & \color{red}{32 } & \color{red}{33 } & \color{red}{34 } & \color{red}{35 } & \color{red}{36 } & \color{red}{37 } & \color{red}{38 } & \color{red}{39 } & \color{red}{40 } & \color{red}{41 } & \color{red}{42 } \\ c_1 & \color{red}{0 } & 0 & 0 & 0 & \color{red}{0 } & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & \color{red}{1 } & 1 & \color{blue}{1 } & 1 & 1 & \color{red}{1 } & 1 & 1 & \color{blue}{1 } & 1 & 1 & 1 & \color{red}{1 } & 1 & \color{blue}{1}\\ c_2 & \color{red}{0 } & 0 & 0 & 0 & \color{red}{0 } & 0 & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & \color{red}{0 } & 0 & \color{blue}{0 } & 1 & 1 & \color{red}{1 } & 1 & 1 & \color{blue}{1 } & 1 & 1 & 1 & \color{red}{1 } & 1 & \color{blue}{1}\\ c_3 & \color{red}{0 } & 0 & 0 & 0 & \color{red}{0 } & 1 & 1 & 1 & 1 & 1 & 0 & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & 1 & 0 & 0 & 0 & 0 & 0 & 1 & 1 & \color{red}{1 } & 1 & \color{blue}{1 } & 0 & 0 & \color{red}{0 } & 0 & 0 & \color{blue}{0 } & 1 & 1 & 1 & \color{red}{1 } & 1 & \color{blue}{1}\\ c_4 & \color{red}{0 } & 0 & 1 & 1 & \color{red}{1 } & 0 & 0 & 1 & 1 & 1 & 0 & 0 & 1 & 1 & 1 & 0 & 0 & 1 & 1 & 1 & 0 & 0 & 1 & 1 & 1 & 0 & 0 & \color{red}{1 } & 1 & \color{blue}{1 } & 0 & 0 & \color{red}{0 } & 1 & 1 & \color{blue}{1 } & 0 & 0 & 0 & \color{red}{1 } & 1 & \color{blue}{1}\\ c_5 & \color{red}{0 } & 1 & 0 & 1 & \color{red}{1 } & 0 & 1 & 0 & 1 & 1 & 0 & 1 & 0 & 1 & 1 & 0 & 1 & 0 & 1 & 1 & 0 & 1 & 0 & 1 & 1 & 0 & 1 & \color{red}{0 } & 1 & \color{blue}{1 } & 0 & 1 & \color{red}{1 } & 0 & 1 & \color{blue}{1 } & 0 & 1 & 1 & \color{red}{0 } & 1 & \color{blue}{1}\\ c_6 & \color{red}{0 } & 0 & 0 & 0 & \color{red}{1 } & 0 & 0 & 0 & 0 & 1 & 0 & 0 & 0 & 0 & 1 & 0 & 0 & 0 & 0 & 1 & 0 & 0 & 0 & 0 & 1 & 0 & 0 & \color{red}{0 } & 0 & \color{blue}{1 } & 0 & 0 & \color{red}{1 } & 0 & 0 & \color{blue}{1 } & 0 & 0 & 1 & \color{red}{0 } & 0 & \color{blue}{1}\\ } ; \end{tikzpicture}\ ] ] + + @xmath175:44c_1 + 44c_2 + 42c_3 + 48c_4 + 52c_5 + 20c_6 \geq 42}$ ] .
+ [ morriscounter ] let @xmath176 .
+ then @xmath177 is a fc - family .
. then @xmath179 is infeasible . by making use of fc - chain inequalities and displaying irreducible infeasible subsets of constraints for the two leaf nodes of the resulting branch and bound tree . ]
@xmath10 is a counterexample to conjecture [ conjecture ] .
@xmath180 .
we exhibit a few more new fc - families verified by our algorithm .
as mentioned earlier , the computations are rechecked with a number of solvers followed by a final check with an exact rational solver and vipr @xcite .
furthermore , for all the families shown here , by adding fc - chain inequalities at the root node , it is possible to reduce the resulting branch and bound tree to a few nodes and check some irreducible infeasible subset of each leaf node ( linear program ) manually .
let @xmath181 .
+ then @xmath160 is a fc - family .
. then @xmath183 is infeasible .
let @xmath184 .
+ then @xmath160 is a fc - family .
. then @xmath183 is infeasible .
in this section , we briefly address the practical behaviour of our algorithm with a focus on the number of calls made to our separation oracle as it sheds light on open questions of interests in vaughan @xcite and morris @xcite . as a result
, we exhibit a counterexample to both closely related questions .
although throughout this work we mostly refer to @xmath97 to keep the exposition clear , our current implementation features @xmath96 . in doing
so we avoid possible numerical trouble by minimizing the sum of the @xmath39 , in addition to selecting the
sharpest cut " whenever the separation oracle is called . yet , without witnessing first - hand computations for fixed uc families @xmath4 that covers @xmath23 , our separation oracle and algorithm appear fraught with theoretical dangers . , and furthermore the number of calls it receives could be exponential .
] however , in practise our method is well - behaved in the described range , and is consequently the currently best available technique for the _ exact _ determinination of fc - families .
furthermore , our implementation _ mostly _ confirms the heuristic intuition of vaughan and morris as will be made explicit in the next paragraphs .
thus in the tested range , our separation oracle is mostly called @xmath6 times . however , in some cases the oracle is called more than @xmath6 ( but less than @xmath186 ) times and a few minutes for @xmath187 , to a few hours for @xmath188 . furthermore verification with exact scip @xcite takes longer , as does testing a non - minimal fc - family .
computations were carried out on machines with 2.40 ghz quad - core processors and 16 gb of ram . ] . among the
later we find counterexamples to open questions of interest which we feature below .
as mentioned in section [ sec : introduction ] , vaughan @xcite implements a heuristic that guides the search for a feasible weight system .
given an uc family @xmath4 , @xmath63 , the heuristic focuses only on uc families @xmath25 with @xmath189 , where @xmath190 \setminus \left\{j\right\ } } \uplus \mathcal{a}$ ] for all @xmath191 $ ] .
if there exists a solution to the system of linear equations @xmath61}c_i|\mathcal{b}_i| = |\mathcal{b}|/2 $ ] in non - negative @xmath39 , with @xmath61}c_i \leq 1 $ ] , then the considered uc family @xmath4 becomes a candidate fc - family .
all of vaughan s candidate fc - families in @xcite are identified as above , followed by tedious case analysis that spans several pages for the proof that the given family is fc .
we precisely state vaughan s question as follows : let @xmath4 , @xmath63 be an uc family that covers @xmath6 elements , and consider uc families @xmath190 \setminus \left\{j\right\ } } \uplus \mathcal{a}$ ] for all @xmath191 $ ] .
suppose the linear system of equations @xmath61}c_i|\mathcal{b}_i|
= |\mathcal{b}|/2 $ ] for all @xmath25 as above has a solution in non - negative @xmath39 for all @xmath8 $ ] , such that @xmath61}c_i \leq 1 $ ] .
denote this by @xmath192 .
does @xmath192 imply that @xmath193 is feasible ?
given an uc family @xmath4 , @xmath63 , morris @xcite also focuses on @xmath25 as above , searching instead for feasible non - negative points of the polytope defined by the @xmath6 given @xmath25 and @xmath61}c_i \geq 1 $ ] when the desired points are integral , or @xmath61}c_i = 1 $ ] when the desired points are rational .
the idea is that the @xmath6 given inequalities could serve as a relaxation of the underlying polytope of @xmath40 ( or equivalently @xmath38 ) .
morris shows that this holds in a number of cases , but is it true in general ?
more precisely , we state it as the following question : let @xmath4 , @xmath63 be an uc family that covers @xmath6 elements , and consider uc families @xmath190 \setminus \left\{j\right\ } } \uplus \mathcal{a}$ ] for all @xmath191 $ ] .
suppose the polytope @xmath194 defined by @xmath61}\bar{c}_i \geq 1 $ ] , @xmath195 for all @xmath25 as above , and @xmath196 for all @xmath8 $ ] , is non - empty .
does this imply @xmath40 is feasible ? a positive answer to vaughan s question implies a positive answer to morris question .
thus , considering the above , we can explicitly describe the structure associated with the nonfc - family that leads to the counterexample in section [ conjecture ] . as above , it sufficies to consider @xmath190 \setminus \left\{j\right\ } } \uplus \mathcal{a}$ ] for all @xmath191 $ ] , where @xmath4 is our given uc family .
this greatly simplifies the tedious task of checking that the algorithm s output is correct .
once the family is constructed according the given @xmath25 , it is easy to see that the necessary conditions for correctness are met . given that the empty set does not make a difference in determining whether an uc family @xmath4 is fc or nonfc , as we saw in proposition [ emptynot ] , we may think the condition @xmath63 in the questions of vaughan and morris can be relaxed .
if this were the case , the structure of the considered @xmath25 with @xmath197 is again simplified , since the cardinality of the new family is at most the cardinality of the original one .
unfortunately , as we shall see , this is not the case .
still , in the next proposition , we show that a feasible @xmath194 implies the feasibility of a polytope arising from smaller structures .
[ smallerstruct ] let @xmath4 be an uc family that covers @xmath6 elements such that @xmath63 .
suppose @xmath194 is non - empty
. then the analogous polytope defined by the constraints derived from @xmath198 \setminus \left\{j\right\ } } \uplus \mathcal{a}\right\ } \setminus s_{[n ] \setminus \left\{j\right\}}$ ] , for all @xmath191 $ ] , is non - empty .
let @xmath4 be an uc family that covers @xmath6 elements such that @xmath63 .
furthermore , let @xmath190 \setminus \left\{1\right\ } } \uplus \mathcal{a}$ ] .
since @xmath63 , it follows that @xmath199 \setminus \left\{1\right\ } } \subset \mathcal{b}$ ] .
define @xmath200 \setminus \left\{1\right\}}$ ] , @xmath201 .
suppose that @xmath194 is non - empty .
from observation [ equivalence ] we focus our arguments on the rational equivalent of @xmath194 from poonen s theorem .
then the following holds , _
|| & _ i 2c_i|_i| + _ i\{1 } 2c_i|_i| || + || + & _ i 2c_i|_i| || .
+ + in the last implication we use @xmath202 \setminus \left\{1\right\ } } c_i \leq 1 $ ] , with @xmath101 for all @xmath8 \setminus \left\{1\right\}$ ] and therefore _ i\{1 } 2c_i|_i| & = ||_i\{1}c_i || .
+ + since the same argument applies to @xmath190 \setminus \left\{j\right\ } } \uplus \mathcal{a}$ ] for all @xmath191 $ ] , the desired result follows .
as we shall see next , the implication does not hold the other way around .
let @xmath203 .
let @xmath204 .
morris @xcite proves that @xmath205 is empty .
let @xmath4 be an uc family that covers @xmath6 elements .
a feasible @xmath206 does not imply a feasible @xmath194 .
we show that @xmath207 is non - empty .
it is easy to see this without explicitely writing down the inequalities .
observe that if we write each set in @xmath208 as a column of a @xmath209 binary matrix @xmath210 , we have more entries with ones than zeros . we conclude similarly for all @xmath191 $ ] such that @xmath190 \setminus \left\{j\right\ } } \uplus \mathcal{a}$ ] .
hence , any vector @xmath196 for all @xmath8 $ ] , with @xmath61}\bar{c}_i \geq 1 $ ] , is a feasible point of @xmath207 .
the reverse implication in proposition [ smallerstruct ] does not hold .
finally , we give a negative answer to morris question , which in turn implies a negative answer to vaughan s question .
+ let @xmath211 .
@xmath205 is feasible .
we simply write down the relevant inequalities and exhibit a feasible point .
the order of display matches @xmath212 in @xmath190 \setminus \left\{j\right\ } } \uplus { \ensuremath{\langle \mathcal{s}\rangle}}$ ] .
[ cols= " < " , ] let @xmath4 be an uc family that covers @xmath6 elements such that @xmath63
. the feasibility of @xmath194 does not imply the feasibility of @xmath40 .
let @xmath4 be an uc family that covers @xmath6 elements such that @xmath63 .
@xmath192 does not imply that @xmath193 is feasible .
in this work we design a cutting - plane algorithm that determines if a given uc family necessarily implies frankl s conjecture for all families that contain it . by employing exact rational integer programming and highly redundant verification routines , we classify more previously unknown miminal non - isomorphic fc - families than the total output of the past twenty - five years of research on the topic .
the effects of _ safely _ automating the discovery of fc - families allow us to answer several open questions of morris @xcite and vaughan @xcite .
in particular , the counterexamples we exhibit to settle open questions of interest require no _ trust _ from the reader , in the sense that they are independent of the complex optimization processeses that led to them , and can be checked by hand .
furthermore , our framework can be used to improve several other results in the following ways : * since our algorithm determines exactly whether a given uc family @xmath4 that covers @xmath6 elements is fc or nonfc for @xmath23 , lower bounds for previously unknown @xmath12 in this range become trivial to obtain . coupled with a computer algebra system or graph isomorphism software to obtain the isomorphism types of generators , upper or exact bounds for previously unknown @xmath12 can be easily obtained in the aforementioned range . *
the approach of morris @xcite for the classification of fc - families on five elements lends itself well to being generalized within our framework .
the number of minimal non - isomorphic generators for fc - families seems to quickly grow for @xmath213 , but we believe a complete classication for @xmath214 is possible with routine work . * proving the 3-sets conjecture of morris @xcite , by recovering the arguments of vaughan @xcite through a classification of @xmath215 for @xmath216 and using morris s lower bound on 3-sets , is within reach . *
each fc - family represents an
optimality " cut for the integer program that relates to frankl s conjecture in the work of pulaj , raymond and theis @xcite .
improved bounds may be achieved , if an efficient separation routine is implemented for uc inequalites on @xmath217 .
alternatively , a counterexample may be found .
finally , we believe that by separating uc and fs inequalities , our framework can possibly classify fc - families for @xmath218 . for larger @xmath6 other advanced techniques such as column generation may be necessary .
more generally , we believe our present work underscores the importance of bringing together computational tools and problems from different fields that rarely cross paths despite being well suited for each other as is often the case with advanced integer programming techniques and fitting problems in extremal combinatorics .
this is particularly useful when the problem at hand lacks apparent stucture and even finding interesting examples becomes non - trivial .
the author would like to thank ambros gleixner and stephen maher for their helpful suggestions about scip and exact scip .
00 polymath11func4 .
last accessed on november 20 , 2016 .
bonjak , i. and markovi , p. : the 11-element case of frankl s conjecture . , * 15*(1 ) , 17 ( 2008 ) , ( research paper 88 ) bruhn , h. , charbit , p. , schaudt , o. , and telle , j.a . : the graph formulation of the union - closed sets conjecture . , ( 2013 )
bruhn , h. and schaudt , o. : the journey of the union - closed sets conjecture . , * 31*(6 ) , 20432074 ( 2015 ) cheung , k.k.h . and gleixner , a. and steffy , d.e . : verifying integer programming results . , * 16 - 58 * , ( 2016 ) cook , w. , koch , t. , steffy , d. , and wolter , k. : a hybrid branch - and - bound approach for exact rational mixed - integer programming . , * 3 * , 305344 ( 2013 ) ibm ilog .
https://www-01.ibm.com/software/commerce/optimization/cplex-optimizer/ [ https://www-01.ibm.com/software/commerce/optimization/cplex-optimizer/ ] .
last accessed on november 20 , 2016 .
gamrath , g. , fischer , t. , gally , t. , gleixner , a.m. , hendel , g. , koch , t. , maher , s.j . ,
miltenberger , m. , mller , b , pfetsch , m.e . , puchert , c. , rehfeldt , d. , schenker , s. , schwarz , r. , serrano , f. , shinano , y. , vigerske , s. , weninger , d. , winkler , m. , witt , j.t . , and witzig , j. the scip optimization suite 3.2 .
zib - report ( 15 - 60 ) ( 2016 ) gao , w. and yu , h. : note on the union - closed sets conjecture . , * 49 * , 280288 ( 1998 ) gurobi .
https://www.gurobi.com/. last accessed on november 20 , 2016 .
mari , f. , ivkovi , m. , and vukovi , b. : formalizing frankl s conjecture : fc - families . , * 7362 * , 248263 ( 2012 ) markovi , p. : an attempt at frankl s conjecture . .
* 81 * 2943 ( 2007 ) morris , r. : fc - families , and improved bounds for frankl s conjecture . , * 27 * , 269282 ( 2006 ) poonen , b. : union - closed families . , *
59*(2 ) , 253268 ( 1992 ) pulaj , j. , raymond , a. , and theis , d. : new conjectures for union - closed families .
( 3 ) , # p23.3 .
( 2016 ) sarvate , d.g . and
renaud , j.c . : on the union - closed sets conjecture .
, * 27 * , 149154 ( 1989 ) vaughan , t.p . : families implying the frankl conjecture . , * 23 * , 851860 ( 2002 ) vaughan , t.p . : a note on the union closed sets conjecture . ,
* 45 * , 95108 ( 2003 ) vaughan , t.p . : three - sets in a union - closed family .
, * 49 * , 7384 ( 2004 ) vukovi , b. and ivkovi , m. : the 12-element case of frankl s conjecture . , ( 2012 ) to check the claims of _ infeasibility _ for the linear systems in this paper it is sufficient to ensure that the vector of values exhibited in square brackets before each row corresponds to the vector @xmath219 below .
let @xmath220 , @xmath221 and @xmath222 .
also let @xmath223 , @xmath224 and @xmath225 .
then the following system of linear equalities and inequalities in @xmath226 : @xmath227 + is infeasible if and only if there exist @xmath228 such that : @xmath229 + _ proof of proposition [ morriscounter]_. we identify sets in @xmath230 with the columns in the matrix below . for each column , the number on the top row represents its corresponding variable index in @xmath231 .
column @xmath232 corresponds to a weight vector for the elements in @xmath5 $ ] .
the columns representing families of sets @xmath150 and @xmath163 are colored red and blue , respectively .
as previously , @xmath233 .
+ @xmath234 \matrix ( m)[matrix of math nodes ] { \color{red}{w } & \color{red}{0 } & \color{red}{1 } & \color{red}{2 } & \color{red}{3 } & \color{red}{4 } & \color{red}{5 } & \color{red}{6 } & \color{red}{7 } & \color{red}{8 } & \color{red}{9 } & \color{red}{10 } & \color{red}{11 } & \color{red}{12 } & \color{red}{13 } & \color{red}{14 } & \color{red}{15 } & \color{red}{16 } & \color{red}{17 } & \color{red}{18 } & \color{red}{19 } & \color{red}{20 } & \color{red}{21 } & \color{red}{22 } & \color{red}{23 } & \color{red}{24 } & \color{red}{25 } & \color{red}{26 } & \color{red}{27 } & \color{red}{28 } & \color{red}{29 } & \color{red}{30 } & \color{red}{31 } \\ 16 & \color{blue}{1 } & \color{red}{1 } & 1 & 1 & 1 & 1 & 1 & 1 & \color{blue}{1 } & 1 & 1 & 1 & \color{red}{1 } & 1 & 1 & 1 & \color{blue}{1 } & 1 & 1 & \color{red}{1 } & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1\\ 8 & \color{blue}{1 } & \color{red}{1 } & 1 & 1 & 1 & 1 & 1 & 1 & \color{blue}{1 } & 1 & 1 & 1 & \color{red}{1 } & 1 & 1 & 1 & \color{blue}{0 } & 0 & 0 & \color{red}{0 } & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0\\ 12 & \color{blue}{1 } & \color{red}{1 } & 1 & 1 & 1 & 1 & 1 & 1 & \color{blue}{0 } & 0 & 0 & 0 & \color{red}{0 } & 0 & 0 & 0 & \color{blue}{1 } & 1 & 1 & \color{red}{1 } & 1 & 1 & 1 & 1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0\\ 20 & \color{blue}{1 } & \color{red}{1 } & 1 & 1 & 0 & 0 & 0 & 0 & \color{blue}{1 } & 1 & 1 & 1 & \color{red}{0 } & 0 & 0 & 0 & \color{blue}{1 } & 1 & 1 & \color{red}{1 } & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & 0 & 0 & 0 & 0\\ 17 & \color{blue}{1 } & \color{green}{1 } & 0 & 0 & 1 & 1 & 0 & 0 & \color{blue}{1 } & 1 & 0 & 0 & \color{red}{1 } & 1 & 0 & 0 & \color{blue}{1 } & 1 & 0 & \color{red}{0 } & 1 & 1 & 0 & 0 & 1 & 1 & 0 & 0 & 1 & 1 & 0 & 0\\ 15 & \color{blue}{1 } & 0 & 1 & 0 & 1 & 0 & 1 & 0 & \color{blue}{1 } & 0 & 1 & 0 & \color{red}{1 } & 0 & 1 & 0 & \color{blue}{1 } & 0 & 1 & \color{red}{0 } & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0\\ } ; \end{tikzpicture}\ ] ] + + @xmath235 \matrix ( m)[matrix of math nodes ] { \color{red}{32 } & \color{red}{33 } & \color{red}{34 } & \color{red}{35 } & \color{red}{36 } & \color{red}{37 } & \color{red}{38 } & \color{red}{39 } & \color{red}{40 } & \color{red}{41 } & \color{red}{42 } & \color{red}{43 } & \color{red}{44 } & \color{red}{45 } & \color{red}{46 } & \color{red}{47 } & \color{red}{48 } & \color{red}{49 } & \color{red}{50 } & \color{red}{51 } & \color{red}{52 } & \color{red}{53 } & \color{red}{54 } & \color{red}{55 } & \color{red}{56 } & \color{red}{57 } & \color{red}{58 } & \color{red}{59 } & \color{red}{60 } & \color{red}{61 } & \color{red}{62 } & \color{red}{63 } \\ 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & \color{red}{0 } & 0 & 0 & 0 & 0 & 0 & 0 & \color{red}{0}\\ 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & \color{red}{0 } & 0 & 0 & 0 & 0 & 0 & 0 & \color{red}{0}\\ 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & 1 & \color{red}{0 } & 0 & 0 & 0 & 0 & 0 & 0 & \color{red}{0}\\ 1 & 1 & 1 & 1 & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & 0 & 0 & 0 & 0 & \color{red}{1 } & 1 & 1 & 1 & 0 & 0 & 0 & \color{red}{0}\\ 1 & 1 & 0 & 0 & 1 & 1 & 0 & 0 & 1 & 1 & 0 & 0 & 1 & 1 & 0 & 0 & 1 & 1 & 0 & 0 & 1 & 1 & 0 & 0 & \color{red}{1 } & 1 & 0 & 0 & 1 & 1 & 0 & \color{red}{0}\\ 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & 1 & 0 & \color{red}{1 } & 0 & 1 & 0 & 1 & 0 & 1 & \color{red}{0}\\ } ; \end{tikzpicture}\ ] ]
+ + we prove that @xmath231 is unfeasible by branching on @xmath236 and showing that the linear relaxations of the two subproblems are infeasible .
we do this by explicitly exhibiting farkas dual values ( shown in square brackets ) for each row of some irreducible infeasible subset of constraints .
define @xmath237 .
then it is clear that if @xmath238 and @xmath239 are infeasible , this implies @xmath240 is unfeasible .
let @xmath241 , and let the following be the linear relaxation of @xmath242 , defined by the following constraints ( trivial ones not shown ) : 1 .
@xmath171:x_{0}=1}$ ] .
uc inequalites : + @xmath243:x_{11 } + x_{45 } - x_{9 } \leq 1}$ ] , @xmath244:x_{13 } + x_{59 } - x_9 \leq 1}$ ] , + @xmath243:x_{14 } + x_{43 } - x_{10 } \leq 1}$ ] , @xmath245:x_{22 } + x_{61 } - x_{20}\leq 1}$ ] , + @xmath246:\tt{x_{23 } + x_{60 } - x_{20}\leq 1}$ ] , @xmath245:x_{35 } + x_{45 } - x_{33}\leq 1}$ ] , + @xmath244:x_{35 } + x_{62 } - x_{34}\leq 1}$ ] , @xmath247:x_{37 } + x_{59 } - x_{33}\leq 1}$ ] , + @xmath245:x_{38 } + x_{43 } - x_{34 } \leq 1}$ ] , @xmath244:x_{38 } + x_{45 } - x_{36 } \leq 1}$ ] , + @xmath245:x_{38 } + x_{61 } - x_{36 } \leq 1}$ ] , @xmath245:x_{39 } + x_{44 } - x_{36 } \leq 1}$ ] , + @xmath243:x_{42 } + x_{55 } - x_{34 } \leq 1}$ ] , @xmath243:x_{53 } + x_{43 } - x_{33 } \leq 1}$ ] , + @xmath248:x_{54 } + x_{43 } - x_{34}\leq 1}$ ] , @xmath244:x_{44 } + x_{55 } - x_{36 } \leq 1}$ ] , + @xmath249:x_{47 } + x_{49 } - x_{33}\leq 1}$ ] .
fs inequalities : + @xmath250:x_{47 } - x_{1 } \leq 0}$ ] , @xmath247:x_{63 } - x_{1 } \leq 0}$ ] , @xmath251:x_{63 } - x_{8 } \leq 0}$ ] , + @xmath245:x_{7 } - x_{4 } \leq 0}$ ] , @xmath252:x_{55 } - x_{4 } \leq 0}$ ] , @xmath253:x_{63 } - x_{12 } \leq 0}$ ] , + @xmath244:x_{14 } - x_{2 } \leq 0}$ ] , @xmath254:x_{46 } - x_{2 } \leq 0}$ ] , @xmath253:x_{47 } - x_{3 } \leq 0}$ ] , + @xmath255:x_{61 } - x_{17 } \leq 0}$ ] , @xmath256:x_{62 } - x_{18 } \leq 0}$ ] , @xmath249:x_{63 } - x_{19 } \leq 0}$ ] , + @xmath254:x_{31 } - x_{24 } \leq 0}$ ] , @xmath245:x_{37 } - x_{32 } \leq 0}$ ] , @xmath249:x_{38 } - x_{32 } \leq 0}$ ] , + @xmath257:x_{39 } - x_{32 } \leq 0}$ ] , @xmath252:x_{47 } - x_{40 } \leq 0}$ ] , @xmath258:x_{55 } - x_{48 } \leq 0}$ ] , + @xmath259:x_{63 } - x_{56}\leq 0}$ ] .
fc inequalities : + @xmath243:x_{15 } + x_{53 } - x_{1 } - x_{5 } \leq 0}$ ] , @xmath260:x_{15 } + x_{57 } - x_{1 } - x_{9 } \leq 0}$ ] , + @xmath261:x_{58 } + x_{15 } - x_{8 } - x_{10 } \leq 0}$ ] , @xmath243:x_{15 } + x_{59 } - x_{1 } - x_{11 } \leq 0}$ ] , + @xmath260:x_{45 } + x_{23 } - x_{1 } - x_{5 } \leq 0}$ ] , @xmath248:x_{60 } + x_{23 } - x_{20 } - x_{16 } \leq 0}$ ] , + @xmath245:x_{61 } + x_{23 } - x_{21 } - x_{16 } \leq 0}$ ] , @xmath248:x_{45 } + x_{27 } - x_{1 } - x_{9 } \leq 0}$ ] , + @xmath244:x_{27 } + x_{61 } - x_{25 } - x_{16 } \leq 0}$ ] , @xmath250:x_{43 } + x_{29 } - x_{1 } - x_{9 } \leq 0}$ ] , + @xmath248:x_{54 } + x_{29 } - x_{20 } - x_{16 } \leq 0}$ ] , @xmath247:x_{29 } + x_{59 } - x_{16 } - x_{25 } \leq 0}$ ] , + @xmath249:x_{43 } + x_{30 } - x_{10 } - x_{8 } \leq 0}$ ] , @xmath243:x_{53 } + x_{30 } - x_{16 } - x_{20 } \leq 0}$ ] , + @xmath260:x_{59 } + x_{30 } - x_{16 } - x_{26 } \leq 0}$ ] , @xmath249:x_{31 } + x_{60 } - x_{16 } - x_{28}\leq 0}$ ] , + @xmath245:x_{43 } + x_{45 } - x_{8}- x_{41 } \leq 0}$ ] , @xmath245:x_{43 } + x_{62 } - x_{8}- x_{42 } \leq 0}$ ] , + @xmath244:x_{47 } + x_{62 } - x_{8}- x_{46 } \leq 0}$ ] , @xmath244:x_{51 } + x_{62 } - x_{16 } - x_{50 } \leq 0}$ ] , + @xmath260:x_{7}+ x_{54}- x_{4}- x_{6 } \leq 0}$ ] , @xmath261:x_{51 } + x_{53 } - x_{48 } - x_{49 } \leq 0}$ ] .
feasibility inequality : + @xmath262:88x_{0}+58x_{1}+54 x_{2}+24 x_{3}+48 x_{4}+18x_{5}+14 x_{6}-16 x_{7}+64 x_{8}}$ ] + @xmath263 + @xmath264 + @xmath265 + @xmath266 + @xmath267 + @xmath268 + @xmath269 .
1 . @xmath272:x_0 = 0}$ ] 2 .
fs inequalities : + @xmath273 : x_1 - x_0 \leq 0}$ ] , @xmath274 : x_6 - x_0 \leq 0}$ ] , @xmath275 : x_{11 } - x_0\leq 0}$ ] , + @xmath250 : x_{19 } - x_0 \leq 0}$ ] , @xmath275:x_{23 } - x_0 \leq 0}$ ] , @xmath249:x_{35 } - x_0 \leq 0}$ ] , + @xmath260:x_{37 } - x_0 \leq 0}$ ] , @xmath261:x_{38 } - x_0 \leq 0}$ ] , @xmath254:x_{39 } - x_0 \leq 0}$ ] , + @xmath276:x_{41 } - x_0 \leq 0}$ ] , @xmath252:x_{44 } - x_0 \leq 0}$ ] , @xmath255:x_{46 } - x_0\leq 0}$ ] , + @xmath277:x_{47 } - x_0 \leq 0}$ ] , @xmath247:x_{50 } - x_0\leq 0}$ ] , @xmath256:x_{55 } - x_0 \leq 0}$ ] , + @xmath278:x_{56 } - x_0 \leq 0}$ ] , @xmath279:x_{59 } - x_0 \leq 0}$ ] , @xmath252:x_{61 } - x_0 \leq 0}$ ] , + @xmath258:x_{62 } - x_0\leq 0}$ ] , @xmath257:x_{63 } - x_0 \leq 0}$ ] , @xmath253:x_{13 } - x_{12 } \leq 0}$ ] , + @xmath280:x_{14 } - x_{2 } \leq 0}$ ] , @xmath280 : x_{22 } - x_{18 } \leq 0}$ ] , @xmath281:x_{62 } - x_{18 } \leq 0}$ ] , + @xmath245:x_{42 } - x_{40 } \leq 0}$ ] , @xmath260:x_{51 } - x_{48 } \leq 0}$ ] , @xmath273:x_{43 } - x_{8}\leq 0}$ ] , + @xmath278 : x_{29 } - x_{17}\leq 0}$ ] , @xmath278:x_{61 } - x_{9 } \leq 0}$ ] , @xmath250:x_{63 } - x_{56 } \leq 0}$ ] .
fc inequalities : + @xmath273:x_{15 } + x_{45 } - x_{1}- x_{13}\leq 0}$ ] , @xmath261:x_{15 } + x_{53 } - x_{1}- x_{5 } \leq 0}$ ] , + @xmath282:x_{15 } + x_{57 } - x_{1}- x_{9}\leq 0}$ ] , @xmath273:x_{23 } + x_{62 } - x_{16 } - x_{22 } \leq 0}$ ] , + @xmath259:x_{27 } + x_{45 } - x_{1}- x_{9 } \leq 0}$ ] , @xmath275:x_{31 } + x_{43 } - x_{1}- x_{11}\leq 0}$ ] , + @xmath245 : x_{31 } + x_{53 } - x_{16 } - x_{21 } \leq 0}$ ] , @xmath282:x_{45 } + x_{57 } - x_{8}- x_{41 } \leq 0}$ ] , + @xmath261:x_{55 } + x_{58 } - x_{16 } - x_{50 } \leq 0}$ ] , @xmath279:x_{7}+ x_{54 } - x_{4}- x_{6 } \leq 0}$ ] , + @xmath248:x_{51 } + x_{53 } - x_{48 } - x_{49}\leq 0}$ ] .
fc - chain inequalities ( it is easy to check that the explicit chains work for all subsets ) : + @xmath283:x_{29 } + x_{47 } + x_{61 } + x_{63 } - x_{8}- x_{13 } - x_{17 } - x_{56 } \leq 0}$ ] , + @xmath284 , @xmath285 , @xmath286 , @xmath287 , @xmath288 , + @xmath289 , @xmath290 .
+ @xmath249:x_{29 } + x_{61 } + x_{62 } - x_{16 } - x_{17 } - x_{28 } \leq 0}$ ] , + @xmath291 , @xmath290 , @xmath292 , + @xmath293 .
+ @xmath273:x_{30 } + x_{31 } + x_{47 } + x_{63 } - x_{8}- x_{14 } - x_{16 } - x_{24 } \leq 0}$ ] , + @xmath285 , @xmath286 , @xmath294 , @xmath295 , @xmath296 , + @xmath297 , @xmath298 , @xmath299 .
+ @xmath249:x_{30 } + x_{31 } + x_{55 } - x_{19 } - x_{22 } - x_{24}\leq 0}$ ] , + @xmath297 , @xmath299 , @xmath300 , @xmath301 , @xmath302 .
+ @xmath260:x_{30 } + x_{31 } + x_{59 } - x_{16 } - x_{24 } - x_{26 } \leq 0}$ ] , + @xmath297 , @xmath299 , @xmath298 , @xmath296 , + @xmath303 , @xmath304 .
+ @xmath250:x_{47 } + x_{54 } + x_{63 } - x_{8}- x_{16 } - x_{38 } \leq 0}$ ] , + @xmath285 , @xmath294 , @xmath286 , @xmath305 , @xmath306 .
+ @xmath253 : x_{47 } + x_{60 } + x_{63 } -x_{8 } -x_{44 } -x_{56}\leq 0}$ ] .
+ @xmath286 , @xmath285 , @xmath288 , @xmath307 , @xmath308 .
feasibility inequality : + @xmath262:58x_{1}+54 x_{2}+24 x_{3}+48 x_{4}+18x_{5}+14 x_{6}-16 x_{7}+64 x_{8}}$ ] + @xmath263 + @xmath264 + @xmath265 + @xmath266 + @xmath267 + @xmath268 + @xmath269 .
[ 1256 , 3456 , 456 , 236 ] * & @xmath311 , @xmath312 , @xmath313 , @xmath314 , @xmath315 , @xmath316 + * & @xmath317 , @xmath318 , @xmath319 , @xmath320 , @xmath321 , @xmath322 + * & @xmath317 , @xmath323 , @xmath324 , @xmath325 , @xmath326 , @xmath327 + * & @xmath328 , @xmath318 , @xmath324 , @xmath329 , @xmath330 , @xmath327 + * & @xmath328 , @xmath318 , @xmath324 , @xmath329 , @xmath330 , @xmath327 + * & @xmath331 , @xmath332 , @xmath333 , @xmath334 , @xmath335 , @xmath336 + * & @xmath337 , @xmath338 , @xmath339 , @xmath340 , @xmath341 , @xmath316 + * & @xmath337 , @xmath342 , @xmath343 , @xmath344 , @xmath345 , @xmath346 + * & @xmath347 , @xmath348 , @xmath349 , @xmath350 , @xmath351 , @xmath352 + * & @xmath353 , @xmath354 , @xmath333 , @xmath329 , @xmath355 , @xmath356 + * & @xmath328 , @xmath357 , @xmath358 , @xmath359 , @xmath360 , @xmath361 + * & @xmath328 , @xmath354 , @xmath319 , @xmath320 , @xmath330 , @xmath327 + * & @xmath362 , @xmath332 , @xmath324 , @xmath325 , @xmath355 , @xmath363 + * & @xmath317 , @xmath323 , @xmath358 , @xmath364 , @xmath365 , @xmath361 + * & @xmath317 , @xmath357 , @xmath366 , @xmath359 , @xmath360 , @xmath367 + * & @xmath317 , @xmath357 , @xmath366 , @xmath359 , @xmath360 , @xmath367 + * & @xmath337 , @xmath368 , @xmath313 , @xmath340 , @xmath369 , @xmath370 + * & @xmath362 , @xmath332 , @xmath324 , @xmath325 , @xmath330 , @xmath327 + * & @xmath317 , @xmath357 , @xmath366 , @xmath359 , @xmath360 , @xmath367 + * & @xmath328 , @xmath323 , @xmath358 , @xmath329 , @xmath355 , @xmath371 + * & @xmath372 , @xmath332 , @xmath319 , @xmath320 , @xmath321 , @xmath356 + * & @xmath328 , @xmath318 , @xmath373 , @xmath364 , @xmath355 , @xmath371 + * & @xmath331 , @xmath354 , @xmath319 , @xmath320 , @xmath330 , @xmath327 + * & @xmath328 , @xmath323 , @xmath373 , @xmath329 , @xmath355 , @xmath371 + * & @xmath362 , @xmath332 , @xmath324 , @xmath329 , @xmath330 , @xmath363 + * & @xmath372 , @xmath354 , @xmath319 , @xmath320 , @xmath321 , @xmath322 + * & @xmath323 , @xmath318 , @xmath373 , @xmath329 , @xmath355 , @xmath363 + * & @xmath318 , @xmath323 , @xmath373 , @xmath329 , @xmath355 , @xmath363 + * & @xmath323 , @xmath318 , @xmath373 , @xmath329 , @xmath355 , @xmath363 + * & @xmath318 , @xmath342 , @xmath374 , @xmath375 , @xmath376 , @xmath377 + * & @xmath378 , @xmath338 , @xmath333 , @xmath314 , @xmath379 , @xmath380 + * & @xmath381 , @xmath382 , @xmath383 , @xmath384 , @xmath385 , @xmath386 + * & @xmath387 , @xmath387 , @xmath388 , @xmath329 , @xmath330 , @xmath327 + * & @xmath317 , @xmath357 , @xmath373 , @xmath320 , @xmath326 , @xmath356 , @xmath389 + * & @xmath317 , @xmath323 , @xmath366 , @xmath364 , @xmath365 , @xmath371 , @xmath390 + * & @xmath317 , @xmath357 , @xmath388 , @xmath364 , @xmath326 , @xmath327 , @xmath389 + * & @xmath317 , @xmath357 , @xmath324 , @xmath359 , @xmath355 , @xmath363 , @xmath390 + * & @xmath317 , @xmath323 , @xmath358 , @xmath359 , @xmath365 , @xmath361 , @xmath390 + * & @xmath317 , @xmath318 , @xmath358 , @xmath359 , @xmath355 , @xmath363 , @xmath390 + * & @xmath317 , @xmath357 , @xmath324 , @xmath329 , @xmath355 , @xmath363 , @xmath391 + * & @xmath328 , @xmath392 , @xmath388 , @xmath364 , @xmath379 , @xmath322 , @xmath393 + * & @xmath317 , @xmath318 , @xmath373 , @xmath329 , @xmath355 , @xmath363 , @xmath394 + * & @xmath317 , @xmath318 , @xmath324 , @xmath359 , @xmath355 , @xmath363 , @xmath394 + * & @xmath328 , @xmath332 , @xmath333 , @xmath320 , @xmath326 , @xmath327 , @xmath393 + * & + * & @xmath317 , @xmath378 , @xmath366 , @xmath364 , @xmath379 , @xmath322 , @xmath395 + * & + * & + * & + * & + * & + * & + * & @xmath362 , @xmath332 , @xmath324 , @xmath325 , @xmath330 , @xmath371 , @xmath393 + * & @xmath317 , @xmath323 , @xmath358 , @xmath364 , @xmath365 , @xmath361 , @xmath391 + * & @xmath372 , @xmath392 , @xmath396 , @xmath397 , @xmath335 , @xmath398 , @xmath399 + & + & + & @xmath400 , @xmath401 , @xmath402 , @xmath403 , @xmath404 , @xmath405 , @xmath406 + & @xmath400 , @xmath407 , @xmath408 , @xmath409 , @xmath410 , @xmath411 , @xmath412 + & @xmath400 , @xmath401 , @xmath402 , @xmath413 , @xmath414 , @xmath405 , @xmath406 + & @xmath415 , @xmath416 , @xmath417 , @xmath413 , @xmath418 , @xmath419 , @xmath420 + & + & @xmath311 , @xmath392 , @xmath388 , @xmath397 , @xmath321 , @xmath356 , @xmath421 + & @xmath400 , @xmath407 , @xmath408 , @xmath403 , @xmath404 , @xmath411 , @xmath412 + & @xmath422 , @xmath401 , @xmath417 , @xmath413 , @xmath414 , @xmath419 , @xmath423 + & @xmath422 , @xmath416 , @xmath417 , @xmath424 , @xmath418 , @xmath419 , @xmath423 + & @xmath400 , @xmath407 , @xmath408 , @xmath409 , @xmath410 , @xmath411 , @xmath412 + & @xmath400 , @xmath407 , @xmath408 , @xmath403 , @xmath404 , @xmath411 , @xmath412 + & @xmath400 , @xmath407 , @xmath408 , @xmath409 , @xmath410 , @xmath411 , @xmath412 + & + & @xmath425 , @xmath401 , @xmath402 , @xmath403 , @xmath404 , @xmath411 , @xmath406 + & @xmath400 , @xmath407 , @xmath402 , @xmath403 , @xmath404 , @xmath411 , @xmath406 + & @xmath400 , @xmath401 , @xmath402 , @xmath409 , @xmath404 , @xmath411 , @xmath406 + & + & @xmath422 , @xmath416 , @xmath426 , @xmath424 , @xmath414 , @xmath419 , @xmath423 + & @xmath422 , @xmath416 , @xmath426 , @xmath424 , @xmath414 , @xmath419 , @xmath423 + & @xmath427 , @xmath428 , @xmath429 , @xmath430 , @xmath431 , @xmath432 , @xmath433 + & @xmath422 , @xmath416 , @xmath426 , @xmath424 , @xmath418 , @xmath405 , @xmath423 + & @xmath427 , @xmath428 , @xmath429 , @xmath430 , @xmath431 , @xmath432 , @xmath433 + & @xmath422 , @xmath416 , @xmath426 , @xmath424 , @xmath418 , @xmath405 , @xmath423 + & @xmath422 , @xmath416 , @xmath426 , @xmath424 , @xmath414 , @xmath419 , @xmath423 + & @xmath422 , @xmath416 , @xmath426 , @xmath424 , @xmath418 , @xmath405 , @xmath423 + & @xmath422 , @xmath416 , @xmath426 , @xmath424 , @xmath418 , @xmath419 , @xmath406 + & @xmath422 , @xmath401 , @xmath426 , @xmath424 , @xmath414 , @xmath405 , @xmath406 + & @xmath415 , @xmath434 , @xmath426 , @xmath435 , @xmath418 , @xmath419 , @xmath423 + & @xmath415 , @xmath434 , @xmath426 , @xmath435 , @xmath418 , @xmath419 , @xmath423 + & @xmath422 , @xmath416 , @xmath417 , @xmath403 , @xmath418 , @xmath419 , @xmath406 + & @xmath422 , @xmath434 , @xmath426 , @xmath413 , @xmath436 , @xmath437 , @xmath423 + & @xmath400 , @xmath401 , @xmath402 , @xmath403 , @xmath414 , @xmath405 , @xmath412 + & @xmath438 , @xmath439 , @xmath440 , @xmath441 , @xmath442 , @xmath443 , @xmath444 , @xmath445 + & @xmath438 , @xmath439 , @xmath446 , @xmath441 , @xmath442 , @xmath447 , @xmath444 , @xmath445 + & @xmath438 , @xmath428 , @xmath448 , @xmath449 , @xmath431 , @xmath432 , @xmath450 , @xmath451 + & @xmath438 , @xmath452 , @xmath446 , @xmath453 , @xmath454 , @xmath455 , @xmath433 , @xmath456 + & @xmath438 , @xmath439 , @xmath440 , @xmath453 , @xmath442 , @xmath443 , @xmath457 , @xmath458 + & @xmath459 , @xmath452 , @xmath429 , @xmath460 , @xmath454 , @xmath411 , @xmath461 , @xmath462 + & + & @xmath459 , @xmath428 , @xmath463 , @xmath430 , @xmath464 , @xmath432 , @xmath450 , @xmath451 + & @xmath459 , @xmath452 , @xmath463 , @xmath460 , @xmath431 , @xmath405 , @xmath450 , @xmath451 + & + & @xmath459 , @xmath465 , @xmath402 , @xmath460 , @xmath464 , @xmath466 , @xmath423 , @xmath467 + & @xmath459 , @xmath452 , @xmath429 , @xmath460 , @xmath454 , @xmath411 , @xmath461 , @xmath462 + & @xmath459 , @xmath468 , @xmath408 , @xmath460 , @xmath464 , @xmath469 , @xmath406 , @xmath467 + & @xmath459 , @xmath465 , @xmath408 , @xmath460 , @xmath410 , @xmath466 , @xmath406 , @xmath467 + & + & @xmath459 , @xmath452 , @xmath429 , @xmath460 , @xmath454 , @xmath411 , @xmath461 , @xmath462 + & @xmath459 , @xmath401 , @xmath402 , @xmath409 , @xmath410 , @xmath411 , @xmath470 , @xmath471 + & @xmath438 , @xmath439 , @xmath446 , @xmath449 , @xmath442 , @xmath432 , @xmath457 , @xmath458 + & @xmath459 , @xmath407 , @xmath408 , @xmath430 , @xmath431 , @xmath466 , @xmath406 , @xmath472 + & @xmath459 , @xmath452 , @xmath463 , @xmath460 , @xmath431 , @xmath405 , @xmath450 , @xmath451 + & @xmath459 , @xmath468 , @xmath463 , @xmath430 , @xmath464 , @xmath466 , @xmath412 , @xmath472 + & @xmath459 , @xmath465 , @xmath463 , @xmath409 , @xmath431 , @xmath432 , @xmath461 , @xmath451 + & @xmath422 , @xmath434 , @xmath408 , @xmath403 , @xmath464 , @xmath419 , @xmath457 , @xmath471 + & + & @xmath422 , @xmath407 , @xmath446 , @xmath453 , @xmath464 , @xmath466 , @xmath433 , @xmath456 + & + & + & + & @xmath477 , @xmath439 , @xmath440 , @xmath449 , @xmath454 , @xmath447 , @xmath474 , @xmath458 , @xmath478 + & + & + & @xmath427 , @xmath428 , @xmath446 , @xmath453 , @xmath410 , @xmath455 , @xmath461 , @xmath451 , @xmath479 + & + & + & + & + & + & + & + & + & @xmath425 , @xmath452 , @xmath446 , @xmath430 , @xmath414 , @xmath443 , @xmath444 , @xmath451 , @xmath480 + @xmath481 & @xmath482 & & + @xmath483 & @xmath484 & @xmath485 & + @xmath486 & @xmath487 & @xmath488 & @xmath489 + @xmath490 & @xmath491 & @xmath492 & @xmath493 + @xmath494 & @xmath495 & @xmath496 & @xmath497 + | we find previously unknown families which imply frankl s conjecture for all families that contain them using an algorithmic framework .
the conjecture states that for any non - empty union - closed ( uc ) family there exists an element in at least half of the sets .
poonen s theorem characterizes the existence of weights which determine whether a given uc family implies the conjecture for all uc families which contain it .
an uc family is _ nonfrankl - complete _ ( nonfc ) , if and only if it does not imply the conjecture in its elements for some uc family that contains it .
we design a cutting - plane method that computes the explicit weights which imply the existence conditions of poonen s theorem .
this method allows us to answer several open questions , including the construction of a counterexample to an eleven - year - old conjecture of morris about the structure of generators for nonfc - families .
* keywords : * frankl s conjecture , union - closed families , integer programming , cutting - plane method , extremal combinatorics |
the tight correlation between luminosity and rotation velocity for spiral galaxies , a.k.a . the tully - fisher relation ( tfr , * ? ? ? * ) , has motivated dozens of studies over the years . in pursuit of the best possible distance indicator ,
most tf analyses have been carried out in red or infrared passbands ( see reviews by * ? ? ?
* ; * ? ? ?
* ) , which offer lower scatter than bluer passbands .
however , recent theoretical work has emphasized the value of studying tf scatter for its own sake , as this scatter holds fundamental clues to the formation and evolution of galaxies ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
observations in bluer passbands are extremely useful for understanding tf scatter , because the scatter arising from recent star formation , differences in stellar populations , and the spread in formation redshifts is most visible in the blue .
physical scatter in the tfr can come from three sources : ( 1 ) variations in the stellar mass - to - light ratio , ( 2 ) variations in the stellar mass fraction ( stellar - to - total mass ratio ) , and ( 3 ) differences in how the observed velocity width relates to the total mass . by identifying physical scatter from any one of these sources we limit the contributions of the other two , simultaneously gaining insight into the star formation histories of galaxies , the relationship between visible galaxies and dark matter halos , and the dynamics and structure of galaxy disks .
a better understanding of the sources of tf scatter may also minimize uncertainties in other types of tf analyses .
for example , if tf slope varies with environment , or if galaxies with different colors define different tf zero points , then a cluster tfr calibrated with a small set of galaxies that have cepheid distances may not be universal ( e.g. as in * ? ? ?
* ; * ? ? ?
correcting for these systematics could reduce scatter and improve the tfr as a distance indicator .
calibrating zero point shifts as a function of color and emission line strength could also aid in the interpretation of evidence for luminosity evolution in the intermediate - redshift tfr ( * ? ? ?
* ; * ? ? ?
* see also kannappan , fabricant , & franx , in preparation ) . here , we use a well - defined sample of emission - line galaxies drawn from the 196-galaxy nearby field galaxy survey ( nfgs , * ? ? ?
* ; * ? ? ?
* ) to examine how offsets from the u , b , and r - band tfrs depend upon the physical properties of galaxies . while previous tf studies have often included hundreds or thousands of galaxies ( e.g. *
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , none of these data sets has both the broadly representative sample demographics and the wide array of supporting data of the nfgs ( [ sc : nfgstf1 ] ) .
thus despite its modest size , the present survey offers unique advantages for studying scatter in the tfr .
the bulk of this paper analyzes a sample of nfgs galaxies consisting of @xmath670 sa sd spiral galaxies brighter than m@xmath5 .
this spiral sample allows us to investigate the behavior of galaxies often excluded from the tfr : sa galaxies and galaxies with peculiarities such as warps , multiple nuclei , or interacting companions .
ultimately , we seek to fit these galaxies into a unified picture , in which we understand tf residuals in terms of continuous physical properties . to this end
we undertake a detailed analysis of third - parameter correlations with tf residuals , and we attempt to form a physical understanding of tf scatter .
the remainder of the paper draws from the full nfgs to define an extended sample of @xmath6110 galaxies including very late types , emission - line e / s0 galaxies , and faint dwarfs .
the extended sample allows us to search for physical drivers of tf scatter in a heterogeneous population .
the only other comparable tf sample is the @xmath640-galaxy ursa major sample of @xcite , which was drawn from a restricted environment .
the virgo and fornax cluster samples of @xcite and @xcite apply morphology restrictions that exclude irregular or interacting galaxies .
although some intermediate - redshift tf analyses have included a wide range of morphologies @xcite , these studies nonetheless favor a select population : typically , moderately bright blue galaxies with strong emission and high surface brightness . and
although the samples studied by @xcite and @xcite both span large ranges in luminosity , the former study is restricted by hubble type and evidence for interaction , while the latter study is drawn from a heterogeneous mix of smaller samples selected by various criteria in different photometric bands .
the nfgs thus provides an unusually broad and unbiased representation of the general galaxy population . in what follows
, we describe the survey and briefly review our data reduction and analysis techniques ( [ sc : nfgstf1][sc : ldata ] ) , then present results for the spiral sample ( [ sc : spiralcal][sc : spiral3 ] ) and for the extended sample ( [ sc : fulltfr ] ) .
we summarize our major conclusions in [ sc : concltf1 ] .
further details on our data analysis methods may be found in appendices [ sc : fitting][sc : asymmeast ] .
the comprehensive nature of the nearby field galaxy survey ( nfgs ) enables us to analyze how tf scatter depends on a broad range of galaxy properties .
the database includes ubr photometry , integrated and nuclear spectrophotometry , and gas and stellar kinematic data for a statistically representative sample of the local galaxy population ( * ? ? ?
* ; * ? ? ?
* kannappan et al . , in preparation ) .
the 196 nfgs galaxies were objectively selected from the cfa 1 redshift survey @xcite without preference for morphology , color , diameter , inclination , environment , or any other galaxy property , as described in @xcite . to counteract the bright galaxy bias of cfa 1
, galaxies were included approximately in proportion to the local b - band galaxy luminosity function ( lf , e.g. * ? ? ?
* ) , with the final sample spanning luminosities from m@xmath7 @xmath6@xmath823 to @xmath815 .
the resulting sample provides an unusually unbiased sampling of the general galaxy population . however , no sample is completely free of selection effects .
the nfgs is subject to the inherent color and surface brightness biases of the parent cfa 1 survey , itself a descendent of the zwicky catalog and therefore essentially b - selected . also , despite the attempt to favor faint galaxies in the selection process
, the nfgs reproduces the local lf imperfectly .
figure [ fg : lf ] shows that the sample luminosity distribution varies slowly over the range @xmath9 and cuts off for brighter and fainter galaxies .
another potential concern is the luminosity - distance correlation built into the nfgs .
the survey was selected with a luminosity - dependent lower redshift limit instead of a maximum diameter limit , in order to increase the odds that galaxies would fit within the length of the 3@xmath10 spectrograph slit without biasing the sample in diameter @xcite . as a result ,
fainter galaxies have greater fractional uncertainties in distances and absolute magnitudes , and such galaxies receive less weight in error - weighted tf fits .
for this reason , we adopt unweighted fits as our primary fitting technique ( see appendix [ sc : fitting ] for a discussion of the differences between techniques ) .
our primary velocity widths are derived from the full set of 153 major - axis ionized gas rotation curves ( rcs ) obtained for the nfgs kinematic database ( kannappan et al .
, in preparation ) .
this sample is complete in the sense that kinematic observations were attempted for all nfgs galaxies ; however , emission line detection limits varied with observing conditions , integration times , and available emission lines ( see [ sc : orcobs ] ) . figure [ fg : emhists ] displays the emission line properties of the optical rc sample in the context of the nfgs as a whole . as illustrated in figure [ fg : typemagdistrib ] , the optical rc sample includes all nfgs galaxies of type sab and later ( including type `` pec '' galaxies ) , except for two galaxies that show largely featureless spectra .
in addition , the sample includes 12 of the 13 sa s in the nfgs and 17 of the e / s0 s . for comparison with the optical data ,
we have also extracted hi 21-cm linewidths for 105 nfgs galaxies from the catalogs of @xcite and @xcite .
these galaxies are not a statistically fair subsample of the nfgs : hi data are more often missing for the higher luminosity nfgs galaxies .
we exclude from analysis one galaxy that has hi data but no optical rc ( ngc 2692 , an sa galaxy ) .
the nfgs includes galaxies at all inclinations @xmath11 , but @xmath12 corrections are uncertain for galaxies with @xmath13 , so most of our analysis considers only the 108 nfgs emission - line galaxies with @xmath14 .
however , we do not restrict inclinations when calibrating optical - to - radio conversions ( [ sc : vcalcs ] ) , because no inclination correction is required .
long slit spectra were obtained during several observing runs between 1996 and 1999 , using the fast spectrograph on the tillinghast telescope @xcite . for most galaxies we observed the h@xmath4 , [ nii ] , and [ sii ] lines between @xmath662007200 , with spectral resolution @xmath15 30 @xmath16 , a 2@xmath17 or occasionally 3@xmath17 wide slit , and a spatial binning of 2.3@xmath17/pixel ( comparable to the typical seeing of 2@xmath17 ) . for some galaxies , in conjunction with stellar absorption line observations
, we also observed the h@xmath18 and [ oiii ] emission lines between @xmath641006100 , again with 2.3@xmath17 spatial binning and a 2@xmath17 slit , but with reduced spectral resolution , @xmath19 @xmath16 .
these lower resolution observations serve as the primary emission line data for 10 galaxies for which we lack high resolution observations .
the alignment between the slit position angle and the galaxy major axis position angle listed in the ugc @xcite was generally tight , @xmath20p.a .
@xmath21 6@xmath22 . for two galaxies
the slit was severely misaligned ( @xmath20p.a .
= 40@xmath22 and 75@xmath22 ) . also , a few galaxies had no ugc p.a . and
were instead assigned p.a.s based on visual inspection of the digitized poss images ; these galaxies have uncertain @xmath20p.a .
all galaxies with large or uncertain @xmath20p.a .
have photometric inclination @xmath13 and so are automatically excluded from our tf samples . in appendix [ sc :
vdefns ] , where we do not employ an inclination restriction , we explicitly exclude galaxies with large or uncertain @xmath20p.a .
all of the data were reduced using standard methods , including bias and dark subtraction , flat - fielding , wavelength calibration , heliocentric velocity correction , sky subtraction , spectral straightening , and cosmic ray removal , using iraf and idl .
we extracted rotation curves by fitting all available emission lines simultaneously with fixed line spacing in @xmath23 , rejecting fits with s / n @xmath24 3 . in cases of severe h@xmath4 or h@xmath18 absorption , these lines were omitted from the fits .
we have considered three possible optical velocity measures : @xmath25 , the single largest velocity in the rotation curve ; @xmath26 , the velocity interpolated at the `` critical radius '' ( 1.3@xmath27 , the peak velocity position for a theoretical exponential disk , * ? ? ?
* ) using a functional fit to the rotation curve ( cf .
* ) ; and @xmath28 , defined as half the difference between the statistical `` probable minimum '' and `` probable maximum '' velocities in the rotation curve ( cf .
the prescriptions for these velocity measures are fully described in appendix [ sc : vdefns ] .
as illustrated in figure [ fg : rcgallery ] , @xmath25 is overly sensitive to details of rc shape .
@xmath26 sucessfully models bright spiral galaxies similar to those in courteau s sample but sometimes fails when confronted with the full variety of rotation curve shapes in the nfgs . for our sample ,
the most robust velocity width estimator is @xmath28 , which makes use of all of the data without imposing a specific model on the rc . except when comparing our data with other optical rc tf samples ,
we convert all of our optical and radio velocity widths to an equivalent w@xmath29 linewidth scale ( where w@xmath30 is the hi linewidth at @xmath31% of peak intensity , see * ? ? ?
* ) . for galaxies with w@xmath32 but no w@xmath29 in the catalog ,
we simply subtract 20 @xmath33 from the w@xmath32 linewidths and use the adjusted w@xmath32 s together with the w@xmath29 s , referring to the combined data set as w@xmath34 .
the number 20 @xmath35 comes from a comparison of w@xmath32 and w@xmath29 where we have both measures ; @xcite derive a similar value from a much larger sample .
appendix [ sc : vdefns ] presents the empirically fitted relations we use to convert raw optical @xmath36 s to w@xmath29-equivalent w@xmath37 s , as determined from the 96 galaxies for which we have both optical rc and w@xmath29 data .
we use the scatter in the conversion relations to set the nominal error bars for the final velocity widths .
these error bars include not only the optical rc measurement errors ( which are small ) but also all other sources of discrepancy between optical and radio velocity widths , including p.a .
misalignment errors as well as `` physical '' errors due to fundamental differences between the kinematics of unresolved global hi fields and the kinematics of hii regions that happen to lie along arbitrary 1d lines .
_ however , contrary to common assumption , the limited extent of the optical rcs does not lead to any significant discrepancy with the global hi linewidths . _
radio w@xmath29 s and optical velocity widths agree well even for the nfgs galaxies whose rcs are truncated at @xmath211.3@xmath27 ( see appendix [ sc : vdefns ] ) . except
when comparing with @xcite , we adopt the w@xmath29-equivalent velocity width w@xmath38 as our default optical velocity width , with a nominal error bar of 20 @xmath35 .
all optical velocity widths are computed from rest - frame rotation curves , so no additional correction for cosmological expansion is necessary .
hi linewidths are corrected for cosmological expansion by dividing by @xmath39 .
we correct for inclination by dividing by @xmath40 , denoting corrected velocity widths with a superscript @xmath11 . here
@xmath41 with @xmath11 set to 90 whenever @xmath42 , the assumed intrinsic minor - to - major axis ratio .
we adopt @xmath43 except when comparing with @xcite , in which case we use his value ( 0.18 ) .
tests with a range of @xmath44 s indicate that tf scatter is remarkably insensitive to the details of computing inclinations .
the errors are propagated assuming errors in @xmath45 and @xmath46 equal to their measurement precision in the ugc ( typically 0.050.1@xmath10 ) .
we have checked the quality of the ugc estimates against estimates derived from both 2mass data and nfgs ccd photometry ( r. jansen , private communication ) : the newer measurements do not improve tf scatter .
we do not apply turbulence corrections to either optical or radio velocity widths , except when calculating extinctions ( see [ sc : extinct ] ) .
turbulence estimates are uncertain , and furthermore turbulence corrections may be inappropriate for small galaxies and non - equilibrium systems .
in fact turbulence may provide an important source of support for such systems , and we might wish to _ add _ a gas pressure correction to the optical velocity widths .
the conversion of optical velocity widths to an equivalent w@xmath29 scale ( [ sc : vcalcs ] ) effectively makes such a correction , averaged over the sample .
simulations by @xcite indicate that slit misalignment effects remain negligible for @xmath20p.a .
@xmath21 15@xmath22 , and observational tests in the range @xmath20p.a .
@xmath21 8@xmath22 confirm this result @xcite .
for galaxies with @xmath14 , our p.a .
alignments are all within 6@xmath22 of the reference p.a.s listed in the ugc ( [ sc : orcobs ] ) .
comparison between the ugc p.a.s and corresponding automated p.a .
measurements from 2mass and poss @xcite suggests typical catalog - to - catalog deviations of under 15@xmath22 .
therefore the combined misalignment from observational mismatch and reference uncertainty should nearly always be within the range tested by @xcite , and we choose to make no correction . however , an rms p.a .
misalignment error term is _ de facto _ included in our errors , because we estimate the optical velocity width errors based on the scatter in the optical - to - radio conversion ( [ sc : vcalcs ] ) .
extrapolated total b and r magnitudes are available in the nfgs photometry database @xcite for all galaxies with optical rcs .
in addition , @xciteprovide u - band magnitudes for all but two of the galaxies .
the database magnitudes include corrections for galactic absorption and have typical errors of @xmath60.020.03 mag .
we compute absolute magnitudes using distances derived from a linear hubble flow model with h@xmath47 , corrected for virgocentric infall . for consistency
, we use the redshifts tabulated by @xcite and adopt their virgocentric infall correction ( from * ? ? ?
comparison with our newly measured redshifts indicates good agreement between the two sets of measurements ( @xmath4835 @xmath49 ) with a few outliers .
we assign a redshift error of the greater of 35 @xmath16 or the difference between the tabulated and newly measured redshifts and add this term in quadrature with an assumed peculiar velocity uncertainty of 200 @xmath49 .
magnitudes corrected for internal extinction are denoted by a superscript @xmath11 . except
when comparing with @xcite , we adopt the velocity width - dependent ( and thus implicitly luminosity - dependent ) internal extinction corrections of @xcite : @xmath50 where @xmath51 is the major to minor axis ratio and @xmath52 is the extinction coefficient .
@xcite measure @xmath52 at b , r , and k. for the u band we use extrapolated coefficients : @xmath53 here @xmath54 is the velocity width on the turbulence - corrected scale of @xcite : @xmath55 where @xmath56 , @xmath57 , and @xmath58 .
tully & fouque introduce the last term to define a dynamical velocity width suitable for dwarf galaxies ; it prevents small linewidths from being `` corrected '' to square roots of negative numbers .
note that we add the 17 @xmath16 `` restored turbulence '' term before the inclination correction , apparently contrary to the formula in tully & fouque , because otherwise the turbulence correction is multiplied by @xmath59 while the restoration is not , which still results in square roots of negative numbers . while we do not generally use turbulence corrections ( see [ sc : vcorrs ] ) , for consistency with tully et al .
we make an exception when computing @xmath60 .
w@xmath32-equivalent linewidths are calculated from optical velocity widths as described in
[ sc : vcalcs ] and appendix [ sc : vdefns ] .
we apply the above extinction corrections to galaxies of type sa and later ; extinction corrections for e / s0 galaxies are set to zero . except for one e galaxy ,
this choice has little impact , since applying the extinction formula to the other e / s0 s would yield corrections of @xmath240.15 mag .
note that we _ do _ apply the correction to dwarf galaxies .
although the @xcite extinction corrections were derived for a sample composed primarily of spiral galaxies , @xcite has shown that they perform well for dwarf galaxies . as prescribed by @xcite ,
the extinction corrections are set to zero whenever they go negative , to avoid having the corrected luminosities come out fainter than the observed luminosities .
for the same reason , the error bars on the corrected magnitudes do not extend fainter than the measurement errors on the observed magnitudes .
@xcite , following @xcite , uses an alternative correction formula with no velocity width or luminosity dependence .
however , if we follow their procedure and choose a luminosity - independent value of @xmath61 to best minimize tf scatter , we find that no value of @xmath61 reduces scatter as well as the @xcite formulation . in the r band ,
the @xcite corrections perform only marginally better than the luminosity - independent corrections , but at b and u the differences become more substantial ( at u , scatter decreases by 18% compared to 8% ) .
we therefore conclude that the general behavior , if not the zero point , of the luminosity - dependent corrections appears correct .
traditional tf samples consist of moderately bright spiral galaxies , chosen to be relatively edge - on to avoid uncertain @xmath12 corrections . in this section and [ sc :
pecandsa][sc : spiral3 ] we analyze the tfr for spiral galaxies brighter than @xmath2 and inclined by more than 40@xmath22 .
the nfgs contains 69 sa sd galaxies that meet these criteria ; optical velocity widths and magnitudes are available for 68 of these ( 67 at u ) .
hi linewidths are available for 46 of the 68 ( or 45 of the 67 ) .
table gives tf fit parameters for the spiral tf sample in u , b , and r , including the observed scatter values and the scatter values predicted to arise from measurement errors .
we provide fit results using three different types of fits , as discussed in appendix [ sc : fitting ] , but in what follows we prefer the inverse fit ( a linear fit that minimizes residuals in velocity width ) , which best avoids slope bias .
rrrrr + u & [email protected] & [email protected] & [email protected] & [email protected] + b & [email protected] & [email protected] & [email protected] & [email protected] + r & [email protected] & [email protected] & [email protected] & [email protected] + + u & [email protected] & [email protected] & [email protected] & [email protected] + b & [email protected] & [email protected] & [email protected] & [email protected] + r & [email protected] & [email protected] & [email protected] & [email protected] + + u & 0.88(0.56 ) & 0.81(0.51 ) & 1.02(0.63 ) & 0.85(0.45 ) + b & 0.78(0.56 ) & 0.73(0.52 ) & 0.82(0.59 ) & 0.77(0.46 ) + r & 0.70(0.56 ) & 0.67(0.53 ) & 0.74(0.58 ) & 0.68(0.42 ) + measurement - error scatter remains fairly constant across the three passbands , while observed scatter increases sharply from r to b to u. subtracting the two in quadrature yields an estimate of the intrinsic scatter in the tfr . using the 46 galaxies in the hi linewidth sample ,
we have directly compared the intrinsic scatter estimates from the hi linewidth and optical rc tfrs .
we find a slightly lower intrinsic scatter for the optical tfr , which probably indicates that the errors on the optical velocity widths have been overestimated , and those on the catalog hi linewidths underestimated ( the hi errors do not account for systematic effects such as confusion in the beam ) .
however , we have not adjusted the predicted scatter values in table , as the listed values would change by only @xmath60.02 mag and the comparison performed here is based on an incomplete subsample
. the steep slopes and high scatter values reported here may be surprising to those familiar with restricted tf samples and/or alternate analysis techniques based on forward fits ( appendix [ sc : fitting ] ) and luminosity - independent extinction corrections .
_ we stress that the slope , zero point , and scatter in different samples can only be meaningfully compared when differences in sample selection and analysis techniques are fully taken into account . _ in addition , it is important to quantify measurement - error scatter , which may be qualitatively different for cluster and field samples .
below , [ sc : field][sc : cluster ] compare the nfgs tf sample with several previous field and cluster tf samples and demonstrate that our results are consistent when samples and analysis techniques are carefully matched .
the reader who wishes to skip directly to new results may go to [ sc : pecandsa ] .
the @xcite and ( * ? ? ?
* hereafter mfb ) samples are ideal for a direct comparison with the nfgs because both employ optical rotation curves . both data sets have been analyzed by @xcite using the @xmath26 velocity width parameter described in appendix [ sc : vdefns ] .
the courteau sample consists of @xmath6300 sb sc galaxies with @xmath63 , typically brighter than @xmath64 ( using courteau s extinction corrections ) and pruned by eye to eliminate peculiar and interacting galaxies @xcite .
the mfb sample includes @xmath6950 sb sd galaxies with @xmath14 , typically brighter than @xmath64 ( assuming @xmath65 colors of @xmath60.5 mag and using our standard extinction corrections ) .
figure [ fg : cmath ] plots the tfrs for two subsamples of the nfgs defined to match the courteau and mfb selection criteria . except for the exclusion of sa sab galaxies , the sb
sd sample defined to match the mfb selection criteria is identical to the spiral sample analyzed throughout
[ sc : spiralcal][sc : spiral3 ] .
the left panels show courteau s tf fits for the mfb and courteau samples overlaid on our matching subsample data points and fits . here
we use courteau s analysis techniques , including forward fits ( minimizing residuals in m@xmath66 ) , the @xmath26 velocity width parameter , and courteau s luminosity - independent extinction corrections . for the 15 nfgs galaxies matching the courteau selection criteria , the biweight scatter of 0.21 mag is lower than courteau s 0.46 , probably fortuitously .
the apparent zero point offset may also represent small - number statistics , especially since no significant zero point shift is apparent between the nfgs and mfb samples.0.40.5 mag from @xcite , in the sense that the pairs sample is brighter than the reference sample .
given the likelihood that many of the barton et al .
galaxies have experienced some luminosity enhancement due to interactions , the result measured with respect to tully & pierce seems more likely to be correct , in which case the courteau zero point would have to be too bright ( as we may be seeing here ) . ] for the 54 nfgs galaxies matching the mfb selection criteria , the biweight scatter of 0.63 mag is quite close to the value courteau measures from the mfb data , 0.56 .
the slightly lower mfb scatter most likely reflects smaller measurement errors ; we return to this point below .
the right panels of figure [ fg : cmath ] demonstrate the effect of switching to our standard analysis techniques .
we use unweighted inverse fits ( minimizing residuals in velocity width ) , the w@xmath38 velocity parameter , and luminosity - dependent extinction corrections @xcite . with these conventions ,
the slope of the tfr steepens considerably , and the measured scatter about the fit increases , despite the smaller scatter perceived by eye .
the eye perceives the scatter in velocity width , whereas we measure the scatter in absolute magnitudes : the scatter in absolute magnitudes increases as a result of the steeper slope of the tfr .
this steeper slope is due in roughly equal measure to the use of inverse fits and to the use of luminosity - dependent extinction corrections .
another factor affecting scatter is measurement errors .
in fact , the good agreement between our scatter and the scatter in the mfb sample should really be taken as evidence that the measurement errors in the two data sets are very similar .
for the nfgs subsample defined by mfb s criteria and analyzed using our standard techniques ( lower right panel of figure [ fg : cmath ] ) , we calculate that 0.60 mag of the observed 0.71 mag of scatter arises from measurement errors , leaving only @xmath60.38 mag of intrinsic scatter after subtraction in quadrature .
the nfgs subsample error budget includes four contributions added in quadrature : photometry errors and hubble - law distance uncertainties ( 0.18 mag , dominated by the contribution of the peculiar velocity field ) ; photometric inclination errors ( 0.42 mag , dominated by the one - decimal place precision of the ugc axial ratios ) ; velocity width uncertainties ( 0.38 mag , dominated by intrinsic scatter in the correspondence between `` true '' rotation velocities and measured optical velocity widths , regardless of high - quality rotation curves ) ; and extinction correction errors ( 0.09 mag , derived from the uncertainties in inclination and velocity width ) . for the mfb sample
, inclination errors should be somewhat smaller .
mfb galaxies are generally larger on the sky than nfgs galaxies , so assuming that the photometric inclinations for the mfb sample were computed using axial ratios with round - off precision at least comparable to that of the ugc ( and with at least comparable accuracy ) , the mfb inclinations will almost certainly be better than ours .
if we ascribe the ( slight ) scatter difference between the mfb and nfgs samples entirely to the difference in inclination errors , then we infer mfb inclination errors of @xmath60.3 mag . except for its restriction to a single environment , the ursa major sample of ( * ? ? ?
* hereafter vs ) is almost as broadly representative as the nfgs . like the nfgs ,
the vs sample is morphology - blind , including both sa galaxies and peculiar and interacting galaxies .
here we restrict both samples to m@xmath67 , type sa
sd , and @xmath68
. within these limits vs s hi data set is nearly complete , with only one sa galaxy missing .
figure [ fg : verhfigscatt ] directly compares the nfgs and vs spiral samples . with distance scales , extinction corrections , and velocity width definitions
matched as described in the figure caption , we detect no zero point offset within the errors .
the slope of the vs sample appears considerably shallower , but inspection of the data reveals that the behavior of a few galaxies at the faint end of the tfr accounts entirely for the difference . in
[ sc : umdwarfs ] we discuss possible reasons for the divergence of the two samples at the faint end .
the sources of measurement error for the ursa major sample are qualitatively different than for the nfgs , so we must compare intrinsic scatter values , formed by subtracting measurement - error scatter from observed scatter in quadrature .
we predict the measurement - error scatter for vs using the same code with which we propagate our own errors , which yields a total predicted r - band scatter of 0.24 mag from four sources : hybrid kinematic - photometric inclination errors ( 0.09 mag ) ; hi linewidth errors ( 0.13 mag ) ; distance uncertainties from cluster depth effects ( 0.17 mag , as estimated by * ? ? ?
* ) ; and errors in photometry and extinction corrections ( 0.06 mag ) .
these numbers are based on the inclination and hi linewidth errors given by vs combined with basic photometry errors of @xmath60.05 mag @xcite . because the nfgs error budget is dominated by inclination errors ,
it is worth mentioning that vs s inclination errors are smaller than ours not so much because they sometimes use kinematic inclinations , as because their photometric inclinations are better than ours .
our inclination errors would probably be comparable to theirs if our galaxies were equally large on the sky : the one - decimal - place precision of the ugc adds greater round - off error for galaxies of smaller angular diameter . to illustrate this
, we have recomputed photometric inclinations for the ursa major sample using ugc axial ratios and the method of [ sc : vcorrs ] . substituting these inclinations for vs s preferred inclinations
only slightly increases tf scatter . after subtracting measurement - error scatter in quadrature
, we obtain very similar intrinsic r - band scatter values for the nfgs and the ursa major sample : 0.42 mag and 0.37 mag .
similar agreement is found in the b band . recomputing the nfgs scatter and errors about the shallower slope of the ursa major sample ,
we find even better agreement , but if the nfgs velocity width errors are slightly overestimated ( [ sc : basiccal ] ) then the intrinsic nfgs scatter about the ursa major slope may still be as high as @xmath60.42 mag .
the moderately higher intrinsic scatter of the nfgs compared to ursa major should not be at all surprising , as the nfgs includes a wide range of environments and star formation histories .
apart from the hint of a systematic offset at the faint end mentioned previously , the two samples are in excellent agreement .
other cluster tf studies quote lower scatter than we find for the complete ursa major sample .
for example , combining data from multiple clusters ( including ursa major , fornax , coma , the pisces filament , and several others ) , @xcite measure an r - band scatter of 0.34 mag .
@xcite obtain the same result with a partially overlapping data set and similar analysis techniques .
however , directly comparing the nfgs spiral sample with these multi - cluster samples would be misleading , for several reasons .
first , each individual cluster in the multi - cluster samples obeys a different definition of `` complete '' and satisfies that definition to a different degree .
for example , the `` complete '' fornax sample rejects interacting or disturbed galaxies and multiple systems @xcite .
second , each cluster represents a different environment with a different star formation history , so each cluster s color magnitude relation may have a different color zero point ( e.g. tully & pierce find @xmath69 color zero point offsets of @xmath60.1 mag when they compare cluster color magnitude relations ; see also * ? ? ?
we will see in [ sc : color ] that if real , such color offsets almost certainly imply tf zero point offsets between the clusters . however , the tf scatter due to these tf zero point offsets is suppressed in the multi - cluster analyses to the extent that the individual cluster tfrs are allowed to shift freely to minimize zero point offsets ( the absolute zero point is generally determined separately using the cepheid calibrator galaxies ) .
a third concern is the effect of a top - heavy luminosity distribution : in combining more distant clusters with more nearby ones , the multi - cluster studies define samples that statistically favor bright galaxies .
this bias may drive down scatter if fainter galaxies have higher scatter . while the existence of such a trend is unclear in our restricted spiral sample ,
dwarf galaxies definitely show higher scatter in the full sample ( [ sc : fulltfr ] ) .
we have shown that our tf results are consistent with previous studies if we reproduce their sample selection criteria .
these criteria often exclude sa galaxies and/or galaxies with morphological peculiarities . here
, we begin the process of analyzing tf scatter for a broad spiral sample by turning the spotlight on sa and peculiar galaxies , highlighting where these galaxies tf residuals fall within the general spiral tfr . for this analysis
we return to the full nfgs spiral tf sample : types sa sd brighter than m@xmath5 and inclined by @xmath7040@xmath22 .
fourteen of the 68 galaxies in the nfgs spiral tf sample are sa
sab galaxies , collectively referred to as `` sa '' galaxies from here on .
as a group , these 14 galaxies sit clearly to one side of the tfr ( figure [ fg : sa ] ) , with offsets toward lower @xmath71/higher @xmath72 of 0.76 mag at r , 0.95 mag at b , and 1.20 mag at u as compared to sb sd galaxies ( where both sa and sb sd offsets are computed with respect to the unweighted inverse tf fits in table 1 , with relative offset uncertainties of @xmath60.17 mag ) .
the possibility that sa galaxies rotate faster at a given luminosity was first suggested by @xcite based on 21 cm data .
sa offsets became controversial after @xcite claimed that the sa tfr lies 2 mag below the sc tfr in the b band .
other studies reported smaller offsets ( e.g. * ? ? ?
@xcite argue that a greater intrinsic spread of luminosities for late - type vs. early - type galaxies at a given velocity width , combined with malmquist bias , may explain the large offset found by @xcite .
however , sa offsets continue to be observed in increasingly complete samples ( e.g. * ? ? ?
* see [ sc : color ] ) .
the scatter properties of our data suggest that the offset we observe is mostly real . in the r band , the sa galaxies and the sb sd galaxies have biweight tf scatter values of 0.63 mag and 0.70 mag , respectively .
these numbers imply that malmquist bias is unlikely to produce the observed 0.76 mag offset between the two populations . in the b and u bands ,
the scatter in each population increases but the offset increases even more .
the offset exceeds the scatter by 0.3 mag at u. we have investigated whether the sa offset in our data may be ascribed to systematic errors in luminosity or velocity width measurements .
for example , large bulges could make sa galaxies appear rounder ( and more face on ) than they really are , leading to underestimated extinction corrections and overestimated @xmath40 corrections .
sa s in the present sample do have more face - on inclination estimates on average ( 61@xmath22 vs. 66@xmath22 for the full spiral tf sample ) ; however , this difference can account for an offset of only @xmath60.2 mag at a tf slope of @xmath810 .
furthermore , even extreme assumptions about axial ratios ( e.g. assigning an intrinsic axial ratio of @xmath73 to sa s and assigning @xmath74 to all other types ) reduce the r - band offset by at most half and have much less effect in the b and u bands .
no other aspect of our analysis is responsible for the offset either .
whether we adopt luminosity - dependent or independent extinction corrections has negligible effect on our results .
we do not add an explicit morphological type dependence to our extinction corrections , but in any case adding such a dependence would probably only increase the sa offset : extinction is likely to be lower in sa galaxies @xcite .
the offset is not appreciably affected when we substitute @xmath26 for @xmath28 or when we omit the optical - to - radio conversion . finally , the available
hi data also seem to confirm the offset ( figure [ fg : sa ] ) , although these data are subject to incompleteness and high scatter for the sa population .
we conclude that the bulk of the sa offset in our sample is real . however , with just 14 sa galaxies , our analysis is subject to small number statistics .
formally , a kolmogorov - smirnov test yields a probability of @xmath75 that the r - band tf residuals of the sa and sb sd subpopulations were drawn from the same parent population ( @xmath63.5@xmath76 ) .
a possible physical explanation for the sa offset and a way to remove the offset to reduce scatter will be discussed in [ sc : color ] .
tf samples are frequently pruned to remove morphologically peculiar galaxies and sometimes also to remove barred galaxies . here
we use the term `` peculiar galaxies '' to mean disturbed spiral galaxies ( i.e. those with warps , tidal features , multiple nuclei , polar rings , or interacting neighbors ) rather than galaxies that can not be reliably classified as spirals .
figure [ fg : pecgallery ] shows rotation curves and images for three nfgs galaxies we have identified as morphologically peculiar . to avoid bias , bars and morphological peculiarities
were independently identified by two of us ( s.k . and m.f . )
without reference to tf residuals , and the identifications were further checked against the notes provided by r.a .
jansen in @xcite .
although a number of weak bars and peculiarities were noted , we report here only strong , unambiguous cases identified by more than one observer .
such cases are broadly distributed across a wide range of spiral types and luminosities .
figure [ fg : mfbprune]a uses two different ( possibly overlapping ) symbols to identify peculiar and barred galaxies within the nfgs tfr .
although the barred galaxies that are _ not _ peculiar do not show a clear systematic offset , the peculiar galaxies as a group tend to lie on the high @xmath71/low @xmath72 side of the tfr .
this pattern suggests that pruning such galaxies may obscure interesting physics hidden in tf scatter . even when the goal is scatter reduction , pruning may not be the most effective strategy .
figure [ fg : mfbprune]c demonstrates that pruning barred and peculiar galaxies from the nfgs spiral sample reduces r - band tf scatter to approximately the level expected from measurement errors . in bluer passbands , however , the reduced scatter still exceeds the level expected from measurement errors : we measure 0.63 mag and 0.71 mag in b and u , respectively , compared to a predicted measurement - error scatter of 0.58 mag in both passbands .
also , the decrease in scatter won by pruning comes at a high price , as @xmath7030% of the sample must be rejected . in [ sc : spiral3 ] we will use more objective , physically motivated techniques to reduce tf scatter to a level comparable to measurement - error scatter in all passbands without discarding galaxies from the sample .
we have also considered independent definitions of peculiarity involving kinematic data : figure [ fg : mfbprune]b identifies galaxies with truncated rotation curves ( @xmath77 , [ sc : vcalcs ] ) , rotation curve asymmetries greater than 5% ( our rc asymmetry index is defined in appendix [ sc : asymmeast ] ) , or discrepancies between w@xmath78 and w@xmath38 greater than 10% .
interestingly , galaxies with high rc asymmetry behave much like those with strong morphological peculiarities , with high @xmath71/low @xmath72 offsets .
figures [ fg : mfbprune]c and [ fg : mfbprune]d show that pruning based on either morphology or kinematics produces a comparable reduction in scatter , although kinematic pruning misses some sa outliers . to test whether our definitions of kinematic peculiarity have physical significance or merely identify faulty rotation curves , we have tried pruning the hi linewidth tfr using kinematic peculiarity information obtained from our optical rcs .
amazingly , optical rc peculiarities are at least as effective as morphological peculiarities in flagging outliers in the hi linewidth tfr .
this result implies that either ( 1 ) optical rc peculiarities are associated with observational errors in the nfgs surface photometry or the ugc - derived photometric inclinations common to both the hi and optical rc tfrs , or ( 2 ) optical rc peculiarities occur in galaxies that lie off the tfr for physical reasons .
we find evidence for the latter in [ sc : spiral3 ] .
we have seen that interacting , merging , and peculiar galaxies often lie on the high @xmath71/low @xmath72 side of the tfr , while sa galaxies lie on the low @xmath71/high @xmath72 side ( [ sc : pecandsa ] ) .
figure [ fg : tfr ] shows that in addition , nfgs galaxies brighter than m@xmath79 display asymmetric scatter about the best - fit tfr for fainter galaxies , with a systematic offset of @xmath60.7 mag toward lower @xmath71/higher @xmath72 . below
we demonstrate that all of these offsets reflect underlying correlations between tf residuals and quantitative galaxy properties , most notably @xmath3 color and h@xmath4 equivalent width ( ew ) .
following tradition , we refer to these properties as `` third parameters , '' i.e. additional variables that control physical scatter in the two - parameter tfr .
@xcite demonstrate that false correlations between tf residuals and candidate third parameters may arise when the tf slope is measured incorrectly , in which case any parameter that varies along rather than perpendicularly to the tfr will produce a correlation .
a true third parameter should vary at least partly perpendicularly to the tfr .
we adopt the following strategy to avoid false detections of third parameters .
first , we apply strict cuts in m@xmath80 to eliminate sections of the data where the velocity width scatter is clearly asymmetric . for the r band ,
these cuts are indicated by the dashed lines in figure [ fg : tfr ] ( see also table ): the upper line excludes galaxies brighter than m@xmath79 , which show a systematic offset as discussed above , and the lower line excludes galaxies that do not meet the m@xmath81 sample definition criterion . after these cuts ,
the data are biased in m@xmath80 but nearly perfectly unbiased in velocity width .
second , we apply an inverse tf fit to the data , which minimizes residuals in velocity width . at this point
the velocity width residuals are completely uncorrelated with luminosity m@xmath80 .
finally , we use spearman rank tests to quantify both the candidate third parameter
tf residual correlation and the candidate third parameter luminosity correlation . for the spearman tests , velocity width residuals and magnitude residuals are interchangeable .
if these tests show that the candidate third parameter correlates with tf residuals but not with luminosity , then we conclude that it varies perpendicularly to the tfr .
llll bright cut & -21 & -21 & -22.5 + standard faint cut & -17 & -17 & -18 + dwarf faint cut & -16 & -16 & -17 + in fact , even if a parameter does vary with luminosity , it may still qualify as a third parameter if it also correlates clearly with tf residuals , because we have in principle eliminated the tf residual luminosity correlation .
nonetheless , one must be wary if the parameter - luminosity correlation is stronger than the parameter
tf residual correlation , as in such a case even small tf slope errors may lead to false correlations .
for luminosity - dependent parameters , one may wish to subtract out the luminosity dependence and measure the correlation between tf residuals and `` parameter residuals , '' i.e. parameter offsets from the fitted parameter luminosity correlation ( cf .
below we give results based on raw parameters , but we have also checked our results using parameter residuals .
the parameter
luminosity correlations are strongest for color and rotation curve asymmetry ( figure [ fg : kandcvslum ] ) ; however , even for these parameters we find that correcting for the luminosity dependence makes little difference , slightly improving the third - parameter correlation if anything .
in general , we use unweighted tf fits to define tf residuals .
unweighted fits avoid the slope bias that would otherwise arise because our errors vary systematically with luminosity and velocity width .
however , we have confirmed all correlations using both weighted and unweighted fits . we have also confirmed all correlations using two definitions of velocity width ( w@xmath78 and w@xmath38 ) and using both luminosity - dependent and luminosity - independent extinction corrections ( [ sc : extinct ] ) .
results quoted in the text use w@xmath38 and the @xcite extinction corrections .
we have tested a wide range of candidate third parameters , making use of photometric and spectrophotometric quantities from the nfgs database @xcite and photometric asymmetries kindly provided by r. jansen ( private communication , see also * ? ? ?
for the luminosity - restricted sa sd sample of figure [ fg : tfr ] , tf residuals correlate strongly with @xmath82 ( effective @xmath3 color measured within the b - band half - light radius ) and global ew(h@xmath4 ) in all bands .
other measures of color and emission - line strength yield somewhat noisier correlations .
we also find correlations with rotation curve asymmetry , photometric asymmetry , and two measures of surface brightness ; these correlations reach @xmath703@xmath76 significance only in certain photometric bands .
tf residuals do not correlate with luminosity ( by construction ) or with inclination , validating our analysis .
likewise , we detect no correlation between tf residuals and rotation curve extent , isophotal radius , effective radius , hi gas mass , gas consumption timescale , or @xmath83 ( we have incomplete data for the last three however , see [ sc : neutral ] ) .
figure [ fg : restspircorr ] displays some of the correlations in the r band and lists spearman rank test results for u , b , and r ( the `` probability of no correlation '' @xmath31 is given , where @xmath84 is the confidence of the result ) .
the symbol coding shows that sa and peculiar galaxies drive most of the statistical signal in the r - band correlations .
if we eliminate both groups of galaxies , no r - band correlation survives with @xmath703@xmath76 confidence .
however , the @xmath82 correlation does survive in b , and both color and emission line correlations remain significant in u. in general , the two groups of galaxies contribute roughly equally to the correlations , with the sa galaxies showing a slightly stronger signal due to lower scatter .
the peculiar galaxies form a more heterogeneous group , including not only interacting , merging , and warped galaxies , but also a few galaxies whose oddities may indicate later evolutionary states , e.g.a bulge - dominated sa galaxy with a prominent polar dust lane ( ngc 984 ) .
although sa and peculiar galaxies dominate the statistical signal we measure , the correlations in figure [ fg : restspircorr ] apply to all galaxies in the sample .
a blue , strongly star - forming galaxy whose image does not appear peculiar nonetheless shows the expected tf residual .
thus the correlations provide a physical , quantitative basis for understanding systematic offsets in the tfr , independent of morphological classification .
furthermore , the correlations are not limited to the optical tfr . despite disproportionate hi catalog incompleteness for bright and peculiar galaxies , mixed hi data quality , and confusion in single - dish linewidths , we still detect 23@xmath76 correlations between radio tf residuals and both @xmath82 and global ew(h@xmath4 ) .
we also find evidence for a tf residual
total color correlation in radio data from the complete ursa major database of @xcite , see [ sc : color ] .
these results emphasize the close link between optical and radio velocity widths already noted in [ sc : vcalcs ] and [ sc : pec ] ( see also appendix [ sc : vdefns ] ) .
however , we do not perform a detailed third - parameter analysis of the radio tfr , as the catalog hi linewidths are incomplete and of mixed quality .
effective @xmath3 color yields a striking correlation with tf residuals , predicting spiral galaxy tf offsets better than any other physical parameter we have tested .
effective @xmath85 color also produces a clear correlation in all three bands . the color tf residual correlations strengthen from r to b to u , as blue peculiar galaxies and red sa galaxies move further apart in the tfr .
although we correct colors for internal extinction ( see [ sc : extinct ] ) , omitting the extinction correction does not significantly affect the correlation strength .
the correlation slope is quite a bit steeper than expected from dust extinction and reddening ( @xmath61.3 in the r band with @xmath3 colors , e.g. * ? ? ?
* table 1 ) .
we stress that the correlation reported here can not be dismissed as merely the `` built - in '' correlation due to the variation of tf scatter with passband .
a weak correlation of u and b - band tf residuals with color might have been expected from the fact that tf scatter is significantly higher in these bands than in the r band , which implies that both @xmath3 and @xmath86 colors may be used to reduce scatter ( most trivially , simply by adding these colors to the u and b magnitudes to recover the r magnitudes ) . in this case
, the color tf residual correlation would have a slope of @xmath61 and there would be no correlation between r - band tf residuals and @xmath3 colors .
however , both the persistence of the correlation in the r band and the steep slope of the correlation point to a more interesting result : we explore this result empirically in this section and discuss its physical significance in
[ sc : causes ] . for our sample ,
effective colors produce much tighter correlations than total colors . in the r band ,
spearman tests yield @xmath87 for the ( @xmath3)@xmath88 correlation , or @xmath64@xmath76 , but only @xmath89 for the total ( @xmath3)@xmath90 correlation , or just over 2@xmath76 .
total colors dilute the correlation signal with observational scatter from profile extrapolation and background subtraction ( r. jansen , private communication ) . however , we see no strong systematic differences between @xmath3 and @xmath82 : on average the two track closely , with @xmath82 slightly redder , by at most @xmath60.1 mag for the reddest galaxies .
the color tf residual correlation offers an alternative to sample pruning for reducing tf scatter .
if we restore spiral galaxies brighter than m@xmath91 to the sample ( they were not used in figure [ fg : restspircorr ] , see figure [ fg : tfr ] ) , then refitting the color tf residual correlation provides a color - correction formula to go with the unweighted inverse tf fit in table . a least - squares forward fit to @xmath20 m vs. @xmath92 yields @xmath93 where @xmath94 is the measured tf residual ( observed magnitude minus fitted relation ) .
if we correct our r - band magnitudes with equation [ eq : colcorr ] and recompute the residuals with respect to the original tf fit , the biweight scatter falls from 0.74 mag to 0.66 mag .
an analogous procedure reduces scatter to 0.64 mag in both the b and u bands , from 0.82 and 1.02 mag respectively , essentially eliminating the differences between wavelength bands ( figure [ fg : scattred ] ) .
the corrected scatter values come quite close to the scatter expected from measurement errors ( table ) .
determining whether these errors hide further non - random behavior in tf scatter would require smaller measurement errors .
full two - dimensional velocity fields , with kinematic inclinations and modeling of disk warps and asymmetries , offer the best hope of progress ( cf . * ? ? ?
* ) . the existence of a strong color
tf residual correlation is not entirely unexpected .
@xcite describe environmental shifts in the b - band tfr for a mixed cluster - field sample that are linked to galaxy colors , which suggests the possibility of a color tf residual correlation .
@xcite demonstrates reduced tf scatter in a two - color tully - fisher relation ( essentially equivalent to our color - corrected tfr , as in figure [ fg : scattred ] , but with the slope of the color tf residual correlation assumed rather than measured ) .
@xcite reduces tf scatter using 60 @xmath95 m data to define `` population corrected '' luminosities .
@xcite present initial evidence for a color tf residual correlation in their intermediate - redshift sample , which spans a wide range of galaxy types ; however , the tf residuals are measured relative to an externally defined tfr , making it harder to eliminate systematic errors ( also a potential problem for @xcite , see @xcite and the discussion in [ sc : thirdpartech ] ) .
most significantly , @xcite reports a @xmath69 color vs. tf residual correlation for a subsample of 21 galaxies drawn from his complete ursa major sample , and our own analysis of his full sa
sd data set indicates that total @xmath3 colors correlate with b - band tf residuals at @xmath633.5@xmath76 confidence ( using unweighted inverse fits , our standard luminosity cuts , and the w@xmath29 and m@xmath96 from * ? ? ?
the r - band correlation for this sample registers at @xmath62@xmath76 , consistent with our expectations given that total rather than effective colors were used .
most other low - redshift tf studies exclude sa and peculiar galaxies and/or use red to near - infrared photometry , making it difficult to observe the color tf residual correlation .
for example , @xcite find only a slight , statistically insignificant color tf residual correlation in a sample of undisturbed sb sc galaxies observed in the r band .
the close galaxy pairs sample of @xcite includes disturbed galaxies but was not designed to evenly sample the range of spiral morphologies ; again no correlation emerges using total colors and r - band tf residuals ( e. barton , private communication ) .
the color tf residual correlation provides a natural explanation for two effects noted previously : the sa galaxy offset ( [ sc : saoffset ] ) and the asymmetric scatter at the bright end of the tfr ( [ sc : spiral3 ] ) .
sa galaxies are typically red and show relatively little scatter in color compared to later type galaxies @xcite . as a consequence , sa galaxies deviate from the tfr as a group , with larger offsets in bluer bands . applying the color correction given in equation [ eq : colcorr ]
reduces the r - band offset between sa and sb sd galaxies from [email protected] to 0.28@xmath97 mag , where we have computed offsets using tf residuals relative to the unweighted inverse fit in table 1 .
the remaining offset is not statistically significant .
similarly , the @xmath60.7 mag offset for galaxies brighter than m@xmath91 shrinks to @xmath60.15 mag following the color correction , suggesting that the asymmetric scatter at the bright end of the tfr may simply reflect the rarity of extremely luminous blue galaxies .
@xcite has also reported a link between color dependence and morphological type dependence in the tf residuals for his pruned ursa major spiral sample .
the sa and sab galaxies are underluminous by @xmath60.5 mag in the b band ( consistent with our result when normalized for slope differences , @xmath86.7 vs. @xmath810 ) , while the same galaxies have normal luminosities at k , suggesting a color effect .
later types show a wide range of positive and negative tf residuals in both verheijen s complete tf sample @xcite and in the nfgs spiral tf sample , consistent with the much greater color dispersion observed for types sb and later in the full nfgs @xcite . we discuss a physical interpretation of the color tf residual correlation in terms of stellar populations in [ sc :
causes ] . global h@xmath4 equivalent width ( integrated over the entire galaxy , see * ? ? ?
* ) correlates with tf residuals for the sa
sd subsample almost as well as @xmath82 ( figure [ fg : restspircorr ] ) .
this result should not be too surprising : the two parameters are highly correlated .
ew(h@xmath4 ) measures recent star formation on timescales of myr relative to integrated past star formation , while @xmath82 measures the ratio between young and old stellar populations and thus traces star formation on timescales of gyr ( e.g. * ? ? ?
as expected , the ew(h@xmath4)tf residual correlation intensifies in bluer tf bands , although not quite so strongly as the color correlation .
substituting either nuclear ( fixed @xmath98 aperture ) h@xmath4 or global [ oii ] 3727 measurements for global h@xmath4 weakens the correlation , but it still reaches @xmath703@xmath76 significance in the b and u bands . restoring spiral galaxies brighter than m@xmath91 to the sample , we can fit an ew(h@xmath4)-correction formula analogous to the color - correction formula in equation [ eq : colcorr ] : @xmath99 for the unweighted inverse fits in table , this formula and its b and u analogues reduce scatter to 0.66 , 0.67 , and 0.78 mag in the r , b , and u bands respectively .
the ew(h@xmath4 ) correction thus performs nearly as well as the color correction in both b and r , although it underperforms in u. to our knowledge , this paper is the first to report a correlation between tf residuals and ew(h@xmath4 ) .
@xcite search unsuccessfully for such a correlation using nuclear ( fixed slit ) h@xmath4 measurements and r - band tf residuals .
these results are consistent , given that ( a ) our sample yields only a 2.9@xmath76 correlation between r - band tf residuals and _ nuclear _ ew(h@xmath4 ) , and ( b ) the barton et al.survey targets a specific population ( galaxies in close pairs ) rather than a broad cross - section of galaxies .
the existence of a correlation between tf residuals and photometric asymmetry has not been firmly established .
@xcite argue that i - band asymmetry and b - band tf residuals are correlated , but @xcite suggest that the bulk of this correlation arises from a systematic error in tf calibration .
we measure no significant correlation of r - band asymmetry with tf residuals in any passband , but we do see a marginal 3@xmath76 correlation between b - band asymmetry and r - band tf residuals , which weakens for b and u - band residuals . we conclude that photometric asymmetry is a weak third parameter at best . since asymmetry arises from multiple causes that may move tf residuals in different directions ( knottiness from dust , unevenness from recent starbursts or accretion , etc . ) , sample selection may partly determine whether or not a correlation is detected .
rotation curve asymmetry ( defined in appendix [ sc : asymmeast ] ) correlates with tf residuals in the b and u bands with @xmath63@xmath76 significance .
the rc asymmetry vs. tf residual plot ( figure [ fg : restspircorr ] ) may also indicate a discrete effect : galaxies with asymmetry less than @xmath62% show lower scatter and a slight offset relative to those with higher asymmetry .
@xcite also report a discrete effect in their close pairs sample : the r - band tf residuals of galaxies identified as severely kinematically distorted differ from the residuals for the remaining galaxies by 3@xmath76 in a k - s test .
we return to the possible link between kinematic distortions and tf offsets in [ sc : voffs ] .
@xcite argue against any surface brightness dependence in the tf residuals of high surface brightness spiral galaxies , while @xcite demonstrate that extreme low surface brightness ( lsb ) galaxies can lie well off the tfr . in our sample
we find no clear correlation in any passband with the most common measure of surface brightness used in the literature , the extrapolated disk central surface brightness ( as measured from exponential profile fits , r. jansen , private communication ) . on the other hand
, we do detect marginal correlations between tf residuals and two other measures of surface brightness . the observed r - band central surface brightness ( measured from raw profiles including both disk and bulge light , and subject to uncertainties due to both seeing and the distance to the galaxy ) correlates with u and b - band tf residuals at @xmath703@xmath76 , but the detection has high uncertainty due to a strong central surface brightness luminosity correlation ( which causes the third parameter correlation to weaken if weighted tf fits are used , see [ sc : thirdpartech ] ) , and it disappears for the full sample . a more robust quantity , the effective surface brightness in the u band ( @xmath100 , the surface brightness measured _ at _
@xmath27 ) displays a 3@xmath76 correlation with r - band tf residuals , but this correlation fades in the b and u bands ( figure [ fg : restspircorr ] ) .
however , the correlation becomes stronger for the full sample ; we discuss it in [ sc : sbeu ] .
the present spiral sample shows no correlation between tf residuals and neutral gas properties such as @xmath101 , gas consumption timescale fluxes ( r. jansen , private communication ) using the calibration of @xcite with no correction for recycling . ] , and @xmath83 , where we determine @xmath71 in the same passband as the tf residuals to eliminate any correlation due to color .
however , we compute gas masses from the catalogs of @xcite and @xcite , and these catalogs do not provide hi masses for @xmath631% of the sa
sd subsample .
the incompleteness preferentially affects sa and peculiar galaxies , which have the largest tf offsets , so our results should be treated with caution .
nonetheless , our failure to detect an @xmath101tf residual correlation is interesting in light of the strength of the @xmath102luminosity correlation . for the sa
sd subsample , the @xmath102luminosity correlation rivals the tfr itself in correlation strength : although weaker at b and r , it is actually stronger at u , with a spearman rank probability of no correlation of @xmath103 .
therefore even a small tf slope error could easily generate a false correlation between tf residuals and hi mass in this sample . the fact that we detect nothing validates our fitting and analysis procedures , as hi masses and tf velocity widths are measured completely independently .
furthermore , since the color
luminosity correlation is much weaker than the hi mass luminosity correlation in our data , this argument supports our claim that slope errors do not drive the strong correlation we observe between color and tf residuals .
the existence of a well - defined tully - fisher relation for spiral galaxies suggests that for large disk galaxies , the total mass of the system ( traced by its rotation velocity ) closely couples to the mass of its stars ( traced by their luminosity ) .
tf residuals may arise from any of three basic relations underlying the overall relation : ( 1 ) the relation between observed rotation velocity and total mass , ( 2 ) the relation between total mass and stellar mass , and ( 3 ) the relation between stellar mass and observed luminosity .
below , we examine the possibility that variations in stellar @xmath104 due to differences in stellar populations can fully explain the color tf residual correlation , with tf offsets interpreted as luminosity offsets .
we then discuss alternative explanations for the color tf residual correlation that interpret tf offsets as velocity width offsets that correlate with color .
variations in stellar @xmath104 due to differences in stellar populations provide the simplest explanation of tf scatter ( e.g. , * ? ? ?
scatter driven by population differences should be higher in bluer passbands , where variations in @xmath104 are greater , and indeed tf scatter increases from r to b to u. if there is negligible scatter in the correlation between rotation velocity and stellar mass ( i.e. , the combination of relations ( 1 ) and ( 2 ) above ) , then the color tf residual correlation follows directly from the dependence of color on @xmath104 , provided that color variations are not dominated by the systematic correlation between color and magnitude ( in agreement with our observations , see [ sc : color ] and figure [ fg : kandcvslum ] ) .
the only caveat is that different star formation histories may yield different ratios of luminosity evolution to color evolution ( @xmath105 ) , producing different tf offsets for a given color change .
thus we can test the hypothesis that stellar population differences drive the color tf residual correlation by comparing the slope of the correlation to the @xmath105 predictions of population synthesis models designed to describe the general population of spiral galaxies .
we perform this comparison using the model predictions of @xcite .
these authors use population synthesis models to evaluate the slope of the color vs. stellar @xmath104 correlation for passively evolving disk galaxies subject to a variety of star formation histories in which stars form in a reasonably smooth fashion with no major bursts .
these slope predictions may be compared to the measured slopes shown in the top two panels of figure [ fg : restspircorr ] .
the top right panel shows the color _
residual_tf residual correlation , where color residuals are defined relative to the color magnitude relation
; this relation offers the most direct measurement of @xmath105 , suitable for comparison with a simple model such as a closed box model .
( one might prefer to define color residuals relative to the color - velocity width relation ; however , because our sample is defined by strict limits in magnitude rather than in velocity width , the color - velocity width relation would be biased . ) the standard color tf residual correlation , uncorrected for the color magnitude relation , may correspond better with models employing mass - dependent star formation histories . depending on the particular star formation history used , bell & de jong s predicted slope varies from 1.82.5 for r - band tf residuals and @xmath3 colors , with a closed box model giving a slope of 2.3 .
for the observed correlation , a forward least - squares fit yields a slope of 2.2(2.5 ) @xmath106 for the uncorrected(corrected ) correlation , while a least - squares bisector fit yields 3.5(3.9 ) @xmath106 ( cf . * ? ? ?
the latter fit technique is more realistic as it assumes intrinsic scatter in both fit variables ; slopes measured by this technique are broadly consistent with the bell & de jong predictions but appear to favor slightly steeper values .
slopes measured in the b band are also somewhat steep : e.g. while the closed box model gives a slope of 3.3 in b , we derive values of 3.2 and 4.4 using forward and least - squares bisector fits , respectively , for the color residual tf residual correlation .
the basic agreement between bell & de jong s predictions and our observations supports our interpretation of the data in terms of stellar populations . with this interpretation
we find a single passband - independent stellar - mass tfr ( i.e. , our `` color - corrected '' tfr , [ sc : color ] ) with lower scatter than the conventional tfr and a form very similar to that predicted by bell & de jong : @xmath107 . to the extent that there is any discrepancy between observed and predicted @xmath105 ratios
, the data would appear to support slightly steeper slopes than the models .
this result is largely insensitive to the choice of extinction corrections , even when we omit such corrections altogether , although it is possible to push the observations closer to the models with selected analysis techniques ( e.g. adopting the type - dependent @xmath108 described in [ sc : saoffset ] and using colors uncorrected for the color
magnitude relation ) .
if real , the steeper slopes might reflect several effects : ( 1 ) variations in dark matter fraction or dark matter structure that correlate with color ; ( 2 ) systematic velocity width offsets due to inclination errors or rotation curve distortions that correlate with color ; or ( 3 ) star formation histories that include recent starburst activity ( @xmath241 gyr ago ) , which would give @xmath105 ratios @xmath1095 ( e.g. , bell & de jong figure 5 ) . in support of option
( 3 ) , many nfgs galaxies show morphological and kinematic evidence of disturbance ( [ sc : pec ] ) , suggesting the likelihood of starbursts driven by interactions , mergers , or global instabilities .
we discuss options ( 1 ) and ( 2 ) in the next section .
we have also analyzed the color tf residual correlation for one other complete tf sample , the ursa major sample of @xcite , and we find much shallower slopes ( @xmath61 in r , @xmath62 in b ) .
these slopes are even shallower than expected purely from population synthesis .
as the ursa major cluster represents a uniform moderate - density environment , the difference may be related to a greater homogeneity of star formation histories compared to what we see in the general field . in this case
the tf residuals in the color tf residual correlation might have less to do with luminosity offsets related to stellar populations and more to do with color - dependent velocity width offsets driven by one of the mechanisms described in [ sc : voffs ] below .
tf residuals are very sensitive to small offsets in velocity width because of the steep slope of the tfr . therefore any systematic velocity width trends in our data must correlate tightly with color , given the success of the color correction in reducing scatter to near measurement - error levels in all bands ( [ sc : color ] ) .
possible sources of velocity width residuals include variations in dark matter fraction or dark matter structure , systematic errors in photometric inclination , and symmetric or asymmetric distortions in rotation curves .
if any of these correlates strongly with color , it may offer an alternative explanation for the color tf residual correlation in terms of velocity width offsets rather than luminosity offsets .
a full consideration of dark matter structure is beyond the scope of this paper , but we can address one plausible scenario with straightforward observable consequences .
halo contraction in bulge - dominated galaxies could lead to increased rotation velocities within the central parts of the galaxies , in turn driving a correlation between color and velocity width residuals insofar as higher bulge - to - disk ratios correlate with redder colors .
however , the high color dispersion for later morphological types @xcite suggests a relatively weak link between colors and bulge - to - disk ratios , except for early types such as sa galaxies .
furthermore , the observed rcs for our sa galaxies do not decrease in velocity in their outer parts , so we have no independent evidence that these galaxies high central concentrations have in fact increased their measured rotation velocities . extended hi data
might reveal such an effect .
alternatively , velocity width offsets might be driven by photometric inclination errors that correlate with color .
a color inclination error correlation could arise indirectly , via a color - morphology connection involving either bulge - to - disk ratios or morphological peculiarities , both of which can affect photometric inclination estimates .
we can not directly test this possibility without kinematic inclinations .
however , raising the inclination cut from @xmath14 to @xmath110 to reduce inclination - related offsets simply tightens the color tf residual correlation , reducing its scatter without significantly changing its slope ( in the r band , the range of acceptable slopes narrows to 2.53.4 ) .
furthermore , we have argued that inclination errors can not explain the sa galaxy offset , although they may contribute to it ( [ sc : saoffset ] ) , and that the link between colors and bulge - to - disk ratios is probably fairly weak .
it is also hard to imagine how a heterogeneous population of peculiar galaxies could yield consistently high inclinations , as required to explain their high @xmath71/low @xmath72 offsets.21.5 in the color - corrected u - band tfr , figure [ fg : scattred ] ( also shown in the top panel of figure [ fg : pecgallery ] ) . its nominal photometric inclination is 42 degrees , but its inner parts appear closer to face on . if its true inclination were 20 degrees it would shift by @xmath62.8 mag , bringing it much more in line with the tfr .
however , ngc 5993 is not at all typical : its offset does _ not _ follow the color tf residual correlation .
in fact , on average , spiral galaxies with peculiarities have slightly lower apparent inclination than the rest of the spiral sample ( 63@xmath22 vs.66@xmath22 ) . ] systematically high random errors for peculiar galaxies would produce a net offset in the wrong direction , due to the nonlinearity of the sine function .
rotation curve asymmetries can not explain the color tf residual correlation either , although these asymmetries do correlate weakly with both tf residuals and @xmath82 ( see [ sc : rcasym ] ) . even supposing that a 5% asymmetry produces a 5% offset in rotation velocity ( certainly an overestimate ) , that velocity offset will in turn generate a tf residual of only @xmath60.2 mag .
relatively few galaxies have asymmetries that large ( see figures [ fg : mfbprune ] and [ fg : restspircorr ] ) .
symmetric rotation curve distortions could be more important , if these distortions correlate with color .
2d velocity fields would be required to evaluate their role . @xcite
have shown that intrinsically elongated disks can produce symmetric kinematic distortions , but random orientations to the line of sight will lead primarily to scatter rather than to a systematic correlation .
galaxy interactions could create distortions that correlate with blue color , but a mechanism for shifting both radio and optical velocity widths in the same direction is not obvious .
@xcite argue that tidal interactions broaden hi linewidths , which would actually weaken the color tf residual correlation for the hi tfr , as would hi confusion . technically
, these authors observations do not distinguish between broadening of hi linewidths and narrowing of co linewidths , which is likely to occur as molecular gas concentrates at the center of a galaxy where it does not sample the full velocity field @xcite .
ionized gas is also likely to concentrate at the center for starburst galaxies , yielding low velocity width tf outliers due to rotation curve truncation ( * ? ? ?
* ; * ? ? ?
* see also [ sc : allmorphthirdpars ] ) . while this effect may create offsets for a few galaxies
, it does not contribute to systematic trends : tf residuals show no significant correlation with rotation curve extent .
we conclude that the color tf residual correlation is most easily explained as being fundamentally a color luminosity residual correlation , although it may be steepened by a secondary color velocity width residual correlation .
we now turn to the full tf sample of nfgs emission - line galaxies , with no restrictions on luminosity or morphology , but still requiring @xmath14 .
table gives fitted tf parameters for this sample in the u , b , and r bands , with optical rc results based on 108 galaxies ( 107 at u ) and hi linewidth results based on 76 galaxies ( 75 at u ) .
the r - band tfr is shown in figure [ fg : allgalsone ] .
rrrrr + u & [email protected] & [email protected] & [email protected] & [email protected] + b & [email protected] & [email protected] & [email protected] & [email protected] + r & [email protected] & [email protected] & [email protected] & [email protected] + + u & [email protected] & [email protected] & [email protected] & [email protected] + b & [email protected] & [email protected] & [email protected] & [email protected] + r & [email protected] & [email protected] & [email protected] & [email protected] + + u & 1.26(0.78 ) & 1.05(0.68 ) & 1.19(0.75 ) & 1.09(0.53 ) + b & 1.20(0.79 ) & 1.02(0.70 ) & 1.11(0.75 ) & 1.03(0.53 ) +
r & 1.16(0.81 ) & 1.02(0.73 ) & 1.07(0.76 ) & 1.00(0.53 ) + as illustrated in figure [ fg : addback ] , scatter is dramatically higher for the full sample than for the spiral sample of [ sc : spiralcal][sc : spiral3 ] . including non - spiral and dwarf galaxies adds 0.77 mag of scatter in quadrature , yielding a total scatter of 1.07 mag .
faint non - spiral galaxies are responsible for most of the new scatter , so the relative importance of the morphology and luminosity extensions is unclear . the two e / s0 s brighter than m@xmath111 behave much like sa galaxies , with a slight low @xmath71/high @xmath72 offset .
@xcite observe a similar tf offset for s0 galaxies using stellar kinematics .
we begin by broadening the spiral sample of [ sc : spiral3 ] to include all morphologies , staying within the luminosity limits used in [ sc : spiral3 ] ( table and figure [ fg : allgalsone ] ) and as usual requiring @xmath14 . in [ sc : dwarfthirdpars ] we extend the luminosity range 1 mag fainter . adding non - spirals brighter than m@xmath5 to the spiral sample
strengthens the correlations between tf residuals and both rotation curve asymmetry and nuclear ew(h@xmath4 ) .
we also see a hint of a new correlation between tf residuals and gas consumption timescale ( based on incomplete data , as discussed in [ sc : neutral ] , but with 88% completeness for the newly added non - spiral galaxies ) .
the tf residual@xmath100 correlation remains about the same .
however , the expanded sample displays _ weaker _ correlations between tf residuals and both effective @xmath3 color and global ( as opposed to nuclear ) ew(h@xmath4 ) .
as shown in figure [ fg : allcorrstdlumcuts ] , the @xmath82 correlation actually falls below 3@xmath76 at r. although most non - spiral galaxies , marked with circled and boxed points in figure [ fg : allcorrstdlumcuts ] , follow the correlation established by spiral galaxies quite well ( albeit with slightly larger scatter ) , three of the six emission - line e / s0 galaxies in the sample are extreme outliers .
these three outliers , labeled by name in figure [ fg : allcorrstdlumcuts ] , have red colors @xmath1091.2 and tf residuals @xmath701 mag toward high @xmath71/low @xmath72 .
hi linewidths are available for two of the three s0 outliers ; these two galaxies are clear outliers from the hi tfr as well , although the hi residuals are smaller than the optical rc residuals .
the large tf residuals for these three s0 galaxies suggest that our observations do not accurately reflect their disk rotation velocities and/or their underlying luminosities . in each case
the data are likely to be influenced by central activity , in the form of either an agn or a central starburst . however , the central luminosity enhancements are too small to account for the tf offsets .
inclination errors are also unlikely to explain the outliers , as all three galaxies show disky morphologies and inclinations of @xmath660 - 70 degrees .
we suspect kinematic effects .
ngc 4117 is a seyfert ii and has a truncated rotation curve that extends to only 0.56@xmath27 ; its stellar kinematics are peculiar and may indicate gas - stellar decoupling .
a11332@xmath1123536 hosts a central starburst and has a truncated rotation curve that extends to 0.92@xmath27 ; the gas seems to lie in an inclined counterrotating disk @xcite .
a12001@xmath1126439 has an unusually bright nucleus and its rotation curve reaches only 1.31@xmath27 ; the gas displays velocity reversals . for the latter two galaxies hi masses are available , from which we calculate rapid gas consumption timescales of @xmath62 gyr .
the irregular kinematics and starburst / agn activity found in these s0 galaxies may result from recent interactions or mergers .
as discussed by @xcite , interacting galaxies may display radially truncated emission due to centrally concentrated starbursts , leading to artificially low velocity widths .
interactions and mergers can also generate non - coplanar gas and stellar rotation in early - type galaxies @xcite , in which case artificially low gas velocities will be measured because the spectrograph slit p.a .
is aligned to the stellar major axis rather than to the gas major axis .
deviations from the color tf residual correlation provide a convenient and physically motivated way to identify such galaxies and prune them from a tf sample . figure [ fg : fullsampcorr ] plots third parameter tf residual correlations for the full tf sample , with luminosity cuts extended to include dwarf galaxies down to m@xmath113 ( see table ) .
adding these dwarfs dilutes the third - parameter correlations for @xmath82 and ew(h@xmath4 ) severely ; for r - band tf residuals they become scatterplots .
the only third - parameter correlations that reach 3@xmath76 significance in the r band involve parameters that showed weak or nonexistent correlations for the spiral sample : rotation curve asymmetry , u - band effective surface brightness , and gas consumption timescale ( see [ sc : rcasym][sc : neutral ] ) .
because these correlations are highly scattered , they are more easily understood in terms of bulk differences in properties for galaxies on either side of the tfr , rather than in terms of a linear trend like the color tf residual correlation .
figure [ fg : twodwarfs ] illustrates these bulk differences with separate histograms for galaxies on either side of the tfr , considering only galaxies fainter than m@xmath114 since these galaxies are the primary drivers of the new correlations .
( unlike figure [ fg : fullsampcorr ] , figure [ fg : twodwarfs ] excludes galaxies with truncated rotation curves , which might appear on the low @xmath72 side of the tfr erroneously ; however , figure [ fg : twodwarfs ] applies no lower limit on luminosity . ) in what follows we refer to galaxies fainter than m@xmath114 as `` dwarfs . ''
figure [ fg : addback ] shows that the non - spiral morphologies typical of dwarfs are most common below m@xmath115 . asymmetries in rotation curves arise from multiple sources including turbulence , bars , tidal distortions , early or late - stage infall , and satellite accretion .
most of these underlying phenomena are likely to enhance star formation , qualitatively explaining the sense of the correlation in figure [ fg : fullsampcorr ] as well as its stronger statistical signal in bluer passbands .
however , the scatter and substructure in the correlation suggest that the rc asymmetry star formation connection may be complex . on the the high @xmath71/low @xmath72 side of the tfr , several dwarf galaxies show rc asymmetries @xmath10910% ( much higher than expected from the rc asymmetry luminosity relation in figure [ fg : kandcvslum ] ) . on the other side of the tfr , rc asymmetries are generally lower and more uniform ( figure [ fg : twodwarfs ] ; see also [ sc : dwarfreal ] ) . if u - band effective surface brightness ( @xmath100 ) measures the surface density of recent star formation in galaxy disks , then high @xmath100 ( numerically small values ) should correlate with high global star formation activity , consistent with the observed tf residual@xmath100 correlation
. however this view may be too simplistic , because galaxies ranging from undisturbed spirals to emission - line s0 s with kinematically decoupled gas all seem to follow the same correlation , albeit with very large scatter . given the constancy of the spearman test significance for this correlation whether using u , b , or r - band tf residuals ( figure [ fg : fullsampcorr ] ) , we suggest that variations in @xmath100 may track not only star formation but also structural differences between classes of galaxies located in different parts of the @xmath71@xmath72 plane , such as differences in the concentration of light or the velocity width at a given luminosity .
dwarfs on opposite sides of the tfr have different @xmath100 distributions ( figure [ fg : twodwarfs ] ) , analogous to the different rotation curve asymmetry distributions discussed in the previous section .
while low to moderate surface brightnesses occur on both sides of the tfr , high surface brightnesses occur only on the high @xmath71/low @xmath72 side .
typical gas consumption timescales also appear to differ for dwarfs on the two sides the tfr , although this result relies on incomplete hi mass data ( 82% complete for the galaxies in figure [ fg : twodwarfs ] , with fainter galaxies preferentially missing ) . in figures [ fg : fullsampcorr ] and [ fg : twodwarfs ] , only a few galaxies have low @xmath71/high @xmath72 ( positive ) residuals and very short gas consumption timescales .
by contrast , short gas consumption timescales are the norm for galaxies with negative residuals .
most of the latter set are low - luminosity non - spiral galaxies , which explains the emergence of a gas consumption timescale correlation in the full sample , where there was no correlation in the spiral sample of [ sc : spiral3 ] . in [ sc : dwarfrcasym][sc : gastime ]
we showed that if we divide the dwarf population into two groups with positive and negative offsets from the tfr , the groups have distinct physical properties ( summarized in figure [ fg : twodwarfs ] ) .
although k - s tests yield only @xmath63@xmath76 confidence that the two populations are distinct based on any one property , the same split is observed for three completely independent parameters : rotation curve asymmetry ( based on high - resolution spectra ) , @xmath100 ( based on surface photometry ) , and gas consumption timescale ( based on catalog hi data and nfgs spectrophotometry ) .
the tf residuals for the two dwarf populations differ by @xmath62 mag in the mean .
we now consider whether systematic effects might create such a split . because the two sets of dwarfs in figure [ fg : twodwarfs ] show similar distributions in inclination and morphology , correlated inclination errors are unlikely . truncated optical rcs
are also unlikely to play a role , as figure [ fg : twodwarfs ] excludes galaxies whose rcs do not reach 1.3@xmath27 .
furthermore , replacing optical velocity widths with hi linewidths preserves the division . rc asymmetries may affect measured velocity widths , but the asymmetries we measure are too small to produce a 2 mag difference between the two groups of dwarfs ( cf .
[ sc : voffs ] ) .
correlated distance errors might be a concern .
both dwarf populations show broad distributions in r.a .- dec .- redshift space , but due to the nfgs selection procedure , the two sets of dwarfs in figure [ fg : twodwarfs ] have systematically different recession velocities : 5001300 and 7001600 @xmath16 for underluminous and overluminous dwarfs respectively , with over half the underluminous set in the range 500700 @xmath16 ( measured with respect to the local group ) . to test whether peculiar velocities or local flows may have affected our distance estimates and produced a spurious bifurcation in luminosities , we have substituted distances estimated by a number of techniques .
these include : ( 1 ) hubble - law distances assigned using group recession velocities rather than individual galaxy recession velocities @xcite ; this substitution yields one large shift ( 0.6 mag toward higher luminosity ) , for a galaxy ( ugc 6446 ) assigned `` questionable '' membership in the ursa major cluster by @xcite .
its projected location lies near the cluster outskirts and its velocity approaches the cluster lower limit . ]
( 2 ) absolute distances derived from the light - to - mass model of @xcite ; ( 3 ) hubble - law distances with and without corrections for virgocentric infall ( [ sc : absmag ] ) and/or for the local anomaly @xcite .
the segregation of the two dwarf populations is robust under all of these substitutions , with the caveat that many of the overluminous ( low velocity width ) dwarf galaxies lack distances from methods ( 1 ) and ( 2 ) .
we conclude that the tf offset between the two groups of nfgs dwarfs is probably real .
these results confirm trends reported for two other heterogeneous tf samples @xcite .
both studies identify a rotationally supported dwarf population on the faint side of the tfr and a population with comparable rotational and random velocities on the bright side ( where we measure high rotation curve asymmetries ) .
the physical mechanisms responsible for the disturbed kinematics and rapid gas consumption of the dwarfs on the bright side of the tfr are uncertain , but interactions are an obvious possibility . without deep wide - field imaging
we can not evaluate the frequency of close companions for the two populations , particularly given the possibility of relatively faint neighbors .
nonetheless , it is intriguing to note that deep survey work by @xcite has demonstrated a statistically higher rate of companions near starbursting dwarfs than near lsb dwarfs ( see also * ? ? ?
the analogy is suggestive rather than exact , as our two dwarf populations do not show a definitive split in ew(h@xmath4 ) , although the three most extreme starbursting dwarfs do fall on the bright side of the tfr ( figure [ fg : twodwarfs ] ) .
the three studies that have reported a difference in dwarf properties on either side of the tfr are all based on heterogeneous galaxy samples representing a variety of environments ( our own ; * ? ? ?
* ; * ? ? ?
here we investigate dwarf properties and their relationship to the tfr for the ursa major sample of @xcite , which represents a single moderate - density environment .
while the vs data set does not include kinematic asymmetries or u - band data , we can examine b - band surface brightnesses and tf offsets .
( although not significant at the 3@xmath76 level , @xmath116 trends basically follow @xmath100 trends with weaker signal . ) as noted in
[ sc : cluster ] ( see also figure [ fg : verhfigscatt ] ) , the ursa major tfr appears to diverge from the nfgs tfr at the faint end .
when we expand the two spiral samples to include late - type galaxies and dwarfs down to m@xmath117 ( the ursa major sample completeness limit ) , we find a tf offset of @xmath60.5 mag between the nfgs and ursa major dwarfs , with the ursa major dwarfs showing higher @xmath71/lower @xmath72 .
values ( ngc 3985 , ugc 7089 , and ngc 4218 ) , so for these we substitute w@xmath32 @xmath8 20 @xmath35 ( cf .
[ sc : vcalcs ] ) .
] for the same nfgs and ursa major dwarfs , we also find a mean effective surface brightness ( @xmath116 ) offset of @xmath60.5 mag arcsec@xmath118 , with the ursa major dwarfs _
fainter_. these results
do not conform to what we observe in the nfgs , where overluminous dwarfs have relatively bright @xmath116 values .
also , figure [ fg : sbcomparison ] shows that the ursa major dwarfs display a very narrow range in @xmath116 compared to the nfgs , with a faint median value of 23.7 .
herein lies a possible explanation for the discrepancy : if lsb dwarfs are more likely to have rising rcs at their outermost observed radii , then the `` true '' halo velocities of the ursa major dwarfs may be systematically underestimated ( cf .
* ) . in this case
the divergence of the two tfrs may be due to real physical differences between the two dwarf populations , and the narrow range of properties seen for the ursa major dwarfs may simply reflect the specific environment of ursa major .
the cluster consists of a moderate overdensity of spiral galaxies that lacks any discernible core , contains no elliptical galaxies , and includes surprisingly few dwarfs relative to bright galaxies @xcite .
@xcite infer from the scarcity of dwarf galaxies in ursa major that the cluster collapsed late , so that its dwarf halos formed after reionization when conditions were poor for star formation .
nonetheless , the environmental differences hypothesis should be treated with caution , because the ursa major dwarfs differ from those in the nfgs at only the 2@xmath76 level , with velocity width data missing for a few dwarfs and with the cluster membership of another dwarf in question ( see note [ fn : umoutlier ] ; this galaxy happens to appear in both data sets ) .
furthermore , we can not completely eliminate small systematic uncertainties in relative distance scale and photometric calibration between the two data sets ; a larger and more uniform data set would be required to definitively establish the existence of environmental effects . as noted by @xcite ,
the existence of two dwarf populations complicates tf slope measurements . for the full sample ,
we find a linear tf slope with no break at low luminosities ( top left panel of figure [ fg : slopebreak ] ) .
however , different sample selection criteria may favor one dwarf population over the other , with consequences for the faint - end slope .
if we consider the tfr as a physical rather than an empirical relation , linking luminosity and rotation velocity for passively evolving disk galaxies , then we should restrict it to undisturbed rotationally supported systems .
dwarf galaxies on the low @xmath71/high @xmath72 side of the tfr best meet this description .
these galaxies show moderate rotation curve asymmetries , indicating that they are probably rotationally supported .
in addition , their long gas consumption timescales and low u - band effective surface brightnesses appear consistent with quiescent evolution .
( we note that photometric `` normalcy '' is not a useful criterion for dwarfs , as almost all are lumpy and asymmetric , so defining what is `` peculiar '' becomes extremely difficult . ) figure [ fg : slopebreak ] demonstrates the effect of selecting restricted tf samples to match the properties of the quiescent dwarfs at all luminosities .
each panel shows separate tf fits for galaxies above and below m@xmath114 , first for the full tf sample , and then for subsamples restricted in turn to have low kinematic asymmetry , long gas consumption timescale , and faint u - band effective surface brightness .
the faint - end fits should not be interpreted as precisely measuring the slope of the dwarf tfr , given the small number statistics and short lever arm .
nonetheless , these fits clearly suggest a faint - end slope discontinuity for tf samples restricted to undisturbed , rotationally supported galaxies : the dwarfs in these samples appear underluminous when compared to the extrapolated bright - galaxy tfr .
breaks in the faint - end slope of the tfr have been previously observed for extremely late type spirals @xcite and lsb dwarfs @xcite .
mcgaugh et al.demonstrate that a straight tf slope ( the `` baryonic tully - fisher relation '' ) can be restored for lsb dwarfs by mathematically converting their gas into stars and correcting the luminosity for these unformed stars .
this procedure does not work very well for our faint galaxies , but mcgaugh et al .
see significant effects primarily for galaxies fainter than our cutoff of m@xmath119 .
the nearby field galaxy survey offers an ideal sample for exploring the link between offsets from the tully - fisher relation and the evolutionary states of galaxies .
galaxies in the nfgs were selected from the cfa 1 redshift survey without preference for morphology or environment and span a wide range of luminosities , @xmath120 .
using ubr photometry , optical rotation curve data , and hi linewidths , we have analyzed tf residuals for two subsamples of the nfgs , a sample of sa sd galaxies brighter than @xmath2 and an extended sample including dwarfs and non - spiral galaxies . within the spiral sample ,
we find strong third - parameter correlations between tf residuals and both @xmath3 color and ew(h@xmath4 ) .
the cleanest correlations are achieved using effective colors ( measured within the effective radius ) and global equivalent widths of h@xmath4 and [ oii ] ( integrated over the entire galaxy ) . using the color tf residual correlation in each tf band to define a color - correction term to the tfr
, we reduce scatter to a single constant value across the r , b , and u bands , eliminating the usual trend of increasing scatter with bluer passbands
. the remaining scatter of @xmath60.65 mag ( at a slope of @xmath810 ) approximately equals the scatter we expect from measurement errors .
an ew(h@xmath4)-correction term performs equally well for the r and b - band tfrs and almost as well for the u - band tfr .
the color and ew(h@xmath4 ) correlations are continuous , but their statistical signal is strongly driven by two morphology classes : peculiar galaxies ( by which we mean recognizable spiral galaxies with oddities such as warps , multiple nuclei , or interacting companions ) that fall on the high luminosity / low velocity width side of the tfr , and sa galaxies that fall on the low luminosity / high velocity width side of the tfr .
this result offers a different perspective on the longstanding debate over the sa offset , by linking the offset simply to the fact that sa galaxies are consistently red .
differences in star formation history offer the simplest explanation for the color tf residual correlation : as stellar populations evolve , changes in their colors and luminosities are intimately connected .
we argue that the steep slope of the correlation can not be explained in terms of a trivial correlation due to passband - dependent scatter . interpreting tf residuals as luminosity residuals ,
the slope of the color tf residual correlation can be directly compared with the predictions of population synthesis models of disk galaxy evolution .
the slopes we find using r and b band tf residuals closely match the slopes predicted from the passively evolving galaxy models of @xcite .
however , we find a slight suggestion of steeper slopes , which might be taken as evidence that in addition to passive evolution , many spiral galaxies experience significant starburst - driven evolution , e.g. triggered by interactions or minor mergers ( as independently suggested by the disturbed morphology and kinematics of many of our bluer galaxies ) . variations in dark matter structure could also steepen the correlation if they were correlated with color , although we find no independent evidence for such an effect .
finally , the present data set can not conclusively rule out systematic trends caused by kinematic distortions or inclination errors that correlate with galaxy color : such trends are unlikely to explain the entire color
tf residual correlation but may affect its slope . removing the spiral morphology restriction , we find that most non - spiral galaxies brighter than @xmath2 follow the color tf residual and ew(h@xmath4)tf residual correlations established by spiral galaxies , but with greater scatter .
a few emission - line s0 galaxies at the faint end of the included range deviate strongly from both the tfr and the color tf residual correlation ; these galaxies all show starburst or agn activity and anomalous gas kinematics .
dwarf galaxies fainter than @xmath121 do not follow the color tf residual correlation .
however , dwarfs on the high luminosity / low velocity width side of the tfr have higher rotation curve asymmetries , higher u - band effective surface brightnesses , and shorter gas consumption timescales than dwarfs on the low luminosity / high velocity width side of the tfr .
these properties suggest that many of the high @xmath71/low @xmath36 dwarfs are `` disturbed , '' while the low @xmath71/high @xmath36 dwarfs are likely to be passively evolving , rotationally supported systems .
this split in dwarf properties implies that the faint - end slope of the tfr depends on sample selection .
we find no break in slope for dwarf galaxies in the full nfgs sample .
however , if we select for rotationally supported , passively evolving galaxies at all luminosities , we find evidence for a break toward steeper slope at the faint end of the tfr , consistent with recent work by @xcite and @xcite .
we thank eric bell for performing additional model calculations to answer our questions and for pointing us to helpful references .
we also thank betsy barton for sharing information on technical aspects of tully - fisher analysis , and rolf jansen for providing both data and general expertise .
finally , we are grateful to margaret geller , douglas mar , and kristin nelson - patel for helpful discussions , and to john huchra , renzo sancisi , and the anonymous referee for critical readings of the manuscript .
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 .
this publication also makes use of the digital sky survey ( poss plates ) , based on photographic data of the national geographic society palomar geographic society to the california institute of technology .
the plates were processed into the present compressed digital form with their permission .
the digitized sky survey was produced at the space telescope science institute under us government grant nag w-2166 .
sjk acknowledges support from a nasa gsrp fellowship .
86 natexlab#1#1 , m. , huchra , j. , mould , j. , schechter , p. l. , & tully , r. b. 1982 , , 258 , 64 , r. g. , tanvir , n. r. , santiago , b. x. , ellis , r. s. , glazebrook , k. , & van den bergh , s. 1996 , , 279 , l47 , e. j. , geller , m. j. , bromley , b. c. , van zee , l. , & kenyon , s. j. 2001 , , 121 , 625 , e. j. & van zee , l. 2001 , , 550 , l35 , t. c. , flynn , k. , & gebhardt , k. 1990 , , 100 , 32 , e. f. & de jong , r. s. 2001 , , 550 , 212 , m. a. & andersen , d. r. 2000 , in asp conf . ser . 197 : dynamics of galaxies : from the early universe to the present , 175 , m. a. , haynes , m. p. , giovanelli , r. , & andersen , d. r. 1999 , in asp conf .
ser . 182 : galaxy dynamics - a rutgers symposium , 499 , l. , gouguenheim , l. , fouque , p. , & paturel , g. 1990 , , 82 , 391 , a. , jimenez , r. , & kamionkowski , m. 2001 , , 322 , 43 , m. , mould , j. r. , & staveley - smith , l. 1996 , , 463 , 60 , s. 1996 , , 103 , 363 . 1997 , , 114 , 2402 , s. .
& rix , h. 1999 , , 513 , 561 , d. j. & loeb , a. 1996 , , 459 , 432 , d. , yepes , g. , kates , r. , mller , v. , & klypin , a. 1999 , , 515 , 525 , d. , cheimets , p. , caldwell , n. , & geary , j. 1998 , , 110 , 79 , d. a. , phillips , a. c. , koo , d. c. , & illingworth , g. d. 1996 , , 462 , 89 , m. & de zeeuw , t. 1992 , , 392 , l47 , k. c. 1970 , , 160 , 811 , z. & gunn , j. e. 1994 , , 108 , 1476 , r. , paturel , g. , petit , c. , marthinet , m. c. , & rousseau , j. 1996 , , 117 , 467 , r. , haynes , m. p. , herter , t. , vogt , n. p. , da costa , l. n. , freudling , w. , salzer , j. j. , & wegner , g. 1997 , , 113 , 53 , r. , haynes , m. p. , herter , t. , vogt , n. p. , wegner , g. , salzer , j. j. , da costa , l. n. , & freudling , w. 1997 , , 113 , 22 , e. 1986 , , 164 , 17 , k. d. , calzetti , d. , & witt , a. n. 1997 , , 487 , 625 , m. & mould , j. 1990 , , 360 , 448 , m. p. , giovanelli , r. , chamaraux , p. , da costa , l. n. , freudling , w. , salzer , j. j. , & wegner , g. 1999 , , 117 , 2039 , m. p. , jore , k. p. , barrett , e. a. , broeils , a. h. , & murray , b. m. 2000 , , 120 , 703 , j. , davis , m. , latham , d. , & tonry , j. 1983 , , 52 , 89 , t. , feigelson , e. d. , akritas , m. g. , & babu , g. j. 1990 , , 364 , 104 , g. h. , branch , d. , clardullo , r. , davies , r. l. , harris , w. e. , pierce , m. j. , pritchet , c. j. , tonry , j. l. , & welch , d. l. 1992 , , 104 , 599 , r. a. 2000 , phd thesis , univ .
groningen , the netherlands , r. a. , fabricant , d. , franx , m. , & caldwell , n. 2000 , , 126 , 331 , r. a. , franx , m. , fabricant , d. , & caldwell , n. 2000 , , 126 , 271 , i. 1994 , , 106 , 967 , s. j. 2001 , phd thesis , harvard university , s. j. & fabricant , d. g. 2001 , , 121 , 140 , r. c. , j. 1998 , , 36 , 189 , r. c. , j. , tamblyn , p. , & congdon , c. e. 1994 , , 435 , 22 , k. & watanabe , m. 1988 , , 96 , 1593 , r. c. , sandage , a. , & tammann , g. a. 1984 , , 283 , 24 , r. o. , huchra , j. p. , & geller , m. j. 1994 , , 428 , 43 , d. s. , ford , v. l. , & buchhorn , m. 1992 , , 81 , 413 , l. d. , van driel , w. , & gallagher , j. s. 1998 , , 116 , 2196 , s. s. , schombert , j. m. , bothun , g. d. , & de blok , w. j. g. 2000 , , 533 , l99 , j. c. & hernquist , l. 1996 , , 464 , 641 , e. , maoz , d. , rix , h. , & tonry , j. l. 1999 , , 117 , 2666 , p. 1973
, uppsala general catalogue of galaxies ( uppsala : astronomiska observatorium , 1973 ) , k. , bothun , g. d. , & schombert , j. 2000 , , 119 , 136 , m. j. & tully , r. b. 1988 , , 330 , 579 . 1992 , , 387 , 47 , d. 1999 , , 352 , 49 , d. & tuffs , r. j. 1999 , , 343 , 751 , s. a. , kniazev , a. y. , lipovetsky , v. a. , & ugryumov , a. v. 2001 , , 373 , 24 , s. , von braun , k. , bernstein , g. m. , & guhathakurta , p. 1997
, , 113 , 2046 , m .- h .
1996 , phd thesis , univ .
groningen , the netherlands , h. , guhathakurta , p. , colless , m. , & ing , k. 1997 , , 285 , 779 , m. s. 1978 , , 83 , 1026 , v. c. , burstein , d. , ford , w. k. , & thonnard , n. 1985 , , 289 , 81 , s. , mould , j. r. , hughes , s. m. g. , huchra , j. p. , macri , l. m. , kennicutt , r. c. , gibson , b. k. , ferrarese , l. , freedman , w. l. , han , m. , ford , h. c. , graham , j. a. , illingworth , g. d. , kelson , d. d. , madore , b. f. , sebo , k. , silbermann , n. a. , & stetson , p. b. 2000 , , 529 , 698 , p. l. 1980 , , 85 , 801 , l. & pritchet , c. j. 1998 , , 505 , 96 , j. m. & israel , f. p. 1998 , in astro - ph/9810151 , m. a. & willick , j. a. 1995 , , 261 , 271 , c. l. , thomas , d. l. , brinks , e. , & skillman , e. d. 1996 , , 107 , 143 , g. , bottinelli , l. , coudreau - durand , n. , gouguenheim , l. , hallet , n. , loulergue , m. , paturel , g. , & teerikorpi , p. 1998
, , 130 , 333 , n. , tully , r. b. , & verheijen , m. a. w. 2001 , , 325 , 385 , r. b. 1988 , nearby galaxies catalog ( cambridge and new york , cambridge university press ) , r. b. & fisher , j. r. 1977 , , 54 , 661 , r. b. & fouque , p. 1985
, , 58 , 67 , r. b. & pierce , m. j. 2000 , , 533 , 744 , r. b. , pierce , m. j. , huang , j. , saunders , w. , verheijen , m. a. w. , & witchalls , p. l. 1998 , , 115 , 2264 , r. b. , shaya , e. j. , & pierce , m. j. 1992 , , 80 , 479 , r. b. , verheijen , m. a. w. , pierce , m. j. , huang , j. , & wainscoat , r. j. 1996 , , 112 , 2471 , y. & sofue , y. 1997 , , 326 , 915 , m. a. w. 1997 , phd thesis , univ .
groningen , the netherlands , m. a. w. 2001 , , accepted , m. a. w. & sancisi , r. 2001 , , 370 , 765 , n. p. , phillips , a. c. , faber , s. m. , gallego , j. , gronwall , c. , guzman , r. , illingworth , g. d. , koo , d. c. , & lowenthal , j. d. 1997 , , 479 , l121 , m. , yasuda , n. , itoh , n. , ichikawa , t. , & yanagisawa , k. 2001 , , 555 , 215 , j. a. 1994 , , 92 , 1 , j. a. , courteau , s. , faber , s. m. , burstein , d. , dekel , a. , & kolatt , t. 1996 , , 457 , 460 , n. , fukugita , m. , & okamura , s. 1997 , , 108 , 417 , d. 1995 , , 448 , l17 , d. & rix , h. 1997 , , 477 , 118
tf samples are generally incomplete at low luminosities and low surface brightnesses , resulting in asymmetric scatter about the fundamental relation . in a fit to magnitude as a function of @xmath122 ( the `` forward '' tfr ) , this asymmetric scatter leads to artificially shallow slopes .
extended luminosity coverage , as in the nfgs , helps to alleviate the problem , but asymmetric faint - end scatter can still bias results .
@xcite describes one way to address tf slope bias , using an iterative analysis .
his technique assumes that the `` missing '' galaxies have the same intrinsic tf slope and scatter as the observed galaxies .
however , if either the scatter or the slope increases at low luminosity , the bias - corrected slope is still too shallow .
alternatively , one can fit the `` inverse '' relation ( minimizing the scatter in velocity width as a function of magnitude ) without any bias correction , assuming that velocity width biases are negligible ( cf .
* ; * ? ? ?
this approach seems appropriate for our sample , because : ( 1 ) optical rc data ( unlike radio data ) are not subject to a velocity width detection bias , and ( 2 ) the nfgs makes no explicit diameter cut that might introduce an indirect velocity width bias through a correlation of rotation speed with surface brightness .
however , intrinsic surface brightness and color bias in the parent survey may still play a role .
tables 1 and 2 present our basic tf calibrations using three different techniques : error - weighted inverse fits , bivariate fits with regression in both magnitude and velocity width errors , and unweighted inverse fits , our preferred technique .
error - weighted inverse fits avoid the usual tf slope bias , but they favor galaxies with higher luminosity and larger velocity width , which have smaller errors .
bivariate fits use all the available information but at the expense of introducing some slope bias ( c.f .
note that the systematic differences between these fitting techniques often exceed the formal errors ( tables 1 and 2 ) .
we compute unweighted inverse fits in two steps .
first we find the 68% confidence interval for the slope , where for each test value of the slope we allow the intercept to assume the value that minimizes deviations for that slope .
once we have the best - fit slope , we fit for the intercept with the slope fixed . although the results are the same as those of a one - step two - parameter fit , the errors are different .
this type of fit is appropriate for comparing the slopes of different samples irrespective of zero point , or conversely for comparing their zero points at fixed slope .
all fits are subject to one round of 3@xmath76 rejection .
such rejection reduces measured scatter in a gaussian distribution by @xmath62% , which we correct for in @xmath123 . using the same slope and zero point we also compute the full scatter @xmath124 , the biweight scatter @xcite , and the predicted scatter from observational errors @xmath125 .
the biweight statistic provides the most robust measure of observed scatter , generally agreeing closely with @xmath123 .
below we present three different optical velocity width definitions , @xmath25 , @xmath26 , and @xmath28 , and place them on an equivalent w@xmath29 scale .
our optical - to - radio conversions are expected to differ from those of @xcite , not just because of differences in sample selection , but also because courteau s conversions are calibrated using turbulence - corrected hi linewidths ( taken from * ? ? ?
* ) . in our case ,
the conversions are calibrated using uncorrected hi linewidths .
we define @xmath25 as the single largest velocity in the rotation curve , relative to the origin defined by minimizing the rc asymmetry ( appendix [ sc : asymmeast ] ) . because this number does not make use of the full information in the rotation curve , we expect it to be somewhat unreliable .
nonetheless , it correlates surprisingly well with the radio linewidth w@xmath29 ( figure [ fg : w50vsopt ] ) , even when the optical rc does not extend as far as 1.3@xmath126 ( the peak velocity position for an exponential disk with effective radius @xmath27 , * ? ? ?
an iterative least - squares fit with automatic 3.5@xmath76 outlier rejection yields @xmath127 with scatter 25 @xmath16 .
following @xcite , we define @xmath26 by first fitting the empirical function @xmath128 to the observed rc , with the origin unconstrained , and then interpolating to find the velocity at a specific radius . here
@xmath129 , @xmath130 defines the origin , @xmath131 gives the velocity scale , @xmath132 is related to the turnover radius , and @xmath61 and @xmath18 are free parameters governing the shape of the rotation curve .
we modify courteau s technique slightly , in that we interpolate the velocity @xmath26 at 1.3@xmath126 rather than at 2.2 disk scale lengths @xmath133 , because 1.3@xmath126 remains well defined even for non - disk galaxies and does not require bulge+disk decomposition .
theoretically the two radii are equivalent , in that both represent the peak velocity position for a pure exponential disk @xcite . for disk - dominated galaxies in the nfgs ,
1.3@xmath126 and 2.2@xmath133 also match well observationally .
( disk scale lengths are courtesy of r. jansen , private communication ; both disk scale lengths and @xmath126 s are from the b band images of @xcite . ) as shown in figure [ fg : rcgallery ] , our modified - courteau technique performs well for bright spiral galaxies comparable to those in courteau s sample , but less regular rcs are poorly modeled .
figure [ fg : w50vsopt ] displays the correlation of @xmath26 with the radio linewidth w@xmath29 , demonstrating reasonable agreement , although some points are simply missing because the fit would not converge .
an iterative least - squares fit with automatic 3.5@xmath76 outlier rejection yields @xmath134 we define the equivalent w@xmath29 linewidth w@xmath78 by this equation and assign it an error equal to the scatter in the fit , 25 @xmath16 .
following @xcite , we define @xmath28 as half the difference between the statistical `` probable maximum '' and `` probable minimum '' velocities implied by the observed rotation curve .
the probable maximum velocity @xmath135 is defined to have a 10 percent chance of exceeding all velocities in the rotation curve : @xmath136 modeling each velocity @xmath137 as a gaussian distribution about the measured value , with @xmath76 equal to the measurement error , we obtain @xmath138 which may be solved numerically for @xmath135 .
the probable minimum velocity @xmath139 is defined by analogy , and @xmath140 .
@xcite suggests that @xmath28 will be very sensitive to outlying points in the rotation curve , while @xcite claim the opposite .
the difference seems to hinge on the size of the error bars associated with the outliers if a high s / n cosmic ray hit masquerades as a very high - velocity data point , it may indeed throw off @xmath28 , but outlying points with large error bars will have relatively little impact . in practice , @xmath28 produces smaller scatter than @xmath26 in the optical - to - hi conversion .
more importantly , as @xcite point out , @xmath28 has the virtue of using all of the data in the rotation curve without imposing any particular model on the data .
the diversity of the nfgs defies modeling through a simple fitting function ( see figure [ fg : rcgallery ] ) .
figure [ fg : w50vsopt ] shows the correlation of @xmath28 with radio linewidth w@xmath29 .
an iterative least - squares fit with automatic 3.5@xmath76 outlier rejection yields @xmath141 we define the equivalent w@xmath29 linewidth w@xmath38 by this equation and assign it an error equal to the scatter in the fit , 20 @xmath16 .
we compute optical rotation curve asymmetries using a technique similar to the photometric technique of @xcite . reflecting the rest - frame rotation curve about its origin , we find the average absolute deviation between the two sides , @xmath142 , and express the result as a percentage of the velocity width @xmath143 ( appendix [ sc : vpmmdefn ] ) .
our asymmetry , velocity width , and central redshift measurements are not completely independent of one another , therefore in practice we perform these calculations iteratively .
asymmetry measurements depend sensitively on the choice of origin for the spatial and velocity axes .
we therefore vary the position of the origin to minimize the asymmetry , keeping the spatial coordinate constrained within the one - sigma error bars of the continuum peak position , while allowing the velocity coordinate to vary freely . at each step in the minimization process
we resample the rotation curve onto a uniform spatial grid about the trial origin . to partially standardize our asymmetry index for rotation curves of differing spatial extent , we define an inner asymmetry using only points within 1.3@xmath126 ( the peak velocity position for a pure exponential disk , appendix [ sc : vfitdefn ] ) . when the rotation curve does not extend this far , we compute the inner asymmetry from the data available .
the rotation curve origin is determined by minimizing the inner asymmetry .
all other analysis makes use of the total asymmetry , which we compute relative to the origin defined by minimizing the inner asymmetry . | we analyze residuals from the tully - fisher relation for the emission - line galaxies in the nearby field galaxy survey , a broadly representative survey designed to fairly sample the variety of galaxy morphologies and environments in the local universe for luminosities from @xmath0 to @xmath1 . for a subsample consisting of the spiral galaxies brighter than @xmath2
, we find strong correlations between tully - fisher residuals and both @xmath3 color and ew(h@xmath4 ) .
the extremes of the correlations are populated by sa galaxies , which show consistently red colors , and spiral galaxies with morphological peculiarities , which are often blue .
if we apply an ew(h@xmath4)-dependent or @xmath3 color - dependent correction term to the tully - fisher relation , the scatter in the relation no longer increases from r to b to u but instead drops to a nearly constant level in all bands , close to the scatter we expect from measurement errors .
we argue that these results probably reflect correlated offsets in luminosity and color as a function of star formation history . broadening the sample in morphology and luminosity
, we find that most non - spiral galaxies brighter than m@xmath5 follow the same correlations between tully - fisher residuals and @xmath3 color and ew(h@xmath4 ) as do spirals , albeit with greater scatter .
however , the color and ew(h@xmath4 ) correlations do not apply to galaxies fainter than m@xmath5 or to emission - line s0 galaxies with anomalous gas kinematics . for the dwarf galaxy population ,
the parameters controlling tully - fisher residuals are instead related to the degree of recent evolutionary disturbance : overluminous dwarfs have higher rotation curve asymmetries , brighter u - band effective surface brightnesses , and shorter gas consumption timescales than their underluminous counterparts . as a result , sample selection strongly affects the measured faint - end slope of the tully - fisher relation , and a sample limited to include only passively evolving , rotationally supported galaxies displays a break toward steeper slope at low luminosities . |
the behavior of the single nucleon in the medium bears clear signatures of the nucleon
nucleon ( nn ) correlations induced by the strong nuclear interaction .
the most celebrated one is the presence of high momentum components in the momentum distribution ( md ) , which may be explained only in terms of short range correlations ( src ) and can not be justified by any independent particle model ( ipm)@xcite .
in addition , both short and long range correlations ( lrc ) are now widely thought to be the origin of the large reduction of the spectroscopic strengths of the hole states@xcite . in general ,
the investigation of src is one of the main topics in the physics of strongly interacting systems , from liquid helium to nuclear matter .
the translational invariance of homogeneous media has made possible a quantitative assesment of their importance in these contexts .
however , the advances in realistic many body theories have only recently allowed to address the same issue in inhomogeneous objects , as atomic nuclei , with a comparable accuracy . to this aim ,
correlated basis functions ( cbf ) theory has emerged as an effective tool to tackle the src problem in nuclear systems @xcite when realistic hamiltonians are employed . when used in conjuction with the fermi hypernetted chain ( fhnc ) integral equations@xcite , cbf has attained in @xmath0o and @xmath1ca the same accuracy as in the best variational studies of nuclear matter@xcite . in order to dissentagle the src effects
, one has to look at effects that go beyond any possible ipm description of the nucleus .
src are mainly produced by the repulsive core of the nn interaction , whereas important intermediate and long range contributions ( essentially of the tensor type ) come from the exchange of one or more pions .
this part of the nuclear interaction is the responsible of the nuclear binding@xcite .
direct evidence of the src are not easy to experimentally single out and only in these last few years , thanks to the advances in the experimental techniques , there have been consistent efforts aimed to their identification . for instance , @xmath5 data in the quasi - elastic region need a large reduction of the ipm hole strength to be reproduced @xcite
. moreover , charge density distributions obtained by elastic electron scattering experiments are smaller than those predicted by the ipm @xcite in the nuclear interior .
these facts can be be explained by assuming occupation probabilities of the single particle levels different from the ipm ones@xcite .
the basic quantity to investigate in order to verify the hypothesis of partial occupation probability is the one body density matrix ( obdm ) , @xmath6 , defined as : @xmath7 where @xmath8 is the ground state a body wave function and @xmath9 is the creation ( annihilation ) operator of a nucleon at the position @xmath10 .
the fourier transform of the obdm gives the momentum distribution ( md ) , @xmath11 , sometimes used in plane wave impulse approximation to study inclusive and exclusive reactions .
the natural orbits ( no)@xcite , with their occupation numbers ( @xmath12 ) , are defined as the basis where the obdm is diagonal . in the ipm ,
the nuclear ground state is described by a slater determinant of fully occupied single particle ( sp ) wave functions below the fermi surface , @xmath13 .
so , the no and the sp w.f .
coincide , with @xmath14=1 and @xmath15=0 .
deviations from this situation are a measure of the correlations , since they deplete the populated orbitals and allow higher no to get @xmath160 .
other important quantities are the quasihole ( qh ) wave functions , @xmath17 , defined as the overlaps between @xmath18 and the hole states , @xmath19 , obtained by removing a nucleon from the position @xmath20 . from @xmath5 experiments it is possible to obtain an accurate determination of the qh overlap functions@xcite , whose normalizations give the spectroscopic factors , @xmath21 .
typical values of @xmath21 , as extracted from the experiments , are @xmath22@xcite , and the deviations from unity ( their ipm value ) come in account of various effects , from center of mass to correlation corrections . in cbf theory the description of the correlations starts with an a body correlated wave function @xmath23 corresponding to a symmetrized product of non commuting two - body correlation operators , @xmath24 , acting on the mean field wave function , @xmath25 , given by a slater determinant of single particle wave functions , @xmath26 .
a realistic choice of @xmath24 is : @xmath27 where @xmath28\otimes \left [ 1 , { \mbox{\boldmath $ \tau$}}_i \cdot { \mbox{\boldmath $ \tau$}}_j \right]\ ] ] and @xmath29 is the tensor operator .
the variational principle provides a natural recipe to determine the correlation functions , @xmath30 , and the single particle wave functions by minimizing the ground state energy . in our calculation
we adopt a non relativistic nuclear hamiltonian of the form : @xmath31 where we have used the @xmath32 reduction of the argonne @xmath33 @xcite potential and the urbana ix@xcite three nucleon interaction .
the one body density matrix ( [ obdm_0 ] ) of doubly closed shell nuclei in ( @xmath34 ) coupling can be evaluated in fhnc@xcite , starting from @xmath35 where @xmath36 and @xmath37 is the mass number .
once the obdm is known , its diagonal part provides the one body density , @xmath38 , and the momentum distribution may be computed by @xmath39 the independent particle model expression for the obdm is well known and reads as : @xmath40 in the case of jastrow correlated wave functions ( when only the first component of the correlation ( [ corr6 ] ) is retained ) , the density matrix may be expanded in powers of the _ dynamical correlations _ ,
@xmath41 ^ 2 - 1 $ ] and @xmath42 , and of the _ statistical correlations_. the expansion generates cluster terms classified according to the number of particles and to the number of the correlation lines .
the fhnc equations allow for summing cluster terms at all orders .
details of the finite systems fhnc theory may be found in ref.@xcite .
the fhnc equations for the more general @xmath43 correlation of ( [ corr6 ] ) were derived in ref.@xcite for the one and two body densities and for the obdm .
the single particle wave functions needed to build the slater determinant @xmath25 have been obtained by solving the single particle schrdinger equation with a woods saxon potential , @xmath44}}}\ , , \label{ws}\ ] ] whose parameters have been fixed to reproduce at best the nuclei empirical densities .
this choice , while spoiling the binding energies by only a few percent with respect to a full minimization , provides an accurate description of the nuclear densities@xcite .
the correlated momentum distributions computed within the cbf scheme are shown in fig.[fig : fig1 ] .
the mds of @xmath0o ( continuous line ) and of @xmath1ca ( dashed line ) are compared with that of nuclear matter with jastrow ( thin solid line ) and full ( dash dotted line ) correlations .
the ipm momentum distributions are given by squares ( @xmath0o ) and stars ( @xmath1ca ) .
we stress that the differences between the jastrow and the @xmath43 correlations are similar in the infinite and finite systems and that the three cases show an analogous behavior in large momentum region , that is dominated by the short range structure ( and the nn correlations ) of the nuclear wave function .
the effect is , to a large extent , independent on the nucleus .
the tensor correlations enhance the tails of the mds by a factor 23 with respect to the jastrow case , slightly smaller than that found in ref.@xcite ( @xmath454 ) .
the difference may be understood in terms of the stronger tensor force of the argonne @xmath46 potential adopted in that reference , and of the presence of spin orbit correlations , omitted in our calculation .
another piece of information that can be extracted from the obdm are the natural orbits and their occupation numbers .
the no are obtained by diagonalizing the obdm : @xmath47 for spherical nuclei in ( @xmath34 ) single particle coupling , saturated in both spin and isospin , this expression can be recast as : @xmath48 where @xmath49 are the legendre polinomials , @xmath50 is the angle between @xmath10 and @xmath51 and @xmath52=4 is the nucleon degeneracy . exploiting again the spherical symmetry , one obtains @xmath53 the first three no of @xmath0o and @xmath1ca are shown in figs .
[ fig : fig2 ] .
the no occupation numbers for jastrow and @xmath43 correlations are given in table [ tab : tab1 ] the effect of the correlations on the shell model orbitals are mainly visible in the short range part of the @xmath54 state , making this no more localized than its ipm counterpart .
the shape of the other ipm states is barely influenced .
the occupation of the no corresponding to the shell model ones is depleted by as much as 22@xmath3 ( the @xmath55 state in @xmath1ca ) .
in contrast , the mean field unoccupied states become sizeably populated .
the two effects are largely due to the tensor correlations .
the green function ( gf ) approach of ref.@xcite found the @xmath0o @xmath561 no more populated than the cbf ones for the occupied shell model states ( @xmath570 , 1 ) , and , consequently , lower occupations for all the remaining orbitals ( @xmath58=.921 , @xmath59=.941 and @xmath60=.017 .
the @xmath61 ( @xmath62 ) occupation numbers are the average of the @xmath63 and @xmath64 ( @xmath65 and @xmath66 ) orbitals of the reference ) .
these discrepancies are probably due to the different potentials adopted , rather than to the methodologies . in fact , in the gf calculation the one boson exchange bonn b potential@xcite was used .
the a8+uix model induces stronger correlations , so giving a larger depletion of the lowest no .
a similar effect was found in @xmath67he atomic drops@xcite , where the strong repulsive interaction between the @xmath67he atoms depletes the shell model occupations by 1546 @xmath3 .
in a fixed center reference frame , the overlap between the a body ground state and the ( a-1)body hole state of the residual system , @xmath68 , is given by @xmath69 in cbf theory @xmath70 is built by substituting @xmath25 in ( [ psi_0 ] ) with the slater determinant obtained by removing a nucleon in the state @xmath71 , @xmath72 . in doubly closed shell nuclei in the @xmath34 coupling scheme the radial dependence of the overlap function can be singled out , as @xmath73 where @xmath74 is the spin isospin single particle wave function . in the ipm ,
the overlaps are simply the shell model functions and @xmath75 , where @xmath76 is the radial part of @xmath26 . in order to develop a cluster expansion for @xmath77 ,
its expression is rearranged as : @xmath78 where @xmath79 and @xmath80 cluster expansions and fhnc like equations are then used to compute @xmath81 and @xmath82@xcite , along the lines given in ref.@xcite to evaluate the overlap matrix elements in the cbf study of the nuclear matter spectral function . finally , the spectroscopic factor , @xmath21 , is given by the quasihole normalization , @xmath83 the independent particle model gives @xmath84 . center of mass ( @xmath85 ) corrections are a first source of correction .
in fact , in the harmonic oscillator model they enhance @xmath21 for the valence hole states ( those having the largest oscillator quantum number , @xmath86 ) by the @xmath87^{n_v}$ ] factor@xcite .
so , the @xmath85corrected @xmath61shell spectroscopic factor of @xmath0o is @xmath88 , while the average between the @xmath55 and @xmath62 states in @xmath1ca is @xmath89 .
similar results are found with the more realistic woods
saxon orbitals@xcite . the correlated spectroscopic factors ( without @xmath85 corrections ) in @xmath0o and @xmath1ca
are shown in table [ tab : tab2 ] for @xmath43 and jastrow correlations .
jastrow correlations marginally reduce @xmath21 ( at most 3@xmath3 ) .
central spin
isospin correlations also provide a few percent depletion in the valence states , whereas the tensor correlations ( @xmath43 ) give most of the reduction .
for instance ,
they lower @xmath90 in @xmath0o to 0.90 and @xmath91 and @xmath92 in @xmath1ca to 0.86 and 0.87 , respectively . the @xmath61 cbf @xmath0o result fully agrees with the variational monte carlo estimate of ref.@xcite .
the operatorial correlations largely influence the low lying states , whose spectroscopic factors are drastically reduced by both central and tensor components : @xmath93 in @xmath0o is 0.70 , @xmath90 and @xmath93 in @xmath1ca are 0.58 and 0.55 , respectively
. the latest experimental values of @xmath94 from the @xmath0o@xmath95n reaction@xcite are @xmath96=0.61 for the 1/2@xmath97 ground state in @xmath98n and @xmath99=0.53 for the lowest 3/2@xmath97 state at 6.32 mev .
this state exhibits 87@xmath3 of the total @xmath100 strength , that is fragmented over three states at 6.32 , 9.93 and 10.70 mev .
the total @xmath100 may be estimated to be @xmath100=0.53/0.87=0.61@xcite . in the @xmath1ca@xmath101k
reaction@xcite the transition to the 1@xmath102 ground state gives @[email protected] .
we recall that the cbf values are @xmath1050.90 in @xmath0o and @xmath1060.87 in @xmath1ca .
the squared overlap functions are shown in fig.[fig : fig3 ] for the jastrow and @xmath43 correlations , and are compared with the ipm ones .
jastrow components have little effect on the overlaps , while the spin
isospin correlations are the main responsible for the quenching of the ipm overlaps and of the spectroscopics factors .
the ipm results obtained by a woods saxon potential fitting the @xmath0o@xmath95n cross section to the 6.32 mev state with @xmath99=0.53 are shown in the @xmath107@xmath0o panel as stars .
we have rescaled @xmath108 by the factor 0.53/0.90 and the result is in nice agreement with the empirical estimate .
in this contribution we have presented some results obtained within the correlated basis functions theory and concerning the properties of the single nucleon in the nuclear medium .
more specifically , we have discussed the behavior of the one - body density matrix and some related quantities , as the momentum distributions and the natural orbits , in @xmath0o and @xmath1ca using the fermi hypernetted chain resummation technique .
in addition , we have addressed the evaluation of the overlap functions and of the spectroscopic factors within the same approach .
the relevance of the short range correlations , both of central and tensor type , has been stressed .
the results presented here have been obtained using the argonne @xmath109 two nucleon potential plus the urbana ix three
nucleon interaction , together with a set of single particle wave functions fixed to reproduce at best the empirical charge distributions of the two nuclei . the high momentum tail of the momentum distribution is largely dominated by correlations , and tensor components
enhance the tail by a factor 34 with respect to the central ones .
the tensor correlation is also important for the occupation of the natural orbits .
in fact , the reduction of the occupation of the levels below the fermi surface and , conversely , the enhancement above is amplified by the tensor terms .
as far as the correlated overlap functions are concerned , they are close to the single particle wave functions if central correlations are used , whereas their shapes are strongly modified by the tensor correlations .
the overall spectroscopic factors are depleted by 10 - 15% , for the valence levels and 30 - 45% for the deeply lying ones . in spite of this reduction , the fhnc approach in @xmath0o
does not reproduce the empirical @xmath100 spectroscopic factor extracted from @xmath5 reactions .
a similar situation was met in ref.@xcite for nuclear matter , where the variational calculation of the one hole strength , @xmath110 , around the fermi level provided @xmath1110.88 , mostly due to tensor correlations .
second order perturbative corrections in a correlated basis , including two hole one particle , @xmath112 , correlated states , brought the strength to @xmath113 .
this decrease explains almost half of the discrepancy with the empirical @xmath114pb spectroscopic factor , @xmath115 .
the missing strength can be attributed to the coupling of the single particle waves to the collective low - lying surface vibrations , not reproducible in infinite nuclear matter .
it is expected that the inclusion of correlated @xmath116 corrections in the finite nuclei calculations can similarly take into account the coupling with surface vibrations .
the results presented in this contribution have been obtained in collaboration with giampaolo co. this work has been partially supported by murst through the _ progetto di ricerca di interesse nazionale : fisica teorica del nucleo atomico e dei sistemi a molticorpi_. j. g. zabolitzky and w , ey , phys .
b * 76 * , 527 ( 1978 ) ; o. bohigas and s. stringari , phys . lett .
b * 95 * , 9 ( 1980 ) ; m. dal r , s. stringari and o. bohigas , nucl . phys . a * 376 * , 81 ( 1982 ) ; f. dellagiacoma , g. orlandini and m. traini , nucl . phys . a * 393 * , 95 ( 1983 ) ; m. jaminon , c , mahaux and h. ng , nucl . phys .
a * 473 * , 509 ( 1987 ) .
a. n. antonov , p. e. hodgson and i. zh .
nucleon momentum and density distributions _
( clarendon press , oxford , 1988 ) .
g. van der steenhoven and p. k. a. de witt huberts , in _ modern topics in electron scattering _ , edited by b. frois and i. sick ( world scientific , singapore , 1991 ) , p.510 .
r. b. wiringa , v. ficks , and a. fabrocini , phys .
c * 38 * , 1010 ( 1988 ) .
g. co , a. fabrocini , s. fantoni , and i. e. lagaris , nucl .
a * 549 * , 439 ( 1992 ) .
f. arias de saavedra , g. co , a. fabrocini , and s. fantoni , nucl . phys .
a * 605 * , 359 ( 1996 ) .
a. fabrocini , f. arias de saavedra , g. co , and p. folgarait , phys .
c * 57 * , 1668 ( 1998 ) .
a. fabrocini , f. arias de saavedra , and g. co , phys .
c * 61 * , 044302 ( 2000 ) .
p. k. a. de witt huberts , j. phys .
g * 16 * , 507 ( 1990 ) .
et al . _ ,
lett . * 49 * , 978 ( 1982 ) . c. papanicolas , in _ nuclear structure at high spin , excitation and momentum transfer _
, h. nann ed . , ( american institute of physics , new york 1986 ) , p. 110 . c. mahaux and r. sartor , adv .
. phys . * 20 * , 1 ( 1991 ) .
j. j. kelly , adv .
* 23 * , 75 ( 1996 ) . l. lapiks , nucl . phys . a * 553 * , 297c ( 1993 ) .
r. b. wiringa , v. g. j. stoks , and r. schiavilla , phys .
c * 51 * , 38 ( 1995 ) .
b. s. pudliner , v. r. pandharipande , j. carlson , s. c. pieper , and r. b. wiringa , phys .
c * 56 * , 1720 ( 1997 ) .
a. fabrocini , and g. co , phys .
c * 63 * , 0443109 ( 2001 ) .
g. co , a. fabrocini , and s. fantoni , nucl .
a * 568 * , 73 ( 1994 ) .
s. c. pieper , r. b. wiringa , and v. r. pandharipande , phys .
c * 46 * , 1741 ( 1992 ) . a. polls , h. mther , and w. h. dickhoff , nucl . phys .
a * 594 * , 117 ( 1995 ) .
r. machleidt , adv .
* 19 * , 1 ( 1989 ) .
d. s. lewart , v. r. pandharipande , and s. c. pieper , phys . rev .
b * 37 * , 4950 ( 1988 ) .
o. benhar , a. fabrocini , and s. fantoni , nucl .
a * 505 * , 267 ( 1989 ) .
a. e. l. dieperink , and t. de forest , phys .
c * 10 * , 543 ( 1974 ) .
d. van neck , m. waroquier , a. e. l. dieperink , s. c. pieper , and v. r. pandharipande , phys .
c * 57 * , 2308 ( 1998 ) .
m. leuscher _
c * 49 * , 955 ( 1994 ) .
o. benhar , a. fabrocini and s. fantoni , phys .
c * 41 * , r24 ( 1990 ) .
ccccc @xmath117 & @xmath118(@xmath43;@xmath0o ) & @xmath118(j;@xmath0o ) & @xmath118(@xmath43;@xmath1ca ) & @xmath118(j;@xmath1ca ) + + + @xmath54 & 0.858 & 0.960 & 0.864 & 0.952 + @xmath55 & 0.019 & 0.005 & 0.780 & 0.962 + @xmath119 & 0.010 & 0.002 & 0.052 & 0.002 + @xmath120 & 0.005 & 0.001 & 0.013 & 0.001 + & & & & + @xmath61 & 0.919 & 0.980 & 0.841 & 0.949 + @xmath121 & 0.021 & 0.004 & 0.024 & 0.009 + @xmath122 & 0.011 & 0.003 & 0.016 & 0.006 + & & & & + @xmath62 & 0.025 & 0.006 & 0.956 & 0.983 + @xmath123 & 0.011 & 0.003 & 0.030 & 0.007 + @xmath124 & 0.006 & 0.001 & 0.019 & 0.006 + + cccccccccc & corr .
& @xmath54 & @xmath61 & @xmath62 & @xmath55 + + + @xmath0o & j & 0.98 & 0.98 & & + & @xmath43 & 0.70 & 0.90 & & + + @xmath1ca & j & 0.98 & 0.99 & 0.97 & 0.98 + & @xmath43 & 0.55 & 0.58 & 0.87 & 0.86 + + | we discuss some single particle properties of @xmath0o and @xmath1ca using correlated basis function theory and fermi hypernetted chain equations with central and tensor correlations . in particular , we concentrate on one body density matrix , momentum distribution , natural orbits and quasi hole states .
the correlations are variationally generated by a realistic hamiltonian containing the argonne @xmath2 two nucleon and urbana ix three nucleon interactions .
the correlated momentum distributions show the well known enhancement at large momenta , with a relative importance of the different correlations ( jastrow and tensor ) similar to that in nuclear matter .
the natural orbits and their occupation numbers are obtained by diagonalization of the density matrix .
the correlated first natural orbits occupations are depleted by more than 10@xmath3 , whereas the first following ones are occupied by a few percent .
the spectroscopic factors of the valence states are lowered by @xmath4 with respect to unity by central and tensor correlations , confirming that short range correlations alone are not able to explain the values extracted from @xmath5 experiments . |
one of the astounding principles of quantum mechanics is that a quantum object can exist in a superposition state , _
e.g. _ , a spin in a superposition of up and down states , @xmath0 .
applying this principle to two quantum particles , a superposition state such as a singlet state @xmath1 allows the instantaneous determination of the quantum state of the second particle after one measures the quantum state of the first particle , as revealed by einstein , podolsky and rosen in 1935 @xcite .
this instantaneous determination seems like a violation of a principle of the special relativity , _ i.e. _ , nothing can travel faster than light , leading to the so called epr paradox
. the incompleteness of quantum mechanics and the existence of hidden variables were therefore suspected . in a seminal paper by bell in 1964 @xcite , an equality was proposed to show that the statistical results of quantum mechanics is incompatible with any local hidden variable theories .
subsequent experiments overwhelmingly support the predictions of quantum mechanics . in particular ,
any local hidden variable theories are excluded by loophole - free experiments in 2015 @xcite . in the exploration of epr paradox ,
the greenberger - horne - zeilinger ( ghz ) states @xcite play an important role .
it concerns with a three - particle entangled state , @xmath2 , which allows a direct contradiction between the predictions of quantum mechanics and local hidden variable theories with a single test , in contrast to the statistical nature of the bell s inequality .
ghz states have also been found useful in the heisenberg limit metrology @xcite .
mesoscopic ghz states with @xmath3 particles @xmath4 can be generated by using the mlmer - srensen approach @xcite in ion traps , with @xmath3 up to 14 in experiments @xcite . to generate ghz states in a wide range of quantum systems ,
new mechanisms are highly wanted .
here we introduce a quantum zipper " to sew ghz states in a systematic way . ) .
5 , send a @xmath5 pulse to flip the second spin .
6 , send in a @xmath5 pulse to flip the fourth spin . 7 ,
introduce the same interaction as in step 4 among the red spins until achieving the state shown in this step .
8 , send a @xmath5 pulse to flip the third spin . repeat the steps 6 - 8 until zipping all spins in the ghz state , each time adding two spins to the ghz chain . for a ghz state with @xmath6 spins ,
we need in total @xmath7 @xmath5 pulses and one @xmath8 pulse . ]
the basic process for generating ghz states is shown in fig . [
we start with an all - down state and send a @xmath5 pulse to the first spin to flip it up .
then we send a @xmath8 pulse to prepare the second spin in a superposition of the up and down states .
the following step 4 is crucial .
we introduce a special interaction between the first three spins , which undergo opposite chiral spin wave rotations for @xmath9 and @xmath10 states , [ eq1 ] |||| , [ ea ] + the spin states move to the right for the states containing one up spin while they move to the left for the states containing two up spins .
the step 4 corresponds to the rotations @xmath11 and @xmath12 in eq.([eq1 ] ) .
we then send a @xmath5 pulse to the second spin and prepare the first three spins in a ghz state , as shown in the step 5 . following and repeating the steps 6 - 8 ( see fig.[scheme ] caption ) , we can zip the following spins into a ghz state .
in a previous work , we studied chiral photon rotations among three cavities by generating a synthetic magnetic field @xcite . the process in eq.([eq1 ] ) can be realized in a similar way by the following hamiltonian , @xmath13 where @xmath14 and @xmath15 are the raising and lowering operators for the @xmath16th spin and the summation over @xmath16 is cyclic .
the hamiltonian in eq.([hspin ] ) commutes with @xmath17 and the number of up spins is conserved .
we first investigate the dynamics in the subspace expanded by @xmath9 , @xmath18 and @xmath19 , in which the eigen frequencies are @xmath20 , @xmath21 and @xmath22 .
the corresponding eigenstates are the evolution of the initial state @xmath23 is @xmath24.\end{aligned}\ ] ] it is clear that at time @xmath25 , @xmath26 , and at time @xmath27 , @xmath28 .
we obtain the chiral spin wave rotation in eq .
( [ eq1]@xmath29 ) .
we can follow the same procedure to calculate the dynamics in the subspace with two up spins , @xmath10 , @xmath30 and @xmath31 .
the spin states move to the left , as shown in eq .
( [ eq1]@xmath32 ) . to understand this surprising result
, we try to know how @xmath10 evolves based on the knowledge that the state @xmath9 rotates to the right . by reversing our definition of up and down ,
the state @xmath10 becomes @xmath9 in the new basis . to express the hamiltonian in this upside down basis
, we need to make the replacement @xmath33 and @xmath34 , which results in @xmath35 .
the state evolves backward in time , _
i.e. _ , the spin wave moves to the left .
the key feature of the hamiltonian in eq.([hspin ] ) is the imaginary interaction strength between the spins .
previously we proposed the synthetic magnetic field for photons by oscillating the frequencies of three cavities that are coupled to the same spin @xcite . here
we consider three spins with oscillating frequencies coupled to the same cavity mode .
let us first consider two spins with frequencies being modulated with different phases .
the interaction hamiltonian can be written as @xmath36+h.c.,\ ] ] where @xmath29 is the annihilation operator of the cavity , @xmath37 , @xmath38 and @xmath39 are the modulation amplitude , frequency and phase for the @xmath16th spin .
we assume the central frequencies of the spins are the same as that of the cavity .
since @xmath40 where @xmath41 is the @xmath42th order bessel function of the first kind , we can expand the interaction hamiltonian @xmath43 , where @xmath44 + i^nj_n(-f)\hbar g [ \sigma^+_1 e^{in\phi_1}+\sigma^+_2 e^{in\phi_2}]a,\ ] ] the effective hamiltonian is @xcite , @xmath45\\ = h_0+i\frac{\hbar g^2}{\nu_d}\eta(\sigma_2^+\sigma_1 ^ --\sigma_1^+\sigma_2 ^ -).\ ] ] where @xmath46/n$ ] .
when @xmath47 , @xmath48 and @xmath49 , we obtain the imaginary interaction strength between the two spins .
we introduce the third spin with the same modulation amplitude @xmath37 and frequency @xmath38 but a different phase @xmath50 , and we set @xmath51 .
we obtain @xmath52 in eq .
( [ hspin ] ) with @xmath53 and @xmath54 . in the experiments ,
the transition frequencies of the spins can be modulated by the dynamic stark shift of a detuned light field with modulated intensities .
the spins can be atoms coupled to a one - dimensional waveguide that can mediate long range interactions between atoms @xcite . | a mechanism of a chiral spin wave rotation is introduced to systematically generate mesoscopic greenberger - horne - zeilinger states . |
hadroproductions of charmed particles have been measured in fixed - target experiments @xcite and indicate the large leading / non - leading asymmetries , i.e. asymmetries between production cross sections of leading particles and those of non - leading particles defined as @xmath11 the leading particle shares a valence quark in common with the incident hadron , while the non - leading one does not .
the leading particles are copiously produced at large feynman @xmath12 as compared with the non - leading particles in the forward region of the incident hadron .
this is called as leading particle effect .
for example , the asymmetry between @xmath13 and @xmath14 in @xmath15 interaction with a nucleon is defined as @xmath16 and increases from zero to nearly one with @xmath12 in the @xmath17 fragmentation region . in the conventional perturbative qcd at leading order ( lo ) ,
the factorization theorem predicts that @xmath18 and @xmath19 quarks are produced symmetrically and then fragment into @xmath6 and @xmath1 independently . as a consequence ,
lo perturbative qcd predicts no asymmetry in contrast with the experimental data .
the next to leading order ( nlo ) calculations generates asymmetries but they are much smaller than the data@xcite . to explain the leading / non - leading asymmetry , many attempts have been investigated : @xmath20 factorization@xcite , string fragmentation@xcite , intrinsic charm contributions@xcite , recombination process with beam remnants@xcite , recombination with participants in the hard scattering process@xcite , recombination with surrounding light quarks in projectile and target@xcite , recombination using valon concept@xcite , light quark fragmentation@xcite , meson cloud model@xcite and so on .
productions of @xmath2 and @xmath3 in @xmath21 @xcite and @xmath22 and @xmath5 collisions@xcite have been measured and analyzed in refs.@xcite and @xcite .
these models explain the leading / non - leading asymmetry of charmed hadrons successfully , however , there are some uncertainties in the perturbative qcd calculations as reviewed in refs.@xcite and @xcite .
hadronization of heavy quarks is still an open problem , since it is intrinsically non - perturbative process .
we have applied the covariant quark - diquark cascade model for hadroproductions of charmed mesons as well as hadrons composed @xmath23 and @xmath24 quarks@xcite . in the present paper , we investigate the hadroproductions of @xmath0 and @xmath1 and the leading / non - leading asymmetry in the framework of the constituent quark - diquark cascade picture and discuss the role of constituent diquark .
we assume that a baryon is composed of a constituent quark and a constituent diquark , and a meson is composed of a constituent quark and a constituent anti - quark .
the constituent quark ( diquark ) is a quark - gluon cluster consisting of the valence quark ( diquark ) , sea quarks and gluons .
hereafter , we simply call them as quark and diquark . when the collision between the incident hadrons @xmath25 and @xmath26 occurs , it is assumed that the incident hadrons either break up into two constituents or emit gluons followed by a quark - antiquark pair creation .
there are four interaction types : a ) non - diffractive dissociation , b ) and c ) single - diffractive dissociations of @xmath27 and @xmath25 , and d ) double - diffractive dissociation types as shown in fig.[fgr : intrctn_type ] .
the probabilities of these types to occur are @xmath28 and @xmath29 , respectively . here
we denote the quark - antiquark pair emitted from @xmath30 ( @xmath31 ) via gluons as @xmath32 ( @xmath33 ) .
the probabilities of @xmath32 ( @xmath33 ) to be @xmath34 and @xmath35 are are denoted by @xmath36 and @xmath37 , respectively .
collision : ( a ) non - diffractive dissociation type , ( b ) , ( c ) single - diffractive and ( d ) double diffractive dissociation type mechanisms.,height=151 ] the light - like variables of @xmath30 and @xmath31 are defined as follows : @xmath38 where @xmath39 is the center of mass energy of the incident hadrons @xmath25 and @xmath26 .
the momentum fractions of @xmath40 and @xmath41 for the process shown in fig.1b are given as follows : the momentum fraction of @xmath41 is fixed by the distribution function @xmath42 and the uniform distribution @xmath43 in the interval from zero to one as , @xmath44 then the momentum fractions of the incident particles @xmath25 and @xmath26 become as follows : @xmath45 @xmath46 where the mass shell condition is considered and transverse momenta are neglected .
analogous expressions describe the momentum fractions of @xmath40 and @xmath47 for the process shown in fig.1c , upon replacing superscripts @xmath48 and subscripts @xmath49 .
the single - diffractive dissociation type mechanisms result in diffractive peaks of spectra of the same kind of the incident particles @xmath25 and @xmath26 at @xmath50 .
the probability of the incident hadron to emit gluons followed by a quark - antiquark pair @xmath51 and the parameters @xmath52 and @xmath53 in ( [ eqn : hmaa ] ) are chosen to reproduce the diffractive peaks of the spectra such as @xmath54 and @xmath55 in @xmath56 and @xmath57 collisions at @xmath50@xcite .
the parameters are chosen as @xmath58 and @xmath59 . however , the main contributions to charmed hadrons come from the non - diffractive dissociation type mechanism
. the distribution functions of the constituents in the projectile @xmath30 composed of @xmath60 and @xmath61 are described as @xmath62 similarly , the distribution function of @xmath63 in @xmath41 is given as @xmath64 .
the dynamical parameters @xmath65 s in ( [ eqn : ha / a ] ) , which determine the momentum sharing of the constituents , are related to the intercepts of the regge trajectories as @xmath66@xcite . for flavor
anti - symmetric diquark @xmath67 $ ] and symmetric diquark @xmath68 , the dynamical parameters @xmath69}$ ] and @xmath70 are chosen as @xmath69}=\gamma_{[\ ] } ( \beta_q+\beta_{q'})$ ] and @xmath71 , respectively .
the commonly shared constituents of @xmath2 with the incident @xmath72 and @xmath17 beams are @xmath73-diquark , @xmath74 and @xmath74-quark respectively .
@xmath3 has a commonly shared @xmath75 quark with @xmath17 and @xmath76 beams and @xmath2 is a non - leading particle in @xmath76 beam .
these distributions of the constituents in the incident hadrons affect the shapes of spectra of produced hadrons significantly .
hadrons are produced on the chain between a constituent valence quark(@xmath77)(anti - quark(@xmath78 ) ) from the incident hadron @xmath79 and the constituent valence diquark(@xmath80)(quark(@xmath81 ) ) from the incident hadron @xmath82 for a meson - baryon collision as illustrated in fig.1 .
the small intrinsic transverse momenta of constituents in the incident hadrons are introduced by changing the direction of the incident constituents by an angle @xmath83 in the rest frame of the chain . here
we choose the distributions for @xmath84 as @xmath85 in the region @xmath86 . here
we choose very small intrinsic transverse momenta and use the value @xmath87 used in ref.@xcite in order to reproduce @xmath88 dependences of @xmath6 mesons .
we take account of the annihilation of the valence antiquark from beam and valence quark from target in a slightly different way from the one pointed out in refs . @xcite and @xcite .
when the incident constituents in @xmath25(@xmath41 ) and @xmath82 are quark and its anti - quark , it is assumed that the annihilation process occurs with the probability @xmath89 and the non - annihilation process occurs with @xmath90 .
let us consider the case of @xmath91 collision .
the annihilation process @xmath92 occurs with the probability @xmath89 , when @xmath75 in @xmath17 interacts with @xmath93 in @xmath55 .
branching ratios of @xmath94 and @xmath95 are chosen to be equal to @xmath96 and @xmath97 for the channels allowed energetically .
the produced @xmath98 and @xmath63 in ( [ eqn : anni ] ) are supposed to be non - free in terms of hadronization mechanism .
thus it is assumed that @xmath98 has tendency to be produced in the forward direction of @xmath75 in @xmath17 beam .
therefore , here we choose the distributions for @xmath99 as @xmath100 in the region @xmath86 , where @xmath101 is the angle between the directions of @xmath75 and @xmath98 . to reproduce the @xmath102
, we choose the value @xmath103 and use a value of @xmath104 which roughly means @xmath105 @xcite .
+ hadrons are produced on the cascade chain by processes @xmath106)+\overline{[q'q '' ] } , b(q\{q'q''\})+\overline{\{q'q''\}},\\ \label{eqn : cscda2b } \overline{[q'q '' ] } & \rightarrow & \overline{b}(\bar{q}\overline{[q'q'']})+q,\\ \label{eqn : cscda2 m } \overline{[q'q '' ] } & \rightarrow & m(q\bar{q}')+\overline{[qq '' ] } , m(q\bar{q}')+\overline{\{qq''\}},\\ \label{eqn : cscds2b } \overline{\{q'q''\ } } & \rightarrow & \overline{b}(\bar{q}\overline{\{q'q''\}})+q,\\ \label{eqn : cscds2 m } \overline{\{q'q''\ } } & \rightarrow & m(q\bar{q}')+\overline{[qq '' ] } , m(q\bar{q}')+\overline{\{qq''\ } } , % \label{eqn : cscds2m}\end{aligned}\ ] ] where @xmath63 denotes @xmath107 and @xmath108 and @xmath109 $ ] does @xmath110 , [ us],[uc],[ds],[dc ] , [ sc]$ ] and so on .
the symbols @xmath111 $ ] and @xmath112 denote the flavor anti - symmetric and symmetric diquarks , respectively .
meson production probabilities from @xmath113 } $ ] and @xmath114 are @xmath115 } $ ] and @xmath116 , respectively .
the parameter @xmath117 is the probability of baryon productions from a quark and is related to the baryon productions from @xmath118 and @xmath119 beams resulting in @xmath120.@xcite the processes ( [ eqn : cscda2 m ] ) and ( [ eqn : cscds2 m ] ) contribute to production of baryons which share no valence quark or diquark in common with the incident hadron .
although both @xmath121 and @xmath122 share no valence quark common with the incident proton , the spectrum of @xmath121 is harder than that of @xmath122 in the proton fragmentation region . in our model
this is described as follows : after emitting @xmath119 mesons through these processes , the incident diquark in the incident proton converts into @xmath123 diquark and produce @xmath121 by the process ( [ eqn : cscds2b ] ) .
the difference between the spectra of @xmath121 and @xmath122 in @xmath57 collision is described by the processes ( [ eqn : cscda2m]),([eqn : cscds2 m ] ) and ( [ eqn : cscds2b ] ) and the parameters are chosen as @xmath124}=\eta_{\{\}}=0.25$].@xcite for the cascade process @xmath125 in the beam ( target ) side , the distribution function of light - like fraction of @xmath126 is assumed as @xmath127 and the light - like fraction of @xmath126 becomes @xmath128 .
we fix the transverse momentum of the hadron @xmath129 as a difference between two random points @xmath130 with their lengths fixed by the distribution function @xmath131 in @xmath132 space , where @xmath133 denotes the mass of @xmath129 .
the parameter @xmath134 is fixed from the experimental data on @xmath132 distributions of pions@xcite .
we use the values of @xmath135 and @xmath136gev@xmath137 , from previous analyses@xcite .
we now see how the cascade processes occurs between the incident constituents , for example , @xmath63 from @xmath25 @xmath138 and @xmath139 from @xmath26 .
we redefine the light - like fractions of @xmath63 and @xmath139 in the rest frame of the incident constituents @xmath63 and @xmath139 .
the momentum sharing of the cascade process @xmath140 from @xmath63 with @xmath141 and @xmath139 with @xmath142 takes place as follows@xcite : using eq.([eqn : fqq ] ) with @xmath143 , we fix the light - like fractions of @xmath126 and @xmath144 as @xmath145 and @xmath146 , respectively and put @xmath147 .
the transverse momentum of @xmath129 is fixed by eq.([eqn : pt2 ] ) .
then @xmath148 is fixed from the onshell condition @xmath149 , where @xmath150 is the subenergy of the incident @xmath63 and @xmath139 system .
the transverse momentum of @xmath151 is @xmath152 .
the light - like fraction of @xmath153 is decreased to @xmath154 . if the energy of the @xmath155 system is enough to create another hadron , the cascade process such as @xmath156 takes place in the opposite side . in the final step ,
we assume that the constituents recombine into one or two hadrons according to the processes : @xmath157 \rightarrow b(q[q'q '' ] ) , ~ \{qq'\ } + \overline{[q''q ' '' ] } \rightarrow m(q \overline{q '' } ) + m(q'\overline{q'''})$ ] and so on .
the momenta of the recombined hadrons are the sum of those of the final constituents and are offshell . in the cms of incident hadrons @xmath25 and @xmath26 , we have energy - momentum conservation relations @xmath158 and @xmath159 , where @xmath160 denotes the three momentum of @xmath161-th produced particle . in order to put recombined hadrons onshell , we multiply the three momenta of all produced hadrons by a common factor @xmath162 so that the summation @xmath163 would be equal to @xmath164@xcite .
we take into account pseudoscalar @xmath165 , vector @xmath166 and tensor @xmath167 mesons .
the production probabilities for them are assumed to be @xmath168 and @xmath169 .
for baryons we consider lower lying baryons : octet ( @xmath170 ) and decuplet ( @xmath171 ) baryons composed of @xmath172 and @xmath173 flavors and the corresponding ones with charm flavor .
octet and decuplet baryons are described as @xmath174 @xmath175 we assume the production probabilities for octet and decuplet baryons from symmetric diquarks to be @xmath176 and @xmath177 , respectively .
then the emission probabilities of individual cascade processes are determined from the above probabilities . for examples ,
the probabilities to produce @xmath178 and @xmath179 from a @xmath180 quark are as follows : + @xmath181 where the factors 1/3 and 1/2 are flavor su(4 ) factors .
directly produced resonances decay into stable particles .
the mixing angle of symmetric and anti - symmetric diquarks in octet baryon @xmath182 and the pair creation probabilities of @xmath183 and @xmath184 are determined from the analyses of non - charmed hadron productions in @xmath185 and @xmath57 collisions . in ref.@xcite , we used @xmath186 and @xmath187 and @xmath188 . to reproduce the data on @xmath189
cross section , we choose @xmath190 and @xmath191@xcite .
the dynamical parameters @xmath192}$ ] and @xmath193 as well as the mixing angle of symmetric and anti - symmetric diquarks in octet baryon is related to the @xmath194 spectra in proton fragmentation region and are chosen as @xmath192}=1.5 $ ] and @xmath195@xcite . we now consider hadron - nucleus interaction .
we regard the hadron - nucleus collision as a sum of hadron - nucleon collisions and neglect the intra - nuclear cascade of produced hadrons .
the projectile hadron successively collides with nucleons inside the nucleus .
the probability @xmath196 that the incident hadron @xmath197 collides with @xmath198 nucleons is calculated by the glauber - type multiple collision model@xcite . the number @xmath198 is determined by the distribution of nucleons in the nucleus and the cross section of the incident hadron with a nucleon @xmath199 .
we take the values @xmath200 and @xmath201 mb in the hadron - nucleus collisions .
the nucleon number density of the nucleus with a mass number @xmath25 is assumed as @xmath202 where we choose @xmath203 ( fm ) and @xmath204 ( fm ) .
distributions of @xmath205 and @xmath206 in a ) @xmath55 , b ) @xmath207 , c ) @xmath17 , d ) @xmath76 and charmed mesons distributions in e ) @xmath208 and f ) @xmath17 collisions with @xmath209 and @xmath210 targets .
, height=491 ] then it is assumed that @xmath211 @xmath212 and @xmath213 interactions take place , where @xmath214 is @xmath215 state emitted from @xmath197 as @xmath41 discussed in section [ sec : quark_dstb ] . at each collision , the projectile hadron @xmath197 looses its momentum . from the data on @xmath25 dependences of the spectra of the projectile hadrons , we put the degradation rate of the momentum of @xmath197 as @xmath216 which is used in ref.@xcite .
in this section , we give the results of our model for charmed hadron productions in @xmath217 and @xmath76 beam interactions with nuclei .
to see the target mass number @xmath218 dependence , we show the @xmath12 distributions of @xmath0 and @xmath1 productions in @xmath217 and @xmath76 interactions with @xmath209 and @xmath210 targets at @xmath219 gev / c in fig.2 .
all spectra increase with @xmath25 in the region @xmath220 .
the spectra of @xmath2 become soft with @xmath25 in the baryon beam fragmentation region .
this is due to the momentum degradation of the incident baryon through the multiple interactions with nucleons in the target .
distributions of @xmath205 and @xmath206 in a ) @xmath55 , b ) @xmath207 , c ) @xmath17 and d ) @xmath221 collisions with @xmath134 and @xmath222 targets in arbitrary unit . the experimental data are taken from ref .
height=302 ] the spectra of @xmath3 scarcely depend on @xmath25 in the baryon beam fragmentation regions . in meson beam fragmentation regions ,
both @xmath2 and @xmath3 slightly depend on @xmath25 and become soft with increasing @xmath25 . in baryon beam fragmentation regions , @xmath25-dependences of the spectra of @xmath1 mesons are smaller than those for @xmath6 mesons .
the shapes of the spectra for @xmath6 mesons in baryon beam fragmentation regions are steep as compared with those of @xmath1 mesons in baryon and @xmath6 and @xmath1 mesons in meson beam fragmentation regions . in fig.3 , we show the @xmath12 distributions of @xmath2 and @xmath3 in @xmath217 and @xmath76 beam interactions with @xmath134 and @xmath222 targets in arbitrary unit and compare the results with the experimental data e781@xcite .
the shapes of @xmath2 and @xmath3 in @xmath72 and @xmath17 fragmentation regions are well explained by our model . in our model , the shapes of the spectra are strongly related to the distributions of valence constituents in the incident hadrons , which are characterized by the parameters @xmath223 s in ( [ eqn : ha / a ] ) .
diquarks in the incident baryons have a tendency to bring a large momentum fraction of the incident baryons . the valence constituents @xmath74 and @xmath75 in @xmath17 beam have the same distribution i.e. energetic @xmath74 and wee @xmath75 or vice versa . the valence constituents @xmath74 and @xmath75 in @xmath17 are harder than the valence quarks in @xmath207 and @xmath55 beams .
the @xmath224 diquark in @xmath207 is harder than the @xmath73 diquark in @xmath55 .
the valence @xmath225 diquark in @xmath55 beam is the hardest among the commonly shared constituents between @xmath2 and the incident @xmath217 and @xmath76 beams .
therefore @xmath12-distribution of @xmath2 in proton beam is harder than those in other hadron beams . and
@xmath206 in a ) @xmath55 b ) @xmath207 , c ) @xmath17 and d ) @xmath221 collisions with @xmath134 and @xmath222 targets . the experimental data are taken from ref .
height=302 ] between leading and non - leading charmed mesons in a ) @xmath226 and b ) @xmath21 collisions .
asymmetries between non - leading charmed mesons in c ) @xmath226 and d ) @xmath21 collisions .
the experimental data are taken from ref .
, height=302 ] the asymmetries between the spectra @xmath2 and @xmath3 in @xmath227 and @xmath228 collisions are shown in fig.4 .
the results are in good agreement with the experimental data except for @xmath17 beam . for @xmath17 beam , @xmath2 and @xmath3 share the common valence constituents @xmath74 and @xmath75 , respectively .
therefore the same spectra are expected for @xmath2 and @xmath3 productions in the region of @xmath17 beam fragmentation . in our model , however , the excess of @xmath3 over @xmath2 is seen at @xmath229 .
this is due to the fact that energetic @xmath2 and @xmath3 are produced in the @xmath74-quark - diquark and the @xmath75-antiquark - quark chains , respectively .
the former chain is slightly shifted to the target fragmentation region as compared with the latter one due to the large momentum fraction of the diquark@xcite .
therefore our model gives the negative asymmetry @xmath230 at @xmath229 .
the valence diquark in @xmath207 beam contributes to @xmath2 production via a processes such as @xmath231 in @xmath207 fragmentation region .
the contributions to @xmath2 production from the valence diquarks dominate the contributions from the valence quarks in @xmath207 beam . in fig.5
, we show the results of asymmetries between charmed and anti - charmed mesons in @xmath55 and @xmath207 interactions with @xmath232 and @xmath222 targets .
the asymmetries @xmath233 and @xmath234 are small as compared with the asymmetries @xmath235 and @xmath236 .
although both @xmath237 and @xmath238 in @xmath55 and @xmath239 and @xmath240 in @xmath207 beam are non - leading particles , the results of asymmetries @xmath241 and @xmath242 are non - zero and have large values .
this is because in our model @xmath1 mesons are produced mainly from the chain between the valence diquark in the incident baryon and the valence quark in the target nucleon .
for example , @xmath238 in proton fragmentation region is produced via a process such as @xmath243 .
the more energetic the valence diquark becomes , the more copiously @xmath1 mesons are produced in the baryon fragmentation region . in our model the valence @xmath224 diquark in @xmath207 beam has a larger momentum fraction than the valence diquarks in @xmath55 beam .
consequently the asymmetries @xmath244 are large as compared with @xmath245 .
the asymmetry @xmath246 is larger than other charmed meson asymmetries in @xmath55 and @xmath207 beams .
this is also due to the large momentum fraction of the valence @xmath224 diquark in @xmath207 . in the target fragmentation region ,
both @xmath6 and @xmath1 mesons increase similarly with the mass number @xmath25 and the asymmetries show small mass number dependences .
since the spectra of @xmath6 mesons increase with @xmath25 and those of @xmath1 mesons change little , there are considerable target mass number dependence in the asymmetries @xmath247 at @xmath248 .
\1 . we have examined hadroproductions of charmed baryons in the frame work of the quark - diquark cascade model .
their spectra are well explained by using the values of the dynamical parameters fixed from regge intercepts .
the large leading / non - leading asymmetries @xmath235 and @xmath249 are naturally explained .
the valence diquark in the incident baryon , which brings a large momentum fraction of the beam , plays an important role for @xmath1 productions as well as for @xmath2 production@xcite .
the non - leading anti - charmed meson productions in baryon beams are largely affected by the valence diquark distribution in the incident baryons through the processes ( 13 ) and/or ( 15 ) .
\3 . in meson beam fragmentation regions , @xmath25-dependences of the spectra of non - leading charmed mesons
are small as compared with those for leading charmed mesons .
there are considerable @xmath25-dependences in @xmath2 and @xmath6 meson productions , while those for @xmath3 and @xmath1 productions are small in baryon beam fragmentation regions .
there are discrepancies between the data and calculations of asymmetries @xmath230 at @xmath17 fragmentation region as seen above .
piskounova discussed this problem successfully by introducing the transfer of string junction in quark - gluon string model@xcite . to explain the experimental data in our model , it is necessary , for example , a energetic @xmath74 quark from @xmath17 beam and a wee @xmath250 diquark from target nucleon to recombine into a @xmath2 . here
the @xmath250 diquark is converted from the valence diquark in the target nucleon by emitting several mesons through the processes ( 13 ) and/or ( 15 ) .
the significant value of the asymmetry @xmath230 as well as the small value of @xmath251 in meson beam fragmentation is a challenging problem .
this work was partly supported by grant of the academic frontier project promoted by the ministry of education , culture , sports , science and technology , japan .
0 m. adamovich et al . , ( wa82 ) : phys .
b305 * , 402(1993 ) ; m. adamovich et al . ,
( wa92 ) : nucl . phys . * b495 * , 3(1997 ) ; g. a. alves et al .
, ( e769 ) : phys . rev . lett . * 72 * , 812(1994 ) ; g. a. alves et al . ,
( e769 ) : phys . rev . lett . *
77 * , 2388(1996 ) ; e. m. aitala et al . , ( e791 ) :
411b * , 230(1997 ) e. m. aitala et al . , ( e791 ) : phys .
* b371 * , 157(1996 ) p. nason , s. dawson , k. ellis : nucl .
* b327 * , 49(1989 ) ; w.beenakker , r.meng , g.a.schulter , j.smith , w.l.van neerven : nucl . phys . *
b351 * , 507(1991 ) s.frixione , m.l.mangano , p.nason , g.ridolfi : nucl . phys .
* b431 * , 453(1994 ) m.g.ryskin , yu.m.shabelski , a.g.shuvaev : z. phys . *
c69 * , 269(1996 ) ; m.g.ryskin , a.g.shuvaev , yu.m.shabelski : phys .
64 * , 120(2001 ) h. u. bengtsson , t. sjstrand : comput .
. commun . * 46 * , 43(1987 ) ; e. norrbin , t. sjstrand : phys.lett . * b442 * 407(1998 ) o. i. piskounova : nucl .
. suppl . * 50 * , 179(1996 ) ; nucl .
proc . suppl . *
93 * , 144(2001 ) ; g. h. arakelyan , p. e. volkovitsky : z. phys . * a353 * , 87(1995 ) ; g. h. arakelyan : phys.atom.nucl .
* 61 * 1570(1998 ) r. vogt , s. j. brodsky , p. hoyer : nucl . phys .
* b383 * , 643(1992 ) ; r. vogt , s. j. brodsky : nucl .
b438 * , 261(1995 ) ; t.gutierrez , r. vogt : nucl . phys . *
b539 * , 189(1999 ) ; r. vogt : z. phys . *
c71 * , 475(1996 ) g. anzivino et al . :
nuovo cim .
* 107a * , 955(1994 ) ; + n. armesto , c. pajares , yu . m. shabelski : hep - ph/9506212 , 1995 v. a. bednyakov : mod .
a10 * , 61(1995 ) e. cuautle , g. herrera , j. magnin : eur .
j. * c2 * , 473(1998 ) a.k.likhoded , s.r.slabospitsky : phys . atom . nucl . *
60 * , 981(1997 ) ; phys .
nucl . * 62 * , 693(1999 ) ; phys . atom .
* 65 * , 127(2002 ) ; a.k.likhoded , s.r.slabospitsky : phys .
* 65 * , 127(2002 ) ; yu.m.shabelski : hep - ph/0011032 c. avila , j. magnin , l.m.mendoza-navas : hep - ph/0307358 e. braaten , yu jia , t. mehen : phys.rev.lett .
* 89 * 122002(2002 ) e. braaten , m. kusunoki , yu jia , t. mehen : hep - ph/0304280 ; t. mehen : hep - ph/0306178 r. rapp , e. v. shuryak : phys
* d67 * , 074036(2003 ) r. c. hwa : phys .
* d51 * , 85(1995 ) ; k. p. das , r. c. hwa : phys . lett . *
68b * , 459(1977 ) j.dias de deus , f.o.duraes : eur .
j. * c13 * , 647(2000 ) f. carvalho , f.o .
duraes , f.s.navarra , m.nielsen : phys .
84 * 5434(2001 ) m. adamovich et al .
( wa89 ) : eur .
j. * c8 * , 593(1999 ) f. g. garcia et al .
( e781 ) : phys
b528 * , 49(2002 ) j. c. anjos , g. herrera , j. magnin , f.r.a .
simao : phys . rev . *
d56 * , 394(1997 ) ; o. i. piskounova : hep - ph/0202005 s. frixione , m. l. mangano , p. nason , g. ridolfi : adv .
direct in high energy phys . *
15 * , 609(1998 ) [ hep - ph/9702287 ] j. c. anjos , e. cuautle : hep - ph/0005057 k. kinoshita , h. noda , t. tashiro : prog .
68 * , 1699(1982 ) ; prog .
* 68 * , 2086(1982 ) ; t. tashiro , h. noda , k. kinoshita , c. iso : z. phys .
* c35 * , 21(1987 ) k. kinoshita , h. noda , t. tashiro , j. nagao : _ int .
symp . on high energy nuclear collisions and quark gluon plasma _ , kyoto 1991 , edited by m. biyajima , h. enyo , t. kunihiro , o. miyamura(world science , singapore 1993)245 ; h. noda , t. tashiro : z. phys .
* c45 * , 85(1989 ) t. tashiro , h. noda , s. nakariki , k. ishii , k. kinoshita : eur .
j. * c2 * , 733(1998 ) t. tashiro , h. noda , k. kinoshita , s. nakariki : eur .
j. * c24 * 573(2002 ) ; hep - ph/9810284 h. minakata : phys . rev . *
d20 * , 1656(1979 ) g. cohen - tannoudji , a.el hassouni , j. kalinowski , r. peschanski : phys . rev . *
d19 * , 3397(1979 ) k. kinoshita , a. minaka : prog .
theor . phys .
* 81 * , 183(1989 ) | we discuss the hadroproductions of @xmath0 and @xmath1 in the framework of the constituent quark - diquark cascade model taking into account the valence quark annihilation .
the spectra of @xmath2 and @xmath3 in @xmath4 and @xmath5 collisions are well explained by the model using the values of parameters used in hadroproductions of @xmath6 and @xmath1 .
it is shown that the role of valence diquark in the incident baryon is important for @xmath1 productions as well as for @xmath2 production . * * ' '' '' leading effects in hadroproductions of @xmath2 and @xmath6 from constituent quark - diquark cascade picture + tsutomu tashiro@xmath7 , shin - ichi nakariki,@xmath7 , hujio noda@xmath8 , kisei kinoshita@xmath9 and shuxin lan@xmath10 + @xmath7department simulation physics , okayama university of science , ridai - cho , okayama 700 - 0005 , japan + @xmath8department of mathematical science , ibaraki university , bunkyou , mito 310 - 0056 , japan + @xmath9physics department , kagoshima university , korimoto , kagoshima 890 - 0065 , japan + @xmath10department physics , okayama university , okayama 700 - 0005 , japan + |
in the preceeding paper @xcite , referred to hereafter as paper i , we presented global representations of the three ( @xmath1 , @xmath2 , and @xmath3 ) complex - valued potential - energy surfaces of the metastable states of h@xmath0o@xmath4 that underlie dissociative electron attachment to water .
this paper is concerned with the calculation of the cross sections for that physical process .
prior experimental and theoretical results @xcite have characterized the various breakup channels and determined the spatial symmetries of the three metastable electronic states of h@xmath0o@xmath4 , the @xmath1 , @xmath2 , and @xmath3 electronic feshbach resonances , which are responsible for production of h@xmath4 and o@xmath4 . as explained in ref .
@xcite and paper i , the energetically lowest h+oh@xmath4 channel does not directly correlate with any of the three feshbach states .
we therefore conclude that oh@xmath4 production is due to nonadiabatic effects .
we pursue this problem theoretically using a coupled born - oppenheimer treatment of the nuclear motion .
the first task , which was described in paper i , is the construction of three - dimensional , complex - valued potential - energy surfaces for these three states , which have a negative imaginary component due to the finite probability of electron autodetachment back to h@xmath0o+@xmath5 .
these complex - valued potential - energy surfaces , which are functions of the nuclear geometry @xmath6 , are defined as @xmath7 where @xmath8 is the resonance position and @xmath9 is the width of the resonance , which is related to the lifetime by @xmath10 .
( we use atomic units throughout this paper . ) the present article , which we label paper ii , is concerned with the use of these potential curves within the local complex potential ( lcp ) model @xcite to calculate the nuclear dynamics leading to dissociation .
the analysis of the dynamics yields the dea cross section as a function of incident electron energy .
we must account for two major nonadiabatic physical effects in calculating the quantum dynamics of the nuclei .
as described in paper i , the three potential - energy surfaces have several degeneracies that lead to coupling among them .
first , the @xmath1 and @xmath2 states become members of a degenerate @xmath11 pair in linear geometry , and for this reason there will be renner - teller coupling between them .
we expect this coupling to be relevant for dea via the @xmath2 state , because the gradient of its potential - energy surface will cause the system to move toward linear geometry after the electron attaches .
second , there is a conical intersection @xcite between the @xmath3 and @xmath2 states which leads to coupling between them .
for this reason , as described in paper i , we constructed a set of diabatic @xmath3 and @xmath2 surfaces , along with a coupling term , which we use in the calculations presented in this paper . in fig .
[ c2vfig ] , we show the real parts @xmath8 of the constructed potential - energy surfaces along a two - dimensional cut which includes the equilibrium geometry of the neutral ( @xmath12 ; @xmath13 ) .
the degeneracies that lead to the nonadiabatic effects listed above can be seen in this figure .
the two - dimensional cut depicted is that for which the two oh bond lengths are equal ( @xmath14 ) , corresponding to @xmath15 symmetry .
( in @xmath15 symmetry , the adiabatic and diabatic @xmath2 and @xmath3 surfaces coincide . )
the backside of this cut lies at @xmath12 , which is the equilibrium value of the bond lengths in neutral h@xmath0o , and is marked with solid lines .
the surfaces extend forward in fig .
[ c2vfig ] along the symmetric stretch direction to geometries at which @xmath16 . the conical intersection comprises the set of points along which the @xmath2 and @xmath3 surfaces intersect . the renner - teller degeneracy between the @xmath1 and @xmath2 states occurs at @xmath17=180@xmath18 .
although fig .
[ c2vfig ] shows only one cut of the potential - energy surfaces , and only their real part , it is useful for introducing certain features of these surfaces and the dynamics that will result .
dissociative attachment via the lower @xmath1 and @xmath2 states leads primarily to the product h@xmath4+oh ( x @xmath11 ) .
the two oh bond lengths for such an arrangement are unequal , and therefore this product arrangement can not be seen in fig . [ c2vfig ] .
however , we can see that at the equilibrium geometry of the neutral , the @xmath1 surface is relatively flat with bend , while the @xmath2 surface slopes steeply downward toward linear geometry ( @xmath17=180@xmath18 ) . as a result
, the dynamics beginning on the @xmath2 surface will lead toward linear geometry , and we expect that the renner - teller coupling between these two states will be more important for dea via the @xmath2 than via the @xmath1 state . the channel h@xmath0+o@xmath4 is the minor channel for dea via the @xmath1 and @xmath2 states , but the major channel for the @xmath3 state .
we can see why this is the case from fig .
[ c2vfig ] ; the gradient of the @xmath3 ( 2 @xmath19 ) surface leads downward toward the conical intersection , where the system may make a nonadiabatic transition to the lower surface and access the clearly visible h@xmath0+o@xmath4 well on the 1 @xmath19 ( lower cone ) surface .
the 2 @xmath19 surface does not have a low energy asymptote in this geometry , instead correlating to o@xmath4+h@xmath0 ( @xmath20 ) .
the outline of this paper is as follows . in sec .
[ priorsect ] , we summarize previous experimental and theoretical work on this problem . in sec .
[ lcpsect ] , we present the local complex potential model , which forms the foundation of our theoretical implementation .
the hamiltonian for the rovibrational nuclear motion of a triatomic molecule , and the additional terms which arise when the renner - teller effect is included , are described in sec .
[ triatomicsect ] . in sec .
[ mctdhsect ] we describe the multi - configuration time - dependent hartree ( mctdh ) method , which we use to calculate the nuclear dynamics , and the formalism for calculating the dea cross sections . in sec .
[ resultssect ] we present the final results of this study : cross sections , as a function of incident electron energy , resolved into the final rovibrational product states .
dissociative electron attachment to water molecules has been the subject of previous experimental investigation , starting as early as 1930 @xcite , and as recently as the past year ( 2006 ) @xcite .
early experiments on dissociative electron attachment to h@xmath0o focused mainly on the identification of the negative ion species formed , the measurement of the total cross sections , and the energy locations of the structures in the resonance process @xcite .
buchelnikova @xcite and schultz @xcite established that the main products of dissociative electron attachment to water are @xmath21 and @xmath22 , with the production of @xmath22 being almost ten times smaller than that of @xmath21 at lower energies , but with @xmath22 dominating at higher electron - impact energies .
both compton and christophorou @xcite and melton @xcite carried out comprehensive studies of negative ion formation in water and measured absolute cross sections for dea .
three resonance peaks were observed .
@xmath21 production was observed at approximately 6.5 ev and 8.6 ev , with the second peak much less intense than the first .
the species @xmath22 was observed in increasing intensities in three peaks at 7.0 ev , 9.0 ev , and 11.8 ev @xcite .
the species oh@xmath4 is also observed in the dissociative electron attachment experiments , though at an intensity one order of magnitude below the minor o@xmath4+h@xmath0 channel , which is itself observed at an intensity approximately one order of magnitude lower than the dominant h@xmath4+oh channel .
melton @xcite argued that oh@xmath4+h was a true channel of dissociative electron attachment to h@xmath0o molecules , while in subseqent studies ( e.g. , ref .
@xcite ) it was argued that oh@xmath4 is produced by dea to water clusters [ h@xmath0o]@xmath23 .
the question of oh@xmath4 production has been reexamined in the recent experimental study of fedor _ _ et al.__@xcite .
these authors have concluded that , indeed , it is a direct product of dissociative electron attachment to water .
this minor channel is not examined in the present treatment , and no mechanism has , as yet , been advanced .
the effects of isotopic substitution have also been an issue of some debate .
the replacement of h@xmath0o by d@xmath0o as the molecular target has the effect of nearly doubling the reduced masses corresponding to oh ( od ) bond motion .
one would expect , at least in a simple one - dimensional picture , that the nuclear dynamics may be substantially altered by such replacement , and in particular , the time to dissociation is increased . a longer dissociation time allows a greater amount of electron autodetachment to take place ; therefore , performing the same experiment with different isotopic variants provides information on the lifetime of the electronic state involved .
the cross sections for dea via both h@xmath0o and d@xmath0o were measured , compared , and discussed in detail by compton and christophoreau @xcite .
a smaller peak cross section for d@xmath4 production than for h@xmath4 production via the lowest - energy @xmath1 state was observed . on the basis of these results
, these authors derived an approximate lifetime of 2.1 @xmath24 10@xmath25 seconds for the lowest - energy @xmath1 feshbach resonance .
we have already published an initial study of dea via this state @xcite , which arrived at results and conclusions much different from those of ref .
the calculated results yielded a higher peak cross section for d@xmath4 production via the @xmath1 resonance than for h@xmath4 production , and a similar energy - integrated cross section , in stark contrast to the results of compton@xcite .
the calculations indicated a larger lifetime of 10.9 @xmath24 10@xmath25 seconds for the @xmath1 state , and the nuclear dynamics that we calculated indicated that only a small portion of the dissociating anion flux is lost to autodetachment .
the recent experimental results of fedor _ _ et al.__@xcite have substantially resolved this controversy .
these authors obtain results different from those of ref .
@xcite , reversing the trend in peak heights for h@xmath4 versus d@xmath4 production via the @xmath1 resonance .
they observe a higher peak for d@xmath4 production than for h@xmath4 production , which brings the current experimental and theoretical results into qualitative agreement . although the peak heights provide considerable information about the physical process of dissociative electronic attachment to water , further information
is gained by resolving the angular dependence of the fragments produced , and the final ( ro)vibrational state of the diatomic fragment .
a series of measurements by trajmar and hall @xcite and belic , laudau , and hall @xcite revealed the energy and angular dependence of @xmath21 in dissociative electron attachment to @xmath26 .
the determination of the angular dependence aided the assignment of the spatial symmetries of the three resonant states , @xmath27 , @xmath28 , and @xmath29 , which had previously been misassigned . by resolving the kinetic energy of the h@xmath4 fragment ,
this experiment yielded information about the vibrational and rotational state distribution of the oh fragments .
curtis and walker @xcite measured cross sections for dissociative electron attachment to d@xmath0o and obtained two important results . by measuring the kinetic energy of recoil of the d@xmath4 fragments produced , these researchers established that both ground state od ( @xmath11 ) and excited od ( @xmath30 ) accompany the d@xmath4 anions produced within the third resonance peak , and that the three - body breakup channel d@xmath4+d+o is observed toward the high - energy tail of the second peak .
the experimental studies determined that there are three metastable electronic resonance states of the h@xmath0o@xmath4 anion , the @xmath1 , @xmath2 , and @xmath3 , which are primarily responsible for dissociative electron attachment to water .
these three electronic states correspond to the three peaks seen in the experimental cross sections . although the third peak is not obvious in the h@xmath4 cross sections , it is present , though much smaller than the first and second peaks .
several salient features of the early experiments suggest that the nuclear dynamics of this process may hold some surprises . for dissociative attachment through the @xmath1 resonance ,
the cross section for producing @xmath31 is roughly 40 times larger at its peak than the cross section for producing the energetically favored products , @xmath32 @xcite .
the lowest - energy atom / diatom arrangement , @xmath33 , is produced in even smaller quantities .
in addition , the branching ratios for the different product states vary greatly depending on which feshbach resonance is formed by the attachment .
these observations indicate that the products of this reaction are determined by the dynamics of the process itself rather than by the energetics of the possible product channels , and that moreover those dynamics are different for each of the resonance states of the water anion .
the detailed experiments of belic , landau and hall @xcite in 1981 indicated that the dissociation dynamics involve correlated motion among multiple degrees of freedom . for instance , the channel producing @xmath31 through the @xmath1 resonance state is accompanied by extensive vibrational excitation of the oh fragment . therefore , given the competition between dissociation channels and the observed product vibrational excitation , one expects that the dynamics of dissociative attachment to this molecule are intrinsically polyatomic , and can only be described theoretically by a treatment using the full dimensionality of nuclear motion .
compared with the large number of experimental measurements , detailed theoretical work on dissociative electron - water collisions has been relatively scarce .
the paucity of theoretical work on da stems from the fact that , in water , da proceeds , not through tunneling shape resonances , but through feshbach resonances that involve changes in the electronic structure of the target .
early theoretical work focused on the electronic structure @xcite and configuration - interaction @xcite calculations on various states of @xmath34 that are possible resonances .
these calculations , together with experimental observations , formed the basis of the assignment of the three feshbach resonances that are responsible for electron - impact dissociation of water in the gas phase .
contemporary theoretical work has included _ ab initio _
complex kohn @xcite and @xmath35-matrix @xcite calculations , at the equilibrium nuclear geometry , of the resonances and excitation cross sections into low - lying dissociative electronic states .
more recently , gorfinkiel , morgan , and tennyson @xcite carried out @xmath35-matrix calculations of dissociative excitation of water through the four lowest excited states ( the @xmath36 and @xmath37 states ) . a limited study of the effects of nuclear motion were included in that work by increasing one of the oh bonds while keeping the equilibrium hoh bond angle and the other oh bond length constant . the only theoretical work on the dynamical aspects of dissociative electron attachment to water
are earlier classical trajectory analyses based on either repulsive @xcite or attractive @xcite model resonace surfaces .
we previously reported calculations of the cross sections for dissociative attachment through the lowest - energy @xmath1 resonance @xcite that incorporated a full quantum treatment of the nuclear motion of the resonant state .
that study found good agreement with experiment for dissociative attachment through the lowest resonance state ( @xmath1 ) of the water anion to produce h@xmath4 , and it established that the associated dynamics are intrinsically polyatomic and thus can not be described successfully by one - dimensional models .
the present treatment supersedes our earlier study and extends the treatment to include the higher resonance states as well .
we have recently presented a qualitative study @xcite of the potential - energy surfaces for the three feshbach resonances , which demonstrated that for these metastable , anion states , there exist numerous intersections and degeneracies within the adiabatic manifold .
this study identified the conical intersection between the @xmath2 and @xmath3 states , as well as a novel degeneracy between the @xmath3 feshbach resonance and a @xmath3 shape resonance .
this degeneracy defines a branch seam , and the two resonance energies are seen to comprise two components of a double - valued adiabatic potential - energy surface .
this seam and the resulting dynamics may have an effect upon the three - body , h+h+o@xmath4 cross section , although we do not include it in the present treatment . finally , in a separate publication @xcite , we derived a `` constant - eigenmode approximation '' and used it to calculate the angular dependence of the h@xmath4 fragment production @xcite via the @xmath1 resonance .
we found excellent agreement with the results of belic , landau and hall @xcite , and demonstrated that the observed angular dependence is a result of partial - wave mixing in the resonance - background coupling .
we treat the nuclear dynamics of dissociative electron attachment within the local complex potential model .
this model is concerned with the proper accounting for the decay of the resonant state , and its effect upon the nuclear dynamics .
the lcp model includes the simplest such accounting , in which the decay rate is a local function of the nuclear geometry .
the local complex potential model @xcite , also known as the `` boomerang '' model when applied to vibrational excitation , describes resonance nuclear motion by an inhomogeneous schrdinger equation and a complex , but purely local potential .
it is perhaps easiest to derive by applying feshbach partitioning @xcite within the born - oppenheimer framework to derive a nuclear wave equation @xcite .
the derivation begins by defining a discrete ( square - integrable ) approximation to the resonant electronic state , @xmath38 which depends parametrically on the nuclear coordinates @xmath6 and which is unit - normalized with respect to integration over the electronic coordinates @xmath39 .
one then defines the geometry - dependent feshbach projection operator @xmath40 , which operates on the electronic degrees of freedom , @xmath41 \big [ \psi_q(\vec{q } ) \big\vert , \ ] ] and its complement @xmath42 : @xmath43 with @xmath44 , @xmath45 and @xmath46 .
( brackets denote integration over the electronic degrees of freedom only . ) partitioning the full wave function for total energy @xmath47 as @xmath48 , we can formally derive the following inhomogenous equation for @xmath49 : @xmath50 where @xmath51 is the sum of the electronic hamiltonian and nuclear kinetic energy , @xmath52 . in view of eq .
( [ q_def ] ) , we can write @xmath53 the function @xmath54 describes the relative motion of the nuclei in the negative - ion resonance state . to derive an equation for @xmath54 , the first approximation that is made
is the born - oppenheimer approximation : we neglect all non - adiabatic couplings arising from the operation of the nuclear kinetic energy upon the adiabatic basis . then multiplying eq .
( [ transcendental2 ] ) from the left by @xmath38 and integrating oven the electronic coordinates gives the nuclear wave equation , @xmath55 where @xmath56\ ] ] and @xmath57 the real - valued potential @xmath58 is the expectation value of the electronic hamiltonian with respect to the discrete state @xmath59 ; the additional , energy - dependent term @xmath60 is called the `` level - shift operator '' and is nonlocal in the nuclear degrees of freedom @xmath6 , owing to the presence of the nuclear green s function .
the residue of this green s function gives the level - shift operator @xmath60 a negative - definite imaginary component . in order to bring eq.([transcendental3 ] ) into the form of the local complex potential model , it is necessary to make a local approximation to the level - shift operator @xmath60 , and also to approximate the driving term .
the assumptions that underlie these approximations are well understood @xcite . a local approximation to the level - shift operator yields @xmath61 where @xmath8 and @xmath9 are the location and _ total _ width of the resonance . a first - order perturbation treatment ( fermi s golden rule ) of the driving term yields @xcite @xmath62 where @xmath63 is the _ partial _ width for decay to the ground electronic state of the target , and
@xmath64 is the initial rovibrational state of the target .
the final working equation of the lcp model then reads @xmath65 the location and widths of the various resonance states were obtained from configuration interaction and fixed - nuclei variational electron scattering calculations , respectively , as detailed in paper i. in the case of the @xmath1 resonance , which generally lies below its @xmath66 neutral parent , the resonance can only decay into the ground electronic channel . in that case , the total and partial widths , @xmath9 and @xmath63 , coincide and can be obtained by fitting the eigenphase sum to a breit - wigner form . for the higher resonances ,
a more elaborate fitting procedure is required to obtain the partial widths , as outlined in ref .
@xcite and in paper i. a direct solution of the differential equations of the local complex potential model can pose significant difficulties for problems with multiple degrees of freedom and , in such cases , a time - dependent formulation of the problem can offer distinct computational advantages . such a formulation can be made , as demonstrated by mccurdy and turner @xcite , by formally writing the solution of eq .
( [ eq : boomerang ] ) as @xmath67 and writing the nuclear green s function as the fourier transform of the propagator for the time - dependent schrdinger equation : @xmath68 where we define the time - dependent nuclear wave function as @xmath69 the driving term @xmath70 of the lcp equation can thus be viewed as the initial value of a wave packet that subsequently evolves on the complex potential surface of the resonance anion .
since the potential surface is complex , the packet decays as a function of time until it effectively escapes the region of the surface where the width is nonzero .
c + the lcp model equations were solved in the coordinate systems depicted in fig .
[ bigcoords ] .
for the three internal degrees of freedom of this triatomic molecule , we employ jacobi coordinate systems , which are depicted at the top of this figure .
the jacobi coordinate system on the left , marked `` ( a ) , '' is used to analyze the oh+h arrangement ; the one on the right , marked `` ( b ) , '' is used for the h@xmath0+o arrangement .
the vector @xmath71 connects the nuclei of the diatomic .
the vector @xmath72 connects the center of mass of the diatomic to the third atom .
@xmath35 is the length of @xmath72 , @xmath73 is the length of @xmath71 , and @xmath74 is the angle between the @xmath72 and @xmath71 vectors . for ( a ) , @xmath75 denotes a linear ohh configuration .
in addition to the three internal degrees of freedom there are also the three euler angles that orient the internal or body - fixed ( bf ) frame with respect to the lab or space - fixed ( sf ) frame .
the origin of both frames is the center of mass .
the space - fixed @xmath76-axis is always chosen to be parallel with the wavevector of the incident electron . for calculations with total rotational angular momentum @xmath77 ,
the hamiltonian only operates on the internal degrees of freedom . for @xmath78
we must take the euler angles into account , and we denote them by @xmath79 . the total ( electronic plus nuclear , ignoring spin ) angular momentum , @xmath80 , and its projection upon the space - fixed @xmath76 axis , @xmath81 , are quantum numbers conserved by the hamiltonian .
we also use the quantum number @xmath82 to specify the projection of the angular momentum on a bf axis .
@xmath82 is not a conserved quantity and there is some flexibility in its definition .
we use the `` @xmath35-embedding '' scheme @xcite in which @xmath72 is taken to be collinear with the bf @xmath83 axis and the angular momentum number @xmath82 is quantized around this axis . with this convention , the euler angles @xmath84 and @xmath85 are the polar angles which orient the @xmath35 vector with respect to the sf frame , and @xmath86 is the third euler angle specifying orientation about the bf @xmath83 axis .
a schematic of the coordinate system is also shown in fig . [ bigcoords ] .
we may write a general expression for the six - dimensional rovibrational wave function for a triatomic with specified @xmath80 and @xmath81 value as follows : @xmath87 where the basis of @xmath88 is the set of normalized wigner rotation matrices ( and bf angular momentum eigenstates ) @xmath89 such that @xmath90 in eqs.([expansion ] ) and ( [ dnormalized ] ) we follow the conventions of zhang @xcite , which for the @xmath91 is the same as that of edmonds @xcite . the standard @xcite bf hamiltonian for the radial solutions @xmath92 of this expansion incorporates coupling among the different @xmath82 values for a given total angular momentum @xmath80 .
the neglect of this coupling is termed the `` coupled states '' or `` centrifugal sudden '' ( cs ) approximation@xcite , and we employ this approximation for our calculations , since the kinetic energy of the recoiling fragments is large compared to their centrifugal energy .
the resulting hamiltonian is thus diagonal in @xmath82 and can be written @xmath93 where @xmath94 and @xmath95 are the reduced masses in either degree of freedom and @xmath96 is the ( coupled set of ) born - oppenheimer potential - energy surface(s ) that we calculate . for dynamics beginning on the @xmath2 ( 1 @xmath19 ) resonance surface , the gradient of that surface will force the wave packet toward linear geometry , at which point this resonance state is degenerate with the @xmath1 resonance ( see fig .
[ c2vfig ] ) .
the renner - teller effect @xcite will therefore couple these two components of the @xmath11 state , and we modify the hamiltonian of eq.([hamiltonian ] ) accordingly .
the quantum numbers @xmath97 and @xmath82 in eq.([hamiltonian ] ) are obtained as eigenvalues of the total angular momentum operators @xmath98 and @xmath99 : @xmath100 etc . , where @xmath99 has a simple form in terms of derivative operators in ( @xmath84 , @xmath85 , @xmath86 ) @xcite .
properly , the operators that appear in the born - oppenheimer hamiltonian for the rovibrational motion of the nuclei , eq.([hamiltonian ] ) , should be not @xmath98 and @xmath99 but the _ nuclear _ angular momentum operators @xmath101 and @xmath102 , where @xmath103 in which expression @xmath104 is an electronic angular momentum operator ; the hamiltonian @xcite with this form is exact except for the omission of the mass - polarization term .
the exact hamiltonian @xcite introduces numerous new diagonal and off - diagonal ( off - diagonal in @xmath82 , electronic state , and both ) coupling terms to the triatomic hamiltonian .
the term that is most commonly labeled the renner - teller coupling comes from the @xmath105 term in eq.([hamiltonian ] ) : @xmath106 it is the @xmath107 term which couples the two components ( sine and cosine , @xmath1 and 1 @xmath19 ) of the @xmath11 state at linear geometry . at such geometries
the operator @xmath108 is diagonalized by @xmath109 the matrix elements of @xmath110 may either be computed @xcite , or approximated by their values at linear geometry @xcite .
we take the latter route , i.e. , we assume that eq.([lsubz ] ) holds everywhere .
this approximation has little effect on the dynamics because only near linear geometry does the coupling become large .
we perform our renner - teller calculations in the ( @xmath111 ) diabatic basis because it allows us to incorporate the boundary condition in @xmath74 using the `` _ _ k__-legendre '' discrete variable representation @xcite . with this assumption , for a given value of @xmath82 , the ( @xmath111 ) diabatic states have @xmath112 . the kinetic energy operator in eq.([angcentrifugal ] )
is diagonal in this diabatic basis .
the coupling then arises from the electronic hamiltonian , which is not diagonal in this basis .
the electronic hamiltonian in this basis takes the form @xmath113 i.e. , the diabatic states are degenerate .
when @xmath114 , there is no renner - teller effect , since the coupling term in eq.[angcentrifugal ] vanishes .
the multiconfiguration time - dependent hartree or mctdh @xcite method is an efficient adaptive scheme for propagating quantum - mechanical wave packets for systems with multiple degrees of freedom .
we use this method to perform the propagation in eq .
( [ timedepwfcn ] ) .
we use the implementation within the mctdh package@xcite , a freely available suite of codes built at the university of heidelberg , germany . in the mctdh method , as in other methods developed for solving the time - dependent schrdinger equation , we start with a time - independent orthonormal product basis set , @xmath115 for a problem with @xmath116 degrees of freedom and nuclear coordinates labeled @xmath117 . for computational efficiency , the basis functions @xmath118 are chosen as the basis functions of a discrete variable representation ( dvr ) @xcite . the central idea of the mctdh technique is the representation of the nuclear wave packet as a sum of separable terms , @xmath119 with @xmath120 .
each `` single particle function '' ( or spf ) @xmath121 is itself represented in terms of the primitive basis , @xmath122 one can determine equations of motion for the parameters @xmath123 and @xmath124 . since both the coefficients @xmath125 and the single - particle functions @xmath126 are time - dependent , the wave function representation is made unique by imposing additional constraints on the single - particle functions which keep them orthonormal for all times @xcite . the evaluation of the hamiltonian matrix , which must be carried out at every time step , may be expedited @xcite if the hamiltonian can be written as a sum of products of single - coordinate operators .
the mctdh package @xcite includes a utility which performs a fit of a given potential to a separable representation of this form .
details can be found in beck _
et al._@xcite .
all potential - energy surfaces used in the current calculation were represented in this manner , using this utility to fit them specifically for each choice of the dvr grids . for calculations on the electronically coupled @xmath2 and @xmath3 states ,
the underlying dvr is the same for each electronic state , but each electronic state has its own set of single - particle functions @xmath126 .
this is referred to as the `` multi - set '' formalism , as opposed to `` single - set . ''
the renner - teller coupled 1 @xmath19 - @xmath1 calculation is performed under the single - set formalism .
the sine dvr bases in the @xmath73 and @xmath35 degrees of freedom incorporate standing wave boundary conditions at their edges .
therefore , when the dissociating wave packet reaches the end of the dvr grid , it must be absorbed to prevent unphysical backward reflections . to this end
we include an artificial negative imaginary component to the surface called a `` complex absorbing potential '' or cap@xcite : @xmath127 a similar expression for the cap in the @xmath73 degree of freedom also applies .
formally , the cap s provide the @xmath128 limit in eq.([fourier ] ) .
we use a value for @xmath129 equal to 0.007 hartree , and place @xmath130 three bohr before the end of our grid , except for the 1 @xmath19 calculations for @xmath131 and @xmath132 , for which we use a strength of 0.0018 hartree and a value of @xmath130 five bohr before the end of the grid .
the cross sections for dissociative attachment can be calculated directly from the time - propagated wave packet by computing the energy - resolved , outgoing projected flux .
the energy resolution is achieved by fourier transform and a final state resolution is achieved by the introduction of appropriate projection operators . for dea to a specific final rovibrational state labeled by rotational ( @xmath133 ) and vibrational ( @xmath134 ) indices , we use the projection operator @xmath135 the flux operator , which measures the flux passing through a surface defined by @xmath136 , is defined as @xmath137,\ ] ] where @xmath138 is a heaviside function .
the energy - resolved projected flux is then given by @xmath139 the mctdh package @xcite includes a utility which computes the outgoing projected flux . in the actual calculations , the flux operator appearing in the equation above is replaced by an expression involving the complex absorbing potential , eq.([eq : cap ] ) .
this formulation of the flux operator is very convenient numerically and entirely equivalent to the traditional formal definition of the operator in this context , in the limit that the cap does not perturb the propagating wave packet beyond first order , which in the present case , given the nuclear masses , holds as a good approximation . for more details on this cap flux formalism
see refs .
@xcite .
the resulting energy - resolved projected flux is that associated with the time - independent solution of the driven schrdinger equation of the lcp model in eq .
( [ eq : boomerang ] ) , @xmath140 in terms of @xmath141 , the dea cross section is @xcite @xmath142 for the h@xmath4+oh channel , an additional factor of two is included in eq .
( [ cross ] ) to account for the fact that in a given calculation we perform the flux analysis for only one of the two @xmath21+oh arrangements , namely the one for which the jacobi coordinates are appropriate .
the definition of the rovibrational states @xmath143 is complicated by the ion - dipole interaction of the fragments . in our earlier study dea to water via the @xmath1 feshbach resonance@xcite , we attempted a complete final state analysis , and projected upon pendular ( restricted rotor ) states @xcite , not free rotational states , and assumed that these pendular states evolve adiabatically to their free rotational state asymptotes .
this analysis did not yield any major insight , and so for the present calculations we simply project upon free rotational states . as a consequence
, there is a small error in our final state resolution , but the magnitude of this error will span a range of states and range of energies approximately equal to the magnitude of the ion - dipole interaction at the edge of our grid , which is small compared to the kinetic energy spread of the fragments . in most of the calculations reported here , we used dvr primitive basis sets for all internal degrees of freedom , choosing the standard sine dvr@xcite for the @xmath73 and @xmath35 degrees of freedom and , for @xmath77 , the legendre dvr@xcite for @xmath74 . for @xmath144 ,
as previously discussed , the dvr for @xmath74 must be modified to account for singularities in the hamiltonian [ see eq.([hamiltonian ] ) ] due to the term @xmath145 .
this is done by using an extended legendre dvr @xcite , which is implemented in the heidelberg mctdh package @xcite .
the initial rovibrational states @xmath64 of eq .
( [ drivdef ] ) were obtained via relaxation and improved relaxation @xcite as implemented within the mctdh package @xcite . in relaxation runs ,
an initial guess @xmath146 for the ground state is propagated in imaginary time , which yields the ground state @xmath147 : @xmath148 in improved relaxation runs , the propagation of the spf expansion coefficients @xmath149 of eq.([mctdh3 ] ) is performed via eq.([relax ] ) , but the configuration coefficients @xmath124 are obtained anew at each time - step via a davidson diagonalization .
for the two - state @xmath3-@xmath2 calculations , the wavefunction is represented in the diabatic basis , each component of which has an expansion of the form of eq.([mctdh2 ] ) , and different sets of time - dependent single - particle functions @xmath121 .
since the adiabatic - to - diabatic transformation angle is not constant with geometry , the single - particle functions which represent the initial state will be different in the diabatic basis than in the adiabatic basis ( in which they would be identical to the single - particle functions of the relaxation run ) . for this reason , an iterative technique @xcite is employed to minimize the error between the diabatic representation and its adiabatic representation .
[ cols="^,^ " , ] the high degree of vibrational excitation ( @xmath150 ) in both the oh ( @xmath151 ) and oh ( @xmath152 ) channels is apparent in the oscillations of the dissociating wave packet within each potential well , visible in the lower panels of figs .
[ b2a1j1prop2 ] and [ b2a1j1prop1 ] .
[ b2a1j1prop1 ] shows that in this case there is additional structure to the dissociating wave packet on the lower 1 @xmath19 surface ; however , this structure is most likely due to the calculation not being fully converged .
the channel h@xmath0+o@xmath4 is the dominant channel observed in experiment for dissociative attachment to water via the highest - energy @xmath3 resonance . as discussed at length in ref .
@xcite and paper i , this channel is not present as an asymptote on the @xmath3 ( 2 @xmath19 ) surface , and therefore , the system must undergo a nonadiabatic transition via the conical intersection to the lower 1 @xmath19 surface in order to reach this product channel . in the context of the representation which we constructed in paper i , the system must follow the _
@xmath3 surface past its crossing with the @xmath2 diabatic surface .
the h@xmath0+o@xmath4 channel is present as an asymptote of the diabatic @xmath3 surface . as described in ref .
@xcite and paper i , the adiabatic @xmath2 surface does not have a bound asymptote in this arrangement ; it correlates to o@xmath4+h@xmath0 ( @xmath20 ) instead .
the calculated peak cross section for this channel , 1.87@xmath2410@xmath153@xmath154 , is smaller than the experimental value , 5.7@xmath2410@xmath153@xmath154 , reported by melton @xcite .
the comparison with experiment is again complicated by the fact that we calculate only the two - body dea cross section , while the available experimental data do not discriminate between production of o@xmath4+h+h and o@xmath4+h@xmath0 .
a possible explanation for the discrepancy between our calculated results and experiment is the presence of a large three - body component to o@xmath4 production via the @xmath3 resonance .
rotational excitation of the target h@xmath0o molecule , on the other hand , can not account for this discrepancy .
we have performed several calculations in which rotational excitation of the target is included .
these include calculations for total angular momentum @xmath155 , projection @xmath114 .
we find that the effect of such rotational excitation is minimal , as fig .
[ h2_o_b2a1.excite ] shows .
we calculate a very high degree of rotational and vibrational excitation in the h@xmath0 or d@xmath0 fragment .
the average degree of vibrational excitation @xmath150 calculated from eq.([psurv ] is 7.75 for the h@xmath0 fragment and 13.0 for the d@xmath0 fragment .
the corresponding values for @xmath156 are 405 and 725 , respectively .
figure [ h2_o_b2a1 ] shows the total cross sections , as well as the cross sections into either rotational or vibrational states , summed over the opposite quantum number .
the degree of vibrational excitation evidently decreases with incident electron energy , while the degree of rotational excitation shows little correlation with incident electron energy .
the high degree of rotational and vibrational excitation of the diatomic fragment reduces the kinetic energy of the atom - diatom recoil .
this is reflected in the cross sections for production of both h@xmath0 and d@xmath0 via the @xmath3 resonance , which have the greatest magnitude nearere the lower range of recoil energy .
two - dimensional plots of these cross sections as functions of both incident electron energy and the kinetic energy of the recoil are shown in the epaps archive@xcite .
plots of the propagated wave packet for dea via the @xmath3 resonance are shown in figs .
[ b2a1j2prop2 ] and [ b2a1j2prop1 ] .
the first of these shows the magnitude - squared of the 1 @xmath19 component to the propagated wave packet , integrated over @xmath74 , and the latter shows the 2 @xmath19 component .
the corresponding potential - energy surfaces , evaluated at @xmath74=90@xmath18 , are also plotted , along with the location of the conical intersection seam which appears as a bold line .
the wave packet begins on the upper surface and proceeds to the lower surface only via nonadiabatic coupling near the conical intersection .
as described in paper i , the gradient of the upper 2 @xmath19 resonance surface leads downhill toward its conical intersection with the 1 @xmath19 resonance , leading the propagated wave packet toward the seam .
this behavior is clearly visible in fig .
[ b2a1j2prop2 ] and [ b2a1j2prop1 ] .
the 2 @xmath19 wave packet follows the conical intersection seam in fig .
[ b2a1j2prop2 ] , until it is consumed by the large imaginary component to that potential - energy surface and by nonadiabatic coupling to the 1 @xmath19 state along the intersection .
the wave packet appears on the 1 @xmath19 surface in fig.[b2a1j2prop1 ] along the conical intersection , and a small portion of it is able to reach the h@xmath0+o@xmath4 well of that state .
the magnitude of the cross section for production of h@xmath0+o@xmath4 from the @xmath3 resonance is therefore controlled by several competing effects .
the shape of the real part of the potential - energy surface determines the dynamically accessible pathways and favors localization of the 2 @xmath19 wave packet near the conical intersection . at the same time , the large imaginary component to this surface consumes the wave packet and decreases the amount of flux available to enter the conical intersection . on the lower 1 @xmath19 surface , the amount of flux that enters the h@xmath0 potential well is determined by the shape of that potential - energy surface , since the conical intersection is outside the potential well and only a fraction of the wave packet is propagated into the well .
we have presented the results of a fully _ ab initio _ study of dissociative electron attachment to h@xmath0o that includes the full dimensionality of nuclear motion .
we have attempted to calculate the cross sections for all the major and minor two - body channels which are present as asymptotes of the born - oppenheimer , @xmath1 , @xmath2 , and @xmath3 adiabatic electronic feshbach resonances .
while we have qualitatively described the principal features that have been experimentally observed , it is clear that a fully quantitative description of this process has yet to be achieved .
the nuclear dynamics calculations were carried out using the mctdh method within the framework of the local complex potential model .
for the major channel dea , h@xmath4+oh ( x @xmath11 ) production through the lowest @xmath1 resonance , the underlying assumptions of the model are well satisfied and we have obtained reasonably good agreement with the experimental observations .
another notable feature of the present study is the quantification of the mechanism in the major channel that leads to production of o@xmath4 through the @xmath3 resonance . our earlier speculation
@xcite that a conical intersection between the @xmath2 and @xmath3 states would play the key role in this process has been confirmed by the present study .
the present treatment has been limited to a consideration of dea only into the final - state two - body channels .
this limitation undoubtedly explains our inability to produce a non - zero cross section for o@xmath4 production via the @xmath2 resonance , which is likely to be dominated by three - body breakup .
three - body breakup may also play a role in o@xmath4 production via the @xmath3 resonance , and its neglect here could explain why our calculated cross sections are smaller than the experimental results , which did not differentiate two- and three - body channels .
physics beyond the local complex potential model may be at work in some of the minor channels .
dissociative electron attachment via the @xmath2 feshbach resonance may involve an even greater variety of complicated resonant as well as non - resonant phenomena not described by the lcp model .
a variety of effects that go beyond the lcp model could be at play in the production of h@xmath4 via @xmath2 feshbach resonance , including coupling to a broader shape resonance and even non - resonant virtual state effects .
the neglect of such effects could well explain our overestimation of the cross section for production of h@xmath4+oh via the @xmath2 state . even for dea via the lowest - energy @xmath1 state , nonlocal physics may be important in the minor channel , which leads to h@xmath0+o@xmath4 .
we have achieved considerable success in describing the mean features of dea to water , clarified the mechanisms for the two - body breakup channels , and found evidence to suggest that three - body breakup to produce o@xmath4 might be important .
nonetheless , many challenges remain before a complete and quantitative understanding of this fundamental , but complicated , system will be realized .
this work was performed under the auspices of the us department of energy by the university of california lawrence berkeley national laboratory under contract de - ac02 - 05ch11231 and was supported by the u.s .
doe office of basic energy sciences , division of chemical sciences .
the authors acknowledge many helpful discussions with h .- dieter meyer concerning various aspects of the mctdh method . | we report the results of a first - principles study of dissociative electron attachment to h@xmath0o .
the cross sections are obtained from nuclear dynamics calculations carried out in full dimensionality within the local complex potential model by using the multi - configuration time - dependent hartree method .
the calculations employ our previously obtained global , complex - valued , potential - energy surfaces for the three ( @xmath1 , @xmath2 , and @xmath3 ) electronic feshbach resonances involved in this process .
these three metastable states of h@xmath0o@xmath4 undergo several degeneracies , and we incorporate both the renner - teller coupling between the @xmath1 and @xmath2 states as well as the conical intersection between the @xmath2 and @xmath3 states into our treatment .
the nuclear dynamics are inherently multidimensional and involve branching between different final product arrangements as well as extensive excitation of the diatomic fragment .
our results successfully mirror the qualitative features of the major fragment channels observed , but are less successful in reproducing the available results for some of the minor channels .
we comment on the applicability of the local complex potential model to such a complicated resonant system . |
with the accumulation of observational data from cosmic microwave background measurements ( cmb ) , large scale structure surveys ( lss ) and supernovae observations and the improvements of the data quality , the cosmological observations play a crucial role in our understanding of the universe , and also in constraining the cosmological parameters , such as the eos of dark energy models , the curvature of universe , the total neutrino masses and those associated with the running of the spectral index and gravitational waves . in our previous work @xcite , we have used the markov chain monte carlo ( mcmc ) method to constrain cosmological models from the astronomical observational data , including the wmap three - year data ( wmap3 ) @xcite , small - scale cmb data , lss data @xcite and snia essence sample @xcite .
we found that the cosmological constant is consistent with the data , however the dynamical dark energy models are still allowed and interestingly the model with its eos getting across @xmath23 the _ quintom _ model @xcite is the best fit model .
and we found no strong significant evidence for the non - flat universe and massive neutrino . within the @xmath0cdm model , the scale - invariant spectrum and the spectra with @xmath24
are disfavored by more than @xmath18 confidence level . due to the degeneracy between the eos of dark energy and tensor fluctuation
, those inflationary models excluded within the @xmath0cdm model , will revive in the framework of dynamical dark energy models .
furthermore , we did not find any significant evidence for the tensor fluctuations and the large running of the spectral index .
given the precision of current observations , these results are not conclusive for the time being .
recently , the wmap experiment has published its five - year data of temperature and polarization power spectra @xcite .
the arcminute cosmology bolometer array ( acbar ) experiment has also published its new cmb temperature power spectrum @xcite .
these new cmb data can strengthen the constraints on the cosmological parameters , especially for the inflationary models @xcite .
furthermore , the supernova cosmology project has made an unified analysis of the world s supernovae datasets and presented a new compilation union " ( 307 sample ) @xcite which includes the recent samples of snia from snls and essence survey , as well as some older datasets , _
etc_. in the literature @xcite , this union compilation combining with the shift parameter derived from cmb and the bao information has been used to constrain cosmological models . however
, in these studies the cmb information considered is just the shift parameter instead of the full cmb data , which will lose some information of cmb and lead to a biased result @xcite .
thus , in this paper , we revisit the issue on the determination of these cosmological parameters and update our previous results with the latest observational data .
our paper is organized as follows : in section ii we describe the method and the latest observational datasets we used ; section iii contains our main global fitting results on the cosmological parameters and the last section is the summary .
in our study , we perform a global analysis using the publicly available mcmc package cosmomc @xcite . we assume the purely adiabatic initial conditions .
our most general parameter space is : @xmath25 where @xmath26 and @xmath27 , in which @xmath28 and @xmath29 are the physical baryon and cold dark matter densities relative to the critical density , @xmath6 is the spatial curvature and satisfies @xmath30 , @xmath31 is the ratio ( multiplied by 100 ) of the sound horizon to the angular diameter distance at decoupling , @xmath32 is the optical depth to re - ionization , @xmath33 is the dark matter neutrino fraction at present , namely , @xmath34 the primordial scalar power spectrum @xmath35 is parameterized as @xcite : @xmath36 where @xmath37 is defined as the amplitude of initial power spectrum , @xmath19 measures the spectral index , @xmath38 is the running of the scalar spectral index and @xmath20 is the tensor to scalar ratio of the primordial spectrum . for the pivot scale
we set @xmath39mpc@xmath40 .
moreover , @xmath41 and @xmath42 are the parameters of dark energy eos , which is given by @xcite : @xmath43 where @xmath44 is the scale factor and @xmath45 characterizes the running " of eos ( runw henceforth ) .
the @xmath0cdm model has @xmath46 and @xmath47 . for the dark energy model with a constant eos , @xmath47 ( wcdm henceforth ) .
when using the global fitting strategy to constrain the cosmological parameters , it is crucial to include dark energy perturbations @xcite . in this paper
we use the method provided in refs.@xcite to treat the dark energy perturbations consistently in the whole parameter space in the numerical calculations . in the computation of cmb
we have included the wmap5 temperature and polarization power spectra with the routine for computing the likelihood supplied by the wmap team .
we also include some small - scale cmb measurements , such as boomerang @xcite , cbi @xcite , vsa @xcite and the newly released acbar data @xcite .
besides the cmb information , we also combine the distance measurements from bao and snia . for the bao information , we use the gaussian priors on the distance ratios , @xmath48 at @xmath49 and @xmath50 at @xmath51 , with a correlation coefficient of @xmath52 , measured from the bao in the distribution of galaxies @xcite . in the calculation of the likelihood from snia we have marginalized over the nuisance parameter @xcite .
the supernova data we use are the recently released union " compilation ( 307 sample ) @xcite . in order to improve the constraint on the total neutrino mass , we include the lyman-@xmath12 forest power spectrum from sloan digital sky survey ( sdss ) @xcite
, however also keep its unclear systematics in mind @xcite .
furthermore , we make use of the hubble space telescope ( hst ) measurement of the hubble parameter @xmath53h km s@xmath40 mpc@xmath40 by a gaussian likelihood function centered around @xmath54 and with a standard deviation @xmath55 @xcite .
in this section we present our global fitting results of the cosmological parameters determined from the latest observational data and focus on the dark energy parameters , curvature of universe , neutrino mass and the inflationary parameters , respectively . , and the present matter density , @xmath56 , from the latest observations , assuming a flat universe .
the blue dash - dot lines and red solid lines are obtained with and without the systematic uncertainties of union compilation , respectively.[fig1 ] ] and @xmath42 from the current observations , cmb+bao+sn .
the red solid lines and the blue dash - dot lines are obtained for the flat and non - flat universe , respectively . and
the black dashed lines are obtained when ( incorrectly ) neglecting dark energy perturbations .
the magenta solid lines stand for @xmath46 and @xmath57 . in this numerical calculation
the systematic uncertainties of union compilation is not considered .
[ fig2 ] ] table i. constraints on the dark energy eos and some background parameters from the latest observations . here
we have shown the mean and the best fit values , which are obtained from the cases with and without the systematic uncertainties of union compilation , respectively .
& & & & + & with sys & w / o sys & with sys & w / o sys & with sys & w / o sys & with sys & w / o sys + @xmath0cdm & bestfit & @xmath58 & @xmath58 & @xmath59 & @xmath59 & @xmath60 & @xmath61 & @xmath62 & @xmath63 + @xmath64 & mean & @xmath58 & @xmath58 & @xmath59 & @xmath59 & @xmath65 & @xmath66 & @xmath67 & @xmath68 + wcdm & bestfit & @xmath69 & @xmath70 & @xmath59 & @xmath59 & @xmath71 & @xmath60 & @xmath72 & @xmath73 + @xmath64 & mean & @xmath74 & @xmath75 & @xmath59 & @xmath59 & @xmath76 & @xmath77 & @xmath78 & @xmath79 + runw & bestfit & @xmath80 & @xmath81 & @xmath82 & @xmath83 & @xmath60 & @xmath71 & @xmath84 & @xmath85 + @xmath64 & mean & @xmath86 & @xmath87 & @xmath88 & @xmath89 & @xmath90 & @xmath77 & @xmath91 & @xmath92 + runw & bestfit & @xmath93 & @xmath94 & @xmath93 & @xmath95 & @xmath93 & @xmath96 & @xmath93 & @xmath97 + @xmath98 & mean & @xmath93 & @xmath99 & @xmath93 & @xmath100 & @xmath93 & @xmath101 & @xmath93 & @xmath102 + runw & bestfit & @xmath93 & @xmath103 & @xmath93 & @xmath104 & @xmath93 & @xmath105 & @xmath93 & @xmath106 + w / o pert .
& mean & @xmath93 & @xmath107 & @xmath93 & @xmath108 & @xmath93 & @xmath109 & @xmath93 & @xmath110 + in table i we list the constraints on the dark energy parameters as well as the hubble constant in different dark energy models . assuming the flat universe , firstly we explore the constraints on the constant eos of dark energy , @xmath111 ( @xmath112 , @xmath113 ) , from the latest observational data . in fig.[fig1
] we show the constraints on @xmath111 and the present dark matter density , @xmath56 .
this result shows that the combination of union compilation ( with systematic uncertainties not included ) and other observational data yield a strong constraint on the constant eos of dark energy , @xmath1 ( @xmath2 ) .
our result is similar to the limit from wmap5 @xcite , @xmath114 ( @xmath2 ) and physically it indicates that @xmath115 is consistent with the data .
furthermore , some of the quintessence models get strongly constrained , for example , the tracker quintessence model which predicts @xmath116 @xcite will be excluded by more than @xmath117 .
however we notice that the systematic uncertainties of union compilation will affect strongly the error estimation of dark energy parameters @xcite . if taking the systematic uncertainties into account , we find the limit on the constant eos of dark energy is @xmath118 ( @xmath2 ) and the error bar is significantly enlarged . for the time evolving eos , @xmath119 , in fig.[fig2 ] we illustrate the constraints on the dark energy parameters @xmath41 and @xmath42 . for the flat universe , from the latest observational data we find that the best fit model is the quintom dark energy model , @xmath4 and @xmath5 , whose @xmath120 can cross the cosmological constant boundary during the evolution .
but then the variance of @xmath41 and @xmath42 are still large , namely , the @xmath8 constraints on @xmath41 and @xmath42 are @xmath121 and @xmath122 .
this result implies that the dynamical dark energy models are not excluded and the current data can not distinguish different dark energy models decisively .
the @xmath0cdm model , however is still a good fit right now .
our results are consistent with the wmap5 group @xcite , while the upper limit on @xmath41 and lower limit on @xmath42 are slightly weaker than theirs .
this difference is mainly from the supernovae datasets used .
the supernovae dataset we use in this paper is the new union compilation with homogeneous analysis of the present world s supernovae data .
but in ref.@xcite they use three supernovae datasets , snls @xcite , hst @xcite and essence @xcite . for each of them
, they marginalize over the absolute magnitude separately , and simply add these three pieces to get the total @xmath123 . because the parametrization of eos of the dark energy used in this paper is assumed to extend to an arbitrary high redshift , it is important to check if the energy density of dark energy component is negligible compared with the radiation density at the epoch of the big bang nucleosynthesis ( bbn ) , @xmath124 . as shown by the red solid lines in fig.[fig2 ] ,
the dynamical dark energy models allowed by the current data within the @xmath8 confidence level safely satisfy the limits of @xmath125 and bbn @xcite to avoid the dark energy domination in the early universe .
furthermore , the null energy condition ( nec ) should also be satisfied for the eos of universe @xmath126 @xcite : @xmath127 where @xmath128 and @xmath129 are the eos and energy density for component @xmath130 in the universe .
violation of nec will lead to the breakdown of causality in general relativity and the violation of the second law of thermodynamics @xcite .
this requirement from @xmath131 will constrain the eos parameters of dark energy models @xcite .
firstly , we consider the wcdm dark energy model . from table
i , the current observational data give the present energy density of dark energy @xmath132 .
if we just assume that the eos of dark energy is constant from the early time of universe to present , using the above eq.([nec ] ) , we obtain that the nec limit requires the eos of dark energy @xmath133 straightforwardly @xcite .
fortunately , the current constraints on @xmath134 with and without the systematic uncertainties of union " compilation satisfy this limit safely . however , if we extend the period of validity of constant eos of dark energy into the future where the dark energy component will dominate the universe entirely ( namely @xmath135 ) , the nec requires @xmath136 .
consequently , the phantom dark energy models ( @xmath137 ) will be excluded and the fate of universe with big rip @xcite could not happen . and @xmath42 from the current observations ( black solid lines ) and the null energy condition ( blue dashed lines ) . in the upper panel , the parametrization of eos in eq.([eos ] ) is assumed to be a description of dark energy for the past until now . in the lower panel ,
we assume that the parametrized eos are also valid for the far future .
[ fignec],title="fig : " ] and @xmath42 from the current observations ( black solid lines ) and the null energy condition ( blue dashed lines ) . in the upper panel , the parametrization of eos in eq.([eos ] )
is assumed to be a description of dark energy for the past until now . in the lower panel ,
we assume that the parametrized eos are also valid for the far future .
[ fignec],title="fig : " ] for the runw dark energy model , we illustrate the constraints on the dark energy parameters @xmath41 and @xmath42 from the current observations ( black solid lines ) and the nec ( blue dashed lines ) in fig.[fignec ] , respectively . in the upper panel ,
the parameterized eos @xmath138 is assumed to be valid until now , not for the future , _
i.e. _ @xmath139 .
the reasons for doing this are i ) experimentally there are no constraints on the dark energy models for the future universe ; ii ) theoretically it is always possible to choose a different type of parametrization of dark energy eos for the future universe , such as @xmath140 for @xmath141 , which matches to @xmath138 at @xmath142 and satisfies the nec @xcite .
in fact , if the current eos of dark energy @xmath143 , this scenario allows the transition from @xmath137 to @xmath144 , consequently avoids the violation of nec .
one can see from the fig.[fignec ] , in this case , the nec limit does improve the constraints on dark energy parameters from the current observations .
for example , the current observational data permit @xmath145 at @xmath18 confidence level , which however violates the nec limit . in the lower panel of fig.[fignec ] ,
the parametrization @xmath138 is assumed to be a valid description of dark energy model at any time from the far past to the far future .
one can see from this figure , the nec puts a stronger constraint on the eos parameters of dark energy models than the observational data . for this case , the regions of quintessence and phantom dark energy models will be shrunk significantly in the ( @xmath41,@xmath42 ) space , and the nec excludes the dark energy models corresponding to the regions labelled by the phantom and quintom a with @xmath137 for @xmath146 .
consequently , the models satisfying the nec include the quintessence and some of the quintom b dark energy models , which can be seen in the lower panel of fig.[fignec ] .
finally we discuss the degeneracy between the dark energy and curvature . as we know
, the eos of dark energy is degenerated with the curvature of universe @xmath6 @xcite . if we do not include the prior that the universe is flat , from table i and the blue dash - dot lines in fig.[fig2 ]
, we can see that the constraints on @xmath41 and @xmath42 are weakened significantly and the two - dimensional distribution extends more towards the quintom b region .
but the main conclusions are unchanged , @xmath147 and @xmath148 , namely , the quintom dark energy model is still mildly favored by the current observational data . moreover , in order to show the importance of dark energy perturbations in the global analysis , we do the calculation by incorrectly neglecting the dark energy perturbations . illustrated as the black dashed lines in fig.[fig2 ] , one can see that the constraints on the dark energy parameters become tighter immediately , similar to our previous results @xcite .
this study shows that how biased the result will be , once the dark energy perturbations are incorrectly neglected in the analysis @xcite .
constraints on cosmological parameters @xmath19 , @xmath149 , @xmath20 , @xmath6 and @xmath150 from the current observations .
we have shown the mean @xmath151 errors .
for the weakly constrained parameters we quote the @xmath8 upper limits instead . [ cols="^,^,^",options="header " , ] the measurements on the position of first acoustic peak of cmb temperature power spectrum
have been used to determine the curvature of universe @xmath6 . however , due to the well - known degeneracy between @xmath56 and @xmath6 , we have to add other cosmological data , such as the large scale structure and supernovae data , to break this degeneracy and improve the constraint . within the @xmath0cdm model , from table ii and fig.[fig3 ]
, we can see that our universe is very close to flatness , namely , the @xmath8 limit is @xmath7 , which is consistent with the prediction of inflation paradigm .
as we mentioned before , dark energy parameters and @xmath6 are correlated via the cosmological distance information . in the framework of dynamical dark energy models , the constraint on @xmath6 should be relaxed .
based on the calculations , we can see that the combination of observational data implies @xmath10 at @xmath18 confidence level .
current constraints on the inflationary parameters @xmath19 , @xmath149 and @xmath20 , as well as the curvature @xmath6 and the total neutrino mass @xmath150 , in different dark energy models : @xmath0cdm model ( blue dash - dot lines ) and runw model ( red solid lines ) . for @xmath150
, we also show the limits combined with the sdss lyman-@xmath12 forest power spectrum .
blue dotted lines and red dashed lines denote the @xmath0cdm and runw models , respectively.[fig3 ] ] detecting the neutrino mass is one of the challenges of modern physics .
currently the neutrino oscillation experiments , such as atmospheric neutrinos experiments @xcite and solar neutrinos experiments @xcite , have confirmed that the neutrinos are massive , but give no hint on their absolute mass scale . fortunately , cosmological observational data can provide the crucial complementary information on absolute neutrino masses , because massive neutrinos leave imprints on the cosmological observations , such as the hubble diagram , cmb temperature power spectrum and lss matter power spectrum @xcite .
within the @xmath0cdm model , from table ii one can read @xmath8 upper limit of the total neutrino mass derived from the current observations , cmb+lss+sn , @xmath152 ( @xmath8 c.l . ) , which is consistent with the recent results from wmap5 group @xcite . however , there are degeneracies between the neutrino mass and other cosmological parameters , such as the eos parameters of dark energy @xcite and the running of spectral index @xcite . due to the degeneracy among dark energy parameters and the neutrino mass @xcite , in the framework of dynamical dark energy models ,
the limit on the neutrino mass can be relaxed to @xmath153 ( @xmath8 c.l . ) significantly , as shown in fig.[fig3 ] .
it is well known that when neutrinos become non - relativistic at late time of the universe , they damp the perturbations within their free streaming scale . thus the massive neutrinos will suppress the matter power spectrum at small scale by roughly @xmath154 @xcite . therefore , lyman-@xmath12 forest data at small scale can significantly improve the constraint on the neutrino mass .
but when we use the ly@xmath12 data , we should keep in mind on their unclear systematics right now @xcite . when including the sdss ly@xmath12 forest power spectrum @xcite
, we can obtain a much more stringent @xmath18 upper limit @xmath155 in the @xmath0cdm model .
moreover , the heidelberg - moscow ( hm ) experiment , which is controversial for the time being , uses the half - life of @xmath156 decay to constrain the effective majorana mass and this translates to the constraint on the sum of neutrino masses under some assumptions @xcite , @xmath157 ( @xmath8 c.l . ) .
we can find an obvious tension on the neutrino mass limits from between the cosmological observations and the hm experiment , which however can be resolved if the neutrino masses vary during the evolution of the universe @xcite . in order to be consistent with the observational data
, the neutrino mass must be very small in the past , but has grown recently in order to agree with the hm experiment data .
again , in the runw model , the limit on the neutrino mass is relaxed , @xmath158 ( @xmath8 c.l . ) , which implies the existence of degeneracy between the dark energy parameters and neutrino mass .
inflation , the most attractive paradigm in the very early universe , has successfully resolved many problems existing in hot big bang cosmology , such as flatness , horizon , monopole problem and so forth @xcite .
its quantum fluctuations turn out to be the primordial density fluctuations which seed the observed large scale structures and the anisotropies of cmb .
inflation theory has successfully passed several non - trivial tests
. currently , the cosmological observational data are in good agreement with a gaussian , adiabatic and scale - invariant primordial spectrum , which is consistent with single field slow - roll inflation predictions . within the @xmath0cdm model , from fig.[fig3 ]
, we obtain the limit on the spectral index of @xmath159 ( @xmath2 ) , which excludes the scale - invariant spectrum , @xmath160 , and the spectra with blue tilt , @xmath24 , at more than @xmath161 confidence level .
when considering the gravitational waves , the latest observational data yield the @xmath8 upper limit of tensor - to - scalar ratio @xmath16 . in fig.[fig4 ] we show the two dimensional constraints in ( @xmath19,@xmath20 ) panel which can be compared with the prediction of the inflation models .
we find that the harrison - zeldovich - peebles scale - invariant ( hz ) spectrum ( @xmath160 , @xmath162 ) is still disfavored more than @xmath18 confidence level . and
many hybrid inflation models and the inflation models with blue " tilt ( @xmath24 ) are also excluded by the current observations .
furthermore , the single slow - rolling scalar field with potential @xmath163 , which predicts @xmath164 , is still well within @xmath18 region , while another single slow - rolling scalar field with potential @xmath165 , which predicts @xmath166 , has been excluded more than @xmath18 @xcite .
however , the tensor fluctuations and the dark energy component , through the isw effect , are correlated , which mostly affect the large scale ( low multipoles ) temperature power spectrum of cmb @xcite . in the framework of dynamical dark energy model
, we find that the upper limit of @xmath20 can be relaxed to @xmath167 ( @xmath8 c.l . ) .
furthermore , the @xmath8 confidence level contour in ( @xmath19,@xmath20 ) panel will be enlarged consequently and the distribution extends towards the hybrid inflation region .
therefore , we can see that the hz spectrum is consistent with the latest observational data and many hybrid inflationary models , the inflationary models with
blue " tilt ( @xmath24 ) , which are excluded in the @xmath0cdm model , have revived in the framework of dynamical dark energy model as illustrated in fig.[fig4 ] . and
@xmath8 constraints on the panel ( @xmath19,@xmath20 ) based on the different dark energy models : @xmath0cdm model ( blue dash - dot lines ) and runw model ( red solid line ) .
the two magenta solid lines delimit the three classes of inflationary models , namely , small - field , large - field and hybrid models .
the star points are predicted by hz spectrum , @xmath168 model and @xmath169 model , respectively .
these predictions assume that the number of e - foldings , @xmath170 , is @xmath171 for @xmath168 model and @xmath172 for @xmath169 model.[fig4 ] ] finally , we explore the constraint on the running of the spectral index from the latest observational data . when combining the wmap1 or wmap3 data with other astronomical data , the previous analysis have found a significant evidence for large running @xcite .
physically this large running would be a great challenge to the single - field inflation models @xcite .
the combination of the current observational data yield the limit on the running of the spectral index of @xmath21 ( @xmath2 ) for the @xmath0cdm model and @xmath22 ( @xmath2 ) for the runw model , respectively .
the error is dramatically reduced compared with the previous results @xcite , beneficial from the more accurate observational data .
given the current data , we find no significant evidence for the large running of the spectral index .
recently many experimental groups have published their new observational data , such as temperature and polarization power spectra of wmap5 @xcite , temperature power spectrum of acbar @xcite and supernovae dataset of union compilation @xcite . in this paper
we report the updated constraints on the cosmological parameters from these latest observational data , such as the eos of dark energy , curvature of universe , neutrino mass and inflation parameters . for dark energy ,
we explore the constraints on two kinds of dynamical models .
assuming a flat universe , within the wcdm model , we find that the latest observational data yield the limit on the constant eos of dark energy , @xmath1 ( @xmath2 ) . for runw model with a flat universe , we find that the best fit model is @xmath4 and @xmath5 described by the quintom theory with eos across the cosmological constant boundary . and because the precision of current observations are not good enough to determine the dark energy eos conclusively , the dynamical dark energy models are not excluded and the @xmath0cdm model remains a good fit . the quintom scenario , with the particular feature that its eos can cross the cosmological constant boundary smoothly , has been applied to many aspects of cosmology theoretically .
firstly , the quintom dark energy models can also give rise to interesting prediction on the fates of the universe , different from the quintessence or phantom models , such as the cyclic universe @xcite , an expanding universe with oscillating eos .
secondly , applying a quintom matter for the early universe can provide a scenario of bouncing cosmology , which can avoid the notorious issue of initial singularity @xcite .
our results also show that the universe is very close to flatness and the upper limit on the total neutrino mass is @xmath173 ev ( @xmath8 c.l . ) , from the combination of cmb , bao and sn data . given the efficiency of ly@xmath12 forest data on constraining the total neutrino mass
, we also perform a calculation with the inclusion of the sdss lyman-@xmath12 forest power spectrum and find that @xmath14 ev ( @xmath8 c.l . ) .
this result might lead to the exclusion of degenerate pattern of neutrino mass , when combining the results of neutrino oscillation experiments .
due to the degeneracy between the neutrino mass and eos of dark energy , however in the presence of dynamics of dark energy , the constraints on @xmath150 can be relaxed by a factor of @xmath15 . finally , for the inflationary models , within the @xmath0cdm framework
, we find that the latest observational data prefer the inflation models with red " tilt , namely , @xmath159 ( @xmath2 ) and small tensor fluctuations , @xmath16 ( @xmath8 c.l . ) . because of the degeneracy between @xmath20 and eos of dark energy , the upper limit on @xmath20 is relaxed to @xmath167 ( @xmath8 c.l . ) and the parameter space in ( @xmath19,@xmath20 ) panel are enlarged in the framework of dynamical dark energy models .
therefore , the inflationary model with hz primordial spectrum ( @xmath160 , @xmath162 ) , some hybrid models and some models with a blue " tilt ( @xmath24 ) , which are excluded more than @xmath18 confidence level in the @xmath0cdm model , will be consistent with the current observations now .
furthermore , in our analysis we do not find any significant evidence for the running of spectrum index .
we acknowledge the use of the legacy archive for microwave background data analysis ( lambda ) . support for lambda is provided by the nasa office of space science .
we have performed our numerical analysis on the shanghai supercomputer center ( ssc ) .
we thank yi - fu cai , jie liu and tao - tao qiu for helpful discussions .
this work is supported in part by national science foundation of china under grant no .
10533010 and 10675136 , and the 973 program no.2007cb815401 , and by the chinese academy of science under grant no .
kjcx3-syw - n2 .
hl is supported by the china postdoctoral science foundation .
gz is supported by national science and engineering research council of canada ( nserc ) .
a. kosowsky and m. s. turner , phys .
d 52 , 1739 ( 1995 ) ; j. e. lidsey , a. r. liddle , e. w. kolb , e. j. copeland , t. barreiro and m. abney , rev .
69 , 373 ( 1997 ) ; s. hannestad , s. h. hansen , f. l. villante and a. j. s. hamilton , astropart .
17 , 375 ( 2002 ) ; s. l. bridle , a. m. lewis , j. weller and g. efstathiou , mon . not .
r. astron .
342 , l72 ( 2003 ) ; b. feng , x. gong and x. wang , mod .
lett . a 19 , 2377 ( 2004 ) . k. s. hirata , _ et al .
_ , phys .
b 280 , 146 ( 1992 ) ; y. fukuda , _ et al .
_ , phys .
b 335 , 237 ( 1994 ) ; y. fukuda , _ et al .
_ , phys .
81 , 1562 ( 1998 ) ; w. w. m. allison , _ et al .
_ , phys .
b 449 , 137 ( 1999 ) ; m. ambrosio , _ et al .
_ , phys .
b 517 , 59 ( 2001 ) .
b. t. cleveland , _ et al .
_ , astrophys . j. 496 , 505 ( 1998 ) ; w. hampel , _ et al .
_ , phys .
b 447 , 127 ( 1999 ) ; j. n. abdurashitov , _ et al .
_ , phys .
c 60 , 055801 ( 1999 ) ; s. fukuda , _ et al .
_ , phys .
86 , 5651 ( 2001 ) ; s. fukuda , _ et al .
_ , phys .
86 , 5656 ( 2001 ) ; q. r. ahmad , _ et al .
_ , phys .
89 , 011301 ( 2002 ) ; s. n. ahmed , _ et al . _ ,
92 , 181301 ( 2004 ) .
y. f. cai , t. qiu , y. s. piao , m. li and x. zhang , jhep 0710 , 071 ( 2007 ) ; y. f. cai , t. qiu , r. brandenberger , y. s. piao and x. zhang , jcap 0803 , 013 ( 2008 ) ; y. f. cai , t. qiu , j. q. xia , h. li and x. zhang , arxiv:0808.0819 . | in this paper , we combine the latest observational data , including the wmap five - year data ( wmap5 ) , boomerang , cbi , vsa , acbar , as well as the baryon acoustic oscillations ( bao ) and type ia supernoave ( sn ) union " compilation ( 307 sample ) , and use the markov chain monte carlo method to determine the cosmological parameters , such as the equation - of - state ( eos ) of dark energy , the curvature of universe , the total neutrino mass and the parameters associated with the power spectrum of primordial fluctuations .
our results show that the @xmath0cdm model remains a good fit to the current data . in a flat universe
, we obtain the tight limit on the constant eos of dark energy as , @xmath1 ( @xmath2 ) . for the dynamical dark energy models with time evolving eos parameterized as @xmath3
, we find that the best - fit values are @xmath4 and @xmath5 , implying the preference of quintom model whose eos gets across the cosmological constant boundary during evolution .
for the curvature of universe @xmath6 , our results give @xmath7 ( @xmath8 c.l . )
when fixing @xmath9 .
when considering the dynamics of dark energy , the flat universe is still a good fit to the current data , @xmath10 ( @xmath8 c.l . ) .
regarding the neutrino mass limit , we obtain the upper limits , @xmath11 ev ( @xmath8 c.l . ) within the framework of the flat @xmath0cdm model .
when adding the sdss lyman-@xmath12 forest power spectrum data , the constraint on @xmath13 can be significantly improved , @xmath14 ev ( @xmath8 c.l . ) . however , these limits can be weakened by a factor of @xmath15 in the framework of dynamical dark energy models , due to the obvious degeneracy between neutrino mass and the eos of dark energy model .
assuming that the primordial fluctuations are adiabatic with a power law spectrum , within the @xmath0cdm model , we find that the upper limit on the ratio of the tensor to scalar is @xmath16 ( @xmath8 c.l . ) and the inflationary models with the slope @xmath17 are excluded at more than @xmath18 confidence level .
however , in the framework of dynamical dark energy models , the allowed region in the parameter space of ( @xmath19,@xmath20 ) is enlarged significantly .
finally , we find no strong evidence for the large running of the spectral index , @xmath21 ( @xmath2 ) for the @xmath0cdm model and @xmath22 ( @xmath2 ) for the dynamical dark energy model , respectively . |
it has been clear for decades that metallicity is the most influential factor governing the hb morphologies of galactic gcs : metallicity is the first parameter .
the earliest photographic color - magnitude diagrams ( cmds ) of gcs by , e.g.@xcite , revealed that hb stars in metal - rich gcs tend to lie on the red side of the rr lyrae instability strip while hb stars in metal - poor gcs lie primarily on the blue side .
as more and more cmds were assembled , exceptions to this rule were uncovered .
@xcite noted that m 13 and m 22 display hb morphologies appropriate for metal - poor gcs , despite the fact both gcs were of intermediate metallicity , and suggested that a difference in age might be responsible .
@xcite noticed that the hb of ngc 7006 is redder than either m 13 or m 3 , despite the fact that all three gcs appeared to have very similar metallicities .
such anomalies suggested the need for a second parameter that could account for differences in hb morphology not obviously caused by metallicity .
@xcite summarized the problem , stating that metallicity is not sufficient to explain the extent of observed hb morphologies but that differences in age or he enrichment could explain the observed variations .
both of these suggestions remain valid up to the present time .
@xcite analyzed the integrated colors of 49 gcs with @xmath2 photometry and concluded that ` at least two parameters ( one of which is metal abundance ) are required to describe globular clusters . '
the second parameter phenomenon took on a greater significance with the seminal work of @xcite .
searle & zinn recognized that gcs with unusually red hbs are relatively rare in the inner regions of the galactic halo ( galactocentric radius , @xmath3 kpc ) but become increasingly common at greater @xmath4 ( see their figure 10 ) .
searle & zinn used this fact to argue that the inner halo was assembled early and in a fairly short time while the outer halo was assembled over an extended period of time .
current galaxy formation scenarios envision the outer regions of the galaxy as the accumulated debris of the many accretion events that shaped the early evolution of the galaxy . in that context , understanding the origin and existence of age and metallicity gradients in the galactic gc population is as relevant now as it was at the time when searle & zinn first introduced their halo formation scenario .
subsequent efforts relating to the second parameter problem fall into two general categories : those that attempt to measure the age difference between two ( or a few ) gcs with similar metallicities but markedly different hb morphologies and those that investigate the ensemble properties of a large sample of gcs .
one canonical pair in the former category is ngc 288 , with a blue hb , and ngc 362 , with a red hb .
the first ccd - based , differential photometric study of these clusters was conducted by @xcite .
bolte concluded that ngc 288 is @xmath13 gyr older than ngc 362 .
the subsequent work of @xcite , @xcite , and @xcite echoed this result .
by contrast , @xcite reached a very different conclusion . using the brightest rgb star in each gc
, @xcite corrected the photometry of ngc 288 and ngc 362 for their relative distances and found their main sequence turnoffs to be roughly coincident , thereby suggesting a negligible age difference .
@xcite found ngc 288 to be 2@xmath51 gyr older than ngc 362 using three different techniques ; in a follow - up , @xcite found that an age difference of 2 gyr was plausible if both gcs have @xmath6}}\sim -1.2 $ ] but that their synthetic hb models were unable to match the detailed hb morphology of either gc using canonical assumptions of average mass loss and dispersion on the hb . in considering differences between second parameter pairs , it is important to understand just how similar the abundances are in both clusters .
@xcite performed a detailed comparison of abundances in red giants with 13 stars in ngc 288 and 12 in ngc 362 .
these authors concluded that ngc 288 has a lower @xmath6}}$ ] than ngc 362 by 0.06 dex , that the average @xmath7}}$ ] ratios are very similar , and that both gcs exhibit variations among o , na , and al .
the other canonical second parameter pair is m 13 , with a blue hb , and m 3 , with an intermediate hb .
m 13 and m 3 are @xmath10.2 dex more metal poor than ngc 288 and ngc 362 .
@xcite found evidence for an age difference but the quality of their cmds did not allow a more definitive statement .
@xcite estimated an upper limit to the age difference of @xmath13 gyr and concluded that it was insufficient to explain the hb morphologies , assuming both gcs have the same chemical composition .
alternatively , @xcite suggested the difference in their hb morphologies was due to a difference in their he abundances .
johnson & bolte suggested that m 13 had a he mass fraction @xmath10.05 greater than m 3 .
however , @xcite showed that a difference of @xmath8y @xmath9 at constant @xmath6}}$ ] would make the level of the hb brighter by @xmath10.2 mag . such a difference between m 3 and
m 13 was ruled out by the photometry of @xcite , who also concluded that m 13 is older than m 3 by @xmath10 gyr .
@xcite compared spectra of 28 red giants in m 3 with 35 in m 13 and found the two gcs mean @xmath6}}$ ] values the same within their 1-@xmath11 error bars .
however , @xcite reported differences in light element abundance variations , particularly o , for which m 13 displays a larger range of variation than m 3 by @xmath10.5 dex .
a discussion of the search for the second parameter in gc - to - gc comparisons would not be complete without including the work of @xcite and @xcite .
these investigations used hst wfpc2 photometry to measure the ages of the outer halo gcs palomar 3 , palomar 4 , and eridanus @xcite and am-1 and palomar 14 @xcite relative to inner halo gcs with similar metallicities , m3 and m5 .
each of the five outer halo gcs has a redder hb morphology than its comparison inner halo gc .
both studies concluded that the outer halo gcs are @xmath11.5 - 2 gyr younger than m3 and m5 , provided the chemical compositions of the outer halo gcs are comparable to their inner halo counterparts . a recent study by @xcite found the abundances of pal 3 are essentially the same as inner halo gcs of similar metallicity and thus the relative age comparison with m 3 by @xcite was justified .
overall , the outer halo gcs chemical compositions remain poorly understood by comparison with the inner halo , see e.g.@xcite .
@xcite found the reported age differences between pal 4 and eridanus , with red hbs , and m 5 , with an intermediate hb , too small to explain the difference in hb morphologies unless all three gcs are younger than 10 gyr , assuming standard assumptions of mass loss on the rgb .
@xcite concluded that it was possible to explain the difference in hb morphology between pal 3 and m 3 if the former has less hb mass dispersion and is younger than the latter .
these examples indicate that if age is the second parameter , then it is our lack of understanding of mass loss that confuses gc - to - gc comparisons as first pointed out by @xcite .
the review by @xcite includes a thorough discussion of different mass loss prescriptions and their efficacy . given the hb morphologies and relatively young ages of the most distant outer halo gcs , it is important to include them when considering the properties of the entire galactic gc population , especially in the context of searle & zinn s halo formation scenario .
the second category of second parameter studies are those that consider the properties of a large sample of galactic gcs , e.g.@xcite . in the first ccd - era study ,
@xcite compiled properties for 31 galactic gcs . among other things
, @xcite examined the variation of hb morphology with age for gcs in a narrow range of metal abundance and showed that gcs with red hbs are significantly younger than those with blue hbs by as much as 5 gyr .
the subsequent studies of @xcite , @xcite , and @xcite updated and reaffirmed this result .
@xcite found that the gcs with @xmath12 kpc are , on average , younger than those with @xmath13 kpc , which is consistent with age as the second parameter .
by contrast , @xcite examined the cmds of 36 gcs , found an age dispersion of @xmath11 gyr with no significant age gradient in the galactic halo , and concluded that this age range was too small for metallicity and age alone to explain hb morphology in gcs .
theoretical efforts have provided further insights into the complexities of hb morphology . in particular ,
several studies have applied the synthetic hb model developed by @xcite .
@xcite , @xcite , and @xcite used synthetic hb models to explore the interplay of hb morphology , metallicity , and age in the cmd .
these studies demonstrated considerable degeneracy in the hb morphology
metallicity diagram and , in particular , @xcite argued that unless absolute values of the he abundance , @xmath7}}$ ] ratios , and rgb mass loss were known , hb morphology did not constitute a reliable age constraint .
nevertheless , @xcite concluded that there was evidence for an age dispersion of @xmath14 gyr among galactic gcs of similar metallicity but markedly different hb morphologies .
a number of theoretical studies have also focused on second parameter pairs or triads but , unfortunately , the lack of a firm theoretical understanding of mass loss along the rgb has impaired the these efforts , as first pointed out by @xcite .
see @xcite for a recent review . as it pertains to hb morphology
, the use of the term second parameter has expanded to cover a broad range of complex behaviors . in particular , this includes the faint extent of the blue hb tail in some gcs .
@xcite analyzed 53 gcs and found that the length of the hb in the cmd and the extent of the blue tail is correlated with central density . @xcite also reported a link between an extended blue tail and central density in gcs and concluded that ` environment is `` a '' second parameter ' .
@xcite showed that there is a correlation between central density and hb morphology for intermediate metallicity gcs ( @xmath15 } } < -1.3 $ ] ) where the second parameter effect is most pronounced .
@xcite analyzed the properties of 54 galactic gcs with homogeneous photometry from an hst wfpc2 snapshot survey @xcite and concluded that more massive gcs tend to have more extended blue hbs .
@xcite recognized a link between the effective temperature of the hottest hb star in a gc and its mass ( as inferred from its integrated luminosity ) and suggested that self - pollution could explain the existence of faint blue tails in preferentially massive gcs .
evidence abounds for the presence of chemical abundance variations in all gcs @xcite and the possible correlation between the degree of abundance variations and the faint extent of the blue hb @xcite . however , it is unlikely that the existence of faint , blue tails in the hbs of some gcs is directly related to the appearance of anomalously red hbs in others .
it is also unlikely that chemical abundance variations will unduly affect age estimates of most gcs , provided the total metal content is constant across all stars @xcite . for cases in which the total metal abundance is not constant , or there are distinct stellar populations present in the cmd ,
age estimates are necessarily more complicated and careful analysis of each of these gcs is needed .
see @xcite for a recent summary of gcs with multiple stellar populations visible in the cmd , several of which were discovered with photometry from the acs survey of galactic gcs .
the complex issue of multiple stellar populations in gcs remains poorly understood and the extent to which multiple - population gcs permeate the galactic gc population is unknown at present .
to summarize , the presence of metal poor gcs with red hbs that primarily reside in the outer galactic halo is well - known observationally .
although the present study is focused on the galactic gcs , including some of those associated with the sagittarius dwarf , there is ample evidence to suggest that metal poor gcs with red hbs are also found in the magellanic clouds @xcite and fornax @xcite .
@xcite summarize our current knowledge of the gc populations in the galaxy and its satellites in the hb morphology metallicity diagram . despite an abundance of observational evidence , no consensus has been reached as to what parameter(s ) are responsible for the appearance of relatively red hbs in metal poor gcs .
age is frequently offered as a candidate but considerable doubt still remains because the age difference claimed or required by theoretical studies of hb morphology is too large to satisfy the observations .
the existence of a homogeneous database of deep , high quality photometry from the acs survey of galactic gcs ( * ? ? ?
* paper i in this series ) has motivated a re - examination of hb morphology and its relation to a variety of gc properties .
the paper is organized as follows : @xmath16[data ] describes the data sample ; @xmath16[hbm ] describes the methods that were employed to determine the hb morphologies ; @xmath16[gcages ] describes the sources of gc ages and provides some discussion of complicating factors in age determination ; @xmath16[results ] presents the analysis performed on the assembled data and discusses the important results ; and , finally , @xmath16[conclusion ] provides a summary of the salient points .
the photometric catalog of the acs survey of galactic gcs ( * ? ? ? * paper v ) consists of 65 gcs observed in @xmath17 and @xmath18 .
most details concerning the data reduction and calibration are provided in paper v but , since its publication , the data have been adjusted to account for updated hst / acs wide field channel zero - points @xcite ; the new @xmath17 and @xmath18 zeropoints are , respectively , 22 and 35 mmag fainter than given by @xcite .
not all gcs in the acs survey catalog will be considered in the analysis performed in the following sections .
@xmath19 cen ( ngc 5139 ) and m 54 ( ngc 6715 ) have been excluded because their complex cmds and abundances indicate that these objects are not gcs in the traditional sense .
@xcite analyzed the m 54 cmd and measured ages for the stellar populations present and also estimated the metallicities of the stellar populations for which no spectroscopic information was available . because m 54 lies at the center of the sagittarius dwarf galaxy , it is not a simple process to distinguish between the stellar population(s ) that belong to m 54 and those of the sagittarius dwarf galaxy .
@xmath19 cen has long been known to contain a range of different metallicities as well as multiple stellar populations in the cmd @xcite .
@xcite presented a spectroscopic analysis of 180 red giants in @xmath19 cen that revealed at least four distinct metallicities and a total range of @xmath20 dex in @xmath6}}$ ] .
other gcs for which multiple stellar population evidence exists will be noted in the following sections .
@xcite provided a summary of the gcs known to exhibit multiple stellar populations in their cmds circa 2008 .
three more gcs were excluded : pal 1 and e 3 for lack of identifiable hb stars and pal 2 due to extreme differential reddening ( see paper i ) . in total , 60 gcs from the acs survey are considered in the following analysis .
the acs survey provides coverage of the galactic gc population out to @xmath120 kpc . in order to give proper consideration to the outer galactic halo ,
where the second parameter phenomenon is most pronounced , the sample was extended to include the following gcs : ngc 2419 @xcite ; pal 3 , pal 4 , and eridanus @xcite ; and am-1 and pal 14 @xcite .
these are the six most distant gcs in the galaxy with @xmath21 kpc @xcite .
unfortunately , no photometry of comparable quality is currently available for the 10 or so gcs that lie between these two groups ( @xmath22 kpc ) kpc : pal 15 , ngc 7006 , and pyxis . ] . in order to distinguish between the homogeneous acs survey sample and the larger , heterogeneous sample , the statistical analysis in section @xmath16[results
] will be presented with and without the six outer halo gcs .
in addition to hb morphologies ( @xmath16[hbm ] ) and ages ( @xmath16[gcages ] ) , several gc parameters were extracted from @xcite .
the @xcite catalog is based on the @xcite catalog but includes several updated quantities and , importantly , provides quantities in absolute physical units .
these units are more appropriate for comparisons than the observational units given in the @xcite catalog .
the quantities obtained from the @xcite catalog are : galactocentric distance ( @xmath4 ) , integrated absolute @xmath23 magnitude ( @xmath24 ) , half - light radius ( @xmath25 ) , and tidal radius ( @xmath26 ) . from the @xcite catalog we extracted central luminosity density ( @xmath27 ) .
neither catalog includes basic measurements for lyng 7 .
inclusion of these quantities in the following analysis makes it possible to explore the possibility that a gc s location in the galaxy , structural properties , and total mass ( assuming the mass - to - light ratio is constant ) can influence its hb morphology .
in this study , hb morphology is characterized by the difference between the median color of the hb stars and the median color of the rgb at the level of the hb , denoted @xmath0 .
this metric is essentially the same as @xmath28 @xcite and is less likely to become saturated than the most frequently used hb morphology parameter , ( b@xmath29r)/(b+v+r ) introduced by @xcite . as a corollary
, the magnitude level of the hb in @xmath17 has also been measured ; it provides a reference for gc distance estimates as discussed below .
the mean level of the horizontal part of the hb was measured from the acs survey data relative to the hb of m 5 ( ngc 5904 ) .
m 5 has a well - populated hb that extends approximately 0.7 mag in @xmath30 from the red side to the blue and , therefore , can overlay the hb of almost any other gc .
measurements were made by matching the cmd of each gc to that of m 5 in the region of the hb by making vertical and horizontal adjustments to the comparison cmd until it most closely overlaid that of m 5 . in the majority of cases ,
the comparison gc had enough overlap to make the best alignment unambiguous .
a few clusters with purely blue hbs , such as ngc 6254 , present some difficulty because even the reddest hb stars do not become ` horizontal ' .
the measured hb levels are listed in table [ morph ] .
typical measurement uncertainties are @xmath10.05 mag and are dominated by the uncertainty inherent in aligning the hbs .
the formal photometric errors of most hb stars are a few thousandths of a magnitude .
the uncertainty is larger in cases where differential reddening is significant or the hb is otherwise difficult to quantify . as an example , figure [ outline ] shows hb levels for ngc 1261 and m 92 ( ngc 6341 ) .
measurements were obtained by drawing an outline around the hb and the rgb at the level of the hb and then calculating the statistics of the stars within each outline .
this process is illustrated in figure [ outline ] for two gcs , ngc 1261 and m 92 .
rr lyrae stars are not given special consideration in this analysis : they are included if they are present in the acs observations .
for the acs survey data , all determinations of hb membership were performed in the acs @xmath30 cmd as shown in figure [ outline ] . in order to allow direct comparison with gcs observed in different bands ,
the color difference presented in table [ morph ] is given in terms of @xmath31 rather than @xmath30 .
this was achieved by transforming the acs magnitudes of each individual hb star into @xmath23 and @xmath32 using the equations provided by @xcite and then calculating the median @xmath31 values of the hb and rgb .
the median was chosen to represent the central tendency of @xmath0 because the hb stars are not normally distributed in color space .
hence , the more commonly - used mean and standard deviation are not the most appropriate choices .
if the hb stars were normally distributed in color space , then the mean and median would be equal .
in fact , about one half ( one quarter ) of the gcs in the sample have means and medians that differ by more than 5% ( 10% ) .
for gcs with intermediate hbs , such as m 3 and ngc 1261 , the mean and median are substantially different .
table [ morph ] provides the quantities used to determine @xmath0 .
in addition to the median @xmath31 of the hb and rgb , 1-@xmath11 errors on these quantities are listed and the quadrature sum is given as the uncertainty on @xmath0 .
uncertainty in the median @xmath31 was estimated by bootstrapping with replacements performed 10,000 times on both the hb and rgb of each gc .
the error bars on the median represent the range of @xmath31 within which 68% of bootstrapped medians lie . to compliment the median and uncertainty , we also give the mean absolute deviation ( mad ) for the hb and rgb .
the mean absolute deviation represents the intrinsic spread in each individual data set .
consider , for example , the hbs of 47 tuc ( ngc 104 ) and m 3 ( ngc 5272 ) .
both have well - populated hbs and thus small uncertainties , @xmath33 determined from bootstrapping , but 47 tuc has a tightly clustered red clump of hb stars and a small mad=0.023 while m 3 has a more broadly distributed hb and a larger mad=0.208 .
perhaps the most frequently used hb morphology metric in the literature is ( b@xmath29r)/(b+v+r ) : the difference between the number of blue hb stars ( b ) and the number of red ( r ) normalized by the total number of hb stars including variables ( v ) .
its major inconvenience is that it saturates if all stars lie on one side or the other of the instability strip .
the left panel of figure [ dvibvr ] compares ( b@xmath29r)/(b+v+r ) from @xcite and @xmath0 from this paper for the 66 gcs in the present sample .
there is an obvious correlation between the two but @xmath0 continues to vary after ( b@xmath29r)/(b+v+r ) saturates . the most recent large - scale study of hb morphology , that of @xcite , used the maximum effective temperature encountered along the hb [ log @xmath34 determined with the use of theoretical zero - age hb sequences from @xcite .
the log @xmath35 metric appears to be complimentary to @xmath0 or ( b@xmath29r)/(b+v+r ) because , while it lacks sensitivity through the middle , it is more sensitive at the extremes .
this is demonstrated in the right panel of figure [ dvibvr ] . in the absence of an ideal hb morphology metric that quantifies every feature of interest , it is beneficial to identify a metric that is well suited to a particular problem .
* paper vii ) presented relative ages measured from the acs survey data .
these ages were derived by first measuring the absolute magnitudes of the main sequence turnoffs ( mstos ) for all gcs and then interpolating in isochrone - based grids of msto as a function of age and metallicity .
the same analysis was performed with four different isochrone libraries and the distribution of relative ages with metallicity was shown to be independent of the isochrone library .
for the purpose of the present paper , relative ages derived from ( * ? ? ?
* paper ii ) isochrones using the ( * ? ? ?
* hereafter zw84 ) metallicity scale have been adopted .
furthermore , we assumed that gcs with @xmath36 } } < -1 $ ] have [ @xmath37/fe]=+0.3 and gcs with @xmath36}}\geq -1 $ ] have [ @xmath37/fe]=+0.2 .
the ages measured in this fashion and with these assumptions were placed on a relative scale by dividing out the average age of the most metal - poor gcs ( [ fe / h ] @xmath38 ) , in this case 13.3 gyr . in the present study ,
this factor has been retained , resulting in absolute ages for the particular isochrone library , metallicity scale , and method of age determination employed in paper vii . in order to put the acs survey ages on the same scale as the six outer halo gcs , and to demonstrate that later results do not depend on the method used , age estimates were determined using isochrone fitting to the cmds of all gcs in the data set .
the fits were performed using isochrones from paper ii for the acs survey data and @xcite for the outer halo gcs .
the same zeropoint corrections applied to the acs data , as described in @xmath16[survey ] , were applied to the isochrones in the acs photometric system as well .
the underlying luminosities , temperatures , surface gravities , and color transformations of the isochrones in paper ii and @xcite are identical .
the only difference between the two is the photometric system .
the outer halo gcs were measured in @xmath39 and @xmath18 , whereas the acs survey data were taken in @xmath17 and @xmath18 .
several gcs were excluded from isochrone fitting because they are known to harbor multiple stellar populations .
the presence of more than one stellar population makes a single age determination insufficient to characterize the gc .
ages were not determined for the following gcs : ngc 1851 ( * ? ? ?
* paper iii ) , ngc 2808 @xcite , ngc 6388 , ngc 6441 , and ngc 6656 @xcite .
it is likely that other gcs with multiple populations exist , even within the acs survey data , but to have avoided detection thus far the separation between the populations must either be so small that they overlap to a great extent or that reddening has obscured the separation in the cmd .
a second population may also represent a small fraction of stars and have little or no apparent influence on the hb morphology .
the recent discovery of a second subgiant branch in 47 tuc ( ngc 104 ) by @xcite is an example : this second population accounts for only @xmath110% of the stars in the core .
further out , the cmd was too sparsely populated to detect this population .
it should also be noted that @xcite found the main population of 47 tuc to have a spread in the cmd greater than can be accounted for by the photometric errors .
the authors found it unlikely that this spread was due to binaries , differential reddening , or a depth effect .
@xcite explored the possibility that the color spread could be caused by an intrinsic variation in the he abundance or @xmath6}}$ ] .
another case is m 4 ( ngc 6121 ) .
@xcite presented spectroscopic evidence for a bimodal abundance distribution along the rgb of m 4 as well as a broadening of the rgb in @xmath40 . whether m 4 contains two distinct stellar populations or a continuous distribution is not clear at present .
if it is the former , the difference is small enough to remain undetected on the main sequence in the exquisite photometry presented by @xcite .
if it is the latter , age determinations are expected to be reasonable ( see the discussion at the end of @xmath16[metage ] ) .
the procedure used to determine ages involved a two step process .
initial estimates for @xmath6}}$ ] , distance modulus , and reddening estimates were taken from the @xcite catalog , 2003 revision .
estimates of [ @xmath37/fe ] were chosen as appropriate for a given value of @xmath6}}$ ] .
all efforts were made to follow these values as closely as possible with minor adjustments made to improve the fit to the unevolved main sequence first and the rgb second . in some instances , isochrones at the catalog values were not able to provide an adequate fit to the cmds and , in these cases , the initial estimates of @xmath6}}$ ] , distance , and reddening were allowed to vary until the fit obtained was acceptable .
this step was a necessity in a number of cases because the published values for some lesser - known gcs are uncertain .
for example , @xcite found that the main sequence and rgb morphologies of am-1 are comparable to those of m 3 although the former is listed at @xmath6}}=-1.8 $ ] while the latter is listed at @xmath6}}=-1.57 $ ] in the harris catalog .
another example is the distance modulus of ngc 6254 : the harris catalog value is @xmath41 but the present study produces a value that is @xmath10.2 mag fainter .
this result is confirmed by both isochrone fitting to the main sequence and the level of the hb from table [ morph ] .
these results are not meant to diminish the value of the gc catalogs , which are invaluable resources , but merely to emphasize that there are still many gcs in the galaxy whose properties are poorly constrained .
the models , which are homogeneous in terms of physical assumptions and ingredients , can prove useful in determining the relative differences between two gcs if one has well - measured properties and another does not .
several of the gcs in the acs survey are heavily obscured by interstellar reddening .
differential reddening confuses the age determinations in such gcs because the reddening line is nearly perpendicular to the most age - sensitive features : the main sequence turnoff and the subgiant branch .
differential reddening causes these features to be spread out and , in light of this trend , we have attempted to correct for its effects in the following manner .
the first step involves the construction of a fiducial sequence for the gc .
the fiducial sequence is centered approximately at the msto .
more stars lie below the msto than above but the differential corrections are weighted based on the angle between the stellar sequence and the reddening line . since the reddening line is almost perpendicular to the subgiant branch , the differential correction is approximately equally weighted by stars above and below the msto .
care is taken to avoid unresolved binaries . from the fiducial sequence ,
each individual star yields a color residual , taken along the reddening direction in the cmd . from these residuals
a reddening map is created by finding the median residual in each @xmath42 pixel square of the image .
then each star is corrected , along its reddening line , by an amount that is interpolated from the @xmath43 points of the reddening map . a full explanation of these procedures , the differentially corrected cmds , and reddening maps will be provided in a forthcoming paper ( i. king , 2009 , in preparation ) .
the best - fit age of a given gc was estimated by determining the isochrone that best fit the cmd from the msto through the subgiant branch .
the uncertainty was derived from the intrinsic scatter in the cmd and/or the inherent mismatch between the models and the data .
the range of ages that allowed the isochrones to envelope the bulk of the stars at the msto and on the subgiant branch are taken to be 1-@xmath11 uncertainties .
the age uncertainty is only based on the fitting procedure described and does not account for uncertainties in the input physics of the stellar models or differences between the chemical composition assumed in the models and those actually present in the stars ( for more on this point , see @xmath16[metage ] ) .
hence the uncertainties include the random component but exclude the systematic .
still , it is worthwhile to consider that incomplete knowledge of chemical composition and incomplete treatment of the physics , such as rotation and convection , in the current generation of 1-d stellar evolution models are primary sources for the systematic error present in the analysis .
the results of isochrone fitting are listed in table [ age ] . if a differential reddening corrected cmd was employed in the fit , an asterisk ( * ) appears after the name in table [ age ] .
figures [ iso6362 ] through [ iso7099 ] demonstrate how the isochrone fits were achieved for three different gcs .
figure [ iso6362 ] shows a case where the isochrones trace the stellar population throughout the cmd ( left panel ) and how the age , [email protected] gyr , was measured ( right panel ) .
figure [ iso3201 ] shows the differential reddening - corrected cmd of ngc 3201 .
the corrected cmd reduces the scatter about the subgiant branch but the age uncertainty is still larger than in the case of ngc 6362 .
figure [ iso7099 ] shows the worst case scenario where the reddening is low and the isochrones match the unevolved main sequence and rgb but not the shape of the age sensitive region .
the slope of the models is shallower than the data along the subgiant branch and , although the data define a narrow sequence , the age is poorly constrained .
this case is likely to be severely affected by the systematic errors described in the preceding paragraph . as a consistency check
, it is useful to plot the absolute magnitude of the hb as a function of metallicity .
the absolute magnitude of the hb is obtained by subtracting the @xmath17 distance modulus derived from the isochrone fits ( table [ age ] ) from the apparent magnitude of the hb ( table [ morph ] ) .
the resulting quantity is plotted in figure [ mvhb ] for the acs survey clusters .
the points define a relatively tight relationship .
a linear , least squares fit to these data for gcs with @xmath6}}\leq -1 $ ] gives : @xmath44}}+ 0.802\pm0.020,\ ] ] which is shown as the solid line in figure [ mvhb ] .
the quoted errors in equation ( [ mveq ] ) are due only to scatter in the data and do not include measurement uncertainties .
the fit is limited to clusters with @xmath6}}\leq -1 $ ] so as to exclude clusters with predominantly red hb morphologies .
the justification for restricting the fit is that red hb gcs often lack rr lyrae variable stars .
some gcs with completely blue hbs were included in the fit even though they may not have rr lyrae stars .
as described in @xmath16[vhb ] , the uncertainty in the apparent magnitude of the hb is approximately 0.05 mag per gc and this is shown by the two dashed lines .
all but a handful of points lie within the dashed lines .
if equation ( [ mveq ] ) is transformed from @xmath17 to @xmath23 using the synthetic color transformations employed in paper ii , the slope increases by @xmath45 because of a very slight metallicity dependence of @xmath46 on @xmath6}}$ ] and , assuming a characteristic temperature for rr lyrae stars of logt=3.83 , the intercept increases by 0.09 mag .
@xcite gave @xmath47 = ( [email protected])(@xmath6}}$]+1.6 ) + ( [email protected] ) .
other recent estimates include @xmath48=([email protected])(@xmath6}}$]+1.5)+([email protected] ) @xcite and a further refinement of the slope to [email protected] mag / dex @xcite . if equation ( [ mveq ] ) is transformed again so that the independent variable is ( @xmath6}}+1.6 $ ] ) and the dependent variable is @xmath49 rather than @xmath50 , it becomes @xmath51}}+1.6 ) + 0.53 $ ] where the error bars have been omitted for brevity .
the relationship between @xmath48 and @xmath6}}$ ] derived here is within the 1-@xmath11 uncertainties from each of the three determinations previously mentioned .
figure [ met ] compares @xmath6}}$ ] from table [ age ] and the zw84 scale with the ( * ? ? ?
* hereafter ki03 ) scale for 47 gcs in common between ki03 and this paper .
( ki03 provided @xmath6}}$ ] values measured with three different model atmospheres ; this paper adopts their measurements based on marcs models . ) for reference , the line of equality is drawn on both panels of figure [ met ] .
paper vii compared ages on the zw84 and ( * ? ? ? *
hereafter cg97 ) scales .
as discussed by ki03 , the cg97 scale is consistently higher than either the zw84 or ki03 scales , by 0.2 - 0.3 dex in general .
it is beyond the scope of this paper to determine which scale is superior to the others0.2 dex lower than that of cg97 .
] . suffice it to say that ages and metallicities from table [ age ] and paper vii on the zw84 scale will be used in the analysis that follows .
however , speaking hypothetically , if it is assumed that the ki03 scale is the ` correct ' @xmath6}}$ ] scale , then the residuals from the solid line in each panel of figure [ met ] can be interpreted as uncertainties in the adopted scales .
we shall return to this hypothesis in @xmath16[2ndp ] .
the ages reported in tables [ gcages ] are based on isochrones that assume an @xmath37-enhanced composition .
they do not explicitly account for variations in the he or c+n+o abundances but , if such variations are present , they should certainly influence age determinations @xcite . while the extent of he and c+n+o variations in gcs are largely unknown at present
, the existence of light element abundance variations is quite clear .
the review by @xcite , the recent results presented by @xcite , and many other studies have shown the widespread presence of chemical abundance variations within individual gcs .
these abundance variations are among the light elements ( c , n , o , na , mg , and al ) and do not extend to heavier elements such as fe in the majority of gcs . models for gc self - enrichment predict that varying degrees of he variation should accompany the other abundance variations @xcite .
the influence of such abundance variations on gc ages is not fully understood at present .
however , the recent results from @xcite lend some insight into the problem of deriving gc ages in the presence of light element abundance variations .
those authors identify three groups of stars within gcs based on their chemistry .
they are : the ` primordial ' stars or stars which have halo - like abundances , characterized by super - solar o and sub - solar na ; the ` intermediate ' stars , with slightly less o and more na than the primordial stars ; and the ` extreme ' stars , characterized by [ o / na ] @xmath520.9 .
@xcite state that all gcs ( measured so far ) contain primordial and intermediate stars but not all gcs contain the extreme component .
furthermore , the majority of gc stars are in the intermediate group
. how do these findings influence gc age estimates ? @xcite compared stellar evolution models from the basti library @xcite with chemical compositions representative of the primordial and extreme groups .
@xcite state that , as long as the total amount of c+n+o remains constant , it is safe to use @xmath37-enhanced isochrones to derive gc ages .
however , if the total amount of c+n+o varies among the stars in a given gc , then it is necessary to use models with the appropriate cno abundances to derive gc ages .
this is all without consideration of a potential change in the he content that is predicted to accompany the light element abundance variations .
the he abundance is important because models of chemical enrichment in gcs predict that the abundance variations from the primordial values should be accompanied by an increase in he , whether the source of the enrichment is massive stars @xcite or intermediate mass agb stars @xcite .
the extreme stars should be most he - enriched , the intermediate stars should be slightly he - enriched , and the primordial stars should have primordial he .
@xcite investigated the he content of 57 gcs using the r parameter : the ratio of the number of hb stars to the number of rgb stars brighter than the hb . within the uncertainties , they found little or no evidence for a spread in the he content of gcs with ( b@xmath29r)/(b+v+r ) @xmath53 0.8 . for gcs with ( b@xmath29r)/(b+v+r )
@xmath54 0.8 , @xcite found a larger spread and a tendency toward higher he abundances .
these authors state that the apparent trend towards higher he abundances in gcs with blue hbs may be related to increased evolutionary timescales for the lowest mass hb stars not entirely accounted for in their calibration of the r parameter or the genuine presence of he - rich stars in the blue hbs of some gcs . as long as the total metal content remains constant ,
a modest spread ( @xmath8y @xmath55 0.05 ) in the he content of a gc will not significantly alter the level of the main sequence turnoff or subgiant branch @xcite and therefore not confuse the age determination .
if the total metal content varies or the spread in he is larger , age estimates will require great care and detailed abundance information .
it is important to keep in mind the complexities of measuring gc ages in the context of chemical abundance variations and their proposed origins in gcs .
the average difference between the ages derived in @xmath16[isoage ] and in paper vii is @xmath56 gyr ; the standard error of the mean is 0.105 gyr and the standard deviation is 0.781 gyr .
figure [ relages ] shows the normalized age differences of 56 gcs .
the age differences are calculated by subtracting the paper vii age from the age listed in table [ age ] and dividing by the quadrature sum of the age uncertainties from both sources . in accordance with paper vii
, metallicity is represented in figure [ relages ] and later figures by [ m / h]=@xmath6}}$]+log@xmath57}}}+0.362)$ ] @xcite .
the ages of 43 of the 55 gcs shown in figure [ relages ] differ by less than 1-@xmath11 .
the primary reason for disagreement in ages is most likely the adopted metallicity scale since the largest systematic deviation occurs around [ m / h]@xmath58 and the metallicity scales deviate there by as much as 0.3 dex ( see figure [ met ] ) .
figure [ amr ] compares the amrs from this paper with those of paper vii ( top panel ) and @xcite ( bottom panel ) . despite the age differences already addressed in figure [ relages ] and the preceding paragraph
, both amrs reveal the same general features .
the main difference is that the ages from table [ age ] show smaller dispersion at low and high metallicities .
the appearance of a separate trend beginning at [ m / h]=@xmath59 and extending to pal 12 and ter 7 ( the two youngest gcs ) is nearly identical in both amrs .
the open circles in the upper panel are the outer halo gcs not present in the acs survey data ; these additional gcs strengthen the trend already present in the acs survey data . given that @xcite carried out his analysis using a heterogeneous collection of data in @xmath60 and @xmath31 , and used distances set by the level of the hb , it is encouraging that all three amrs reveal the same basic trends .
hb morphology is most strongly influenced by the first parameter , metallicity .
the metallicity-@xmath0 diagram is shown in figure [ dvimh ] . for the acs survey data , the inner halo gcs ( @xmath61 kpc )
are plotted as circles and the outer halo gcs ( @xmath12 kpc ) as squares .
the six most distant gcs are plotted as triangles and only appear in the left panel because they were not considered in paper vii .
the left panel shows the metallicity scale from table [ age ] and the right panel shows the zw84 metallicity scale from paper vii .
the error bars shown are only due to measurement errors in @xmath0 as listed in table [ morph ] . in order to more clearly identify the second parameter
, an attempt was made to remove the influence of metallicity from the hb morphology data , thereby exposing the second parameter effect . to accomplish this ,
a function was fit to the inner halo gcs which cover the full range of metallicity and exhibit relatively little scatter in the [ m / h]-@xmath0 diagram , with the notable exception of ngc 6584 .
ngc 6584 is presently located at @xmath62 kpc but @xcite noted its position in the metallicity - hb morphology diagram and large , positive heliocentric velocity and suggested it is actually an outer halo gc .
this suggestion is supported by @xcite who calculated its apogalacticon at @xmath63 kpc . as a result ,
m 5 ( see @xmath16[intro ] ) and ngc 6584 were excluded from the fit .
the fitting function is made up of two parts with @xmath64=[m / h ] , @xmath65 the first part , @xmath66 , resembles a fermi - dirac function that transitions from blue to red as metallicity increases , as suggested by @xcite .
the second part , @xmath67 , is a quadratic that allows for the remaining curvature , in particular that the most metal - poor gcs turn back to the red with decreasing metallicity as can be seen in figure [ dvimh ] .
the two parts are @xmath68\ ] ] and @xmath69 the @xmath70 s and @xmath71 s were determined from a least squares fit . for comparison , and to demonstrate the robustness of the method , the same functional form was fit to the inner halo @xmath0 data using the metallicity scale from table [ age ] ( shown in the left panel of figure [ dvimh ] ) and the zw84 metallicity scale from paper vii ( right panel ) .
the fitting function coefficients are reproduced in table [ coeff ] .
the search for correlations focuses on the difference between the measured @xmath0 of an outer halo gc and the @xmath0 value of the inner halo trend ( as represented by the fitting function ) at that gc s metallicity .
this quantity will henceforth be referred to as @xmath72 .
figure [ dhbmh ] shows how @xmath72 varies with metallicity ; the symbols are the same as in figure [ dvimh ] . with the exception of ngc 6584 the inner halo gcs exhibit residuals of @xmath10.1 or less .
altogether , metal - rich gcs ( [ m / h]@xmath73 ) show very little scatter in figure [ dhbmh ] . evidently , metallicity alone is almost sufficient to describe the hb morphology of the majority of metal - rich gcs and thus they are of little use in the search for the second parameter .
the solid error bars shown in figure [ dhbmh ] represent the measurement errors listed in table [ morph ] .
metallicity uncertainty plays an important role in this diagram and thus the error bars in figure [ dhbmh ] are only a lower limit to the total uncertainty in @xmath0 .
unfortunately , it is not possible to provide realistic metallicity errors for all of the gcs in the sample . for the large deviation shown for @xmath74 [ m / h ] @xmath75 in figure [ dhbmh ] to be completely erased , metallicities of the gcs in this region that stand out would have to be systematically under - estimated by 0.25 - 0.5 dex .
such a conspiracy seems highly unlikely .
however , it is possible to approximate the effect of metallicity errors on @xmath0 in the following manner .
first , the fitting function described by equations [ fit1 ] and [ fit2 ] was applied to the inner halo gcs with metallicities on the ki03 scale .
next , the ki03-based fitting function was used to calculate a @xmath0 but using the metallicity from table [ age ] or the zw84 scale , thereby introducing a metallicity error .
finally , the difference between the erroneous @xmath0 value obtained in this manner and the value obtained from the appropriate fitting function provides an estimate of how a metallicity error propagates into a @xmath0 error . under the hypothesis that the ki03 @xmath6}}$ ] scale is the ` correct ' one , this procedure gives an indication of how a metallicity error ( a vertical displacement in figure [ dvimh ] ) influences the significance of the distance between a gc and the inner halo fitting function ( the horizontal distance between a point and the fit line in figure [ dvimh ] ) which we refer to as @xmath72 . the dashed error bars in figure [ dhbmh ] are the quadrature sum of the @xmath0 measurement errors from table [ morph ] and the metallicity errors estimated as described in the preceding paragraph . the metallicity effect is small except in the vicinity of the transition from red to blue that occurs around [ m / h ] @xmath58 ( see figure [ dvimh ] ) where a small change in metallicity corresponds to a large change in @xmath0 .
it should be stressed that this approach is not a rigorous treatment of the influence of metallicity uncertainty on @xmath0 or @xmath72 because it assumes the difference between one scale and another is the error .
nevertheless , it serves to illustrate that the existence of several gcs with @xmath76 [ m / h ] @xmath77 and large , positive values of @xmath72 can not readily be attributed to metallicity errors alone .
figures [ dhbrgc ] through [ dhbage ] show how the @xmath0 difference depends on @xmath4 , @xmath24 , @xmath25 , @xmath26 , @xmath27 , and age for the outer halo gcs . in order to quantify the information provided in the plots , spearman s rank correlation coefficient
was calculated for the data presented in figures [ dhbrgc ] through [ dhbage ] and included in the figures .
the spearman rank correlation coefficient ( spearman s @xmath78 ) measures the degree of correlation between two variables but makes no assumption about the functional form of their relationship other than monotonicity .
as such , it is a very general measure of correlation ; a perfect correlation has a value of @xmath79 , a perfect anti - correlation has @xmath80 , and complete lack of correlation has @xmath81 .
the left panels show the 6 most distant gcs , which were not part of the acs survey , as triangles .
since these 6 gcs are not part of the homogeneous acs data set , and were not considered in paper vii , we have calculated the spearman correlation coefficients with and without them in the left panels .
apart from ngc 2419 , the other 5 most distant gcs tend to strengthen the correlation coefficients because they are more distant , intrinsically fainter , more extended , and younger .
@xmath0 error bars are included in figures [ dvimh ] through [ dhbage ] , and age errors in figure [ dhbage ] , but uncertainties in the other quantities are neither readily available nor simple to estimate given the wide range of distances and reddenings present in the sample .
it is , however , important to consider that the quantities in figures [ dhbrgc ] through [ dhbrho ] were converted to physical units after adopting a particular distance modulus . as distance increases ,
a given error in the distance corresponds to a larger uncertainty in the physical quantity . while @xmath4 ,
@xmath24 , @xmath25 , @xmath26 , and @xmath27 reveal no obvious visible trends , figure [ dhbage ] shows a trend with age .
the spearman coefficients support this conclusion : @xmath24 , @xmath25 , @xmath26 , and @xmath27 show no strong correlations .
@xmath4 shows a mild correlation but that is mostly likely due to a large anti - correlation between age and @xmath4 .
age shows a significant anti - correlation and the age trend works as anticipated : the hbs grow redder with decreasing age and so the distances between the measured @xmath0 values and the inner halo trend increase as age decreases .
the results presented in this section demonstrate that , of all the parameters considered here , age has the most significant correlation with @xmath72 .
figure [ dhbage ] and the accompanying slopes and coefficients do not include pal 12 or terzan 7 , the two youngest gcs in the sample
. both gcs are metal rich and so their hb morphologies are almost entirely determined by their metallicity alone .
the presence or absence of pal 12 and ter 7 does not significantly alter the correlation coefficients ( @xmath82 ) for any of the other quantities .
figure [ dhbage ] also includes a linear fit to each data set that describes the rate at which hb morphology varies with age .
the ages and metallicities listed in table [ age ] suggest that hb morphology changes over a smaller age interval than those from paper vii , but only by about @xmath10.5 gyr .
the larger errors that include the metallicity effect were not used in the least squares fits to the age-@xmath72 relations shown in figure [ dhbage ] .
if they had been , the slope of the relation derived using the paper vii ages ( shown in the right panel ) would be reduced to @xmath83 while the slope derived from the ages given in table [ age ] would not change significantly . in @xmath16[2ndp ] , it was shown that when the inner halo relationship between metallicity and @xmath0 is subtracted from the outer halo gcs , a correlation between age and @xmath72 appears .
the inner halo gcs were chosen to define the trend because they exhibit a tight relation with little scatter in the hb morphology
metallicity diagram ( * ? ? ?
* see also figure [ dvimh ] in this paper ) .
unfortunately , attempts to use this technique a second time to subtract off the age trend were unsuccessful because the remaining residuals are generally smaller than the measurement uncertainties in @xmath0 .
thus , to identify a potential third parameter , it is necessary to remove the effect of metallicity and age to the fullest extent possible .
the most metal poor gcs ( [ m / h ] @xmath84 ) in the sample are an ideal choice because they have a small range in ages ( @xmath16[gcages ] ) and only a weak metallicity dependence on hb morphology ( see figure [ dvimh ] ) .
comparisons of the metal poor gcs @xmath72 values with @xmath24 , @xmath25 , @xmath26 , and @xmath27 were performed . of these ,
central density ( log @xmath27 ) produced the most significant trend .
figure [ rho ] demonstrates how central density relates to @xmath72 among the metal poor gcs using the two fitting functions employed in @xmath16[2ndp ] and the same symbols used in figures [ dvimh ] through [ dhbage ] .
there is an evident trend with the highest central density gcs having the bluest hb morphologies and hence the smallest @xmath72 values . assuming that the contributions of age and metallicity are minimal in figure [ rho ] , the influence that central density has on hb morphology in the metal poor gcs is evident and its magnitude is @xmath85 in @xmath72 .
although it is difficult to disentangle the effects of age and central density in gcs with intermediate metallicities ( @xmath74 [ m / h ] @xmath75 ) , where the second parameter is most pronounced , these gcs span essentially the same range of central densities as the metal poor gcs .
thus central density should only account for a small portion of the hb morphology variation seen in the intermediate metallicity group .
the influence of central density on hb morphology has already been demonstrated by @xcite , @xcite , and @xcite .
however , its effect is not as pronounced as that of age ( compare figures [ dhbage ] and [ rho ] ) .
therefore central density is the most likely candidate for the third parameter influencing hb morphology ( as characterized by @xmath0 ) that has been considered by this study .
it is worthwhile to demonstrate that the use of the inner halo gcs to define the hb morphology
metallicity relation does not include an unintended bias .
in particular , since claims have been made in the preceding sections that , after metallicity , age and central density are the two most significant factors influencing hb morphology [ as represented by @xmath0 ] , it is important to demonstrate that the hb morphology metallicity trend derived from the inner halo gcs does not include an implicit dependence on either of these quantities .
indeed , as figure [ inner ] shows , the inner halo gcs exhibit some dispersion inside 8 kpc in both central density ( left panel ) and age ( right panel ) but there is no clear , systematic variation of either parameter as a function of @xmath4 inside 8 kpc ( denoted by the dotted line in the figure ) . figure [ dvimh ] demonstrates that the hb morphologies of the inner halo gcs , which have a fairly homogeneous age distribution , transition from blue to red as [ m / h ] rises through @xmath58 .
the metal rich gcs have red hbs ( though ngc 6388 and ngc 6441 also have blue extensions ) while the majority of metal poor gcs have blue hbs .
it is at metallicities that are lower than , but within a few tenths of a dex of , [ m / h]=@xmath87 where a gc is most likely to move a large distance in the @xmath0metallicity diagram due to a relatively small change in metallicity or another parameter .
it is within this same narrow metallicity range that the amr of the outer halo gcs branches off from that of the inner halo gcs ( see figure [ amr ] ) and so it is no coincidence that the classic second parameter effect is most pronounced in the outer halo .
the analysis presented in @xmath16[2ndp ] indicates that a gc with @xmath74[m / h]@xmath75 requires 2 - 2.5 gyr to transition from a red hb to a blue one .
this result is consistent with a number of studies mentioned in the introduction .
we confirm the assertion by @xcite that the presence of metal poor gcs with red hbs at large @xmath4 is primarily an age effect .
however , our result indicates that lee et al.s claim that the necessary age difference is greater than @xmath13 gyr is slightly exaggerated . that a gc with @xmath74 [ m / h ] @xmath75 transitions from a red to a blue hb in 2 - 2.5 gyr poses a challenge to our understanding of mass loss during the red giant phase of evolution .
rgb mass loss has been essentially a free parameter in synthetic hb models since they were first constructed by @xcite .
a number of mass loss rates that vary as a function of global stellar properties such as luminosity , mass , radius , or some combination of the three have been proposed with varying degrees of success , see @xcite for a detailed discussion . following @xcite , @xcite showed that a simple relationship between global metallicity and average rgb mass loss , along with stochastic variations , can reproduce the general trend observed in old , roughly coeval galactic gcs in the hb morphology - metallicity diagram .
recent progress from observations by , e.g.@xcite , @xcite , @xcite , and @xcite , should lead to a better understanding of rgb mass loss as a function of composition , evolutionary status , and pulsational properties .
the recent , large - scale study of hb morphology by @xcite suggests that , just as age can explain anomalously red hbs in some gcs , total gc mass is linked to the degree of abundance variation and the extent of the faint blue tail observed in other gcs .
( of course , there is no reason why both effects can not be operating in the same gc . )
recall that figure [ dvibvr ] showed the hb morphology metric adopted by @xcite was very sensitive to the extremes but lacked sensitivity through the middle of the distribution . in the absence of an ideal hb morphology metric , a large - scale study of hb morphology should choose a metric that is well - suited to the effect(s ) of interest to that investigation .
hb morphologies characterized by @xmath0 of 66 galactic gcs were examined to determine the sensitivity of hb morphology to a variety of different factors .
deep , homogeneous photometry from the acs survey of galactic gcs accounts for 60 of these while the remaining 6 gcs are the most distant galactic gcs known .
the complete sample is the largest examined to date and consists solely of high - quality hst photometry .
it spans the full range of @xmath4 and almost the entire range of metallicity present in the galactic gc population .
@xmath0 values and two independently measured sets of ages were joined with other quantities from the literature to assess the relative importance of these quantities on hb morphology .
the data were split into two groups , roughly equal in number , based on @xmath4 .
the tight relationship between metallicity and hb morphology in the inner halo group ( @xmath13 kpc ) was characterized by a fitting function and this trend was subtracted off of the outer halo group . the difference between fit and measured @xmath0
was then compared to a variety of parameters , of which only age showed a significant correlation .
the age correlation does not rely on the presence of the 6 most distant gcs in the analysis , though their presence does strengthen the result .
hence we conclude that , after metallicity , age has the most influence on @xmath0 .
the age spread among the bulk of gcs in the galactic halo was found to be 2 - 2.5 gyr , though there are a few younger outliers such as pal 12 and ter 7 .
further analysis , in which both metallicity and age were restricted , provided strong evidence that central luminosity density ( @xmath27 ) is the third most influential parameter on @xmath0 .
we wish to express our gratitude to the anonymous referee for a thoughtful , thorough report that improved the accuracy and focus of the paper .
we thank ivan king for sharing his differential reddening corrected cmds in advance of publication and the following people for sharing their hst photometry : peter stetson ( pal 3 , pal 4 , and eridanus ) , bill harris ( ngc 2419 ) , and eric sandquist ( ngc 2419 ) .
ad acknowledges support from a cita national fellowship and an nserc grant to d. vandenberg .
further support for this work ( proposal number go-10775 ) was provided by nasa through a grant from the space telescope science institute which is operated by the association of universities for research in astronomy , incorporated , under nasa contract nas5 - 26555 .
anderson , j. et al .
2008 , , 135 , 2055 ( paper v ) anderson , j. piotto , g. , king , i. r. , bedin , l. r. , & guhathakurta , p. 2009
, , 697 , l58 arp , h. c. , baum , w. a. , & sandage , a. r. 1952 , , 57 , 4 bedin , l. r. , salaris , m. , piotto , g. , anderson , j. , king , i. r. , & cassisi , s. 2009 , , 697 , 965 bellazzini , m. , fusi pecci , f. , ferraro , f. r. , galleti , s. , catelan , m. , landsman , w. b. 2001 , , 122 , 2569 bohlin , r. c. 2007 , instrument science report acs 2007 - 06 bolte , m. 1989 , , 97 , 1688 buonanno , r. , corsi , c. , bellazzini , m. , ferraro , f. r. , & pecci , f. fusi 1997 , , 113 , 706 buonanno , r. , corsi , c. , castellani , m. , marconi , g. , fusi pecci , f. , & zinn , r. 1999 , , 118 , 1671 carretta , e. , bragaglia , a. , gratton , r. , dorazi , v. , & lucatello , s. 2009c , in press ( arxiv:0910.0675 ) carretta , e. , bragaglia , a. , gratton , r. g. , & lucatello , s. 2009b , , 505 , 117 carretta , e. et al . 2009a , , 505 , 139 carretta , e. & gratton , r. g. 1997 , 1997 , , 121 , 95 ( cg97 ) carretta , e. , recio - blanco , a. , gratton , r. , piotto , g. , & bragaglia , a. 2007 , , 671 , 125 cassisi , s. , castellani , v. , deglinnocenti , s. , piotto , g. , & salaris , m. , 2001 , , 366 , 578 catelan , m. 2000 , , 531 , 826 catelan , m. 2009 , , 320 , 261 catelan , m. , bellazini , m. , landsman , w. b. , ferraro , f. r. , fusi pecci , f. , & galleti , s. 2001 , , 122 , 3171 catelan , m & de freitas pacheco , j. a. 1993 , , 106 , 1858 catelan , m & de freitas pacheco , j. a. 1995 , 297 , 345 catelan , m. , ferraro , f. r. , & rood , r. t. 2001 , , 560 , 970 chaboyer , b. 1999 , assl , 237 , 111 chaboyer , b. , demarque , p. , & sarajedini , a. 1996 , , 459 , 558 chaboyer , b. , sarajedini , a. , & demarque , p. 1992
, , 394 , 515 decressin , t. , meynet , g. , charbonnel , c. , prantzos , n. , & ekstrm , s. 2007 , , 464 , 1029 dantona , f. , caloi , v. , montalbn , j. , ventura , p. , gratton , r. 2002 , , 395 , 69 dantona , f. , stetson , p. b. , ventura , p. , milone , a. p. , piotto , g. , & caloi , v. 2009 , , 399 , l151 dinescu , d. i. , girard , t. m. , & van altena , w. f. 1999 , , 117 , 1792 dotter , a. 2008 , , 687 , 21 dotter , a. et al .
2007 , , 134 , 376 ( paper ii ) dotter , a. chaboyer , b. , kostov , v. , jevermovc , d. , baron , e. , & ferguson , j. w. 2008a , , 178 , 89 dotter , a. , kaluzny , j. , & thompson , i. b. 2009 , iaus , 258 , 171 dotter , a. , sarajedini , a. & yang , s .- c . , 2008b , , 136 , 1407 dupree , a. k. , smith , g. , & strader , j. 2009 , , 138 , 1485 fusi pecci , f. , ferraro , f. r. , bellazzini , m. , djorgovski , s. , piotto , g. , & buonanno , r. 1993 , , 105 , 1145 glatt , k. , et al .
2008 , , 136 , 1703 gratton , r. g. , bragaglia , a. , carretta , e. , clementini , g. , desidera , s. , grundahl , f. , & lucatello , s. , 408 , 529 gratton , r. g. , bragaglia , a. , clementini , g. , carretta , e. , di fabrizio , l. , maio , m. , & taribello , e. , 421 , 937 gratton , r. , sneden , c. , & carretta , e. 2004 , , 42 , 385 green , e. m. & norris , j. e. 1990 , , 353 , l17 harris , w. e. 1996 , , 112 , 1487 harris , w. e. et al .
1997 , , 114 , 1030 johnson , j. a. & bolte , m. 1998 , , 115 , 693 johnson , j. a. , bolte , m. , stetson , p. b. , hesser , j. a. , & somerville , r. a. 1999 , , 527 , 199 johnson , c. i. , pilachowski , c. a. , simmerer , j. , & schwenk , d. 2008 , , 681 , 1505 koch , a. , ct , p. , & mcwilliam , a. 2009 , , 506 , 729 kraft , r. p. & ivans , i. i. 2003 , , 115 , 143 ( ki03 ) lee , y .- w .
1989 , ph.d .
thesis , yale university , new haven , ct .
lee , y .- w . , 1991 , , 373 , 43 lee , y .- w . , demarque , p. , & zinn , r. j. 1990 , , 350 , 155 lee , y .- w .
, demarque , p. , & zinn , r. j. 1994 , , 423 , 248 lee , y .- w . , joo , j .-
m . , sohn , y .-
j . , rey , s .- c . ,
lee , h .- c . , & walker , a. r. 1999 , , 402 , 55 mackey , a. d. & gilmore , g. f. 2004 , , 355 , 504 mackey , a. d. & van den bergh , s. 2005 , , 360 , 631 marn - franch , a. et al .
2009 , , 694 , 1498 ( paper vii ) marino , a. f. , villanova , s. , piotto , g. , milone , a. p. , momany , y. , bedin , l. r. , & medling , a. m. 2008 , , 490 , 625 mcdonald , i , van loon , j. th .
, decin , l. , boyer , m. l. , dupree , a. k. , evans , a. , gehrz , r. d. , & woodward , c. e. , , 394 , 831 mszros , s. , avrett , e. h. , dupree , a. k. , , 138 , 615 milone , a. et al .
2008 , , 673 , 241 ( paper iii ) origlia , l. rood , r. t. , fabbri , s. , ferraro , f. r. , fusi pecci , f. , & rich , r. m. 2007 , , 667 , 850 pietrinferni , a. , cassisi , s. , salaris , m. , & castelli , f. 2004 , , 612 , 168 pietrinferni , a. , cassisi , s. , salaris , m. , & castelli , f. 2006 , , 642 , 797 pietrinferni , a. , cassisi , s. , salaris , m. , percival , s. & ferguson , j. w. 2009 , , 697 , 275 piotto , g. 2009 , iaus , 258 , 233 piotto , g. et al .
2007 , , 661 , 53 piotto , g. et al .
2002 , , 391 , 945 pritzl , p. j. , venn , k. a. , & irwin , m. 2005 , , 130 , 2140 recio - blanco , a. , aparicio , a. , piotto , g. , de angeli , f. , djorgovski , s. g. 2006 , , 452 , 875 rey , s. c. , yoon , s .- j . , lee , y .- w . , chaboyer , b. , & sarajedini , a. 2001 , , 122 , 3219 richer , h. b. et al .
1996 , , 463 , 602 rood , r. t. 1973 , , 184 , 815 rosenberg , a. , saviane , i. , piotto , g. , & aparicio , a. , 118 , 2306 salaris , m. , chieffi , a. & straniero , o. 1993 , , 414 , 580 salaris , m. , riello , m. , cassisi , s. , & piotto , g. 2004 , , 420 , 911 sandage , a. r. 1953 , , 58 , 61 sandage , a. & wallerstein , g. 1960 , , 131 , 598 sandage , a. & wildey , r. 1967 , , 150 , 469 sandquist , e. & hess , j. m. 2008 , , 136 , 2259 sarajedini , a. 1999 , aspc , 165 , 295 sarajedini , a. et al .
2007 , , 133 , 1658 ( paper i ) sarajedini , a. & demarque , p. 1990
, , 365 , 219 sarajedini , a. & king , c. r. 1989 , , 98 , 1624 sarajedini , a. , lee , y. w. , & lee , d. h. 1995 , , 450 , 712 scholz , r .- d . , odenkirchen , m. , hirte , s. , irwin , m. j. , borngen , f. , & ziener , r. , 1996 , , 278 , 251 searle , l. & zinn , r. 1978 , , 225 , 357 shetrone , m. d. & keane , m. j. 2000 , , 119 , 840 siegel , m. et al .
2007 , , 667 , 57 sirianni , m. et al .
2005 , , 117 , 1049 smith , g. 2004 , observatory , 124 , 365 sneden , c. , kraft , r. p. , guhathakurta , p. , peterson , r. c. , fulbright , j. p. 2004
, , 127 , 2162 stetson , p.
al . 1999 , , 117 , 247 sweigart , a. v. & catelan , m. 1998 , , 501 , l63 van den bergh , s. 1965 , , 59 , 151 van den bergh , s. 1967 , , 72 , 70 vandenberg , d. a. 2000 , , 129 , 315 vandenberg , d. a. , bolte , m. , & stetson , p. b. 1990 , , 100 , 445 vandenberg , p. b. & durrell , p. r. 1990 , , 99 , 221 ventura , p. , caloi , v. , dantona , f. , ferguson , j. , milone , a. , & piotto , g. 2009 , mnras , 399 , 934 zinn , r. & west , m. j. 1984 , , 55 , 45 ( zw84 ) cccccc arp 2 & 18.05 & 0.316 @xmath5 0.017 & 0.200 & 1.056 @xmath5 0.018 & 0.740 @xmath5 0.025 + lyng7 & 17.10 & 1.863 @xmath5 0.003 & 0.034 & 1.974 @xmath5 0.015 & 0.111 @xmath5 0.015 + ngc 104 & 13.90 & 0.916 @xmath5 0.002 & 0.023 & 1.069 @xmath5 0.002 & 0.153 @xmath5 0.003 + ngc 288 & 15.35 & -0.018
@xmath5 0.023 & 0.127 & 1.004 @xmath5 0.011 & 1.022 @xmath5 0.025 + ngc 362 & 15.33 & 0.786 @xmath5 0.003 & 0.083 & 0.981 @xmath5 0.001 & 0.195 @xmath5 0.003 + ngc 1261 & 16.60 & 0.762 @xmath5 0.004 & 0.129 & 0.963 @xmath5 0.003 & 0.201 @xmath5 0.005 + ngc 1851 & 16.00 & 0.764 @xmath5 0.011 & 0.272 & 0.998 @xmath5 0.002 & 0.234 @xmath5 0.011 + ngc 2298 & 16.00 & 0.332 @xmath5 0.018 & 0.123 & 1.203 @xmath5 0.007 & 0.871 @xmath5 0.019 + ngc 2808 & 16.10 & 0.235 @xmath5 0.025 & 0.439 &
1.201 @xmath5 0.002 & 0.966 @xmath5 0.025 + ngc 3201 & 14.60 & 0.767 @xmath5 0.108 & 0.270 & 1.290 @xmath5 0.011 & 0.523 @xmath5 0.108 + ngc 4147 & 16.85 & 0.103 @xmath5 0.031 & 0.136 & 0.941 @xmath5 0.011 & 0.838 @xmath5 0.033 + ngc 4590 & 15.65 & 0.291 @xmath5 0.011 & 0.147 & 0.991 @xmath5 0.003 & 0.700 @xmath5 0.011 + ngc 4833 & 15.50 & 0.451 @xmath5 0.029 & 0.154 & 1.351 @xmath5 0.003 & 0.900 @xmath5 0.029 + ngc 5024 & 16.80 & 0.090 @xmath5 0.009 & 0.088 & 0.941 @xmath5 0.003 & 0.851 @xmath5 0.010 + ngc 5053 & 16.60 & 0.145 @xmath5 0.041 & 0.158 & 0.927 @xmath5 0.003 &
0.782 @xmath5 0.041 + ngc 5272 & 15.55 & 0.207 @xmath5 0.014 & 0.208 & 0.943 @xmath5 0.002 & 0.736 @xmath5 0.014 + ngc 5286 & 16.40 & 0.372 @xmath5 0.011 & 0.182 & 1.245 @xmath5 0.003 & 0.873 @xmath5 0.011 + ngc 5466 & 16.45 & 0.208 @xmath5 0.024 & 0.124 & 0.929 @xmath5 0.006 & 0.721 @xmath5 0.025 + ngc 5904 & 14.90 & 0.120 @xmath5 0.024 & 0.230 & 0.994 @xmath5 0.001 & 0.874 @xmath5 0.024 + ngc 5927 & 16.35 & 1.490 @xmath5 0.002 & 0.025 & 1.602 @xmath5 0.006 & 0.112 @xmath5 0.007 + ngc 5986 & 16.40 & 0.357 @xmath5 0.011 & 0.136 & 1.311 @xmath5 0.004 & 0.954 @xmath5 0.012 + ngc 6093 & 16.20 & 0.275 @xmath5 0.008 & 0.153 & 1.198 @xmath5 0.003 &
0.923 @xmath5 0.008 + ngc 6101 & 16.50 & 0.203 @xmath5 0.007 & 0.091 & 1.061 @xmath5 0.005 & 0.858 @xmath5 0.009 + ngc 6121 & 13.17 & 0.987 @xmath5 0.106 & 0.252 & 1.528 @xmath5 0.010 & 0.541 @xmath5 0.106 + ngc 6144 & 16.20 & 0.612 @xmath5 0.009 & 0.100 & 1.491 @xmath5 0.007 & 0.879 @xmath5 0.011 + ngc 6171 & 15.40 & 1.296 @xmath5 0.013 & 0.104 & 1.499 @xmath5 0.009 & 0.204 @xmath5 0.015 + ngc 6205 & 14.85 & -0.091 @xmath5 0.010 & 0.122 & 0.956 @xmath5 0.002 & 1.047 @xmath5 0.010 + ngc 6218 & 14.60 & 0.227 @xmath5 0.009 & 0.099 & 1.213 @xmath5 0.004 & 0.986 @xmath5 0.009 + ngc 6254 & 14.80 & 0.269 @xmath5 0.015 & 0.113 & 1.290 @xmath5 0.004 & 1.022 @xmath5 0.016 + ngc 6304 & 15.95 & 1.556 @xmath5 0.002 & 0.020 & 1.662 @xmath5 0.004 & 0.105 @xmath5 0.004 + ngc 6341 & 15.05 & 0.036 @xmath5 0.013 & 0.093 & 0.946 @xmath5 0.004 & 0.910 @xmath5 0.013 + ngc 6352 & 14.95 & 1.249 @xmath5 0.004 & 0.025 & 1.372 @xmath5 0.013 & 0.123 @xmath5 0.013 + ngc 6362 & 15.18 & 0.817 @xmath5 0.009 & 0.193 & 1.064 @xmath5 0.007 &
0.247 @xmath5 0.012 + ngc 6366 & 15.25 & 1.817 @xmath5 0.010 & 0.030 & 1.957 @xmath5 0.011 & 0.141 @xmath5 0.014 + ngc 6388 & 17.00 & 1.344 @xmath5 0.002 & 0.172 & 1.470 @xmath5 0.002 & 0.126 @xmath5 0.003 + ngc 6397 & 12.90 & 0.217 @xmath5 0.007 & 0.066 & 1.162 @xmath5 0.010 & 0.944 @xmath5 0.012 + ngc 6441 & 17.65 & 1.501 @xmath5 0.002 & 0.154 & 1.611 @xmath5 0.002 & 0.110 @xmath5 0.003 + ngc 6496 & 16.20 & 1.224 @xmath5 0.004 & 0.023 & 1.331 @xmath5 0.007 &
0.107 @xmath5 0.008 + ngc 6535 & 15.65 & 0.554 @xmath5 0.041 & 0.127 & 1.457 @xmath5 0.011 & 0.903 @xmath5 0.042 + ngc 6541 & 15.15 & 0.112 @xmath5 0.010 & 0.107 & 1.076 @xmath5 0.004 & 0.964 @xmath5 0.011 + ngc 6584 & 16.40 & 0.640 @xmath5 0.061 & 0.215 & 1.048 @xmath5 0.003 & 0.408 @xmath5 0.062 + ngc 6624 & 15.85 & 1.248 @xmath5 0.003 & 0.025 & 1.383 @xmath5
0.012 & 0.135 @xmath5 0.012 + ngc 6637 & 15.75 & 1.092 @xmath5 0.002 & 0.020 & 1.230 @xmath5 0.002 & 0.138 @xmath5 0.003 + ngc 6652 & 15.77 & 1.030 @xmath5 0.004 & 0.021 & 1.171 @xmath5 0.004 & 0.141 @xmath5 0.006 + ngc 6656 & 14.15 & 0.438 @xmath5 0.011 & 0.141 & 1.373 @xmath5 0.004 & 0.935 @xmath5 0.012 + ngc 6681 & 15.63 & 0.074 @xmath5 0.010 & 0.112 & 1.060 @xmath5 0.002 & 0.986 @xmath5 0.010 + ngc 6717 & 15.53 & 0.268 @xmath5 0.021 & 0.128 & 1.216 @xmath5 0.003 & 0.948 @xmath5 0.021 + ngc 6723 & 15.30 & 0.704 @xmath5 0.039 & 0.301 & 1.075 @xmath5 0.003 & 0.371 @xmath5 0.039 + ngc 6752 & 13.70 & -0.058 @xmath5 0.009 & 0.102 & 1.024 @xmath5 0.003 & 1.082 @xmath5 0.010 + ngc
6779 & 16.15 & 0.305 @xmath5 0.007 & 0.110 & 1.220 @xmath5 0.003 & 0.914 @xmath5 0.008 + ngc 6809 & 14.35 & 0.154 @xmath5 0.021 & 0.097 & 1.060 @xmath5 0.004 & 0.906 @xmath5 0.021 + ngc 6838 & 14.21 & 1.185 @xmath5 0.009 & 0.021 & 1.291 @xmath5 0.005 & 0.106 @xmath5 0.010 + ngc 6934 & 16.78 & 0.355 @xmath5 0.046 & 0.251 & 1.066 @xmath5 0.005 & 0.711 @xmath5 0.046 + ngc 6981 & 16.73 & 0.537 @xmath5 0.122 & 0.247 & 1.008 @xmath5 0.007 & 0.471 @xmath5 0.122 + ngc 7078 & 15.75 & 0.131 @xmath5 0.028 & 0.196 & 1.007 @xmath5 0.002 & 0.876 @xmath5 0.028 + ngc 7089 & 15.95 & 0.043 @xmath5 0.009 & 0.148 & 0.993 @xmath5 0.002 & 0.951 @xmath5 0.009 + ngc 7099 & 15.12 & 0.087 @xmath5 0.005 & 0.077 & 0.960 @xmath5 0.004 & 0.872 @xmath5 0.006 + pal 12 & 16.90 & 0.908 @xmath5 0.005 & 0.008 & 1.019 @xmath5 0.016 & 0.111 @xmath5 0.017 + terzan 7 & 17.67 & 1.028 @xmath5 0.004 & 0.012 & 1.115 @xmath5 0.005 & 0.087 @xmath5 0.007 + terzan 8 & 17.90 & 0.230 @xmath5 0.011 & 0.111 & 1.095 @xmath5 0.011 & 0.865 @xmath5
0.015 + am-1 & 20.92 & 0.799 @xmath5 0.008 & 0.021 & 0.991 @xmath5 0.005 & 0.192 @xmath5
0.009 + eridanus & 20.23 & 0.829 @xmath5 0.015 & 0.017 & 1.009 @xmath5 0.033 & 0.180 @xmath5 0.036 + ngc 2419 & 20.35 & 0.198 @xmath5 0.008 & 0.155 & 1.017 @xmath5 0.002 & 0.819 @xmath5 0.008 + pal 3 & 20.40 & 0.688 @xmath5 0.021 & 0.087 & 1.002 @xmath5 0.019 & 0.314 @xmath5 0.028 + pal 4 & 20.65 & 0.816 @xmath5 0.008 & 0.018 & 0.999 @xmath5 0.006 & 0.183 @xmath5 0.010 + pal 14 & 20.00 & 0.828 @xmath5 0.021 & 0.035 & 1.056 @xmath5 0.027 & 0.228 @xmath5 0.034 + lrrrrr arp 2 & -1.80 & 0.2 & 17.55 & 0.113 & 13.00 @xmath5 0.75 + lyng7 & -0.60 & 0.2 & 15.80 & 0.713 & 12.50 @xmath5 1.00 + ngc 104 & -0.70 & 0.2 & 13.30 & 0.023 & 12.75 @xmath5 0.50 + ngc 288 & -1.40 & 0.4 & 14.85 & 0.013 & 12.50 @xmath5 0.50 + ngc 362 & -1.30 & 0.2 & 14.80 & 0.023 & 11.50 @xmath5 0.50 + ngc 1261 & -1.35 & 0.2 & 16.10 & 0.013 & 11.50 @xmath5 0.50 + ngc 2298 * & -1.90 & 0.2 & 15.43 & 0.237 & 13.00 @xmath5 1.00 + ngc 3201 * & -1.50 & 0.2 & 13.90 & 0.268 & 12.00 @xmath5 0.75 + ngc 4147 & -1.70 & 0.2 & 16.48 & 0.018 & 12.75 @xmath5 0.75 + ngc 4590 & -2.30 & 0.2 & 15.30 & 0.056 & 13.00 @xmath5 1.00 + ngc 4833 * & -2.30 & 0.4 & 14.84 & 0.353 & 13.00 @xmath5 1.25 + ngc 5024 & -2.00 & 0.2 & 16.43 & 0.023 & 13.25 @xmath5 0.50 + ngc 5053 & -2.40 & 0.2 & 16.32 & 0.021 & 13.50 @xmath5 0.75 + ngc 5272 & -1.60 & 0.2 & 15.08 & 0.018 & 12.50 @xmath5 0.50 + ngc 5286 * & -1.70 & 0.0 & 15.75 & 0.263 & 13.00 @xmath5 1.00 + ngc 5466 & -2.10 & 0.2 & 16.12 & 0.023 & 13.00 @xmath5 0.75 + ngc 5904 & -1.30 & 0.2 & 14.38 & 0.033 & 12.25 @xmath5 0.75 + ngc 5927 * & -0.50 & 0.2 & 15.30 & 0.393 & 12.25 @xmath5 0.75 + ngc 5986 * & -1.60 & 0.2 & 15.73 & 0.295 & 13.25 @xmath5 1.00 + ngc 6093 * & -1.70 & 0.2 & 15.55 & 0.213 & 13.50 @xmath5 1.00 + ngc 6101 & -1.80 & 0.2 & 16.03 & 0.113 & 13.00 @xmath5 1.00 + ngc 6121 * & -1.20 & 0.4 & 12.20 & 0.423 & 12.50 @xmath5 0.50 + ngc 6144 * & -1.80 & 0.2 & 15.40 & 0.448 & 13.50 @xmath5 1.00 + ngc 6171 * & -1.00 & 0.2 & 14.45 & 0.418 & 12.75 @xmath5 0.75 + ngc 6205 & -1.60 & 0.2 & 14.47 & 0.019 & 13.00 @xmath5 0.50 + ngc 6218 & -1.30 & 0.2 & 13.90 & 0.191 & 13.25 @xmath5 0.75 + ngc 6254 * & -1.55 & 0.2 & 14.15 & 0.261 & 13.00 @xmath5 1.25 + ngc 6304 * & -0.50 & 0.2 & 14.87 & 0.473 & 12.75 @xmath5 0.75 + ngc 6341 & -2.40 & 0.2 & 14.80 & 0.031 & 13.25 @xmath5 1.00 + ngc 6352 * & -0.80 & 0.4 & 14.10 & 0.253 & 13.00 @xmath5 0.50 + ngc 6362 & -1.10 & 0.4 & 14.55 & 0.070 & 12.50 @xmath5 0.50 + ngc 6366 & -0.70 & 0.2 & 14.00 & 0.718 & 12.00 @xmath5 0.75 + ngc 6397 & -2.10 & 0.2 & 12.40 & 0.183 & 13.50 @xmath5 0.50 + ngc 6496 & -0.50 & 0.2 & 15.35 & 0.213 & 12.00 @xmath5 0.75 + ngc 6535 & -1.90 & 0.2 & 14.85 & 0.443 & 13.25 @xmath5 1.00 + ngc 6541 * & -1.90 & 0.4 & 14.68 & 0.118 & 13.25 @xmath5 1.00 + ngc 6584 & -1.40 & 0.2 & 15.85 & 0.078 & 12.25 @xmath5 0.75 + ngc 6624 * & -0.50 & 0.0 & 15.05 & 0.253 & 13.00 @xmath5 0.75 + ngc 6637 * & -0.70 & 0.2 & 15.05 & 0.163 & 12.50 @xmath5 0.75 + ngc 6652 & -0.75 & 0.2 & 15.05 & 0.113 & 13.25 @xmath5 0.50 + ngc 6681 & -1.50 & 0.2 & 15.05 & 0.098 & 13.00 @xmath5 0.75 + ngc 6717 * & -1.10 & 0.2 & 14.78 & 0.203 & 13.00 @xmath5 0.75 + ngc 6723 & -1.00 & 0.2 & 14.67 & 0.073 & 12.75 @xmath5 0.50 + ngc 6752 & -1.50 & 0.2 & 13.26 & 0.053 & 12.50 @xmath5 0.75 + ngc 6779 * & -2.20 & 0.2 & 15.65 & 0.248 & 13.50 @xmath5 1.00 + ngc 6809 & -1.80 & 0.2 & 13.88 & 0.113 & 13.50 @xmath5 1.00 + ngc 6838 * & -0.70 & 0.2 & 13.40 & 0.223 & 12.50 @xmath5 0.75 + ngc 6934 & -1.55 & 0.2 & 16.23 & 0.108 & 12.00 @xmath5 0.75 + ngc 6981 & -1.50 & 0.2 & 16.20 & 0.048 & 12.75 @xmath5 0.75 + ngc 7078 & -2.40 & 0.2 & 15.40 & 0.083 & 13.25 @xmath5 1.00 + ngc 7089 & -1.60 & 0.2 & 15.48 & 0.048 & 12.50 @xmath5 0.75 + ngc 7099 & -2.40 & 0.2 & 14.82 & 0.053 & 13.25 @xmath5 1.00 + pal 12 & -0.80 & 0.0 & 16.40 & 0.033 & 9.50 @xmath5 0.75 + terzan 7 & -0.60 & 0.0 & 17.15 & 0.073 & 8.00 @xmath5 0.75 + terzan 8 & -2.40 & 0.4 & 17.50 & 0.133 & 13.50 @xmath5 0.50 + am-1 & -1.50 & 0.2 & 20.41 & 0.020 & 11.10 @xmath5 0.50 + eridanus & -1.30 & 0.2 & 19.78 & 0.060 & 10.50 @xmath5 0.50 + ngc 2419 & -2.00 & 0.2 & 20.05 & 0.062 & 13.00 @xmath5 1.00 + pal 3 & -1.50 & 0.2 & 19.89 & 0.060 & 11.30 @xmath5 0.50 + pal 4 & -1.30 & 0.2 & 20.14 & 0.055 & 10.90 @xmath5 0.50 + pal 14 & -1.50 & 0.2 & 19.51 & 0.045 & 10.50 @xmath5 0.50 + ccc @xmath88&@xmath89 & @xmath90 + @xmath91&@xmath92 & @xmath92 + @xmath93&@xmath94 & @xmath95 + @xmath96&@xmath97 & @xmath98 + @xmath99&@xmath100 & @xmath101 + @xmath102&@xmath103 & @xmath104 + @xmath105&@xmath106 & @xmath107 + | the horizontal branch ( hb ) morphology of globular clusters ( gcs ) is most strongly influenced by metallicity .
the second parameter phenomenon , first described in the 1960 s , acknowledges that metallicity alone is not enough to describe the hb morphology of all gcs .
in particular , astronomers noticed that the outer galactic halo contains gcs with redder hbs at a given metallicity than are found inside the solar circle .
thus , at least a second parameter was required to characterize hb morphology .
while the term ` second parameter ' has since come to be used in a broader context , its identity with respect to the original problem has not been conclusively determined .
here we analyze the median color difference between the hb and the red giant branch ( rgb ) , hereafter denoted @xmath0 , measured from hubble space telescope ( hst ) advanced camera for surveys ( acs ) photometry of 60 gcs within @xmath120 kpc of the galactic center .
analysis of this homogeneous data set reveals that , after the influence of metallicity has been removed from the data , the correlation between @xmath0 and age is stronger than that of any other parameter considered . expanding the sample to include hst acs and wide field planetary camera 2 ( wfpc2 ) photometry of the 6 most distant galactic gcs lends additional support to the correlation between @xmath0 and age .
this result is robust with respect to the adopted metallicity scale and the method of age determination , but must bear the caveat that high quality , detailed abundance information is not available for a significant fraction of the sample .
furthermore , when a subset of gcs with similar metallicities and ages are considered , a correlation between @xmath0 and central luminosity density is exposed . with respect to the existence of gcs with anomalously red hbs at a given metallicity
, we conclude that age is the second parameter and central density is most likely the third .
important problems related to hb morphology in gcs , notably multi - modal distributions and faint blue tails , remain to be explained . |
cool white dwarfs are typically identified through the combination of proper motion and apparent magnitude called reduced proper motion @xcite . at a given brightness and color
, white dwarfs will have larger proper motions than main sequence stars owing to closer distances from the earth implied by their small radii .
metal - poor stars , or subdwarfs , also stand out well in reduced proper motion diagrams due to their lower luminosities ( for the same color class ) and typically higher velocities .
subluminous main sequence stars which distinguish themselves in reduced proper motion are kinematic members of the thick disk and halo @xcite .
subdwarfs have long been a contaminant in surveys aiming to find cool white dwarfs because both stellar types are mixed in a reduced proper motion diagram at intermediate colors .
cool white dwarfs must be confirmed spectroscopically @xcite .
current complete and ongoing large - scale proper motion surveys are discovering numerous exotic objects including extreme subdwarfs , brown dwarfs , and cool white dwarfs ( see @xcite and references therein ) .
most if not all of these searches are utilizing available large - scale optical photographic sky survey catalogs .
this is potentially problematic for the identification of cool white dwarfs because optical photographic bandpasses do not yield broad color baselines , and photographic photometry has intrinsically large errors .
these facts lead to a large overlap in the white dwarf and subdwarf sequences in reduced proper motion diagrams @xcite .
yet even with spectroscopy , subdwarfs and other weak - lined objects can appear featureless in low resolution and/or low signal - to - noise ( s / n ) measurements potentially mistaken as dc white dwarfs @xcite .
identification of degenerates with @xmath1 k should include good s / n spectroscopy covering blue - green optical wavelengths @xcite . contrary to the claim of @xcite , this paper presents evidence that the high proper motion star ssspm j1549@xmath03544 ( @xmath2 @xmath3 @xmath4 , @xmath5 @xmath6 @xmath7 , j2000 ) is neither the nearest cool white dwarf nor a degenerate star , but rather a very high velocity halo star passing through the solar neighborhood .
optical @xmath8 photometric data were obtained 14 february 2005 with the orthogonal parallel transfer imaging camera ( optic ; @xcite ) on the university of hawaii 2.2 meter telescope at mauna kea . observing conditions were photometric with seeing @xmath9 and some gusty winds @xmath10 mph .
three identical exposures of ssspm j1549@xmath03544 were taken at each bandpass for a total integration time of @xmath11 seconds , yielding s / n @xmath12 at all wavelengths in a @xmath13 aperture radius .
a standard landolt field @xcite was observed similarly immediately following the science target .
the images were flat fielded , cleaned of bad pixels in the region of interest , and combined into a single frame at each wavelength for photometry .
calibration was performed with three stars and aperture photometry was executed using standard iraf packages .
calibrator stars were measured in a @xmath14 aperture radius and the science target data were corrected to this standard aperture and adjusted for relative extinction .
the results are listed in table [ tbl-1 ] .
optical @xmath15 spectroscopic data were obtained 18 march 2005 with the double beam spectrograph ( dbs ; @xcite ) on the australian national university 2.3 m telescope at siding spring observatory .
only the red arm of the spectrograph was operated , with no dichroic in the beam .
the 158 l / mm grating used gave a spectral resolution of 10 ( 2.5 pixels ) with the @xmath16 slit .
the seeing was @xmath17 and the sky was clear .
ssspm j1549@xmath03544 and the flux calibrator ltt 6248 were each observed for 600 seconds , yielding s / n @xmath18 over the majority of the science target wavelength shown in figure [ fig1 ] .
the spectral images were flat fielded , cleaned of bad pixels in the region of interest , and flux calibrated spectra were extracted using standard iraf packages .
the spectra were trimmed to show the region of interest , while no attempt was made to remove telluric features .
the combined optical and near - infrared colors of ssspm j1549@xmath03544 are consistent with a main sequence star of @xmath19 k , late k spectral type @xcite .
the near - infrared magnitudes are too bright for known cool and ultracool white dwarfs @xcite .
for example , the reddest known degenerates have colors which peak near @xmath20 , after which they become bluer again as shown by observation and models ( excepting pure helium models at @xmath21 k ; @xcite ) , while ssspm j1549@xmath03544 has @xmath22 .
the spectrum in figure [ fig1 ] confirms the status of ssspm j1549@xmath03544 as a nondegenerate star , and the deep mgh feature near 5200 indicates it is metal - poor @xcite .
other unlabelled spectral features present include : weak h@xmath23 , several fe and ca lines , plus some weak cah and tio .
these latter features are difficult to distinguish as they are located , for the most part , within and around the prominent telluric bands @xcite which are also seen in the calibrator star spectrum . following the methodology of @xcite ,
a spectral type and subdwarf class of ssspm j1549@xmath03544 was estimated by measuring the tio5 , cah1 , cah2 , & cah3 spectroscopic indices , which are listed in table [ tbl-2 ] . using equations @xmath24 in @xcite , a spectral type of k5
is found consistently across all three relations , and type sdk5 is found using equation 7 in the same work . however , given the fact that the spectral resolution achieved here is @xmath25 times lower than that of @xcite , the spectroscopic measurments may be unreliable but are presented here as a rough guide .
there are no distinctions between sdk and esdk stars before spectral type k7 , and , combined with the uncertain spectroscopy , it is difficult to say whether or not ssspm j1549@xmath03544 is an extreme subdwarf without a direct metallicity measurment . the mgh band near 5200 is quite strong and an extreme subdwarf classification might be appropriate @xcite . in order to assess the kinematics of ssspm j1549@xmath03544
, a distance must be estimated . conservatively assuming this star lies two magnitudes below the main sequence at spectral type k5
, it would have @xmath26 mag @xcite .
this would place the star at @xmath27 pc with a tangential speed of 430 km s@xmath28 based on the astrometic data in @xcite .
furthermore , fourteen spectral lines ( many without high s / n ) were used to measure the radial velocity , the crude result being @xmath29 km s@xmath28 , corrected for the motion of the earth along the line of sight on the date of observation . combining all this data , a galactic @xmath30 space motion was calculated , corrected for the solar motion , @xmath31 km s@xmath28 @xcite , relative to the local standard of rest ( @xmath32 positive toward the galactic anticenter , @xmath33 positive in the direction of galactic rotation , @xmath34 positive toward the north galactic pole ) .
the result is @xmath35 km s@xmath28 , making ssspm j1549@xmath03544 a halo star @xcite .
if we instead assume @xmath36 due to the potential unreliability of the measured radial velocity , the result becomes @xmath37 km s@xmath28 , which is still consistent with halo membership . despite a fairly conservative assumption ,
the 114 pc distance estimate and space motions should be considered preliminary as subdwarfs can span a wide range of absolute magnitudes .
if ssspm j1549@xmath03544 is closer to 150 pc , or around 1.5 magnitudes below the main sequence , it would have a total heliocentric velocity in the range @xmath38 km s@xmath28 . at one magnitude below the main sequence , or 180 pc , ssspm
j1549@xmath03544 would have a total galactocentric velocity of @xmath39 km s@xmath28 and would be among the fastest moving stars ever seen @xcite .
the star ssspm j1549@xmath03544 is shown to be a metal - poor sdk star rather than a cool white dwarf as previously claimed @xcite .
it is possible this star is an extreme k subdwarf , but in any case it is a very high velocity halo star passing through the solar neighborhood . with many new proper motion objects being discovered and studied ,
it is critical to correctly distinguish cool white dwarfs from subdwarf contaminants .
there is great science potential in the oldest degenerates .
some data used in this paper are part of the two micron all sky survey , 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 ( nasa ) and the national science foundation .
j. farihi acknowledges support by grants from nasa .
prw acknowledges the support provided by a grant from the australian research council . | spectroscopy and photometry demonstrate that ssspm j1549@xmath03544 is not a cool white dwarf but a high velocity metal - poor halo star . |
nonrotating black holes with both mass and charge have two horizons : the outer horizon is the actual event horizon of the black hole , while the inner horizon is unstable to small perturbations , and is conjectured to become singular in any realistic situation.@xcite for such black holes the charge is always less than or equal to the mass , with the special case of charge equals mass called an extreme black hole . in the limit as the charge approaches the mass , the two horizons merge .
thus , as pointed out by marolf@xcite one might expect that in some sense extreme black holes have singular event horizons .
more precisely , it is claimed in @xcite that an observer crossing the event horizon encounters a singularity a short time later and that this short time goes to zero in the limit of extremality .
the treatment of @xcite used an eternal black hole , whereas physical black holes are formed by gravitational collapse . in this paper
we treat the problem of nearly extreme black holes formed by gravitational collapse . for simplicity ,
we use collapse of a charged thin shell .
we then consider a free fall observer who falls into such a black hole at late times and calculate the time it takes for him to reach the singularity .
methods and results are presented in section 2 and conclusions are presented in section 3 .
a charged , nonrotating , black hole is described by the reissner - nordstrom metric @xmath0 where the quantity @xmath1 is given by @xmath2 where @xmath3 and @xmath4 are respectively the mass and charge of the black hole .
the quantity @xmath1 vanishes at the outer and inner horizons of the black hole , which are located respectively at @xmath5 and @xmath6 where @xmath7 the coordinate system in eqn.([rn ] ) becomes singular at the horizons , thus to describe the process of gravitational collapse and black hole formation , we need a different coordinate system .
we therefore use eddington - finkelstein coordinates ( @xmath8 ) where @xmath9 is given by @xmath10 in these coordinates the reissner - nordstrom metric becomes @xmath11 the coordinate @xmath9 is constant along ingoing light rays .
one can also define an analogous coordinate @xmath12 which is constant along outgoing light rays . with a series of coordinate patches of this type
, one can cover the extended reissner - nordstrom spacetime , which is shown in fig .
( [ fig1 ] ) . however , the extended reissner - nordstrom spacetime is too large to represent what actually happens in gravitational collapse .
this is for two different reasons : first when a charged , spherical object collapses , it is only the spacetime outside the object that is described by the reissner - nordstrom metric .
thus , to describe gravitational collapse , one must find the path of the outer surface of the collapsing object , and then cut out all parts of the diagram of fig .
( [ fig1 ] ) that are outside that surface and replace them with something appropriate for the interior of the collapsing object .
the second reason is that small perturbations of the reissner - nordstrom solution blow up as @xmath13 ( which is also a null surface at @xmath14 ) .
therefore , to get the appropriate diagram , one must replace this null surface with a null singularity and cut out all portions of the diagram to the future of this surface . in the end one
obtains a restricted portion of the extended reissner - nordstrom spacetime , which can be covered by a single coordinate patch of the sort used in eqn .
( [ rnef ] ) . the extended reissner - nordstrom spacetime . here
the outer and inner horizons are represented by solid lines at 45 degree angles to the vertical , while the timelike singularity is denoted by a dashed line ] we want a simple model for gravitational collapse to form the black hole , so we choose a thin shell of charged dust .
thin shells in general relativity are described by the israel formalism@xcite in which the history of the shell is a timelike hypersurface @xmath15 separating two spacetimes ( labeled @xmath16 and @xmath17 and which are respectively the exterior and the interior of the shell ) .
the intrinsic metric @xmath18 is continuous , while the extrinsic curvature @xmath19 is discontinuous and satisfies @xmath20 = - 8 \pi \left ( { t_{ab } } - { \textstyle { \frac 1 2 } } t { h_{ab } } \right ) \label{israelshell}\ ] ] where @xmath21 is the delta function stress - energy of the shell . here ,
for any quantity @xmath22 we define @xmath23= { \phi _ + } -{\phi _ -}$ ] where the quantity @xmath24 is @xmath22 on @xmath15 evaluated from the @xmath16 side ( and correspondingly for @xmath25 ) .
the worldsheet of the shell is given by @xmath26 where @xmath27 is the proper time of an observer in the shell .
the stress - energy of the dust shell takes the form @xmath28 where @xmath29 is the four - velocity of the shell and @xmath30 is the shell s mass per unit area . as shown by de la cruz and israel@xcite , kuchar@xcite and chase@xcite eqn.([israelshell ] ) can be used to find the equation of motion of the shell as follows : the exterior of the shell is a reissner - nordstrom spacetime .
spherical symmetry implies that the extrinsic curvature takes the form @xmath31 combining eqns .
( [ israelshell ] ) , ( [ shellstress ] ) and ( [ kdecomp ] ) yields @xmath32 = - 4 \pi \sigma \label{bjump}\ ] ] @xmath33 = - 4 \pi \sigma \label{cjump}\ ] ] a straightforward calculation using eqn.([rnef ] ) shows that the quantities @xmath34 and @xmath35 are given by @xmath36 where an overdot denotes derivative with respect to @xmath37 .
since the interior of the shell is minkowski spacetime , expressions for @xmath34 and @xmath35 for the shell interior are given by eqns ( [ beqn ] ) and ( [ ceqn ] ) with @xmath1 replaced by @xmath38 .
applying eqn .
( [ ceqn ] ) to eqn .
( [ cjump ] ) yields @xmath39 then using eqns .
( [ bjump ] ) , ( [ beqn ] ) , and ( [ motion1 ] ) yields @xmath40 which means that there is a constant @xmath41 such that @xmath42 using this result in eqn .
( [ motion1 ] ) and solving for @xmath43 yields @xmath44 path of the thin shell in the reissner - nordstrom spacetime . here
nonsingular horizons are denoted by solid lines , and the null singularity by a dashed line ] we start the shell at a large radius with a speed @xmath45 in the inward direction .
then it follows from eqn .
( [ motion2 ] ) that @xmath46 we also introduce the quantity @xmath47 defined by @xmath48 since we are considering nearly extremal black holes , @xmath49 .
expressing eqn .
( [ motion2 ] ) in terms of @xmath50 and @xmath45 we obtain @xmath51 \right ) } ^2 } \label{motion3}\ ] ] we have @xmath52 , so @xmath53 is just slightly less than @xmath54 and therefore the shell will cross @xmath53 a short ( proper ) time after crossing @xmath54 .
it is also true that the perturbed black hole becomes singular at @xmath53 .
however , it is here that the use of eddington - finkelstein coordinates becomes essential . in the collapse process , there are two different null surfaces at @xmath14 , an outgoing null surface where @xmath9 is finite , and which is not singular , and an ingoing null surface which occurs in the limit as @xmath13 , and which is singular
. the question then becomes , which of these surfaces does the shell cross first ? to answer this question , we note that there is a function @xmath55 such that on the worldsheet of the shell @xmath56 . then from eqn .
( [ rnef ] ) it follows that @xmath57 and therefore that @xmath58 however , the null coordinate @xmath9 always increases towards the future , fixing the sign above to be plus , so we obtain @xmath59 note that in this equation , if @xmath60 when @xmath1 vanishes , then @xmath61 remains finite ; but if @xmath62 when @xmath1 vanishes then @xmath63 .
since the shell starts out ingoing , and since it can not change direction between the two horizons ( where @xmath64 is a timelike coordinate ) then it follows that the shell first crosses @xmath53 at a place where @xmath9 remains finite , and where therefore the spacetime remains nonsingular . after crossing @xmath65 ,
the shell reaches a minimum radius @xmath66 and then re - expands .
it follows from eqn .
( [ motion3 ] ) that @xmath67 where we have neglected any terms higher than zeroth order in @xmath47 . after reaching @xmath66 the shell re - expands , and eventually reaches @xmath53 .
note , however , that this time the shell approaches @xmath53 with @xmath68 .
it then follows from eqn .
( [ dotv ] ) that @xmath9 diverges as @xmath69 and thus that the shell approaches the null singularity .
the path of the shell in the usual penrose diagram of the reissner - nordstrom spacetime is shown in fig .
( [ fig2 ] ) .
the shell clearly takes a non - neglible amount of proper time to encounter the singularity , and this will also be true of an observer who falls into the black hole shortly after it forms .
however , as we will see , things are somewhat different for an observer who falls in at late times .
since for such an observer , @xmath70 when the outer horizon is crossed , and since @xmath9 increases along future directed timelike curves , it follows that @xmath70 when the observer encounters the shell .
therefore , we are led to consider the behavior of the shell at large @xmath9 . we know that as @xmath13 we have @xmath69 and therefore @xmath71
however , we will need to know how rapidly @xmath71 at large @xmath9 . from eqn .
( [ dotv ] ) it follows that on the shell we have @xmath72 then using eqns .
( [ f ] ) and ( [ dvdr ] ) and neglecting any terms that are higher order in the small quantities @xmath1 and @xmath47 we find that on the shell at large @xmath9 we have @xmath73 the solution of this equation is @xmath74 where @xmath75 is a constant . note that there are two separate regimes where this equation simplifies .
for @xmath76 we have @xmath77 while for @xmath78 we have @xmath79 $ ] .
we now consider the fate of a free - fall observer who falls into the black hole at late times .
just as the shell has four - velocity @xmath80 , denote the four - velocity of the observer by @xmath81 . due to the time translation symmetry of the reissner - nordstrom spacetime
, the observer has a conserved energy per unit mass @xmath82 given by @xmath83 since @xmath84 is a unit timelike vector , the same reasoning that led to eqn .
( [ dotv ] ) yields @xmath85 using eqns .
( [ energy ] ) and ( [ dotv2 ] ) we find @xmath86 at late times , the shell is approximately at @xmath53 and for the region in between @xmath54 and @xmath53 we have @xmath87 .
thus the observer crosses the shell approximately a proper time @xmath88 after crossing the outer horizon . to find the amount of proper time that it takes for the observer to encounter the singularity , we must also find the behavior of the observer in the interior of the shell . to begin with ,
we calculate the quantity @xmath89 when the observer crosses the shell . from eqn .
( [ rn ] ) we find @xmath90 then using eqns . ( [ dotv ] ) , ( [ dotv2 ] ) and ( [ vdotu ] ) we obtain @xmath91 since when the observer crosses the shell we have @xmath68 and @xmath92 it follows that to lowest order in @xmath47 @xmath93 and then using eqn .
( [ motion3 ] ) we obtain @xmath94 path of the geodesic observer ( thin curve ) in the spacetime formed by the collapse of the thin shell . here
nonsingular horizons are denoted by solid lines , and the null singularity by a dashed line ] since @xmath1 is very small when the observer crosses the shell , it follows that @xmath89 is very large there .
however , since @xmath84 satisfies the geodesic equation , and since the christoffel symbols in the thin shell spacetimes have only jumps rather than delta functions , it follows that @xmath84 is continuous across the shell .
since the thin shell formalism also requires that the metric and @xmath29 are continuous across the shell , it follows that immediately upon entering the interior of the shell @xmath89 is still given by the expression in eqn .
( [ vdotu4 ] ) .
but the interior of the shell is minkowski spacetime , and the shell has @xmath95 so it follows from eqn .
( [ vdotu4 ] ) that in the rest frame of the center of the shell , the observer is an ingoing geodesic with @xmath96 this is a very large @xmath97 factor , so even though the observer must travel a distance of @xmath98 ( to the center and back ) to encounter the singularity , this only takes a proper time of @xmath99 since @xmath1 is given by eqn .
( [ flargev ] ) it follows that this is a very small proper time . the motion of the observer is shown in fig .
( [ fig3 ] )
though the pictures drawn are somewhat different , our calculations agree with the main conclusion of @xcite .
for a nearly extremal black hole formed by gravitational collapse an observer crossing the event horizon at late times almost immediately encounters a curvature singularity .
note however , that this statement involves two different small numbers : @xmath47 which expresses how close to extremality the black hole is , and @xmath100 which measures how late the observer crosses the black hole event horizon .
( here @xmath101 is the value of @xmath9 at which the observer crosses the horizon ) . in all cases , the proper time from event horizon
to shell is of order @xmath102 .
for @xmath103 the proper time from shell to singularity is of order @xmath104 , while for @xmath105 the proper time from shell to singularity is completely negligible .
we now consider the effect the nature of the singularity could have on this result . throughout
, we have assumed that the singularity is null .
however , it is generally thought that when black holes form the singularity starts out null near the horizon and then becomes spacelike further inside the black hole .
however , any such change would only serve to decrease the amount of time it would take for an observer to encounter the singularity , so the conclusion that the observer encounters the singularity almost immediately after crossing the event horizon remains true .
finally note that the collapsing thin shell is only one simple model for the formation of a charged black hole by gravitational collapse .
it would be interesting to consider other models ( _ e.g. _ the collapse of a charged scalar field ) to see whether the features found in the shell model continue to hold .
i would like to thank ratin akhoury , ryo saotome , eric poisson , gary horowitz , and don marolf for useful discussions .
this work was supported by nsf grant phy-0855532 to oakland university .
poisson e and israel w 1990 _ phys .
rev . _ * d41 * 1796 | we examine the properties of nearly extremal black holes produced by gravitational collapse .
it is shown that an observer who crosses the black hole horizon at late times rapidly encounters a singularity . |
we are very grateful to rob van der voo , who kindly provided us with invaluable pre - cambrian apwps .
we thank paul renne , yudong he , and buford price for useful discussions , suggestions , and support . v. trimble , ann .
. astrophys . * 25 * , 425 ( 1987 ) e.w . kolb and m.s .
turner , _ the early universe _ ( addison - wesley , reading , 1988 ) ; p.j.e .
peebles , _ principles of physical cosmology _ ( princeton university press , princeton , 1993 ) .
j. r. primack , d. seckel , and b. sadoulet , ann .
nucl . part .
sci . * 38 * , 751 ( 1988 ) ; p.
f. smith and j. d. lewin , phys .
187 * , 204 ( 1990 ) .
see for instance , m. white , d. scott , j. silk , and m. davis , mon . not .
r. astron .
soc . * 276 * , l69 ( 1995 ) . for a recent review
see , e. diehl , g.l .
kane , c. kolda , and j.d .
wells , phys .
d * 52 * 4223 ( 1995 ) .
d. p. snowden - ifft , e. s. freeman and p. b. price , phys .
lett * 74 * , 4133 ( 1995 ) .
d. p. snowden - ifft and m. k. y. chan , nucl .
inst . and meth .
b * 101 * , 247 ( 1995 ) .
d. n. spergel , phys .
d * 37 * , 1353 ( 1988 ) .
ziegler , j.p .
biersack , and u. littmark , _ the stopping and range of ions in solids _ ( pergamon press , oxford , 1985 ) .
s. p. ahlen et al . , phys .
b * 195 * , 603 ( 1987 ) .
p. m. mathews and i. i. shapiro , ann .
earth planet .
* 20 * , 469 ( 1992 ) .
j. laskar and p. robutel , nature * 361 * , 608 ( 1993 ) j. laskar , f. joutel , and p. robutel , nature * 361 * , 615 ( 1993 ) .
d. h. tarling , _ paleomagnetism : principles and applications in geology , geophysics and archaeolog _
( chapman and hall , london , 1983 ) . m. w. mcelhinny , _ paleomagnetism and plate tectonics _ 188 ( cambridge press , cambridge , 1973 ) . rob van der voo , _ paleomagnetism of the atlantic , tethys and iapetus oceans _
( cambridge university press , cambridge , 1993 ) . manoel s. dagrella - filho _ et al .
_ , earth and planet .
* 101 * , 332 ( 1990 ) .
a. blaauw _
et al . _ , mon . not .
. soc . * 121*(2 ) , 10 ( 1960 ) .
f. j. kerr and d. lynden - bell , mon . not .
221 * , 1023 ( 1986 ) .
m. j. reid , ann .
. astrophys . * 31 * , 345 ( 1993 ) . c. klein and c.s .
hurlbut jr . , _ manual of mineralogy , 20th ed . _
485 - 486 ( john wiley and sons , new york , 1985 ) . unpublished data . .
calculated values of track accumulation rate @xmath67 , angular coefficient of track accumulation rate @xmath16 , and signal contrast @xmath19 for several values of wimp mass .
the error bars arise from the monte carlo technique used to calculate these points . [ cols="^,^,^,^ " , ] | mica can store ( for @xmath0 gy ) etchable tracks caused by atoms recoiling from wimps . because a background from fission neutrons will eventually limit this technique , a unique signature for wimps in ancient mica
is needed .
our motion around the center of the galaxy causes wimps , unlike neutrons , to enter the mica from a preferred direction on the sky .
mica is a directional detector and despite the complex rotations that natural mica crystals make with respect to this wimp `` wind , '' there is a substantial dependence of etch pit density on present day mica orientation .
# 1[#1 ] # 1\{*#1 * } decades of research have established that at least 90% of the mass of our universe is not emitting any light @xcite . in recent years
the nature of this dark matter has come under considerable experimental and theoretical scrutiny . as a result of this scrutiny
the list of potential candidates has dwindled from dozens a decade ago to only a few today @xcite . among the surviving candidates
is the weakly interacting massive particle ( wimp ) . from a theoretical standpoint wimps
enjoy a great deal of support .
they fit well into our understanding of structure formation in the universe @xcite and into many supersymmetric theories of particle physics @xcite .
because of this theoretical support an increasing number of experimentalists employing a variety of techniques have set out to find these particles .
one such technique consists of looking for evidence of recoils from wimps in ancient mica@xcite .
this technique exploits the fact that atoms recoiling from interactions with wimps ( `` wimp recoils '' ) would produce permanent chemical changes in the mica which later would produce etch pits when immersed in hydrofluoric acid .
stringent limits on wimp cross sections as a function of mass were obtained @xcite by scanning only a small area , due to the fact that ancient mica integrates this signal for order 1 gyr .
unfortunately other particles , e.g. fission neutrons , can also cause recoils which produce etch pits very similar to the etch pits we expect from wimp recoils .
eventually this background will limit the technique and ultimately preclude the possibility of actually discovering wimps using this method .
a signature for wimps is needed . in this letter
we discuss a possible signature and calculate its magnitude .
the signature arises from the fact that our motion around the center of the galaxy through a stationary wimp halo causes a wimp `` wind '' to arrive at earth preferentially from one direction , the so - called ram direction .
the ancient mica technique can exploit this asymmetry because the mica is itself asymmetric .
mica is a layered crystalline mineral which is easily cleaved along the @xmath1 crystallographic plane . in order to distinguish
wimp recoils from the large population of etch pits produced by alpha - decay in the @xmath2u and @xmath3th decay chains , we require recoils to cross the cleavage plane @xcite . in order to quantify the asymetry of the mica we define , @xmath4 to be the signal contrast where n is the number of matched etched pits created by wimp recoils and the subscripts , min and max , refer to the maximum and minimum of n over all possible mica orientations .
the signal contrast is the maximum assymetry possible . to determine the magnitude of this parameter for a wimp halo , we performed the following calculation .
starting with the recoil spectrum@xcite , energy and angle with respect to the ram direction @xmath5 , and using @xmath6 and @xmath7 with a cutoff at @xmath8 , we randomly generated recoils of the constituent atoms of mica .
ranges and stopping powers of the recoils were calculated with the trim92 computer code @xcite . as discussed by snowden - ifft and chan @xcite ,
not every ion which crosses the cleavage plane produces an etch pit .
the appearance of an etch pit is determined by a stochastic process which depends on the type of stopping power ( nuclear or electronic , characterized by two constants @xmath9 and @xmath10 ) , the magnitude of the stopping power and the angle the recoil makes with the @xmath1 plane .
the angular dependence of the etching model has been well tested @xcite and it is found that ions normally incident on the cleavage plane are less likely to create an etch pit than those crossing at grazing angles .
each recoil that crossed the cleavage plane was etched according to this etching model @xcite with @xmath11 , @xmath12 and @xmath13 nm .
a coherence term @xcite was included in the calculation .
this procedure was carried out for a variety of wimp masses @xmath14 at a variety of angles @xmath5 .
we find that the rate of accumulation of matched etched pits with depths of 2 nm or greater can be adequately characterized by @xmath15 @xmath16 was positive for all wimp masses considered , so mica oriented at @xmath17 has a higher rate of accumulation than mica oriented at @xmath18 .
two pieces of mica constantly held at @xmath17 and @xmath18 will therefore have the largest possible signal contrast , @xmath19 , given by @xmath20 in table [ table1 ] we show @xmath19 for several values of wimp mass .
the lower the mass of the wimp the larger @xmath19 although for very low mass wimps the expected density decreases rapidly .
no piece of mica has pointed in the same direction over geologic time , so to compute the wimp track density as a function of contemporary orientation and age we require a detailed history of an ancient mica s orientation with respect to the ram direction .
this history is complex , but quantifiable .
several rotations contribute to it ; we discuss each below in order of increasing timescale .
the diurnal rotation of the vector normal to the @xmath1 mica plane , the mica normal vector , about the celestial pole has the shortest timescale : 24 hours . over timescales much larger than this
the track density will be a function only of the declination of the mica normal vector , or the _ mica normal declination_. we assume that the interaction length of wimps is much larger than the radius of the earth .
the motion of the celestial pole with respect to the ecliptic pole is comprised of two general rotations : nutation and precession .
nutation has an amplitude of less than 10 arcseconds@xcite , so can be neglected .
precession is the smooth rotation of the celestial pole about the ecliptic pole with fixed angle
the obliquity between the two .
the obliquities of most of the inner planets have undergone large and chaotic changes on timescales of order 1 - 10 my @xcite since solar system formation , with excursions from @xmath21 to @xmath22 .
fortunately , both for this calculation and for the climatic stability of the earth , the presence of the moon has had a dramatic stabilizing effect , so that the earth s obliquity has varied from its mean value of @xmath23 by only @xmath24 since the moon was captured@xcite .
the slow orbital motion of the solar system around the center of the galaxy has the longest timescale of the celestial rotations .
this motion has the effect of slowly rotating the ram direction within the galactic plane .
we assumed that the solar system describes a circular orbit about the galactic center , with a velocity of 220 km sec@xmath25@xcite and a period of 224 my@xcite .
in addition to these celestial rotations , tectonic drift , with a timescale between 10 - 100 my , also affects a mica crystal s orientation with respect to the ram direction .
it is not necessary to know the absolute longitude of the body , but a history of the latitude and orientation of the body is required .
paleomagnetic data directly provide this history . in a simplified description of the technique ,
the direction of the terrestrial magnetic field is recorded by the magnetization direction of a magnetic mineral as it cools below the so called `` blocking '' temperature .
the local magnetic field direction at that time can thus be determined by a careful measurement of the magnetization direction of the mineral , and this determines the position of the terrestrial magnetic pole with respect to the sample .
the age of the mineral is typically derived from the fossil record in the same stratum as the mineral for phanerozoic ( younger than 590 my ) samples , and by radioisotope or fission - track dating for pre - cambrian ( older than 590 my ) samples . in practice , there are many complications , but the technique is well - developed , and the position for an ancient magnetic pole can be determined with an accuracy of @xmath26 .
the determination of a mean paleomagnetic pole is equivalent to a determination of the geographic pole position relative to the sample@xcite .
the discrepancy between the mean paleomagnetic pole and the modern geographic pole is attributed to tectonic drift , yielding the latitude and orientation of the body containing the sample .
a large body of time sequences of paleomagnetic pole measurements generally known as apparent polar wander paths ( apwps ) have been compiled by several authors ( e.g. , @xcite ) with the general goal of the reconstruction of ancient geography .
this dataset is ideal for our purposes .
the best - determined apwps have been done for the major continental cratons ( bodies which have been geologically stable and have undergone only rigid - body rotations over long geological times ) . for each craton , we arbitrarily choose a point near the geographic center of the craton for the purpose of calculation . as a general statement ,
the reconstruction of apwps is best for phanerozoic time because the fossil record allows for accurate dating ; for pre - cambrian times apwps are in general rather poorly known , although the situation is improving rapidly particularly for the north american and european continents ( e.g. , @xcite ) .
we briefly summarize here the method used for signal integration . the track density for a piece of mica of age @xmath27 , zenith angle @xmath28 and azimuth angle @xmath29 ( true north corresponding to @xmath30 , true east to @xmath31 , etc . )
can be calculated from eq .
[ n - eqn ] to be @xmath32 t , \label{n - integrated}\ ] ] where @xmath33 is the average value of @xmath34 over the age of the mica sample . for a given wimp mass and mica age
this function will achieve a maximum value , @xmath35 , in an orientation specified by @xmath36 and @xmath37 and a minimum value , @xmath38 , for some other orientation specified by @xmath39 and @xmath40 .
combining eqs .
[ s - first ] and [ n - integrated ] with the fact that @xmath41 allows us to factor @xmath42 and write it as @xmath43 where @xmath44 ( eq . [ smax - eqn ] ) characterizes the asymmetry of the mica response and is independent of time , and @xmath45 characterizes the mica s history and is independent of mass . the track density integrated over timescales longer than the precessional timescale ( @xmath46 y ) but shorter than the timescale for tectonic drift and galactic rotation ( @xmath47 my ) is a function only of mica normal declination @xmath48 and of the ecliptic declination of the ram direction @xmath49 , since rapid diurnal rotation and polar precession have the effect of averaging over celestial right ascension and ecliptic longitude . in fig .
[ contour ] , we show @xmath50 , the calculated average of @xmath34 over these intermediate timescales as a function of mica normal declination and ram ecliptic declination . to calculate @xmath51 over geologic timescales , we first determine the ram ecliptic declination @xmath52 for a given short geological time interval @xmath53 .
we then determine the continental latitude and orientation using an linear interpolation on the sphere between the two paleomagnetic pole measurements closest in time to @xmath54 .
the ancient values of mica normal declination @xmath55 are determined for each contemporary mica sample orientation ( @xmath28,@xmath29 ) , using the paleomagnetically determined rotations from the current position .
symbolically , the functional dependence of the calculations is summarized by : @xmath56.\ ] ] @xmath57 in eq .
6 may then be calculated using eq .
7 .
although much of this integration can be in principle be done analytically , tectonic drift requires that the calculation be done numerically . in fig .
[ xifigure ] , we show @xmath57 as a function of mica age for several major continental cratons . for this calculation , we used time intervals of 5 my , a @xmath58 grid of diurnal and precessional rotations , and a @xmath58 grid of mica orientations .
because of the symmetry of the mica mica orientation is defined only up to a parity transformation it is only necessary to define the signal over the upper hemisphere , ( @xmath59 ) .
the periodicity due to galactic rotation is clearly visible .
it is desirable to choose locations and ages for which @xmath60 is large and , because of uncertainties in measuring the age of the mica , for which the signal amplitude and direction are not changing rapidly .
we intend to address in a future paper the effect of the uncertainties in our calculation , including the uncertainty in the length of the galactic year ( @xmath61 ) , the effect of uncertainties in mica ages , and so on .
for old mica the signal is smaller but the integration time is longer and for a given area of mica scanned the statistical significance of an observed signal turns out to be roughly independent of age .
the signal contrast for 50 gev wimps ( @xmath62% ) in mica 500 my old ( @xmath63% ) is of order one percent ( @xmath64% ) .
this signal contrast is small but it is `` state - of - the - art . ''
no wimp limits have been published with detectors having a substantially larger signal contrast .
finally we need to address the feasibility of such an experiment .
first in order to utilize this idea we will need to obtain several pieces of mica oriented in the proper directions .
fortunately the large nearly perfect crystals required of such a search are almost always found in coarse grained rock formations known as pegmatites .
a typical pegmatite will contain many large mica crystals oriented in random directions@xcite .
the existence of these structures virtually guarantees that we will be able to obtain mica of the proper quality having the same composition and same history and oriented in nearly the optimal directions .
large concentrations of uranium in the pegmatite will need to be surveyed as the fast neutrons that emanate from them could potentially mimic our signal .
the second issue is the difficulty of detecting a @xmath64% signal . if wimps exist at cross - sections just below our current limits we estimate that it will take a @xmath65 improvement in scanned area to see a signal at the 3@xmath66 level . with an automated afm an area equivalent to that of our first mica search can be scanned in a little less than 4 hours .
an additional factor of 3 is gained by etching for twice as long . as mentioned above the appearance of an etch pit
is governed by a stochastic process . as shown in @xcite , a longer etching time results in more opportunities for an etch pit to form .
this effect has been confirmed experimentally@xcite .
the @xmath65 improvement can therefore be achieved in about 6 years of machine time . as discussed in [ 6 ] large crystals of mica
will allow us to achieve this improvement without running into neutron background .
if a signal is observed it can be confirmed by similar experiments on other pegmatites .
we conclude by first stressing the two main points of this letter . because of the asymmetries inherent in etching latent tracks in ancient mica our detector can sense the direction of the wimp `` wind . ''
moreover despite the complex rotations that mica crystals makes with respect to this wimp `` wind , '' there is a substantial dependence of etch pit density on present day mica orientation .
this dependence can be used as a signature for wimps which in turn will allow us to detect wimps if they exist or rule them out more efficiently if they do not . |
observations have established the presence of a supermassive black hole ( smbh ) in the center of virtually every massive galaxy in the local universe @xcite .
there is a large body of circumstantial evidence suggesting that feedback from smbhs during luminous accretion episodes active galactic nuclei or quasars plays prominent roles in galaxy evolution ( * ? ? ?
* and references within ) .
quasar activity also helped to reionize and heat the intergalactic medium @xcite , which may have influenced the formation and evolution of low - mass galaxies and their central bhs @xcite .
the origins of these cosmic behemoths remain a fundamental unsolved problem ( see reviews by * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
particularly puzzling are the smbhs with masses in excess of @xmath11 powering luminous quasars at redshifts @xmath12 @xcite , less than @xmath13 after the big bang . to reach such masses in so short a time , these smbhs must have either accreted nearly continuously near the eddington limit ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) or undergone episodes of super - eddington accretion @xcite regardless of whether they formed as the remnants of the first generation of stars @xcite or through the ` direct collapse ' of atomic - cooling gas @xcite . the notion that smbhs accreted nearly uninterrupted runs counter to expectations from observations at lower redshifts ( @xmath14 ) , where only a small fraction of smbhs are undergoing quasar episodes , which are estimated to last for @xmath15 to @xmath16 ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in this paper ,
i show that steady and prolific growth of nuclear bhs at @xmath17 can be reconciled with their relative inactivity at lower redshifts if gas accretion episodes near the eddington limit are triggered by major mergers of the bh s host galaxy or dark matter halo .
major mergers of galaxies have long been associated with quasar activity @xcite .
( note , however , that luminous accretion can also be triggered by secular processes ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
previously , li et al .
* see [ sec : mergers ] ) argued , by examining the hierarchical growth of an exceptionally massive dark matter halo a large cosmological simulation , that host major mergers provide a plausible explanation for the growth of @xmath3 quasar smbhs .
here , i use a semi - analytic monte carlo technique @xcite to argue that this is the case generally
i.e . that the most massive dark matter halos experience a rapid succession of major mergers prior to @xmath18 .
the rate of major galaxy mergers per unit time per dark matter halo evolves extremely rapidly with redshift , roughly as @xmath19 , where the proportionality holds at the large redshifts of interest here .
this is a much steeper dependence than most physical timescales that plausibly govern the duration of a bh feeding event for example , the dynamical time at the virial radius of dark matter halos scales as @xmath20 .
put another way , we can write the duty cycle of bh growth at any epoch as f_duty=_trig t_feed , [ eq : fduty ] where @xmath21 is the frequency of trigger events per bh and @xmath22 is the typical duration of each feeding episode . both of the quantities on the right hand side can depend on factors such as redshift , bh mass , the masses of the merging galaxies , and so on i will return to this point later .
for feeding episodes triggered by major mergers of galaxies , the trigger rate @xmath23 increases so rapidly with redshift that there is a large range of functions @xmath24 that would allow for nearly continuous bh growth ( @xmath25 ) at high @xmath26 .
i motivate a simple parametrization for @xmath22 for the narrow subpopulation of smbhs of interest ( @xmath3 , masses @xmath27 ) , and delineate the region of parameter space that can explain their formation at the observed number densities .
an example of a successful growth model is one where major galaxy mergers trigger fast - feeding episodes lasting for a timescale comparable to the dynamical time of the host halo .
this paper is organized as follows . in
2 , i present additional background by summarizing general properties of luminous quasars and by detailing the argument that @xmath3 quasar must have experienced nearly continuous growth if their accretion was eddington - limited .
i present in 3 results from merger - tree simulations of hierarchical structure formation , showing the prolific growth histories of the massive dark matter halos that are likely to host the @xmath17 quasars . in
4 , i motivate a specific parametrization of the growth episodes triggered by galaxy mergers , and model the durations of such episodes required to explain the @xmath28 quasar smbhs without super - eddington accretion .
i conclude in 5 . throughout this work ,
i adopt the cosmological parameters @xmath29 , @xmath30 , @xmath31 , @xmath32 and @xmath33 ; these values are chosen based on the latest published empirical values @xcite .
while quantities such as the age of the universe and the mass function of dark matter halos are sensitive to these parameters , the methods and results presented here are qualitatively robust .
the majority of the mass in smbhs in the local universe appears to have been accumulated during luminous quasar episodes @xcite .
the brightest quasars have luminosities on the order of @xmath34 times the eddington luminosity of the smbh engine ( e.g. * ? ? ?
* ; * ? ? ?
* ) , @xmath35 .
the luminosity can be expressed in terms of the mass energy of the accreted fuel and a radiative efficiency factor @xmath36 as l = c^2 . by comparing the cumulative quasar luminosity density at @xmath37 with the mass density in nuclear smbhs
, the cosmic mean value of @xmath36 is found to be @xmath38 @xcite , in rough agreement with theoretical expectations of luminous accretion flows ( * ? ? ? * and refs .
therein ) .
the quasar duty cycle , or the fraction of time smbhs spend as quasars , is @xmath39 at @xmath40 ( where the number density of quasars peaks ) ( e.g. * ? ? ?
* ) , and decreases with redshift ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
estimates of quasar lifetimes vary , but tend to fall in the range @xmath41 ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the radiative and kinetic output from the luminous quasar is thought to act as a negative feedback , making growth intermittent ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
although luminous accretion activity is associated with major and minor mergers , as well as possibly secular processes , the detailed mechanism that channels the gas to the nuclear smbh remains an open topic of study ( e.g. * ? ? ?
there is uncertainty as to how much of the smbh growth occurs when it is observable as a quasar , as opposed to other stages such as in the midst of a galaxy merger during which the central smbh is heavily obscured . for a fixed value of @xmath36
, eddington - limited accretion implies exponential growth ( @xmath42 ) , with an @xmath43-folding timescale t_edd=450 , with @xmath44 for @xmath45 . for the adopted cosmological parameters , the age of the universe at @xmath46 ( @xmath47 ) is @xmath48 myr ( @xmath49 myr ) .
if the @xmath3 smbhs grew from the remnants of the first generation of stars ( population iii or ` popiii ' stars ) at @xmath50 , then they would have had approximately @xmath51 to grow to @xmath0 .
( note that the available growth time @xmath52 is only marginally longer if the ` seed ' bh began to grow at , say , @xmath53 . )
nuclear bhs can also grow through hierarchical bh mergers , but the efficiency of this avenue is limited @xcite by the gravitational recoil effect ( the momentum imparted by asymmetric gravitational - wave emission on the product of a bh merger , e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) that is , nature can not simply throw together a thousand seed bhs to form a bh a thousand times more massive .
optimistic estimates suggest that mergers between popiii remnants could contribute a factor @xmath54 toward assembling a @xmath55 smbh by @xmath56 @xcite , regardless of whether the seeds formed and began to grow in halos with virial temperatures @xmath57 or @xmath58 .
the mean eddington ratio required to form a smbh with mass @xmath59 from a seed bh with mass @xmath60 can be written as ( equation reproduced from * ? ? ?
* ) @xmath61\nonumber\\ & \approx & \left[0.676 + 0.045\ln\left(\frac{m_{\rm smbh}}{3\times 10 ^ 9\,{\;\mathrm{m}_{\odot } } } \frac{30\,{\;\mathrm{m}_{\odot}}}{m_{\rm seed}}\frac{30}{x_{\rm merge}}\right ) \right]\nonumber\\ & & \qquad \times\left(\frac{\eta}{0.07}\right)\left(\frac{t_{\rm avail}}{700\,{\;\mathrm{myr}}}\right)^{-1}. \label{eq : feddpopiii}\end{aligned}\ ] ] the masses of the @xmath3 quasars could be explained if they began as popiii remnants and grew at the eddington limit for @xmath62 of the available time , or at @xmath62 of the eddington limit for the entire available time .
we can repeat the above exercise for the ` direct collapse ' seed scenario , in which nuclear bhs form in the gravitational collapse of massive , atomic - cooling ( temperature @xmath63 ) gas clouds .
such an event could occur if the gas has low metallicity and is inundated by a strong ultraviolet ( uv ) flux that photo - dissociates molecular hydrogen and thus prevents fragmentation of the cloud into stars of ordinary mass ( * ? ? ?
* ; * ? ? ?
* additional references in [ sec : intro ] ) .
direct collapse could form bhs with masses @xmath64 , but only after redshifts @xmath65 @xcite ( when the universe is @xmath66 old ) if the mechanism is requires the prior emergence of powerful uv sources ( see * ? ? ? * for a discussion ) .
in other words , direct collapse seeds are expected to begin with a head start in mass , but a delayed start in time . moreover , because these seeds can only form in rare massive halos under specific circumstances , their opportunities to grow via major mergers is limited ( i.e. smaller @xmath67 than a scenario where popiii seeds form and merge ) .
we have @xcite @xmath68\nonumber\\ & \qquad \times\left(\frac{\eta}{0.07}\right)\left(\frac{t_{\rm
\label{eq : fedduv}\end{aligned}\ ] ] equations [ eq : feddpopiii ] and [ eq : fedduv ] imply that _ both the popiii and direct collapse seed scenarios require the @xmath3 quasar smbhs to have grown nearly continuously _
( i.e. accreting more than half of the time ) , if the growth occurred at rates near the eddington limit .
this statement is still true even if @xmath69 bhs were to have formed as early as @xmath70 ( see * ? ? ?
* for an example of such a scenario ) .
the requirement of such large duty cycles poses a stark contrast with the rarity of luminous quasar activity observed at @xmath14 .
it is this contrast that i address in this paper .
i begin by discussing the frequency of major merger events for massive galaxies at redshifts @xmath3 .
the goal here is to answer a simple question : supposing that a galaxy major merger triggers a feeding episode of the central bh , how often do the hosts of the @xmath3 quasar smbhs undergo such triggers ?
in other words , i seek to quantify one of the two factors , @xmath21 on the right hand side of interest in equation [ eq : fduty ] ; i will turn to the other factor , @xmath22 , in the next section . strictly speaking , throughout this paper i am referring to mergers of host dark matter halos , not host galaxies however , at redshifts of interest , the masses of the most massive halos ( @xmath71 ) are comparable to the largest galaxy masses ( e.g. * ? ? ? * ; * ? ? ?
* ) , but not to galaxy clusters
. therefore , in this paper i assume that galaxy counts and halo counts are equivalent , and use the two terms interchangeably .
the number density of @xmath0 smbhs at @xmath28 is @xmath72 @xcite , or of order one or a few in the millennium simulation @xcite volume .
the masses of dark matter halos with comparable abundances are @xmath73 , and the naive expectation is that these smbhs are hosted by the most massive class of halos ( see , however * ? ? ?
earlier , li et al . ( 2007 ) @xcite examined a particularly massive dark matter halo in a cosmological @xmath74-body simulation and noted that it experienced seven major mergers ( which they defined as mergers with progenitor mass ratio @xmath75 ) in rapid succession between @xmath76 and @xmath46 .
they showed that this prolific merger history provides a plausible explanation for the growth of a @xmath3 , @xmath27 smbh without super - eddington accretion .
angulo et al .
( 2012 ) @xcite noted that the most massive @xmath18 halos ( @xmath73 ) doubled their masses in the preceding 100 myr .
i present that the explanation provided by li et al . , if qualitatively correct , is likely to hold quite generally for any comparable volume in the universe .
figure [ fig : merge1 ] shows a graphical representation of the merger history of a dark matter halo that reaches a mass of @xmath77 at @xmath46 .
the merger history is generated using a monte carlo merger tree algorithm whose underlying mathematical formalism provides excellent matches to cosmological @xmath74-body simulations , especially for large halo masses ( e.g. * ? ? ?
* ) , and whose numerical implementation has been shown to be highly accurate @xcite . the growth history of one dm halo with @xmath78 .
the thickness of the gray shaded region is proportional to @xmath79 $ ] , where @xmath80 is the most massive progenitor halo ( the `` trunk '' of the merger tree ) .
the diamonds denote significant merger events ( progenitor masses more equal than 1:9 ) , with the center of the diamond marking the time of the merger and the width - to - height ratio equal to the progenitor mass ratio . ] the horizontal axis of figure [ fig : merge1 ] shows the age of the universe , in linear scale to emphasize the rapid hierarchical growth of the halo .
the width of the shaded region represents the mass of the most massive progenitor of the halo ( i.e. the ` trunk ' of the merger tree ) at any given time ; the width of the region is proportional to @xmath81 $ ] .
the diamonds mark merger events ( down to mass ratios of @xmath82 ) , with the center of the diamond indicating the time of the merger and the ratio of the axes showing the mass ratio of the merging progenitors . between @xmath83 and @xmath84 , this particular halo experiences 10 mergers with mass ratios @xmath85 , 16 mergers with @xmath86 , and 10 additional mergers with @xmath87 .
the physical time between mergers is much shorter at higher redshifts , showing a steep evolution in the merger rate with redshift ( more on this point shortly ) .
figure [ fig : merge40 ] presents the same information , but for 40 simulated dark matter halos , all with @xmath88 .
the merger tree sample is the same as the one used in tanaka & li ( 2014 ) @xcite . it is readily apparent from the figure that the points made in the previous paragraph hold generally for this mass class of halos .
all of these halos undergo @xmath89 major merger events ( regardless of whether one defines a major merger with a minimum mass ratio of 1/3 or 1/5 ) between @xmath65 and @xmath18 , a span of @xmath90 .
by contrast , the typical massive galaxy experiences @xmath8 major merger between @xmath91 and @xmath92 ( e.g. * ? ? ?
* ) , a span of over @xmath93 . the same as fig.1 , except that this figure shows the growth histories of 40 simulated dm halos with @xmath78 . the thickness of each gray region is proportional to @xmath79 $ ] .
the diamonds mark merger events with progenitor masses more equal than 1:9 , with the width - to - height ratio equal to the progenitor mass ratio .
the `` trunk '' of each merger tree has a mass of @xmath94 at @xmath95 .
, width=451 ] to further emphasize the fact that the prolific merger history found by li et al .
is generic for all halos of similar mass at similar redshift , and in an effort to quantify this trend , i show in figure [ fig : mergerates ] the mean merger rates @xmath96 ( per unit time per halo ) of massive halos .
the halo sample includes 40 halos with @xmath97>12.9 $ ] , 100 halos with @xmath98>12.5 $ ] , and 100 halos in each mass bin with @xmath99>12.0 $ ] , @xmath100>11.5 $ ] etc .
, down to @xmath101>8.0 $ ] .
the counting statistics for each mass bin are scaled up to match the abundances expected in a 50 comoving gpc@xmath102 volume ( halo cloning ; see , e.g. * ? ? ?
this method is effective for examining the assembly histories of the most massive dm halos , which are sparsely sampled by even the largest cosmological @xmath74-body simulations ; on the other hand , it offers much weaker statistical power for all but the most massive objects ( i.e. the lower - mass bins are represented by a hundred halos here , but represented in the thousands in large @xmath74-body simulations ) . in figure
[ fig : mergerates ] , the thick blue , medium green , and thin red curves show merger rates of dark matter halos whose ( post - merger ) masses are @xmath103>11.5 $ ] , @xmath104>9.5 $ ] and @xmath105>7.5 $ ] , respectively .
the solid curves show the rate of mergers whose progenitor mass ratios are greater than @xmath106 , and the dashed lines show the rate of mergers with @xmath107 .
the grey lines show the mean merger rate in the main progenitor ( the merger tree ` trunk ' ) of the most massive halos that have @xmath108 at @xmath46 .
this is the same sample of 40 halos whose merger histories are graphically represented in figure [ fig : merge40 ] .
the thickness of the grey curves and their colored accents denote where the mean mass for this sample lies within the mass bins described above ( i.e. thick blue denotes where @xmath109>11.5 $ ] , and so on ) .
note that whereas the blue , green and red curves show the mean merger rates for halos in the entire simulation sample , the grey curves show the merger rates only for the main progenitors of halos that end up with @xmath88 .
the mean merger rates @xmath96 ( per unit time per halo ) , of dm halos in the merger tree simulation .
the thick blue lines denote mergers for halos with @xmath110 , the green lines for halos with @xmath111 , and the thin red lines for halos with @xmath112 .
the solid and dashed lines show mergers with progenitor mass ratios @xmath106 and @xmath107 , respectively .
the grey curves show the mean merger rates for the main ( ` trunk ' ) progenitor for halos that have @xmath108 at @xmath46 ( the same 40 halos as shown in the previous figure ) , with the line thickness and color accents indicating when the mean mass of the trunk lies within the three bins described above . for reference ,
i have also plotted the quantities @xmath113 ( solid black line ) and @xmath114 ( dotted black line ) . ]
figure [ fig : mergerates ] shows some noteworthy trends , most of which have been noted in previous studies utilizing semi - analytic methods ( e.g. * ? ? ?
* ; * ? ? ?
* ) and @xmath74-body simulations ( e.g. * ? ? ?
* ; * ? ? ?
first , more massive halos have somewhat higher merger rates .
second , mergers of ratios @xmath85 are almost as common as mergers of @xmath107 .
this piece of information is useful because the definition of what mass ratios constitute a ` major ' merger is arbitrary ; previous works have used @xmath115 , @xmath116 , and so on .
this figure suggests that the choice of a minimum @xmath117 value for defining a major merger ( i ) does not qualitatively affect the ( steep ) redshift evolution of the inferred major merger rate , and ( ii ) affects it quantitatively only up to a factor of order unity .
( note that there are other uncertainties in defining and interpreting merger rates of halos and galaxies @xcite ) .
third , and most importantly for the topic of this paper and as suggested by the previous figures , the major merger rate per galaxy evolves very rapidly with redshift . across the range of halo masses considered here ,
the rate at @xmath118 is higher by an order of magnitude than the rate at @xmath119 , which is in turn @xmath120 times greater than the predicted major merger rates of quasar hosts at @xmath121 ( * ? ? ?
* see below ) . finally , the cosmic mean merger rate for _ halos in a fixed mass range at all redshifts _ ( colored curves ) and the mean merger rate for the main progenitors of the @xmath78 halos _ across five orders of magnitude of mass growth _ ( grey curves )
exhibit similar redshift evolution . for comparison ,
i have plotted the quantity @xmath122 , or the rate at which the universe ages per unit redshift , as a dotted black curve . at the redshift values of interest
, @xmath123 scales approximately as @xmath4 .
the evolution of the halo merger rate roughly follows this power law , consistent with theoretical expectations and confirming the fidelity of the merger trees .
this simply reflects the fact that the theoretical merger rate per redshift @xmath124 depends weakly on @xmath26 . these semi
- analytic results can be compared to those of fakhouri et al .
( 2010 ) @xcite , who investigated the halo merger rates in the two millennium simulations @xcite .
although the cosmological parameters adopted here differ slightly from those in those simulations , the two sets of results are broadly consistent with each other : those authors also found major merger rates of @xmath125 for massive halos at @xmath3 , and that @xmath96 scaled roughly as @xmath126 .
the reader may wish to juxtapose figure [ fig : mergerates ] in this work to the right - hand panel of figure 3 in fakhouri et al . note that whereas that study focused mostly on halo mergers at @xmath127 , here i m interested in the range @xmath128 .
having quantified the host merger rates , the rate of trigger events in equation [ eq : fduty ] , i now turn to the effective duration of the smbh ` feeding time . '
it is worth re - emphasizing that the exact manner in which galaxy mergers deliver gas to the central bh(s ) is not fully understood ( cf .
* ; * ? ? ?
assuming constant growth at fixed eddington rate is ( predictably ) problematic @xcite .
while numerous studies semi - analytic models , as well as simulations that employ sub - grid prescriptions for smbh growth have used bondi - hoyle accretion , this prescription is known to break down in the presence of radiative feedback from the bh @xcite , angular momentum @xcite , inhomogeneous gas cooling and dynamics @xcite , advection @xcite , thermal conduction @xcite , etc . indeed
, many examples of luminous bh activity do not appear to be represented by bondi accretion ( e.g. * ? ? ? * ) .
prescriptions where smbh growth is coupled to the baryonic properties of the host can be successful with specific model prescriptions ( e.g. * ? ? ? * ; * ? ? ?
* ) , as are some models with local ( e.g. * ? ? ?
* ; * ? ? ?
* ) , and global @xcite regulatory feedback . in discussing the host merger - triggered feeding times ,
it is important to clarify two points .
first , the growth episode is not a step - function event ( e.g. * ? ? ?
* ) , but rather a continuous process during which the smbh accretion rate varies with time . here
, i refer to as the ` feeding time ' the effective period over which the average accretion rate is equal to eddington .
second , i distinguish the terms ` feeding time ' and ` quasar lifetime ' to emphasize that the two need not be the same .
although quasar episodes are associated with the final stages of smbh growth via rapid gas accretion , at high redshifts the growth could be preferentially more obscured due to greater gas column densities particularly so , if obscuration by host mergers are commonplace .
depending on the host morphology , as well as the line of sight , timing and wavelength of the observation , a given growth episode may or may not be classified as a quasar . in general , the feeding time should depend on a number of variables : the total mass and mass ratio of the merging host , the redshift , the mass and spin of each nuclear bh , the gas metallicity , the geometry of a given host merger and the angular momentum of the gas , and perhaps the dynamics of the nuclear bhs as they form a binary and evolve .
however , in this particular instance , we re concerned with a very specific subpopulation of massive dark matter halos that share similar growth histories . all of these halos were selected to have similar masses at @xmath46 , and their mass growth histories @xmath129 are highly uniform , with the masses of their main progenitors typically only varying by a factor of a few out to @xmath2 ; see fig .
[ fig : merge40 ] .
( note that this trend does not extend to lower redshifts ; the most massive halos at @xmath28 do not necessarily grow into the most massive halos at @xmath14 ( e.g. * ? ? ?
) i will also restrict the following analysis to major mergers with mass ratios @xmath130 .
in addition , these halos always have masses well above the cosmological jeans ( filtering ) scales , even in hypothetical scenarios where the intergalactic medium is heated prolifically by early mini - quasar activity @xcite .
thus , at any given redshift , these halos are the least sensitive ( compared to lower - mass halos ) to spatial fluctuations in local radiative backgrounds .
i consider a feeding time of the form t_feed(m_halo , m_bh , z , , ... ) ~t_feed ( z=6 ) ( ) ^a .
[ eq : tfeed ] that is , because in this particular instance the variations in @xmath131 , the merger rates , and the mass ratio @xmath117 are small , i suppose that the duration of the smbh feeding episode can be characterized by a characteristic mean value @xmath132 that follows a redshift evolution @xmath133 .
i assume that at any given redshift , the feeding episode duration can be characterized by a log - normal distribution with scatter @xmath134 .
many studies have sought to empirically estimate quasar lifetimes , with large uncertainties .
the best constraints come from data at @xmath14 ; the limited redshift range , and the fact that the overall quasar sample contains a wide variety of host galaxies and smbh masses , makes it difficult to evaluate how smbh growth depends on the redshift and the host properties .
for example , wyithe & loeb ( 2009 ) @xcite suggest that the quasar lifetime scales with the dynamical time of the host dark matter halo , @xmath135^{-3/2 } { \;\mathrm{myr}}$ ] ( e.g. * ? ? ? * ) . for reference ,
i have plotted @xmath113 alongside the halo merger rates in figure [ fig : mergerates ] ( solid gray line ) .
i perform a parameter space survey to quantify what combinations of feeding time parameters
normalization @xmath136 , redshift evolution slope @xmath137 , and scatter @xmath134can explain the observed population of @xmath28 , @xmath27 quasars via host merger - triggered , eddington - limited growth .
i assume that a nuclear bh is in place in each massive halo , and that following a major merger @xmath130 the bh accretes for @xmath22 , drawn randomly out of the @xmath26-dependent log - normal distribution , over which time the mean accretion rate is eddington .
i take a conservative model , in which a subsequent major merger during @xmath22 does not extend the feeding episode ( i.e. rapid mergers can not counteract feedback by ` blowout ' ) ; an alternative model would be one where the bh accretes for @xmath22 since the last major merger .
this exercise is repeated 100 times ( to statistically average different possible monte carlo realizations of @xmath22 ) for each of the main progenitors of the sample of halos with @xmath138 .
molecular - cooling halos that have just formed popiii stars will have shallow potentials , and may be more susceptible to negative feedback from bh activity @xcite .
therefore , i only allow bhs to grow if their halos are atomic - cooling ( virial temperatures @xmath139 ) , which may be a crucial threshold that allows for dense cooling flows to carry gas to the central bh @xcite .
note that this sample of particularly massive halos becomes atomic - cooling at @xmath50 , and exist at the same abundances as the @xmath3 quasars ( @xmath140 ) as early as @xmath2 ( see * ? ? ?
this is much earlier than the typical redshift for popiii formation ( @xmath141 ) or proto - galaxy formation ( @xmath142 ) .
this implies that if the first ` monster ' smbhs grew from popiii seeds , then the numerical simulations focusing on typical halos at @xmath143 may not be representative of their cradles .
the collected output is the number density of host halos whose central bhs would have grown to @xmath0 by @xmath46 . in figure
[ fig : npop3 ] , i plot this quantity for population synthesis realizations resulting from different combinations of the three model parameter values .
i evaluate the number density of halos whose bhs have grown by a factor @xmath144 from when the main progenitor of the host halo is atomic - cooling to @xmath46 .
that is , the seed bh in the main progenitor must grow by a factor of @xmath145 via host merger - triggered gas accretion episodes , and acquire @xmath146 from a combination of the initial mass and bh mergers ( e.g. by having @xmath147 at formation and growing by a factor @xmath148 via bh mergers ) .
the white grids in the figure represent model parameter combinations that do not produce any such quasar smbhs .
the dark grey and black grids show models that overproduce massive bhs at @xmath46 .
a model may be said to be viable if the predicted number density of @xmath28 , @xmath27 smbhs matches the observed estimate of @xmath149 @xcite .
the blue solid lines in figure [ fig : npop3 ] show , for reference , the combination of parameters where @xmath150 would be @xmath151 and @xmath16 , representing the approximate quasar lifetimes derived from observations at @xmath14 .
these blue lines are meant only as rough guides .
again , ( i ) quasar lifetimes are not ( necessarily ) bh feeding times , and ( ii ) the halos studied here belong to a very narrow subset of the most massive @xmath17 halos and have quantitatively similar mass growth histories , and may be quite different from hosts of low - redshift quasars . the number density of bhs in halos with @xmath152 , and that have grown by a factor @xmath144 between when the main progenitors of host halo becomes atomic - cooling ( virial temperature @xmath153 , @xmath154 ) and @xmath46 .
these criteria are representative of the growth requirements from a population iii bh seed ( see text ) .
the different grids show outcomes for different combinations of parameters for the feeding time : the normalization at @xmath46 and the redshift dependence exponent @xmath137 ( where @xmath155 ) .
the three panels show results for different assumptions of log - normal scatter in the feeding times , from top to bottom , 0.1 , 0.3 and 0.5 dex .
white grids show models that have no smbhs at @xmath46 , and dark gray grids show those that exceed the observed quasar number density ( @xmath149 ) . ]
i repeat the above parameter survey for direct - collapse seed models , which may form by @xmath65 and have initial masses as large as @xmath156 [ see [ sec : quasars ] ] .
whereas for population iii seed models i considered bhs that have grown by @xmath144 between when the host becomes atomic - cooling and @xmath46 , here i consider bhs that have grown by @xmath157 between @xmath158 .
the results are plotted in figure [ fig : ndc ] .
the parameter requirements do not differ very much from the popiii case , which is not surprising given that similar mean eddington ratios are required for both the popiii and direct collapse families of models [ see equations [ eq : feddpopiii ] and [ eq : fedduv ] ] .
roughly speaking , both models require bhs of masses @xmath159 to be in place by @xmath160 , and grow by a factor @xmath161 by @xmath18 .
same as the previous figure , but here i plot the number density of bhs that have grown by a factor @xmath157 between @xmath162 and @xmath46 , representative of the growth requirements for uv - aided direct collapse seed models . ] figures [ fig : npop3 ] and [ fig : ndc ] show that the @xmath3 quasar population could be explained , for both population iii and direct - collapse seed models , by host merger - triggered accretion near the eddington limit if feeding episodes last for @xmath163 and if the slope of the redshift evolution is @xmath164 .
large values of the scatter @xmath165 in the distribution of feeding times can enhance the number of particularly massive bhs ( i.e. if it is more common for major mergers to trigger long feeding episodes ) .
i have investigated the merger histories of dark matter halos at @xmath3 , focusing on halos that are massive enough ( @xmath138 ) to plausibly host the @xmath3 quasars .
below is a summary of the findings . 1
. the mean major merger rate of the main progenitors of the @xmath78 halos is approximately equal to @xmath114 , all the way up to redshifts @xmath2 .
this approximation is valid within a factor of two , whether one defines a major merger by a mass ratio threshold @xmath107 or by @xmath106 .
this result is consistent with expectations from semianalytic theory and results from large cosmological @xmath74-body simulations .
the steep evolution of the merger rate @xmath96 directly implies that for a wide range of physical bh feeding mechanisms with duration @xmath22 , the duty cycle @xmath166 increases with redshift . while @xmath22 can in general depend on a myriad factors , i took advantage of the fact that the halos of interest here share closely similar assembly histories to conjecture that the feeding times for these halos can be characterized as a function of redshift with a reasonably small scatter .
the formation of the @xmath3 quasar smbhs can be explained without super - eddington accretion , for both popiii and direct collapse seed models , if @xmath22 is greater than several @xmath151 at @xmath18 and scales with redshift as @xmath133 with @xmath167 , for scatter in @xmath22 of @xmath168 dex .
this parameter space includes @xmath169^{-3/2 } { \;\mathrm{myr}}$ ] .
this finding suggests that the smbh growth scenario suggested by li et al .
@xcite may be viable for a plurality of all dark matter halos at this mass scale that host a nuclear bh .
3 . the main progenitors of the @xmath3 quasar hosts become molecular - cooling ( atomic - cooling ) very early , at @xmath170 ( @xmath50 ) , significantly earlier than halos in detailed cosmological simulations that focus on the formation of typical first stars and galaxies , at @xmath171 ( @xmath142 ) . such simulations may not be well - suited for studying the the evolution of the @xmath3 quasar smbhs , as they would underestimate the effects of halo growth and mergers , especially for popiii seed models . when considering the origins of the @xmath3 quasar smbhs ,
it is important to keep in mind that the observed properties of @xmath12 _ galaxies _ are remarkably similar to what is found at later cosmological epochs : their masses are comparable to ( but somewhat lower than ) those of the most massive smbhs and galaxies in the local universe ; the quasar spectra and metallicities appear identical to what is found at lower redshifts @xcite ; they exhibit strong star formation @xcite and winds @xcite associated with post - merger quasar activity . in a similar vein ,
massive galaxies at @xmath172 appear to already have metal - enriched stellar populations with ages of @xmath173 ( e.g. * ? ? ?
* ; * ? ? ?
the masses of both the @xmath3 quasar smbhs and their hosts are comparable to the largest masses of smbhs and galaxies found at lower redshifts . indeed , the simple expectation from the theory of hierarchical structure formation is that the abundance of galaxies approaching the mass ceiling of galaxies @xcite reaches @xmath140 by @xmath3 .
that is , the observation of smbhs at the highest mass scales at @xmath3 is coincident with the emergence of the most massive class of galaxies expected from theory , and observed to be fully evolved
. it also follows from theory that after their relatively early arrival on the cosmic stage , the population of massive galaxies should evolve increasingly slowly as the universe ages at the largest masses due to the inhibition of gas cooling , and in general due to the precipitous drop in the frequency of major mergers ( the triggers of starbursts and smbh growth ) toward low redshifts .
this is qualitatively consistent with observed trends in the emergence and growth of the most massive smbhs and galaxies throughout cosmic time @xcite .
one then wonders : aside from their early formation , what is extraordinary about the @xmath3 quasars and their hosts ? that their hosts are more compact , gas - rich and rapidly merging despite having similar masses to their low - redshift counterparts may result in their bhs having somewhat larger eddington ratios @xcite and being heavily obscured for large periods of time ( in addition to their being possibly obscured at birth @xcite ) .
the latter possibility could negatively affect the detectability of @xmath174 quasars by the james webb space telescope and athena+ at rest - frame uv and x - ray frequencies , respectively , while making large contributions to the cosmic infrared background ( see , e.g. * ? ? ?
* ; * ? ? ?
the correlations between smbhs and galaxy properties could differ from what is found in the local universe ( e.g. * ? ? ?
* ) , depending on the details of the seeding and growth mechanisms @xcite . the formation and evolution of low - mass galaxies and their nuclear bhs may have proceeded quite differently at high redshifts , as gas cooling in small ( proto- ) galaxies is known to be more sensitive to the temperature and ionization state of the inter- and circumgalactic media .
the transition of the igm from neutral @xmath175 gas at @xmath176 , to ionized gas with temperatures comparable to protogalactic virial temperatures and the atomic - cooling threshold at @xmath177 , would have affected star formation and bh growth in low - mass galaxies .
it is interesting to note that the hosts of the @xmath3 smbhs were likely to be the least affected by this upheaval of the intergalactic environment ( e.g. * ? ? ?
* ) . a minimalistic , zeroth - order ansatz would be that once their immediate environments are ionized and provided that they are above the cosmological jeans ( filtering ) mass scale , the formation and evolution of galaxies smbhs , metallicities , winds and all are driven primarily by the gravitational environment of their dark matter halos , at high and low redshift ; that while the earliest galaxies and smbhs at the largest mass scales arise and evolve rapidly , they do so without processes that are either rare or absent ( e.g. highly super - eddington accretion ) in galaxies of similar mass at lower redshift . or , we can turn this ansatz into a question : is there a redshift above which the evolution of massive galaxies and their smbhs is qualitatively different from what is observed at @xmath14 ? just as turner ( 1991 ) @xcite noted for @xmath178 quasars , the formation of the @xmath3 quasars could be explained within the confines of ` conventional cosmic structure formation . ' +
i am grateful to miao li for graciously allowing me to re - use merger tree data from an earlier collaboration .
i thank eugene churazov for stimulating discussions , and zoltn haiman for encouraging conversations and comments on the manuscript .
+ x. fan , v. k. narayanan , r. h. lupton , m. a. strauss , g. r. knapp , r. h. becker , r. l. white , l. pentericci , s. k. leggett , z. haiman , j. e. gunn , .
ivezi , d. p. schneider , s. f. anderson , j. brinkmann , n. a. bahcall , a. j. connolly , i. csabai , m. doi , m. fukugita , t. geballe , e. k. grebel , d. harbeck , g. hennessy , d. q. lamb , g. miknaitis , j. a. munn , r. nichol , s. okamura , j. r. pier , f. prada , g. t. richards , a. szalay , and d. g. york . . ,
122:28332849 , december 2001 .
x. fan , m. a. strauss , d. p. schneider , r. h. becker , r. l. white , z. haiman , m. gregg , l. pentericci , e. k. grebel , v. k. narayanan , y .- s .
loh , g. t. richards , j. e. gunn , r. h. lupton , g. r. knapp , .
ivezi , w. n. brandt , m. collinge , l. hao , d. harbeck , f. prada , j. schaye , i. strateva , n. zakamska , s. anderson , j. brinkmann , n. a. bahcall , d. q. lamb , s. okamura , a. szalay , and d. g. york . . ,
125:16491659 , april 2003 .
d. j. mortlock , s. j. warren , b. p. venemans , m. patel , p. c. hewett , r. g. mcmahon , c. simpson , t. theuns , e. a. gonzles - solares , a. adamson , s. dye , n. c. hambly , p. hirst , m. j. irwin , e. kuiper , a. lawrence , and h. j. a. rttgering . . ,
474:616619 , june 2011 .
n. a. grogin , c. j. conselice , e. chatzichristou , d. m. alexander , f. e. bauer , a. e. hornschemeier , s. jogee , a. m. koekemoer , v. g. laidler , m. livio , r. a. lucas , m. paolillo , s. ravindranath , e. j. schreier , b. d. simmons , and c. m. urry . .
, 627:l97l100 , july 2005 .
m. cisternas , k. jahnke , k. j. inskip , j. kartaltepe , a. m. koekemoer , t. lisker , a. r. robaina , m. scodeggio , k. sheth , j. r. trump , r. andrae , t. miyaji , e. lusso , m. brusa , p. capak , n. cappelluti , f. civano , o. ilbert , c. d. impey , a. leauthaud , s. j. lilly , m. salvato , n. z. scoville , and y. taniguchi . . ,
726:57+ , january 2011 .
g. hinshaw , d. larson , e. komatsu , d. n. spergel , c. l. bennett , j. dunkley , m. r. nolta , m. halpern , r. s. hill , n. odegard , l. page , k. m. smith , j. l. weiland , b. gold , n. jarosik , a. kogut , m. limon , s. s. meyer , g. s. tucker , e. wollack , and e. l. wright . .
, 208:19 , october 2013 . y. shen , m. a. strauss , m. oguri , j. f. hennawi , x. fan , g. t. richards , p. b. hall , j. e. gunn , d. p. schneider , a. s. szalay , a. r. thakar , d. e. vanden berk , s. f. anderson , n. a. bahcall , a. j. connolly , and g. r. knapp . .
, 133:22222241 , may 2007 .
d. richstone , e. a. ajhar , r. bender , g. bower , a. dressler , s. m. faber , a. v. filippenko , k. gebhardt , r. green , l. c. ho , j. kormendy , t. r. lauer , j. magorrian , and s. tremaine .
, 395:a14 + , october 1998 .
p. oliva - altamirano , s. brough , c. lidman , w. j. couch , a. m. hopkins , m. colless , e. taylor , a. s. g. robotham , m. l. p. gunawardhana , t. ponman , i. baldry , a. e. bauer , j. bland - hawthorn , m. cluver , e. cameron , c. j. conselice , s. driver , a. c. edge , a. w. graham , e. van kampen , m. a. lara - lpez , j. liske , a. r. lpez - snchez , j. loveday , s. mahajan , j. peacock , s. phillipps , k. a. pimbblet , and r. g. sharp . . , 440:762775 ,
may 2014 .
v. springel , s. d. m. white , a. jenkins , c. s. frenk , n. yoshida , l. gao , j. navarro , r. thacker , d. croton , j. helly , j. a. peacock , s. cole , p. thomas , h. couchman , a. evrard , j. colberg , and f. pearce . . , 435:629636 , june 2005 .
f. hopkins , d. croton , k. bundy , s. khochfar , f. van den bosch , r. s. somerville , a. wetzel , d. keres , l. hernquist , k. stewart , j. d. younger , s. genel , and c .- p . ma . .
, 724:915945 , december 2010 .
r. wang , j. wagg , c. l. carilli , f. walter , l. lentati , x. fan , d. a. riechers , f. bertoldi , d. narayanan , m. a. strauss , p. cox , a. omont , k. m. menten , k. k. knudsen , r. neri , and l. jiang . .
, 773:44 , august 2013 .
j. s. dunlop , a. b. rogers , r. j. mclure , r. s. ellis , b. e. robertson , a. koekemoer , p. dayal , e. curtis - lake , v. wild , s. charlot , r. a. a. bowler , m. a. schenker , m. ouchi , y. ono , m. cirasuolo , s. r. furlanetto , d. p. stark , t. a. targett , and e. schneider . . ,
432:35203533 , july 2013 .
c. a. collins , j. p. stott , m. hilton , s. t. kay , s. a. stanford , m. davidson , m. hosmer , b. hoyle , a. liddle , e. lloyd - davies , r. g. mann , n. mehrtens , c. j. miller , r. c. nichol , a. k. romer , m. sahln , p. t. p. viana , and m. j. west . . , 458:603606 , april 2009 | the formation mechanism of supermassive black holes ( smbhs ) in general , and of @xmath0 smbhs observed as luminous quasars at redshifts @xmath1 in particular , remains an open fundamental question .
the presence of such massive bhs at such early times , when the universe was less than a billion years old , implies that they grew via either super - eddington accretion , or nearly uninterrupted gas accretion near the eddington limit ; the latter , at first glance , is at odds with empirical trends at lower redshifts , where quasar episodes associated with rapid bh growth are rare and brief . in this work
, i examine whether and to what extent the growth of the @xmath1 quasar smbhs can be explained within the standard quasar paradigm , in which major mergers of host galaxies trigger episodes of rapid gas accretion below or near the eddington limit . using a suite of monte carlo merger tree simulations of the assembly histories of the likely hosts of the @xmath1 quasars ,
i investigate ( i ) their growth and major merger rates out to @xmath2 , and ( ii ) how long the feeding episodes induced by host mergers must last in order to explain the observed @xmath3 quasar population without super - eddington accretion . the halo major merger rate scales
roughly as @xmath4 , with quasar hosts typically experiencing @xmath5 major mergers between @xmath6 ( @xmath7 ) , compared to @xmath8 for typical massive galaxies at @xmath9 ( @xmath10 ) .
an example of a viable sub - eddington smbh growth model is one where a host merger triggers feeding for a duration comparable to the halo dynamical time .
these findings suggest that the growth mechanisms of the earliest quasar smbhs need not have been drastically different from their counterparts at lower redshifts . |
the totem experiment at lhc provides , in fact , the first measurements of soft pomeron contribution ( more exactly of pomeron and odderon ) because the contributions of secondary or subasymptotic reggeons are very small at such high energies .
it means that the precise measurement of @xmath1 total and differential cross sections @xcite gives a possibility to discriminate the various pomeron models by comparing their predictions with new data . generally speaking the totem results are going to be a strong restriction for various phenomenological pomeron models @xcite .
many of them being directly extrapolated to 7 tev could not describe the recent totem data though they are in a good agreement with data at @xmath61.96 tev . moreover , even basing on the total cross section measurement at 7 and 8 tev only one can conclude that @xmath7 rises with energy faster than @xmath8 .
therefore some of the models have to be rejected despite its excellent agreement with the data available before lhc epoch .
for example , the dipole pomeron model with trajectory intercept @xmath9 ( see @xcite and references therein ) is unable to give @xmath10 mb at @xmath11 tev .
the current state of elastic scattering model market is given in @xcite and in the comprehensive review @xcite .
obviously , all the models must be verified for a consistency with new data .
many of them , if not the all , require some modification to be successfully extrapolated to new data . in @xcite
such a procedure was performed for the eikonal @xcite and @xmath12-matrix @xcite models .
it was shown that adding the second odderon term solves the problem only partially , a further modification of the model is necessary .
eikonal and @xmath12-matrix models are treated as infinite series of multiple input reggeon exchanges .
however another approach to constructing model is also possible .
one can just take into account unitarity and analyticity requirements from the beginning as well as experimental information on the cross sections ( e.g. increase of total cross sections ) to determine the form of amplitude .
following this idea in @xcite we have suggested the tripole pomeron - odderon model of elastic @xmath1 and @xmath2 scattering amplitudes and compared it with data .
amplitudes in the model at very high energies are dominated by pomeron contribution which corresponds to pair of the branch points colliding at @xmath0 and at the angular momentum @xmath13 to generate a triple pole . besides the set of the preasymptotic terms , even and odd ,
have been incorporated in the model .
structure of leading singularities is different from those of the well - known the maximal pomeron and odderon model @xcite .
it was shown that the model @xcite is in a good agreement with the data in a wide interval of energy and momenta transferred , but now we have found that those model without changes fail to describe the recent totem @xmath14 data . in the present paper we are focused on our parametrization of the tripole pomeron - odderon model @xcite exploring its simplified form ( neglecting the cut contributions unimportant at high energies ) .
this model in ref .
@xcite was applied at @xmath15 6 gev@xmath5 .
we add two additional terms decreasing at high @xmath16 as @xmath17 in order to extend model for larger @xmath16 .
such terms allow us ( as it was shown in @xcite ) to describe the @xmath14 data at @xmath18 6 gev@xmath5 ) .
we remind the important properties of the tripole pomeron - odderon model , argue its modification for high @xmath16 , then we present the results of the fit and give predictions for the totem measurements at higher energies and @xmath16 .
as it was stressed in @xcite the contribution to the partial amplitude of a triple pole with linear trajectory @xmath19^{3}}\ ] ] violates the unitarity inequality @xmath20 .
the correct pomeron singularity dominating the partial wave is the following @xmath21^{3/2}}.\ ] ] the black disk model holds the same singularity of the partial amplitude .
the imaginary part of black disk amplitude in the impact parameter representation has a step - like form with the height 1/2 and the width equal to @xmath22 .
it leads to @xmath23 gev@xmath5 .
it seems to be natural to keep the same structure of singularity for subleading terms changing only their multiplicity .
our choice of subleading terms is the following @xmath24},\ ] ] @xmath25^{1/2}},\ ] ] @xmath26 thus the leading pomeron contribution to partial amplitude has a form @xmath27 taking into account that @xmath28 @xmath29 @xmath30 where @xmath31 are the bessel functions , one can write the main pomeron part of amplitude in the @xmath32-representation @xmath33 where @xmath34 are the vertex functions with @xmath35 , @xmath36 is defined below by exp.([eq : zvar ] ) , and @xmath37 , @xmath38 gev@xmath5 , @xmath39 and @xmath40 are constants . strictly speaking the amplitudes in regge models depend on @xmath41 through the @xmath42 , cosine of the scattering angle in @xmath43-channel , which for @xmath1 scattering in the center mass system has the form @xmath44 in the considered kinematical region the @xmath43 in the numerator of eq .
[ eq : costetha ] can be neglected .
however we keep @xmath45 because for @xmath0 we take into account the data on cross sections at low energies , @xmath46 5 gev . absorbing factor @xmath47 into the vertex functions we define the energy variable as the following @xmath48 we propose to keep the same singularity structure for the leading terms of crossing - odd part of amplitude .
however , taking into account the fact that there is no visible odderon contribution at @xmath0 we multiply each odderon term by factor @xmath43 @xmath49 where @xmath50 and @xmath51 are the constants .
we would like to emphasize that similar ( but not the same ) models for @xmath52 and @xmath53 were considered in @xcite .
these models have the same dominating even term @xmath54 but the different subasymptotic terms @xmath55 @xmath56 .
the odderon terms included in those amplitudes do not vanish at @xmath0 .
the properties of the models @xcite and their defects have been discussed in detail in @xcite .
they were modified in @xcite , defects were eliminated and a good description of data at energies up to 1.8 tev has been obtained . however , as was noticed in @xcite such a maximal odderon term ( @xmath57 in our notations ) gives rise to the contradiction with unitarity in the models where @xmath58 at @xmath59 . in despite of this fact we refit the model @xcite ( without the cuts important only at low energies ) at the energies @xmath60 19 gev , including totem data , but we are fail to describe the data qualitatively . in @xcite
the described model was applied to analysis of the @xmath14 data at @xmath60 6 gev , where not only contributions of the pomeron and secondary reggeons but also their rescatterings ( or cuts ) are very important . besides , the considered momenta transfer were restricted by @xmath61 gev@xmath5 . here
we would like to check a principal possibility of the model to describe an interpolation between gev and tev energy region .
we investigate as well which amplitude terms are important for that and which form of vertex functions @xmath62 is more suitable for @xmath14 at @xmath63 gev@xmath5 .
thus we consider high energy @xmath1 and @xmath2 elastic scattering , starting from fnal energy 19 gev .
we do hope that in this energy interval we can neglect at least all cuts reducing the number of adjustable parameters .
these terms are important at lower energies but their parametrization should be chosen being consistent with good description of high energy data .
we keep in amplitudes also the standard `` soft '' pomeron and odderon , simple @xmath64-poles with linear trajectories @xmath65 @xmath66 @xmath67 to describe the behaviour of the @xmath68 at low energy and low @xmath16 the usual secondary crossing - even and odd reggeons ( @xmath69 reggeons with intercepts @xmath70 ) have to be included in the amplitudes .
however , analyzing @xmath1 and @xmath2 amplitudes only it is sufficient to consider at @xmath60 5 gev one effective even reggeon and one effective odd reggeon @xmath71 a preliminary analysis of the data was performed to understand how a behavior of @xmath14 ( at @xmath63 gev@xmath5 ) can be described .
we found that a correct description of these data is provided by the terms behaving as @xmath72 at large @xmath16 ( which are almost independent on @xmath41 ) .
therefore we add to the amplitudes crossing even and odd terms @xmath73 and @xmath74 with arbitrary phases .
to decrease an influence of these terms on the amplitudes at small @xmath16 we introducing the factor @xmath75
. it would be reasonable to construct them in a more customary ( regge - like ) form at small @xmath43 and power - like behaving at large @xmath43 .
but as we noticed above our aim is , first of all , to check some principal possibilities of the tripole pomeron and odderon models .
therefore to avoid an extra number of parameters we choose these terms in a simplified form @xmath76 @xmath77 thus the @xmath1 and @xmath2 amplitudes in a general case are defined as following @xmath78 where @xmath79 are given by eqs .
( [ eq : tripom - st ] , [ eq : triodd - st ] , [ eq : softpom ] , [ eq : softodd ] , [ eq : sec - reg ] , [ eq : power+ ] , [ eq : power- ] ) correspondingly .
we have considered two options for the vertex functions @xmath80 in the pomeron ( @xmath81 ) and odderon ( ( @xmath82 ) ) terms .
* model i * , exponential form @xmath83 * model ii * , power form @xmath84 the behaviour with -5 for odderon vertices is taken because odderon terms of amplitudes ( eqs.([eq : triodd - st],[eq : softodd ] ) ) have been multiplied by the factor @xmath43 . in both models
the vertices for secondary reggeons @xmath85 are chosen in an exponential form @xmath86 @xmath85 are negligible at high @xmath16 because of large slopes of their trajectories .
however these terms of amplitude become important at small @xmath16 where exponential vertices are reasonable .
the following normalization of @xmath87 amplitude is used @xmath88 where @xmath89 and @xmath90 is the momentum of initial proton ( antiproton ) in laboratory system of another proton .
the amplitudes and couplings @xmath91 are dimensionless with this normalization .
the constructed models were compared with the @xmath1 and @xmath2 data on @xmath92 and @xmath93 in the following region of @xmath41 and @xmath43 @xmath94 the cosmic ray data on the @xmath1 total cross sections were not included in the fit procedure .
the data set we used for adjusting the model parameters has been proposed in @xcite , where a coherent set of all existing data for @xmath95 gev and @xmath96 gev@xmath97 has been built .
a detailed study of the systematic errors of the data from more than 260 subsets of the data from more than 80 experimental papers have performed ( original data and the corresponding references can be found in the hep data system @xcite ) .
the corrected data are collected and written in a common format .
we suggest to use this data set as a _
standard data set_. the latest ( with some corrections ) updated version of it including totem @xcite and d0 @xcite data is available online @xcite .
the set used for the given analysis contains 2384 points in the region described above ( numbers of points for measured quantities are given in the table [ tab : exppoints ] ) .
31 points from the 3 groups , 8 points at @xmath9826.946 gev , 11 points at 30.7 gev and 12 points at 53.018 gev only for @xmath99 were excluded from the all data presented in @xcite because these groups are strongly deviated from the rest data points and can slightly distort the fit .
.number of experimental points used for the fitting [ cols="^,>,>,>",options="header " , ] as one can see from the figs .
[ fig : sigtot]-[fig : pap - dsdt ] and from the table [ tab : chi2 ] all models , even the simplest one with @xmath100 , describe quite well the experimental data .
the theoretical curves in the models differ in some details but nevertheless they are globally in agreement with experimental data .
moreover , one can obtain very similar description of the data with the following vertices @xmath101 this first analysis gives a ground for three important inferences .
* there are no any unusual or unexpected phenomena in new totem data .
they correspond to the standard regge - like behaviour at low and intermediate @xmath16 . * the data at relatively low energies but at highest measured @xmath16 require in amplitude the terms decreasing like @xmath72 .
it can be an indication of a hard scattering .
however , taking into account that slow decreasing with @xmath43 is obtained in the eikonal and @xmath12-matrix models @xcite , one can think that a power - like behaviour may be imitated by series of rescatterings or multi - reggeon exchanges . *
we believe that successful description of the data in the considered models results from the well tuned structure of pomeron and odderon singularities ( eqs .
( [ eq : tripom - st],[eq : triodd - st ] ) ) rather than from a choice of vertex functions .
the theoretical curves for @xmath102 and integrated elastic cross sections , @xmath103 , as well as for inelastic cross section , @xmath104 , are shown in figs .
[ fig : sigtot],[fig : rho],[fig : sigint ] .
the values of the total cross sections obtained at 7 and 8 tev are little bit less than the totem data but they are within the measured errors .
the both models predictions for higher lhc energies are given in the tables [ tab : lhc - predict - e ] , [ tab : lhc - predict - p ] and @xmath2 total cross sections ] and @xmath2 amplitudes ] cross sections ] description of the differential cross section in the models is demonstrated in the figs .
[ fig : pp - dsdt ] , [ fig : pap - dsdt ] . for the simplest model , i.e. at @xmath100 , we have shown curves only for model with exponential vertices in order to avoid a meshing figure . from the table [ tab : chi2 ]
one can see that quality of description in the both models is almost the same .
the totem @xmath14 data are described with @xmath105 and @xmath106 in the models i and ii ( with @xmath107 ) , correspondingly . elastic scattering ] elastic scattering ] we would like to stress that in despite of a widespread opinion that at high energy and high momentum transfers an odderon contribution is dominating we found that in the considered models it is not the case .
one can see in figs .
[ fig : evenodd ] , [ fig : isrlhcpart ] that the dominating partial even and odd components at large @xmath16 have comparable values ( at least for @xmath10815 gev@xmath5 ) . however , fig .
[ fig : isrlhcpart ] shows that cumulative even contribution to @xmath109 at @xmath16 outside of the dip positions ( see below ) is larger of the odd one .
the cumulative even contribution in this region at 7 tev is few times larger than odd one .
at the same time a role of odd contributions is important at low and intermediate @xmath43-values and , especially , in the regions of dips .
elastic scattering amplitudes at 53 gev and 7 tev ( calculated in the model ii ) .
factor @xmath110 is defined in eq .
[ eq : norma ] . ]
+ at 53 gev and 7 tev ( calculated in the model ii ) .
factor @xmath110 is defined in eq .
[ eq : norma ] . ] in fig .
[ fig : lhc - dsdt ] we give the predictions of the considered models for the totem experiment at higher transferred momenta and higher lhc energies .
the most interesting point is an existing / absence of the second dip in @xmath14 .
model with exponential formfactors leads to smooth behaviour with no the dips .
whereas , models with power vertices predict a dip structure in @xmath14 . if @xmath100 then the shoulder at @xmath111 gev@xmath5 and @xmath98 7 tev is transformed into a dip moving to @xmath112 gev@xmath5 at 14 tev .
moreover , in this model the third dip is visible at the energy 14 tev near @xmath113 10 gev@xmath5 . if @xmath107 then well pronounced dip develops at @xmath114 8 gev@xmath5 moving to @xmath114 6 gev@xmath5 when energy is increasing from 7 to 14 gev .
accurate measurements at such @xmath16 would be the excellent test for the considered models . [ h ! ] differential cross sections at the lhc energies ]
analysis performed above demonstrated a high credibility of the tripole pomeron - odderon model developed in @xcite . we have made its minor improvement only by adding two terms , crossing even ( @xmath115 ) and odd ( @xmath74 ) , in order to apply the model to high @xmath43 .
we have considered two choices of @xmath43-dependence in vertex functions and found out in all the cases a good description of the data in wide region of @xmath41 and @xmath43 even in the simplified versions of the model without the standard simple pole pomeron and odderon contributions .
thus , it allow us to conclude that the model is quite stable under variations of the preasymptotic components and the form of vertex functions .
apparently such the model stability is stipulated by the well tuned structure of the leading pomeron @xmath52 and odderon @xmath53 singularities .
we have predicted the values of the total , elastic and inelastic cross sections , as well as the ratio @xmath116 and differential cross sections at higher lhc energies .
the amizing prediction is made for large @xmath16-region at the lhc energies .
model with an exponential @xmath43-dependence of vertices leads to a smooth behaviour of differential @xmath1 cross sections while in the model with power vertices a dip structure moving with energy is generated at large @xmath16 .
we hope that future measurements of totem collaboration will allow us to discriminate within possibilities for the developed model .
i d like to thank my colleagues j.r .
cudell , v. petrov , a. prokudin , j.p .
revol , s. troshin and g. zinovjev for numerous fruitful discussions of the regge approaches and high - energy models , for reading the manuscript and valuable remarks .
g. antchev _ et al_. , totem collaboration , europhys .
* 95 * ( 2011 ) ( 41001 ) .
g. antchev _ et al_. , totem collaboration , europhys
* 101 * ( 2013 ) ( 21002 ) .
g. antchev _
et al_. , totem collaboration , europhys .
* 101 * ( 2013 ) ( 21004 ) .
e. martynov , phys . rev . *
d76 * , 074030 ( 2007 ) , [ arxiv : hep - ph/0703248 ] .
godizov , aip conf . proc . * 1523 * 145 ( 2012 ) , [ arxiv:1203.6013 ] .
dremin , phys.usp . * 56 * 3 ( 2013 ) .
a.alkin , o. kovalenko , e. martynov , europhysics letters , * 102 * ( 2013 ) 31001 , [ arxiv:1304.0850 ] .
petrov and a.v .
prokudin , eur .
j. , * c23 * , ( 2002 ) ( 135 ) , [ arxiv : hep - ph/0105209 ] . v.a .
petrov and a.v .
prokudin , e - print : arxiv:1212.1924 , 2013 .
troshin and n.e .
tyurin , phys .
* 35 * 555 ( 2004 ) ; [ arxiv : hep - ph/0308027 ] and references therein .
p. gauron , e. leader and b. nicolescu , phys . lett .
* b238 * 406 ( 1990 ) .
r. avila , p. gauron and b. nicolescu , eur .
j. * c49 * 581 ( 2007 ) [ arxiv : hep - ph/0607089 ] .
evgenij martynov , basarab nicolescu , eur.phys.j .
* c56 * 57 ( 2008 ) , [ arxiv:0712.1685 ] .
troshin , phys.lett .
* b682 * 40 ( 2009 ) , [ arxiv:0903.4042 ] j.r .
cudell , a. lengyel and e. martynov , phys . rev . *
d73 * 034008 ( 2006 ) [ arxiv : hep - ph/0511073 ] .
j. beringer _ et al_. , phys .
, * d86 * ( 2012 ) ( 010001 ) , http://pdg.lbl.gov/2012/hadronic - xsections/. durham database group ( uk ) , m.r . whalley _
et al_. , http://durpdg.dur.ac.uk/hepdata/ v.m .
et al . _ ( the d0 collaboration ) . phys.rev . *
d86 * ( 2012 ) 012009 , [ arxiv:1206.0687 ] .
file with the data `` alldata-v2.zip '' is available at the address : http://www.theo.phys.ulg.ac.be/alldata-v2.zip | the model of elastic scattering amplitudes dominated by the triple ( at @xmath0 ) pomeron pole suggested earlier is modified to confront to existing experimental data on @xmath1 and @xmath2 total and differential cross sections at @xmath3 gev and @xmath4 gev@xmath5 including the newest totem data .
predictions for the future totem measurements at 13 and 14 tev are given . |
recently , graphene monolayers have attracted much attention motivated by their unusual twodimensional ( 2d ) dirac energy spectrum of electrons . in graphite , which is a graphene multilayer composed of weakly coupled graphene sheets , the interlayer interaction converts the 2d electron energy spectrum of graphen into the three - dimensional ( 3d ) spectrum of graphite .
the application of the tilted magnetic field is a typical method to distinguish between 2d and 3d electron systems , as in 3d systems the orbital effect of the in - plane magnetic field should be observable .
this problem has been touched in two recent theoretical studies of the graphene multilayer energy spectrum in magnetic fields parallel to the layers @xcite , and of the landau levels ( lls ) of bilayer graphene in magnetic fields of arbitrary orientation @xcite .
both papers conclude that the very strong in - plane field components are necessary to induce observable effects on the electronic structure . here
we show that at the h point of the hexagonal brillouin zone of graphite the application of the tilted magnetic field leads to experimentally observable splitting of lls .
bulk graphite is composed of periodically repeated graphene bilayers formed by two nonequivalent bernal - stacked graphene sheets .
there are two sublattices , a and b , on each sheet and , therefore , four atoms in a unit cell .
the distance between the nearest atoms a and b in a single layer is @xmath0 , the interlayer distance between nearest atoms a is @xmath1 . to describe the graphite band structure
, we employ the minimal nearest - neighbor tight - binding model introduced in ref .
@xcite and successfully applied in refs .
the wave functions are expressed via four orthogonal components @xmath2 , @xmath3 , @xmath4 , @xmath5 , which are , in zero magnetic field , bloch sums of atomic wave functions over the lattice sites of sublattices a and b in individual layers .
the tight - binding hamiltonian @xmath6 includes only the intralayer interaction @xmath7 between nearest atoms a and b in the plane , and the interlayer interaction @xmath8 between nearest atoms a out of plane .
the continuum approximation is used in the vicinity of the @xmath9 axis of the graphite brillouin zone , for small @xmath10 measured from the axis . then the electron wavelength is larger than the distance between atoms , and the non - zero matrix elements of @xmath6 can be written as @xmath11 the fermi velocity , @xmath12 , is defined by @xmath13 and will be used as an intralayer parameter instead of @xmath14 in the subsequent consideration .
when solving the schrdinger equation for the above four wave functions , we can make use of the structure of @xmath6 , which allows to express the function @xmath18 via the function @xmath19 from the same layer , and leaves us with two , , interlayer
equations for @xmath19 and @xmath20 .
note , that graphite is no longer periodic in the @xmath21-direction ( parallel to the @xmath22-axis ) but becomes periodic in the direction of the tilted magnetic field .
here we apply an approach developed in @xcite .
the new periodicity implies that @xmath23 can be written as @xmath24 @xmath25 . here
@xmath26 and @xmath27 can be associated with cyclotron orbits in two layers .
the dimensionless variable @xmath28 is defined by @xmath29 , where @xmath30 is the cyclotron orbit center , @xmath31 , and @xmath32 is the magnetic length , @xmath33 .
the small parameter @xmath34 means the shift of the cyclotron orbit center in the @xmath35-layer due to the in - plane component of the magnetic field , @xmath36 . introducing the shift operator by @xmath37 and employing @xmath38-representation , two ,
, interlayer equations can be given the form @xmath39 \phi_2^a+ 2 t\ , e\cos{(\kappa\eta_d+k_zd ) } \phi_1^a= 0 , \label{eq1tt } \\ \label{main - eq2 } \left[\frac{\mathcal{b}}{2}\left(-\frac{\partial^2}{\partial\kappa^2 } + \kappa^2 -1\right)-e^2\right ] \phi_1^a+ 2 t\ , e\cos{(\kappa\eta_d+k_zd ) } \phi_2^a = 0,\label{eq2tt } \end{aligned}\ ] ] where @xmath40 stands for @xmath41 .
( [ main - eq1],[main - eq2 ] ) represent the main result of this work .
it is evident that @xmath42 and @xmath43 can not be too far from the eigenfunctions of the harmonic oscillator , @xmath44 .
if we write @xmath45 we arrive to @xmath46 a_{2,n}+ 2 e \sum_{n'=0}^{\infty}t_{n , n'}a_{1,n'}=0,\label{eq1p } \\
\left[\mathcal{b}n - e^2\right ] a_{1,n}+ 2 e \sum_{n'=0}^{\infty}t_{n , n'}a_{2,n'}=0,\label{eq2p}\end{aligned}\ ] ] where @xmath47 in the perpendicular magnetic field @xmath48 and @xmath49 .
this corresponds to the previously discussed case @xcite . at the @xmath50 point ,
@xmath51 , we get the energy spectrum of an effective bilayer . at the @xmath52 point ,
@xmath53 , the coupling between layers disappears , and we obtain the lls corresponding to graphene dirac fermions , namely @xmath54 for the first layer , and @xmath55 for the second layer , @xmath56 .
let us discuss the magnetic field of an arbitrary direction .
it is obvious that the influence of the small parameter @xmath57 is negligible at the point k when @xmath51 and @xmath58 .
this case corresponds to the bilayer subject to the tilted field discussed in ref .
@xcite , with the result that corrections induced by @xmath36 are negligible .
let us consider the influence of @xmath36 at the h point , when @xmath59 and @xmath60 . then , unlike the case of the perpendicular field , the interlayer interaction is not reduced to zero , but remains finite .
note , that the small perturbation @xmath57 couples the unperturbed states @xmath61 and @xmath62 , which belong to the same unperturbed eigenvalues @xmath63 .
consequently , the perturbation approach suitable to describe the coupling of these degenerated states must be applied , which yields equations @xmath64 a_{2,n}+2 e t_{n , n+1 } a_{1,n+1}=0,\label{eq1pp } \\
\label{eq2 } \left[\mathcal{b}(n+1)-e^2\right ] a_{1,n+1}+2 e t_{n+1,n}a_{2,n}=0,\label{eq2pp}\end{aligned}\ ] ] where @xmath65.\ ] ] the secular equation derived from eqs.([eq1],[eq2 ] ) reads @xmath66 ^2 - 4 e^2 t^2_{n , n+1}=0,\ ] ] and from here we get the four eigenenergies @xmath67 the eigenenergies , @xmath68 and @xmath69 , originated from @xmath70 remain the same as in the perpendicular magnetic field . in that case
the degeneracy is not removed .
the authors benefited from discussions with milan orlita .
the support of the european science foundation epigrat project ( gra/10/e006 ) , av cr research program avoz10100521 and the ministry of education of the czech republic project lc510 is acknowledged . | the minimal nearest - neighbor tight - binding model with the peierls substitution is employed to describe the electronic structure of bernal - stacked graphite subject to tilted magnetic fields .
we show that while the presence of the in - plane component of the magnetic field has a negligible effect on the landau level structure at the k point of the graphite brillouin zone , at the h point it leads to the experimentally observable splitting of landau levels which grows approximately linearly with the in - plane field intensity .
address = institute of physics , academy of science of the czech republic , v. v. i. , + cukrovarnick 10 , 162 53 praha 6 , czech republic |
the one - point probability distribution function ( pdf ) of a random mass density field @xmath3 , or the counts - in - cells ( cic ) statistics of a discrete random distribution such as galaxies , is probably the first statistics used to reveal the clustering feature of galaxy distribution . in the famous work by edwin hubble ( 1934 ) , he showed that the frequency distribution of galaxy count @xmath4 in angular cells is not gaussian , but lognormal .
this result indicates that the pdf of galaxy distribution is fundamental in characterizing the cosmic mass and velocity fields .
although current samples applied for large - scale structure study have much deeper redshift and much wider angular size than available in hubble s time , the one - point statistics is still frequently applied .
this is because the one - point distribution and its moments contain complete information of the field , which might not be easily detected by other conventional methods . unlike the fourier amplitude ,
cic is not subject to the central limit theorem .
it can detect the non - gaussianity of a field consisting of randomly distributed non - gaussian clumps , while the pdf of the fourier amplitudes is still gaussian due to the central limit theorem ( fan and bardeen 1995 ) .
even the 2nd moment of one - point distribution is different from the 2nd moment of the fourier decomposition power spectrum .
the former contains perturbations on scales larger than the size of the observed sample , while the latter does not .
therefore , one - point statistics is applied on various samples of large - scale structures , including galaxy surveys ( e.g. hamilton 1985 ; alimi et al 1990 ; gaztaaga 1992 ; szapudi et al 1996 ; kim & strauss 1998 ) , transmitted flux of quasars ly@xmath5 absorption spectrum ( e.g. meiksin & bouchet 1995 ; gaztanaga & croft 1999 ; zhan & fang 2002 ) , and n - body simulation samples ( e.g. coles & jones 1991 ; taylor & watts 2000 ) .
one - point statistics of a density fluctuation field @xmath6 is given by the distribution of the one - point variable @xmath7 , which is a sampling of the field by a window function @xmath8 around position @xmath9 and on scale @xmath10 , i.e. @xmath11 the variable @xmath7 is actually a mean value of the field at position @xmath9 and on scale @xmath10 .
the distribution of @xmath7 gives a pdf description of the field .
eq.(1 ) is a space(@xmath9)-scale(@xmath10 ) decomposition of the field @xmath6 with bases @xmath12 .
most popular windows @xmath8 are gaussian and top - hat filters .
the gaussian windows generally are not orthogonal , i.e. the function @xmath8 does not satisfy @xmath13 , if @xmath14 .
thus , the variables @xmath7 are either incomplete or redundant . in turn
, these may lead to 1 ) loss of information of the field @xmath6 ; and 2 ) false correlation .
it is possible to construct an orthogonal and complete set of bases from top - hat windows .
however , they are not localized in the fourier space
. the index @xmath10 does not refer to a well defined scale @xmath15 . as a consequence ,
the one - point statistics with conventional windows are not suitable for problems with a scale - dependence .
for instance , the redshift distortion on the one - point statistics can be properly estimated only if we know how the redshift distortion depends on the scale and shape of the window function .
therefore , a better algorithm for the one - point statistics is needed , for example , to recover the real - space rms density fluctuation @xmath16 from redshift distortion . in this paper , we show that these problems can be solved with a multiresolutional decomposition via the discrete wavelet transform ( dwt ) .
the bases of the dwt decomposition are complete , orthogonal , and localized in both physical and fourier spaces .
it was shown in the last few years that the dwt can be employed as an alternative representation in most conventional statistics of cosmic mass and velocity fields , including power spectrum ( pando & fang 1998 ; fang & feng 2000 ; yang et al 2001a , 2002 ) , high order correlations ( pando & fang 1998 ; pando , feng & fang 2001 ; feng , pando & fang 2001 ) , bulk and pair - wise velocity ( zhan & fang 2002 ; yang et al .
2001b ) , and identification of halos and clusters ( xu , fang & wu 2000 ) .
the dwt representation is also able to reveal statistical features , which might not be easily detected without a set of decomposition bases localized in both physical and scale spaces . with the dwt analysis ,
the intermittency of the fluctuations of quasars ly@xmath5 transmitted flux has been detected ( jamkhedkar , zhan & fang 2000 ; pando , feng & fang 2000 ; zhan , jamkhedkar & fang 2001 ; pando et al .
this property can not be simply detected by the popular @xmath8 or fourier decompositions .
we also show that the dwt decomposition is very useful for one - point statistics .
we establish the algorithm of the one - point variable and its moments in considering various corrections , such as poisson sampling and selection function . to demonstrate the advantage of the dwt one - point statistics , we show that one can map between the one - point statistics in real and redshift spaces scale - by - scale , taking account the distortion due to bulk velocity , velocity dispersion and selection function . with these results
one can construct @xmath1 ( redshift distortion parameter ) estimators with moments of the dwt one - point variables .
unlike conventional estimators , the dwt estimators do not need extra - information or _
ad hoc _ assumption when the non - linear redshift distortion is not negligible .
this paper is organized as follows .
2 presents the algorithm for the pdf of one - point variables with the dwt space - scale decomposition .
3 discusses the second moment of the one - point variables in considering the effects of poisson sampling and selection function . in
4 we develop the theory of redshift distortion of the dwt one - point variables .
5 tests the theoretical results in 4 on n - body simulation samples .
the emphases are the redshift - to - real - space mapping of the diagonal and off - diagonal second moments .
6 shows the application in estimating the redshift distortion parameter . finally , the conclusions and discussions are given in 7 .
the appendix provides relevant formulae for quantities defined in 4 .
we release the codes for calculating these quantities via anonymous ftp at samuri.la.asu.edu/pub/zhan/dwtcic.tgz .
as emphasized in 1 , a key problem for one - point statistics is to have proper window functions @xmath17 , and define the variable with eq.(1 ) . in the dwt analysis
, this is done by the so - called scaling function .
let us first briefly introduce the dwt - decomposition of a random field .
for the details of the mathematical properties of the dwt see mallat ( 1989a , b ) ; meyer ( 1992 ) ; daubechies ( 1992 ) , and for physical applications see fang & thews ( 1998 ) . to simplify the notation
, we first consider a one - dimensional ( 1-d ) density fluctuation @xmath18 on a spatial range from @xmath19 to @xmath20 .
we first divide the space @xmath20 into @xmath21 segments labeled by @xmath22 .
each segment is of size @xmath23 .
the index @xmath24 can be a positive integer .
it stands for a length scale @xmath23 . the higher the @xmath24 , the smaller the length scale .
even though we often refer some properties to a ( range of ) @xmath24 in the analyses below , it must be read as an association of the properties with the length scale @xmath23 .
the index @xmath25 is for position , and it corresponds to the spatial range @xmath26 .
that is , the space @xmath20 is decomposed into cells @xmath27 . in the dwt analysis ,
each cell @xmath27 supports two compact functions : the scaling function @xmath28 and the wavelet @xmath29 .
one example of such functions is the daubechies 4 ( d4 ) wavelets , which have to be constructed recursively ( daubechies 1992 ; see also fang & thews 1998 ) .
the properties of d4 wavelets are listed below and in the appendix a.1 .
we show , in fig . 1 , examples of the d4 scaling function and wavelet , and their fourier transform @xmath30 and @xmath31 .
one can see from fig .
1 that both the scaling function and the wavelet are localized in fourier space as well as in physical space .
generally , @xmath28 and @xmath29 are localized in cell @xmath27 , @xmath30 is localized in the scale range @xmath32 , and @xmath31 in the range @xmath33 , where @xmath34 .
obviously , the scaling function @xmath28 is a window function on scale @xmath24 and around the segment @xmath25 . it can be used to measure the mean field in cell @xmath27 @xmath35 where @xmath36 is called scaling function coefficient ( sfc ) , given by @xmath37 in analogous to eq.(1 ) , eqs.(2 ) and ( 3 ) show that the sfc @xmath36 or @xmath38 can be employed as the variable for one - point statistics .
1 also demonstrates that wavelets are admissible , i.e. @xmath39 , and therefore , they are used to measure the fluctuations of a field with respect to its mean in cells @xmath27 @xmath40 where @xmath41 is called the wavelet function coefficient ( wfc ) .
the scaling function @xmath28 and the wavelet @xmath29 satisfy a set of orthonormal relations as @xmath42 where @xmath43 is the kronecker delta function . to be consistent with eq.(5 ) , the scaling function @xmath28 is normalized to @xmath44 with these properties , a 1-d density field @xmath18 can be decomposed into @xmath45 the second term on the r.h.s of eq.(9 ) contains only the information of the field @xmath18 on scales equal to and less than @xmath23 . therefore
, the @xmath21 variables ( sfcs ) @xmath36 , with @xmath46 , completely describe the behavior of the field on scales larger than @xmath23 .
for instance , if the resolution of the sample is @xmath47 , @xmath18 can be described by the @xmath48 sfcs @xmath49 .
because of the orthogonal relation eqs.(5 ) and ( 7 ) , the @xmath21 variables ( sfcs ) @xmath36 are independent and irredundant . in a functional space consisting of functions , which are smoothed on scales equal to and less than @xmath23 , the completeness of the scaling function @xmath28 can be expressed as @xmath50 where @xmath51 is the dirac delta function .
all the above - mentioned results can be easily generalized to three - dimensional ( 3-d ) fields .
let us consider a 3-d distribution @xmath6 in a volume @xmath52 to @xmath53 .
similar to the 1-d case , we divide the volume @xmath54 into cells @xmath55 , where @xmath56 refers to the length scale @xmath57 , and @xmath58 refers to the spatial range of the cell , @xmath59 and @xmath60 , where @xmath61 .
the one - point statistical variable of the field @xmath6 is @xmath62 where the 3-d scaling function ( 3-d window ) @xmath63 is given by a direct product of 1-d scaling functions as @xmath64 one can also construct 3-d wavelets @xmath65 as a direct product of 1-d scaling functions and wavelets ( see yang et al 2002 ) , and generalize eq.(9 ) into 3-d as @xmath66 since the second term on the r.h.s .
is not needed in calculating the one - point statistics , therefore , we do not show the details of @xmath67 , but only give the orthogonal relations @xmath68 if we consider only functions smoothed on scales equal to and less than the scale of @xmath69 in each dimension , the set of bases @xmath63 is orthogonal and complete . in one - point statistics ,
the statistical feature of the field @xmath6 is characterized by the moments of the distribution of @xmath70 .
the @xmath71th moment is @xmath72 , where @xmath73 stands for an ensemble average . using eq.(11 ) , we have @xmath74 for a homogeneous field , these moments are @xmath75-independent .
moreover , when the fair sample hypothesis " ( peebles 1980 ) holds , or equivalently , when the random field is ergodic , the @xmath76 sfcs @xmath70 , @xmath60 , @xmath61 for a given scale @xmath69 can be considered as @xmath76 independent measurements , because they are measured by projecting onto the mutually orthogonal basis @xmath63 .
accordingly , the @xmath76 sfcs form a statistical ensemble on the scale @xmath69 .
this ensemble represents actually the one - point distribution of @xmath70 over the dwt modes at a given scale @xmath69 .
thus , the average over @xmath75 is a fair estimation of the ensemble average , i.e. @xmath77
we now consider the 2nd moment @xmath78 . from eq.(15 ) , the 2nd moment is determined by the usual two point correlation function @xmath79 .
we define a dimensionless 2nd moment on scale @xmath69 by @xmath80 using eqs.(12 ) and ( 15 ) , one can rewrite eq.(17 ) as @xmath81 where @xmath82 is the fourier transform of the basic scaling function ( see eq.(a1 ) in appendix a ) .
the term @xmath83 in eq .
( 18 ) is the fourier power spectrum defined by @xmath84 where @xmath85 with @xmath86 , and @xmath87 one can compare the 2nd moment @xmath88 with the dwt power spectrum given by fang & feng ( 2000 ) as @xmath89 where @xmath90 is the fourier transform of the basic wavelet .
therefore , both the dwt power spectrum and the dwt one - point statistics can be calculated with a single dwt decomposition .
the former relies on the wavelet , while the latter on the scaling function .
the fourier transform of the scaling function @xmath82 is non - zero in @xmath91-space where @xmath92 , and @xmath90 is mainly in @xmath93 ( fig . 1 and eq.(a4 ) ) .
therefore , the dwt power spectrum @xmath94 is actually a banded fourier power spectrum in the wavenumber range @xmath95 , while the 2nd moment @xmath96 contains all powers with @xmath97 .
it is well known that the 2nd moment of one - point statistics , such as @xmath88 or @xmath16 , is sensitive to long wavelength behavior of the perturbations .
this can also be seen from the relation between the scaling functions and wavelets as follows @xmath98 that is , the factor @xmath99 in eq.(22 ) extracts all powers on scales larger than @xmath23 .
eq.(22 ) holds only for compactly supported discrete wavelets such as daubechies wavelets .
we now consider distributions of discrete objects , such as simulation particles and observed or mock galaxies .
if the position measurement of particles or galaxies is perfectly precise , the number density distribution of these samples can be written as @xmath100,\ ] ] where @xmath101 is the total number of the particles or galaxies , @xmath102 is the position of the @xmath71th galaxy , @xmath103 is its weight , and @xmath104 is the selection function , which is given by the mean number density of galaxies when galaxy clustering is absent , and @xmath6 is the density fluctuation in the distribution
. one can subject @xmath105 to a dwt decomposition .
similar to eq.(13 ) , we have @xmath106 where @xmath107 the dwt one - point variables of @xmath6 is now given by @xmath108 \phi_{\bf j , l}({\bf x } ) d{\bf x}.\ ] ] since the selection function varies slowly , we have @xmath109 where @xmath110 is the mean of the selection faction in the cell @xmath55 .
the algorithm for @xmath110 is given in next subsection .
substituting eq.(27 ) into eq.(26 ) we have @xmath111 the second term on the r.h.s .
is due the normalization of the scaling function eq.(8 ) . by definition of equation ( 23 ) ,
selection function @xmath104 is the galaxy distribution if galaxy clustering @xmath6 is absent . in the plane - parallel approximation
, selection function depends only on @xmath112 , i.e. the coordinate in the redshift direction or the line - of - sight ( los ) .
thus , from equation ( 23 ) the selection function @xmath113 can be approximated by an average of @xmath105 over the plane @xmath114 , which depends upon the geometry of a real survey . in a simple case , for example , a mock survey in a simulation box
, it reads @xmath115 with eq.(24 ) , eq.(29 ) yields @xmath116 by definition , @xmath110 is the mean of @xmath113 in the cell @xmath55 . using eq . (
30 ) , we have @xmath117 that is , the selection function can be approximately expressed by the sfc of the ( observable ) galaxy distribution @xmath105 .
the observed galaxy distributions @xmath105 are considered to be a poisson sampling with an intensity @xmath118 $ ] . in this case , the characteristic function of the galaxy distribution @xmath105 is @xmath119 \rangle_p= \exp\left \ { \int d{\bf x } n({\bf x})[e^{iu({\bf x})}-1 ] \right \},\ ] ] where @xmath120 is the average for the poisson sampling . the @xmath71-point correlation functions of @xmath105 are given by @xmath121_{u=0}.\ ] ] we have then @xmath122 and @xmath123 substituting @xmath118 $ ] into eq.(35 ) , we have @xmath124 subjecting eq.(36 ) to a scaling function projection , and using eq.(27 ) , we have approximately @xmath125 the last term in eq.(37 ) is due to the normalization eq.(8 ) . from eq.(37 ) , the 2nd moment @xmath88 is given by @xmath126 -1.\ ] ] the second term on the r.h.s . under the summation is the correction due to poisson noise , and the rest is the normalized 2nd moment of @xmath6 .
one can also calculate the poisson correction for higher order moments with eq.(33 ) .
actually , astronomical data often gives only the distribution in redshift space , i.e. @xmath127= \bar{n}^s({\bf s})[1+\delta^s({\bf s})],\ ] ] where @xmath128 is the radial ( @xmath129 ) component of the velocity of the @xmath130th galaxy , @xmath131 is the selection function in redshift space , and @xmath132 is the hubble constant at the corresponding redshift .
thus , we can only calculate the cic moments in redshift space . for a given mass field @xmath6
, the galaxy velocity @xmath133 is a random field with mean @xmath134 where @xmath135 is the average over the ensemble of velocities .
the mean velocity @xmath136 is also called bulk velocity at @xmath137 , which is assumed irrotational ( bertschinger & dekel 1989 ; dekel , bertschinger & faber 1990 ; dekel et al .
1999 ) . in linear regime ,
the bulk velocity is related to the density contrast by @xmath138 where the parameter @xmath139 at present , i.e. redshift @xmath140 , and @xmath141 is the linear bias parameter .
the _ rms _ deviation of velocity @xmath133 from the bulk velocity @xmath136 is @xmath142 ^ 2\rangle_v = [ \sigma^{v_i}({\bf x})]^2 , \mbox { $ i=1,2,3$}.\ ] ] in order to express the scale dependence of the bulk velocity @xmath143 and the variance @xmath144 , we can also decompose the velocity field @xmath133 with a dwt , i.e. @xmath145 obviously , the variance @xmath144 on scale @xmath69 is given by @xmath146^{1/2}.\ ] ] although @xmath147 is @xmath75-dependent , @xmath148 should be @xmath75-independent if the random field @xmath133 is ergodic . on large scales , say @xmath149h@xmath0mpc , the velocity field @xmath133 is roughly _ gaussian _ , and it can be described by its mean and variance , which are generally scale - dependent . because of eq.(39 ) , the directly measurable dwt variable is not given by eqs .
( 26 ) or ( 28 ) , but @xmath150 \phi_{\bf j , l}({\bf s})d{\bf s } \\ \nonumber & = & \frac{1}{\bar{n}^s_{\bf j , l } } \sum_{m=1}^{n_g}w_m \phi_{\bf j ,
l}({\bf x}_m + \hat{\bf r } v_r({\bf x}_m)/h ) - \sqrt{\frac{l_1l_2l_3}{2^{j_1+j_2+j_3}}}.\end{aligned}\ ] ] where @xmath151 is calculated in the same way as eq.(31 ) , but with @xmath152 to replace @xmath105 . in this section
, we do not consider the effect of selection function , i.e. @xmath153 .
in the plane - parallel approximation , i.e. @xmath154 is along the @xmath155-direction , eq.(45 ) becomes @xmath156 if the velocity field is gaussian , subjecting eq.(46 ) to an average over the ensemble of velocities , we have @xmath157 e^{v_3({\bf x})\frac{1}{h } \frac{\partial}{\partial x_3 } + \frac{1}{2}(\sigma_{\bf j}^{v_3}/h)^2 \left ( \frac{\partial } { \partial x_3 } \right)^2 } \phi_{\bf j , l}({\bf x } ) d{\bf x } - \sqrt{\frac{l_1l_2l_3}{2^{j_1+j_2+j_3}}}.\end{aligned}\ ] ] for clarity , the angle brackets @xmath158 are dropped hereafter without causing any confusion .
if we consider only the linear effect of the bulk velocity , equation ( 47 ) is approximately @xmath159 \left [ 1+v_3({\bf x})\frac{1}{h } \frac{\partial}{\partial x_3 } \right ] e^{\frac{1}{2}(\sigma_{\bf j}^{v_3}/h)^2 \left ( \frac{\partial } { \partial x_3 } \right)^2 } \phi_{\bf j , l}({\bf x } ) d{\bf x } - \sqrt{\frac{l_1l_2l_3}{2^{j_1+j_2+j_3}}}.\ ] ] neglecting the terms of the order of @xmath160 , and using the linear relation between @xmath6 and @xmath136 , we have @xmath161 \frac{\partial } { \partial x_3 } \right\ } e^{\frac{1}{2}(\sigma_{\bf j}^{v_3}/h)^2 \left ( \frac{\partial } { \partial x_3 } \right)^2 } \phi_{\bf j , l}({\bf x } ) d{\bf x } - \sqrt{\frac{l_1l_2l_3}{2^{j_1+j_2+j_3 } } } \\ \nonumber & = & \int e^{\frac{1}{2}(\sigma_{\bf j}^{v_3}/h)^2 \left ( \frac{\partial } { \partial x_3 } \right)^2 } \phi_{\bf j , l}({\bf x } ) d{\bf x } + \sum_{{\bf l } ' } \epsilon_{{\bf j , l } ' } \int \phi_{{\bf j , l}'}({\bf x } ) e^{\frac{1}{2}(\sigma_{\bf j}^{v_3}/h)^2 \left ( \frac{\partial } { \partial x_3 } \right)^2 } \phi_{\bf j , l}({\bf x } ) d{\bf x}\\ \nonumber & & + \beta \sum_{\bf l ' } \epsilon_{\bf j , l ' } \int \phi_{\bf j , l'}({\bf x } ) e^{\frac{1}{2}(\sigma_{\bf j}^{v_3}/h)^2 \left ( \frac{\partial } { \partial x_3 } \right)^2}\nabla^{-2 } \frac{\partial^2}{\partial x_3 ^ 2 } \phi_{\bf j , l}({\bf x } ) d{\bf x } - \sqrt{\frac{l_1l_2l_3}{2^{j_1+j_2+j_3}}}.\end{aligned}\ ] ] in the last step , integration by parts for @xmath162 and the completeness relation equation ( 10 ) are used .
the summation of @xmath75 in eq.(49 ) is over @xmath163 . to simplify eq.(49 ) , we define the following matrices @xmath164 where the differential operators @xmath165 and @xmath166 are @xmath167 obviously @xmath168 depends on @xmath169 , not @xmath75 or @xmath170 individually .
appendix a provides the algorithms to calculate @xmath168 . thus , eq.(49 ) becomes @xmath171 using the so - called `` partition of unity '' ( fang & feng 2000 ) , one can show that @xmath172 therefore , eq.(52 ) gives @xmath173 this is the mapping of one - point variables between real and redshift spaces .
the second moment of one - point statistics in redshift space is @xmath174 [ \sum_{\bf l''}(\gamma^{1,0}_{\bf j , l , l '' } + \beta \gamma^{1,1}_{\bf j , l , l '' } ) \epsilon_{\bf j , l '' } ] \big\rangle.\ ] ] one can show that , even for a weakly non - linear field , both @xmath175 and @xmath176 are symmetric and quasi - diagonalized with respect to @xmath75 and @xmath170 , and @xmath177 should be @xmath75-independent .
the wavelet counterpart , i.e. @xmath178 , has been shown by pando , feng & fang ( 2001 ) .
thus , eq.(55 ) becomes @xmath179 using the completeness relation equation ( 10 ) , we have @xmath180 thus , eq.(56 ) gives @xmath181 where we have used @xmath182 .
therefore , we have finally the mapping of @xmath88 between real and redshift spaces @xmath183 in linear approximation , it is reasonable to estimate the effect of redshift distortion of @xmath152 and @xmath131 separately . in this case
, one can still use eq.(54 ) as the mapping from @xmath184 to @xmath185 .
we only need to study the effect of the mapping between @xmath131 and @xmath104 , which is given by @xmath186 \rangle_v \simeq \bar{n}({\bf x } ) + \frac{1}{h}v_r({\bf x } ) \hat{\bf r}\cdot \nabla n^g({\bf x}),\ ] ] where we have used @xmath187 . with the plane - parallel approximation of selection function ( 3.2 ) , eq.(60 )
becomes @xmath188 where we have dropped @xmath189 for @xmath131 .
> from equation ( 41 ) , @xmath190 can be represented by @xmath6 , so we have @xmath191 \\ \nonumber & \simeq & \bar{n}(x_3 ) \left \ { 1 - \beta\frac{d\ln \bar{n}(x_3)}{dx_3 } \frac{\partial}{\partial x_3 } \nabla^{-2}\left [ \frac{n^g ( { \bf x})}{\bar{n}(x_3 ) } - 1 \right ] \right \}.\end{aligned}\ ] ] combining eqs.(54 ) and ( 62 ) , we have finally @xmath192 \frac { \epsilon_{\bf j , l'}}{\bar{n}_{\bf j , l ' } } + \left .
\beta \frac{d\ln \bar{n}(x_3)}{dx_3}\right|_{\bf j , l } \sum_{l_3 } q_{\bf j , l , l ' } \frac{\epsilon_{\bf j , l'}}{\bar{n } _ { \bf j , l ' } } , \ ] ] where @xmath193 stands for the mean value of @xmath194 in the cell @xmath55 , and @xmath110 is given by eq.(31 ) .
in the last term of eq.(63 ) , @xmath195 and @xmath196 , and the summation runs only over @xmath197 .
the coefficient @xmath198 is defined by @xmath199 the calculation of @xmath198 is given in appendix a. because all @xmath75-diagonal elements of @xmath200 are zero ( appendix a ) , and @xmath201 are quasi-@xmath75-diagonal , the first and the second terms on the r.h.s .
of equation ( 63 ) are not correlated .
we have then @xmath202 ^ 2 \sum_{\bf l ' } q^2_{\bf l , l ' } \left \langle \left ( \frac{\epsilon_{\bf j , l'}}{\bar{n}_{\bf j , l ' } } \right ) ^2 \right \rangle.\end{aligned}\ ] ] for a uniform field , @xmath203 and @xmath204 are @xmath75-independent .
thus , equation ( 65 ) gives @xmath205 ^ 2 \sum_{\bf l ' } q^2_{\bf 0 , l ' } \right ) d_{\bf j}.\ ] ] using the inequality equation ( a16 ) , we can show that if @xmath206 we have @xmath207 ^ 2 \sum_{l_3-l'_3 } q^2_{{\bf j } , l_3-l'_3 } < \beta^2 s^2_{\bf j}.\ ] ] that is , the selection function term in equation ( 66 ) is even less than the second order terms @xmath208 if the selection function is slowly varying with @xmath112 .
we use n - body simulation samples to test the dwt algorithms for recovering the one - point statistics from redshift space to real space .
the model parameters are listed in table 1 .
notice that the simulation boxes are relatively large to minimize the effect on the bulk velocities due to the missing power at long wavelengths ( tormen & bertschinger 1996 ) .
we use a modified p@xmath209 m code ( jing & fang 1994 ) to evolve @xmath210 ( @xmath211 for lcdm2 ) cold dark matter ( cdm ) particles in a periodic cube of length @xmath20 on each side .
the linear power spectrum is given by the fitting formula in bardeen et al .
zeldovich approximation is applied to set up the initial perturbation .
the particles evolve 600 ( 800 for lcdm2 ) integration steps from @xmath212 down to @xmath213 for all the models .
table 1 [ cols="^,^,^,^,^,^,^,^",options="header " , ] in addition to @xmath1 at @xmath214 , we can also test the redshift distortion algorithms at high redshifts . in this case , @xmath1 is a function of @xmath215 , @xmath216 and @xmath217 as ( lahav et al .
1991 ) @xmath218^{0.6}.\ ] ] the simulation code is modified to generate light - cone outputs from @xmath219 to @xmath220 by a similar method in the _ hubble volume simulations _ ( evrard et al . 2001 ) .
instead of producing spherical light - cones , we apply the plane - parallel approximation in which a plane ( light - front ) sweeps through the simulation box at the speed of light .
let the los be along the @xmath112-axis , and the position of the light plane at time @xmath221 be @xmath222 .
the position and velocity of a particle is recorded when it crosses the plane , i.e. when the position @xmath223 of the particles satisfies @xmath224 since the time step @xmath225 in the simulations is finite , it is computationally impractical to use eq.(70 ) directly .
the position of the plane in time interval from step @xmath130 to step @xmath226 is approximately @xmath227,\ ] ] where @xmath5 is from 0 to 1 . on the other hand , for a particle we have @xmath228 , where @xmath229 is the velocity of the particle along the los .
thus , if we can find a solution of @xmath230 from the following equation @xmath231,\ ] ] we have @xmath232 , i.e. the particle crosses the plane at time @xmath233 .
we then record its position @xmath234 and velocity @xmath235 .
the accuracy is satisfactory .
for example , the difference between @xmath236 and @xmath237 is typically less than 50 @xmath238kpc out to redshift @xmath239 in the lcdm1 model . in the actual simulations ,
the integration variable is @xmath240 , the cosmic expansion factor , instead of @xmath221 . since the size of the simulation box is less than the distance swept by the plane from redshift @xmath219 to 0 , the motion of the plane is realized by periodic extensions of the simulation box when the plane meets the boundary .
this treatment should not have significant effects on our analyses , because the simulations already impose a periodic boundary condition , and the largest scale analyzed for each simulation is only a quarter of its box size .
once the real - space light - cone output is obtained , the observed redshift of a particle is given by @xmath241 , where @xmath215 is the actual redshift , and @xmath242 is the peculiar velocity .
to avoid negative @xmath243 caused by peculiar velocities at low redshifts , a lower cut is set to @xmath244 .
this results in less than a hundred particles bearing a negative @xmath243 in each run , which are then removed from the samples .
three light - cone outputs along the three orthogonal axes are produced for each simulation , which effectively increases the number of realizations by a factor of 3 . before testing the theory of recovering @xmath88 ( eq.(59 ) ) , we need to study the factors @xmath245 , @xmath246 , and @xmath247 , because the first two are the key of the redshift distortion of bulk velocity , while the last one the distortion of velocity dispersion .
below we drop @xmath248 in subscripts for simplicity . by the definition of eq.(50 ) , the factors @xmath249 and @xmath250 are given by @xmath251 we plot @xmath249 and @xmath250 in fig . 2 for the d4 wavelet .
it should be pointed out that @xmath252 depends only on the _ shape _ , not the scale @xmath69 of the scaling function .
for instance , @xmath253 , because the 3-d cells of @xmath254 have the same shape , i.e. all the ratios of length : width : height of these cells are 2:2:1 , even though they have different volumes ( scales ) . for cubic cells , i.e. @xmath255 , we have @xmath256 , and @xmath257 .
the factor @xmath258 is given by @xmath259 it depends on both the scaling function and the velocity dispersion @xmath260 , and therefore it is model - dependent .
the velocity variance @xmath260 is scale - dependent .
it is shown in fig . 3 for all the models , which is similar to fig . 6 in yang
actually , @xmath261 is given by all the power of pairwise velocity on scales less than the scale of @xmath69 , and therefore , it is larger on larger scales . to avoid crowding the figure , we only show @xmath262 for lcdm1 in fig .
4 . other modes , such as @xmath263 ,
follow @xmath262 closely . an interesting feature seen in fig .
4 is that @xmath264 and @xmath265 are almost equal to 1 from @xmath266 to 6 , which corresponds to spatial scales from 200 to 12.5 h@xmath0 mpc on the celestial sphere , and 200 to 100 h@xmath0 mpc along the los .
even for @xmath267 , the dependence on @xmath24 is only mild .
moreover , this property is model - independent , i.e. for other models , we also have the @xmath24-independence of @xmath264 and @xmath265 .
the factor @xmath263 is sensitive to @xmath261 ( as a variable ) when the scale along the los is small , but insensitive when the scale is large .
for example , when @xmath261 varies from 0 to 500 km s@xmath0 , @xmath265 drops from 1 to 0.97 , i.e. the change is only 3% , while @xmath268 from 1 to 0.46 , i.e. the change is by a factor of 2 .
if we try to use a scale - independent @xmath269 to recover the real space power spectrum , we should chose @xmath269 to give a good fitting on small scales , because the value of @xmath269 does not significantly affect the large scales .
in addition to @xmath270 , @xmath258 depends on @xmath271 and @xmath272 ( collectively @xmath273 ) through @xmath260 , which is less than 500 km s@xmath0 ( fig .
3 ) . therefore ,
when the los scale is above 100 h@xmath0 mpc , @xmath258 is almost a constant with respect to @xmath273 . on the other hand ,
when the scale is below 50 h@xmath0 mpc , @xmath258 shows a mild dependence on @xmath273 due to the scale dependence of @xmath260 . in a word
, we benefit from the dwt representation to see the different behavior of the two types of the redshift distortion .
the former is sensitive to the shape of the mode , not the scale , while the latter just the contrary .
we neglect the effect of selection function here for the simulation samples , which is similar to that on the dwt power spectrum tested in yang et al .
( 2002 ) . since the matrix @xmath168 is quasi - diagonalized with respect to @xmath274 , we have approximately @xmath275 .
thus , eq.(59 ) yields @xmath276 in eq.(75 ) , the contributions of the operators @xmath165 and @xmath166 to the redshift distortion are separated .
usually terms of @xmath166 ( @xmath249 and @xmath250 ) are called linear redshift distortion , and that of @xmath165 ( @xmath258 ) non - linear distortion .
these terminologies may cause a confusion . here
`` linear '' means only that @xmath166 is from the linear term of @xmath190 in the approximation eq.(48 ) , and `` non - linear '' means that @xmath165 is non - linear of the velocity dispersion @xmath277 . it does not imply that the terms of @xmath166 is enough for estimating the redshift distortion of a field that is in the linear regime . since the velocity dispersion @xmath269 is non - zero in the linear regime , the non - linear distortion term of @xmath165 may also play a non - negligible role even when the field is linear . in this paper , we follow the tradition to call the terms of @xmath166 and @xmath165 the linear and non - linear effects of redshift distortion , respectively . however , it should be kept in mind that these names do not reflect whether the field is linear or non - linear . for @xmath69-diagonal modes , equation ( 75 )
reduces to @xmath278 .
the factors within the parentheses are known as the linear redshift - to - real space mapping for two - point correlation function ( kaiser 1987 ) .
figure 5 shows the recovery of @xmath69-diagonal @xmath88 .
it is evident that the recovery equation ( 75 ) works very well on all scales and redshifts considered , except for the scdm model at @xmath279 . due to greater @xmath1 parameters , the distortions of the tcdm and
the scdm models are generally stronger than that of the rest .
for these two models , @xmath280 is almost parallel to @xmath88 , which indicates a very weak non - linear redshift distortion on all scales .
this is consistent with the fact that the velocity dispersions of the two models are small on all scales compared to the others . since the volume of a cubic cell with @xmath281h@xmath0mpc on each side is approximately the same as that of a sphere with a radius of @xmath282h@xmath0mpc , the values of @xmath283 are consistent with @xmath284 at corresponding redshifts for each model .
the differences , nevertheless , are due to the difference in the window functions .
similar to @xmath88 , one can generalize @xmath16 to @xmath285 , which is the _ rms _ density fluctuation in a sphere of radius @xmath10 .
since the behavior of @xmath286 is the same as @xmath287 , one can recover the real space @xmath285 from the redshift space @xmath288 in a similar way as eq.(75 ) , i.e. @xmath289^{1/2 } \sigma_r$ ] , where @xmath24 is chosen to match the volumes of the dwt cell and the spherical top - hat window of radius @xmath10 .
figure 6 demonstrates the recovery of @xmath285 for all the models .
it is a coincidence that @xmath290 for the low density models .
figures 7 and 8 give the recovery of off - diagonal @xmath88 .
this recovery works well on scales above 50 h@xmath0 mpc . on scales
less than 50 h@xmath0 mpc , the error is noticeable but still small .
this is partially due to the approximation of the quasi - diagonality of the covariance @xmath291 in equation ( 56 ) .
moreover , the separation between the factors involving operators @xmath165 and @xmath166 in eq.(75 ) is not perfect as well .
the errors of recovering @xmath88 can be more clearly seen via a comparison with the recovery of the dwt power spectrum from redshift space @xmath292 to real space @xmath293 .
it is ( yang et al .
2002 ) @xmath294 where @xmath295 is defined by eq.(50 ) but with replacement of @xmath296 to @xmath297 .
the factors @xmath298 and @xmath299 also satisfy @xmath300 and @xmath301 .
the non - linear term @xmath302 behaves the same as @xmath258 , but with a little stronger dependence on @xmath69 .
we find that the recovery of @xmath88 by eq.(75 ) has an error within 10% in most cases , while the dwt power spectrum recovery eq.(76 ) is accurate to 5% or better .
this is because that the quasi - diagonality of the sfc s covariance @xmath303 is poorer than the quasi - diagonality of the wfc s covariance @xmath304 .
as we know from the dwt analysis , the covariance @xmath305 is always quasi - diagonal for a gaussian field , regardless the power spectrum . on the other hand ,
the quasi - diagonality of the covariance @xmath306 is not a generic mathematical property , but only an approximation .
all the above algorithms can be applied to real samples as well . however , there are several problems to be considered in the analysis on real data .
data assignment on grid _ to calculate the moments of counts - in - cells , one has to cast a grid on the data set either explicitly or implicitly , and assign the data in cells defined by the grid .
the arbitrariness of the assignment may lead to significant error ( colombi , bouchet & schaeffer 1995 ) .
one way to reduce this error is to shift the grid .
it is well known that this assignment may also result in spurious features of the power spectrum on scales around the nyquist frequency of the grid . in the dwt analysis
, one can use different grids , but no shift is needed . for a given grid ,
the assignment is realized by the same scaling function used for the data decomposition . since scaling functions
are orthogonal and complete in each given scale @xmath69 , the spurious features and false correlations can be completely avoided ( fang & feng 2000 ) . to test the effect of the assignment on the one point statistics , we calculate @xmath287 for a snapshot sample of lcdm1 with randomly shifted grids .
the snapshot sample is take at @xmath307 . in fig .
9 , snap-0 represents @xmath287 without grid shifting , while snap-32 and snap-1024 , respectively , the average over 32 , and 1024 random shifts for grids on all scales .
no significant deviation is detected among the results .
_ 2 . boundary condition and edge effect _ in a real survey , the edge effect is unavoidable due to the geometry ( e.g. szapudi & colombi 1996 ) .
since the dwt bases are localized in physical space , the effect of edges can be effectively suppressed by dropping modes that are close to the boundary of the samples . in other words ,
when calculating @xmath88 , the averaging in eq.(17 ) runs only over modes that are not significantly affected by the edge effect , and the normalization factor @xmath308 is replaced by the volume over which the average takes place .
this method has been tested numerically in the power spectrum detection ( pando & fang 1998b ) .
it shows that the power spectrum can be fairly reconstructed regardless whether applying a periodic boundary or zero padding outside of the samples .
the treatment of dropping edge modes has also been successfully employed to obtain the power spectrum for the las campanas redshift survey of galaxies , which has a slice - like geometry ( yang et al .
when the interested scale is close to the dimension of the sample , one can not afford to drop all the edge modes , and then a more thorough treatment is needed .
the edge effect exists in our simulation sample because the light - cone output is not periodic along the redshift axis , while we have used periodic d4 wavelets in the analysis . for comparison , the real - space light - cone @xmath287 from fig . 5
is included in fig .
9 . there is no significant difference between the results from the snapshot , which is truly periodic , and that from the light - cone around the same redshift .
the slightly larger standard deviation of the light - cone result at 200 @xmath238mpc ( a quarter of the simulation box ) is due to a larger fraction of edge cells .
non - poisson sampling _ in 3.3 , we have considered the correction for poisson sampling .
it is sufficient for simulation samples .
however , it may not be typical for real samples . for instance , some galaxy catalogs may be given by sub - poisson sampling on small scales , or small halos ( e.g. bullock , wechsler & somerville , 2002 ) .
the sub - poisson distribution is simply due to very low mass of the considered halo , so that it can not host any additional objects . in other words ,
the sub - poisson distribution is significant on small scales on which galaxies are anticorrelated .
therefore , this sub - poisson sampling is similar to the sub - poisson distribution of the fermi - dirac statistics ( one state can host no more than one particle ) .
the effect of this sub - poisson has been extensively studied in quantum optics ( e.g. martin & landauer , 1992 ) .
it is possible to extend the poisson sampling eq.(35 ) to include fermi - like sampling .
the algorithm for a modified poisson sampling has been developed by jamkhedkar , bi and fang ( 2001 ) ( see their append b ) .
we first consider a @xmath1 estimator without non - linear effect .
for instance , @xmath309 ( fig .
4 ) , i.e. the non - linear effects are small above 100 h@xmath0 mpc along the los , eq.(75 ) gives @xmath310 it is easy to construct a @xmath1 estimator with eq.(77 ) , because the quantities @xmath88 , @xmath280 , and @xmath311 are not rotationally invariant , but they satisfy the following symmetries with respect to the triple indices @xmath312 . 1 . if the cosmic density and velocity fields are statistically isotropic , @xmath88 in real space is invariant with respect to cyclic permutations of index @xmath56 , i.e. @xmath313 2 . in the plane - parallel approximation ,
i.e. the coordinate @xmath112 is in the redshift direction , we have @xmath314 @xmath315 using eqs.(77)-(80 ) , we have a @xmath1 estimator as follows @xmath316 eq.(81 ) looks very similar to the @xmath1 estimator with quadrupole - to - monopole ratio , or the multipole moments of two - point correlation function .
however , eq.(81 ) contains not only the information of shape ( like multipole moments ) , but also scales .
all the dwt quantities in eq.(77 ) are on scale @xmath69 .
that is , the @xmath1 in eq.(77 ) refers to mode @xmath56 .
therefore , if @xmath1 is scale - dependent , the @xmath1 estimated by eq.(81 ) is its value referring to @xmath69 . on the other hand ,
if @xmath1 is scale - free , the values of @xmath1 given by estimator eq.(81 ) with different mode @xmath69 should be the same .
figure 10 plots the results in the case of a scale - free @xmath1 for the lcdm1 model using estimators with @xmath317 .
the results indeed show that the estimated @xmath1 is independent of the @xmath69 used in the estimator , and it follows the theoretical curves in the entire redshift range considered .
therefore , we expect that the estimator ( 81 ) would be useful to study the scale - dependence of bias parameters of galaxies . for comparison , fig 11 shows the @xmath1 estimated from quadrupole - to - monopole ratio as @xmath318 where @xmath319 and @xmath320 are the quadrupole moment and the monopole moment of the power spectrum @xmath321 respectively ( cole , fisher & weinberg 1994 , hereafter cfw ; hamilton 1997 ) .
we average the results from wavelength 50 h@xmath0mpc to 400 h@xmath0mpc for each realization .
thus the error bars are among different realizations of each model .
as noticed in cfw , this estimator gives a lower @xmath1 than the true value at @xmath214 even for wavelengths @xmath322 h@xmath0mpc .
in addition , we find that it progressively overestimates @xmath1 at higher and higher redshifts .
it is suggested in cfw that the underestimate at short wavelengths is due to the impact of non - linear gravitational clustering .
however , this explanation is still difficult to apply for the underestimate at long wavelengths , and the overestimate at high redshift , as cosmic field on long wavelengths or high redshifts should be linear .
this discrepancy occurs because the quadrupole - to - monopole ratio eq.(82 ) is scale - dependent even when the field is in linear regime .
this scale - dependence is given by the non - linear redshift distortion , or the distortion of velocity dispersion .
as mentioned in 5.3 , @xmath277 is non - zero in linear clustering regime , the non - linear redshift distortion ( or the distortion of @xmath277 ) should be considered regardless whether the field is linear or non - linear .
therefore , the discrepancy in cfw can not be eliminated for long wavelengths or high redshift . to solve this problem , extra information , such as the power - spectrum index of the density fluctuations , or the real - space correlation function ,
is added in the quadrupole - to - monopole ratio estimator eq.(82 ) ( peacock et al .
2001 ) .
different from the quadrupole - to - monopole ratio , the dwt @xmath1 estimators are scale - decomposed .
it is useful to consider scale - dependent effect .
the dwt estimators are able to consider the non - linear redshift distortion ( @xmath277 redshift distortion ) without assuming extra information .
actually , the result with estimator eq.(81 ) has already shown the effect of the non - linear redshift distortion .
we can see from fig .
10 that the @xmath1 estimated by eq.(81 ) are slightly dependent on @xmath24 , and it is progressively higher in the order @xmath323 .
this can be explained with fig .
4 , which shows that the factor @xmath258 is less than one , and is smaller for a smaller scale .
thus the estimator eq.(81 ) , which ignores the factor @xmath258 in eq.(77 ) , leads to a progressively higher @xmath1 in the order @xmath323 .
thus , a simplest way to estimate the non - linear effect is to take an average over @xmath1 given by eq.(81 ) with different @xmath24 .
the left panel of figure 12 presents the averaged @xmath1 from fig .
10 for the four models .
the error bars contain the contribution of the non - linear effect .
that is , the dwt algorithm can do a self - test on whether the result is largely affected by the non - linear effect .
more delicate @xmath1 estimators can be constructed if we consider the following properties : the non - linear factor @xmath258 depends only on @xmath270 , but independent of @xmath273 if the los scale is above 100 h@xmath0mpc , while the linear factors @xmath249 and @xmath250 depends only on the shape of the dwt mode ( 5.2 ) .
these properties apply to @xmath302 , @xmath298 , and @xmath299 as well .
thus , we can combine modes with similar @xmath270 but different shapes to cancel the non - linear factors , such as @xmath324 thus , from eq.(76 ) , we have a @xmath1 estimator as @xmath325 this estimator is similar to that in yang et al .
it can be used for any pairs ( @xmath24 @xmath326 @xmath327 ) even when the scale of @xmath24 is small .
the right panel of figure 12 shows that for the four models , the @xmath1 estimated by eq.(84 ) has no more than about 15% error at all redshifts @xmath328 .
one can also use the modes @xmath329 , @xmath330 to replace modes @xmath331 and @xmath332 in eqs.(83 ) and ( 84 ) , which gives similar results to fig .
that is , the weakly non - linear redshift distortion can be considered by the dwt @xmath1 estimators on scales until about @xmath333 h@xmath0 mpc when the quadrupole - to - monopole estimator shows significant errors ( fig .
the counterpart of eq.(84 ) with @xmath88 , @xmath334 is not as good as eq.(84 ) .
it causes an error of about 40% in @xmath1 because again the quasi - diagonality of the sfc s covariance is poorer than that of wfc s . since a dwt decomposition of a random field yields variables for calculating both the power spectrum and one - point statistics , the @xmath1 estimations with eqs.(81 ) and ( 84 )
can be done in the same time .
an accurate @xmath1 estimator requires the knowledge of a precise recovery of @xmath293 or @xmath88 , which accounts for the non - linear redshift distortion due to the velocity dispersion , and even the second order effect of the bulk velocity ignored in eq.(48 ) . in other words ,
the non - linear redshift distortion exists even on scales that are linear in the sense of structure formation , and it must be corrected to achieve a reliable and accurate @xmath1 .
we have developed the one - point statistics of a perturbed density field with the multiresolutional decomposition based on discrete wavelet transform .
since the scale and shape of the dwt bases are well defined , this frame work is very effective to deal with problems of how the one - point distribution and its moments depend on the scale and shape of the window function . with this property ,
we have established the algorithm of one - point variable and its moments in considering the effects of poisson sampling and selection function .
we have also established the algorithm for recovering the dwt one - point statistics from the redshift distortion due to bulk velocity , velocity dispersion and selection function .
because the recovery of the real - space dwt one - point variable and its moments can be realized scale - by - scale , one can design @xmath1 estimators which are sensitive to the scale - dependence of @xmath1 , for instance , caused by the scale - dependence of bias parameter of galaxies .
these @xmath1 estimators are effective in avoiding the difficulty caused by the scale - dependence of the non - linear redshift distortion . compared with conventional @xmath1 estimators ( peacock et al 2001 )
, the dwt @xmath1 estimators do not need to assume that the velocity dispersion is scale - independent , or to add extra information , such as the power - spectrum index or the real - space correlation function of the field .
numerical tests by n - body simulation samples show that the proposed estimators can yield the correct value of @xmath1 with about 15% uncertainty for all popular cdm models in the redshift range @xmath335 . since dwt decomposition contains two sets of bases , the scaling function @xmath28 and wavelet @xmath29 , a dwt decomposition of a density field @xmath18 actually yields variables for one - point statistics @xmath336 as well as the variables for calculating the power spectrum @xmath337 ( fang & feng 2000 ) .
in this sense , one can say that the dwt decomposition unifies the algorithm of cic and power spectrum , which are only two aspects of the statistics with the dwt variables sfcs and wfcs respectively . for a finite size sample , these two aspects of statistics can play different roles : the former contains information of the perturbations on scales larger than the size , while the latter does not .
however , the latter generally has diagonalized covariance , while the former does not . therefore ,
the @xmath1 estimators with wfcs are better than sfcs , while sfcs are useful to estimate the effect of perturbations on large scales .
actually , the dwt decomposition can provide more types of statistics variables . for 2-d or 3-d samples
, we have variables defined as @xmath338 @xmath339 obviously , these variables are sampled partially by the scaling function @xmath28 , and partially by the wavelet @xmath29 .
statistics with these variables are not typical cic , or power spectrum . they are , however , useful to study the one - point statistics , or power spectrum of 2-d and 3-d samples . with the method developed in this paper , it is not difficult to calculate various corrections ( poisson sampling , selection function , redshift distortion ) on the one - point statistics with variables ( 86 ) or ( 87 ) .
hz is grateful to daniel eisenstein for extensive discussions on this paper and facilitating the lcdm2 simulation .
hz would also like to thank david burstein for hosting the utility codes in this paper .
let us consider the plane - parallel approximation , i.e. coordinate @xmath112 is in the redshift direction . by definition eq.(50 )
, we have @xmath340 because 1-d scaling function @xmath28 is given by dilating and translating the basic wavelet @xmath341 as @xmath342 the fourier transform of @xmath29 is @xmath343 and @xmath344 where @xmath82 is the fourier transform of the basic scaling function @xmath345 the function @xmath346 is shown in fig
. 1 .
if @xmath354 is independent of @xmath137 , @xmath258 is @xmath355.\ ] ] in the plane - parallel approximation , we have @xmath356.\ ] ] the summation of equations ( a8 ) and ( a9 ) also runs only over numbers of @xmath357 . using the results for @xmath252 and @xmath258 ,
it is easy to find @xmath358 \exp[-(a/2)(\sigma^{v_3 } k_3/h)^2 ] |\hat{\phi}(u_1)\hat{\phi}(u_2)\hat{\phi}(u_3)|^2.\end{aligned}\ ] ] it is obvious from equation ( a10 ) that @xmath176 is symmetric with respect to @xmath75 and @xmath170 , and it depends only on the difference @xmath359 .
consequently , @xmath360 is independent of @xmath75 . in addition , the elements @xmath360 are dominant over @xmath176 with @xmath361 , because the latter sums over oscillating terms . since @xmath367\sin[2\pi u'_3(l_3-l_3 ' ) ] \simeq \left
\ { \begin{array}{ll } 1/2\pi u_3 & \mbox{if $ u_3=u'_3 $ } \\ 0 & \mbox{otherwise } \end{array } \right .,\ ] ] we have @xmath368 ^ 2 [ \gamma^{0,1}_{\bf j}]^2.\ ] ] | in this paper , we develop a method of performing the one - point statistics of a perturbed density field with a multiresolutional decomposition based on the discrete wavelet transform ( dwt ) .
we establish the algorithm of the one - point variable and its moments in considering the effects of poisson sampling and selection function .
we also establish the mapping between the dwt one - point statistics in redshift space and real space , i.e. the algorithm for recovering the dwt one - point statistics from the redshift distortion of bulk velocity , velocity dispersion , and selection function .
numerical tests on n - body simulation samples show that this algorithm works well on scales from a few hundreds to a few h@xmath0 mpc for four popular cold dark matter models . taking the advantage that the dwt one - point variable is dependent on both the scale and the shape ( configuration ) of decomposition modes , one can design estimators of the redshift distortion parameter ( @xmath1 ) from combinations of dwt modes . comparing with conventional @xmath1 estimators , such as quadrupole - to - monopole ratio , the dwt @xmath1 estimators are scale - decomposed .
it is useful to consider scale - dependent effects . when the non - linear redshift distortion is not negligible , the quadrupole - to - monopole ratio is a function of scale .
this estimator would not work without adding information about the scale - dependence , such as the power - spectrum index or the real - space correlation function of the random field .
the dwt @xmath1 estimators , however , do not need such extra information .
moreover , the scale - decomposed @xmath1 estimators would also be able to reveal the scale - dependence of the bias parameter of galaxies .
numerical tests show that the proposed dwt estimators are able to determine @xmath1 robustly with less than 15% uncertainty in the redshift range @xmath2 . |
type 2 quasars are powerful ( @xmath3 erg / s ) active galactic nuclei ( agns ) whose central regions are shielded from the observer by large amounts of gas and dust . until recently , the existence of dust - enshrouded quasars was in question , but now these objects are routinely discovered at x - ray ( e.g. , @xcite ) , infrared @xcite and optical @xcite wavelengths .
a significant controversy remains regarding their space density and contribution to the agn population luminosity budget .
some analyses suggest that there are at least as many obscured as unobscured quasars even at the highest luminosity end @xcite , whereas others argue for a strong decline of obscured quasar fraction with luminosity @xcite .
if an agn is entirely embedded in an optically thick dusty cloud , virtually all of its original radiation ( emitted primarily at optical to soft x - ray wavelengths ) is absorbed by dust and is then re - emitted thermally in the ir . in this case , there is no optical signature of the agn , but its presence may be revealed when hard x - rays and a large ir luminosity are detected . in this way , some ultra - luminous infrared galaxies ( ulirgs ) have been shown to contain embedded agns @xcite .
the dominant energy source in ulirgs remains a matter of controversy .
some authors argue that they are powered by star formation , with the agn contributing only a small fraction @xcite , while others suggest that buried agns are energetically dominant in many ulirgs @xcite .
an alternative possibility is that the nucleus is only partially covered by the obscuring material . in this case , the agn is seen as obscured or unobscured depending on the line of sight , making its uv , optical and x - ray emission highly anisotropic @xcite . in particular , uv and x - ray emission can make its way out through the holes in the obscuring cloud and photo - ionize extended regions , which in turn produce powerful optical emission lines . using the sloan digital sky survey ( sdss ; @xcite ) we have selected hundreds of type 2 quasars at redshifts @xmath5 on the basis of their optical emission line properties @xcite .
ir observations are key to determining the physical conditions of the obscuring material and its geometric properties such as its covering factor and extent . in this paper
we present space telescope @xcite infrared spectrograph ( irs ; @xcite ) observations of twelve type 2 quasars from the sdss sample . in section [ sec_data ]
we describe our sample selection and data reduction . in section [ sec_spectra ]
we present a detailed analysis of the spectral features .
we compare our sample with other samples of type 2 quasars in section [ sec_compare ] , we discuss our results in the context of radiative transfer models in section [ sec_models ] and we conclude in section [ sec_conc ] . throughout the paper ,
we adopt a cosmology with @xmath6 , @xmath7 , @xmath8 .
objects are identified as sdss [email protected] in figure [ pic1 ] and table [ tab_summary ] and are shortened to sdss jhhmm@xmath9ddmm in the text .
we refer to type 1 and type 2 agns based on their optical classification ( in type 1 objects , broad lines are detected and in type 2 , they are not ) .
when discussing the luminosity dependence of agn phenomenon , we distinguish between seyfert galaxies ( @xmath10 erg / s ) and quasars ( @xmath3 erg / s ) based on their bolometric luminosities .
the type 2 quasars discussed in this paper were selected from the spectroscopic database of the sdss on the basis of their optical emission line properties @xcite . briefly , type 2 agns are required to show narrow emission lines with ratios characteristic of a non - stellar ionizing continuum .
we then use the luminosity of the [ oiii]5007 emission line as a proxy for the intrinsic luminosity of the obscured nucleus @xcite and classify type 2 agns as quasars if @xmath11[oiii]@xmath12 and as seyfert galaxies otherwise .
@xcite obtained a sample of 887 type 2 quasars at redshifts @xmath13 by applying these criteria to the most recent sdss data .
[ oiii ] luminosities were not corrected for reddening , and they are thus likely to be underestimated @xcite . our extensive follow - up work has demonstrated that the objects that we classified as type 2 quasars based on their optical properties indeed fit operational definitions of type 2 quasars used at other wavelengths . in particular , they typically have high x - ray luminosities , @xmath14 erg / s , and large amounts of neutral gas along the line of sight , @xmath15 @xmath16 @xcite .
furthermore , using optical spectropolarimetry , we demonstrated that at least some of the objects in our sample harbor broad - line ( type 1 ) agns in their centers @xcite . in cycle 1 of the space telescope we conducted seven - band irac+mips photometry of 25 type 2 quasars from @xcite and @xcite .
these objects were selected primarily based on their [ oiii]5007 line luminosity ( specifically , @xmath11[oiii]@xmath17 ) .
photometric observations ( zakamska et al . , in prep . ) , as well as archival iras data @xcite , show that type 2 quasars from our sample typically have ir luminosities in excess of @xmath18 erg / s , placing them into the classical quasar luminosity regime and validating our [ oiii]5007 - based luminosity criterion .
we selected ten sources from this sample with the highest irac-8 and mips-24 fluxes ( @xmath19>1.5 $ ] mjy and @xmath20>6 $ ] mjy ) for spectroscopic follow - up with the irs ( table [ tab_summary ] ) .
in addition , we matched the most recent catalog of sdss type 2 quasars @xcite against the archive and found the following two sources with archival irs observations : 1 .
sdss j091345.49 + 405628.2 ( @xmath21iras b09104 + 4109 ) program 1018 ( in - orbit checkout / science verification ) , pi armus , z=0.442 ; 2 .
sdss j005009.81 - 003900.6 ( @xmath21iras b00476@xmath220054 ) program 105 ( guaranteed time observations ) , pi houck , z=0.729 .
the first of these sources is a well - known luminous type 2 quasar @xcite , whereas the second source was originally classified as an ultra - luminous infrared galaxy ( ulirg ) .
both objects show high - ionization optical line ratios , which is why they were selected by the optical procedure described in @xcite , so we include them in our analysis . the ten sources in our targeted program were observed with the first order of the short wavelength low resolution module ( sl1 ; 484 sec exposure for each object ) and with the second order of the long wavelength low resolution module ( ll2 ; 488 sec ) .
these observations cover the wavelength range @xmath23 with resolution @xmath24 .
we started our analysis from the basic calibrated data ( bcd ) provided by the science center ( pipeline ver .
s13.2.0 ) , which applies dark corrections , flat fields , and droop and nonlinearity corrections , and provides flux calibration .
we removed the sky background and most of the hot pixels by a pairwise subtraction of frames at different nod positions .
we then extracted positive and negative spectra from the subtracted images with the smart package @xcite , using the ` interactive / column / no skysub ' method .
additionally , we applied hot pixel corrections to the subtracted two dimensional spectra using irsclean . the resulting spectra ( 4 for sl1 and 8 for ll2 for each object ) were averaged into the final spectra shown in figure [ pic1 ] .
when necessary , the spectral errors were estimated from pixel - to - pixel variations by subtracting a smooth model from the spectrum .
we found that our reduced spectra are in agreement with irac-8.0 and mips-24 photometry and conclude that the absolute flux calibration is accurate to better than 15% .
we reduced the two archival sources in the same way ( pipeline ver .
s14.0.0 ) , and they are also shown in figure [ pic1 ] .
the archival observations had total exposures 2 - 4 times shorter than the targeted observations , since the archival targets are iras - detected and therefore are relatively bright in the ir .
we have included sl1 , ll2 and ll1 observations ( total wavelength coverage @xmath25 ) . in the source sdss
j0913 + 4056 ( @xmath21iras b09104 + 4109 ) , one of the two available ll2 datasets produced fluxes which were systematically lower ( by about 5% ) than those from the other ll2 dataset and those extrapolated from the ll1 dataset .
we multiplied the fluxes of the discrepant ll2 dataset by a linear function of wavelength to correct for this difference .
some of the analysis in this paper makes use of ir photometric data .
for the ten targeted sources , the ir photometry is from our cycle 1 program ( zakamska et al .
, , in prep . ) .
for sdss j0913 + 4056 ( @xmath21iras b09104 + 4109 ) , we use iras catalog fluxes ( @xmath26=130 $ ] mjy , @xmath27=333 $ ] mjy , @xmath28=525 $ ] mjy ) and fluxes derived from irac observations in program 105 .
using aperture photometry , we calculate @xmath29=5.0\pm 0.7 $ ] mjy . for sdss j0050 - 0039
( @xmath21iras b00476@xmath220054 ) , we use iras data from @xcite ( @xmath28=260 $ ] mjy ) . in all cases where the data overlap , the ir photometry is in agreement with the fluxes derived from the irs spectra within the uncertainties of the photometric data .
from figure [ pic1 ] it is apparent that the ir spectra of type 2 quasars display a wide range of properties .
some are nearly featureless , whereas others are busy with absorption and emission features , some of which are identified in the top panels of figure [ pic1 ] .
the difficulties of placing the continuum to identify detected features , as well as of disentangling blends of emission and absorption are well known for ir spectra ( e.g. , @xcite ) , and some of our conclusions may suffer from the uncertainties in this rather subjective step .
nevertheless we attempt to minimize these uncertainties or at least obtain their measure by using several different approaches to spectral analysis . in the first step ,
we use pahfit @xcite , a public idl - based routine that decomposes low - resolution irs spectra into a combination of starlight ( represented by a black body spectrum ) , featureless continuum ( represented by a sum of thermal components at several fixed temperatures ) , emission features ( pah dust emission features , fine - structure lines of ions , rotational lines of h@xmath0 ) , and dust extinction ( modeled as a screen of cold silicate dust affecting all other components ) .
the code was developed primarily for the analysis of polycyclic aromatic hydrocarbon ( pah ) features in high signal - to - noise ratio spectra of star forming galaxies , so the relative strengths of pahs are allowed to vary .
the advantage of using this code for our objects is that it produces excellent global fits in every case .
the disadvantage is that the large number of fit parameters ( 30 - 40 ) makes it hard to assess which of the weak features included in the best pahfit model are actually detected . even for the stronger features ,
some fit parameters are degenerate , as was illustrated by @xcite in the case of the interplay between pah emission and silicate ( si ) absorption .
we use pahfit as a guide to indicate significant pah feature detections . in the second step ,
we use a much simpler global fit to the data consisting of a featureless continuum ( represented by a second order polynomial ) and a pah template with fixed ratios of feature strengths .
the sum of these components is then multiplied by @xmath30 $ ] , where @xmath31 is the silicate opacity curve , whose normalization is a free parameter which is allowed to be positive or negative . the pah template that we use is derived using pahfit from a composite spectrum of galaxies from smith et al .
( 2007 ; solid line in figure [ pic1 ] , top ) , and the si opacity curve is based on iso observations of the galactic center ( @xcite ; dotted line in figure [ pic1 ] , top ) .
in some cases we masked the strongest emission lines in the fitting ( e.g. , [ nevi]7.63 , h@xmath0 s(3)9.67 , [ siv]10.51 and [ neii]12.8 in sdss j1641 + 3858 ) .
this model has only five free parameters , but still produces good global fits for our spectra .
using these two models , we calculate the si feature strength , @xmath32/f_{\nu}[{\rm cont}])$ ] , where @xmath33 $ ] is the model flux density at rest - frame 9.7 and @xmath34 $ ] is the flux density extrapolated from the model continuum outside the feature . in table
[ tab_summary ] and figure [ pic1 ] , we quote si strengths from the second method .
the 1@xmath35 uncertainties shown in figure [ pic_tau ] are taken to be half the difference in si strengths measured by the two methods .
we tried a range of model milky way and small magellanic cloud ( smc ) opacity curves from @xcite , as well as the sgr a * opacity curve from @xcite and the galactic dust opacity curve from @xcite .
the variation in @xmath36 measured using different opacity curves amounts to no more than 10% , although the conversion between ir optical depth and optical extinction or hydrogen column density varies by large factors ( @xmath37=1 $ ] corresponds to @xmath38 mag for the milky way opacity curve and to @xmath39 mag for the smc curve ) .
we find that the shape of the 10 si feature in type 2 quasars is well - represented by the @xcite opacity curve , which we use hereafter . in the third step
we examine detections of emission lines , such as those from fine - structure transitions in ions ( section [ sec_atomic ] ) and those from rotational transitions of molecular hydrogen ( section [ sec_hydro ] ) .
as shown in the top panel of figure [ pic1 ] , we consider all transitions listed in @xcite , as well as [ nevi]7.63 and [ nev]14.33 .
all these lines are expected to be unresolved , so their widths are determined by the instrumental resolution ( [email protected] for sl1 , 0.14 for ll2 and 0.34 for ll1 in the observed frame ) .
for each line , we calculate the error per pixel by fitting a low order polynomial to a cut - out from the spectrum excluding a region 3@xmath41fwhm@xmath42 wide on both sides of the line centroid .
we then include the region containing the line into the fit and examine the reduced @xmath43 with or without a fixed - width gaussian representing the line to determine the confidence of the detection .
similarly , we examine detections of pah emission features by fitting a second order polynomial + fixed - width drude ( damped harmonic oscillator ) profiles @xcite .
the line fluxes that we obtain from these simple fitting procedures are then compared to pahfit line fluxes , allowing us to estimate measurement confidence and uncertainty . for unresolved lines ,
the uncertainties are in good agreement with those calculated from the measured error per pixel .
silicates are a major constituent of interstellar and circumstellar dust .
their presence is indicated in many classes of galactic and extragalactic objects by two strong features , one due to a si - o stretching mode and peaking at around 10 and another due to a o - si - o bending mode and peaking at around 18 @xcite . in a recent study of about 200 ulirgs , seyfert 2 galaxies , seyfert 1
galaxies and type 1 quasars , @xcite find that there exists a strong relationship between the optical classification of these objects and their ir properties .
we show their sample together with our type 2 quasars and a compilation of ulirgs and agns from different datasets in a si strength vs. mid - ir color diagram ( figure [ pic_tau ] ; model lines shown in this figure are described in detail in section [ sec_models ] ) .
for example , the si feature always appears strongly in absorption in ulirgs , implying a deeply embedded source of energy with the temperature steeply declining toward the observer @xcite .
the total amount of obscuring material is such that is optically thick even in mid - ir . on the contrary
, it was recently confirmed that type 1 quasars show the si feature in emission @xcite .
the emission feature is thought to be produced by the thin hot layer of dust on the inner side of the obscuring material , where it is illuminated and heated by the central engine .
the hot dust emission region is seen without intervening absorption in type 1 quasars . in type 2
quasars from our sample the silicate feature centered at 9.7 is detected in eight spectra in absorption , and in the remaining four it is not detected , with upper limits of @xmath44 . as shown in figure [ pic_tau ] , the same is true for other samples of type 2 quasars .
a possible exception is sdss j0920 + 4531 ( figure [ pic_0920 ] ) .
this object has a hump in its mid - ir spectrum centered at about 11 , which may be due to si emission . with this assumption ,
we estimate the color temperature of this emission that is , we fit the continuum - subtracted spectrum @xmath45 with thermal emission models @xmath46 , where @xmath47 is the black - body spectrum with temperature @xmath48 .
our best fit ( @xmath49k with @xcite opacity curve ) is shown in figure [ pic_0920 ] .
this temperature is significantly lower than those typically derived for type 1 quasars showing si emission ( @xmath50k , @xcite ) ; such a low temperature is required to account for the shift of the feature centroid to long wavelengths .
the peak strength of the emission is @xmath51 , but there is no flux excess at 9.7 due to si emission , so we use @xmath52 for this object .
we now investigate the relationship between the strength of si feature and ir colors .
@xcite find a strong correlation between the depth of the si feature and the @xmath53/f_{\nu}[27.5\micron])$ ] color in their compilation of ulirgs , seyfert galaxies and type 1 quasars , as shown in figure [ pic_tau ] .
we therefore look for relationships between the values of @xmath36 and ir colors within our sample of type 2 quasars . to this end , for each object we compile all available ir photometric data from our observations ( zakamska et al .
, in prep . ) , from iras where applicable ( section [ sec_data
] ) , and from the irs spectra themselves ( from which we extract fluxes at rest - frame 14.5 ) .
we then model the observed spectral energy distribution as a third - order polynomial in the @xmath54 space , interpolate to calculate model fluxes at rest - frame 3.0 , 4.0 , 5.5 , 8.0 , 14.5 , 27.5 and 50.0 , and compute all 21 rest - frame flux ratios .
we then calculate the spearman rank correlation coefficient ( @xmath55 ) between values of @xmath36 and these ir colors for our sample of twelve objects .
we find that the values of @xmath55 range from 0.2 to 0.75 , depending on the color in question , with the most common value around 0.5 .
the sense of these correlations is that the si 9.7 absorption feature is stronger in colder objects , as @xcite found .
most correlations are rather weak ( the values @xmath56 imply that there is a @xmath57 chance to find such relationships in uncorrelated datasets ) , but two are significant at a @xmath58 level .
these are @xmath36 vs. @xmath59/f_{\nu}[8.0\micron]$ ] and @xmath36 vs. @xmath60/f_{\nu}[50\micron]$ ] ; they are shown in the bottom panels in figure [ pic_tau ] . in section [ sec_models ]
we discuss the implications of these findings and the geometrical and physical factors that determine the positions of objects in this diagram .
it would be interesting to determine the relationship between gas column density and dust column density in the obscuring material in agns .
the column density of neutral gas can be probed directly by x - ray observations .
the strength of the si feature would be a direct measure of the dust optical depth only if the dust were so cold that its emission at 10 were negligible . in practice
dust exists at a range of temperatures and both emits and absorbs in the mid - ir , so that @xmath36 reflects some combination of optical depth and temperature distribution .
@xcite find a correlation ( with a large scatter ) between the strength of the si feature and the neutral gas column density measured using x - ray observations . for 11
out of 12 objects in our sample , pointed _
chandra _ and _ xmm _ observations are available ( table [ tab_summary ] )
. in most cases , only a few photons are detected , and we can not derive a reliable measurement of the column density from x - ray spectra . instead , in figure [ pic_lx ] we show the strength of the si feature vs. the x - ray flux of our objects , normalized to their [ oiii]5007 flux presumed to be an isotropic measure of the bolometric luminosity .
we calculate x - ray fluxes expected in a type 1 quasar based on the [ oiii]5007 flux .
the mean ( thick solid line ) and the 1@xmath35 range ( dotted lines ) of the @xmath61[oiii ] ratio are taken from @xcite .
we then use the observed x - ray/[oiii]5007 ratio as a crude measure of the neutral gas column density ( the grey lines show the expected ratio for two values of the column density ) .
we see that half of our sample likely has very high column densities , @xmath62 @xmath16 @xcite , including the two objects with the deepest si absorption .
although our method of estimating @xmath63 is rather crude , we also see that type 2 quasars occupy a similar range of the @xmath36-@xmath63 parameter space as do seyfert 2 galaxies @xcite .
mid - ir spectra of some deeply absorbed sources , both galactic ( e.g. , protostars ) and extragalactic ( e.g. , ulirgs ) , display absorption due to organic molecules and inorganic ices ( e.g. , @xcite and references therein ) .
as we show in figure [ pic_ices ] , in sdss j0815 + 4304 we detect a broad absorption feature between 6 and 8 and within it a pair of narrow features an absorption complex quite similar to the one seen in the ulirgs iras f00183 - 7111 @xcite and iras 08572 + 3915 @xcite .
we identify the two narrow absorption features centered at 6.85 and 7.27 with bending ( deformation ) modes of ch bonds in the ch@xmath0 and ch@xmath64 subgroups of the aliphatic ( chain - like ) component of organic molecules within the dust . in figure
[ pic_ices]a , we compare the observed absorption with the opacity curve of hydrogenated amorphous carbon measured in the laboratory @xcite . using normalizations from @xcite and @xcite and peak optical depths of the two features ( @xmath65=0.19 $ ] and @xmath65=0.12 $ ] ) , we find column densities of @xmath66 @xmath16 .
this number is quite high ; indeed , taking the abundance of carbon to be @xmath67 relative to hydrogen and assuming that 10% of it is locked into aliphatic chains @xcite , we get a neutral hydrogen column density of @xmath68 @xmath16 .
this value may be compared with @xmath62 @xmath16 estimated from the lack of an x - ray detection in this object ( figure [ pic_lx ] and @xcite ) .
the difference between the two measurements is indicative of large uncertainties in all the steps of our calculation ( such as the conversion between optical depth of the features and @xmath69 , @xmath70 , as well as carbon abundance and the fraction of carbon locked in the chains ) .
it is interesting that the peak optical depth at 6.85 that we derive for sdss j0815 + 4304 ( which is a good compton - thick candidate ) is rather modest in comparison with values @xmath71 seen in some ulirgs @xcite which are therefore likely to have even higher column densities .
water ice absorption and organic molecules ( including the stretching modes of cc bonds in ring - like aromatic components ) can contribute at wavelengths @xmath72 .
the precise shape of the opacity curve at @xmath73 is difficult to determine because it is affected by the choice of the continuum and by assumptions about the strength of the 7.7 pah feature .
assuming a power - law continuum and no pah emission , we derive the opacity curve shown in figure [ pic_ices]b .
for the organic component , a broad range of opacity curves has been obtained in the laboratory at these wavelengths @xcite , but none appears to have the required shape and strength relative to the narrow aliphatic features ( a typical opacity curve of laboratory hydrocarbons is shown as the solid line in figure [ pic_ices]b ) .
a water ice featurelab / databases/ ] with a peak optical depth @xmath74=0.30 $ ] can reproduce the absorption at @xmath75 , but fails at longer wavelengths ( dotted line in figure [ pic_ices]b ) .
more importantly , there is no evidence for the stronger 13 water ice absorption feature in our spectrum ( @xmath76/\tau[6\micron]\ga 2 $ ] for the water ice opacity curve ; @xcite ) , which rules out water ice as the dominant source of opacity at 6 .
the specific carrier of the excess @xmath77 opacity ( which has also been observed in ulirgs , @xcite ) remains unidentified .
many galactic and extragalactic objects show strong broad emission features at 313 which are thought to originate from small carbonaceous dust particles ( @xcite ; see top panels of figure [ pic1 ] for the template of this emission ) .
in contrast to the aliphatic ( chain - like ) molecules responsible for the absorption bands discussed in the previous section , these emission features are due to polycyclic aromatic hydrocarbon ( pah ) molecules in which @xmath78 carbon atoms are arranged on a hexagonal lattice .
the different features arise from different c - c and c - h bending modes @xcite , and the precise spectral positions and ratios of the bands depend on which specific molecules are present , on their ionization state and on the spectral shape of the illuminating emission .
pah emission is powered by uv photons , and therefore can in principle be powered either by star - forming regions or by the active nucleus .
several arguments favor the hypothesis that pah emission predominantly traces star formation rather than agn activity .
in particular , 7.7 pah emission is spatially extended in nearby agns and is suppressed near the nucleus @xcite , presumably because of the destruction of pah molecules by extreme - uv and x - ray photons from the agn @xcite .
furthermore , the near - to - far - ir spectral energy distribution of agns with pah emission is accurately represented by a sum of a non - pah agn template and a star - burst galaxy template @xcite .
finally , for quasars with co emission , the co - pah correlation is in agreement with that of star - forming galaxies @xcite .
since pah[7.7 ] is blended with pah[8.6 ] , si absorption , the [ nevi]7.63 emission line and potentially ice / hydrocarbon absorption as well , we use the strength of the 11.3 pah feature as a star formation indicator . based on the fits described in section [ sec_features ]
, six objects were judged to have a detected 11.3 pah feature . in table
[ tab_summary ] , we list the average of the fluxes obtained by the two methods ( global fitting with pahfit and local fitting of a continuum+drude profile model ) and assume that they are accurate to no better than 30% .
we also assume that the pah emission is not affected by si absorption ( i.e. , that it originates from outside the region affected by absorption ) .
if this is not the case , then the 11.3 luminosities listed in table [ tab_summary ] are underestimated .
@xcite proposed a classification scheme in which they used the ratio of peak ( continuum subtracted ) pah[7.7 ] flux density to 7.7 continuum to distinguish between starburst- and agn - powered sources , and a similar idea is employed in the ` fork diagram ' of @xcite .
a low pah / mid - ir continuum ratio is characteristic of agns , because the spectral energy distributions of star - forming galaxies peak at much longer wavelengths than those of agns ( figure [ pic_tau ] ) . by the criteria of @xcite , all our objects are agn - dominated in the mid - ir .
the pah / continuum ratio can then serve as a crude estimate of the relative contribution of star formation in the host galaxy to the total ir energy budget .
figure [ pic_pah ] shows an anti - correlation between the [ oiii]5007/[oii]3727 line ratio and the pah[11.3]/continuum[27.5 ] ratio , as well as between [ oiii]5007 / h@xmath79 and pah[11.3]/continuum[27.5 ] , which are detected with about 95% confidence .
these anti - correlations suggest that for quasars [ oiii]/[oii ] and [ oiii]/h@xmath79 decrease as the star formation in the host increases , in agreement with classical work on optical line diagnostics @xcite .
we roughly estimate the total luminosity associated with star formation using the relation @xmath80/10^{10}l_{\odot})=1.16\times\log
\left(l[{\rm pah}11.3\micron]/10^{10}l_{\odot}\right)+2.38 $ ] @xcite .
these values are listed in table [ tab_summary ] . in figure
[ pic_lfsf ] , we show the comparison between the distribution of estimated star - formation luminosity in our objects ( histograms ) and the luminosity function of star formation in other agns and in field galaxies from a variety of samples as described in the figure caption . the median star - formation luminosity of our sample ( @xmath1 , which includes objects with upper limits on pah[11.3 ] ) is higher than that of field galaxies and of any other agn sample . as high as the star formation luminosities are , in all but
one case they are less than half of the total infrared luminosity obtained by directly integrating the observed ir spectral energy distribution . thus most of the ir luminosity is due to the dust - reprocessed emission from the quasar ( the one exception is discussed below )
. models of agn obscuration ( e.g. , @xcite ) show that dust - reprocessed emission is anisotropic even at mid - ir wavelengths , and the luminosities that we obtain in type 2 quasars by integrating the ir emission , multiplying by @xmath81 and subtracting the star - formation contribution underestimate the quasar bolometric luminosity , by a factor which depends on the viewing angle ( see figure 4 of @xcite ) . when we average model fluxes from @xcite over all angles such that the central source is not directly seen , we find that this factor ranges from 2.2 to 4.3 .
this implies that the true bolometric luminosities of the quasars in our sample are likely much larger than the star - formation luminosities of their hosts .
furthermore , in all cases the observed radio luminosity is stronger than what is expected from star formation alone @xcite , and most of the radio luminosity is due to the quasar , even though all the objects in our sample ( except sdss j0812 + 4018 ) are radio - quiet ( based on their position in the @xmath82[oiii ] diagram ; @xcite ) . in the radio - loud type 2 quasar sdss j0812 + 4018 , the star - formation luminosity estimated from pah[11.3 ] is similar to the total ir luminosity . in this object , the 11.3 feature appears anomalously strong compared to the 7.7 one , and therefore the star - formation luminosity is probably overestimated .
although it is usually thought that high - luminosity agns destroy pahs in their vicinity @xcite , the processes that determine pah feature ratios are not well - understood , and it is possible that the emission from the quasar itself boosts the 11.3 feature in this case , as is seen in some low - luminosity agns @xcite .
an anomalously strong 11.3 pah feature is seen in some other extragalactic objects ( e.g. , in a buried agn studied by @xcite ) .
we detect a number of ionized elements through their fine - structure emission lines ( the lines that we have considered are enumerated in the top panels of figure [ pic1 ] ) .
just as certain optical line ratios are diagnostic of the physical conditions in the medium where they are produced and of the intensity and spectrum of the ionizing radiation field @xcite , some mid - ir line ratios can also be used as such diagnostics @xcite . with ionization potentials of 126 ev and 97 ev ,
the nevi and nev species are produced exclusively by the active nucleus , whereas neii ( ionization potential 22 ev ) can be excited both by the nucleus and by massive stars . therefore , the [ nevi]/[neii ] and [ nev]/[neii ] line ratios are diagnostic of the agn contribution to the line emission . in addition , for a fixed shape of the agn spectrum , these line ratios increase with increasing ionization parameter ( i.e. , radiation density , @xcite ) .
we detect the [ nevi]7.63 emission line in 11 out of our 12 spectra . in objects with pah emission , identification of [ nevi]7.63
is somewhat complicated by the fact that it is right on top of the pah[7.7 ] feature . the median [ nevi]7.63/[neii]12.8 ratio in our objects is @xmath83 , similar to the sample of seyfert galaxies by @xcite , where the median is 0.9 . the [ nev]14.3 emission line
is also detected in 6 out of the 7 objects in which the wavelength coverage of this line is available .
the median [ nev]14.3/[neii]12.8 ratio is about 1 , similar to that of the sample of @xcite and somewhat larger than that of seyfert galaxies from @xcite who found a median ratio of 0.5 .
the simple ` linear mixing ' model of an agn+starburst spectrum by @xcite shows that our median value of [ nev]14.3/[neii]12.8@xmath84 roughly corresponds to a 100% agn contribution , and our lowest value of this ratio ( 0.5 in sdss j0050 - 0039 ) implies an agn contribution of 50% . since [ nev ] and [ nevi ] are produced in close proximity to the nucleus , they are more affected by extinction in the host galaxy or in the circumnuclear material than [ neii ] .
thus , our measured [ nev]/[neii ] and [ nevi]/[neii ] ratios are in fact lower limits , and so is the agn contribution calculated from the ` linear mixing ' model .
we do not find a strong correlation between the ne line ratios and the pah equivalent widths .
one likely reason is that a range of ionization parameters is present within our sample , so the dependence of ne line ratios on the star formation contribution is washed out . the one emission line that shows significantly different properties in type 2 quasars and seyfert galaxies is [ siv]10.51 ( ionization potential 35 ev ) .
this line is strongly detected in 11 of our 12 objects , and the median ratio is [ siv]10.51/[neii]12.8@xmath85 ( uncorrected for si absorption ) .
this ratio is significantly higher than that of the seyfert galaxies in the sample of @xcite ( median@xmath86 ) and higher still than that of @xcite ( median@xmath87 ) .
perhaps this difference is due to the larger ionization parameter , since the quasars in our sample are significantly more luminous ( median @xmath11(ir)@xmath88erg / s ) than the seyfert galaxies in the comparison samples ( e.g. , @xmath89 erg / s in @xcite ) .
rotational transitions of molecular hydrogen ( h@xmath0 ) allow us to directly observe warm ( @xmath90k ) molecular gas and estimate its excitation temperature and mass ( e.g. , @xcite ) .
the molecule is symmetric , so the transitions are quadrupolar and therefore weak , and their detectability is limited by the low signal - to - noise ratio of our spectra .
furthermore , some of the hydrogen lines blend with other spectral features , which is especially problematic because of our poor spectral resolution . despite these difficulties ,
we detect several rotational h@xmath0 lines in two objects ( table [ tab_hydro ] and figure [ pic_h2 ] ) .
the confidence of the detections , uncertainties in fluxes and upper limits are all calculated using simulations of weak unresolved emission lines in noisy spectra . for a line s@xmath91 ,
the transition is from the upper level with angular momentum @xmath92 to the lower level with angular momentum @xmath93 . under the assumption that emission is purely spontaneous and without correcting for extinction , the line flux @xmath94 can be directly related to the population of the @xmath92 level : @xmath95 where @xmath96 is the luminosity distance to the source , @xmath97 are the einstein coefficients ( e.g. , @xcite ) and
@xmath98 is the energy of the transition .
the rotational energy levels are given by @xmath99 where @xmath100 is the degeneracy of the levels , @xmath101 for even @xmath93 and @xmath102 for odd @xmath93 ( assuming equilibrium ortho - to - para ratio ) and @xmath103 is the boltzmann constant @xcite . in this expression ,
the first term is the quantum - mechanical analog of the energy of a rotator with angular momentum @xmath93 . as @xmath93 increases
, the energy of the ground vibrational level is changed by rotational stretching , which is taken into account by the second term @xcite .
if different levels are populated in a boltzmann equilibrium , the line ratios are indicative of the excitation temperature which is defined as @xmath104 in the spectrum of sdss j1641 + 3858 , we detect s(3)9.665 , s(4)8.026 and possibly s(5)6.909 , allowing us to calculate an excitation temperature for these transitions . the s(1)17.03 and s(4 ) lines are detected in sdss j0050@xmath220039 with a lower confidence . in figure
[ pic_h2 ] , we show the excitation diagrams for both objects , i.e. , we plot @xmath105 from equation ( [ eq_population ] ) as a function of the energy of the @xmath93-th level . in sdss
j1641 + 3858 , we find an excitation temperature of about 700k , somewhat higher than the temperatures found by @xcite in a large sample of ulirgs ( 300 - 400k ) . this difference should not be interpreted as a difference in physical conditions , but rather as a selection effect specifically , our lack of longer wavelength coverage precludes us from making observations of lower @xmath93 transitions which trace colder gas . indeed , in sdss j0050@xmath220039 the excitation temperature derived from s(1 ) and s(4 ) is lower ( @xmath106420k ) . using these excitation temperatures
we can calculate the mass of the gas .
levels are populated according to equation ( [ eq_exc ] ) with the normalization given by equation ( [ eq_population ] ) .
summing up molecules with all @xmath93 yields a mass of @xmath107 in sdss j1641 + 3858 ( emitting at temperature 700k ) and of @xmath108 in sdss j0050@xmath220039 ( emitting at temperature 420k ) . for comparison , in several ulirgs with five or more detected h@xmath0 rotational lines
@xcite , the gas is found at a range of excitation temperatures , with @xmath109 of warm gas ( @xmath110k ) and @xmath111 of cool gas ( @xmath112k )
. there could be large amounts of even colder gas which would be invisible in rotational transitions of h@xmath0 .
for example , if one of our objects contains @xmath113 of cold gas with @xmath114k , the population of @xmath115 levels is so low that even the lowest energy transition s(0)28.22 ( @xmath116 ) is well below the typical noise level of our data , as well as of the data by @xcite .
such cold gas can be traced only through lower energy transitions of other molecules , such as co(1 - 0 ) .
although h@xmath0 rotational line emission is often associated with star formation ( and then it is strongly correlated with pah emission , @xcite ) , there is no evidence for pah emission in sdss j1641 + 3858 .
if h@xmath0 emission originates in the circumnuclear region ( for example , @xmath117pc ) , then the column density of the molecular gas responsible for the observed emission is rather high , @xmath118\ga 5\times 10^{22}(\frac{d}{100{\rm pc}})^{-2}$ ] @xmath16 , and yet it maintains an excitation temperature of 700k . as in sdss j1641
+ 3858 , h@xmath0 emission in lower - luminosity agns is often in excess of what is expected from star formation in the host galaxy @xcite .
there is at the moment no conclusive model for h@xmath0 emission in agns .
in this section we compare the ir properties of optically selected type 2 quasars with those selected by other methods .
we discuss ir - selected ( section [ sec_comp_ir ] ) , radio - loud ( section [ sec_comp_rad ] ) and x - ray selected type 2 quasars ( section [ sec_comp_x ] ) .
low - redshift ( @xmath119 ) quasars show roughly power - law spectral energy distributions in the irac bands , unlike nearby star - forming galaxies whose strong pah emission dominates the 8 band .
the consequent difference in the mid - ir colors allowed @xcite and @xcite to select tens of candidate type 2 quasars , which were then confirmed by optical spectroscopy .
@xcite present irs data for some of the objects from that sample ; their si absorption and ir colors are shown in figure [ pic_tau ] as cyan squares .
they also calculate the total pah luminosity for the objects in their sample . on the basis of the pah template described in section [ sec_features ]
, we estimate that the luminosity of the 11.3 feature is 0.12 times that of the total pah luminosity , and therefore the median 11.3 luminosity of the ir - selected objects is @xmath120 the same as for our sample .
combining irs and hst imaging data for the ir - selected type 2 quasars , @xcite find that the depth of the si feature is correlated with the disk orientation , in the sense that objects that appear more edge - on have a significantly deeper feature .
these authors conclude that some of the obscuration occurs in the plane of the host galaxy , rather than in the circumnuclear material on much smaller spatial scales ( unless the circumnuclear material is aligned with the plane of the host galaxy ) . in figure [ pic_orientation ]
we show both samples ( lacy s and ours ) on the @xmath36 vs. axis ratio diagram .
morphologies and axis ratios are available from hst images for six of our objects @xcite
. the ir - selected objects ( open squares ) tend to show that the si absorption is stronger for the more inclined galaxies , but the trend disappears when our optically - selected type 2 quasars are included ( solid circles ) .
there is a significant difference between the host properties of the ir - selected type 2 quasars and the objects in our sample .
the former lie in disk - dominated , disturbed galaxies with dust lanes and active star formation . on the contrary ,
all our objects are hosted by bulge - dominated galaxies ( 2/3 of the hosts are ellipticals ) .
since the ir - selected sample is hosted by disk galaxies , the axis ratio is a good measure of inclination in the ir - selected sample , but not in ours , and the lack of correlation in figure [ pic_orientation ] is unsurprising .
additionally , [ neii]12.8/[oii]3727 ratio is on average 0.8 dex higher in the sample of @xcite than in our objects .
this difference is suggestive of a higher optical extinction in the host galaxies of the ir - selected objects than in ours , as supported by observations of their optical morphology .
why is there such a difference between the host properties of the two samples ?
the redshift ranges of both samples are similar , and so are the values of si absorption , ir colors and star - formation luminosities . optical properties ( such as line ratios ) are also indistinguishable .
the only significant difference seems to be in the luminosities of the two samples .
indeed , the median [ oiii]5007 luminosity of the objects from @xcite is @xmath121[oiii]/@xmath122 ( calculated from their published [ oii]3727 luminosities using the [ oii ] vs. [ oiii ] correlation by @xcite ) , while it is @xmath121[oiii]/@xmath123 for the objects in our sample .
similarly , the median 27.5 monochromatic ir luminosity for lacy s sample is @xmath124 erg / s , while it is @xmath125 erg / s for our sample .
it appears that there is a qualitative change in the host properties of type 2 quasars occurring at a luminosity bracketed by the two samples .
as it is possible that some subtle selection effect leads to the observed difference in host properties , this luminosity dependence clearly warrants further study . at higher redshifts ( @xmath126 ) , type 2 quasars may be selected via a combination of ir and radio properties .
strong radio emission ensures that the object is an active nucleus ( in the ` radio intermediate ' regime ) rather than a star - forming galaxy , while a large mid - ir / near - ir ratio is used to pre - select obscured rather than unobscured objects @xcite .
the irs spectra of type 2 quasars selected using this method are presented by @xcite ; about half of these objects were confirmed by optical spectroscopy .
none of the objects show the 10 si feature in emission , and roughly a third appear featureless , similar to our sample . because of the higher redshift , the irs spectra of these sources do not cover wavelengths longward of the 10 si absorption feature , so we do not show these objects on the si strength - ir color diagram , but the range of si absorption in these objects seems to be comparable to that of our sample .
the ` high - pah ' subsample of the objects from @xcite has pah[11.3 ] luminosities of @xmath127 similar to the highest star - formation luminosities in our sample ( table [ tab_summary ] ) .
this subsample appears to be well - matched in ir luminosity to the objects in our sample , while the ` no - pah ' subsample is 3 - 4 times more luminous .
@xcite present a spectroscopic study of very luminous ( @xmath128>10^{45.6}$ ] erg / s ) high - redshift ( @xmath129 ) obscured quasars .
these objects were selected from large - area surveys based on their rest - frame nir colors and a large ir - to - optical ratio , and therefore optical spectroscopy is not available for most sources .
those of the sources that showed 10 si absorption made up their sample of 21 obscured quasars .
on the basis of this sample , the authors find that obscured quasars with si absorption constitute more than half of the quasar population even at the highest luminosity end .
featureless ir spectra are common in the optically - selected and ir - selected samples of type 2 quasars ( and in the following sections we show that they are common in radio - selected and x - ray selected samples as well ) , and therefore it is possible that yet more obscured quasars remain to be discovered at these high luminosities and redshifts . in figure [ pic_tau ]
we show si strengths and ir colors of radio - loud agns from the sample of @xcite , with filled magenta triangles for broad - line radio galaxies and filled blue triangles for narrow - line radio galaxies .
si strengths are estimated from the published spectra by computing @xmath130/f_{\nu}[9.7\micron,\mbox{extrapolated continuum}])$ ] . again , although a quarter of these objects show spectra with no si absorption or emission , in the remaining cases the broad - line objects show si in emission and the narrow - line objects show si in absorption .
the median [ oiii]5007 luminosity of the narrow - line radio galaxies in this sample is @xmath131[oiii]/@xmath132 , so these objects are roughly matched in luminosity to the ir - selected sample of @xcite and are less luminous than ours .
a larger sample of radio - loud agns with irs observations was compiled by @xcite .
more than half of their sample of 28 sources have no detectable si feature . in the remaining objects
, si appears in emission in radio quasars and in weak absorption in radio galaxies ( which in the optical either show narrow lines or no lines at all ) .
there are two exceptions to this trend : 3c343 , a radio quasar , shows si weakly in absorption , and 3c325 , a radio galaxy , shows si weakly in emission .
@xcite find weak broad lines and a red continuum in the optical spectrum of 3c325 , and they therefore classify this object as a reddened quasar rather than a radio galaxy . since the rest - frame uv opacity of dust is much larger than its ir opacity , a small amount of reddening could produce strong uv extinction , while the ir spectrum would be transmitted unaffected .
we further comment on the correspondence between the optical classification and the appearance of the si feature in the next section .
the range of si absorption appears to be limited in both samples of radio - loud objects , with more than half showing no feature at all and no objects having @xmath133 .
the distributions of @xmath36 in radio - quiet type 2 quasars ( taking our sample and that of @xcite ) and in radio galaxies in the sample of @xcite are different at @xmath134 level ( determined using the kolmogorov - smirnov test ) .
one possibility is that the contribution of jet to the mid - ir emission dilutes the spectrum , making the si feature weaker relative to the total continuum . a larger sample of radio galaxies with well - measured si strength is needed to investigate this difference in more detail
few , if any , objects in the samples of @xcite and @xcite show pah emission . in a large sample of 3cr objects ( including those from @xcite and @xcite ) ,
@xcite find a median star - formation luminosity of @xmath135 among objects with detectable pahs , and @xmath136 when the upper limits on pahs are accounted for .
this level of star formation is lower than in pg quasars or 2mass agns , and more than an order of magnitude lower than that of optically and ir - selected type 2 quasars .
ir spectra of seven x - ray selected type 2 agns were presented by @xcite .
a wide range of ir luminosities is represented , from @xmath137=10^{43.5}$ ] erg / s to @xmath138 erg / s . unlike our sample , the ir spectra of these objects are mostly featureless .
pah features are firmly detected in only one object , and in all sources the star - formation luminosity is constrained to be @xmath139 a few times @xmath140 significantly smaller than that of the objects in our sample .
none shows the 10 si feature in absorption ( in one of the seven cases the spectrum does not cover the wavelength range affected by the si feature ) .
one object ( lh 28b ) may have si in emission .
this source shows broad lines with a large balmer decrement in its optical spectrum @xcite , so it may be another case of a reddened agn , similar to 3c325 described above .
one possible reason for the difference in the appearance of the si feature is that the two samples probe different ranges of column densities . of the seven objects in the x - ray selected sample , only one ( cdf - s 202 ) has @xmath141 @xmath16 , and that is the object in which the wavelength coverage of the si feature is not available .
the remaining objects have column densities below @xmath142 @xmath16 . although we do not have direct measurements of the column density in our sample , more than half of our objects ( and in particular , two objects with the deepest si absorption ) are probably compton - thick , as we discussed in section [ sec_si ] and as we show in figure [ pic_lx ] @xcite .
none of the four type 2 quasars in our sample with @xmath143 @xmath16 shows strong si absorption ( top right corner of figure [ pic_lx ] ) .
@xcite also find that deep si absorption is more likely to occur in compton - thick sources .
therefore , a bias against compton - thick sources in x - ray selected samples may result in a bias against deep si absorption in their ir spectra .
in this section we discuss the position of various types of sources ( agns and ulirgs ) on the @xmath36-ir color diagram ( figure [ pic_tau ] ) .
we start with the type 1/type 2 quasar sequence . in our hst data ,
some type 2 quasars clearly show conically shaped scattering regions in blue continuum light , with opening angles ( projected on the sky ) ranging from 5 to 60 @xcite . if we could determine the inclination of the scattering cones relative to the line of sight , we would know our viewing angle relative to the obscuring material .
for example , @xcite argued that in sdss j1039 + 6430 the scattering region to the south of the nucleus is oriented toward the observer because the opposing northern scattering region is fainter .
if so , we may be able to see down to hotter areas where dust is heated by the quasar , even if not deep enough to see the broad - line region itself
. this would be consistent with the absence of a si absorption feature in this object ( @xmath144 ) .
on the other hand , the two objects that have rather symmetric scattering regions in the hst images , sdss j1323@xmath220159 and sdss j1106 + 0357 , may be seen almost exactly edge - on .
both these objects show the si feature in absorption ( @xmath145 and @xmath146 , correspondingly ) .
sdss j0920 + 4531 , the only object in our sample with a possible si emission , is also the only object in which an extended scattering region could not be identified from the hst images @xcite .
it is likely that the compact blue source in the center of this object is the one responsible for the observed polarization , and it is possible that the scattering cone appears compact because it is viewed close to its axis ( almost face - on ) .
thus , we may be seeing a dependence of si strength on the viewing angle within the type 2 quasar population .
we also saw in the previous section that the appearance of the 10 si feature ( in emission or in absorption ) is well - correlated with the optical classification ( type 1 or type 2 ) .
these findings imply that the type 1/type 2 quasar sequence in figure [ pic_tau ] is driven at least in part by the viewing angle .
this strong relation between the appearance of the si feature and the viewing angle is rather surprising .
first , the broad - line region is much more compact [ @xmath147 from the virial argument for bulk motions ; cf .
reverberation mapping sizes by @xcite ] than the hot dust which is on scales larger than the region of dust sublimation [ @xmath148 from the heating vs. cooling energy balance for dust particles ] .
therefore , there might be plenty of lines of sight for which the broad line region is obscured while the hot dust region is not , with the si feature appearing in emission , as seen in some geometries in radiative transfer models by @xcite .
however , very few , if any , of the optically - classified type 2 quasars show the feature in emission .
the second complication is that the appearance of the si feature is very sensitive to the geometry of the torus , its total optical depth and the distribution of dust within it .
for example , models of compton - thick tori always show the feature in absorption @xcite , independent of orientation , because the torus is too optically thick to be heated through in the direction parallel to the axis of symmetry , and large temperature gradients are present along all lines of sight .
in clumpy tori @xcite the si appearance depends mostly on the distribution and sizes of clumps , rather than on the orientation of the observerclumpyweb/. ] .
when clumpiness is reduced and the volume - filling factor of clumps is increased , si shows a stronger dependence on the observer s orientation @xcite .
therefore , the observed relation between the optical classification and the appearance of the si feature may prove useful in constraining clumpy models of agn obscuration .
another characteristic of clumpy models of @xcite is that the shape of the si feature in the final spectrum is often significantly different from the shape of the feature in the opacity curve , due to a superposition of emission from regions with different temperatures .
this diversity is not seen in those of our spectra that show si in absorption , while the very low color temperature of the possible si emission in sdss j0920 + 4531 is a challenge to all existing models .
the high fraction of featureless spectra , the centroid and the shape of the 10 feature , the relative strengths of the 10 and the 18 features ( not available for most our objects , but employed in a ulirg study of @xcite ) and the relationship between si strength and the observed @xmath63 @xcite may prove to be useful diagnostics of agn models . the observed mid - ir @xmath149/f_{\nu}[27.5]$ ] color of all quasars in figure [ pic_tau ] has a median of [email protected] ( with 60% of objects falling within this range ) .
this is not well reproduced by either smooth or clumpy models , which typically predict @xmath149/f_{\nu}[27.5]>1 $ ] .
however , the models do predict that this color should be rather similar in type 1 and type 2 quasars , as observed ( type 1 quasars in figure [ pic_tau ] have a ratio of @xmath151 , while type 2 quasars have a ratio of @xmath152 ) . in order to bring the model color in agreement with observations
, one could postulate a cold dust component in the host galaxy .
as long as the temperature of this component is @xmath153k , its flux density at 14.5 is more than an order of magnitude smaller than that at 27.5 ( and it makes an even smaller contribution to the wavelengths of the si feature ) .
we can estimate the minimal mass of this component by assuming that it is an optically thin emitter [ @xmath154 and by requiring that it boosts the luminosity at 27.5 ( taken to be @xmath18 erg / s ) by a factor of two ( see figure [ pic_tau ] ) , which is sufficient to bring the @xmath149/f_{\nu}[27.5]$ ] from 1 to 0.5 . the milky way opacity curve from @xcite is used to convert between optical depth and hydrogen column density ( @xmath155 $ ] @xmath16 ) .
the minimal hydrogen mass is @xmath156 if the cold component has @xmath157k , but the required mass grows rapidly as the assumed temperature is decreased ( e.g. , @xmath158 at @xmath159k ) , since more mass is required to produce a given amount of flux at a colder temperature .
the mass may be underestimated by another factor of 4 if the gas - to - dust ratio is similar to the smc value rather than to the milky way value @xcite . because the radiative transfer models of circumnuclear obscuration in agns predict similar colors for type 1 and type 2 objects , it is impossible to distinguish between multi - temperature circumnuclear obscuration and intervening cold absorption on the basis of ir spectra alone .
indeed , the opacity curve of dust is rather wavelength - independent ( outside the strong si features at 10 and 18 ) , so intervening cold absorption does not change the colors in a significant way ( see the ` screen of cold dust ' absorption vector in figure [ pic_tau]a ) .
in other words , the ir spectra of type 2 quasars look just like they would if there was a screen of cold dust in front of a type 1 quasar .
the correlations between @xmath36 and ir colors found in our sample of 12 objects shown in figures [ pic_tau]b , c also seem consistent with the ` screen of cold dust ' model .
our hst observations of scattering cones strongly suggest obscuration close to the nucleus @xcite , so perhaps the outer parts of the obscuring torus are acting as this screen , while contributing emission at long wavelengths as described in the previous paragraph .
obscuration by intervening cold dust in the host galaxy seems to be important in the less luminous sample of @xcite , as we described above . in type 1
quasars , there is a well - defined maximum value of the strength of si emission , around @xmath160 .
to understand the origin of this maximum , we use dustymoshe / dusty/. ] to calculate emission of optically - thick slabs illuminated by the quasar spectrum on one side .
all incident radiation is absorbed in a thin layer near the illuminating source .
some of it is transmitted through the slab , while some of it is re - emitted toward the source .
the spectrum of this ` reflected ' component is almost independent of the optical depth , but depends somewhat on the assumed temperature at the illuminated surface , with si strength @xmath161 .
so if the temperature on the illuminated surface is close to the dust sublimation temperature ( @xmath162k ) , the model reproduces the observed maximum si strength .
the observed emission from quasars includes emission from the much colder outer parts of the obscuring material and thus is colder than the predicted reflected emission from the slab ( @xmath149/f_{\nu}[27.5]\simeq 2 $ ] at @xmath163k ) .
the seyfert galaxies in the sample of @xcite show a range of si absorption and emission similar to that of the higher - luminosity quasars , as well as similar ir colors ( @xmath149/f_{\nu}[27.5]=0.44^{+0.14}_{-0.20}$ ] ) .
however , the optical classification of seyferts does not correlate with the appearance of the si feature nearly as well as it does for quasars .
in particular , more than 50% of seyfert 1 galaxies show si in absorption , while only 10% of type 1 quasars do .
one possible explanation for this difference is that even a modest amount of star formation in a seyfert galaxy can significantly modify its ir spectrum , since the nucleus is not all that luminous .
this possibility is illustrated with a dotted line in figure [ pic_tau ] , where we show a locus of linear combinations of quasar and ulirg spectra to mimic an agn in a strongly star - forming galaxy .
we see that when a star formation contribution is increased , the ir colors rapidly become much colder , contrary to what is observed .
indeed , seyfert 2 galaxies from the sample of @xcite have a median color @xmath149/f_{\nu}[27.5]=0.38 $ ] and a median si strength @xmath164 , so they are warmer than the composite quasar+ulirg objects on the dotted curve .
another possible explanation for the difference between ir properties of seyfert galaxies and quasars is that there is some systematic geometric or physical difference between the two classes because of their different bolometric luminosity .
for example , it is possible that the opening angle of obscuration or the degree of clumpiness depend on luminosity , resulting in different relationships between the ir properties and the viewing angle in seyferts and quasars .
further theoretical work may clarify this interesting difference .
type 2 quasars are significantly warmer than are ulirgs for a given strength of si absorption .
radiative transfer models show that this difference is not due to the different shapes of the illuminating spectra , since all incident emission is absorbed within a thin layer close to the source @xcite .
the total optical depths are also thought to be comparable in quasars and ulirgs . instead
, the differing ir colors probably reflect different geometric distributions of dust in quasars and ulirgs .
if the energy sources in ulirgs ( starbursts or agns ) are completely embedded ( that is , covered by obscuring material along all lines of sight ) , we would not see any warm dust .
using dusty , we explore optically thick spherical shells illuminated on the inside for a range of shell geometries , optical depths and mass distribution , following @xcite .
we find that the strength of si absorption seen in ulirgs is reproduced by models with ir optical depth , @xmath37 $ ] , of order a few .
there is a trend of colder ir colors for less centrally concentrated shells , but all model colors ( for reasonable si strengths ) are warm , with @xmath149/f_{\nu}[27.5]$ ] between 0.1 and 1 .
it appears that a significant fraction of ulirgs ( those with colder colors ) require a cold dust component just as agns do .
we have conducted irs spectroscopy of ten type 2 quasars and obtained spectra of two more objects from the archives . the sample was selected from the sdss spectroscopic database on the basis of optical emission line properties and was required to have a minimal [ oiii]5007 luminosity of @xmath165 and a minimal ir flux to obtain good quality spectra , but was otherwise unbiased with regard to any multi - wavelength properties . in particular , the sample appears to represent a wide range of column densities along the line of sight , and half of the objects are likely to be compton - thick .
optically - selected type 2 quasars show a diversity of ir spectral properties .
most objects show high - ionization fine structure emission lines such as [ nevi]7.63 , essentially ruling out star formation as the dominant energy source .
the 10 si feature appears in absorption in eight of the objects with a large range of strengths , and in another three the feature is not detected . in the one remaining object ( sdss j0920 + 4531 )
the feature may have been detected in emission , peaking at 11 , which corresponds to an unusually low color temperature of @xmath166k .
half of the sample show pah features .
one type 2 quasar ( sdss j0815 + 4304 ) shows absorption due to organic carbonaceous grains and possibly water ice absorption .
we detect three rotational lines of molecular hydrogen in one object ( sdss j1641 + 3858 ) , with excitation temperature @xmath167k and mass @xmath168 .
molecular hydrogen is also detected ( with a lower confidence ) in sdss j0050@xmath220039 .
the median ir star - formation luminosity of type 2 quasar host galaxies inferred from pah emission is @xmath1 , but the bolometric output of our objects is dominated by the obscured quasars .
these levels of star formation are similar to those found by @xcite and @xcite in other samples of type 2 quasars .
the high star - formation luminosities of type 2 quasars much higher than in field galaxies suggest that quasar activity is associated with intensified star formation in its host galaxy , as suggested by merger - driven models of quasar evolution @xcite .
moreover , the median star - formation luminosity that we find is higher than that of type 1 quasar host galaxies , @xmath169 , depending on the quasar luminosity @xcite .
this trend supports the suggestion that the obscured fraction evolves with time , in the sense that the probability to see the quasar as obscured decreases with time after the merger @xcite . in this case
type 2 quasars would appear on average at an earlier stage of the merger and therefore with a higher star - formation rate than type 1 quasars , as observed .
objects in our sample occupy luminous bulge - dominated or elliptical hosts , whereas ir - selected type 2 quasars from the sample of @xcite inhabit dusty disk galaxies .
the two samples have different luminosities , with the objects from our sample being about 0.5 dex more luminous ( @xmath170=10^{45.3}$ ] erg / s in our sample vs. @xmath124 erg / s in the sample of @xcite ) .
@xcite argue that at @xmath171 most agn activity at low luminosities is due to gas instabilities in the host galaxy , while high luminosity quasars are fueled in major mergers which result in formation of bulges .
one possible explanation for the observed difference between host properties is that the two samples bracket a transition between different quasar fueling mechanisms .
both optically - classified type 2 and type 1 quasars often show spectra with no sign of the 10 si feature .
the fraction of such objects depends on how the sample was selected ; overall , about a third of all type 1 and type 2 quasars discussed in this paper do not show the feature .
most quasars that show si in absorption are optically classified as type 2 , while those with si emission are type 1 .
some models of clumpy agn tori show that the appearance of the feature is primarily determined by the geometrical distribution of the material and its total optical depth , while the viewing angle acts as a secondary parameter .
therefore , the observed correlation between the optical classification and the appearance of the ir spectrum may be a useful discriminator of models of agn obscuration .
we further find that the correlation is significantly weaker in seyfert galaxies than it is in quasars specifically , half of seyfert 1 galaxies show the 10 si feature in absorption .
this suggests that either the geometry or the physical properties of obscuring material vary with luminosity .
is supported by the space telescope fellowship provided by nasa through a contract issued by the jet propulsion laboratory , california institute of technology . l.g .
is grateful to the institute for advanced study and the princeton university observatory for their hospitality and financial support during the visit . l.g .
also acknowledges the support of conacyt , mxico .
m.a.s . acknowledges the support of nsf grant ast-0707266 .
we would like to thank bruce draine , brent groves , martin haas , lei hao , philip hopkins , mark lacy , yong shi , henrik spoon and the anonymous referee for the useful comments ; and lei hao , masatoshi imanishi and edward pier for providing data and models in electronic form .
this work is based on observations made with the infrared spectrograph ( irs ) onboard space telescope , which is operated by the jet propulsion laboratory , california institute of technology under a contract with nasa .
the irs is a collaborative venture between cornell university and ball aerospace corporation funded by nasa through the jet propulsion laboratory and ames research center .
the smart software was developed by the irs team at cornell university and is available through the science center at caltech .
funding for the sdss and sdss - ii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , the u.s .
department of energy , the national aeronautics and space administration , the japanese monbukagakusho , the max planck society , and the higher education funding council for england .
the sdss web site is http://www.sdss.org/. adelman - mccarthy , j. , et al .
2008 , , 175 , 297 allamandola , l.j . ,
tielens , a.g.g.m . , & barker , j.r .
1989 , , 71 , 733 antonucci , r. 1993 , ara&a , 31 , 473 armus , l. , soifer , b.t . , & neugebauer , g. 1999 , ap&ss , 266 , 113 armus , l. , et al .
2006 , , 640 , 204 baldwin , j.a .
, phillips , m.m . , & terlevich , r. 1981 , pasp , 93 , 5 becker , r.h . , white , r.l . , & helfand , d.j .
1995 , , 450 , 559 chiar , j.e . &
tielens , a.g.g.m .
2006 , , 637 , 774 cleary , k. , lawrence , c.r . , marshall , j.a . ,
hao , l. , & meier , d. 2007 , , 660 , 117 dartois , e. , et al .
2007 , a&a , 463 , 635 draine , b.t .
2002 , `` astrophysics of dust in cold clouds '' , in `` the cold universe '' , eds .
pfenninger & revaz ( berlin : springer - verlag ) draine , b.t .
2003 , , 41 , 241 franceschini , a. , et al .
2003 , , 343 , 1181 gerakines , p.a . , schutte , w.a . ,
greenberg , j.m .
, & van dishoeck , e.f . 1995 , a & a , 296 , 810 grimes , j.a . , rawlings , s. , & willott , c.j .
2005 , , 359 , 1345 groves , b. , dopita , m. , & sutherland , r. 2006 , a & a , 458 , 405 haas , m. , siebenmorgen , r. , schulz , b. , krgel , e. , & chini , r. 2005 , a & a , 442 , l39 hao , l. , et al . 2005a , , 129 , 1783 hao , l. , et al . 2005b , , 625 , l75 hao , l. , weedman , d.w . ,
spoon , h.w.w . ,
marshall , j.a . , levenson , n.a . ,
elitzur , m. , & houck , j.r .
2007 , , 655 , l77 heckman , t.m . ,
ptak , a. , hornschemeier , a. , & kauffmann , g. 2005 , , 634 , 161 helou , g. , soifer , b.t .
, & rowan - robinson , m. 1985 , , 298 , l7 higdon , s.j.u . , armus , l. , higdon , j.l . , soifer , b.t . , & spoon , h.w.w .
2006 , , 648 , 323 higdon , s.j.u . , et al .
2004 , , 116 , 975 hopkins , p.f . , hernquist , l. , cox , t.j .
, di matteo , t. , robertson , b. , & springel , v. 2006 , , 163 , 1 hopkins , p.f . , hernquist , l. , cox , t.j . , & kere , d. 2008 , , 175 , 356 houck , j.r . , et al .
2004 , , 154 , 18 huber , k.p . &
herzberg , g. 1979 , molecular spectra and molecular structure .
constants of diatomic molecules .
van nostrand - reinhold , new york imanishi , m. , dudley , c.c .
, maiolino , r. , maloney , p.r . , nakagawa , t. , & risaliti , g. 2007 , , 171,72 kemper , f. , vriend , w.j . , &
tielens , a.g.g.m .
2004 , , 609 , 826 kleinmann , s.g . , hamilton , d. , keel , w.c .
, wynn - williams , c.g . , eales , s.a . , becklin , e.e . , & kuntz , k.d .
1988 , , 328 , 161 knacke , r.f . &
thomson , r.k .
1973 , , 85 , 341 lacy , m. , et al .
2004 , , 154 , 166 lacy , m. , sajina , a. , petric , a.o . ,
seymour , n. , canalizo , g. , ridgway , s.e .
, armus , l. , & storrie - lombardi , l.j .
2007 , , 669 , l61 landau , l.d . &
lifshitz , e.m .
1991 , quantum mechanics : non - relativistic theory ( butterworth - heinemann oxford , boston ) le floch , e. , mirabel , i.f .
, laurent , o. , charmandaris , v. , gallais , p. , sauvage , m. , vigroux , l. , & cesarsky , c. 2001 , a & a , 367 , 487 le floch , e. , et al .
2005 , , 632 , 169 lehmann , i. , et al .
2000 , a&a , 354 , 35 levenson , n.a .
, sirocky , m.m .
, hao , l. , spoon , h.w.w . , marshall , j.a . , elitzur , m. , & houck , j.r .
2007 , , 654 , l45 lutz , d. , spoon , h.w.w . ,
rigopoulou , d. , moorwood , a.f.m . ,
& genzel , r. 1998 , , 505 , l103 maiolino , r. , ruiz , m. , rieke , g.h . , keller , l.d .
1995 , , 446 , 561 maiolino , r. , shemmer , o. , imanishi , m. , netzer , h. , oliva , e. , lutz , d. , & sturm , e. 2007 , a & a , 468 , 979 martnez - sansigre , a. , rawlings , s. , lacy , m. , fadda , d. , jarvis , m.j . ,
marleau , f.r . ,
simpson , c. , & willott , c.j .
2006 , , 370 , 1479 martnez - sansigre , a. , lacy , m. , sajina , a. , & rawlings , s. 2008 , , 675 , 676 nenkova , m. , ivezi , .
, & elitzur , m. 2002 , , 570 , l9 netzer , h. , mainieri , v. , rosati , p. , & trakhtenbrot , b. 2006 , a & a , 453 , 525 pendleton , y.j .
& allamandola , l.j .
2002 , a & a , 138 , 75 peterson , b.m .
1993 , pasp , 105 , 247 pier , e.a . & krolik , j.h .
1992 , , 401 , 99 polletta , m. , weedman , d. , hoenig , s. , lonsdale , c.j . ,
smith , h.e . , & houck , j. 2008 , , 675 , 960 ptak , a. , zakamska , n.l . , strauss , m.a . ,
krolik , j.h . , heckman , t.m . ,
schneider , d.p . , & brinkmann , j. 2006 , , 637 , 147 reyes , r. , zakamska , n.l . ,
strauss , m.a . ,
green , j. , krolik , j.h . ,
shen , y. , richards , g.t . , anderson , s. , & schneider , d.p .
2008 , , submitted ( arxiv:0801.1115 ) rigopoulou , d. , kunze , d. , lutz , d. , genzel , r. , & moorwood , a.f.m .
2002 , a & a , 389 , 374 risaliti , g. , gilli , r. , maiolino , r. , salvati , m. 2000 , a & a , 357 , 13 roussel , h. , et al .
2007 , , 669 , 959 schartmann , m. , meisenheimer , k. , camenzind , m. , wolf , s. , tristram , k.r.w . ,
& henning , t. 2008 , a&a , submitted ( arxiv:0802.2694 ) shi , y. , et al .
2006 , , 653 , 127 shi , y. , et al .
2007 , , 669 , 841 siebenmorgen , r. , haas , m. , krgel , e. , & schulz , b. 2005 , a & a , 436 , l5 sirocky , m.m . ,
levenson , n.a .
, elitzur , m. , spoon , h.w.w . , & armus , l. 2008 , , 678 , 729 smith , j.d.t . , et al . , 2007 , , 656 , 770 spoon , h.w.w . ,
keane , j.v . , tielens , a.g.g.m . , lutz , d. , moorwood , a.f.m . , & laurent , o. 2002 , a & a , 385 , 1022 spoon , h.w.w . , et al .
2004 , , 154 , 184 spoon , h.w.w . ,
marshall , j.a . ,
houck , j.r .
, elitzur , m. , hao , l. , armus , l. , brandl , b.r . , & charmandaris , v. 2007 , , 654 , l49 stanford , s.a . , stern , d. , van breugel , w. , de breuck , c. 2000 , , 131 , 185 stern , d. , et al .
2005 , , 631 , 163 stoughton , c. , et al .
2002 , , 123 , 485 sturm , e. , lehmann , i. , mainieri , v. , genzel , r. , lehnert , m.d . ,
lutz , d. & tacconi , l.j .
, 2006 , , 642 , 81 sturm , e. , lutz , d. , verma , a. , netzer , h. , sternberg , a. , moorwood , a.f.m . , oliva , e. , & genzel , r. 2002 , a&a , 393 , 821 szokoly , g.p . ,
2004 , , 155 , 271 teng , s.h . , wilson , a.s . , veilleux , s. , young , a.j . , sanders , d.b . , & nagar , n.m .
2005 , , 633 , 664 tommasin , s. , spinoglio , l. , malkan , m.a . ,
smith , h. , gonzalez - alfonso , e. , & charmandaris , v. 2008 , , submitted ( arxiv:0710.4448 ) treister , e. , krolik , j.h . , & dullemond , c. 2008 , , in press ( arxiv:0801.3849 ) turner , j. , kirby - docken , k. , & dalgarno , a. 1977 , , 35 , 281 ueda , y. , akiyama , m. , ohta , k. , & miyaji , t. 2003 , , 598 , 886 veilleux , s. & osterbrock , d.e .
1987 , , 63 , 295 vignali , a. , alexander , d.m .
, & comastri , a. 2004 , , 354 , 720 vignali , a. , alexander , d.m . , & comastri , a. 2006 , , 373 , 321 voit , g.m .
1992 , , 258 , 841 weingartner , j.c . & draine , b.t .
2001 , , 548 , 296 werner , m.w . , et al .
2004 , , 154 , 1 york , d.g . ,
2000 , , 120 , 1579 zakamska , n.l . ,
2003 , , 126 , 2125 zakamska , n.l . ,
strauss , m.a . , heckman , t.m . ,
ivezi , . , & krolik , j.h .
2004 , , 128 , 1002 zakamska , n.l . ,
2005 , , 129 , 1212 zakamska , n.l . ,
2006 , , 132 , 1496 ccccccccccc sdss [email protected] & 0.728 & 10.06 & -0.47 & 44.6 ( 3 ) & 46.09 & 46.37 & 41.0 & 10.1 & 12.5 & 46.4 + sdss j005621.72 + 003235.8 & 0.484 & 9.25 & -2.65 & 43.1 ( 1 ) & 45.25 & 45.37 & 40.9 & 9.5 & 11.9 & 45.6 + sdss j012341.47 + 004435.9 & 0.399 & 9.14 & 0.00 & 44.3 ( 1 ) & 44.75 & 44.65 & 40.9 & @xmath1398.8 & @xmath13911.0 & 45.0 + sdss j081253.09 + 401859.9 & 0.551 & 9.39 & -0.14 & 44.2 ( 1 ) & 44.99 & 44.95 & 43.2 & 9.7 & 12.0 & 45.3 + sdss j081507.42 + 430427.2 & 0.510 & 9.44 & -2.21 & @xmath13942.5 ( 3 ) & 45.37 & 45.31 & 40.8 & @xmath1399.2 & @xmath13911.5 & 45.8 + sdss j091345.49 + 405628.2 & 0.441 & 10.33 & -0.32 & 43.9 ( 1 ) & 46.48 & 46.48 & 40.8 & @xmath1399.3 & @xmath13911.6 & 46.7 + sdss j092014.11 + 453157.3 & 0.403 & 9.15 & 0.00 & 42.4 ( 1 ) & 45.10 & 45.13 & @xmath13939.8 & 9.4 & 11.7 & 45.5 + sdss j103951.49 + 643004.2 & 0.402 & 9.43 & 0.00 & 42.6 ( 1 ) & 45.33 & 45.29 & @xmath13939.8 & 9.3 & 11.6 & 45.7 + sdss j110621.96 + 035747.1 & 0.242 & 9.01 & -0.49 & 42.6 ( 1 ) & 44.44 & 44.43 & @xmath13939.4 & @xmath1398.4 & @xmath13910.6 & 44.5 + sdss j115718.35 + 600345.6 & 0.490 & 9.61 & 0.00 & @xmath13942.2 ( 3 ) & 45.78 & 45.78 & 40.2 & 10.0 & 12.3 & 46.3 + sdss [email protected] & 0.350 & 9.27 & -0.31 & n / a & 44.74 & 44.81 & 39.8 & @xmath1399.0 & @xmath13911.3 & 45.1 + sdss j164131.73 + 385840.9 & 0.596 & 10.04 & -0.45 & 44.8 ( 2,3 ) & 45.43 & 45.46 & 40.8 & @xmath1399.6 & @xmath13912.0 & 45.8 + [ tab_summary ] cccccc sdss j0050@xmath220039 & 5@xmath1502 & @xmath1395 & @xmath1393 & [email protected] & + sdss j1641@xmath93858 & n / a & @xmath1392.1 & [email protected] & [email protected] & [email protected] [ tab_hydro ] | type 2 quasars are luminous active galactic nuclei whose central engines are seen through large amounts of gas and dust .
we present spectra of twelve type 2 quasars selected on the basis of their optical emission line properties . within this sample ,
we find a surprising diversity of spectra , from those that are featureless to those showing strong pah emission , deep silicate absorption at 10 , hydrocarbon absorption , high - ionization emission lines and h@xmath0 rotational emission lines .
about half of the objects in the sample are likely compton - thick , including the two with the deepest si absorption .
the median star - formation luminosity of the objects in our sample measured from the strength of the pah features is @xmath1 , much higher than for field galaxies or for any other agn sample , but similar to other samples of type 2 quasars .
this suggests an evolutionary link between obscured quasars and peak star formation activity in the host galaxy . despite the high level of star formation
, the bolometric output is dominated by the quasar in all cases . for a given strength of 10 si absorption ,
ulirgs are significantly colder than are type 2 quasars ( their @xmath2/f_{\nu}[27.5\micron]$ ] ratio is 0.5 dex lower ) , perhaps reflecting different obscuration geometries in these sources .
we find that the appearance of the 10 feature ( i.e. , whether it shows in emission or in absorption ) is well - correlated with the optical classification in type 1 and type 2 quasars , contrary to some models of clumpy obscuration .
furthermore , this correlation is significantly stronger in quasars ( @xmath3 erg / s ) than it is in seyfert galaxies ( @xmath4 erg / s ) . |
although the concept of entropy plays important roles in diverse fields ranging from classical thermodynamics of clausius , @xcite quantum mechanics and uncertainty , @xcite black holes , @xcite coding and computation,@xcite to information technology , @xcite it does not seem to have a standard definition . here , we are interested in its application to nonequilibrium statistical thermodynamics . in classical thermodynamics , it is defined as a _ thermodynamic _ quantity with no association of any notion of a microstate and its probability , while modern approach to statistical thermodynamics , primarily due to boltzmann and gibbs , requires a probabilistic approach in terms of microstates . in this work ,
we are primarily interested in an isolated system @xmath10 .
quantities for the isolated system will carry a suffix @xmath11 ; quantities without the suffix will refer to any body , which need not be isolated , such as a part @xmath12 of @xmath10 ; see fig .
[ fig.sys ] .
the microstates for @xmath10 are determined by the set @xmath13 of extensive observables ( the energy @xmath14 , the volume @xmath15 , the number of particles @xmath16 , etc . ) specifying the isolated system .
while temporal evolution is not our primary interest in this work , we still need to remember the importance of temporal evolution in thermodynamics .
we will say that two microstates belonging to the microstate subspace @xmath17 are `` connected '' if one evolves from the other after some time @xmath18 . before this time
, they will be treated as `` disconnected . ''
let @xmath19 denote the maximum @xmath20 over all pairs of microstates .
the space @xmath17 is _ simply connected _ for all times longer than @xmath19 in that each microstate can evolve into another microstate @xmath21 in due time . for @xmath22 ,
the space @xmath17 will consist of _ disjoint components , _ an issue that neither boltzmann nor gibbs has considered to the best of our knowledge .
but the issue , which we consider later in sec .
[ marker_disjoint space ] , becomes important in considering nonequilibrium states .
[ ptb ] system_modified_1.eps boltzmann assumes _ equal probability _ of various microstates in the simply connected set @xmath17 .
thus , @xmath19 can be identified with the equilibration time @xmath23 for @xmath10 . under the equiprobable assumption
, boltzmann identifies the entropy in terms of the number of microstates@xcite @xmath24 in @xmath25 : @xmath26 we will set the boltzmann constant to be unity throughout the work so that the entropy will always be a _
pure number_. the idea behind the above formula implicitly appears for the first time in a paper @xcite by boltzmann , and then appears more or less in the above form later in his lectures@xcite where he introduces the combinatorial approach for the first time to statistical mechanics .
the formula itself does not appear but is implied when he takes the logarithm of the number of combinations.@xcite ( there is another formulation for entropy given by boltzmann,@xcite which is also known as the boltzmann entropy,@xcite that we will discuss later and that has a restricted validity ; see eq .
( [ boltzmann_s_1 ] ) . )
gibbs , also using the probabilistic approach , gives the following formula for the entropy in a _
canonical ensemble_:@xcite@xmath27 where @xmath28 is the canonical ensemble probability of the @xmath29th microstate of the system , and the sum is over all microstates corresponding to all possible energies @xmath30 ( with other elements in @xmath31 held fixed ) ; their set is denoted by @xmath32 in the above sum .
the gibbsian approach assumes an ensemble at a given instant , while the boltzmann approach considers the evolution of a particular system in time ; see for example a recent review.@xcite in equilibrium , both entropy expressions yield the same result . in quantum mechanics ,
this entropy is given by the von neumann entropy formulation@xcite in terms of the density matrix @xmath33:@xmath34 the entropy formulation in the information theory@xcite has a form that appears to be similar in form to the above gibbs entropy even though the temperature has no significance in the information theory .
there is also another statistical formulation of entropy , heavily used in the literature , in terms of the phase space distribution function @xmath35 , which follows from boltzmann s celebrated h - theorem:@xmath36 here , @xmath37 denotes a point in the phase space .
this quantity is not only not dimensionless but , as we will show later , is not the correct formulation in general ; see eq .
( [ discrete_s ] ) .
the classical thermodynamics entropy @xmath38 is oblivious to the microstates and their probabilities and deals with @xmath39 , etc . as the observables of the system . in equilibrium
, the entropy @xmath40is a state function , and can be expressed as a function of the observables .
this functional dependence results in the gibbs fundamental relation @xmath41 in terms of the observables @xmath42 . for a lattice model
, @xmath38 is non - negative in accordance with the boltzmann definition of @xmath43 , but is known to become negative for a continuum model such as for an ideal gas .
the latter observation implies that such continuum models are not realistic as they violate nernst s postulate(the third law ) and require _ quantum mechanics _ to ensure non - negativity of the entropy.@xcite even the change @xmath8 , the heat capacity , etc .
do not satisfy thermodynamic consequences of nernst s postulate . by invoking nernst s postulate ( the equilibrium entropy vanishes at absolute zero )
, one can determine the equilibrium entropy everywhere _
uniquely_. the consensus is that in equilibrium , the thermodynamic entropy is not different from the above statistical entropies due to boltzmann and gibbs .
however , there is at present no consensus when the system is out of equilibrium .
there is also some doubt whether the nonequilibrium thermodynamic entropy has any meaning in classical thermodynamics .
we will follow clausius and take the view here that the thermodynamic entropy is a well - defined notion even for an irreversible process going on in a body for which clausius@xcite writes @xmath44 in terms of the exchange heat @xmath45 with the medium .
the question that arises is whether the two statistical definitions can be applied to a body out of equilibrium .
we find the answer to be affirmative .
the next question that arises is the following : do they always give the same results ?
we will show that under certain conditions , they give the same results .
this is important as the literature is not very clear on this issue.@xcite for an isolated system , which we mostly consider here , we are not concerned with any thermostat or external working source . as a consequence ,
observables @xmath14,@xmath15,@xmath16 , etc . in @xmath13
must remain constant even if the system is out of equilibrium . while this will simplify our discussion to some extent
, it will also create a problem as we will discuss later .
we discuss the concept of a nonequilibrium state , state variables and state functions in the next section .
we introduce the statistical entropy formulation in sec .
iii and show its equivalence with thermodynamic nonequilibrium entropy when the latter is a state function . in sec .
iv , we carry out an explicit quantum calculation of the nonequilibrium statistical entropy . in sec .
v , we consider a @xmath2-d lattice model appropriate for tonks gas in continuum so that the statistical lattice entropy can be calculated rigorously .
we take the continuum limit and compare the resulting entropy with the continuum entropy of the tonks gas and obtain an interesting result .
we discuss the semi - classical approximation of the statistical entropy in sec .
vi and show that the formulation in eq .
( [ sf_s ] ) does not determine the entropy .
a brief summary and discussion is presented in the final section .
the second law states that the irreversible ( denoted by a suffix i ) entropy generated in any infinitesimal physical process going on in a body satisfies the inequality@xmath46 the equality occurs for a reversible process . for an isolated system , there is no exchange ( denoted by a suffix e ) entropy change @xmath47 with the medium so that @xmath48 in any arbitrary process . in an isolated system
satisfies@xmath49 the law refers to the _ thermodynamic entropy_. it is not a state function in the conventional sense ( @xmath50 ) if @xmath10 is not in equilibrium simply because as a state function @xmath50 must remain constant for constant @xmath13 ; see below also .
as the thermodynamic entropy is not measurable except when the process is reversible , the second law remains useless as a computational tool .
in particular , it says nothing about the rate at which the irreversible entropy increases .
therefore , it is useful to obtain a computational formulation of the entropy , the _ statistical entropy_. this will be done in the next section .
the onus is on us to demonstrate that the statistical entropy also satisfies this law if it is to represent the thermodynamic entropy .
this by itself does not prove that the two are the same .
it is not been possible to show that the statistical entropy is identical to the thermodynamic entropy in general . here , we show their equivalence only when the nonequilibrium thermodynamic entropy is a _ state function _ of nonequilibrium state variables to be introduced below . for an isolated body in equilibrium
, the entropy can be expressed as a function of its observables ( variables that can be controlled by an observer ) , as is easily seen form the gibbs fundamental relation in eq .
( [ gibbs_fundamental ] ) .
the thermodynamic state , also known as the macrostate @xmath51 of a body in equilibrium remains the same unless it is disturbed .
therefore , we can identify the equilibrium state @xmath52 of the body by its observables .
accordingly , its equilibrium entropy @xmath50 can be expressed as a function of its observables @xmath13 .
this is expressed by saying that the equilibrium entropy is a _ state function _ and @xmath13 is the set of _ state variables_. the above conclusion is most certainly not valid for a body out of equilibrium , which we take to be isolated .
if the body is not in equilibrium , its ( macro)state @xmath53 will continuously change , which is reflected in its entropy change ( increase ) in time ; this requires expressing its entropy as @xmath54 with an _ explicit time - dependence _ , since @xmath55 for an isolated body .
the change in the entropy and the macrostate must come from the variations of additional variables , distinct from the observables , that keep changing with time until the body comes to equilibrium as explained elsewhere.@xcite these variables can not be controlled by the observer .
once the body has come to equilibrium , the entropy has no explicit time - dependence and becomes a state function . in this state
, the entropy has its maximum possible value for given @xmath13 .
in other words , when the entropy becomes a state function , it achieves the maximum possible value for the given set of state variables , here given by @xmath13 .
this conclusion about the entropy will play an important role below .
we assume that there is a set @xmath56 of additional variables , known as the _ internal variables _ ( sometimes also called hidden variables )
. we will refer to the variables in @xmath13 and @xmath56 as ( _ nonequilibrium _ )
_ state variables _ ( see below for justification ) and denote them collectively as @xmath57 in the following . from theorem 4 presented
elsewhere,@xcite it follows that with a proper choice of the number of internal variables , the entropy can be written as @xmath58 with no explicit @xmath59-dependence .
the situation is now almost identical to that of an isolated body in equilibrium : the entropy is a function of @xmath60 with no explicit time - dependence .
this allows us to identify @xmath60 as the set of nonequilibrium state variables .
thus , @xmath53 can be specified by @xmath60 so that the entropy becomes a _
state function_. this allows us to extend eq .
( [ gibbs_fundamental ] ) to@xmath61 in which the partial derivatives are related to the fields of the system:@xmath62 these fields will change in time unless the system has reached equilibrium . as @xmath60 changes in time , @xmath53 changes , but at each instant the ( nonequilibrium ) entropy as a state function , has a maximum possible value for given @xmath60 even though @xmath63 . in our previous work,@xcite
we have identified this particular state as an _ internal equilibrium state _ , but its physical significance as presented above was not discussed . for a state that is not in internal equilibrium
, the entropy must retain an explicit time - dependence . in this case , the derivatives in eq .
( [ fields_isolated ] ) can not be identified as state variables like , temperature , pressure , etc .
it may appear to a reader that the concept of entropy being a state function is very restrictive .
this is not the case as this concept , although not recognized by several workers , is implicit in the literature where the relationship of the thermodynamic entropy with state variables is investigated . to appreciate this
, we observe that the entropy of a body in internal equilibrium@xcite is given by the boltzmann formula@xmath64 in terms of the number of microstates corresponding to @xmath65 . in classical nonequilibrium thermodynamics,@xcite the entropy is always taken to be a state function . in the edwards approach@xcite for granular materials , all microstates are equally probable as is required for the above boltzmann formula .
bouchbinder and langer@xcite assume that the nonequilibrium entropy is given by eq .
( [ boltzmann_s_extended ] ) .
lebowitz@xcite also takes the above formulation for his definition of the nonequilibrium entropy . as a matter of fact
, we are not aware of any work dealing with entropy computation that does not assume the nonequilibrium entropy to be a state function . this does not , of course , mean that all states of a system are internal equilibrium states . for states that are not in internal equilibrium ,
the entropy is not a state function so that it will have an explicit time dependence .
but , as shown elsewhere,@xcite this can be avoided by enlarging the space of internal variables .
the choice of how many internal variables are needed will depend on experimental time scales and can not be answered in generality at present .
we hope to come back to this issue in a future publication . for a general body that is not isolated
, the concept of its internal equilibrium state plays a very important role in that the body can come back to this state several times in a nonequilibrium process . in a cyclic nonequilibrium process
, such a state can repeat itself in time after some cycle time @xmath66 so that all state variables and functions including the entropy repeat themselves:@xmath67 this ensures @xmath68 in a cyclic process .
all that is required for the cyclic process to occur is that the body must start and end in the same internal equilibrium state ; however , during the remainder of the cycle , the body need not be in internal equilibrium .
we provide a very general formulation of the statistical entropy , which will also demonstrate that the entropy is a _ statistical average_. we consider a macrostate @xmath69 of @xmath10 at a given instant @xmath59 . in the following ,
we suppress @xmath59 unless necessary .
the macrostate @xmath70 refers to the set of microstates @xmath71 and their probabilities @xmath72 . for the computation of combinatorics ,
the probabilities are handled in the following abstract way .
we consider a large number @xmath73 of independent _ replicas _ or _
samples _ of @xmath10 , with @xmath74 some large constant integer and @xmath75 the number of distinct microstates @xmath76 .
the samples should be thought of as identically prepared experimental samples.@xcite let @xmath77 denote the sample space spanned by @xmath76 . as @xmath78 @xmath17 , @xmath77
is simply connected if @xmath17 is , which we assume in this section .
let @xmath79 denote the number of @xmath76-samples ( samples in the @xmath76-microstate ) so that@xmath80 the above sample space is a generalization of the _ ensemble _ introduced by gibbs , except that the latter is restricted to an equilibrium system , whereas our sample space refers to the system in any arbitrary state so that @xmath81 may be time - dependent
. in the semi - classical approximation , see sec .
[ marker_semiclassical ] , one can similarly take the sample space to represent the classical phase space of boltzmann
. the ( _ sample _ or _ ensemble _ ) _ average _ of some quantity @xmath82 over these samples is given by@xmath83 where @xmath84 is the value of @xmath82 in @xmath76 .
the samples are , by definition , _ independent _ of each other so that there are no correlations among them .
because of this , we can treat the samples to be the outcomes of some random variable , the macrostate @xmath53 .
this independence property of the outcomes is crucial in the following , and does not imply that they are equiprobable .
the number of ways @xmath85 to arrange the @xmath86 samples into @xmath75 distinct microstates is@xmath87 taking its natural log to obtain an _ additive _ quantity per sample@xmath88 and using stirling s approximation , we see easily that @xmath89 , which we hope to identify later with the entropy @xmath58 of the isolated system , can be written as the average of the negative of@xmath90 what gibbs @xcite calls the _ index of probability : _
@xmath91 where we have also shown an explicit time - dependence for the reason that will become clear below .
the above derivation is based on fundamental principles and does not require the system to be in equilibrium ; therefore , it is always applicable . to the best of our knowledge
, even though such an expression has been extensively used in the literature , it has been used _ without _ any derivation ; one simply appeals to this form by invoking it as the information entropy ; however , see sec .
[ marker_summary ] . because of its similarity in form with @xmath92 in eq .
( [ gibbs_s ] ) , we will refer to @xmath93 as the gibbs statistical entropy from now on .
as the nonequilibrium thermodynamic entropy for a process in which the system is always in internal equilibrium can be determined by integrating the gibbs fundamental relation in eq .
( [ gibbs_fundamental_extended ] ) , we can compare it with the statistical entropy introduced above .
however , such an integration is not possible for a process involving states that are arbitrary ( not in internal equilibrium ) .
therefore , there is no meaning to compare @xmath93 with the corresponding thermodynamic entropy whose value can not be determined . to identify @xmath94 with the nonequilibrium
thermodynamic entropy requires the following additional steps : 1 .
it is necessary to establish that @xmath0 satisfies eq .
( [ second_law ] ) .
2 . for an equilibrium canonical system
, it is necessary to establish that @xmath95 is identical to the equilibrium thermodynamic entropy given by @xmath92.@xcite 3 .
it is necessary to show that @xmath0 is identical to the nonequilibrium thermodynamic entropy of the system that is out of equilibrium but whose entropy is a state function .
there are several proofs available in the literature@xcite for ( 1 ) .
therefore , we will not be concerned with ( 1 ) anymore .
we will prove ( 2 ) and ( 3 ) in sect .
[ marker_open system ] .
the maximum possible value of @xmath0 for given @xmath60 occurs when @xmath76 are _ equally probable_:@xmath96 in this case , the explicit time dependence in @xmath0 will _ disappear _ and we have @xmath97 which is identical in form to the boltzmann ( thermodynamic ) entropy in eq .
( [ boltzmann_s ] ) for an isolated body in equilibrium , except that the current formulation has been extended to an isolated body out of equilibrium ; see also eq . (
[ boltzmann_s_extended ] ) .
the only requirement is that all microstates in @xmath98 are equally probable .
the statistical entropy in this case becomes a state function . applying the above formulation to a macrostate characterized by a given @xmath13 and consisting of microstates @xmath99 forming the set @xmath100 with probabilities @xmath101 , we find that @xmath102 is the entropy of this macrostate , where @xmath103 is the number of distinct microstates @xmath104
. it should be obvious that@xmath105 again , under the equiprobable assumption @xmath106 @xmath17 denoting the space spanned by microstates @xmath107 , the above entropy takes its maximum possible value@xmath108 which is identical in value to the boltzmann ( thermodynamic ) entropy in eq .
( [ boltzmann_s ] ) for an isolated body in equilibrium .
the maximum value occurs at @xmath109 .
it is evident that@xmath110\leq\mathcal{s}_{0}[\mathbf{z}_{0}(t)]\leq\mathcal{s}_{0}(\mathbf{x}_{0 } ) .
\label{entropy_inequalities0}\ ] ] the anticipated identification of nonequilibrium thermodynamic entropy with @xmath89 under some restrictions allows us to identify @xmath89 as the _ statistical entropy _ formulation of the thermodynamic entropy . from now on
, we will refer to the general entropy in eq .
( [ gibbs_formulation ] ) in terms of microstate probabilities as the _ time - dependent gibbs formulation _ of the entropy or simply the _ gibbs entropy _ , and will not make any distinction between the statistical and thermodynamic entropies .
accordingly , we will now use the regular symbol @xmath38 for @xmath89 throughout the work , unless clarity is needed .
we will refer to @xmath58 in terms of microstate number @xmath111 in eq .
( [ s_boltzmann0])as the _ time - dependent boltzmann formulation _ of the entropy or simply the boltzmann entropy,@xcite whereas @xmath50 in eq .
( [ s_boltzmann ] ) represents the equilibrium ( boltzmann ) entropy .
it is evident that the gibbs formulation in eqs .
( [ gibbs_formulation ] ) and ( [ conventional_entropy0 ] ) supersedes the boltzmann formulation in eqs.([s_boltzmann0 ] ) and ( [ s_boltzmann ] ) , respectively , as the former contains the latter as a special limit .
however , it should be also noted that there are competing views on which entropy is more general.@xcite we believe that the above derivation , being general , makes the gibbs formulation more fundamental .
the continuity of @xmath112 follows directly from the continuity of @xmath113 .
the existence of the statistical entropy @xmath112 follows from the observation that it is bounded above by @xmath114 and bounded below by @xmath11 , see eq .
( [ s_boltzmann0 ] ) .
it should be stressed that @xmath115 is not the number of microstates of the @xmath116 replicas ; the latter is given by @xmath117^{\mathcal{n}}$ ] .
thus , the entropy in eq .
( [ ensemble_entropy_formulation ] ) should not be confused with the boltzmann entropy , which would be given by @xmath118 .
it should be mentioned at this point that boltzmann uses the combinatorial argument to obtain the entropy of a gas , see eq .
( [ boltzmann_s_1 ] ) , resulting in an expression similar to that of the gibbs entropy in eq .
( [ gibbs_s ] ) except that the probabilities appearing in his formulation represents the probability of various discrete states of a particle , and should not be confused with the microstate probabilities used here ; see sec .
[ marker_jaynes ] .
the approach of boltzmann is _ limited _ to that of an ideal gas only and is not general as it neglects the correlations present due to the interactions between particles.@xcite on the other hand , our approach is valid for any system with any arbitrary interactions between particles as all microstates in the collection are _
independent_. using the above formulation of @xmath119 , we have determined@xcite the statistical formulation of the entropy for a system @xmath12 , which is a small but macroscopically large part of @xmath10 ; see fig .
[ fig.sys ] .
it is assumed that the system and the medium are _ quasi - independent _ so that @xmath1 can be expressed as a sum of entropies @xmath120 and @xmath121 of the system and the medium , respectively : @xmath122 the two statistical entropies are given by an identical formulation @xmath123 respectively . here
, @xmath124 with probability @xmath125 denotes a microstate of @xmath12 and @xmath126 with probability @xmath127 that of the medium . in the derivation,@xcite
we have neither assumed the medium nor the system to be in internal equilibrium ; only quasi - independence is assumed .
the above formulation of statistical entropies will not remain valid if the two are not quasi - independent .
the same will also be true of the thermodynamic entropies .
for the system to be in internal equilibrium , its statistical entropy @xmath120 must be maximized under the constraints imposed by the medium .
the constraints are on the average values of the state variables : @xmath128 where @xmath129 is the value of @xmath130 in @xmath124 .
the condition for internal equilibrium is obtained by varying @xmath125 without changing the microstates , i.e. @xmath129 . using the lagrange multiplier technique , it is easy to see that the condition for this in terms of the lagrange multipliers whose definitions are obvious is @xmath131 the lagrange multipliers are the same for all microstates and the scalar product is over the elements in the set @xmath129
. it now follows that@xmath132 using the same scalar product as above .
it is now easy to identify the lagrange multipliers by observing that @xmath133 comparing this relation with the gibbs fundamental relation for the system , which follows from eq .
( [ gibbs_fundamental_extended ] ) when applied to the system , we find@xmath134 accepting this identification now allows us to conclude that the statistical entropy @xmath120 in eq .
( [ entropies ] ) is no different than the nonequilibrium thermodynamic entropy of the same system in internal equilibrium but in a medium . a special case of such a system is the ( equilibrium ) canonical ensemble of gibbs .
this proves ( 2 ) mentioned in sect .
[ marker_probabilities ] . in equilibrium ,
the lagrange multipliers associated with the internal variables vanish and eq . ( [ general entropy differential ] )
reduce to@xmath135 the significance of @xmath136 is quite obvious . in internal equilibrium , it is given by @xmath137 moreover , as the nonequilibrium entropy in internal equilibrium is a state function , it can in principle be measured or calculated by integrating eq .
( [ general entropy differential ] ) . therefore
, its value can be compared with the statistical entropy .
the above identification in eq .
( [ general entropy differential ] ) then proves ( 3 ) .
if the thermodynamic entropy is not a state function , it can not be measured or computed .
thus , while the statistical entropy can be computed in principle in all cases , as shown below explicitly , there is no way to compare its value with the thermodynamic entropy in all cases .
thus , no comment can be made about their relationship in general .
we merely conjecture that as the two entropies are the same when the thermodynamic entropy is a state function , it is no different from its statistical analog even when it is not a state function . the consideration of dynamics resulting in the simple connectivity of the sample ( or phase ) space has played a pivotal role in developing the kinetic theory of gases,@xcite where the interest is at high temperatures.@xcite as dynamics is very fast here , it is well known that the ensemble entropy agrees with its temporal formulation . however , at low temperatures , where dynamics becomes sluggish as in a glass,@xcite the system can be _ confined _ into disjoint components . sample ( or phase ) space confinement at a phase transition such as a liquid - gas transition is well known in equilibrium statistical mechanics.@xcite it also occurs when the system undergoes symmetry breaking such as during magnetic transitions , crystallizations , etc .
but confinement can also occur under nonequilibrium conditions , when the observational time scale @xmath138 becomes shorter than the equilibration time @xmath139,@xcite such as for glasses , whose behavior and properties have been extensively studied .
the issue has been recently considered by us,@xcite where only energy as an observable was considered .
the discussion is easily extended to the present case when confinement occurs for whatever reasons into one of the thermodynamically significant number of disjoint components @xmath140 , each component corresponding to the same set @xmath57 or @xmath130 ( we suppress the dependence for simplicity ) , depending on whether the body is isolated or not . such a situation arises , for example , in ising magnets at the ferromagnetic transition . ,
where the system is either confined to @xmath141 with positive magnetization or @xmath142 with negative magnetization .
even a weak external magnetic field @xmath143 , that we can _ control _ as an observer , will allow the system to make a choice between the two parts of @xmath144 .
it just happens that in this case @xmath145 and is thermodynamically insignificant .
the situation with glasses or other amorphous materials is very different.@xcite in the first place , @xmath144 is a union of _ thermodynamically significant _
number @xmath146 disjoint components .
in the second place , there is no analog of a symmetry breaking field .
therefore , there is no way to prepare a sample in a given component @xmath147 .
thus , the samples will be found in all different components .
taking into consideration disjointness of the components generalizes the number of configurations in eq .
( [ combinations ] ) to@xmath148 where @xmath149 denotes the number of sample in the microstate @xmath150 in the @xmath151-th component . in terms of @xmath152 , this combination immediately leads to@xmath153 for the statistical entropy of the system and has already been used earlier@xcite by us ; see sec .
4.3.3 there . from what has been said above ,
this statistical entropy is also the thermodynamic entropy of a nonequilibrium state under component confinement for which the entropy is a state function of @xmath57 .
therefore , as before , we take @xmath89 to be the general expression of the nonequilibrium thermodynamic entropy and use @xmath38 in place of @xmath89 .
introducing@xmath154 it is easy to see@xcite that @xmath155 here , the entropy of the component @xmath147 in terms of the reduced microstate probability @xmath156 is@xmath157 so that the first contribution is its average over all components .
the second term is given by@xmath158 and represents the component entropy .
it is this entropy that is related to the residual entropy@xcite in disordered systems .
the same calculation for a system in a medium will result in an identical formulation for the entropy as in eq . ( [ s_component ] ) except that the sum is over components and microstates of the system .
we consider a gas of non - interacting identical structureless particles with no spin , each of mass @xmath159 , confined to a @xmath2-dimensional box of initial size @xmath160 with impenetrable walls ( infinite potential well ) .
initially , the gas is in thermodynamic _ equilibrium _ with a medium at fixed temperature @xmath161 and pressure @xmath162 .
the gas is then isolated by disconnecting it from the medium . in time , the isolated gas expands , may be in a nonequilibrium fashion .
we wish to calculate its entropy as a function of the box size @xmath163 . as there are no interactions between the particles , the wavefunction @xmath164 for the gas is a product of individual particle wavefunctions @xmath165 .
thus , we can focus on a single particle to study the nonequilibrium behavior of the gas.@xcite the simple model of a particle in a box has been extensively studied in the literature but with a very different emphasis .
@xcite the particle only has non - degenerate eigenstates whose energies are determined by @xmath166 , @xmath167 , and a quantum number @xmath168 .
we use the energy scale @xmath169 to measure the energy of the eigenstate so that@xmath170 the corresponding eigenfunctions are given by@xmath171 the pressure generated by the eigenstate on the walls is given by @xcite @xmath172 in terms of the eigenstate probability @xmath173 , the average energy and pressure are given by [ particle energy pressure]@xmath174 the entropy follows from eq .
( [ gibbs_formulation ] ) and is given for the single particle case by @xmath175 the time dependence in @xmath176 or @xmath177 is due to the time dependence in @xmath125 and @xmath178 .
even for an isolated system , for which @xmath179 remains constant , @xmath125 can not remain constant when the gas is not in equilibrium if @xmath180 is held fixed after expansion .
this follows directly from the second law@xcite and creates a conceptual problem because the eigenstates are mutually orthogonal and there can be no transitions among them to allow for a change in @xmath125 . a way to change @xmath125 in an isolated system is to require the presence of some stochastic interactions , whose presence allows for transitions among eigenstates.@xcite as these transitions are happening within the system , we can treat them as `` chemical reactions '' between different eigenstates@xcite by treating each eigenstate @xmath168 as a chemical species . during the transition , these species undergoes chemical reactions to allow for the changes in their probabilities .
we follow this analogy further and extend the traditional approach@xcite to the present case . for the sake of simplicity
, our discussion will be limited to the ideal gas in a box ; the extension to any general system is trivial .
therefore , we will use microstates instead of eigenstates in the following to keep the discussion general .
let there be @xmath181 particles in the @xmath168th microstate at some instant @xmath59 so that @xmath182 at all times , and @xmath183 .
we will consider the general case that also includes the case in which final microstates refer to a box size @xmath184 different from its initial value @xmath180 .
let us use @xmath185 to denote the reactants ( initial microstates ) and @xmath186 to denote the products ( final microstates ) .
for the sake of simplicity of argument , we will assume that transitions between microstates is described by a single chemical reaction , which is expressed in stoichiometry form as @xmath187 let @xmath188 and @xmath189 denote the population of @xmath185 and @xmath186 , respectively , so that @xmath190 .
accordingly , @xmath183 for the reactant and @xmath191 for the product . the single reaction is described by a single extent of reaction @xmath192 and we have @xmath193 it is easy to see that the coefficients satisfy an important relation@xmath194 which reflects the fact that the change @xmath195 in the reactant microstates is the same as in the product microstates .
the _ affinity _ in terms of the chemical potentials is given by@xmath196 and will vanish only in `` equilibrium , '' i.e. only when @xmath125 s attain their equilibrium values .
otherwise , @xmath197 will remain non - zero .
it acts as the thermodynamic force in driving the chemical reaction.@xcite but we must wait long enough for the reaction to come to completion , which happens when @xmath197 and @xmath198 both vanish .
the extent of reaction @xmath192 is an example of an internal variable .
for the ideal gas under consideration , there does not seem to be any other internal variable as particles have no internal structures . in the following
, we will assume only one internal variable @xmath199 .
[ ptb ] entropy_comparison.eps the box expands as a function of time , which need not be quasi - static ( extremely slow ) so there is no reason to assume that the gas remains in equilibrium after expansion .
the entropy of the gas per particle can be obtained by calculating @xmath200 for the particle under consideration .
henceforth , we will call @xmath201 the entropy of the particle , which shares the property that the irreversible entropy change @xmath202 will never be negative .
all the above discussion about the chemical reaction is easily translated to the study of a particle in box without any change .
the change @xmath203 is caused by the transitions between different eigenstates .
we consider the gas in _ equilibrium _ at some initial temperature @xmath161 in a box of length @xmath204 , which we take to be @xmath205 .
this is obtained by keeping the box in a medium of temperature @xmath161 .
the corresponding microstate probabilities follow the boltzmann law ( @xmath206):@xmath207 where @xmath208 denotes the equilibrium partition function .
the energy per particle in this gas is denoted by @xmath209 obtained by replacing @xmath210 by @xmath211 in eq .
( [ particle energy pressure ] ) ; the corresponding pressure is @xmath212 .
the equilibrium entropy can be obtained by using @xmath211 in @xmath213 given in eq .
( [ entropies ] ) .
the initial temperature @xmath161 for @xmath205 is taken to be @xmath214 so that the initial energy @xmath215 .
we now consider equilibrium states having the same initial thermodynamic energy @xmath209 for different values of @xmath180 , even though the eigenstate energies and their boltzmann probabilities vary with @xmath180 .
the corresponding equilibrium entropy @xmath216 as a function of @xmath180 is shown in fig .
[ fig.entropy ] by the continuous curve .
since the energy is constant , the product @xmath217 is also constant ; see eq .
( [ particle energy pressure ] ) .
thus , @xmath218 is a decreasing function of @xmath180 . from the slope of the upper curve in fig .
[ fig.entropy ] which decreases with @xmath180 , we also conclude that @xmath219 is a decreasing function of @xmath180 .
thus , @xmath220 is an increasing function of @xmath180 .
however , our calculation to be presented elsewhere @xcite shows that @xmath221 is also an increasing function of @xmath180 .
the eigenstates for a box of size @xmath180 are given in eq .
( [ eigenfunctions ] ) .
we now consider nonequilibrium states .
for this , we isolate the box from its medium and consider its free expansion as it expands suddenly from @xmath204 to a new size @xmath222 .
because of its isolation , its energy remains @xmath209 during this expansion .
as the expansion is sudden , the initial eigenfunctions @xmath223 for @xmath204 have no time to change , but are no longer the eigenfunctions of the new size @xmath180 ; the latter are given by @xmath224 in eq .
( [ eigenfunctions ] ) for @xmath180 .
however , @xmath223 can be expanded in terms of @xmath224 as a sum over @xmath168 .
we call this the quantum superposition principle .
the corresponding expansion coefficients @xmath225 are easily seen to be@xcite@xmath226 using @xmath225 and @xmath227 , we can determine the probability @xmath228 for the @xmath168th microstate in the new box , which allows us to determine all thermodynamic averages for the new box .
we have checked that the new probabilities add to @xmath2 and that the ( average ) energy after the free expansion is equal to @xmath209 to within our computational accuracy .
thus , @xmath229 in the sudden expansion .
this is consistent with the fact that the gas does no external work and that no external heat is exchanged .
despite this , the free expansion is spontaneous once the confining walls have moved .
therefore , the ( thermodynamic ) entropy of the gas must increase in this process in accordance with the second law .
we use @xmath228 to evaluate the nonequilibrium statistical entropy , which is shown by the dashed curve in fig .
[ fig.entropy ] .
the significance of this curve is as follows : choose a particular value @xmath180 in this graph .
then , the nonequilibrium entropy for this @xmath180 is given by numerically evaluating the sum @xmath230 this is the entropy after the sudden expansion from the initial state at @xmath205 and follows from the quantum superposition principle .
evidently , this entropy is higher that the initial equilibrium entropy @xmath231 .
it is also obvious that this entropy has a memory of the initial state at @xmath205 and @xmath214 .
therefore , it does not represent the equilibrium entropy .
if we now wait at the new value of @xmath180 , the isolated gas in the new box will relax to approach its equilibrium state in which its nonequilibrium entropy will gradually increase until it becomes equal to its value on the upper curve .
a careful reader would have realized by this time that the proposed entropy form in eq .
( [ gibbs_formulation ] ) is not at all the same as the standard classical formulation of entropy , such as for the ideal gas , which can be negative at low temperatures or at high pressures .
the issue has been discussed elsewhere@xcite but with a very different perspective . here , we visit the same issue from a very different perspective that allows us to investigate if and how the entropy in continuum models is related to the proposed entropy in this work . for this , we turn to a very simple continuum model in classical statistical mechanics : the tonks gas,@xcite which is an athermal model and contains the ideal gas as a limiting case when the rod length @xmath232 vanishes .
we will simplify the discussion by considering the tonks gas in one dimension .
the gas consists of @xmath233 impenetrable rods , each of length @xmath232 lying along a line of length @xmath180 .
we will assume @xmath233 to be fixed , but allow @xmath232 and @xmath180 to change with the state of the system , such as its pressure .. the configurational entropy per rod determined by the configurational partition function is found to be@xcite@xmath234 , \label{tonks_s}\ ] ] where @xmath235 is the `` volume '' available per rod @xmath236 .
even though the above result is derived for an equilibrium tonks gas , it is easy to see that the same result also applies for the gas in internal equilibrium .
the only difference is that the parameters in the model are also functions of internal variables now .
the entropy vanishes when @xmath237 and becomes negative for all @xmath238 .
indeed , it diverges to @xmath239 in the incompressible limit @xmath240 .
this is contrary to the boltzmann approach in which the entropy is determined by the number of microstates ( cf .
( [ boltzmann_s ] ) ) or the gibbs approach ( cf .
( [ gibbs_formulation ] ) ) and can never be negative .
can we reconcile the contradiction between the continuum entropy and the current statistical formulation ?
we now demonstrate that the above entropy for the tonks gas is derivable from the current statistical approach under some approximation , to be noted below , by first considering a lattice model for the tonks gas and then taking its continuum limit .
it is in the lattice model can we determine the number of microstates . in a continuum , this number is always _ unbounded _ ( see below also ) . for this
we consider a @xmath2-d lattice @xmath241 with @xmath242 sites ; the lattice spacing , the distance between two consecutive sites , is given by @xmath6 .
we take @xmath243 so that @xmath244 is the length of the the lattice @xmath241 .
we randomly select @xmath233 sites out of @xmath242 .
the number of ways , which then represents the number of configurational microstates , is given by@xmath245 after the choice is made , we replace each selected site by @xmath246 consecutive sites , each site representing an atoms in a rod , to give rise to a rod of length @xmath247 .
it is clear that @xmath6 also changes with the state of the system .
the number of sites in the resulting lattice @xmath248 is @xmath249 so that the length of @xmath250 is given by @xmath251 since @xmath252 .
we introduce the number densities @xmath253 , @xmath254 and @xmath255 .
a simple calculation shows that @xmath256 is given by@xmath257.\ ] ] this result can also be obtained by taking the athermal entropy for a polydisperse polymer solution a bethe lattice@xcite by setting the coordination number @xmath258 to be @xmath259 .
we now take the continuum limit @xmath260 for _ fixed _ @xmath261 and @xmath33 , that is _ fixed _
@xmath262 and @xmath263 , respectively . in this limit , @xmath264 , and @xmath265 .
use of these limits in @xmath213 yields@xmath266 the continuum limit of the entropy from the boltzmann approach has resulted in a diverging entropy regardless of the value of @xmath261,@xcite a well known result . by introducing an arbitrary _ constant _ @xmath267 with the dimension of length ,
we can rewrite @xmath213 as@xmath268 in which the first term remains finite in the continuum limit , and the second term contains the divergence .
the diverging part , although explicitly independent of @xmath261 , still depends on the state of the gas through @xmath6 , and can not be treated as a constant unless we assume @xmath6 to be independent of the state of the gas .
it is a common practice to approximate the lattice spacing @xmath6 as a constant . in that case , the diverging term represents a constant that can be subtracted from @xmath269 . recognizing that @xmath270 , we see that the first term in eq .
( [ cont_manipulation ] ) is nothing but the entropy of the tonks gas in eq .
( [ tonks_s ] ) for the arbitrary constant @xmath271 .
however , this equivalence only occurs in the state independent constant-@xmath6 approximation .
as the second term above has been discarded , the continuum entropy @xmath272 also has _ no _ simple relationship with the number ( @xmath273 ) of microstates in the continuum limit , which means that the continuum entropy can not be identified as the boltzmann entropy in eq .
( [ s_boltzmann ] ) . to see this more clearly ,
let us focus on the centers of mass of each rod , which represent one of the @xmath233 sites that were selected in @xmath241 .
each of the @xmath168 sites @xmath274 , @xmath275 , is free to move over @xmath262 .
volume @xmath276 , also called the probability and denoted by @xmath277 by boltzmann,@xcite of the corresponding phase space @xmath278 is @xmath279 . however , contrary to the conventional wisdom,@xcite @xmath280 does not yield @xmath272 .
the correct expression is given by the gibbs - modified adimensional volume @xmath281 , i.e. @xmath282 the presence of @xmath283 is required to restrict the volume due to indistinguishability of the rods la gibbs . for large @xmath233 , this quantity correctly gives the entropy @xmath272 .
however , this quantity is not only not an integer , it also can not be _ always _ larger than or equal to unity , as noted above .
the analog of a quantum microstate in classical statistical mechanics is normally obtained by recognizing that in the _ adiabatic approximation _ , each small phase space cell of volume element @xmath284 in terms of generalized coordinates @xmath258 and momenta @xmath285 of size
@xmath286 corresponds to a microstate,@xcite where @xmath287 is the degrees of freedom of the system .
the latter follows from the bohr - sommerfeld quantization rule for a periodic motion .
the adiabatic approximation requires the parameters characterizing the system to vary extremely slowly .
we will assume one such parameter @xmath151 ( such as the volume @xmath288 ) so that @xcite @xmath289 where @xmath290 is the period of oscillation for constant @xmath151 ; the system would be isolated in the latter case . in the above approximation , the energy of the system will vary very slowly and can be taken to be constant over a period of oscillation .
the action taken over the closed path for the constant value of @xmath151 and the energy is quantized @xcite:@xmath291 this observation is the justification of the above cell size of a classical microstate .
thus , the number of `` classical '' microstates is given by@xmath292 where @xmath293 is the phase space volume corresponding to the system .
this allows us to divide the phase space into @xmath294 cells , index by @xmath168 , of volume @xmath295 and `` centered '' at @xmath274 at some time which we call the initial time @xmath296 .
we will denote the evolution of the cell at time @xmath59 by the location of its `` center '' @xmath297 and its volume element @xmath298 . in terms of the distribution function @xmath35 in the phase space , the @xmath168th microstate probability therefore is given by@xmath299 evidently , @xmath300 at all times .
the entropy and the average of any observable @xmath301 of the system are given by @xmath302,\label{discrete_s}\\ \overline{o}(t ) & \equiv{\textstyle\sum\nolimits_{k } } o(x_{k},t)f(x_{k},t)dx_{k}(t ) , \label{discrete_o}\ ] ] the sum being over all @xmath294 microstates . while it is easy to see that continuum analogs for eqs .
( [ f - normalization ] ) and ( [ discrete_o ] ) are easily identified , this is not so for the entropy in eq .
( [ discrete_s]).@xcite however , it should be obvious that @xmath303 in eq .
( [ sf_s ] ) can not be a candidate for the statistical entropy @xmath304 it is well known @xcite that the system in the adiabatic limit remains in the same quantum state . for hamiltonian dynamics , the conservation of the phase space cell volume under evolution ensures @xmath295 for each cell so that @xmath305 .
this results in the incompressibility of the phase space . in equilibrium , @xmath306 is not only uniform but also constant , and we conclude from eq .
( [ f - normalization ] ) that this value is @xmath307 .
accordingly , @xmath308 in equilibrium as expected and we obtain the equilibrium entropy @xmath309 .
the situation is far from clear for nonequilibrium states . as the example of expansion of the box shows ,
the system is no longer restricted to be in the same microstate , which means that the microstate energy is no longer a constant and the phase space trajectory is no longer closed .
thus , the suitability of the bohr - sommerfeld quantization is questionable , and care must be exercised to identify the microstates .
we will adopt the following prescription .
we consider some equilibrium state ( uniform @xmath310 ) of the isolated system to identify the cell volume @xmath311 @xmath286 .
once the identification has been done , we will no longer worry about its relationship with @xmath286 , and only deal with the cells .
we then follow the evolution of each cell in time in a nonequilibrium process during which @xmath305 may not hold .
thus the volume of each cell may no longer be constant .
the process may also result in changes in @xmath306.@xcite indeed , it is quite possible that @xmath312 diverges at the same time that @xmath313 vanishes.@xcite however , their product , which determines the microstate probability , must remain strictly bounded and @xmath314 .
in particular , as the cell volume shrinks to zero , @xmath312 must diverge to keep the product bounded .
thus , the divergence@xcite of @xmath312 alone does not imply that the entropy diverges to negative infinity .
this is easily seen by the following 1-d damped oscillator , the standard prototype of a dissipative system:@xcite @xmath315 with a ( positive ) damping coefficient , which is chosen such that @xmath316 just for simplicity .
we have the case of aperiodic damping .
we will only consider the long time behavior .
it is easy to see that in this limit ( @xmath317 ) @xmath318 and their product remains bounded , as expected .
boltzmann@xcite provides the following alternative expression of the entropy@xcite in terms of a single particle probability @xmath319 for the particle to be in the @xmath320th state:@xmath321 not to be confused with that in eq .
( [ boltzmann_s ] ) .
boltzmann is only interested in the maximum entropy , which occurs when all states are equally probable . in this case , @xmath322 where @xmath323 is the number of possible states of a single particle in the gas . in general ,
particles are not independent due to interactions and number of possible states @xmath324 .
accordingly , maximum gibbs entropy @xmath325 per particle is _ less _ than the corresponding equiprobable boltzmann entropy @xmath326 .
however , jaynes@xcite gives a much stronger results , see his eq .
( 5):@xmath327 the equality occurs only if there are no interactions between the particles , as we have asserted above .
recognizing that there does not exists a first principles statistical formulation of nonequilibrium thermodynamic entropy for an isolated system in terms of microstate probabilities , we have attempted to fill in the gap .
we use a formal approach ( frequentist interpretation of probability ) by extending the equilibrium ensemble of gibbs to a nonequilibrium ensemble , which is nothing but a large number @xmath86 of samples of the thermodynamic system under consideration .
accordingly , we refer to the ensemble as a sample space .
the formal approach enables us to evaluate the combinatorics for a given set of microstate probabilities .
the resulting statistical entropy is independent of the number of samples and depends only on the probabilities as is seen from eqs .
( [ gibbs_formulation ] ) and ( [ s_component ] ) .
thus , the use of a large number of samples is merely a formality and is not required in practice .
we have shown that in equilibrium , the statistical entropy is the same as the equilibrium thermodynamic entropy : @xmath328 .
but we have also shown that the statistical entropy is equal to the nonequilibrium thermodynamic entropy , provided the latter is a state function of the nonequilibrium state variables @xmath60 : @xmath329 .
we can not make any comment about the relationship between @xmath93 and @xmath119 for the simple reason that there is no way to measure or calculate a non - state function @xmath119 .
we should remark here that the standard approach to calculate nonequilibrium entropy is to use the classical nonequilibrium thermodynamics@xcite or its variant , which treats the entropy at the local level as a state function .
some readers may think that our statistical formulation is no different than that used in the information theory . we disagree .
for one , there is no concept of internal variables @xmath56 in the latter theory .
because of this , our approach allows us to consider three levels of description so that we can consider three different entropies @xmath330 and @xmath50 satisfying the inequalities in eq .
( [ entropy_inequalities0 ] ) .
the information theory can only deal with two levels of entropies .
there is also no possibility of a residual entropy in the latter . for an isolated system in internal equilibrium ( @xmath331 for @xmath332 )
, just a single sample will suffice to determine the entropy as samples are _ unbiased_. the entropy in this case is no different than the `` entropy '' @xmath333 of a single sample:@xcite@xmath334 where @xmath335 represents @xmath111 or @xmath103 .
however , this simplicity is lost as soon as the system is not in internal equilibrium . here
, one must consider averaging over all microstates .
changes in microstate probabilities result in changes in the entropy .
there are two ways probabilities can change within an isolated system , both of them being irreversible in nature .
one cause of changes is due to the quantum nature as seen in the sudden expansion of the box . here
, the parameter @xmath151 ( @xmath336 ) changes non - adiabatically and creates irreversibility .
the resulting irreversible change in the entropy for the @xmath2-@xmath337 gas has been calculated and shown by the lower curve in fig .
[ fig.entropy ] .
the other cause of probability changes is due to the `` chemical reaction '' going on among the microstates that brings about equilibration in the system .
the corresponding irreversible rise in the entropy for the gas is shown by the difference between the two curves in fig .
[ fig.entropy ] .
the interaction of a body with its medium can also result in the changes in microstate probabilities , and has been considered elsewhere.@xcite we consider the continuum analog of the statistical formulation of entropy and show that the standard formulation , @xmath303 in eq . ( [ sf_s ] ) , is not a good candidate of the nonequilibrium entropy .
it is then argued that the divergence of @xmath312 in some cases , see the discussion above , makes @xmath303 diverge to @xmath239 , even though the statistical entropy remains finite and positive .
thus , @xmath303 can not be equated with our statistical formulation , a generalization of the gibbs formulation .
we suggest that our statistical gibbs formulation can be applied to any nonequilibrium state . the number of combinations in eq .
( 35 ) on p. 56 in boltzmann@xcite is denoted by @xmath277 , but it is not the number of microstates .
the two become the same only when @xmath277 is maximized as discussed on p. 58 . | we consider an isolated system in an arbitrary state and provide a general formulation using first principles for an _ additive _ and _ non - negative _ statistical quantity @xmath0 that is shown to reproduce the equilibrium thermodynamic entropy of the isolated system .
we further show that @xmath0 represents the nonequilibrium thermodynamic entropy @xmath1 when the latter is a state function of nonequilibrium state variables ; see text .
we consider an isolated @xmath2-d ideal gas and determine its non - equilibrium statistical entropy @xmath0 as a function of the box size as the gas expands freely isoenergetically , and compare it with the equilibrium thermodynamic entropy @xmath3 .
we find that @xmath4 @xmath3 in accordance with the second law , as expected . to understand how @xmath0 is different from thermodynamic entropy of classical continuum models that is known to become _ negative _ under certain conditions , we calculate @xmath5for a @xmath2-d lattice model and discover that it can be related to the thermodynamic entropy of the continuum @xmath2-d tonks gas by taking the lattice spacing @xmath6 go to zero .
however , @xmath7 @xmath8 since @xmath6 is state - dependent .
we discuss the semi - classical approximation of our entropy and show that the standard quantity @xmath9 in the boltzmann s h - theorem , see eq .
( [ sf_s ] ) , does not directly correspond to the statistical entropy . |
lattice spin systems are idealized mathematical models of magnetic materials . in the absence of external disturbances such systems are in equilibrium , and governed by the gibbs - boltzmann distribution @xmath4 , where @xmath5 is the energy ( or hamiltonian ) of a micro - state @xmath6 , @xmath7 is the inverse rescaled temperature , and @xmath8 is the partition function . if the set @xmath9 of individual spin states is continuous , the sum in @xmath10 becomes an integral .
the averages of macroscopic functions of micro - states ( observables ) , such as the total energy or the magnetization , can be obtained by differentiation from the free energy @xmath11 , which is related to the internal energy @xmath12 and the gibbs - shannon entropy @xmath13 via the thermodynamic relation @xmath14 .
however , computing @xmath15 analytically for an interacting system of macroscopic size is difficult , and to date only few lattice spin models models have been solved exactly @xcite . to circumvent the above problem one often approximates the true micro - state distribution @xmath16 with a simpler alternative @xmath17 , which retains only some characteristics of the original model .
this approximation can usually be interpreted as a deformation of the true topology of the lattice such that short loops are removed and the analytical computations of thermal averages become easier , in combination with a variational approach that utilises the inequality @xmath18\geq 0 $ ] .
the variational mean - field ( v - mf ) approximation , see e.g. @xcite , uses the probability distribution @xmath17 of a non - interacting system in this inequality . for @xmath19
its results are equal to the results of exactly solvable ( ferromagnetic ) spin systems on complete graphs @xcite , but for @xmath0-dimensional lattices they are unreliable .
it predicts incorrectly , for instance , a phase transition in the @xmath20-dimensional ising model .
however , its predictions for critical exponents are correct for ferromagnetic spin systems when @xmath21 ; see @xcite for a unified proof of this result and references to relevant earlier work . in the bethe - peierls ( bp )
approximation ( also known as belief propagation in computer science @xcite , or the cavity method in the spin glass community @xcite ) one replaces the original lattices by tree - like graphs , which enables a recursive computation of thermal averages .
one such graph is the bethe lattice @xcite , usually defined as the ` central ' part of an infinitely large cayley tree @xcite .
closely related to the bethe lattice is the random regular graph ( rrg ) , defined as a maximally random graph in which all vertices have the same degree @xcite .
rrgs do have loops , but these are typically of length @xmath22 as @xmath19 , so rrgs are locally tree - like
. for ferromagnetic models , bethe lattices and rrgs give the same results @xcite , but in antiferromagnetic and spin - glass models the loops in the rrgs generate frustration , and can not be ignored @xcite .
the bp approximation is more reliable than the mf approximation @xcite , since it involves a less brutal deformation of the original lattice , and it is exact for ferromagnetic ising models on locally tree - like random graphs @xcite ; it is interesting that , despite the fact that they can be solved relatively easily , the behaviour of ising models on trees is more complex than in @xmath0-dimensional lattices @xcite .
further improvements of the bp approximation were obtained by correcting the bp solution for rare loops @xcite ; the improved theory is exact for a bethe lattice with exactly one loop @xcite . in this paper
we study spin models in which not only the correct coordination numbers of @xmath0-dimensional cubic lattices are retained , but ( unlike the v - mf , bp , and rrg approximations ) also the statistics of short loops and many of their nestings .
the spins in our models occupy the vertices of bethe - type lattices constructed from @xmath0-dimensional hypercubes , i.e. from the cells ( squares , cubes , etc . ) of the conventional @xmath0-dimensional cubic lattice .
these hypercubic bethe lattices can be seen as generalisations of husimi lattices @xcite , which are bethe lattices constructed from loops such that no edge lies in more than one loop .
we use the bethe - peierls method to derive equations for the average magnetization , the specific heat , and the internal energy per spin . from these
we compute phase diagrams for the ferromagnetic ising version of the model with @xmath23 .
our phase diagrams are compared with onsager s exact result for the @xmath24 square lattice , and with high- and low - temperature expansions and monte carlo ( mc ) simulation results for cubic lattices with @xmath25 .
we know that the interaction topology of any spin system of size @xmath26 can be represented by a a bipartite factor - graph @xmath27 @xcite , with @xmath28 variable - nodes and @xmath29 factor - nodes , in which the micro - state energy takes the general form @xmath30 here we denote with @xmath31 the microscopic spin state of the system , where @xmath32 for all @xmath33 . for ising systems we would have @xmath34 .
a factor - tree is a special type of factor - graph in which there are no loops , see figure [ figure : factor - tree ] .
( 100,80 ) ( 0,0 ) .
all spins are represented by circular ` variable ' nodes , and each term in the energy ( [ def : e ] ) corresponds to a square ` factor ' node .
a link between variable node @xmath35 and factor node @xmath36 implies that @xmath37 acts as an argument of @xmath38.,title="fig : " ] ( 36,64)@xmath39 ( 34,47)@xmath40 ( 35,27)@xmath36 ( 33,10)@xmath37 the energy @xmath38 of each factor - node @xmath41 in a factor - graph is a specific function of the states of a subset @xmath42 of the spins . we denote similarly with @xmath43 the set of all factor - nodes connected to variable - node @xmath33 , i.e. all energy terms in ( [ def : e ] ) that depend explicitly on spin @xmath33 .
further examples of such systems are given in figure [ figure : squarelattice ] .
from now on we will consider factor - trees only , as in figure [ figure : factor - tree ] .
we denote with with @xmath39 the factor - node at the root of the tree . the full partition function @xmath10 of the spin system ( [ def : e ] ) on our factor tree
can be written as @xmath44.\end{aligned}\ ] ] here @xmath45 $ ] denotes the partition function of the sub - tree descending from node @xmath46 , where @xmath46 descends from the root factor node @xmath39 , @xmath47&= & \prod_{\nu\in\partial_j\setminus \mu}\big ( \sum_{s_\ell,\ell\in \partial_\nu\setminus j } \rme^{- \beta e_\nu(s)}\prod_{\ell\in\partial_\nu\setminus j}z^{(1)}_{\nu \ell}[s_\ell]\big).\end{aligned}\ ] ] continuation of this argument gives similar expressions for any sub - tree function @xmath48 $ ] at distance @xmath49 from the root in terms of the sub - tree - functions at distance @xmath50 , using the tree - like nature of the graph : @xmath51&=&\prod_{\lambda\in\partial_j\setminus \nu}\big ( \sum_{\{s_\ell , \ell\in\partial_\lambda \setminus j\}}\!\!\rme^{-\beta e_\lambda(s)}\prod_{\ell\in \partial_\lambda\setminus j } z^{(r+1)}_{\lambda \ell}[s_\ell ] \big ) .
\label{eq : zr}\end{aligned}\ ] ] we can also calculate marginal spin probability distributions , starting from the top of the tree , and find @xmath52}{\sum_{\{\tilde s_j , j\in\partial_\mu\}}\rme^{-\beta e_\mu(\tilde s ) } \prod_{j\in \partial\mu}p^{(0)}_{\mu j}[\tilde s_j]},\label{def : p } \ ] ] where we have defined the probability distribution @xmath53&=&\frac{z^{(0)}_{\mu j}[s_j]}{\sum_{\tilde s_j}z^{(0)}_{\mu j}[\tilde s_j]}. \label{def : p0}\end{aligned}\ ] ] @xmath54 $ ] is the marginal distribution of spin @xmath46 in the cavity graph , that is obtained upon removing the link from @xmath46 to factor node @xmath39 .
it follows directly from ( [ eq : zr ] ) that also this distribution can be computed recursively via @xmath55&= & \frac{\prod_{\lambda\in \partial j\setminus\nu}\big ( \sum_{\{s_\ell,\ell\in\partial_\lambda\setminus j\ } } \rme^{-\beta e_\lambda(s)}\prod_{\ell\in \partial\lambda\setminus j } p^{(r+1)}_{\lambda \ell}[s_\ell]\big ) } { \sum_{\tilde{s}_j}\prod_{\lambda\in \partial j\setminus\nu}\big ( \sum_{\{\tilde{s}_\ell,\ell\in\partial_\lambda\setminus j\ } } \rme^{-\beta e_\lambda(\tilde{s})}\prod_{\ell\in \partial\lambda\setminus j } p^{(r+1)}_{\lambda \ell}[\tilde{s}_\ell]\big ) } , \label{eq : pr}\end{aligned}\ ] ] which is to be solved with boundary conditions @xmath56 \}$ ] . from ( [ def : p ] ) one obtains the ensemble averages of energies and magnetisations in layer @xmath57 , i.e. for @xmath58 , via @xmath59 if in a homogeneous system such as a regular ferromagnet there exists an @xmath60 such that @xmath61=p^c[s]$ ] for all @xmath62 , i.e. if the iteration ( [ eq : pr ] ) for the spin distribution on the cavity graph has converged to an invariant measure @xmath63 $ ] , this defines a region in the tree where all factor - nodes and all variable - nodes have become equivalent .
any factor - node @xmath39 and any variable node @xmath46 belonging to this region will contribute @xmath64}{\sum_{\{s_j , j\in\partial_\mu\}}\rme^{-\beta e_\mu(s ) } \prod_{j\in \partial\mu}p^c[s_j ] } \label{eq : e2 } \\
\langle s_i \rangle&= & \frac{\sum_{\ { s_j , j\in\partial\mu\ } } s_i \rme^{-\beta e_\mu(s ) } \prod_{j\in \partial\mu}p^c[s_j]}{\sum_{\ { s_j , j\in\partial_\mu\}}\rme^{-\beta e_\mu(s ) } \prod_{j\in \partial\mu}p^c[s_j ] } \label{eq : m2}\end{aligned}\ ] ] to the total energy and the total magnetization , respectively . in inhomogeneous systems this can not happen
; the above averages would involve solutions of ( [ eq : pr ] ) that would also depend on the realisation of the disorder , and one would have to turn to the population dynamics algorithm @xcite to average out this disorder . for ising spins , where @xmath65 we can write ( [ def : p0 ] ) in the following form , with parameters @xmath66 that in the spin - glass literature are known as cavity fields @xcite : @xmath67&=&\frac{\rme^{\beta s_jh_{\mu j}(r)}}{2\cosh(\beta h_{\mu j}(r))}. \label{def : h}\end{aligned}\ ] ] using the identity @xmath68 $ ] in the left - hand side of ( [ eq : pr ] ) gives us a simple recursive equation for the cavity fields : @xmath69 the thermodynamics of homogeneous ising systems of the type under consideration , where all spins are equivalent , would thus be governed by the solution of @xmath70 we can insert the solution of ( [ def : h ] ) into equations ( [ eq : e2],[eq : m2 ] ) , and use the fact that for homogeneous tree - like systems we could have chosen any factor node @xmath39 as our root , so that ( [ eq : e2 ] ) and ( [ eq : m2 ] ) must apply to all factor- and variable nodes .
hence we obtain for the energy density @xmath71 and the magnetisation per spin @xmath72 : @xmath73 as a simple example we can recover from the above equations the known results for the ferromagnetic ising model on a bethe lattice with connectivity @xmath74 ( see e.g. @xcite ) . here
@xmath75 , @xmath76 , and @xmath77 where @xmath78 .
the cavity field equation ( [ eq : h_homogeneous ] ) now reduces to @xmath79^{k-1 } \nonumber\\ & = & \frac{k\!-\!1}{\beta}\tanh^{-1}(\tanh(\beta h)\tanh(\beta j ) ) .
\label{eq : bethe - h}\end{aligned}\ ] ] for the energy density ( [ eq : e3 ] ) and the magnetisation per spin ( [ eq : m3 ] ) we find @xmath80}{1+\tanh^2(\beta h)\tanh(\beta j)}.\end{aligned}\ ] ] using ( [ eq : bethe - h ] ) the latter equation can be rewritten as @xmath81 for @xmath82 the system is always paramagnetic ( @xmath83 ) . if @xmath84 it is paramagnetic for @xmath85 but ferromagnetic ( @xmath86 ) for @xmath87 , where @xmath88 is the critical inverse temperature of the system @xcite .
( 130,56 ) ( 0,0 ) bethe lattice , and its corresponding factor graph representation ( on the right).[figure : squarelattice ] , title="fig : " ] ( 23,35)@xmath89 ( 33,35)@xmath90 ( 23,25)@xmath91 ( 33,25)@xmath92 ( 86,35)@xmath89 ( 101,35)@xmath90 ( 86,20)@xmath91 ( 101,20)@xmath92 we now turn to ferromagnetic ising models on hypercubic bethe lattices .
these lattices are constructed recursively from a single @xmath0-dimensional hypercube , by attaching exactly @xmath93 hypercubes to its corners , thereby producing the centre and the first shell of the lattice ( see figure [ figure : squarelattice ] ) .
the second shell is constructed by attaching @xmath94 hypercubes to the ` available ' corners in the first shell .
this process of attaching @xmath0-dimensional hypercubes to available corners is continued _
ad infinitum_. we note that each vertex in the hypercubic bethe lattice is shared between two adjacent hypercubes , ensuring that each vertex is connected to exactly @xmath95 neighbouring vertices , exactly as in conventional @xmath0-dimensional lattices . to minimise notation clutter we will from now on choose units such that @xmath96 , and transform @xmath97 .
let us first study the simplest case of @xmath24 . here
the hyper cubic bethe lattice is constructed from squares , see figure [ figure : squarelattice ] .
each square contributes @xmath98 to the total energy of the system .
using this in equation ( [ eq : h ] ) gives us @xmath99 + h(r+1)\sum_{j=1}^3s_j } } { \sum_{\{\tilde s_j\}}\rme^{\beta\left[-\tilde s_1 + \tilde s_1\tilde s_2 + \tilde s_2\tilde s_3 - \tilde s_3\right ] + h(r+1)\sum_{j=1}^3\tilde s_j } } \right)\label{eq:2d - h}\end{aligned}\ ] ] for the cavity field acting on a spin @xmath89 living in the @xmath49-th shell of our lattice , see figure [ figure : square ] .
( 100,50 ) ( 0,0 ) . left : computation of the cavity fields .
right : evaluation of the marginal probability @xmath100 , using the cavity fields .
[ figure : square ] , title="fig : " ] ( 10,35)@xmath89 ( 38,35)@xmath90 ( 10,8.5)@xmath91 ( 38,8.5)@xmath92 ( 0,0)@xmath101(42,0)@xmath101 ( 43,43)@xmath101 ( 66,35)@xmath89 ( 94,35)@xmath90 ( 66,8.5)@xmath91 ( 94,8.5)@xmath92 ( 55,0)@xmath101 ( 54,43)@xmath101 ( 97,0)@xmath101 ( 98,43)@xmath101 if we solve this equation from a distant boundary at @xmath102 , we obtain the following equation for the cavity fields in the bulk of the system , describing the fixed - point of the iterative map ( [ eq:2d - h ] ) : @xmath103 the cavity field @xmath101 acts on the spins living in the central part of the square bethe lattice , and can be used to compute the probability distribution of the four spins interacting on the square , see figure [ figure : square ] , being @xmath104 + h\sum_{j=0}^3s_j } } { \sum_{\tilde{s}_0 , \tilde{s}_1 , \tilde{s}_2 , \tilde{s}_3}\rme^{\beta\left[\tilde s_0\tilde s_1 + \tilde s_1\tilde s_2 + \tilde s_2\tilde s_3 + \tilde s_3\tilde s_0\right ] + h\sum_{j=0}^3\tilde s_j } } .
\label{eq : marginal}\end{aligned}\ ] ] since our spin system is homogeneous , this distribution can subsequently be used to compute the magnetisation per spin @xmath105 , the energy density @xmath106 and the specific heat @xmath107 ( see [ app : c ] for details ) .
the results are shown in figure [ figure:2d ] .
the point @xmath108 is always a solution of equation ( [ eq:2d - h - fp ] ) , and corresponds to the @xmath83 paramagnetic ( pm ) state .
this solution becomes unstable at the critical inverse temperature @xmath109 , where a new solution @xmath110 appears .
this new solution is stable for @xmath87 , and corresponds to the @xmath86 ferromagnetic ( fm ) state .
the phase transition at @xmath111 is of second order ( see figure [ figure:2d ] ) .
( 200,95 ) ( 0,0 ) , specific heat @xmath112 and magnetisation per spin @xmath113 as a function of the inverse temperature @xmath114 in @xmath24 lattices .
solid lines : square ( hypercubic ) bethe lattice . dashed lines : regular ( tree - like ) @xmath115 bethe lattice .
dotted lines : the exact result for the ising model in @xmath24 dimensions , with the critical inverse temperature @xmath116 @xcite ( vertical line in the left panel ) .
, title="fig : " ] ( 57,37 ) , specific heat @xmath112 and magnetisation per spin @xmath113 as a function of the inverse temperature @xmath114 in @xmath24 lattices .
solid lines : square ( hypercubic ) bethe lattice . dashed lines : regular ( tree - like ) @xmath115 bethe lattice .
dotted lines : the exact result for the ising model in @xmath24 dimensions , with the critical inverse temperature @xmath116 @xcite ( vertical line in the left panel ) .
, title="fig:"](53,71 ) ( 115,0 ) , specific heat @xmath112 and magnetisation per spin @xmath113 as a function of the inverse temperature @xmath114 in @xmath24 lattices .
solid lines : square ( hypercubic ) bethe lattice . dashed lines : regular ( tree - like ) @xmath115 bethe lattice .
dotted lines : the exact result for the ising model in @xmath24 dimensions , with the critical inverse temperature @xmath116 @xcite ( vertical line in the left panel ) .
, title="fig : " ] ( 55,-8)(170,-8 ) ( 110,55 ) ( -5,55 ) we compare these results with those of the ordinary @xmath115 bethe lattice ( [ eq : bethe - m],[eq : bethe - e ] ) , and with the exact results for the @xmath117-dimensional ferromagnetic ising model @xcite .
we then find that the transition point of our square bethe lattice , i.e. @xmath118 , is an improved estimate of the critical point @xmath119 @xcite of the 2d ising model , compared to the estimate @xmath120 of the ordinary bethe lattice .
we also observe this improvement over the ordinary bethe estimates in terms of the average magnetisation and the energy density ( see figure [ figure:2d ] ) , when comparing these quantities to the exact results for the 2d ising model which are given by @xmath121 here @xmath122 , and @xmath123 is the complete elliptic integral of the first kind @xcite .
finally , we note that the specific heat @xmath112 of the square bethe lattice has a jump at @xmath124 ( see figure [ figure:2d ] and [ app : c ] ) this is in contrast to the conventional square lattice where the specific heat is diverging at @xmath125 .
this behaviour of the specific heat is also found in @xmath25 bethe lattices ( see [ app : c ] for details ) studied in the next two sections .
( 200,95 ) ( 0,0 ) and magnetisation per spin @xmath113 as a function of the inverse temperature @xmath114 for @xmath126 lattices .
solid lines : cubic bethe lattice .
dashed lines : regular ( tree - like ) @xmath127 bethe lattice .
the results for the true ising model in @xmath126 dimensions are low temperature data for @xmath113 ( dotted line ) obtained by mc simulations in @xcite , with the critical inverse temperature estimate @xmath128 @xcite ( arrow and vertical line ) , predictions for @xmath129 obtained from high temperature @xcite and low temperature @xcite series ( both as dotted lines ) . the energy density @xmath130 ( horizontal line ) , assuming that @xmath131 , was computed by mc simulations in @xcite . , title="fig : " ] ( 115,0 ) and magnetisation per spin @xmath113 as a function of the inverse temperature @xmath114 for @xmath126 lattices .
solid lines : cubic bethe lattice . dashed lines : regular ( tree - like ) @xmath127 bethe lattice .
the results for the true ising model in @xmath126 dimensions are low temperature data for @xmath113 ( dotted line ) obtained by mc simulations in @xcite , with the critical inverse temperature estimate @xmath128 @xcite ( arrow and vertical line ) , predictions for @xmath129 obtained from high temperature @xcite and low temperature @xcite series ( both as dotted lines ) .
the energy density @xmath130 ( horizontal line ) , assuming that @xmath131 , was computed by mc simulations in @xcite . , title="fig : " ] ( 55,-8)(170,-8 ) ( 110,55 ) ( -5,55 ) in the ising model on the hypercubic @xmath126 bethe lattice , a cube , which one can regard as an ` upper ' square ( formed of variable nodes 0,1,2,3 ) that is is connected to a ` lower ' square ( formed of variable nodes 4,5,6,7 ) , contributes @xmath132 to the total energy . substituting this into ( [ eq : h ] ) gives us after some straightforward algebra the following equation for the cavity field @xmath101 : @xmath133 .
\label{eq:3d - cavity}\end{aligned}\ ] ] the paramagnetic solution @xmath108 of this equation becomes unstable at @xmath134 , and for @xmath135 equation ( [ eq:3d - cavity ] ) has two @xmath136 ferromagnetic solutions .
substitution of the the factor node energy function ( [ eq:3d - e1 ] ) into equations ( [ eq : m ] ) and ( [ eq : e ] ) gives us the corresponding magnetisation per spin and the energy density : @xmath137 e&=&-3\big ( 4 { \rme^{6 \beta+6 h}}+{\rme^{12 \beta-8 h}}+4
{ \rme^{6 \beta-6 h}}+4 { \rme^{4 \beta-4 { \it h}}}+4 { \rme^{2 \beta-2 h } } \nonumber\\ & & ~~+2 { \rme^{4 \beta}}-4 { \rme^{-2 \beta-2 h}}-4 { \rme^{-6 \beta- 2 h}}-10 { \rme^{-4 \beta}}+4 { \rme^{2 \beta+2 h } } \label{eq:3d - e2 } \\ & & ~~-4 { \rme^{-2 \beta+2 h}}+4 { \rme^{4 \beta+4 h}}-2 { \rme^{-12 \beta}}-4 { \rme^{-6 \beta + 2 h}}+{\rme^{12 \beta+8
h } } \big)/\mathcal{n } , \nonumber\end{aligned}\ ] ] where @xmath138 in figure [ figure:3d ] we compare the above values for @xmath113 and @xmath129 with those of the simple tree - like @xmath127 bethe lattices and with the predictions ( obtained via simulations and expansions ) for the true @xmath126 ising model .
again we find that the hypercubic bethe lattice is a more accurate proxy for the true ising system than the tree - like bethe lattice with the same coordination number . in figure [ figure:4d ]
we compare the values found for @xmath113 and @xmath129 with those of the simple tree - like @xmath145 bethe lattices and with the predictions ( obtained via simulations and expansions ) for the true @xmath139 ising model . as was the case with @xmath146
we find that , although the differences between the different model versions become smaller as @xmath0 increases , also in @xmath139 the hypercubic bethe lattice is a more accurate proxy for the true ising lattice than the tree - like bethe lattice with the same coordination number .
( 200,95 ) ( 0,0 ) and magnetisation per spin @xmath113 as a function of the inverse temperature @xmath114 for @xmath139 lattices .
solid lines : hypercubic ( tesseract ) bethe lattice . dashed lines : regular ( tree - like ) @xmath145 bethe lattice .
the results shown for the ising model in @xmath139 dimensions are obtained from low temperature series for @xmath129 @xcite ( dotted line ) , low temperature series for @xmath113 @xcite ( @xmath147 ) , and high temperature series for @xmath129 @xcite ( dotted line ) . the energy density at the critical point @xmath148 ( horizontal line ) and the critical inverse temperature @xmath149 were computed in mc simulations @xcite ( arrow and vertical line ) .
, title="fig : " ] ( 115,0 ) and magnetisation per spin @xmath113 as a function of the inverse temperature @xmath114 for @xmath139 lattices .
solid lines : hypercubic ( tesseract ) bethe lattice . dashed lines : regular ( tree - like ) @xmath145 bethe lattice .
the results shown for the ising model in @xmath139 dimensions are obtained from low temperature series for @xmath129 @xcite ( dotted line ) , low temperature series for @xmath113 @xcite ( @xmath147 ) , and high temperature series for @xmath129 @xcite ( dotted line ) .
the energy density at the critical point @xmath148 ( horizontal line ) and the critical inverse temperature @xmath149 were computed in mc simulations @xcite ( arrow and vertical line ) .
, title="fig : " ] ( 55,-8)(170,-8 ) ( 110,55 ) ( -5,55 )
in this paper we introduced hypercubic bethe lattices , which are constructed from the cells of regular @xmath0-dimensional cubic lattices , and we analysed the equilibrium properties of spin systems defined on such lattices .
these topologies can be seen as a further generalisations of ordinary bethe lattices , that , unlike tree - like graphs , retain many of the loops of the interaction topologies of more realistic spin systems .
we used the bethe - peierls method to derive equations for the magnetisation per spin and the energy density for ising spin systems on hypercubic bethe lattices , using the factor - tree representation . with these equations we computed phase diagrams for the ferromagnetic ising model on the hypercubic bethe lattice with @xmath146 and @xmath2
the results for the critical temperatures are summarised in table [ table : tc ] , and compared with the values found for true cubic lattices and for ordinary bethe lattices .
hypercubic bethe lattices are found to be more accurate proxies for the true @xmath0-dimensional lattices than regular ( tree - like ) @xmath3 bethe lattices , in terms of the predicted transition temperatures and the values of observables .
however , for @xmath150 one finds , as expected , that the differences between the predictions of all three models become increasingly small .
we expect that for @xmath151 the hypercubic bethe lattice is equivalent to its random graph version , which is a maximally random graph constructed from hypercubes , and with vertices of equal degree , at least for ferromagnets @xcite .
.critical temperatures @xmath152 of ising models on true cubic lattices , hypercubic bethe lattices and @xmath3 bethe lattices , for @xmath23 .
the values for the true @xmath24 lsing lattice and for all hypercubic and ordinary bethe lattices are calculated analytically .
the values of @xmath153 for the cubic lattice with @xmath25 are computed in mc simulations ( see captions of previous figures for references ) . [ cols="<,^,^,^",options="header " , ] our results for the ising model were obtained by computing the sums in the equations ( [ eq : h],[eq : e2],[eq : m2 ] ) directly .
however , even for the hypercube in @xmath154 ( the penteract ) this involves @xmath155 summations ( since the number of corners in the @xmath0-dimensional hypercube is @xmath93 ) , which although in principle easy becomes painful in practice . one interesting future direction following this work would be to consider the one loop correction method of @xcite , which was introduced and used to improve the bethe ( @xmath3 ) estimates for the critical temperature of the ferromagnetic ising model on the @xmath0-dimensional lattice .
generalisation of their method to the hypercubic bethe lattice may further improve our estimates of @xmath153 .
in addition it would be interesting to study a spin - glass on the hypercubic random regular graph ; the spin - glass model on loopy graphs , to the best of our knowledge , was studied only for the husimi lattices @xcite and for a single @xmath0-dimensional hypercube @xcite in the limit of large @xmath0 .
the non - negligible frustration in spin - glasses on random graphs is due to the presence of long @xmath156 loops @xcite , but in our present model it would already enter via the short loops that are present .
the authors would like to thank alessia annibale for interesting discussions .
10 baxter r j 1982 ( academic press ) opper m and saad d 2001 ( mit press ) heydenreich m , van der hofstad r and sakai a 2008 * 132 * 1001 yedidia j s , freeman w t and weiss y 2005 * 51 * 2282 m and parisi g 2001 * 20 * 217 bethe h a 1935 * 150 * 552 bender e a and canfield e r 1978 * 24 * 296 bollobs b 1980 * 1 * 311 johnston d a and plech p 1998 * 31 * 475 gujrati p d 1995 * 74 * 809 dembo a and montanari a 2010 * 20 * 565 hggstrm o 1996 * 104 * 231 montanari a and rizzo t 2005 * 2005 * p10011 parisi g and slanina f 2006 * 2006 * l02003 juri inov e and jur i in m 2012 * 147 * 1077 m and montanari a 2009 ( oxford university press ) onsager l 1944 * 65 * 117 montroll e w , potts r b and ward j c 1963 * 4 * 308 talapov a l and blte h w j 1996 * 29 * 5727 arisue h and fujiwara t 2003 * 67 *
066109 bhanot g , creutz m , horvath i , lacki j , and weckel j 1994 * 49 * 2445 hggkvist r , rosengren a , lundow p , markstrm k , andrn d and kundrotas p 2007 * 56 * 653 vohwinkel c 1993 * 301 * 208 sykes m f 1979 * 12 * 879 lundow p h and markstrm k 2009 * 80 * 031104 lage - castellanos a and mulet r 2008 * 65 * 117 yokota t 2008 * 387 * 3495 fernndez l a , martin - mayor v , parisi g and seoane b 2010 * 81 * 134403 domb c and green m s 1972 academic press
using the energy ( [ def : e ] ) in the definition @xmath157 gives us the specific heat ( density ) equation @xmath158\label{def : c(t)}\ ] ] or equivalently @xmath159 the ising model on a hypercubic bethe lattice is a homogeneous system , i.e. all its nodes are equivalent , so @xmath160 , where @xmath161 and @xmath162 the magnetization @xmath163 is given by the equation @xmath164 and the cavity field @xmath101 satisfies the equation @xmath165 the magnetization and the cavity field are related by the equation @xmath166 which follows from the equations ( [ eq:<s0 > ] ) , ( [ def : f(h ) ] ) and the equality @xmath167 valid for @xmath168 . using the equations ( [ eq:<e0 > ] ) and ( [ def : f(h ) ] ) in the formula ( [ eq : c(1/t ) ] )
gives us the specific heat @xmath169 \nonumber\\ & & - \sum_{j}\left[\langle e_0(s ) s_j\rangle -\langle e_0(s ) \rangle \langle s_j \rangle\right]\frac{\partial h}{\partial\beta}\big\}.\nonumber\end{aligned}\ ] ] let us now assume that @xmath170 ( this property is satisfied in our model ) and consider @xmath171 thus @xmath172 is an odd function of @xmath101 and hence @xmath108 is a solution of the cavity field equation ( [ def : f(h ) ] ) .
furthermore , let us assume that @xmath173 is ferromagnetic , i.e. @xmath174 ( @xmath175 is a set of indices and @xmath176 with @xmath177 ) , and @xmath178 then @xmath172 is a monotonic non - decreasing function of @xmath101 .
to show this we first note that @xmath179}{p^c[-1]}\right)\label{eq : h(p)}\\ p^c[s_0]&=&\frac{\sum_{\{s_j\}}\rme^{-\beta e_0(s ) + h\sum_{j\neq0 } s_j } } { \sum_{\tilde s_0}\sum_{\{\tilde s_j\}}\rme^{-\beta e_0(\tilde s ) + h\sum_{j\neq0 } \tilde s_j}}\label{eq : p^c}\end{aligned}\ ] ] and @xmath180=\frac{1}{2}[1+s_0\langle s_0\rangle_c]$ ] , where the average @xmath181 is generated by the boltzmann weight @xmath182 , then @xmath183.\nonumber \ ] ] secondly the inequality @xmath184 is true by the griffiths - kelly - sherman ( gks ) theorem @xcite and hence @xmath185 . for @xmath108 the cavity magnetizations @xmath186 vanishes and @xmath187 .
furthermore , the gradient of @xmath172 at @xmath108 is a monotonic increasing function of @xmath114 .
this follows from the calculation @xmath188\nonumber\\ & = & \sum_{j\neq0}\sum_a j_a\left [ \langle s_0s_js_a \rangle_c - \langle s_0s_j \rangle_c \langle s_a \rangle_c\right]\geq0,\nonumber\end{aligned}\ ] ] where the inequality is true by the gks theorem . for a @xmath0-dimensional hypercube we have that @xmath189 and
hence there exists @xmath190 such that @xmath191 and the paramagnetic solution @xmath108 becomes unstable when @xmath192 .
if the function @xmath172 is convex in @xmath101 then for @xmath135 there is a unique ( up to @xmath193 ) stable solution @xmath194 of the cavity field equation ( [ def : f(h ) ] ) which corresponds to the ferromagnetic @xmath86 phase . for @xmath24
this solution for @xmath195 is given by @xmath196 but for @xmath25 we were able to obtain these solutions only numerically .
it follows from the equation ( [ eq : c - hyper ] ) that in the paramagnetic phase ( @xmath108 ) the specific heat is given by the equation @xmath197.\label{eq : c - hyper - high - t}\end{aligned}\ ] ] and that it is bounded on the interval @xmath198 $ ] . for @xmath135 the system is in the ferromagnetic phase ( @xmath199 ) and the specific heat is given by the equation @xmath200 \label{eq : c - hyper - low - t}\\ & & - \sum_{j}\left[\langle e_0(s ) s_j\rangle -\langle e_0(s ) \rangle \langle s_j \rangle\right]\frac{\partial h}{\partial\beta}\big\}.\nonumber\end{aligned}\ ] ] let us consider the last line of the above equation .
@xmath201=\sum_a j_a \left[\langle s_a s_j\rangle -\langle s_a \rangle \langle s_j \rangle\right]$ ] is bounded and positive ( by the gks theorem @xcite ) .
furthermore , it follows from the magnetization equation ( [ eq : m(h ) ] ) that the derivative @xmath202 is also positive ( @xmath113 is a monotonic non - decreasing function of @xmath114 ) but is not necessary bounded .
for @xmath24 it diverges as @xmath203 , but the @xmath204 is bounded and as @xmath205 the @xmath206 , so the specific heat has a jump ( see figure [ figure:2d ] ) at @xmath207 .
the specific heat @xmath208 is bounded for any finite @xmath0 as @xmath205 ( this follows from the equation ( [ eq : c - hyper - high - t ] ) ) , but for @xmath203 we were able to verify this behaviour ( as in figure [ figure:2d ] ) only numerically in @xmath25 .
for @xmath139 the formulae for the cavity field @xmath101 , the magnetisation per spin @xmath113 and the energy density @xmath129 become , respectively , @xmath209\nonumber \\[2 mm ] m&=&\big(-12 { \rme^{-16 \beta-2 h}}+12 { \rme^{16 \beta+8 h}}-16 { \rme^{-12 \beta-4 h}}+60 { \rme^{16 \beta+10 h}}\label{eq:4d - m}\\ & & + 174 { \rme^{2 h-8 \beta}}+24 { \rme^{20 \beta+12 h}}+732 { \rme^{4 h}}+258{\rme^{8 \beta+6 h}}\nonumber\\ & & + 48 { \rme^{-8 \beta+6 h}}+2 { \rme^{-24 \beta+2 h}}+14 { \rme^{4 h-16 \beta}}+114 { \rme^{8 h}}+130 { \rme^{8 \beta+10 h}}\nonumber\\ & & + 66 { \rme^{16 \beta+12 h}}+392 { \rme^{4 h-4 \beta}}+624 { \rme^{4 \beta+6 h}}+128 { \rme^{12 \beta+8 h}}\nonumber\\ & & + 12 { \rme^{-16 \beta+2 h}}+192 { \rme^{4 h-8 \beta}}+468 { \rme^{6 h}}+64 { \rme^{2 h-12 \beta}}\nonumber\\ & & + 14 { \rme^{24 \beta+14 h}}+168 { \rme^{6 h-4 \beta}}+16{\rme^{-12 \beta+4 h}}\nonumber\\ & & + 288 { \rme^{4 \beta+8 h}}+160 { \rme^{12 \beta+10 h}}+{\rme^{32 \beta+16 h}}\nonumber\\ & & -{\rme^{32 \beta-16 h}}-14 { \rme^{24 \beta-14 h}}-24 { \rme^{20 \beta-12 h}}\nonumber\\ & & -66 { \rme^{-12 h+16 \beta}}-60 { \rme^{16 \beta-10 h}}-12 { \rme^{16 \beta-8 h}}\nonumber\\ & & -160 { \rme^{12 \beta-10 h}}-128 { \rme^{12 \beta-8 h}}-176 { \rme^{-4 h+8 \beta}}\nonumber\\ & & -8 { \rme^{12 \beta-2 h}}-130 { \rme^{8 \beta-10 h}}-368 { \rme^{8 \beta-8 h}}\nonumber\\ & & -72 { \rme^{12 \beta-6 } } -258 { \rme^{8 \beta-6 h}}-24 { \rme^{12 \beta-4h}}\nonumber\\ & & -114 { \rme^{-8 h}}-624 { \rme^{4 \beta-6 h}}-288 { \rme^{4 \beta-8 h}}-468 { \rme^{-6 h}}\nonumber\\ & & -456 { \rme^{4 \beta-4 h}}-732 { \rme^{-4 h}}-78 { \rme^{8 \beta-2 h}}-48 { \rme^{-8 \beta-6 h}}\nonumber\\ & & -392 { \rme^{-4 \beta-4 h}}-240 { \rme^{-2 h+4 \beta}}-460 { \rme^{-2 h}}-168 { \rme^{-4 \beta-6 h}}\nonumber\\ & & + 8 { \rme^{12 \beta+2 h}}-192 { \rme^{-8 \beta-4 h}}-392 { \rme^{-4 \beta-2 h}}+78 { \rme^{8 \beta+2 h}}\nonumber\\ & & + 240 { \rme^{4 \beta+2 h}}+24 { \rme^{12 \beta+4h}}-14 { \rme^{-16 \beta-4 h}}-174 { \rme^{-8 \beta-2 h}}\nonumber\\ & & -64 { \rme^{-12 \beta-2 h}}+460 { \rme^{2 h}}+176 { \rme^{8 \beta+4 h}}+392 { \rme^{2 h-4 \beta}}\nonumber\\ & & + 456 { \rme^{4 \beta+4 h}}+72 { \rme^{12 \beta+6 h}}-2 { \rme^{-24 \beta-2 h}}+368 { \rme^{8 \beta+8 h}}\big)/\mathcal{n}\nonumber \\[2 mm ] e&= & -4\big(12 { \rme^{24 \beta+14 h}}+{\rme^{32 \beta+16 { h}}}+{\rme^{32 \beta-16 h}}+12 { \rme^{24 \beta- 14 h}}\label{eq:4d - e2}\\ & & ~~+20 { \rme^{20 \beta-12
h}}+44 { \rme^{-12 h+16
\beta}}+48 { \rme^{16 \beta-10 h}}+12 { \rme^{16 \beta-8 h}}\nonumber\\ & & ~~+96 { \rme^{12 \beta-10 { h}}}+96 { \rme^{12 \beta-8 h}}+52 { \rme^{8 \beta-10 h}}+184 { \rme^{8 \beta-8 h}}\nonumber\\ & & ~~+72 { \rme^{12 \beta-6 h}}+172 { \rme^{8 \beta-6 { h}}}+36 { \rme^{12 \beta-4 h}}+208 { \rme^{4 \beta-6 h}}\nonumber\\ & & ~~+176 { \rme^{-4 h+8 \beta}}+24 { \rme^{12 \beta-2 h}}+4 { \rme^{16 \beta}}+72 { \rme^{4 \beta-8 h}}\nonumber\\ & & ~~+228 { \rme^{4 \beta-4 { h}}}+156 { \rme^{8 \beta-2 h}}-32 { \rme^{-8 \beta-6h}}-196 { \rme^{-4 \beta-4 h}}\nonumber\\ & & ~~+240 { \rme^{-2 h+4 \beta}}+192 { \rme^{8 \beta}}-56 { \rme^{-4\beta-6 h}}+200 { \rme^{4 \beta}}\nonumber\\ & & ~~+24 { \rme^{12 \beta+2 h}}-192 { \rme^{-8 \beta-4 h}}-392 { \rme^{-4 \beta-2 h}}+156 { \rme^{8 \beta+2 { h}}}\nonumber\\ & & ~~+240 { \rme^{4 \beta+2 h}}+36 { \rme^{12 \beta+4 h}}-28 { \rme^{-16 \beta-4 h}}-348 { \rme^{-8 \beta-2 h}}\nonumber\\ & & ~~-192 { \rme^{-12 \beta-2 { h}}}-408 { \rme^{-4 \beta}}-528 { \rme^{-8 \beta}}+176 { \rme^{8 \beta+4 h}}\nonumber\\ & & ~~-392 { \rme^{2 h-4 \beta}}+228 { \rme^{4 \beta+4 h}}+72 { \rme^{12 \beta+6 h}}+172 { \rme^{8 \beta+6 h}}\nonumber\\ & & ~~+12 { \rme^{16 \beta+8 h}}-24 { \rme^{-12 \beta-4 { h}}}-48 { \rme^{-16 \beta-2 h}}-216 { \rme^{-12 \beta}}\nonumber\\ & & ~~-348 { \rme^{2 h-8 \beta}}-196 { \rme^{4 { h}-4 \beta}}+208 { \rme^{4 \beta+6 h}}+96 { \rme^{12 \beta+8 h}}\nonumber\\ & & ~~-12 { \rme^{-24 \beta-2 { h}}}-60 { \rme^{-16 \beta}}-40 { \rme^{-20 \beta}}-192 { \rme^{2 h-12 \beta}}\nonumber\\ & & ~~-48 { \rme^{-16 \beta+2 { h}}}-192 { \rme^{4 h-8 \beta}}+184 { \rme^{8 \beta+8 h}}+48 { \rme^{16 \beta+10 h}}\nonumber\\ & & ~~-56 { \rme^{6 h-4 \beta}}-24 { \rme^{-12 \beta+4 { h}}}+72 { \rme^{4 \beta+8 h}}+96 { \rme^{12 \beta+10 h}}\nonumber\\ & & ~~+20 { \rme^{20 \beta+12 h}}-2 { \rme^{-32 \beta}}-12 { \rme^{-24 \beta+2 h}}-28 { \rme^{4 h-16 \beta}}\nonumber\\ & & ~~-32 { \rme^{-8 \beta+6 h}}+52
{ \rme^{8 \beta+10 h}}+44 { \rme^{16 \beta+12 { h}}}\big)/\mathcal{n}\nonumber \ ] ] with the normalisation factor @xmath210 | we study spin systems on bethe lattices constructed from @xmath0-dimensional hypercubes .
although these lattices are not tree - like , and therefore closer to real cubic lattices than bethe lattices or regular random graphs , one can still use the bethe - peierls method to derive exact equations for the magnetization and other thermodynamic quantities .
we compute phase diagrams for ferromagnetic ising models on hypercubic bethe lattices with dimension @xmath1 and @xmath2 .
our results are in good agreement with the results of the same models on @xmath0-dimensional cubic lattices , for low and high temperatures , and offer an improvement over the conventional bethe lattice with connectivity @xmath3 . |
the introduction of submillimeter ( submm ) , imaging arrays such as scuba ( submillimeter common user bolometer array ; holland et al 1999 ) , heralds a revolution in our understanding of the dust properties of external spiral galaxies . for the first time
, we can map cold ( 15 - 20k ) interstellar grains with high sensitivity and good spatial resolution ( @xmath4 ) ( chini et al 1995 ; hughes et al 1997 ; clements et al 1993 ) .
previous studies of radiation emitted by dust have relied heavily on measurements carried out by the infrared astronomical satellite ( iras ) at 60 and @xmath5 m . these wavebands are sensitive primarily to warm ( @xmath3 30k ) dust , and as such , are unlikely to give an accurate impression of either the true quantity or distribution of interstellar grains in external disks ( alton et al 1998a ; devereux & young 1990 ) .
furthermore , satellite missions such as iras , and the recently completed infrared space observatory ( iso ) project , are constrained at present to using small mirrors ( 0.6-m c.f . scuba 15-m dish ) .
this limits their resolving capability to @xmath6 in the far - infrared ( fir ) which is only adequate for probing disk structure in the very closest spiral galaxies ( alton et al 1998b ) .
+ one of the first studies of nearby galaxies , using scuba , was carried out by alton et al ( 1998c ) who mapped the edge - on galaxy ngc 891 at 450 and @xmath0 m ( see also israel et al 1999 ) . by carefully comparing the submm emission over the same region as the fir flux ( @xmath7 ) , they found that this galaxy contains about an order of magnitude more dust than would have been inferred from iras observations alone .
this same conclusion was reached by xilouris et al ( 1999 ) who fitted the optical and near - infrared surface photometry of both ngc 891 and 6 other nearby , edge - on spirals with a sophisticated radiation transfer simulation .
their model , which takes account of both scattering and absorption by interstellar grains , only produces a self - consistent solution across several wavebands if the extinction lane harbours about 10 times more dust than might be expected from iras measurements . + in this paper
, we relate the extinction modelling carried out for ngc 891 ( xilouris et al 1998 ; hereafter xad ) with the corresponding submm images acquired from scuba ( alton et al 1998c ; hereafter abr ) . in doing so
, we address some important issues , namely : 1 .
whether the submm emission , which accounts for the bulk of grain material in ngc 891 , correlates closely with the dust evident in the extinction lane ( [ rtm ] ) ; 2 . in the case of a positive correlation ,
whether the submm emissivity in ngc 891 ( or , equivalently , the grain mass - absorption coefficient ) is comparable to that recorded for our own galaxy ( [ emissivity ] ) ; 3 .
given that we recover most of the dust mass from our submm images , how the gas - to - dust ratio varies along the major axis of ngc 891([gdratio ] ) ; 4 . how the submm emission is distributed with respect to the neutral gas phases ( hi , h@xmath8 ) and the stars in the disk ( [ gdratio ] ) .
this information will prove crucial to the growing number of submm surveys , where sources are poorly resolved and subsidary observational data are limited .
we begin by summarizing the salient results from xad and abr ( [ submm ] and [ rtm ] ) before seeking to relate the submm to the extinction dust lane .
abr obtained the deepest images yet of a nearby galaxy in the submillimeter waveband .
they detected emission from @xmath9 of the optical disk , down to a noise - limited surface brightness of 3.5mjy/@xmath10beam and 13mjy/@xmath11beam at 850 and @xmath12 m respectively . at both submm wavelengths ,
the flux exhibits a pronounced peak at the nucleus and secondary maxima @xmath13 either side of the centre ( equivalent to 3 kpc for a distance of 9.5 mpc , adopted by both abr and xad , to ngc 891 ) .
the secondary maxima may be attributable to a ring structure seen edge - on or limb - brightening associated with spiral arms .
we do not present the scuba maps here , but refer to abr for both these images and more technical details _ vis - a - vis _ the observing procedure .
the submm maps are much more radially - extended than the corresponding emission detected by iras , indicating that longer wavelength fir radiation dominates at larger galactic radii .
+ abr made a careful estimate of the ratio of cold dust to warm grains by smoothing all their submm / fir images of ngc 891 ( 60,100,450 & @xmath0 m ) to the same resolution and then dividing the major axis into spatially - independent radial bins .
this uniform treatment of the data was considered of major importance because there had been a tendency , in the past , to compare photometry at different wavelengths without regard for , or knowledge of , differences in the size of the relevant emitting region(s ) ( e.g. guelin et al 1993 ; israel et al 1999 ; clements et al 1993 ) .
the flux density emanating from each radial bin was matched by the superposition of 2 greybody curves .
thus : @xmath14 where @xmath15 is the flux density at wavelength @xmath16 ( in wm@xmath17hz@xmath18 ) , @xmath19 is the geometrical cross - sectional area of an interstellar grain , @xmath20 is the distance to ngc 891 and @xmath21 is the submm emissivity . @xmath22 and @xmath23 are the number of emitters ( grains ) in the warm and cold dust components respectively whilst @xmath24 and @xmath25 represent the corresponding blackbody intensities .
@xmath21 is sometimes refered to as the emission efficiency and can be considered as the ratio of the emission cross - section to the geometrical cross - section of the grain ( spitzer 1978 ; whittet 1992 ) .
+ abr found that all radial bins could be fitted well with @xmath26=15 - 20k was given as @xmath276k but the more correct range is 15 - 20k ] and @xmath28=30 - 37k , and @xmath29 increased from about 26 at the nucleus to 54 at @xmath30 .
the _ ratio _
@xmath29 is fairly secure .
_ however _ , the exact dust masses pertaining to @xmath23 and @xmath22 depend critically on the absolute level of @xmath21 which , as we shall see , is highly uncertain .
at this stage , it is important to emphasize that over 90% of the @xmath0 m emission arises , at all radii , from the cold ( 15 - 20k ) dust component and that this material , in turn , constitutes the bulk of galactic dust . in other words
, our @xmath0 m scuba map indicates , almost completely , how galactic dust is distributed within ngc 891 .
moreover , if we could avail ourselves of a well - determined value for the grain emissivity at @xmath0 m , @xmath31 , our scuba image would yield , to a good approximation , the total dust mass in the disk .
+ the submm / fir emissivity , @xmath21 , is a poorly known quantity and virtually the only sources of information stem from hildebrand ( 1983 ) and draine & lee ( 1984 ) , which are cited almost exclusively in the literature .
the measurement of hildebrand ( 1983 ) is based on kuiper observations ( 55 and @xmath32 m ) of a single galactic reflection nebula , ngc 7023 , where attenuation of ultraviolet light from the central star is effectively used to constrain the amount of dust present . in order to account for the observed fir flux densities ,
the emissivity at @xmath32 m was estimated to lie in the range @xmath33 .
although this represents little better than an order of magnitude determination , the emissivity of hildebrand has been applied by fir observers almost religiously since its derivation nearly 2 decades ago .
this is at least partially attributable to a dearth of measurements in the interim years , in particular observations applicable to the diffuse interstellar medium ( ism ) where grains may well differ from their particulate counterparts in reflection nebulae ( pendleton et al 1990 ; mathis 1990 ) .
we note that casey ( 1991 ) has repeated the hildebrand procedure for 5 galactic reflection nebulae ( and derived an emissivity approaching the upper limit of the hildebrand determination ) , but this work has been seldom cited in the literature .
draine and lee ( 1984 ) have used a _ pot pourri _ of laboratory measurements and astronomical data to construct dielectric functions appropriate to graphite and silicate interstellar grains . in this way
, they have derived a fir emissivity close to the mean value given by hildebrand .
finally , we note that bianchi et al ( 1999 ) recently determined an emissivity value of @xmath34 for the diffuse dust in the milky way at @xmath5 m .
their technique relied on the tight spatial correlation observed between interstellar extinction and the fir radiation detected by cobe .
this allowed them , in a similar fashion to hildebrand , to constrain the amount of dust responsible for the fir emission .
depending on the exact behaviour of @xmath35 with wavelength ( discussed below ) , the bianchi et al value is somewhere towards the higher end of the hildebrand range .
+ in the following sections , we attempt an independent determination of the submm / fir emissivity in ngc 891 by using extinction observed in the disk to constrain the amount of grain material present .
this approach differs from previous treatments of extragalactic submm / mm observations where , to one extent or another , workers have tended to make use of dust properties in the milky way in order to infer quantities in external galaxies .
for example , guelin et al ( 1993 ) use a galactic emissivity to turn their 1.3 mm observations into a dust column density . assuming a solar gas - to - dust ratio
they then quantify how much ( molecular ) gas is present in the system . in a slightly different approach , neininger et al ( 1996 ) use 21 cm emission to infer a dust column density ( again via a galactic gas - to - dust ratio ) and then compare this quantity with their observed submm / mm emission in order to derive the corresponding emissivity .
our current method is based on the technique pioneered by hildebrand ( 1983 ) .
in essence , the dust column density is measured by means of optical / ultraviolet extinction and this quantity is then compared to the observed fir / submm emission in order to infer the long wavelength emissivity . in deriving @xmath21
we make no direct use of galactic properties as such . whilst we follow the same general method as hildebrand , we believe that the submm window produces more robust values of emissivity than fits made to the emission near the greybody peak ( hildebrand fits near @xmath5 m ) .
the reason for this is that , for any observed level of fir / submm flux , the implied dust mass will vary at least as strongly as @xmath36 for emission near the peak but only @xmath37 in the rayleigh - jean tail ( where @xmath38 is the grain temperature ) .
if the dust mass is fixed _ a priori _ by the amount of extinction , the corresponding uncertainties are then transfered to the derived values of emissivity .
thus @xmath39 will be much more dependent on the grain temperature than @xmath40 and , as a consequence , far more uncertain ( see hughes et al 1997 for an indepth discussion ) .
+ before we are able to embark on this process for ngc 891 , we have to be sure that the submm , which traces the bulk of the dust in this galaxy , corresponds closely to the grain material responsible for the absorption lane .
to this end , we briefly describe the photometric fit made to ngc 891 by xad which yields directly the optical depth at each point in the disk .
the approach adopted by xad ( see also xilouris et al 1997;1999 ) is to simulate the optical / nir appearance of edge - on spirals by creating a model galaxy consisting of an exponential stellar disk , a r@xmath41 bulge and an exponential dust distribution .
a pixel - to - pixel comparison is made between the real and simulated object in order to fit scale - heights ( z - direction ) and scale - lengths ( radial direction ) to both the stellar and dust disks and to determine the optical depth through the centre of the galaxy ( as seen face - on ) .
in addition to radiation absorption , the model takes account of photon scattering in the disk by adopting the henyey - greenstein phase function ( henyey & greenstein 1941 ; bohren & huffman 1983 ) .
our confidence in the model is significantly boosted by the fact that the output parameters are highly consistent across several optical and nir wavebands .
thus for ngc 891 , the radial scale - length for the grain distribution is calculated to be 8.1 kpc in the v - band whilst a determination in the remaining filters ( k , j , i , b ) strays by only 5 - 10% from this value .
simulations for 7 edge - on galaxies indicate , in all cases , an extinction lane consistent with galactic - type reddening .
it is important to emphasize that our modelling of the dust lane vastly transcends simplistic screen models ( which have often been applied in the past ) and takes full account of mixing between stars and dust . admittedly , one obvious limitation to our simulation is that no account is taken of dust clumping or indeed spiral structure within the disk .
kylafis et al ( 1999 ) have tested the influence of spiral structure on the radiation transfer in ngc 891 and found that it produces almost no change to the large - scale properties inferred from the model .
clumping , on the other hand , may well ` hide ' large amounts of dust so that it can not be detected by a fit to the large - scale extinction .
we postpone a discussion on the effects that clumping might have on our results until later ( [ discussion ] ) .
+ since we shall be concentrating on the v - band when making a comparison between extinction in ngc 891 and the submm emission detected by abr , we reproduce in table [ tab1 ] the relevant parameters fitted by xad at this wavelength . using the tabulated parameters , we can construct a synthetic map of v - band optical depth ( @xmath42 ) , which , after smoothing to the same spatial resolution as the @xmath0 m scuba map ( @xmath10 fwhm ) , can be compared directly with the submm image . in figures [ fig2 ] and [ fig3 ] , we display profiles of @xmath42 along the minor and major axes of the galaxy model after the smoothing process .
the corresponding @xmath0 m cross - sections are also shown . in all cases ,
a width of @xmath43 ( @xmath44 beam fwhm ) is used to sample the emission / optical - depth perpendicular to the profile direction . the submm profile , in fig .
[ fig2 ] , is the mean of several transects across the major axis taken at various distances from the galactic nucleus . the average profile has been normalized , along with the @xmath42 curve , to unity at zero z - height .
we can see that the observed and model profiles are extremely similar in this direction ( a deconvolved size of @xmath45 or 610 pc for both fwhm ) . to some extent , this similarity may be attributable to the point spread function ( psf ) being quite large with respect to the intrinsic dust layer .
however , by artificially varying the scale - height of @xmath42 , before smoothing , we have assured ourselves that the thickness of the absorption lane must lie within 30% of the submm width . for z - heights of @xmath46 ( 900 - 2300 pc ) , there is an excess at @xmath0 m over what would be expected purely from the extinction model .
the scuba beam is principally gaussian but slight wings in the psf may contribute to this observed excess .
in fact , until a more careful comparison is made between the high - z ` emission ' and the psf wings ( alton et al 1999a ) , we are not in a position to attribute this excess to the ` vertical ' dust chimneys evident in optical images of ngc 891 ( howk & savage 1997 ; dettmar 1990 ) .
+ for profiles along the major axis , we plot the optical depth given by the model against the observed @xmath0 m surface brightness ( fig .
[ fig3 ] ) .
it should be noted that when the comparison is made here , it is after the opacity model has been smoothed to the same resolution as the submm image ( thus @xmath42 is a factor of 4 lower than the optical depth directly inferred from the original optical image ) .
the correspondance along the major axis is not as good as that for the minor axis and can only be described as fair .
due to its simplicity , we clearly can not expect the xad model to reproduce the local fluctuations in the submm profile .
however , on average , the @xmath42 curve still appears to be somewhat more extended than the @xmath0 m emission .
in fact , if we could choose a @xmath42 curve that would match the behaviour of scuba data perfectly , the dust model would have a scale - length of @xmath475.3 kpc as opposed to 8.1 kpc derived by xad . under such circumstances the dust layer would possess the same radial fall - off as the v - band stars .
the ratio between optical depth and @xmath0 m surface brightness , averaged over the profiles in fig .
[ fig3 ] , is as follows : @xmath48 where @xmath49 is the submm surface brightness , in jy/@xmath10beam , and the uncertainty represents the standard deviation of a linear fit between @xmath49 and @xmath42 .
+ although the major axis fit between submm emission and optical attenuation is only fair , we will assume that in both cases the same population of grains manifests itself .
certainly , simulations of dust bathed in interstellar radiation fields suggest that the classical ` big grains ' , which are responsible for extinction ( size @xmath50 m ) , increasingly dominate the emission process at wavelengths beyond @xmath51 m ( desert et al 1990 ) . and
indeed , observationally , bianchi et al ( 1998 ) have recently established a convincing correlation between optical extinction and submm emission in the nearby spiral ngc 7331 .
for ngc 891 , we therefore fix the @xmath0 m surface brightness with respect to the visual optical depth , @xmath42 , according to eq .
[ eq1 ] . by doing this
, we will be able to derive the dust emissivity at @xmath0 m .
this calculation is carried out in the next section .
as mentioned in [ submm ] , the @xmath0 m waveband is completely dominated by cold dust ( 15 - 20k ) at all positions in the disk .
in fact , the rayleigh - jeans tail of the warm grain component contributes about 8% to the flux density in this filter .
it is possible , therefore , to rewrite eq .
[ eq1 ] approximately as follows : @xmath52 where @xmath23 , the number of cold grains , comprises the bulk of interstellar dust in ngc 891 .
in order to be able to use the ratio @xmath53 ( eq . [ eq2 ] ) in this relation , we have to refer to the definition of optical depth , namely : @xmath54 where @xmath55 is the dust column density ( grains per unit surface area of galaxy ) , @xmath19 , as before , represents the geometrical grain cross - section and @xmath56 constitutes the emissivity , or extinction efficiency , in the v - band .
+ if we can derive the flux density per unit surface area , @xmath23 in eq .
[ eq3 ] becomes @xmath55 and we will be in a position to substitute @xmath42 , from eq .
[ eq4 ] , into eq .
[ eq3 ] . to do this we use @xmath49 , the @xmath0 m surface brightness ( in jy/@xmath10beam ) . dividing @xmath49 by the beam area ( in m@xmath57 ) , and multiplying by @xmath58 to convert from jansky to wm@xmath17hz@xmath18
, we obtain the flux density per unit surface area ( as required ) .
this can be substituted into the left - hand side of eq .
[ eq3 ] thus : @xmath59 where @xmath55 has now replaced @xmath23 .
+ substituting @xmath55 with @xmath60 ( eq . [ eq4 ] ) , we can eliminate the geometrical grain cross - section @xmath19 ( the distance d also falls out ) , leaving : @xmath61 for the blackbody intensity , @xmath62 , we adopt a grain temperature of @xmath63 k. then , comparing with eq .
[ eq2 ] , we obtain the emissivity at @xmath0 m with respect to the v - band emissivity as : @xmath64 to derive @xmath65 directly from eq .
[ eq7 ] , we must have some knowledge of @xmath56 . for ` classical ' grains of radius @xmath66 m , which are responsible for optical extinction in the diffuse ism of the milky way , @xmath56 lies between 1 and 2 ( whittet 1992 ; alton 1996 ; spitzer 1978 ) .
the interstellar grains of ngc 891 are unlikely to be much different in this respect , given that extinction within the disk corresponds extremely well to a galactic reddening law ( [ rtm ] ) .
thus we use @xmath67 in eq .
[ eq7 ] to infer : @xmath68 the grain emissivity within the milky way has usually been estimated at shorter fir wavelengths ( e.g. hildebrand finds @xmath69 ) .
it is difficult , therefore , to compare galactic data with eq .
[ eq8 ] unless we assume some sort of wavelength behaviour for the dust emissivity .
fortunately , a large amount of recent observational evidence seems to be converging on an emissivity , _ for the general ism of non - active spirals _ , which varies as @xmath70 where the spectral index @xmath71 for @xmath72 m .
indeed , rigorous fir spectral measurements of milky way dust , using both cobe and balloon - borne instruments , imply a spectral index between 1.5 and 2.0 for the diffuse ism in the galaxy ( masi et al 1995 ; reach et al 1995 ) .
likewise , the relatively high flux density at @xmath12 m compared to emission at @xmath0 m , for galaxies such as ngc 891 and ngc 7331 , is indicative of an index @xmath73 ( bianchi et al 1998 ) .
accordingly , we plot in fig .
[ fig4 ] the dust emissivity for ngc 891 ( as a ratio @xmath74 ) assuming a wavelength dependency @xmath75 ( @xmath76 ) .
estimates for the fir emissivity of milky way dust , as derived by hildebrand ( 1983 ) , casey ( 1991 ) , draine & lee ( 1984 ) and bianchi et al ( 1999b ) ( see [ submm ] ) , are annotated on this plot .
we also add the emissivity measured by agladze et al ( 1994 ) for ` astronomical - type ' silicates tested at 20k under laboratory - controlled conditions .
+ the fir grain emissivity in ngc 891 appears to be somewhat higher than most estimates of the same property in the milky way . in particular , the ` canonical ' values of draine & lee ( 1984 ) and hildebrand ( 1983 ) are about a factor of 2 - 3 lower than the quantity we have derived here .
the discrepancy is admittedly smaller for an emissivity law following @xmath77 , but both the draine and lee model and the bianchi determination are consistent with @xmath78 and should , therefore , strictly speaking , be compared with the steepest line in fig .
similarly , the lowest values of @xmath79 , as given by hildebrand , are derived under the assumption that @xmath78 .
therefore , the agreement in this case can only be described as marginal . in this respect
, the ngc 891 emissivity seems more consistent with the casey ( 1991 ) observations ( although why there should be such a difference between the hildebrand ( 1983 ) and casey ( 1991 ) measurements , when both apply to galactic reflection nebulae , is not altogether clear ) .
although the fir emissivity we predict for ngc 891 is somewhat higher than most values used for the galaxy , it is still much lower than the inference from laboratory - controlled experiments .
indeed , the values of @xmath65 , implied by agladze et al , are astonishingly high , although the authors themselves do not offer an explanation as to why there is such an incongruence with astronomical observations .
+ it would certainly be something of a surprise if interstellar grains in ngc 891 possess a submm emissivity grossly different from milky way dust ( particularly since xad find that extinction within ngc 891 is characterized by a galactic reddening law ) .
the velocity fields and distribution of neutral gas within the disks of both systems has been noted by several authors as being strikingly similar ( guelin et al 1993 ; garcia - burillo et al 1992 ; scoville et al 1993 ) .
however , it is possible that ngc 891 may display a somewhat more active halo with ` thick disks ' of both neutral and ionized gas extending to several kpc above the midplane ( dettmar 1990 ) .
since this extended disk only contains a few percent of the total dust in ngc 891 ( howk & savage 1997 ; alton et al 1999a ) , the activity at the disk - halo interface should not necessarily imply any general pecularities for the dust properties of the _ main _ disk .
although the error bar in fig . [ fig4 ] represents only a 30% standard deviation in the fit between the submm and opacity measurements , the total uncertainty in our estimate of @xmath65 is quite likely to be about a factor of 2 ( due to @xmath80% error in both @xmath81 and @xmath82 ) .
moreover , measurements of fir emissivity for galactic grains are likely to be uncertain by an order of magnitude once extrapolated to submm wavelengths ( hughes et al 1993 )
. therefore we should not be unduly worried by the factor 2 - 3 difference between the emissivity derived for ngc 891 and estimates of the corresponding property for the milky way .
it is interesting to note that grain models which successfully reproduce the extinction law within our own galaxy , can display radically different submm emissivities .
for example , the fluffy agglomerates of silicate and carbon , that have been postulated by matthis & whiffen ( 1989 ) , satisfy the near - infrared to ultraviolet reddening law in much the same way as the draine & lee ( 1984 ) model .
however in comparison to the latter , they possess an emissivity nearly 3 times larger at @xmath83 .
+ as a confirmatory _ addendum _ to the preceding paragraph , we note that dunne et al ( 1999 ) estimate the submm mass - absorption coefficient in ngc 891 ( @xmath84 ) by using our reported values of @xmath0 m flux density ( abr ) and fixing the dust mass via depletion arguments rather than observed extinction ( a constant fraction of interstellar metals are assumed to be bound up in grains ) .
the dunne et al analysis yields @xmath85 m@xmath57kg@xmath18 which , for grains of radius @xmath66 m and material density 3000 kgm@xmath86 , corresponds to @xmath87 , i.e. almost exactly the value we derive in eq . [ eq8 ] . ] + as mentioned previously , one cause for concern in our derivation of submm emissivity is the fact that the opacity simulations of xad do not account for dust clumping .
this will tend to _ underestimate _ the amount of dust causing extinction within the disk .
witt et al ( 1999 ; also witt priv .
have analysed v - k excesses from spirals disks using homogeneous and clumpy - phase models and found that the latter typically require 50% more dust mass for the same level of extinction .
similarly , the monte carlo radiative transfer simulations of bianchi et al ( 1999b ) , which assign dust clumps according to the distribution of molecular gas , indicate that homogeneous models , such as xad , will underestimate the dust mass by a factor of 2 for spiral galaxies viewed edge - on .
we have attempted to assess the effects of clumping in ngc 891 ourselves , by estimating the amount of dust in gas clouds too optically thick to be detected satisfactorily by the xad homogeneous model .
our reddest waveband is k , therefore our technique becomes insensitive to environments with @xmath88 ( @xmath89 ) . within the galaxy
, this corresponds to a column density @xmath90 of @xmath91 cm@xmath17 ( bohlin et al 1978 ) which is close to the column density within most giant molecular clouds ( gmcs ) [ see for example larson ( 1981 ) , who found @xmath93 cm@xmath17 / @xmath94 for cloud sizes of @xmath95=0.05 - 100pc ] .
it is possible , then , that we may ` miss ' a sizeable fraction of grains residing in gmcs . by comparing relative emission strengths detected from @xmath96co and @xmath97co molecules in various gas clouds , polk et al ( 1988 )
estimate that 50% of molecular gas resides in galactic gmcs and the remainder constitutes a more diffuse medium surrounding gmcs .
this information , then , suggests that the xad model will ` overlook ' about 50% of the h@xmath8 phase and , in total , perhaps 30% of grains associated with the ( hi+h@xmath8 ) gas phase of ngc 891 . since the submm is optically thin , any increase in the amount of dust inferred from extinction would decrease @xmath65 in eq .
indeed , a 50 - 100% rise in @xmath42 , due to clumping , will yield a value of @xmath98
@xmath99 for ngc 891 .
this is much closer to extrapolations of the galactic emissivity into the submm waveband ( @xmath100 ) .
intriguingly , if dust clumps are preferentially located towards the centre of ngc 891 ( where the molecular gas is known to be concentrated ; guelin et al 1993 ) , a correction for clumping would steepen the @xmath42 profile in fig .
[ fig3 ] to a shape more akin to the submm distribution .
+ the submm emissivity we have derived for ngc 891 ( eq . [ eq8 ] ) can now be used , in conjunction with our scuba images , to create dust maps for the disk . in the following section ,
these maps are employed to derive the gas - to - dust mass ratio along the major axis . before embarking on this part of the analysis , however , we recap the method we have used to find dust masses in ngc 891 . in lieu of extrapolating the fir emissivity estimated for the milky way to submm observations of ngc 891
, we have taken the visual optical depth calculated from radiation transfer modelling ( xad ) , and effectively constrained the number of grains , @xmath23 , appearing in eq .
this technique relies on knowledge of the extinction efficiency in the v - band , @xmath56 .
however , this quantity is relatively well known ( @xmath101 ) , compared with the near order of magnitude uncertainty in @xmath65 assumed for the milky way ( hughes et al 1993 ; hughes et al 1997 ) .
fixing @xmath23 , allows us to use eq .
[ eq3 ] to derive the @xmath0 m emissivity from the @xmath0 m flux densities in our scuba maps . + now that we have a robust estimate of @xmath65 , we use the following formula to derive the dust mass , @xmath102 ( kg ) , at each point in the disk : @xmath103 where @xmath104 and @xmath105 are the radius and material density of the interstellar grains , respectively , and @xmath20 is the distance to ngc 891 ( see , for example , hildebrand ( 1983 ) for derivation of eq .
[ eq9 ] ) .
the determination of @xmath102 _ will inevitably _ rely on assumptions about the grain properties ( typically @xmath106 m and @xmath107 kgm@xmath86 are applied ) , and , furthermore , some uncertainty is introduced by @xmath20 appearing as the square in eq .
however , the evaluation of @xmath102 is greatly improved by our newly - acquired knowledge of @xmath65 .
moreover , when we consider the neutral gas in ngc 891 , we can expect the gas - to - dust mass ratio to be independent of distance . as a final note , we underline the fact that the blackbody intensity in eq .
[ eq9 ] , @xmath108 , varies linearly with temperature because we are in the rayleigh - jeans tail of the greybody curve .
this means that @xmath108 will change by as little as @xmath109 over the suggested temperature range of 15 - 20k .
+ it is important to point out that , whilst we use eq . [ eq9 ] to infer the dust distribution from our @xmath0 m map of ngc 891 , the overall dust mass has already been normalized to @xmath42 .
this is because we have substituted @xmath42 into eqs .
[ eq6]-[eq8 ] in order to determine @xmath65 .
thus , whilst our scuba map provides the detailed spatial information concerning the dust distribution , the total amount of dust will not be any different from what we would have directly inferred from the optical depth model of xad .
we produce a dust map of ngc 891 by using d=9.5 mpc , @xmath106 m and @xmath107 kgm@xmath86 in equation [ eq9 ] .
the grain distribution can then also be expressed as : @xmath110 where @xmath49 is the @xmath0 m surface brightness in jy/@xmath10beam .
+ due to the proximity of ngc 891 the neutral gas in this galaxy has been reasonably well investigated .
rupen et al ( 1991 ) mapped out the atomic hydrogen with a high sensitivity and an effective beam of @xmath111 .
the atomic gas distribution can be described as a ` plateau ' which extends approximately out to the @xmath112 but is sharply truncated at this radius . within the inner few kpc of the stellar disk there is a relative dearth of hi .
the molecular gas , on the other hand , appears to be fairly centrally concentrated , with a distribution similar to the submm emission detected by scuba .
scoville et al ( 1993 ) employed high resolution interferometry ( @xmath113 ) to map ngc 891 out to radius of 10 kpc in the co ( j=1 - 0 ) line .
emission appears to be restricted to a thin disk ( @xmath114 fwhm ) with a radial distribution reminiscent of our own galaxy ( young & scoville 1991 ) . +
in deriving the gas - to - dust mass ratio in ngc 891 we make use of these neutral gas measurements but ensure that all the relevant data are smoothed to a common spatial resolution .
thus , we took the co interferometry data of scoville et al and the dust map based on the @xmath0 m observations , and convolved them with a gaussian of @xmath115 and @xmath116 fwhm respectively .
this produced images of comparable resolution to the rupen et al map at 21 cm .
we then profiled along the major axis of each image using a bin width of @xmath43 and a sampling interval of @xmath117 . to convert the co emission to a h@xmath8 column density we used a value of @xmath118 cm@xmath17 k km / s for the conversion factor @xmath119 .
there is considerable controversy surrounding the exact value of @xmath119 ( estimates vary from about 1.5 - 6 @xmath120 cm@xmath17 k km / s for quiescent , giant spiral disks ) but @xmath121 seems appropriate for the general ism in the milky way ( maloney 1990 ) .
ultimately , we wish to compare the gas - to - dust ratio in ngc 891 with that prevailing in the galaxy so , in lieu of any direct measurement of ngc 891 , we adopt @xmath122 cm@xmath17 k km / s . at this stage
, we refrain from allowing @xmath119 to vary with radius ( as might be the case for a steep metallicity gradient between the centre and edge of the ngc 891 disk ) .
maloney ( 1990 ) provides a cogent argument as to why corrections to @xmath119 may only be necessary for environments with a metallicity _ below _ solar .
virtually nothing is known about the metallicity in ngc 891 ( its edge - on orientation precludes optical spectroscopy ) but , so long as ngc 891 constitutes a typical giant spiral galaxy , the heavy element abundance is unlikely to be lower than solar for those parts of the disk where scoville et al detect co emission ( garnet 1998 ) .
we check our assumptions about the metallicity in ngc 891 in the next section ( [ metallicity ] ) .
+ in fig .
[ fig5 ] , we plot the gas - to - dust ratio along the major axis of ngc 891 . for comparison
, we also show the corresponding ratio if either of the neutral gas components ( hi or h@xmath8 ) is excluded from the calculation .
we can see that , at most radii , the molecular hydrogen makes the dominant contribution to the gas mass .
futhermore , the distribution of h@xmath8 in ngc 891 seems to mimic that of the grain material very closely ( see fig .
[ fig6 ] ) .
there are significant fluctuations in the gas - to - dust ratio brought about by an imperfect matching between the submm and co major axis profiles .
secondary maxima present in both the h@xmath8 and dust profiles , possibly attributable to spiral arms seen edge - on , are somewhat displaced with respect to each other .
` misalignments ' between co emission and dust extinction lanes in more face - on spirals such as m83 and m100 have already been noted by other observers , although it is still unclear whether this is a more subtle effect of density wave propogation or simply co acting as a poor tracer of h@xmath8 ( rand & kulkarni 1990 ; rand 1995 ; rand et al 1999 ) . in the case of ngc 891 ,
[ fig6 ] seems to indicate that the overall grain distribution is also more radially extended than the layout of molecular gas .
indeed , it is possible that the submm traces the distribution of _ atomic _ gas for radii @xmath123 .
submm observations closer to the edge of the optical disk would be necessary to confirm whether this is a continuing trend since our @xmath0 m data extend to only @xmath124 whilst 21 cm emission is detected out to @xmath125 ( @xmath126 ) .
the gas - to - dust ratio at the ` edge ' ( @xmath127 ) of spiral disks remains something of a mystery .
however , background galaxies viewed through such outlying regions appear to be systematical reddened by atomic gas possessing a relatively high gas - to - dust ratio ( @xmath128 ; lequeux and guelin 1996 ) .
likewise , @xmath129 m imaging of nearby spirals , using iso , are also indicative of quite large dust scale - lengths , suggesting an association between cold dust and atomic hydrogen at large radii ( alton et al 1998 ) .
our present observations indicate a gas - to - dust ratio @xmath130 towards the optical edge of spiral disks .
+ the prominent association between co emission and submm flux density within ngc 891 ( and its implied correspondance between h@xmath8 and galactic dust ) , is echoed by recent results from the scuba nearby galaxy survey .
dunne et al ( 1999 ) have imaged @xmath131 objects of typical optical size @xmath132 known to possess strong @xmath51 m flux densities ( @xmath1 5.24 jy ) .
notably , they establish a strong correlation between global @xmath0 m and co emission . whilst the same trend
is observed in both ngc 891 and the nearby , face - on spiral ngc 6946 ( bianchi et al 2000 ) , it is possible that an even stronger correlation with hi+h@xmath8 will emerge as submm observations become more sensitive .
+ the error - weighted gas - to - dust ratio in ngc 891 is 260 ( fig .
[ fig5 ] ) .
the corresponding value for the solar neighborhood is 150 , where the dust mass is estimated from either stellar reddening or precipitation of grain elements from the gas phase ( spitzer 1978 ; whittet 1992 ) .
sodroski et al ( 1994 ) analyzed the large - scale 140 and @xmath133 m from the galaxy , as detected by cobe , and thereby established a mean gas - to - dust ratio of 160 in the diffuse ism .
very recently , a gas - to - dust ratio of 200 - 500 has been determined in several nearby galaxies using either radiation transfer modelling ( xilouris et al 1999 ; block et al 1994 ) or @xmath129 m iso imaging observations ( alton et al 1998b ) .
our determination for ngc 891 is consistent with all these values given that sizeable uncertainties exist in the derived dust masses ( via the fir / submm emissivity ) and the quantity of molecular gas ( via the @xmath119 parameter ) .
+ the total dust mass in ngc 891 , for that part of the disk covered by our scuba maps ( @xmath134 ) , is @xmath135 m@xmath136 .
this is towards the low end of the range given in abr ( @xmath137 m@xmath136 ) .
in this section , we attempt to establish the heavy element abundance along the disk of ngc 891 . a determination of metallicity may cause us to modify the value of @xmath119 , chosen in the previous section , to convert co line emission to h@xmath8 column density .
since the gas - to - dust ratio , at least in the inner half of the disk , is dominated by the molecular rather than atomic gas , some justification for our chosen value of @xmath119 would seem in order . due to the obscuring effects of the ngc 891 dust lane , we resort to fir cooling lines from highly ionized metal species ( e.g. oiii @xmath138 m ) which are expected to be optically thin even for a galaxy viewed edge - on .
accordingly , we have searched the iso archive for long wavelength spectrometer ( lws ) and short wavelength spectrometer ( sws ) observations of ngc 891 and , although scientifically - validated spectra of several hii regions were procured , we could not relate the line strengths nor their ratios to any of the diagnostic models currently available in the literature ( statinska 1990 ; spinoglio & malkan 1992 ) .
+ we then explored more indirect means of inferring the metallicity in ngc 891 .
garnet ( 1998 ) has shown that there is a clear trend , spanning 4 magnitudes in b - band brightness , between relative oxygen abundance o / h and absolute galaxy luminosity @xmath139 ( see also garnett & shields 1987 ) .
although extinction effects are severe in ngc 891 ( due to its edge - on orientation ) , the radiative transfer model of xad may be used to derive the intrinsic blue luminosity of the galaxy .
the object can then be located on the o / h vs. @xmath139 plot of garnet . following garnet
, we adopt @xmath140=50 km / s(mpc)@xmath18 in order to determine the distance modulus and hence absolute b magnitude of ngc 891 .
thus , we derive a value of -21.0 for @xmath139 . on the garnet plot this corresponds to an oxygen abundance of 12+log(@xmath141)[email protected] at 1 b - band scale - length from the nucleus ( where 0.2 constitutes the dispersion in the metallicity - magnitude ` relation ' ) .
the milky way has an o / h abundance of 9.1 at 1 b - band scale - length from the centre ( shaver et al 1983 ) , suggesting a very similar metallicity between ngc 891 and our own galaxy . + as mentioned in the previous section , we might expect to modify our value chosen for the conversion factor @xmath119 if the metallicity in ngc 891 falls below solar ( o / h=8.9 ) . for an average metallicity gradient of 0.08 dex / kpc in spiral disks ( vila - costas & edmunds 1992 ) , we expect the oxygen abundance to be [email protected] at the maximum radius at which we convert co line emission to h@xmath8 column density ( r=9.2 kpc ) .
this suggests that the vast majority of molecular gas in ngc 891 has a heavy element abundance similar to , or indeed above , the solar level .
thus the value of @xmath122 cm@xmath17 k
km / s , generally adopted for the general ism inside the solar ring ( [ gdratio ] ) , can be used with some justification for ngc 891 . in fig .
[ fig6b ] , we reproduce the plot of issa et al ( 1990 ) showing the trend of increasing metallicity with higher dust - to - gas ratio for galaxies close to or in the local group .
we incorporate the corresponding quantities derived for ngc 891 which , although subject to appreciable errors , show pleasing agreement with the correlation .
part of the motivation for studying the dust content of external galaxies is to establish whether spiral disks are optically thick or essentially transparent to optical radiation .
there exists a whole history of debate concerning the opacity of spiral galaxies .
indeed , various investigative techniques have produced quite disparate answers in the past ( davies & burstein 1995 ; valentijn 1990 ; disney et al 1989 ) .
more recent studies seem to be in favour of a v - band optical depth @xmath143 through the centre of face - on disks ( xilouris et al 1999 ; kuchinski et al 1998 ) , although some authors still maintain spirals are optically thick right out to their @xmath112 ` edge ' ( valentijn 1990 ) .
one reason why disk opacity is considered such an important topic is that we view the distant universe through , what amounts to , a foreground ` screen ' of spiral galaxies .
this has the effect of both attenuating and reddening light emitted by high redshift systems .
+ alton et al ( 1999b ) have begun to assess the impact of nearby dusty disks on observations of high redshift ( @xmath144 ) galaxies . for one of the disks which they had observed with iso ( ngc 6946 )
, they adopted a bi - model grain distribution which followed the hi and h@xmath8 gas components according to a gas - to - dust ratio of 150 .
this gave an extended dust disk , following the atomic gas , with an exponential fall - off of 14 kpc .
combined with this was a compact grain component , tracing the h@xmath8 , with a radial exponential scale - length of 4.3 kpc .
a simulation was run with a universe populated with disks of this parameterization and assuming a density for field spirals of @xmath145 per mpc@xmath146 ( loveday et al 1992 ) .
although this constituted a rather simple model ( no account was taken of either galaxy clustering or a possible change in gas - to - dust ratio with look - back time ) it produced some instructive results .
notably , for an einstein - de sitter cosmology and a value of 75 kms@xmath18mpc@xmath18 for the hubble constant , alton et al ( 1999b ) predicted significant attenuation of light received in the b - band from high redshifts .
indeed , they concluded that 30 - 40% of the light originating from @xmath147 would fail to reach the present - day observer due to intervening galactic disks .
this attenuation arises principally from the extended dust component which is assumed to follow the atomic hydrogen .
the scale - length and , indeed , total extent of this component is far more critical than the central optical depth associated with h@xmath8 because the peripheral regions of spirals occupy a relatively large filling factor in the plane of the sky . in the absence of information suggesting a radial cut - off for the dust disk ,
none was applied for this simulation .
+ ngc 6946 is a very gas - rich spiral galaxy ( tacconi & young 1990 ) and , as such , might produce somewhat misleading results when determining high - z visibility ( if we simply apply a constant gas - to - dust ratio ) .
therefore , we wish to repeat the calculations of alton et al ( 1999b ) using the disk of ngc 891 as a model .
we recognize that we will be basing our estimate on a single galaxy but it will be many years before the dust distribution has been accurately mapped out in a large number of nearby galaxies . thus provisionally , we use the information contained in fig .
[ fig5 ] regarding the gas - to - dust ratio along the major axis of ngc 891 .
one component of dust is given a radial exponential scale - length of 3.9 kpc , so that it follows the distribution of h@xmath8 with a gas - to - dust ratio of 260 .
we also prescribe a second grain component , with an exponential scale - length of 150 kpc .
this traces the hi plateau with a gas - to - dust ratio of 350 .
the maximum radial extent of both components is set at 14 kpc , the point at which the hi column density appears to drop off sharply in the 21 cm image of rupen et al ( 1991 ) .
our calculation only considers field spirals ( using the density given by loveday et al 1992 ) , but allows for the fact that light received in the b - band , by the observer , has been affected by the uv - fuv part of the extinction curve at higher redshifts . for each slice in redshift ,
the real density of spiral galaxies is calculated and , accordingly , their contribution to the optical depth at that redshift is determined .
finally , we compute a covering factor , @xmath148 , which is defined as the fraction of light emitted at redshift @xmath144 which fails to reach a b - band observer due to the sum of intervening disks . this quantity will naturally vary according to our chosen cosmology but we show in fig .
[ fig7 ] our estimates for the important values of @xmath149 and @xmath150 ( h@xmath151=75 kms@xmath18mpc@xmath18 ) .
there is also a small dependance of the solution on the form chosen for the uv reddening law ( e.g. galactic or lmc - type ) .
however , this is generally neglible for @xmath152 . for more detail concerning our computation of @xmath148
we refer to trewhella ( 1997 ) and heisler & ostriker ( 1988 ) which adopt a similar approach to the current method . + it is apparent from fig . [ fig7 ]
that , in general , foreground attenuation is not a significant problem for galaxies less than a redshift of 4 .
indeed , for galaxies detected in the hubble deep field ( assuming typically @xmath144=1 - 2 here ) , less than 5% of the light appears to be lost . for quasars and related objects , at @xmath153 , small corrections are required in order to learn the true luminosity and colour from optical observations .
these predictions rely heavily on the cut - off radius of the extended dust component ( assumed to follow the hi in the present calculation ) . when using the ngc 891 scuba data , we have extrapolated to @xmath154 , slightly beyond the furthest radius at which we detect @xmath0 m emission ( @xmath155 ) .
submm observations at the optical edge of spiral galaxies are required to ascertain the true gas - to - dust ratio and opacity within the outer hi envelope .
in contrast , we do not expect the obscuration calculations to be affected severely by our neglect of dust clumping .
gmcs have an optical depth of @xmath156 whilst the diffuse ism is characterized by an optical depth close to unity .
thus , for equal amounts of gas in the molecular and atomic phase , dust clumped into gmcs will occupy a filling factor of only 10% . as a point of reference , our opacity model for ngc 891 , if it were not truncated at @xmath154 , would give a face - on optical depth of @xmath157 at the @xmath112 ( @xmath158 ) .
+ the calculations carried out above take no account of spirals belonging to clusters . at the same time
, it is conceivable that a line - of - sight intercepting a foreground cluster may well be subject to significant attenuation .
a proper examination of such effects would require inclusion of the correlation function of galaxies and a description of large - scale structure in our computation . to gauge the size of such effects we have calculated the opacity of various lines of sight through the nearby virgo cluster .
sandage et al ( 1985 ) classify 180 cluster spirals of type sa @xmath159 sd within a radius of @xmath160 ( 2 mpc ) of the virgo centre .
clearly , there will be a whole range of optical depths through the cluster according to how the light path intercepts the cluster members .
nevertheless , we can evaluate filling factors for regions where @xmath161 , @xmath162 etc . to do this
, we use once again the optical depth inferred from our submm observations of ngc 891 i.e. @xmath163 for radius @xmath16414 kpc and @xmath42=0 for @xmath16514 kpc . for @xmath166 ,
the effective cross - section of 180 spiral disks , randomly orientated in the cluster , is 180 @xmath167 ( @xmath168 3.5kpc@xmath57 ) or 7000 kpc@xmath57 .
the total surface area of the virgo cluster is 4.4@xmath167@xmath169 kpc@xmath57 , implying that the probability of a line - of - sight possessing an optical depth greater than 1 , p(@xmath1701.0 ) , is 0.2% . at the lower end , p(@xmath1700.2 )
is limited by the radial cut - off for the dust ( 14 kpc in eq .
[ eq11b ] ) .
we derive a probability of only 2% for a line - of - sight with @xmath1700.2 .
this appears to be incompatible with observations of systematic reddening of background galaxies and qsos by the intergalactic medium ( igm ) in nearby clusters .
such studies imply @xmath171 for most line of sights ( boyle et al 1988 ; romani & maoz 1992 ) .
the discrepancy here between calculation and observation would be virtually removed if the radial grain cut - off were relaxed or if significant quantities of dust were believed to be expelled from spiral disks into the igm during the lifetime of the cluster ( alton et al 1999c ; davies et al 1998 ) .
the first of these issues could be addressed by submm observations of cold dust near the optical edge of spiral disks where the hi column density is low ( @xmath172 cm@xmath17 ) .
we have compared the @xmath0 m emission detected from the nearby , edge - on spiral ngc 891 ( alton et al 1998 ) with the corresponding optical depth predicted from a sophisticated radiation transfer model ( xilouris et al 1998 ) .
morphologically , both model and submm emission show a very similar distribution along the minor axis , with most of the grain mass contained within a 600pc layer ( fwhm ) .
profiles along the major axis show fair agreement between optical extinction and submm radiation .
the @xmath0 m emission is shown to trace the bulk of interstellar dust within ngc 891 . therefore ,
assuming both extinction and submm emission are attributable to the same grains , we can derive the emissivity ( or , equivalently , mass absorption coefficient ) at @xmath0 m and , thence , the distribution of dust mass along the disk .
we believe that this technique is more robust than taking estimates of fir emissivity for galactic grains ( hildebrand 1983 ; draine and lee 1984 ) , and extrapolating them to submm wavelengths .
the @xmath0 m emissivity we derive for ngc 891 is @xmath173 , which is about a factor of 2 - 3 higher than the draine and lee ( 1984 ) model of galactic dust .
uncertainty in our technique means that the emissivity in ngc 891 could be a factor of 2 higher or lower than the quoted value .
importantly , given that the employed radiation transfer model neglects dust - clumping , a _ decrease _ of factor 2 seems quite likely .
+ we have computed the gas - to - dust ratio along the major axis of ngc 891 using our newly - acquired submm emissivity .
we derive a value of 260 , which is consistent , within the uncertainties , with estimates for the diffuse ism of the milky way and nearby spirals .
the dust in ngc 891 appears to be closely associated with the molecular gas phase , at least for the inner half of the disk . at radii @xmath19 kpc ( @xmath174 ) ,
it is possible the grains begin to follow the distribution of atomic hydrogen .
we use the opacity of the ngc 891 disk to calculate the degree with which light emitted at high redshift is attenuated by foreground field spirals . for galaxies typically associated with the hubble deep field ( z=1 - 2 ) ,
the fraction of light lost , in this way , is only 5% .
since this estimate is based on a single galaxy as a model , further submm measurements will be required to support our conclusions .
in particular , observations at the periphery of spiral disks could establish the extent to which cold dust is associated with the expansive hi envelopes surrounding spiral galaxies ( it is the maximum radial extent of the dust which primarily controls the level of foreground extinction ) .
futhermore , we point out that certain lines - of - sight through nearby clusters are likely to be much more optically thick ( than our figure of 5% would suggest ) , particularly if grains are expelled into the intergalactic medium via starburst winds ( alton et al 1999c ) or dispersed through tidal interactions ( yun et al 1994 ) .
+ finally , we note the strong correspondance between dust and molecular gas in ngc 891 .
further submm imaging of face - on spirals ( both h@xmath8 rich and h@xmath8 poor ) would confirm the trend indicated by the current observations , that the bulk of galactic dust appears to be associated with the molecular gas phase .
we would like to thank loretta dunne and steve eales for revealing early results from their scuba nearby galaxy survey .
it is also a pleasure to thank adolf witt for his comments on dust clumping in spiral disks .
we acknowledge the assistance of sunil sidher in the reduction of iso lws spectra . as
always , pba acknowledges pparc for its continued financial support .
agladze , n. , sievers , a. , jones , s. , burlitch , j. , beckwith , s. , 1994 , nature , 372 , 243 parameter & value + + r.a . &
2h 19 24.3 + dec . &
42@xmath175 07 17 + @xmath176 ( kpc ) & 0.39 + @xmath177 ( kpc ) & 0.26 + @xmath178 ( kpc ) & 5.71 + @xmath179 ( kpc ) & 8.10 + @xmath180 & 0.70 + @xmath181 & 89.8 + + | we compare @xmath0 m scuba images of ngc 891 with the corresponding v - band optical depth predicted from radiation transfer simulations .
these two tracers of dust show a very similar distribution along the minor axis and a reasonable agreement along the major axis . assuming that the grains responsible for optical extinction are also the source of @xmath0 m emission we derive a submillimeter emissivity ( emission efficiency ) for dust in the ngc 891 disk .
this quantity is found to be a factor of 2 - 3 higher than the generally - accepted ( but highly uncertain ) values adopted for the milky way .
it should be stated , however , that if a substantial fraction of dust in ngc 891 is clumped , the emissivity in the two galaxies may be quite similar .
we use our newly - acquired emissivity to convert our @xmath0 m images into detailed maps of dust mass and , utilizing 21 cm and co - emission data for ngc 891 , derive the gas - to - dust ratio along the disk .
we compute an average ratio of 260 a value consistent with the milky way and external spirals within the uncertainties in deriving both the dust mass and the quantity of molecular gas .
the bulk of dust in ngc 891 appears to be closely associated with the molecular gas phase although it may start to follow the distribution of atomic hydrogen at radii @xmath19 kpc ( i.e. @xmath2 ) . using the optical depth of the ngc 891 disk , we quantify how light emitted at high redshift is attenuated by dust residing in foreground spirals . for b - band observations of galaxies typically found in the hubble deep field , the amount of light lost is expected to be small ( @xmath3 5% )
this value depends critically on the maximum radial extent of cold dust in spiral disks ( which is poorly known ) .
it may also represent a lower limit if galaxies expel dust over time into the intergalactic medium . |
starting from the early nineties , boris dubrovin and many of his collaborators have studied the relation between gromov - witten theory and the theory of integrable systems of pdes , which was first noticed by witten in @xcite .
the axioms of frobenius manifold , cf .
@xcite , encode all the properties that are satisfied by the algebraic structure generated by rational gromov - witten theory and that , in particular , generate an integrable hamiltonian system of evolutionary pdes .
the structure of frobenius manifold has proven to be central in many different areas of mathematics , from algebra to singularity theory , and provides for instance the most immediate approach to mirror symmetry .
+ one of the consequences of the axioms of a ( homogenous ) frobenius manifold is that the associated hamiltonian system of pdes is actually bihamiltonian .
bihamiltonian structures where introduced by magri in @xcite in the analysis of the so - called lenard scheme ( see e.g @xcite ) to construct the kdv integrals .
they consist of a manifold endowed with two poisson tensors @xmath1 and @xmath2 , mutually compatible in the sense that their schouten - nijenhuis bracket @xmath3 $ ] vanishes . under the condition that the poisson pencil @xmath4 ( a one - parametric family of poisson tensors ) has constant co - rank there exists a simple ricursive procedure for constructing a sequence of commuting integrals for both poisson structures ( see also @xcite and the author s survey @xcite ) .
let p be a manifold endowed with compatible poisson tensors @xmath1 , @xmath2 and associated poisson brackets @xmath5 , @xmath6 .
let @xmath7 for arbitrary sufficiently small @xmath8 .
then the coefficients of the taylor expansion @xmath9 of the casimirs @xmath10 , @xmath11 of poisson tensor @xmath12 commute with respect to both poisson brackets , @xmath13 moreover @xmath14 , @xmath15 .
+ while in the literature this procedure is often called bihamiltonian recursion , we will use the term _ bihamiltonian reconstruction _
, to avoid confusion with the property of a sequence of symmetries @xmath16 of being in _ bihamiltonian recursion _ if @xmath17 for some constant coefficients @xmath18 , @xmath19 , @xmath20 .
+ in the context of gromov - witten theory and the associated frobenius manifolds such reconstruction can be used to recover at least part of the symmetries for the hamiltonian system .
actually , with great generality , given a homogeneous frobenius manifold , finding a fundamental solution for the so called deformed flat connection ( a special case of topological recursion relations for rational one - descendant gw invariants , see @xcite ) is a more powerful way to reconstruct the algebra of symmetries and , moreover , with this method the solution is automatically normalized to match the generating series of rational one - descendant gw invariants ( often called @xmath21-function in the gromov - witten literature ) . besides
, the symmetries found this way are in bihamiltonian recursion anyway .
+ in contrast , bihamiltonian reconstruction is not , in general , directly related with enumerative geometry , which results in discrepancies with the @xmath21-function .
even worse , even when the hypothesis of the above theorem are satisfied , it might happen that the two poisson tensors have some casimir in common .
this means that , when starting with such common casimirs as @xmath22 , in the above taylor expansion , all of the other coefficients trivially vanish .
this is precisely what happens in the case of the gromov - witten theory of the projective line @xmath23 , where bihamiltonian recursion is only capable of recovering the symmetries generated by descendants of the khler class , but not those of the unity class . this pathology of the poisson pencil is called _ resonance_. + this paper deals with symplectic field theory ( sft ) and a recursion procedure for descendants that has much in common with bihamiltonian recursion in gromov - witten theory . introduced by h. hofer , a. givental and y. eliashberg in 2000 @xcite , sft is a very large project and can be viewed as a topological quantum field theory approach to gromov - witten theory . besides providing a unified view on
established pseudoholomorphic curve theories like symplectic floer homology , contact homology and gromov - witten theory , it sheds considerable light on the appearence of infinite dimensional hamiltonian systems in the theory of holomorphic curves ( see @xcite for an review on this topic which includes sft ) .
+ indeed , symplectic field theory leads to algebraic invariants with very rich algebraic structures and in particular , as it was pointed out by eliashberg in his icm 2006 plenary talk ( @xcite ) , the integrable systems of rational gromov - witten theory very naturally appear in rational symplectic field theory by using the link between the rational symplectic field theory of prequantization spaces in the morse - bott version and the rational gromov - witten potential of the underlying symplectic manifold ( see the recent papers @xcite , @xcite ) . after introducing gravitational descendants ( see @xcite ) along the lines of gromov - witten theory , it is precisely the natural algebraic structure of sft that provides a natural link between holomorphic curves and ( quantum ) integrable systems .
+ in this paper we explore the potentiality of an intrinsic difference between gromov - witten and symplectic field theory : the moduli spaces of holomorphic maps studied by the latter carry special evaluation maps controlling the relative gluing angle of different components of a multi - floor configuration ( the sft generalization of nodal curves , see @xcite ) .
these can be used to define new correlators that were not present in the original theory , but give interesting recursive formulas for one - point descendants ( and probably beyond ) , which are similar but not equivalent to bihamiltonian reconstruction , topological recursion and , ultimately , to the integrability properties of the sft infinite dimensional hamiltonian system .
+ in many senses these extra correlators which control the gluing angles of different components of curves in the boundary strata are a natural generalization of the non-@xmath0-equivariant differential for the non - equivariant linearized contact homology of bourgeois - oancea , @xcite ( see also @xcite ) , to non - linearized contact homology and full rational sft . with this in mind
we conclude that the non - equivariant differential plays a role similar to the second poisson structure of gromov - witten theory with respect to gravitational descendants .
more precisely we show how our generalized non - equivariant differential ( the potential encoding such new correlators ) , denoted by by @xmath24 , satisfies ( in contact homology ) the axioms of a nijenhuis operator .
+ recall from magri and fuchsteiner , @xcite@xcite , that on a manifold @xmath25 , a nijenhuis ( or hereditary ) operator @xmath26 ( where @xmath27 denotes the space of @xmath28-tensor fields on @xmath25 ) is one whose nijenhuis torsion vanishes : @xmath29 this condition ensures that , given a sequence of commuting vector fields @xmath30=0,\qquad \alpha,\beta=1,\ldots , n\ ] ] that are also symmetries of the operator @xmath24 , i.e. @xmath31 , we can enlarge the commuting system by recursively applying the operator @xmath24 : @xmath32=0,\qquad k , j\in\mathbb{n}.\ ] ] given a poisson tensor @xmath33 on @xmath25 , in @xcite magri and morosi further studied the compatibility conditions of a nijenhuis operator @xmath24 with the poisson structure .
@xmath24 is said to be compatible with @xmath34 if the following two conditions hold : @xmath35 @xmath36 when these equations are satisfied , the pair @xmath37 is called a poisson - nijenhuis structure on @xmath25 . the main property ,
then , is that one can define a sequence of @xmath38-tensors @xmath39 , @xmath40 which are poisson and are pairwise compatible in the sense that they pairwise form poisson pencils . in particular we get the bihamiltonian structure @xmath41 .
+ the theory of poisson - nijenhuis structure is very well developed and its relation with integrability is deep . in particular the study of the spectrum of
@xmath24 plays a fundamental role , in that the eigenvalues of @xmath24 form a system of commuting symmetries in bihamiltonian recursion ( see for instance @xcite,@xcite,@xcite ) .
+ the first part of the paper deals with the contact homology case . in contact homology , thanks to the total absence of non - constant nodal configurations ( due to the maximum principle for holomorphic curves in symplectizations ) , we prove that the knowledge of @xmath24 and of the primary theory ( contact homology differential @xmath42 , with no descendants ) is sufficient to completely reconstruct the descendant vector fields @xmath43 as differential operators in the variables associated to reeb orbits on contact homology . + in the case of full rational sft ( of any target stable hamiltonian structure ) ,
our bihamiltonian recursion is slightly less effective because of the presence of non - constant nodal curves .
formally the result is similar , i.e. the descendant hamiltonians @xmath44 satisfy recursion relations which are completely analogus to bihamiltonian recursion for a pair of bivectors @xmath45 , where @xmath34 is the natural poisson structure on the sft homology algebra and @xmath46 is a _ graded symmetric _ even bivector which is the sft - generalization of @xmath24 .
+ in any case this looks like a fundamental step in understanding the relation between completeness of the contact homology vector field system , or the sft hamiltonian system , and the underlying symplectic topology of the target manifold .
indeed the full information is contained in the non - equivariant correlators forming @xmath24 and @xmath46 , and we plan to study the consequences in a subsequent publication
. + * aknowledgements .
* + part of the work was conducted while i was a postdoc of fondation de sciences mathematiques de paris at the institut de mathematiques de jussieu , paris vi and part during my postdoc at the institute for mathematics of the university of zurich ( partially supported by snf grant no .
pdamp2_137151/1 ) .
i would like to thank y. eliashberg , o. fabert and d. zvonkine for useful discussions .
the final part of the work was actually completed during my visit at stanford university at the beginning of 2012 .
i am grateful to yakov eliashberg for the invitation .
finally i would like to thank the american institute of mathematics for allowing me to organize the workshop on _ integrable systems in gromov - witten and symplectic field theory _ , in january 2012 , during which i could speak to boris dubrovin , who then noticed how my equations for @xmath24 where equivalent to vanishing of nijenhuis torsion and who pointed me to the correct reference ( magri , morosi and fuchsteiner ) to see the big picture . +
symplectic field theory ( sft ) , introduced by y. eliashberg , a. givental and h. hofer in @xcite , consists in a unified and comprehensive approach to the theory of holomorphic curves in symplectic and contact topology . in the spirit of a topological field theory ,
it assigns algebraic invariants to closed manifolds with a stable hamiltonian structure .
we recall here the main ideas from @xcite,@xcite,@xcite .
+ a hamiltonian structure ( see @xcite ) on a closed @xmath47-dimensional manifold @xmath48 is a closed two - form @xmath49 on @xmath48 of maximal rank @xmath50 .
this means that @xmath51 is a one - dimensional distribution .
a hamiltonian structure is called stable if there exists a one - form @xmath52 and a vector field @xmath53 ( called reeb vector field ) on @xmath48 such that @xmath53 generates @xmath54 , @xmath55 and @xmath56 .
notice that , this way , @xmath57 is a symplectic hyperplane distribution .
also , notice that if @xmath53 exists , it is completely determined by @xmath49 and @xmath52 .
+ any contact form @xmath52 on @xmath48 provides a stable hamiltonian structure @xmath58 on @xmath48 where the symplectic hyperplane distribution coincides with the contact structure .
+ given a principal circle bundle @xmath59 over a closed symplectic manifold @xmath60 and any connection @xmath61-form @xmath52 , @xmath62 is a stable hamiltonian structure on @xmath48 .
given a closed symplectic manifold @xmath60 and a symplectomorphism @xmath63 , consider the symplectic mapping torus @xmath64 .
the natural splitting @xmath65 allows to define the lift @xmath49 of @xmath66 to @xmath48 .
then @xmath67 , where @xmath68 is the natural @xmath0-coordinate , is a stable hamiltonian structure ( with integrable symplectic distribution @xmath69 ) .
given a stable hamiltonian structure @xmath70 , consider a complex structure @xmath71 on the hyperplane ditribution @xmath57 which is @xmath72-compatible ( i.e. @xmath73 is a metric on @xmath74 ) .
such complex structures form a non - empty , contractible set .
we extend @xmath71 uniquely from @xmath74 to an almost complex structure @xmath75 on the cylinder @xmath76 by requiring that @xmath75 is @xmath77-invariant and @xmath78 , @xmath79 being the @xmath77-direction .
+ symplectic field theory assigns algebraic invariants to closed manifolds @xmath48 with a stable hamiltonian structure .
the invariants are defined by counting @xmath75-holomorphic curves in @xmath76 with finite energy .
+ let us recall the definition of moduli spaces of holomorphic curves studied in rational sft .
let @xmath80 be two ordered sets of closed orbits @xmath81 of the reeb vector field @xmath53 on @xmath48 , i.e. , @xmath82 , @xmath83 , @xmath84 , where @xmath85 denotes the period of @xmath81 . here
we assume that the stable hamiltonian structure is morse in the sense that all closed orbits of the reeb vector field are nondegenerate in the sense of @xcite ; in particular , the set of closed reeb orbits is discrete . given a closed reeb orbit @xmath81 of any multiplicity
, we will denote by @xmath86 its underlying simple reeb orbit .
at each simple reeb orbit we will fix a closed form @xmath87 generating @xmath88 .
+ then the ( parametrized ) moduli space @xmath89 consists of tuples @xmath90 , where @xmath91 are three disjoint ordered sets of points on @xmath92 , which are called positive and negative punctures , and additional marked points , respectively . we will also fix an asymptotic marker at each puncture , i.e. a ray originating at the puncture .
the map @xmath93 starting from the punctured riemann surface @xmath94 is required to satisfy the cauchy - riemann equation @xmath95 with the complex structure @xmath96 on @xmath92 .
assuming we have chosen cylindrical coordinates @xmath97 around each puncture @xmath98 in the sense that @xmath99 , the map @xmath100 is additionally required to show for all @xmath101 the asymptotic behaviour @xmath102 with some @xmath103 and the orbits @xmath104 , where @xmath105 denotes period of @xmath106 . in order to assign an absolute homology class @xmath107 to a holomorphic curve @xmath108 we have to employ spanning surfaces @xmath109 connecting a given closed reeb orbit @xmath81 in @xmath48 to a linear combination of circles @xmath110 representing a basis of @xmath111 , @xmath112 in order to define @xmath113 + [ u({\dot{s } } ) ] - [ u_{\gamma^-}],\ ] ] where @xmath114 = \sum_{n=1}^{s^{\pm } } [ u_{\gamma^{\pm}_n}]$ ] viewed as singular chains .
+ observe that the group @xmath115 of moebius transformations acts on elements in @xmath116 in an obvious way , @xmath117 and we obtain the moduli space @xmath118 studied in symplectic field theory by dividing out this action and the natural @xmath77-action on the target manifold @xmath119 .
furthermore it was shown in @xcite that this moduli space can be compactified to a moduli space @xmath120 by adding moduli spaces of multi - floor curves with nodes .
in particular , the moduli space has codimension - one boundary given by ( fibre ) products @xmath121 of lower - dimensional moduli spaces .
+ let us now briefly introduce the algebraic formalism of rational sft as described in @xcite : + let us fix a trivialization of the symplectic bundle @xmath122 over each curve @xmath123 .
this induces a trivialization a hoomotopically unique trivialization of the same bundle over each periodic reeb orbit @xmath81 via the spanning surface @xmath124 .
let us use this trivialization to define the conley - zehnder index of the reeb orbit ( the maslov index of the path in @xmath125 given by the linearized reeb flow along @xmath81 ) . recall that a multiply - covered reeb orbit @xmath126 is called bad if @xmath127 , where @xmath128 denotes the conley - zehnder index of @xmath81 . calling a reeb orbit @xmath81 _ good _
if it is not bad , denote by @xmath129 the space of good reeb orbits .
we assign to every good reeb orbit @xmath81 two formal graded variables @xmath130 with grading @xmath131 when @xmath132 .
+ assuming we have chosen a basis @xmath133 of @xmath134 , we assign to every @xmath135 a formal variable @xmath136 with grading @xmath137 . in order to include higher - dimensional moduli spaces we further assume that a string of closed ( homogeneous ) differential forms @xmath138 on @xmath48 is chosen and assign to every @xmath139 a formal variables @xmath140 with grading @xmath141 with this let @xmath142 be the poisson algebra of formal power series in the variables @xmath143 and @xmath144 with coefficients which are polynomials in the variables @xmath145 and laurent series in the novikov ring variables @xmath146 , with poisson bracket given by @xmath147 where @xmath148 is the multiplicity of the orbit @xmath81 .
+ consider the union @xmath149 of all moduli spaces @xmath150 for @xmath151 and @xmath152 .
as in gromov - witten theory we want to organize all moduli spaces @xmath153 into a generating function @xmath154 , called _
hamiltonian_. in order to include also higher - dimensional moduli spaces , in @xcite
the authors follow the approach in gromov - witten theory to integrate the chosen differential forms @xmath155 and @xmath156 $ ] ( the canonical basis of @xmath157 ) over the moduli spaces after pulling them back under the evaluation maps @xmath158 , @xmath159 , @xmath160 , @xmath161 at the marked points and punctures to the target manifold @xmath48 and the space of ( positive or negative ) good reeb orbits @xmath129 , respectively .
consider furthermore evaluation maps @xmath162 , @xmath161 defined by the asymptotic markers at each puncture .
let @xmath163 @xmath164\ ] ] @xmath165\ ] ] the hamiltonian @xmath166 is then defined by @xmath167 with @xmath168 for @xmath169 .
we recall the definition of gravitational descendants in symplectic field theory ( see @xcite ) . in complete analogy to gromov - witten theory
we can introduce @xmath170 tautological line bundles @xmath171 over each moduli space @xmath150 , as the pull - back of the relative dualizing sheaf of @xmath172 under the canonical section @xmath173 mapping to the @xmath96-th marked point in the fibre . + as in gromov - witten theory we would like to consider the integration of ( powers of ) the first chern class of the tautological line bundles over the moduli space , which by poincar duality corresponds to counting common zeroes of sections of such bundles .
however , in symplectic field theory the moduli spaces can have codimension - one boundary , so we need to replace integration of the first chern class of the tautological line bundle over a single moduli space with a construction involving all moduli space at once , which preserves the algebraic structure of sft . + following the compactness statement in [ behwz ] , the codimension - one boundary of a moduli space @xmath120 of sft holomorphic curves consists of curves with two levels ( in the sense of [ behwz ] ) .
more precisely , each component of the boundary has the form of a fibred product @xmath174 @xmath175 of moduli spaces ( of strictly lower dimension ) , where the marked points distribute on the two levels .
consider a boundary component where the @xmath96-th marked point sits , say , on the first level @xmath176 : it directly follows from the definition of the tautological line bundle @xmath177 at the @xmath96-th marked point over @xmath178 that , over such boundary component , @xmath179 where @xmath180 denotes the tautological line bundle over the moduli space @xmath176 and @xmath181 is the projection onto the first factor . with this
we can now give the definition of coherent collections of sections in tautological line bundles from @xcite .
assume that we have chosen sections @xmath182 in the tautological line bundles @xmath177 over all moduli spaces @xmath178 of @xmath75-holomorphic curves of sft
. then these collections of sections @xmath183 are called _ coherent _ if for every section @xmath182 of @xmath177 over a moduli space @xmath178 the following holds : over each codimension - one boundary component @xmath184 of @xmath178 , the section @xmath182 agrees with the pull - back @xmath185 ( @xmath186 ) of the chosen section @xmath187 ( @xmath188 ) of the tautological line bundle @xmath180 over @xmath176 ( @xmath189 over @xmath190 ) , assuming that the @xmath96-th marked point sits on the first ( second ) level .
since in the end we will again be interested in the zero sets of these sections , we will assume that all occuring sections are sufficiently generic , in particular , transversal to the zero section .
furthermore , we want to assume that all the chosen sections are indeed invariant under the obvious symmetries like reordering of punctures and marked points . in order to meet both requirements
, it follows that actually need to employ multi - sections ( in the sense of branched manifolds ) . on the other hand ,
it is clear that one can always find coherent collections of ( transversal ) sections @xmath191 by using induction on the dimension of the underlying moduli space .
+ for every tuple @xmath192 of natural numbers we choose @xmath193 coherent collections of sections @xmath194 of @xmath177 .
then we define for every moduli space @xmath120 , @xmath195 note that by choosing all sections sufficiently generic , we can assume @xmath196 is a branched - labelled submanifold of the moduli space @xmath150 . note that by definition @xmath197 and it follows from the coherency condition that the codimension - one boundary of @xmath198 is given by the products @xmath199 or @xmath200 ( depending on the @xmath96-th marked points sitting on the first or second level ) .
+ with this we can define the descendant hamiltonian of sft , which we will denote by @xmath201 , while the hamiltonian @xmath202 defined in @xcite will from now on be called _
primary_. in order to keep track of the descendants we will assign to every chosen differential form @xmath203 now a sequence of formal variables @xmath204 with grading @xmath205 and form an extended poisson algebra @xmath206 accordingly
. then the descendant hamiltonian @xmath207 of ( rational ) sft is defined by @xmath208 where @xmath209 and @xmath168 for @xmath169 .
+ symplectic field theory assigns to every contact manifold not only a poisson algebra , rational sft homology , but also , thanks to gravitational descendants , a hamiltonian system in it with an infinite number of symmetries . differentiating the rational hamiltonian @xmath207 with respect to the formal variables @xmath210 defines a sequence of classical hamiltonians @xmath211 _ in the rationall sft homology algebra with differential @xmath212 , which commute with respect to the bracket on @xmath213 , _
@xmath214 note that everything is an immediate consequence of the master equation @xmath215 , which can be proven in the same way as in the case without descendants using the results in @xcite . while the boundary equation @xmath216
, @xmath217 is well - known to follow directly from the identity @xmath215 , the fact that every @xmath218 , @xmath219 defines an element in the homology @xmath213 follows from the identity @xmath220 which can be shown by differentiating the master equation with respect to the @xmath221-variable and using the graded leibniz rule , @xmath222 on the other hand , in order to see that any two @xmath218 , @xmath223 commute _ after passing to homology _ it suffices to see that by differentiating twice ( and using that all summands in @xmath166 have odd degree ) we get the identity @xmath224 we now turn to the question of independence of these nice algebraic structures from the choices like contact form , cylindrical almost complex structure , representatives for the classes @xmath225\in h^*(v)$ ] and @xmath226\in h^*(s^1)$ ] , abstract polyfold perturbations and , of course , the choice of the coherent collection of sections .
this is the content of the following theorem proven in @xcite . for different choices of contact form @xmath227 , cylindrical almost complex structure @xmath228 ,
representatives for the classes @xmath225\in h^*(v)$ ] and @xmath226\in h^*(s^1)$ ] , abstract polyfold perturbations and sequences of coherent collections of sections @xmath229 the resulting systems of commuting functions @xmath230 on @xmath231 and @xmath232 on @xmath233 are isomorphic , i.e. there exists an isomorphism of the poisson algebras @xmath234 and @xmath233 which maps @xmath235 to @xmath236 .
this theorem is an immediate extension of the theorem in @xcite which states that for different choices of auxiliary data the poisson algebras @xmath237 and @xmath238 with @xmath239 are isomorphic .
in particular the extension in @xcite is about the coherent collection of sections @xmath229 . here
one needs a notion of a collection of sections @xmath240 in the tautological line bundles over all moduli spaces of holomorphic curves in the cylindrical cobordism interpolating between the auxiliary structures , which are _ coherently connecting _ the two coherent collections of sections @xmath229 .
+ we want to point out the fact that the primary poisson sft homology algebra can be thought of as the space of functions on some abstract infinite - dimensional poisson super - space .
indeed , consider the poisson super - space @xmath241 underlying the poisson algebra @xmath142 .
then the kernel @xmath242 can be seen as the algebra of functions on the space @xmath243 of orbits in @xmath241 of the hamiltonian @xmath77-action given by @xmath166 , that is , the flow lines of the hamiltonian vector field @xmath244 associated to @xmath166 . even in a finite dimensional
setting the space @xmath243 can be very wild .
anyhow the image @xmath245 is an ideal of such algebra and hence identifies a sub - space of @xmath243 given by all of those orbits @xmath246 at which , for any @xmath247 , @xmath248 .
but such orbits are simply the constant ones , where @xmath244 vanishes .
hence the poisson sft - homology algebra @xmath249 can be regarded as the algebra of functions on @xmath250 , seen as a subspace of the space @xmath243 of orbits of @xmath166 , endowed with a poisson structure by singular , stationary reduction .
in particular the descendant hamiltonians @xmath251 are examples of functions on such space .
+ finally we recall a result from @xcite that states that , besides commutativity , the sft hamiltonians satisfy analogues of the well - known string , dilaton and divisor equations of gromov - witten theory .
such equations hold , after passing to sft homology , for any auxiliary choice used to define the hamiltonians .
[ sdd ] for any choice of differential forms and coherent sections the following _ string , dilaton and divisor equations _
after _ passing to sft - homology @xmath252 where @xmath253 is the @xmath68-variable associated to the @xmath254-th descendant of the unity class @xmath255 , @xmath256 is the one associated with @xmath257 and @xmath258 the corresponding novikov ring variable , and @xmath259 is the linear differential operator @xmath260 + we end this introductory section with a short discussion of the analytical foundations of sft .
all the algebraic results we prove for holomorphic curves rely on the fact that all appearing moduli spaces are ( weighted branched ) manifolds with corners of dimension equal to the fredholm index of the cauchy - riemann operator . in order to equip the zero set of the cauchy - riemann operator with nice manifold structures , one applies an infinite - dimensional version of the classical implicit function theorem , where the crucial step is to prove a transversality result for the cauchy - riemann operator . while it is well - known that transversality holds for a generic choice of almost complex structure as long as all holomorphic curves are simple ,
several problems appear when the curve is multiply - covered .
since this problem is already present in ( symplectic ) gromov - witten theory and floer homology , very involved tools like virtual moduli cycles , kuranishi structures and polyfolds were developed to solve the transversality problem for holomorphic curves in general .
while the polyfold approach of hofer , wysocki and zehnder , see @xcite , seems to be the approach that solves all the challenges in the most satisfactory way ( see also the survey @xcite ) , it is not yet fully completed . since they promise to prove transversality for symplectic field theory and gromov - witten theory in one of their upcoming papers , we follow other papers in the field in proving everything up to transversality and state it nevertheless as a theorem . since ,
in contrast to relative sft and other papers about sft using morse - bott techniques , we only need to add a transversality result for sections in our finite - dimensional tautological line bundles , our results are indeed rigorous when hofer and his collaborators have completed their work . +
contact homology is a reduced version of symplectic field theory that only uses curves with at most one positive puncture . in case
the target manifold @xmath48 is contact , the maximum principle for holomorphic curves in symplectizations forbids that a sft - holomorphic curve with @xmath261 positive punctures can degenerate to a multi - floor curve where any of the components has more than @xmath261 positive punctures .
the absence of local maxima further implies that any curve with no positive punctures must be constant . in particular
this means that we can study moduli spaces of sft - curves with only one positive puncture and be safe that the boundary only involves moduli spaces of the same type .
besides , in such moduli spaces nodal degeneration only involve the appearance of constant bubbles ( when different marked points come together ) .
+ the algebraic structure we obtain is the linear part in the @xmath262-variables of the one described for full sft . in particular we get a complex formed by the graded commutative algebra @xmath263 generated by the variables @xmath264 over the power series in the variables @xmath265 with coefficients in the novikov ring of variables @xmath266 ( i.e. the evaluation at @xmath267 of the poisson algebra @xmath142 ) , and differential given by the ( odd ) vector field @xmath268 the master equation @xmath269 reduces to @xmath270=0 $ ] ( the square bracket here stands for lie bracket on vector fields ) , the resulting homology will be denoted by @xmath271 ( or simply @xmath272 when there is no danger of confusion ) and the system of commuting ( on sft - homology ) descendant hamiltonians @xmath273 , @xmath274 induce a system of lie - commuting vector fields on contact homology @xmath275 + consider now a moduli spaces of punctured @xmath0-parametrized cylinders with marked points .
we start with the fully parametrized space @xmath276 consisting of tuples @xmath277 , where @xmath278 are two disjoint ordered sets of points on @xmath279 ( namely negative punctures with asymptotic markers , and @xmath170 additional marked points ) .
the map @xmath93 from the punctured riemann surface @xmath280 is required to satisfy the cauchy - riemann equation @xmath95 with the complex structure @xmath96 on @xmath92 .
assuming we have chosen cylindrical coordinates @xmath281 around each puncture @xmath282 , in the sense that @xmath283 , the map @xmath100 is additionally required to show for all @xmath284 the asymptotic behaviour @xmath285 with some @xmath103 and the orbits @xmath286 , where @xmath287 denotes period of @xmath288 , and analogous asymptotic behaviour at @xmath289 and @xmath290 for @xmath291 and @xmath292 respectively , for orbits @xmath293 and @xmath294 and with respect to the natural coordinates on @xmath295 .
we assign to each curve an absolute homology class @xmath107 employing as usual a choice of spanning surfaces . in order to obtain the @xmath0-parametrized space @xmath296 we only divide out the @xmath77-component of the @xmath297 group of automorphisms of the cylinder @xmath298 and , as usual , the @xmath77-action coming from the cylindrical target @xmath299 as well . +
this moduli space can be compactified to @xmath300 by adding moduli spaces of @xmath0-parametrized multi - floor curves with ghost bubbles , where the puncture at @xmath289 is always the positive puncture of the top floor , the puncture at @xmath290 can be on any floor and the @xmath0-parametrization is remembered when going through connecting punctures as explained in @xcite ( compactification of the space of curves with _ decorations _ ) .
more explicitly , in such compactification , a @xmath301-floor curve with the @xmath290-puncture on its @xmath254-th component from the top , has the upper @xmath254 components that are @xmath0-parametrized curves where , at each connecting puncture , the @xmath0-coordinates of different components match , while the lower @xmath302 are non-@xmath0-parametrized .
also , both type of components possibly have stable constant bubbles . + the space @xmath303 carries , besides the usual evaluation maps at marked points , orbits and asymptotic markers , extra evaluation maps at the punctures at @xmath289 and @xmath290 to the corresponding target simple reeb orbits given by the special @xmath0-coordinate on the curve , @xmath304 @xmath305 + we form the space @xmath306 by taking the union over @xmath307 of all the spaces @xmath308 with @xmath152 . + this moduli space was actually already introduced in @xcite to define the non-@xmath0-equivariant linearized contact homology differential
. we will proceed in a similar way defining a @xmath309-tensor @xmath310 depending on parameters @xmath311 , on the super - space @xmath312 underlying the algebra @xmath263 .
a point @xmath313 is a cohomology class @xmath314 $ ] on @xmath129 , with the notations of section @xmath61 .
we will write @xmath315 instead of @xmath264 to be coherent with the notion that @xmath264 is treated here as a coordinate for the space @xmath312 , while we will treat the @xmath68-variables as parameters on which the functions on @xmath312 can depend . using such coordinates
we define @xmath316 where we sum over repeated indices , and @xmath317 with the usual grading of the sft variables , and assigning degree @xmath289 to the exterior differential @xmath318 on the superspace @xmath241 , from the index formula for the dimension of the moduli space of sft - curves , we deduce @xmath319 notice that the no - descendant ( or primary ) contact homology differential @xmath42 ( still parametrized by the primary variables @xmath320 ) induces a differential @xmath321 ( lie derivative along the vector field @xmath42 ) on the space of @xmath28-tensor fields @xmath322 ( again with parameters @xmath320 ) on the super - space @xmath312 .
the resulting homology , which we denote @xmath323 , is a module over @xmath324 and is an invariant of the contact structure on @xmath48 , as it can easily be proved with the same procedure as for @xmath271 . in particular , for two different choices of contact form @xmath227 , cylindrical almost complex structure @xmath228 , representatives for the classes @xmath225\in h^*(v)$ ] and @xmath226\in h^*(s^1)$ ] , abstract polyfold perturbations and sequences of coherent collections of sections @xmath229 , there exist an isomorphism @xmath325 which is simply the lift to the tensor algebra of the isomorphism @xmath326 constructed in @xcite by studying curves in the cobordims @xmath327 interpolating between the two different choices ( see also the discussion on invariance for satellites there ) .
@xmath328 and , denoting by @xmath329 the two @xmath309-tensors resulting from two different choices of contact form @xmath227 , cylindrical almost complex structure @xmath228 , representatives for the classes @xmath225\in h^*(v)$ ] and @xmath226\in h^*(s^1)$ ] , abstract polyfold perturbations and sequences of coherent collections of sections @xmath229 , @xmath330 so that @xmath331 is an invariant of the contact structure on @xmath48 . for the proof of the equation
@xmath332 we need to study the boundary of the moduli spaces involved in the definition of @xmath24 , i.e. moduli spaces of punctured @xmath92 with a marked @xmath333 line connecting two punctures @xmath289 and @xmath290 mapped to orbits @xmath334 and @xmath335 , at which we pull back @xmath61-forms from the underlying simple orbits . in the picture below we represent such moduli by drawing the corresponding generic element ( the curve with the marked red @xmath333 line ) and we rapresent the constraining of the endpoints via @xmath61-forms by the small triangles ( a triangle pointing towards an orbit @xmath81 means that the red line is @xmath0-constrained at that orbit by integrating the pull - back of the form @xmath87 ) . the three kind of curve degenerations that form this ( codimension @xmath61 ) boundary are drawn as the corresponding @xmath336-floor curves with constrained or unconstrained red lines ( here we use the fact that , when a curve splits at a puncture @xmath81 through which the @xmath333 line passes , we need to pull back a representative of the diagonal class in @xmath337 , which we express as @xmath338 ) ) .
+ all we need to notice , at this point , is that , taking orientation into account for the right signs , the three terms on the righ - hand side of the equation represented in the picture exactly correspond to the three summands in the coordinate expression of the lie derivative @xmath339 , which hence vanishes by stokes theorem applied to the boundary of our moduli space . + for the second part of theorem , about invariance with respect to auxiliary choices in the definition of @xmath263 , @xmath42 and @xmath24 a similar approach is needed , where we study the boundary of the moduli spaces of the same type of curves , but this time in the cobordism interpolating between two different choices of auxiliary data .
drawing again the same kind of pictures ( only remembering , as explained in @xcite , that the boundary of moduli spaces of connected curves in the cobordism is formed by @xmath336-floor curves in which one of the floors is a connected curve in the cylindrical manifold over one of the boundaries and the other is a possibly disconnected curve in the cobordism ) .
algebraically this gives precisely the transformation rule for @xmath24 described in the statement with respect to the lift @xmath340 of the isomorphism @xmath341 to the homology tensor algebras of @xmath342 .
for any @xmath343 and any @xmath344 @xmath345 simply expand the equation @xmath346 in powers of the @xmath68-variables and consider the linear terms .
consider the case @xmath347 with @xmath348 , @xmath349 and @xmath350 where @xmath351 is the angular coordinate on @xmath0 .
it is easy to compute @xmath352 . writing just @xmath254 for the index @xmath353 associated to the @xmath254-th multiple of the orbit @xmath354 , from the dimension formula for the moduli space of sft - curves and an easy curve counting
we immediately see that @xmath355 @xmath356 [ masterch ] @xmath357 @xmath358 for the proof we need to study the boundary of the moduli spaces of contact homology curves with three special punctures , the positive one and two of the negative ones . without loss of generality , in the interior of such moduli spaces
, the positive puncture can be the image of @xmath359 , and the two negative ones of @xmath360 and @xmath361 .
then the real half - line @xmath333 and the half - circle determine asymptotic directions at the three points ( the one at @xmath289 pointing along @xmath333 , the one at @xmath96 and @xmath290 pointing towards @xmath289 along the half - circle and @xmath333 respectively ) .
these will be constrained to asymptotic markers at the corresponding reeb orbits by pulling back classes @xmath362 , @xmath363 and @xmath364 at the orbits @xmath334 , @xmath335 and @xmath365 respectively
. the above picture , with the usual notations , represents the possible codimension-@xmath61 boundary degenerations in such moduli spaces .
the two red lines , with common tangent direction at the positive puncture , represent the half - circle and the @xmath333 line .
+ the second , third , fourth and sixth term of the right - side form the lie derivative along @xmath42 of the @xmath366-tensor whose correlator counts the holomorphic curves with three special punctures described above , and hence disappear when taking homology with respect to @xmath367 .
the first term spells out as @xmath368 .
indeed the curves represented in the lower floor of the first summand have two red lines on them , whose tangents at the positive puncture match but are not constrained to meet any asymptotic marker on the reeb orbit .
the matching condition can be once more expressed by pulling back to the moduli space the form @xmath369 ) ) representing the anti - diagonal class in @xmath337 ( the class of the diagonal in @xmath370 , where the minus in the second factor denotes reversed orientation ) , which immediately gives the factor @xmath371 .
+ the remaining summands , the fifth and the seventh , give the remaining part of the master equation .
+ for any @xmath372 , @xmath373 simply spell out in components the difference @xmath374 to see that it is proportional to the left - hand side of the master equation for @xmath24 .
the following result shows how the non - equivariant nijenhuis endomorphism @xmath24 is related to the geometry of gravitational descendants and the combined knowledge of the primary vector fields @xmath375 and of the endomorphism @xmath376 allows for completely recovering all of the descendant vector fields @xmath377 , @xmath378 .
+ [ omegarecursionch ] @xmath379 where @xmath380 once more we need to study the codimension-@xmath61 boundary of a moduli space of curves . in this case
we consider contact homology curves with three special points : the positive puncture at the orbit @xmath81 , a marked point at which we pull back the unity class @xmath255 and another marked point carrying the @xmath301-nth descendant of the class @xmath381 ( and no other point carries gravitational descendants ) .
mapping these three points to @xmath382 we obtain an asymptotic direction at the positive puncture given by the @xmath333-line and we constrain such direction as usual via the asymptotic marker at the corrisponding positive reeb orbit . in the usual way the above
picture shows the different types of codimension-@xmath61 boundary degeneration for such moduli space .
notice that , since the two special marked points are constrained to a line which is in turn @xmath0-constrained at the reeb orbit , a special kind of codimension-@xmath61 phenomenon appears , which is not anymore a @xmath336-floor curve , but is instead a @xmath61-floor curve with a constant sphere - bubble carrying the two special marked points ( which corresponds to the limit where the point carrying the class @xmath255 moves along the @xmath333-line to reach the other marked point carrying the descendant ) , represented as the last term in the right hand side in the picture .
+ now we notice that the first and fourth terms in the right - hand side correspond to the lie derivative along @xmath42 of a vector field on @xmath312 whose component along @xmath383 are given by the correlator counting the curves described above .
notice further that the factor corresponding to top floor in the second summand is zero unless the curve is a constant cylinder , because the marked point carrying the class @xmath255 is always unconstrained along the red line and the only way to achieve a zero - dimensional moduli space is by quotienting out the vertical symmetry in constant cylinders over the reeb orbit @xmath81 .
taking homology , what is left can be spelled out as @xmath384 where @xmath385 is the term accounting for the constant bubbles with one psi - class to the power @xmath301 .
such term is easily calculated from the well known fact that , on the deligne - mumford space of genus @xmath289 curves with @xmath170 marked points , @xmath386 finally we need to use the string equation of theorem [ sdd ] on the above equation to obtain the statement .
@xmath387 where @xmath380 just apply theorem [ omegarecursionch ] @xmath301 times to @xmath43 . naturally the above theorem and corollary hold for any choice of auxiliary data given the completely covariant behaviour of the equations .
in particular the above corollary shows how the descendant vector fields @xmath43 are expressed in closed form in terms of the primary vector fields @xmath388 and the endomorphism @xmath24 .
consider again the example of @xmath347 . in this case
we have @xmath389 and @xmath390 , with @xmath391 , where as before we write @xmath254 for the index @xmath392 associated to the @xmath254-th multiple of the orbit @xmath354 .
+ applying the above theorem [ omegarecursionch ] we obtain @xmath393 . here
@xmath394 , hence we obtain @xmath395 and , with the same procedure we obtain , applying @xmath24 @xmath301 times and denoting @xmath396 , @xmath397 ( one needs to use the following trick @xmath398 to take the numerical coefficient out of the sum ) .
an approach similar to the one we used above for contact homology also works in the case of full rational sft , the main difference coming from the presence of non - constant nodal curves which is very naturally incorporated in the algebraic formalism of lie derivatives and tensor fields by trading the nijenhuis operator @xmath24 for a bivector @xmath46 well defined on sft homology . for
any target stable hamiltonian structure @xmath399 , consider the following moduli spaces of punctured @xmath0-parametrized cylinders with marked points .
we start with the fully parametrized space @xmath400 consisting of tuples @xmath90 , where @xmath401,@xmath402,@xmath403 are three disjoint ordered sets of points on @xmath279 ( positive and negative punctures , and @xmath170 additional marked points ) .
the map @xmath93 from the punctured riemann surface @xmath404 is required to satisfy the cauchy - riemann equation @xmath95 with respect to the complex structure @xmath96 on @xmath92 .
assuming we have chosen cylindrical coordinates @xmath97 around each puncture @xmath98 , in the sense that @xmath99 , the map @xmath100 is additionally required to show for all @xmath101 the asymptotic behaviour @xmath102 with some @xmath103 and the orbits @xmath104 , where @xmath105 denotes period of @xmath106 , and analogous asymptotic behaviour at @xmath289 and @xmath290 for @xmath405 ( the signs here correpond to the signs in @xmath406 in the notation for the moduli space ) for orbits @xmath293 and @xmath294 with respect to the natural coordinates on @xmath295 .
we assign to each curve an absolute homology class @xmath107 employing as usual a choice of spanning surfaces . in order to obtain the @xmath0-parametrized space @xmath407 we only divide out the @xmath77-component of the @xmath297 group of automorphisms of the cylinder @xmath408 and , as usual , the @xmath77-action coming from the cylindrical target @xmath299 as well .
+ the compactification @xmath409 is obtained as usual by adding multifloor @xmath0-parametrized curves . in genus zero
each floor has only one non - trivial connected component , all the others being trivial cylinders over reeb orbits ( possibly @xmath0-parametrized ) .
if the @xmath289 and @xmath290 punctures determining the @xmath0-parametrization appear on the @xmath254-th and @xmath410-th floor of a @xmath301-floor curve , it means that all non - trivial curves appearing on the @xmath411-th floor are @xmath0-parametrized when @xmath412 and ordinary unparametrized curves when @xmath413 or @xmath414 .
nodal curves must also be added to the picture , as usual , with the only remark that , when a node separates the @xmath289 and @xmath290 punctures on a given floor , each component carries his own @xmath0-parametrization with respect to the node and the @xmath289 or @xmath290 puncture respectively .
+ as in the contact homology case , the space @xmath303 carries , besides the usual evaluation maps at marked points , orbits and asymptotic markers , extra evaluation maps at the punctures at @xmath289 and @xmath290 to the corresponding target simple reeb orbits given by the special @xmath0-coordinate on the curve , @xmath415 @xmath416 + we form the space @xmath417 by taking the union over @xmath418 and @xmath307 of all the spaces @xmath419 with @xmath420 and @xmath152 .
+ we will now define a @xmath309-tensor on the poisson super - space @xmath421 underlying the poisson subalgebra @xmath422 generated by @xmath262 and @xmath423-variables and even @xmath68-variables only .
in other words , if @xmath424 , with @xmath425 and @xmath426 , we denote by @xmath320 the ( even ) formal variable associated to the class @xmath203 , @xmath343 , and by @xmath427 the ( odd ) formal variable associated to @xmath428 , @xmath429
. then @xmath430 .
correspondingly we define @xmath431 and @xmath432 .
notice that the hamiltonians @xmath433 are always even elements in @xmath434 .
+ we will denote globally by @xmath435 any of the coordinates @xmath320 , @xmath436 or @xmath437 ( again , to avoid confusion , we have raised the indices of @xmath262 and @xmath423 variables , coherently with their interpretation as coordinates for @xmath421 ) .
we will always use lower case roman indices ( e.g. @xmath438 ) to refer indistinctly to a @xmath262 or @xmath423 variable , greek indices ( e.g. @xmath439 ) for @xmath68 variables and checked greek indices ( e.g. @xmath440 ) for @xmath441-variables .
also , for convenience , we let @xmath442 and @xmath443 , so that the roman upper case indices @xmath107 , @xmath444 , etc . can take the values @xmath445 , @xmath446 , etc .
when the corresponding variable is a @xmath68-variable , or the values @xmath447 , @xmath448 , etc . or again simply @xmath449 and @xmath450 when the corresponding variable is a @xmath262 or @xmath423-variable . notice also that , as opposed to what we did for the space @xmath312 , the @xmath68-variables are treated here as genuine coordinates and not as parameters .
in particular , the poincar metric @xmath451 will split into two blocks ( one the transpose of the other ) always pairing even with odd cohomology classes .
we denote the matrix corresponding to each of such blocks by @xmath452 .
+ using such coordinates we define the ( graded ) symmetric bivector @xmath453 where we sum over repeated indices , with @xmath454 and @xmath455 and zero otherwise .
+ as in contact homology , from the index formula for the virtual dimension of the moduli space of sft curves , we have @xmath456 algebraically , the sft differential @xmath457 , defined as the vector field @xmath458 induces a differential @xmath459 on the space of @xmath28-tensor fields @xmath460 on the poisson super - space @xmath421 underlying @xmath434 .
the resulting homology , which we denote by @xmath461 , is a module over @xmath462 and is an invariant of the stable hamiltonian structure on @xmath48 . in particular , for two different choices of form @xmath227 , cylindrical almost complex structure @xmath228 , representatives for the classes @xmath225,[\theta_{\check{\alpha}}]\in h^*(v)$ ] and @xmath226\in h^*(s^1)$ ] , abstract polyfold perturbations and sequences of coherent collections of sections @xmath229 , there exist an isomorphism @xmath463 which is simply the lift to the tensor algebra of the isomorphism @xmath464 constructed in @xcite by studying curves in the cobordims @xmath327 interpolating between the two different choices ( see also the discussion on invariance for satellites there ) .
+ moreover the descendant hamiltonians @xmath465 induce covariant ( with respect to @xmath340 ) hamiltonian vector fields @xmath466 , @xmath429 , @xmath467 .
+ we proceed exactly as in the contact homology case , only keeping in mind that , this time , nodal configurations can appear in codimension @xmath61 when studying the moduli spaces , relevant for @xmath24 , of curves with a doubly @xmath0-constrained line joining the two special @xmath289 and @xmath290 punctures . indeed such extra boundary , containing nodal curves where the node separates the @xmath289 and @xmath290 puncture , corresponds to the term @xmath469 and @xmath470 in the lie derivative @xmath471 .
the following result is the analogue of theorem [ omegarecursionch ] ( and proved in completely similar way ) for the rational sft case , and shows how the non - equivariant bivector @xmath46 is related to the geometry of gravitational descendants and the combined knowledge of differential of the hamiltonian @xmath472 and of the graded symmetric bivector @xmath473 allows to recover the descendant vector fields @xmath474 , @xmath475 .
however how , in general , this is not equivalent to recovering the hamiltonians @xmath476 themselves .
+ the statement is proved precisely in the same way as for theorem [ omegarecursionch ] .
notice only that the analogue of the term containing the constants @xmath478 , counting nodal curves , in this case is absorbed in the lie derivative that vanishes in homology .
notice that the above recursion makes sense for @xmath479 too if we define @xmath480 .
then all of our sequences of hamiltonians @xmath433 satisfy to a recursion which starts from a casimir at level @xmath479 .
this allows to deduce commutativity @xmath481 , which we know to hold on homology , simply from the recursion , since @xmath482 consider again the case @xmath347 with @xmath348 , @xmath349 and @xmath350 where @xmath351 is the angular coordinate on @xmath0 . here , as in any other circle bundle over a symplectic manifold with even cohomology , @xmath483 and everything happens at chian level .
even in the full rational sft case , it is straightforward to compute @xmath46 .
we write @xmath484 for the index @xmath485 associated to the @xmath254-th multiple of the positive or negative orbit @xmath354 and we use the index @xmath289 to refer to the component along @xmath486 ( or , in other words , @xmath487 ) . from the dimension formula for the moduli space of sft - curves and
some easy curve counting we immediately see that @xmath488 applying @xmath46-recursion we can recover the @xmath301-th descendant hamiltonian .
indeed , let us start with @xmath489 recursion tells us @xmath490 from which we deduce @xmath491 notice that , actually , for @xmath492 , the above equation is void , as we expected .
the same procedure can be reiterated to find @xmath493 f. magri , _ a geometrical approach to the nonlinear solvable equations _ , in m. boiti , f. pempinelli , and g. soliani , editors , _ nonlinear evolution equations and dynamical systems _ ,
volume 120 of _ lecture notes in physics _
, springer
verlag , 1980 .
s. piunikhin , d. salamon and m. schwarz , _
symplectic floer - donaldson theory and quantum cohomology .
_ proceedings of the symposium on symplectic geometry , held at the isaac newton institute in cambridge in 1994 , edited by c.b .
thomas , lms lecture note series . | in this paper we investigate the algebraic structure related to a new type of correlator associated to the moduli spaces of @xmath0-parametrized curves in contact homology and rational symplectic field theory .
such correlators are the natural generalization of the non - equivariant linearized contact homology differential ( after bourgeois - oancea ) and give rise to an invariant nijenhuis ( or hereditary ) operator ( la magri - fuchsteiner ) in contact homology which recovers the descendant theory from the primaries .
we also show how such structure generalizes to the full sft poisson homology algebra to a ( graded symmetric ) bivector .
the descendant hamiltonians satisfy to recursion relations , analogous to bihamiltonian recursion , with respect to the pair formed by the natural poisson structure in sft and such bivector . |
a surface is called minimal if its mean curvature vanishes everywhere carmo.minimal surfaces are among the most important objects studied in differential geometry .
the theory of minimal surfaces in euclidean spaces @xmath2 has its roots in the calculus of variations developed by euler and lagrange in the 18-th century and in later investigations by enneper , scherk , schwarz , riemann and weierstrass in the 19-th century . during the years
, many great mathematicians have contributed to this theory .
early , two problems were posed from the belgian physicist plateau and from the swedish mathematician bjrling in the 19-th century for minimal surfaces passing a curve .
plateau set the problem as find a minimal surface @xmath3 having a curve @xmath4 as boundary. this problem was a natural outgrowth of his physical experiments on soap films .
plateau s problem was first solved by douglas and rado .
bjrling considered is to find the minimal surface passing through a given non - closed analytic curve @xmath4 , with given tangent planes along @xmath4 .
the problem was inquired and solved by bjrling .
a solution of this problem always exists and is unique.this solution of bjrling s problem always applicable one to find the minimal surface whenever one of its geodesic lines or one of its asymptotic lines or one of its lines of curvature is known .
one can obtain and detailed description of minimal surfaces , plateau s and bjrling s problems from @xcite and oprea recently , parametric representation of a surface family passing a given geodesic curve was studied in @xcite which using the ferenet frame of a given curve . in like manner ,
@xciteand bayram et al.@xcite studied the necessary and sufficient condition for line of curvature and asymptotic curve , respectively .
li et al . approximation studied the approximation minimal surface with geodesics by using dirichlet function . in this paper , without limitation for curve such that it becomes geodesic lines or asymptotic lines or lines of curvature on minimal surface
, we investigate the parametric representation of minimal surface family passing a given curve .
we utilize the frenet frame of given curve and isothermal parameter on minimal surface .
we derive the necessary and sufficient condition for a surface which satisfying both passing a given curve and becomes minimal .
in section two , we give definitions of isothermal parameter of a surface and frenet frame of a curve . in section three
, we derive the necessary and sufficient condition for minimal surface . in section four and five
, we present parametric representations of minimal surface families passing a given circle and a helix , respectively .
we have following tools for minimal surfaces and curves .
@xcitelet @xmath5 be a reguler surface where @xmath6 @xmath2and @xmath7 @xmath2 .
a parametrization @xmath5 is called isothermal if @xmath8 we utilize isothermal parameters for parameterization of minimal surfaces .
this requirement is not restriction because of the following lemma .
@xciteisothermal parameters exist on any minimal surface in @xmath2 .
the mean curvature of a surface can be calculated by @xmath9 where @xmath10 @xmath11 and @xmath12 is unit normal of the surface defined by @xmath13if we apply isothermal conditions on [ mean ] then we have the following corollary .
@xcitelet @xmath14 is isothermal .
then @xmath14 is minimal if and only if its coordinate functions are harmonic i.e. @xmath15 let @xmath16 be an arclenght parametrized curve that is @xmath17 .
we assume that @xmath18 for some @xmath19 , otherwise the curve @xmath20 corresponds to a straight lines.the frenet frame of a regular curve @xmath20 is defined @xmath21 where @xmath22 , @xmath23 , @xmath24 , are called unit tangent , unit normal and binormal of @xmath20 , respectively .
the derivatives of the franet frame are given by @xmath25 where @xmath26 and @xmath27 are the curvature and torsion of the curve @xmath20 , respectively . by utilizing the frenet frame ,
wang et al.@xcite , construct a surface pencil that posseses @xmath20 as a common geodesic .
they gave parametric representation of surface as @xmath28 where @xmath29 are smooth functions with @xmath30 the unit normal of the surface is @xmath31where @xmath32 they derived the following identities for the necessary and sufficient condition for the curve @xmath20 as a geodesic on the surface @xmath14 .
@xmath33 @xmath34 @xmath35 in @xcite , bayram et al . derived the necessary and sufficient condition for the curve @xmath20 as an asymptotic line on the surface @xmath36
we express minimal surfaces with this parametric representation .
in this section , we derive necessary and sufficient condition for minimal surface family passing the curve @xmath16 .
if we calculate the partial derivatives of @xmath5 then we have , @xmath38 from [ isothermal ] and [ harmonic ] we have following corollaries .
if the surface @xmath14 is isothermal then @xmath39 if the surface @xmath14 is harmonic then@xmath40 if we consider the surface satisfy the isothermal and harmonic conditions together then we have the following theorem .
let @xmath16 is be an arclenght parameterized curve .
the surface [ surpen ] is minimal surface passing the curve @xmath20 if and only if there exist the smooth functions @xmath29
satisfying,@xmath41@xmath42 if the conditon [ isopara ] is satisfied then @xmath43let the surface @xmath44 is be minimal surface . from lemma 1 and corollary 1 there
exist isothermal parameters and the surface @xmath14 is harmonic .
thus , there exist the smooth functions @xmath29 satisfying the conditions [ isot]-[har3 ] . on the other hand , if there exist the smooth functions @xmath45 satisfying the conditions [ isot]-[har3 ] then the surface @xmath44 is isothermal and harmonic .
thus the surface @xmath44 is minimal .
in this section , we investigate a minimal surface family passing a circle . let the curve @xmath16 is be a circle,@xmath46the curvature and torsion of the circle @xmath20 are become @xmath47 if we choose @xmath48 , @xmath49 , @xmath50 and @xmath51 then from [ isot]-[har3 ] we have @xmath52 from [ c5 ] @xmath53from [ c3 ] and
[ c4]@xmath54@xmath55from [ c1 ] and [ c2 ] @xmath56@xmath57@xmath58where @xmath59 $ ] .
the frenet frame of the circle @xmath20 is @xmath60 consequently , we derive the minimal surface family passing the circle @xmath20 as @xmath61 its obvious that each members of this minimal surface family is a surface of revolution .
we know that only surface of revolution whic is minimal is a piece of catenoid or a plane .
thus each member of this family is a piece of catenoid or a plane . for @xmath62 ,
the circle @xmath20 is a geodesic on the surface @xmath63 for @xmath62 , [ geo1 ] and [ geo2 ] are satisfied .
thus , the circle @xmath20 is a geodesic on the surface @xmath64 .
we show that some members of minimal surface family in figure [ fig : figure1]- [ fig : figure4 ] for @xmath65 @xmath66 @xmath67 and @xmath68 , @xmath69 .
+ and minimal surface @xmath70 and minimal surface @xmath71 and minimal surface @xmath72 ]
in this section , we investigate a minimal surface family passing a helix . let the curve @xmath73 is be a helix,@xmath74the curvature and torsion of the circle @xmath20 are become @xmath75 if we choose @xmath48 , @xmath49 , @xmath50 and @xmath51 then from [ isot]-[har3 ] we have @xmath76 from [ h3 ] and [ h5 ] @xmath77 @xmath78thus,@xmath79@xmath80from [ h4 ] @xmath81where @xmath82 . from [ h1 ] and
[ h2]@xmath83@xmath84@xmath85 the frenet frame of the helix @xmath86 is @xmath87 consequently , we derive minimal surface family passing the helix @xmath86 as @xmath88where @xmath89 . for @xmath90 ,
the helix @xmath86 is a geodesic on the surface @xmath91 for @xmath92 , the helix @xmath86 is an asymptotic line on the surface @xmath93 for @xmath92 , [ asymp ] is satisfied .
thus , the helix @xmath86 is a asymptotic line on the surface @xmath94 .
we show that some members of minimal surface family in figure [ fig : figure5 ] - [ fig : figure8 ] for @xmath95 @xmath96 @xmath97 and @xmath98 , @xmath99 . and minimal surface @xmath100 and minimal surface @xmath101 and minimal surface @xmath102 .
we derive necessary and sufficient condition for minimal surface passing a given curve . also , we present parametric representation of minimal surface families passing a circle and a helix . | we investigate minimal surfaces passing a given curve in @xmath0@xmath1 . using the frenet frame of a given curve and isothermal parameter
, we derive the necessary and sufficient condition for minimal surface .
also we derive the parametric representation of two minimal surface families passing a circle and a helix as examples . |
the authors thank s.v .
borisenko , j.c .
davis , i. eremin , t. hanaguri , d.k .
morr , a.w .
rost for valuable discussions and comments .
furthermore , we thank s.v .
borisenko for providing band structure data of lifeas . this work has been supported by the deutsche forschungsgemeinschaft through the priority programme spp1458 ( grant no . be1749/13 and gr3330/2 ) and the graduate school grk 1621 . | several angle resolved photoemission spectroscopy ( arpes ) studies reveal a poorly nested fermi surface of lifeas , far away from a spin density wave instability , and clear - cut superconducting gap anisotropies . on the other hand a very different
, more nested fermi surface and dissimilar gap anisotropies have been obtained from quasiparticle interference ( qpi ) data , which were interpreted as arising from _ intraband _ scattering within hole - like bands . here
we show that this arpes - qpi paradox is completely resolved by _ interband _ scattering between the hole - like bands .
the resolution follows from an excellent agreement between experimental quasiparticle scattering data and @xmath0-matrix qpi calculations ( based on experimental band structure data ) , which allows disentangling interband and intraband scattering processes .
the band structure of the iron - based superconductor lifeas became recently a matter of strong controversial interest @xcite since its precise knowledge is crucial for an understanding of the unusual superconductivity of lifeas .
iron - arsenides have the well - known property that experimentally observed factors such as fermi surface ( fs ) nesting and van - hove singularity - like behavior have a decisive impact on the type of spin fluctuations and on the superconducting ground state : a poorly nested model band structure of lifeas which mimics essential findings from angle resolved photoemission spectroscopy ( arpes ) experiments yields prevailing ferromagnetic fluctuations with an instability towards triplet superconductivity @xcite , whereas a calculated electronic band structure with a stronger nesting has been shown to result in an @xmath1-wave superconducting ground state driven by antiferromagnetic fluctuations @xcite . several arpes studies on lifeas indeed consistently reveal a quite poor nesting , an ( almost ) van hove singularity for one hole - like band at the fermi level , and clear - cut momentum dependencies of the superconducting gap @xmath2 for each band @xmath3 @xcite . a much different normal state band structure which corresponds to a stronger nesting and other superconducting gap anisotropies @xmath2 have , however , recently been deduced from a spectroscopic imaging scanning tunneling microscopy ( si - stm ) study of the quasiparticle interference ( qpi ) of lifeas @xcite .
this finding is based on the central assumption that quasiparticles inducing the qpi are scattered only internally within separate hole - like bands . in this letter , we show that this seeming arpes - qpi paradox is completely resolved when _ interband _ scattering is taken into account .
to be specific , we compare the experimentally observed qpi pattern at small wave vectors @xmath4 with qpi calculations which _ i ) _ are based on experimental band structure data and _ ii ) _ allow to consider individual scattering processes separately .
we find an excellent agreement between the experimental and calculated data when the mentioned interband scattering is taken into account , whereas _ intraband _ scattering alone ( the essential hypothesis in ref . )
fails to yield such an agreement .
this explains the observed qpi s of ref . and that of another qpi study of lifeas
@xcite without having to resort to low - energy band structure models that are contradicting arpes experiments .
lifeas is a _
stoichiometric _ superconductor , i.e. superconductivity emerges already without doping , at a relatively high critical temperature @xmath5 k @xcite .
already this renders it radically different from the canonical 1111 and 122 iron - arsenide superconductors , where strong nesting between hole - like and electron - like fs pockets causes an antiferromagnetic spin density wave ( sdw ) ground state in the undoped parent compounds .
there , superconductivity emerges only if the sdw state is weakened e.g. by doping @xcite , and nesting - related antiferromagnetic fluctuations have been suggested to drive the superconductivity with an @xmath1-wave order parameter @xcite .
lifeas , however , appears to be far away from an sdw instability : all doping attempts lead to a decrease of @xmath6 @xcite , and according to the arpes studies @xcite , the fs of lifeas consists of two similarly sized electron - like fs sheets around the @xmath7-point , and two hole - like fs sheets around the @xmath8-point arising from hole - like bands ( labelled @xmath9 and @xmath10 , see fig .
[ fs_schem ] ) .
one of the two bands has a relatively large fs ( @xmath10-band , @xmath11 ) , whereas the other is very small ( @xmath12 ) such that the corresponding @xmath9-band just touches the fermi level yielding a structure close to a van - hove singularity .
due to the lack of surface states @xcite these findings and those from si - stm are expected to be bulk representative , and in fact the arpes electronic structure has been shown to be well consistent with bulk sensitive hall effect results @xcite . in their qpi study @xcite , allan et al . used the extracted scattering vectors of the qpi to construct three hole - band dispersions along high - symmetry directions , where they assumed the observed scattering vectors to stem from intraband scattering .
one of the resulting bands ( labeled @xmath13 in ref . ) is consistent with the @xmath10-band observed in arpes @xcite , and another ( @xmath14 ) matches quite well the @xmath9-band ( see fig .
[ fs_schem ] ) .
however , the third of the suggested bands ( @xmath15 ) lacks such a correspondence since its fermi wave vector @xmath16 @xcite neither matches that of the large nor the small hole - like fs observed in arpes ( see fig .
[ fs_schem ] ) .
its length is , however much closer to that of the electron - like bands and thus is the one which would result in a stronger nesting . and @xmath13 as obtained from qpi by allan et al .
@xcite .
note that allan et al .
extract the fermi wave vectors along high - symmetry directions and that the shown fs sheets are respective isotropic two - dimensional extrapolations .
the shown unfolded brillouin zone ( bz ) refers to the one - fe unit cell .
the bz related to the two - fe unit cell is indicated by dotted lines .
b ) lifeas band structure at negative energies .
circles represent the extracted quasiparticle dispersions @xmath17 of the hole - like bands @xmath14 , @xmath15 , @xmath13 obtained in ref . , where @xmath18 is given by half of the length of the observed scattering vectors @xmath19 .
grey shaded contours represent @xmath9 and @xmath10 bands as observed by arpes @xcite .
the extended width of the @xmath9 band indicates @xmath20 dependence obtained from different photon energies @xcite . the indicated dispersion shown as a dashed line corresponds to data obtained at a photon energy @xmath21 ev . ] as is evident in refs . and ,
a prominent feature of the fourier transformed experimental qpi pattern at negative bias is a bright squarish contour at small @xmath22 .
fig .
[ exp_schem](a ) shows our si - stm data at @xmath23 mev , which we have obtained on high - quality superconducting lifeas single crystals @xcite in a home - built scanning tunneling microscope at @xmath24 k ( see ref . for details of the measurement and the full topographic and spectroscopic data set ) .
the chosen energy value is significantly larger ( @xmath25 mev ) than the superconducting gap value @xcite .
thus gap anisotropies do not play a strong role in the qpi pattern .
the observed square structure can be straightforwardly understood by considering backscattering between the small ( @xmath9 ) and the large ( @xmath10 ) hole band @xcite : due to the high dos of the @xmath9-band and the relatively small momentum spread of this band at a given energy close to the band - top ( see fig . [ fs_schem](b ) ) , the corresponding scattering vectors ( labeled @xmath26 ) just map the constant energy contour ( cec ) of the @xmath10-band ( cf . fig . [ exp_schem](b ) ) . in order to underpin this statement
, we show in fig .
[ fs_schem](b ) that the diagonal @xmath27 with @xmath28 @xcite with a length @xmath29 ( with @xmath30 the shortest fe - fe distance , and @xmath31 the in - plane lattice constant @xcite ) just connects states in the @xmath9 and @xmath10 bands along the @xmath32 direction as observed at this energy by arpes @xcite .
ref . reports a similar squarish structure , the size of which along the @xmath32 direction allan et al . used to extract the @xmath15-dispersion .
in fact , we find that the diagonal @xmath26 perfectly matches the reported @xmath33 at @xmath34 @xcite .
this shows unambiguously that _
i ) _ the present and the reported qpi patterns of allan et al . are geometrically well consistent , and _
ii ) _ as a consequence , the suggested @xmath15-band by allan et al . does not describe a quasiparticle dispersion but has to be reinterpreted as the dispersion of the scattering wave vectors @xmath26 which connect the @xmath9 and @xmath10 hole - like bands .
mev @xcite .
b ) simplified constant energy contours ( cec ) of the hole - like @xmath9 and @xmath10 bands at @xmath23 mev .
@xmath26 represents interband scattering processes which connect states on both bands .
it has the same length as @xmath35 ( i.e. , the diameter of the dashed @xmath15-cec ) reported in ref . .
@xmath36 represents intraband scattering within the @xmath10 band .
[ exp_schem ] ] having established this most important finding we corroborate it further by qpi simulations which we use to analyze the influence of the individual possible scattering processes separately .
the calculation of the qpi intensity distribution arising from the local density of states ( ldos ) is based on a standard @xmath0-matrix approach .
thereby the underlying electronic band structure of lifeas is modeled by a tight - binding approximation of the arpes results of refs . and .
from the latter work we have derived an appropriate model for the superconducting gap function @xmath37 which is able to model the measured momentum dependence of the superconducting gap along the fermi surface of the hole pockets @xmath9 and @xmath10 of fig . [ fs_schem](a ) .
the model describes anisotropic @xmath38-wave pairing where the gap value along @xmath10 varies between @xmath39 mev and @xmath40 mev being minimal at the direction towards the electron - like fs .
the gap along @xmath9 is almost isotropic with a value of approximately @xmath41 mev .
-matrix approach applied to a tight - binding model of the arpes results @xcite and anisotropic @xmath38-wave pairing ( see text ) .
a ) only intraband scattering within the small hole - like band @xmath9 and the large hole - like band @xmath10 is considered .
all scattering processes between @xmath9 and @xmath10 and those processes involving the electron bands are suppressed in the calculation .
b ) only interband scattering between @xmath9 and @xmath10 is considered .
c ) contributions displayed in panels a ) and b ) are summed up in order to enable the comparison of the intensities .
the scattering vectors @xmath26 , @xmath36 ( see fig . [ exp_schem ] ) are indicated . ] for the numerical evaluation we use a square lattice of @xmath42 lattice points and periodic boundary conditions . to simplify the treatment
we assume that the isotropic scattering potential acts locally without coupling to the spin of conduction electrons ( potential scattering ) .
furthermore , the cleanliness of the sample used for the stm measurements @xcite allows us to neglect multiple scattering processes ( born approximation ) .
fig .
[ theo ] shows calculated qpi patterns for bias voltage @xmath23 mev . in order to identify the origin of the different features in the measured intensity distribution displayed in fig .
[ exp_schem](a ) we faded out specific scattering processes such that only intraband scattering within the two hole pockets @xmath9 and @xmath10 ( panel ( a ) ) and only interband scattering between @xmath9 and @xmath10 ( panel ( b ) ) entered the calculation separately . in panel ( c ) the intensities of both scattering types are summed up . as it is found already from geometrical interpretation of the experimental result
our numerical simulation confirms that the bright squarish structure described by vector @xmath26 has to be assigned clearly to interband scattering .
moreover , its intensity is significantly larger than the intraband contributions which are seen as _
i ) _ a small region around the @xmath8 point indicating scattering within the small hole pocket @xmath9 and _ ii ) _ a large contour indicated by scattering vectors @xmath36 connecting momenta within the large hole fs . the comparison between fig . [ theo](c ) and
fig .
[ exp_schem](a ) shows that the size and shape of most of the qpi features can be reproduced by our numerical simulation in excellent agreement with the experimental results . beyond the bright structures we find also several features showing minor intensity .
for example , the intraband scattering within @xmath10 shows a pronounced maximum at momenta @xmath43 along the lines connecting the points @xmath44 and @xmath45 which is seen very well in the experimental result ( @xmath36 ) . on the other hand structures within the bright @xmath26-contour
appear with less intensity as they show up in the experimental result fig .
[ exp_schem](a ) . this difference might be traced back to the influence of coherence factors entering the numerical simulation .
it is well - known that a possible variation of the superconducting phase along the fermi surface strongly affects the intensity distribution of the qpi at bias energy close to the superconducting gap @xcite .
note that within our simple anisotropic @xmath38-wave order parameter model though the gap value is consistent with arpes studies a possible variation of the phase is neglected completely .
such phase effects are interesting and will be subject to a forthcoming publication . to sum up
, we have shown that the quasiparticle interference in lifeas is dominated by interband scattering between two different hole - like bands around the @xmath8 point .
this interpretation of the experimental results is supported by a comparison to a numerical simulation of the qpi based on a realistic tight - binding model that has been carefully extracted from recent arpes data . within this theoretical analysis we considered the different scattering processes separately and we obtained an excellent agreement with the experimental qpi data from refs . and .
this implies that a consistency between the results of the two experimental stm approaches and arpes is obtained when going beyond considering only intraband scattering and include interband scattering explicitly .
this changes fundamentally the interpretation of the extracted @xmath15-band and the proposed gap anisotropy @xmath46 of ref . , and explains in both cases their strong discrepancy with arpes data .
for the @xmath14 and @xmath13 bands the consistency with the arpes results and also the present qpi calculations is apparently very good .
this is compelling evidence that the band structure suggested by apres is indeed correct .
it is interesting to note , that this band structure is also well compatible with a recent de haas van alphen study @xcite , where significant fs nesting in lifeas is claimed , despite only electron - like orbits in agreement with arpes have been observed by this bulk sensitive technique so far . |
we thank joshua r. wickman for valuable discussions . the work of s. k.
was partially supported by the science and technology development fund ( stdf ) project i d 437 and the ictp project i d 30 .
q.s and m. r acknowledge partial support from an nsf collaborative grant with egypt , award number 0809789 , and support from the doe under grant no .
de - fg02 - 91er40626 ( q.s . and m.r . ) , and by the university of delaware ( m.r . ) .
a. d. linde , phys . rev .
d * 49 * , 748 ( 1994 ) [ arxiv : astro - ph/9307002 ] .
g. r. dvali , q. shafi and r. k. schaefer , phys .
lett . * 73 * , 1886 ( 1994 ) [ arxiv : hep - ph/9406319 ] .
e. j. copeland , a. r. liddle , d. h. lyth , e. d. stewart and d. wands , phys .
d * 49 * , 6410 ( 1994 ) [ arxiv : astro - ph/9401011 ] .
v. n. senoguz and q. shafi , phys .
b * 567 * , 79 ( 2003 ) [ arxiv : hep - ph/0305089 ] .
v. n. senoguz and q. shafi , phys .
d * 71 * , 043514 ( 2005 ) [ arxiv : hep - ph/0412102 ] .
r. jeannerot and m. postma , jhep * 0505 * , 071 ( 2005 ) [ arxiv : hep - ph/0503146 ] . c. pallis , arxiv:0710.3074 [ hep - ph ] .
s. antusch , k. dutta and p. m. kostka , phys .
b * 677 * , 221 ( 2009 ) [ arxiv:0902.2934 [ hep - ph ] ] .
s. antusch , m. bastero - gil , j. p. baumann , k. dutta , s. f. king and p. m. kostka , jhep * 1008 * , 100 ( 2010 ) [ arxiv:1003.3233 [ hep - ph ] ] .
d. h. lyth and a. riotto , phys .
rept . * 314 * , 1 ( 1999 ) [ arxiv : hep - ph/9807278 ] ; g. lazarides , lect . notes phys . *
592 * , 351 ( 2002 ) [ arxiv : hep - ph/0111328 ] ; a. mazumdar and j. rocher , arxiv:1001.0993 [ hep - ph ] , for a recent review .
m. kawasaki , m. yamaguchi and j. yokoyama , phys .
d * 68 * , 023508 ( 2003 ) [ arxiv : hep - ph/0304161 ] .
m. bastero - gil , s. f. king and q. shafi , phys .
b * 651 * , 345 ( 2007 ) [ arxiv : hep - ph/0604198 ] .
m. ur rehman , v. n. senoguz and q. shafi , phys .
d * 75 * , 043522 ( 2007 ) [ arxiv : hep - ph/0612023 ] .
q. shafi and j. r. wickman , phys .
b * 696 * , 438 ( 2011 ) [ arxiv:1009.5340 [ hep - ph ] ] ; m. u. rehman , q. shafi and j. r. wickman , arxiv:1012.0309 [ astro-ph.co ] .
l. covi , g. mangano , a. masiero and g. miele , phys .
b * 424 * , 253 ( 1998 ) [ arxiv : hep - ph/9707405 ] .
e. d. stewart , phys .
d * 51 * , 6847 ( 1995 ) [ arxiv : hep - ph/9405389 ] .
l. boubekeur and d. h. lyth , jcap * 0507 * , 010 ( 2005 ) [ arxiv : hep - ph/0502047 ]
. k. kohri , c. m. lin and d. h. lyth , jcap * 0712 * , 004 ( 2007 ) [ arxiv:0707.3826 [ hep - ph ] ] . c. m. lin and k. cheung , phys .
d * 79 * , 083509 ( 2009 ) [ arxiv:0901.3280 [ hep - ph ] ] . c. pallis , jcap
* 0904 * , 024 ( 2009 ) [ arxiv:0902.0334 [ hep - ph ] ] .
k. s. babu , i. gogoladze , m. u. rehman and q. shafi , phys .
d * 78 * , 055017 ( 2008 ) [ arxiv:0807.3055 [ hep - ph ] ] .
martin , phys .
d * 81 * , 035004 ( 2010 ) [ arxiv:0910.2732 [ hep - ph ] ] .
p. w. graham , a. ismail , s. rajendran and p. saraswat , phys .
d * 81 * , 055016 ( 2010 ) [ arxiv:0910.3020 [ hep - ph ] ] . | we analyze the adjoint field inflation in supersymmetric ( susy ) @xmath0 model . in minimal
susy @xmath0 hybrid inflation monopoles are produced at the end of inflation .
we therefore explore the non - minimal model of inflation based on susy @xmath0 , like shifted hybrid inflation , which provides a natural solution for the monopole problem .
we find that the supergravity corrections with non - minimal khler potential are crucial to realize the central value of the scalar spectral index @xmath1 consistent with the seven year wmap data .
the tensor to scalar ratio @xmath2 is quite small , taking on values @xmath3 .
due to @xmath4-symmetry massless @xmath5 octet and @xmath6 triplet supermultiplets are present and could spoil gauge coupling unification . to keep gauge coupling unification intact ,
light vector - like particles are added which are expected to be observed at lhc .
s. khalil@xmath7 , m. u. rehman@xmath8 , q. shafi@xmath8 and e. a. zaakouk@xmath9 _ @xmath10ain shams university , faculty of science , cairo , 11566 , egypt + @xmath11center for theoretical physics at the british university in egypt , + sherouk city , cairo 11837 , egypt + @xmath12bartol research institute , department of physics and astronomy , + university of delaware , newark , de 19716 , usa + _ [ [ section ] ]
inflation is one of the most motivated scenarios for the early universe , which is consistent with the recent cosmological observations on the cosmic microwave background radiation and the large - scale structure in the universe . in order to construct a consistent model of inflation , an extension of the standard model ( sm )
is required .
the supersymmetric ( susy ) grand unified theory ( gut ) models provide a natural framework for hybrid inflation @xcite-@xcite . on the other hand ,
supersymmetric @xmath0 is the simplest extension of the sm which may realize hybrid inflation through the adjoint scalar field , which is responsible for breaking the @xmath0 gauge symmetry into sm gauge group @xmath13 .
however , in the standard minimal version of susy hybrid inflation the gauge symmetry is broken at the end of inflation , and topological defects are copiously formed . to overcome this problem ,
we consider the leading order non - renormalizable term in the superpotential of susy @xmath0 hybrid inflation .
this class of inflationary models is known as shifted hybrid inflation , which have been introduced in ref .
@xcite in the context of a supersymmetric pati - salam model .
the inclusion of the non - renormalizable term introduces a non - trivial flat direction along which @xmath0 gauge symmetry is spontaneously broken through the appropriate higgs fields acquiring non - zero vacuum expectation values ( vevs ) .
this direction then can be used as an inflationary trajectory with one - loop radiative corrections providing the necessary slope for slow - roll inflation . however , since @xmath0 is broken during inflation , one finds that for a certain range of parameters , the system always passes through the above mentioned inflationary trajectory before falling into the susy vacuum .
therefore , the magnetic monopole problem is solved for all initial conditions .
the scalar spectral index @xmath14 in shifted hybrid inflation , driven solely by radiative corrections , is typically of order @xmath15 with the number of e - foldings @xmath16 and lies outside the seven year wilkinson microwave anisotropy probe ( wmap7 ) 1-@xmath17 bounds @xcite . including supergravity ( sugra ) corrections with minimal khler potential
further enhances the scalar spectral index to exceed unity and a blue - tilted scalar spectral index is obtained as emphasized in refs . @xcite .
the non - minimal khler potential plays a crucial role in reducing @xmath14 and making it compatible with the wmap7 data as shown in refs . @xcite . in susy
@xmath0 shifted hybrid inflation we show that the central value of the scalar spectral index @xmath18 can be realized in the presence of a small negative mass term generated from a non - minimal khler potential .
a salient feature of above mentioned @xmath0 inflationary model is that the octet and triplet components of the adjoint scalar field remain massless below gut scale after the breaking of @xmath0 symmetry .
the masses of these particles are related to the scale of @xmath4-symmetry breaking .
we discuss the effect of these particles on the gauge coupling unification along with the extra vector - like particles which should be added to restore unification .
if @xmath4-symmetry is broken at tev scale , these particles can be observed at the large hadron collider ( lhc ) with clear signatures .
the paper is organized as follows . in section 2
we discuss @xmath0 shifted hybrid inflation . here
we explore conditions which can lead inflaton to the shifted minimum .
we also include sugra corrections with the minimal and non - minimal khler potential . in section 3 ,
we discuss the impact of octet and triplet , which remain massless after @xmath0 symmetry breaking , on gauge coupling unification and show how to restore unification by introducing extra vectorlike particles at tev scale .
finally our conclusions are given in section 4 .
[ [ section-1 ] ] in susy @xmath0 , the matter fields are assigned to the @xmath19 , @xmath20 and @xmath21 dimensional representations , while the higgs fields belong to the adjoint scalar @xmath22 and fundamental scalars : @xmath23 and @xmath24 .
the @xmath4-symmetric inflation with @xmath4-symmetry violating terms has been considered in ref .
@xcite . ]
@xmath0 superpotential , with the leading order non - renormalizable term , is given by w= s+|hh+|h h + + y_ij^(u)10_i10_jh + y_ij^(d , e)10_i|5_j|h +
y_ij^()1_i|5_jh + m__ij1_i1_j , where @xmath25 is a gauge singlet superfield , @xmath26 is some suitable cutoff scale , @xmath27 is the neutrino mass matrix and @xmath28 , @xmath29 , @xmath30 are the yukawa couplings for quarks and leptons . only the terms linear in @xmath25 are relevant for inflation , and their role in realizing successful inflation will be discussed in detail .
the other two terms in the first line in eq .
( 1 ) are involved in the solution of the doublet - triplet problem .
fine tuning is required to implement doublet - triplet splitting and thereby adequately suppress proton decay .
the second line contains terms that generate masses for quarks and leptons .
the @xmath4-charge assignments of the various superfields are as follows : ( r_s , r _ , r_h , r_|h , r_10 , r_|5 , r_1 ) = ( 1 , 0 , , , , , ) . in component form , the above
superpotential takes the following formw s+t_a b^i^i|h_ah_b+|h_ah_a,[superpot - shift]where we express the scalar field @xmath31 in the @xmath0 adjoint basis @xmath32 with @xmath33 and @xmath34 . here
the indices @xmath35 , @xmath36 , @xmath37 run from 1 to 24 whereas the indices @xmath38 , @xmath39 run from 1 to 5 .
moreover , the repeated indices are summed over .
the scalar potential obtained from this superpotential is given by @xmath40note that the scalar components of the superfields are denoted by the same symbols as the corresponding superfields .
vanishing of the @xmath41-terms is achieved with @xmath42 and @xmath43 .
we restrict ourselves to this @xmath41-flat direction and use an appropriate @xmath4 transformation to rotate @xmath25 complex field to the real axis , @xmath44 , where @xmath17 is a real scalar field .
the supersymmetric global minimum of the above potential lies at^0 = h_a^0=|h_a^0=0 , with @xmath45 satisfying the following equation : _
i=1 ^ 24(_i^0)^2 + d_ijk ^0_i ^0_j ^0_k = 2m^2 .
the superscript ` 0 ' denotes the field value at its global minimum . using @xmath0 transformation , one can express the vev matrix @xmath46 in diagonal form with @xmath47 for the diagonal generators only .
now in order to break @xmath0 gauge symmetry into the sm gauge group @xmath48 , the vevs of all @xmath45 components must vanish except the one which is invariant under @xmath48 .
therefore , we choose @xmath49 to have a non - vanishing vev : @xmath50 where @xmath51 satisfies the following equation : v^2 - v^3 = 4 m^2.[vev ] here , we have used the fact that @xmath52 .
we can rewrite the scalar potential in eq .
( [ scalarpot - shift ] ) in terms of the dimensionless variablesy = , z = as , = = ( 1-y^2 + y^3 ) ^2 + 2 z^2 y^2 ( 1- 3y/2)^2 , where @xmath53 .
thus , for a constant value of @xmath54 , this potential has the following three extrema : y_1 & = & 0 , + y_2 & = & , + y_3 & = & - + & + & .
the first two extrema at @xmath55 and @xmath56 are independent of @xmath54 ( or @xmath57 ) and correspond to the ` standard ' and the ` shifted ' inflationary trajectories .
the extremum at @xmath55 is a local minimum ( maximum ) for @xmath58 , while the shifted extremum at @xmath56 is a local minimum ( maximum ) for @xmath59 .
these trajectories are shown in fig .
[ fig1 ] where we have plotted the dimensionless potential @xmath60 as a function of @xmath54 and @xmath61 for a typical value of @xmath62 .
the potential at @xmath56 is @xmath63 , which is lower than the potential @xmath64 at @xmath55 for @xmath65 .
inflation takes place when the system is trapped along the @xmath56 minimum .
also , we restrict ourselves to @xmath66 , so that the inflationary trajectory at @xmath56 can be realized before @xmath54 reaches zero . therefore , the interesting region for the parameter @xmath67 in our analysis is given by @xmath68 .
moreover , the @xmath0 gauge symmetry is always broken during the inflationary trajectory and hence no magnetic monopoles are produced at the end of inflation . versus @xmath61 and @xmath54 for @xmath69 .
the red curve at @xmath70 clearly shows one maximum at @xmath71 and two minima at @xmath72 ( standard trajectory ) and @xmath73 ( shifted trajectory ) .
these minima become maxima near @xmath74 as are shown by the brown curve.,width=377 ] along the shifted trajectory susy is broken due to the presence of a non - zero vacuum energy density @xmath75 .
this in turn generates the radiative corrections which can lift the flatness of the @xmath56 trajectory while providing the necessary slope for driving inflation . in order to calculate the one - loop radiative correction at @xmath56 we need to compute the mass spectrum of the model along this path where
both @xmath0 gauge symmetry and susy are broken . during inflation ,
the field @xmath31 acquires a vev in the @xmath76 direction which breaks the @xmath0 gauge symmetry down to sm gauge group @xmath48 .
perturbing around this vacuum @xmath77 and replacing @xmath78/\sqrt{2}$ ] , the potential in eq .
( [ scalarpot - shift ] ) yields the following masses for real scalar fields , @xmath79 and @xmath80:m^2_24_=^2m^2(-1)+^2|s|^2.the superpotential in eq .
( [ superpot - shift ] ) generates a weyl fermion with mass - squared : m_24 ^ 2=^2|s|^2 .
similarly , we obtain the following masses for the real scalar fields @xmath81 and @xmath82 with @xmath83 : m^2_i_=5 ^2m^2(-1)+25 ^2|s|^2,and from the following terms of the superpotentialw = s ( m^2 - _ i^2 - d_ii24 _ i^2 _ 24),the @xmath84 weyl fermions @xmath85 , @xmath83 , acquire a universal mass - squared : m^2_i= 25 ^2|s|^2 .
it is worth noting that the susy breaking along the inflationary trajectory , which is due to the non zero vacuum energy @xmath86 , generates a mass splitting in @xmath76 supermultiplet and in @xmath87 , @xmath83 , supermultiplets .
this is the only place where mass splitting between fermions and bosons appears .
the @xmath41-term contribution to the masses of scalar fields @xmath87 , @xmath88 , is obtained from the following term : g ^2 ( f^ijk _ j ^_k ) ( f^ilm _ l ^_m),where @xmath89 is the @xmath0 gauge coupling . as an example , we obtain the mass of @xmath90 field as follows : g ^2 ( f^92410 v_2 _ 10+f^91024 v_2 ^_10 ) ^2 , which leads to mass - squared @xmath91 using @xmath92 and @xmath93 for @xmath94 . thus , the d term contributes with a universal mass - squared @xmath91 for the above mentioned @xmath95 real scalar fields .
the mixing between chiral fermions @xmath85 , @xmath96 , and the gauginos @xmath97 , @xmath96 , gives rise to dirac mass term : i g f^ijk ( ^_k _ i _
j + |_k |_i _ j).this leads to @xmath95 dirac fermions with mass - squared @xmath98 .
finally , the gauge bosons @xmath99 , @xmath88 , acquire masses from the following term : g^2 f^ijk f^ilm a_^ja^l ^_k _
m. this generates a universal mass - squared @xmath100 for all @xmath95 gauge bosons . in table 1 , we summarize the results of the mass spectrum of the model along the shifted inflationary trajectory . as noted above , the mass splitting only occurs in @xmath76 and @xmath101 , @xmath83 , supermultiplets which contain 12 majorana fermions with two degrees of freedom and 24 real scalars , whereas there is no mass splitting in @xmath101 , @xmath88 , supermultiplets which consist of 12 massive dirac fermions with four degrees of freedom , 12 massive gauge bosons with 3 degrees of freedom , and 12 real scalars .
it can be readily checked that all these supermultiplets satisfy the supertrace rule @xmath102 .
the inflationary effective potential with 1-loop radiative correction is given by v_1 loop & = & ^2 m_^4 ( 1 + ) , [ vloop ] where f(m_^2,x^2 ) = ( ( x^4 + 1 ) + 2x^2 + 2 - 3 ) , @xmath103 , @xmath104 and @xmath105 is the renormalization scale .
the spectrum of the model at the @xmath0 breaking susy minimum is given by the massless octet @xmath101 , @xmath106 , and triplet @xmath107 , @xmath108 scalars / fermions , while the fields @xmath109 , @xmath110 , acquire mass - squared of order @xmath111 .
finally @xmath76 and @xmath25 fields acquire masses of order @xmath112 . as will be shown later
, these octets and triplets spoil the gauge coupling unification and we , therefore , need to add some vector - like matter to preserve unification . before discussing unification we consider the contribution from sugra corrections .
[ [ section-2 ] ] we take the following general form of khler potential : k & = & s
^2 + tr ^2 + h ^2 + |h^2 + & & + _
s + _ s h + _ s |h + & & + _ h + _ |h + _ h |h + & & + _ s + _ + _ h + _ |h + & & + _ ss + _ + _ h h + _ |h |h + , [ k ] where @xmath113 gev is the reduced planck mass . additionally , for the sake of simplicity , the contribution of many other terms e.g. , of the form c_2[tr(^2 ) + h.c . ]
+ c_3[tr(^3)/m_p + h.c . ] + is assumed to be zero .
alternatively we can effectively absorb these extra contributions coming from the @xmath31 superfield into various couplings of the above khler potential as only the @xmath57 field plays an active role during inflation .
the sugra scalar potential is given by @xmath114 with @xmath115 being the bosonic components of the superfields @xmath116 and where we have defined d_z_iw + m_p^-2w , k_i|j , and @xmath117 now in the inflationary trajectory with the d - flat direction ( @xmath118 , @xmath119 ) and using eqs .
( [ superpot - shift ] ) , ( [ vloop ] ) , ( [ k ] ) , and ( [ vf ] ) , we obtain the following form of the full potential : v & = & ^2m_^4 ( 1 + . + & & + ( -_sx^2 ) ( ) ^2 + & & . + ( . . + & & . .
+ + ) ( ) ^4 + ) + , [ vt ] where @xmath120 , @xmath104 and @xmath121 .
we restrict ourselves to @xmath122 and neglect contribution of soft susy breaking terms assuming soft masses of order 1 tev @xcite . before proceeding further ,
let us consider the possible mass contribution from the non - minimal terms of the khler potential .
in particular , as we will see , the @xmath6 triplet and @xmath5 octet multiplets remain massless as a consequence of both @xmath4 and @xmath0 gauge symmetries . to see this explicitly ,
consider the following general form of the fermionic mass matrix : ( m_f)_ij = e^k/2 ( w_ij + k_ijw + k_iw_j + k_j w_i + k_i k_j w - k^kk_ij d_k w ) .
since , in the susy minimum @xmath123 and @xmath124 essentially vanish due to r - symmetry , the triplet and octet multiplets therefore do not acquire masses from the above contribution .
the other possible contribution to fermionic masses come through the mixing of chiral fermions and gauginos , igf^ijk(_k_|lk_|lj_i + |_|lk_|lk|_i_j ) . due to @xmath0 gauge symmetry
, these terms also vanish for the triplet and octet multiplets .
( note that @xmath125 for the triplet and octet states ) .
therefore , both the fermionic and bosonic masses of these multiplets vanish as a consequence of @xmath4-symmetry and susy @xmath0 .
however , these multiplets can acquire tev scale masses as is discussed in more detail in the third section .
on the other hand , in the shifted trajectory case a non - zero mass contribution is expected from the nonminimal terms of the khler potential .
but these contributions are expected to have a negligible effect on the inflationary predictions as they appear inside the logarithmic functions of radiative corrections .
therefore , in numerical calculations we can safely ignore these contributions . before starting our discussion of this model
it is useful to recall here the basic results of the slow roll assumption .
the inflationary slow - roll parameters are given by = ( ) ^2 , = ( ) . in the slow - roll approximation ( i.e. @xmath126 , @xmath127 ) ,
the scalar spectral index @xmath128 and the tensor to scalar ratio @xmath2 are given ( to leading order ) by n_s 1 + 2 - 6 , r 16 .
the number of e - folds during inflation @xmath129 has crossed the horizon is given by @xmath130 where @xmath131 is the field value at the comoving scale @xmath132 , and @xmath133 denotes the value of field at the end of inflation , defined by @xmath134 ( or @xmath135 ) . during inflation , this scale exits the horizon at approximately @xmath136 where @xmath137 is the reheat temperature and for a numerical work we will set @xmath138 gev .
this could easily be reduced to lower values if the gravitino problem is regarded to be an issue .
the subscript ` 0 ' denotes the comoving scale corresponding to @xmath139 mpc@xmath140 .
the amplitude of the curvature perturbation is given by @xmath141 where @xmath142 is the wmap7 normalization at @xmath143 @xcite .
note that , for added precision , we include in our calculations the first order corrections @xcite in the slow - roll expansion for the quantities @xmath128 , @xmath2 , and @xmath144 . including sugra correction , in general ,
introduces a mass squared term of order @xmath145 , where @xmath146 is the hubble parameter .
this in turn makes the slow parameter @xmath147 and spoils inflation .
this is the well known @xmath148 problem .
however , in the case of supersymmetric shifted hybrid inflation with minimal khler potential this problem is naturally resolved due to a cancellation between the mass squared terms of the exponential factor and the other part of the potential in eq .
( [ vf ] ) .
the linear dependence of @xmath123 in @xmath25 due to @xmath4-symmetry guarantees this cancellation to all orders @xcite . with non - minimal khler potential the mass squared term can be approximated as @xmath149 which can spoil inflation for @xmath150 , but for reasonably natural values of parameters @xmath151 and @xmath152 we can obtain successful inflation consistent with wmap7 data .
the predicted values of various parameters for shifted hybrid inflation are displayed in figs .
[ fig2 ] , [ fig3 ] and [ fig4 ] . for minimal khler potential
the scalar spectral index @xmath153 lies outside the 1-@xmath17 bounds of wmap7 data . with non - minimal khler potential
only @xmath154 and @xmath155 play the important role of reducing the scalar spectral index to the central value of wmap7 data , i.e. @xmath18 . as can be seen in fig .
[ fig2 ] we can obtain @xmath128 within the 1-@xmath17 bounds of wmap7 data for @xmath156 or @xmath157 with all other non - minimal parameters equal to zero . the @xmath158 case has been considered previously in ref .
@xcite for standard and smooth hybrid inflation .
the results we obtain here are quite similar to the one obtained in ref .
@xcite .
in particular , the scalar spectral index is given by n_s 1 - 2_s + ( + 6_s x_0 ^ 2 ) ( ) ^2 - | ^2_x_0 f(5x_0 ^ 2 ) | ( ) ^2 . from fig .
[ fig3 ] and fig .
[ fig4 ] it is clear that the value of parameters @xmath159 , @xmath160 and @xmath161 increases with @xmath162 . therefore , the sugra contribution in the above expression raises the value of the scalar spectral index @xmath128 with @xmath162 .
the radiative correction , on the other hand , does not vary much with @xmath162 since both @xmath163 ( @xmath164 for large @xmath165 ) and @xmath166 tries to compensate the increase in @xmath167 term , in above expression . for small values of @xmath162 sugra correction
is suppressed as compared to radiative correction and @xmath168 factor , which reduces the scalar spectral index within the 1-@xmath17 bounds of wmap7 data . for @xmath169 case
we obtain following result for the scalar spectral index : n_s 1 + ( + 6(1 - 3_ss)x_0 ^ 2 ) ( ) ^2 - | ^2_x_0 f(5x_0 ^ 2 ) | ( ) ^2 . in this case with @xmath170 sugra term is positive and raises the value of @xmath128 with @xmath162 . for small values of @xmath162 ( or @xmath165 ) radiative correction becomes important and a small reduction in @xmath128 is observed as shown in fig .
[ fig2 ] . on the other hand , for @xmath171 ,
@xmath172 becomes negative and we obtain a reduction in @xmath128 with @xmath162 which is consistent with the wmap7 data ( see fig .
[ fig3 ] ) . with @xmath171 quadratic and quartic terms of the potential in eq .
( [ vt ] ) are positive and negative respectively .
this provides a nice example of hilltop inflation @xcite and a similar kind of potential has been analyzed in refs .
@xcite .
the value of the tensor to scalar ratio @xmath3 remains small in the above model of shifted hybrid inflation . for realizing observable @xmath2 values in supersymmetric hybrid inflationary models
see ref .
@xcite .
[ [ section-3 ] ] in this section we discuss the impact of the massless @xmath5 octet and @xmath6 triplet multiplets on the gauge coupling unification .
after @xmath0 breaking , these multiplets remain massless in the limit of exact supersymmtery and may spoil gauge coupling unification .
after inclusion of soft susy breaking mass terms taking account of @xmath173 tev , these particles acquire masses of order tev . in order to achieve gauge coupling unification
we can add suitable vectorlike particles with tev scale masses .
these vectorlike particles have recently been studied to solve the little hierarchy problem in the mssm @xcite .
the requirement that the three gauge couplings should remain perturbative at least up to the unification scale and the value of strong coupling should lie within the experimental uncertainties @xcite , greatly reduces the choices of vectorlike combinations .
furthermore , in order to avoid fast proton decay we do not consider the triplet scalar higgs vectorlike combination @xmath174 . taking into account these considerations
we choose the following combination of extra vectorlike particles : [ comb ] l(1,2,1/2)+(1,2,-1/2 ) + 2(e(1,1,1)+(1,1,-1 ) ) .
the sum of 1-loop beta function of octet @xmath175 , triplet @xmath176 and the above vectorlike combination is @xmath177 .
the 2-loop beta functions and rges for sm , mssm and these extra vectorlike particles can be found in refs .
@xcite .
gev ( left panels ) , @xmath178 tev ( right panels ) and @xmath179 .
dotted ( solid ) lines correspond to @xmath180 ( @xmath181 ) .
the masses of @xmath182 , @xmath123 and extra vectorlike particles are set equal to @xmath183 tev ( left panels ) and @xmath184 tev ( right panels).,title="fig:",width=302 ] gev ( left panels ) , @xmath178 tev ( right panels ) and @xmath179 . dotted ( solid ) lines correspond to @xmath180 ( @xmath181 ) .
the masses of @xmath182 , @xmath123 and extra vectorlike particles are set equal to @xmath183 tev ( left panels ) and @xmath184 tev ( right panels).,title="fig:",width=302 ] the evolution of three gauge couplings with and without the extra vectorlike particle combination ( eq .
( [ comb ] ) ) is shown in fig .
[ fig7 ] .
here we have used 2-loop rges with the susy breaking scale @xmath185 gev and @xmath178 tev and the masses of vectorlike particles @xmath183 tev and @xmath184 tev .
the value of the strong gauge coupling is fixed by the gauge unification condition and is required to lie within the experimental uncertainties @xcite .
the gut scale is found to lie in the range @xmath186 gev for @xmath187 tev and @xmath188 tev .
these extra particles may be detected at the lhc provided their masses are less than or of order a tev @xcite .
as an example , the gluon - gluon fusion channel can lead to @xmath5 octet pair production at the lhc : @xmath189 where @xmath190 .
this coupling can be generated from the kinetic energy term of @xmath31 .
[ [ section-4 ] ] to summarize , we have analyzed the adjoint field hybrid inflationary model in supersymmetric @xmath0 .
since the minimal susy @xmath0 hybrid inflation suffers from the monopole problem we have discussed @xmath0 shifted hybrid inflation . in this model
the @xmath0 gauge symmetry is broken during inflation and monopoles are inflated away before inflation ends . in minimal shifted hybrid inflation
the spectral index @xmath153 lies outside the wmap7 1-@xmath17 bounds with @xmath122 and tev scale soft susy breaking masses .
however , with a nonminimal khler potential the wmap7 1-@xmath17 region nicely compatible with shifted hybrid inflation .
a tiny value of @xmath3 is obtained in this model in these supersymmetric models see ref .
@xcite ] . furthermore , as a consequence of @xmath4-symmetry , the @xmath6 octet and @xmath5 triplet supermultiplets lie in the tev range .
in order to preserve gauge coupling unification we include additional tev scale vectorlike particles which may be observed at the lhc . |
particle production at forward rapidities in p(d)-a reactions probes partons in the target at low - x values . at sufficient high energy , large rapidities or large nucleii
the initial gluon distribution will saturate , and is expected to modify the particle pseudo rapidity densities , as well as the @xmath1 spectra of hadrons .
a theory based on qcd has been developed for dense low - x systems , termed the color glass condensate ( cgc ) @xcite .
this description has inspired much theoretical and experimental work , and was also a motivation for the brahms measurements of charged hadrons at forward rapidities in 200 gev d - au collisions at rhic .
the data reported here were all obtained with the brahms spectrometers .
the brahms forward spectrometer consists of 4 dipole magnets , 5 tracking chambers , two time - of - flight systems and a ring imaging chrenkov detector ( rich ) for particle identification . the angular coverage of the spectrometer extends from 2.3@xmath2 to 15@xmath2 with solid angle of 0.8 msr .
the mid - rapidity spectrometer covers angles from 40@xmath2 and 90@xmath2 .
details of experimental setup can be found in @xcite .
the collision vertex is determined from timing measurements done with a set of symmetricaly placed scintillator counters around the beam pipe at 1.5 , 4.15 and 6.7 meters @xcite .
the resolution of the vertex determination is @xmath3 cm .
this set of counters also provides the minimum bias normalization .
it is estimated that for pp collisions they record @xmath4 of the inelastic cross section .
additional details of the setup as well as the analysis can be found in @xcite where most of the data discussed here were first published .
spectra of charged hadrons in d - au and pp collisions at 200 gev are presented in fig . 1 for several pseudorapidities
. the data at @xmath5 and 1 are for the average of the positive and negatives charges , while the high rapidity data are for negative only . both the pp and dau cross sections becomes steeper with increasing @xmath6 .
the data collected from d - au collisions is compared to p - p using the nuclear modification factor defined as : @xmath7 . where @xmath8 is the number of binary collisions estimated to be equal to @xmath9 for minimum biased d+au collisions .
the pt dependence of the factor is shown in fig.[fig : ratio ] .
each panel shows the ratio calculated at a different @xmath6 value . at mid - rapidity ( @xmath10 )
, the nuclear modification factor exceeds 1 for transverse momenta greater than 2 gev / c in a similar , although less pronounced way as cronin s p+a measurements performed at lower energies @xcite .
nuclear modification factor for charged hadrons at pseudorapidities @xmath11 .
statistical errors are shown with error bars .
systematic errors are shown with shaded boxes with widths set by the bin sizes .
the shaded band around unity indicates the estimated error on the normalization to @xmath12 .
dashed lines at @xmath13 gev / c show the normalized charged particle density ratio @xmath14 . ]
a shift of one unit of rapidity is enough to make the cronin type enhancement disappear , and as the measurements are done at higher rapidities , the ratio becomes consistently smaller than 1 indicating a suppression in dau collisions compared to scaled pp systems at the same energy .
in all four panels , the statistical errors , shown as error bars ( vertical lines ) , are dominant specially in the most forward measurements .
these results have been described within the context of the color glass condensate @xcite ; the evolution of the nuclear modification factor with rapidity and centrality is consistent with a description of the au target where the rate of gluon fusion becomes comparable with that of gluon emmission as the rapidity increases and it slows down the overall growth of the gluon density .
the measured nuclear modification factor compares the slowed down growth of the numerator to a sum of incoherent p+p collisions , considered as dilute systems , whose gluon densities grow faster with rapidity because of the abscence of gluon fusion in dilute systems @xcite .
other explanations for the measured suppression have been proposed and they also reproduce the data @xcite . some of these other explanations particluar focus on the observation that the charged particle pseudo - rapidity density distributions exhibits a change in shape vs. centrality .
this is illustrated in fig.2 by the dashed lines at low @xmath15 where the ratios where obtained from the brahms data@xcite ( shown in fig.[fig : dndeta ] ) and the ua5 pp results@xcite .
thus already the overall soft spectrum shows suppression when going to forward angles .
the nuclear modification factor @xmath16 calculated for anti - protons ( filled squares ) and negative pions ( filled triangles ) at @xmath17 .
the same ratio calcutated for negative particles is shown with filled circles , and the systematic error for that measurement is shown as grey band .
, title="fig : " ] the nuclear modification factor of baryons is different from the one calculated with mesons , whenever the factor shows the so called cronin enhancement , baryons show a stronger enhancement . such difference , seen at lower energies , has also been found at rhic energies at all rapidities , in particular , fig .
[ fig : identifiedrda ] presents the minimum bias nuclear modification @xmath16 for anti - protons and negative pions at @xmath17 .
these ratios were obtained making use of ratios of raw counts of identified particles compared to those of charged particles in each @xmath18 bin .
this nuclear modification factor was calculated from the measured ratios between anti - protons and pions to changed particles for dau and pp and applying these factors to the minimum bias nuclear modification factor for negatively charged hadrons@xcite .
no attempt was made to estimate the contributions from anti - lambda feed down to the anti - proton result .
the remarkable difference between baryons and mesons has been related to parton recombination @xcite for heavy ion reactions .
it is though surprising that this effect in the da system at forward rapidities where only a small soft parton component should be able to account for this increase . in summary ,
particle production from dau and pp collisions at @xmath19 200 gev and at different rapidities with the brahms setup offers a window to the small - x components of the au wave function .
the suppression found in the particle production at high rapidities from d+au collisions may be the first indication of the onset of saturation in the gluon distribution function of the au target .
this work is supported by the division of nuclear physics of the office of science of the u.s .
department of energy under contract de - ac02 - 98-ch10886 , the danish natural science research council , the research council of norway , the jagiellonian university grants and the romanian ministry of education and research .
i thank r.debbe for providing valuable input to this talk and conbtribution .
l.mclerran and r.venugopalan,_phys .
rev . _ * d49 * 1994 , 2233 .
j.adamczyk et.al .
meth . _ * a499 * 2003 , 437 i. arsene _ et al . _
[ brahms collaboration ] , _ phys .
lett . _ * * 9**3 , 2004 , 242303 i. arsene at.al .
letter _ , * 94 * 2005 , 032301 d. antreasyan et . al .
rev . _ * d19 * 1979 , 764 | the rapidity dependence of nuclear modifications factor in d - au collisions at @xmath0 = 200 gev at rhic is discussed .
address = physics department , brookhaven national laboratory |
binary classification problems appear from diverse practical applications , such as , financial fraud detection , spam email classification , medical diagnosis with genomics data , drug response modeling , among many others . in these classification problems ,
the goal is to predict class labels based on a given set of variables .
suppose that we observe a training data set consisting of @xmath0 pairs , where @xmath1 , @xmath2 , and @xmath3 .
a classifier fits a discriminant function @xmath4 and constructs a classification rule to classify data point @xmath5 to either class @xmath6 or class @xmath7 according to the sign of @xmath8 .
the decision boundary is given by @xmath9 .
two canonical classifiers are linear discriminant analysis and logistic regression .
modern classification algorithms can produce flexible non - linear decision boundaries with high accuracy .
the two most popular approaches are ensemble learning and support vector machines / kernel machines . ensemble learning such as boosting @xcite and random forest @xcite combine many weak learners like decision trees into a powerful one . the support vector machine ( svm )
@xcite fits an optimal separating hyperplane in the extended kernel feature space which is non - linear in the original covariate spaces . in a recent extensive numerical study by @xcite ,
the kernel svm is shown to be one of the best among 179 commonly used classifiers .
motivated by data - piling " in the high - dimension - low - sample - size problems , @xcite invented a new classification algorithm named distance weighted discrimination ( dwd ) that retains the elegant geometric interpretation of the svm and delivers competitive performance . since then much work has been devoted to the development of dwd .
the readers are referred to @xcite for an up - to - date list of work on dwd . on the other hand
, we notice that dwd has not attained the popularity it deserves .
we can think of two reasons for that .
first , the current state - of - the - art algorithm for dwd is based on second - order - cone programming ( socp ) proposed in @xcite .
socp was an essential part of the dwd development .
as acknowledged in @xcite , socp was then much less well - known than quadratic programming , even in optimization .
furthermore , socp is generally more computationally demanding than quadratic programming .
there are two existing implementations of the socp algorithm : @xcite in matlab and @xcite in r. with these two implementations , we find that dwd is usually more time - consuming than the svm .
therefore , socp contributes to both the success and unpopularity of dwd .
second , the kernel extension of dwd and the corresponding kernel learning theory are under - developed compared to the kernel svm . although @xcite proposed a version of non - linear dwd by mimicking the kernel trick used for deriving the kernel svm , theoretical justification of such a kernel dwd is still absent . on the contrary ,
the kernel svm as well as the kernel logistic regression @xcite have mature theoretical understandings built upon the theory of reproducing kernel hilbert space ( rkhs ) @xcite .
most learning theories of dwd succeed to @xcite s geometric view of hdlss data and assume that @xmath10 and @xmath0 is fixed , as opposed to the learning theory for the svm where @xmath11 and @xmath12 is fixed .
we are not against the fixed @xmath0 and @xmath10 theory but it would be desirable to develop the canonical learning theory for the kernel dwd when @xmath12 is fixed and @xmath11 .
in fact , how to establish the bayes risk consistency of the dwd and kernel dwd was proposed as an open research problem in the original dwd paper @xcite .
nearly a decade later , the problem still remains open . in this paper , we aim to resolve the aforementioned issues .
we show that the kernel dwd in a rkhs has the bayes risk consistency property if a universal kernel is used .
this result should convince those who are less familiar with dwd to treat the kernel dwd as a serious competitor to the kernel svm . to popularize the dwd
, it is also important to allow practitioners to easily try dwd collectively with the svm in real applications . to this end
, we develop a novel fast algorithm to solve the linear and kernel dwd by using the majorization - minimization ( mm ) principle . compared with the socp algorithm ,
our new algorithm has multiple advantages .
first , our algorithm is much faster than the socp algorithm . in some examples ,
our algorithm can be several hundred times faster .
second , dwd equipped with our algorithm can be faster than the svm .
third , our algorithm is easier to understand than the socp algorithm , especially for those who are not familiar with semi - definite and second - order - cone programming .
this could help demystify the dwd and hence may increase its popularity . to give a quick demonstration
, we use a simulation example to compare the kernel dwd and the kernel svm .
we drew 10 centers @xmath13 from @xmath14 . for each data point in the positive class , we randomly picked up a center @xmath15 and then generated the point from @xmath16 .
the negative class was assembled in the same way except that 10 centers @xmath17 were drawn from @xmath18 .
for this model the bayes rule is nonlinear . ]
figure [ fig : svm_dwd ] displays the training data from the simulation model where 100 observations are from the positive class ( plotted as triangles ) and another 100 observations are from the negative class ( plotted as circles ) . we fitted the svm and dwd using gaussian kernels .
we have implemented our new algorithm for dwd in a publicly available r package ` kerndwd ` .
we computed the kernel svm by using the r package ` kernlab ` @xcite .
we recorded their training errors and test errors . from figure
[ fig : svm_dwd ] , we observe that like the kernel svm , the kernel dwd has a test error close to the bayes error , which is consistent with the bayes risk consistency property of the kernel dwd established in section [ sec : learningtheory0 ] .
notably , the kernel dwd is about three times as fast as the kernel svm in this example .
the rest of the paper is organized as follows .
to be self - contained , we first review the svm and dwd in section [ sec : review ] .
we then derive the novel algorithm for dwd in section [ sec : computation ] .
we introduce the kernel dwd in a reproducing kernel hilbert space and establish the learning theory of kernel dwd in section [ sec : kerdwd ] .
real data examples are given in section [ sec : real_data ] to compare dwd and the svm .
technical proofs are provided in the appendix .
the introduction of the svm usually begins with its geometric interpretation as a maximum margin classifier @xcite . consider a case when two classes are separable by a hyperplane @xmath19 such that @xmath20 are all non - negative . without loss of generality , we assume that @xmath21 is a unit vector , i.e. , @xmath22 , and we observe that each @xmath23 is equivalent to the euclidean distance between the data point @xmath24 and the hyperplane .
the reason is that @xmath25 and @xmath26 , where @xmath27 is any data point on the hyperplane and @xmath21 is the unit normal vector .
the svm classifier is defined as the optimal separating hyperplane that maximizes the smallest distance of each data point to the separating hyperplane .
mathematically , the svm can be written as the following optimization problem ( for the separable data case ) : @xmath28 the smallest distance @xmath29 is called the _ margin _ , and the svm is thereby regarded as a _ large - margin classifier_.
the data points closest to the hyperplane , i.e. , @xmath30 , are dubbed the _ support vectors_. in general , the two classes are not separable , and thus @xmath31 can not be non - negative for all @xmath32 . to handle this issue ,
non - negative slack variables @xmath33 , are introduced to ensure all @xmath34 to be non - negative . with these slack variables ,
the optimization problem is generalized as follows , @xmath35 to compute svms , the optimization problem is usually rephrased as an equivalent quadratic programming ( qp ) problem , @xmath36,\\ \text { subject to } & \;\;\ ; y_i(\beta_0 + { \boldsymbol } x_i^t{\boldsymbol } \beta ) + \xi_i \ge 1,\ \xi_i \ge 0 , \
\forall i , \label{eq : quad_svm } \end{aligned}\ ] ] and it can be solved by maximizing its lagrange dual function , @xmath37 , \\
\text { subject to } & \;\;\ ; \mu_i \ge 0 \text { and } \sum_{i=1}^n \mu_i y_i = 0 .
\label{eq : dual } \end{aligned}\ ] ] by solving , one can show that the solution of has the form @xmath38 @xmath39 being zero only when @xmath24 lies on the support vectors .
one widely used method to extend the linear svm to non - linear classifiers is the kernel method @xcite , which replaces the dot product @xmath40 in the lagrange dual problem with a kernel function @xmath41 , and hence the solution has the form @xmath42 some popular examples of the kernel function @xmath43 include : @xmath44 ( linear kernel ) , @xmath45 ( polynomial kernel ) , and @xmath46 ( gaussian kernel ) , among others
. distance weighted discrimination was originally proposed by @xcite to resolve the _ data - piling _ issue .
@xcite observed that many data points become support vectors when the svm is applied on the so - called high - dimension - low - sample - size ( hdlss ) data , and @xcite coined the term data - piling to describe this phenomenon .
we delineate it in figure [ fig : dp ] through a simulation example .
let @xmath47 be a @xmath48-dimension vector .
we generated @xmath49 points ( indexed from @xmath6 to @xmath49 and represented as triangles ) from @xmath50 as the negative class and another @xmath49 points ( indexed from @xmath51 to @xmath52 and represented as circles ) from @xmath53 as the positive class .
we computed @xmath54 and @xmath55 for svm . in the left panel of figure [ fig : dp ] , we plotted @xmath56 for each data point , and we portrayed the support vectors by solid triangles and circles .
we observe that @xmath57 out of @xmath52 data points become support vectors .
the right panel of figure [ fig : dp ] corresponds to dwd ( will be defined shortly ) , where data - piling is attenuated .
a real example revealing the data - piling can be seen in figure 1 of @xcite .
are plotted for svm and dwd .
indices 1 to 50 represent negative class ( triangles ) and indices 51 to 100 are for positive class ( circles ) . in the left panel ,
data points belonging to the support vectors are depicted as solid circles and triangles . ]
@xcite viewed data - piling " as a drawback of the svm , because the svm classifier is a function of only support vectors .
another popular classifier logistic regression does classification by using all the data points .
however , the classical logistic regression classifier is derived by following the maximum likelihood principle , not based on a nice margin - maximization motivation penalized logistic regression approaches the margin - maximizing hyperplane for the separable data case .
dwd was first proposed in 2002 . ] .
@xcite wanted to have a new method that is directly formulated by a svm - like margin - maximization picture and also uses all data points for classification .
to this end , @xcite proposed dwd which finds a separating hyperplane minimizing the total inverse margins of all the data points : @xmath58,\\ \text { subject to } & \;\;\ ; d_i = y_i(\omega_0 + { \boldsymbol } x_i^t{\boldsymbol } \omega ) + \eta_i \ge 0 , \ \eta_i \ge 0,\ \forall i , \text { and } { \boldsymbol } \omega^t { \boldsymbol } \omega=1 . \end{aligned } \label{eq : nonsep_dwd}\ ] ]
there has been much work on variants of the standard dwd .
we can only give an incomplete list here .
@xcite introduced the weighted dwd to tackle unequal cost or sample sizes by imposing different weights on two classes . @xcite
extended the binary dwd to the multiclass case .
@xcite proposed the sparse dwd for high - dimensional classification .
in addition , the work connecting dwd with other classifiers , e.g. , svm , includes but not limited to lum @xcite , dwsvm @xcite , and flame @xcite .
@xcite provided a more comprehensive review of the current dwd literature .
@xcite solved the standard dwd by reformulating as a second - order cone programming ( socp ) program @xcite , which has a linear objective , linear constraints , and second - order - cone constraints .
specifically , for each @xmath59 , let @xmath60 , @xmath61 , and then @xmath62 , @xmath63 , and @xmath64 .
hence the original optimization problem becomes @xmath65,\\ \text { subject to } & \;\;\ ; { \boldsymbol } \rho - { \boldsymbol } \sigma = \tilde{{\boldsymbol } y } { \boldsymbol } x { \boldsymbol } \omega + \omega_0 \cdot { \boldsymbol } y + { \boldsymbol } \eta , \\ & \;\;\ ; \eta_i \ge 0 , \
( \rho_i ; \sigma_i , 1 ) \in s_3 , \
\forall i , \ ( 1 ; { \boldsymbol } \omega)\in s_{p+1 } , \end{aligned } \label{eq : primal_dwd}\ ] ] where @xmath66 is an @xmath67 diagonal matrix with the @xmath59th diagonal element @xmath68 , @xmath69 is an @xmath70 data matrix with the @xmath59th row @xmath71 , and @xmath72 is the form of the second - order cones . after solving @xmath73 and @xmath74 from , a new observation @xmath75
is classified by @xmath76 .
note that the kernel svm was derived from applying the kernel trick to the dual formulation ( [ eq : soln_svm ] ) .
@xcite followed the same approach to consider a version of kernel dwd for achieving non - linear classification .
the dual function of the problem is @xcite @xmath77,\\ \text { subject to } & \;\;\ ; { \boldsymbol } y^t { \boldsymbol } \alpha = 0 , \ { \boldsymbol } 0 \le { \boldsymbol } \alpha \le c\cdot { \boldsymbol } 1 , \end{aligned } \label{eq : dual_dwd}\ ] ] where @xmath78 .
note that only uses @xmath79 , which makes it easy to employ the kernel trick to get a nonlinear extension of the linear dwd . for a given kernel function @xmath43 ,
define the kernel matrix as @xmath80 , @xmath81 .
then a kernel dwd can be defined as @xcite @xmath82,\\ \text { subject to } & \;\;\ ; { \boldsymbol } y^t { \boldsymbol } \alpha = 0 , \ { \boldsymbol } 0 \le { \boldsymbol } \alpha \le c\cdot { \boldsymbol } 1 .
\end{aligned } \label{eq : dual_kerdwd}\ ] ] to solve ( [ eq : dual_kerdwd ] ) , @xcite used the cholesky decomposition of the kernel matrix , i.e. , @xmath83 and then replaced the predictors @xmath69 in with @xmath84 .
@xcite also carefully discussed several algorithmic issues that ensure the equivalent optimality in and . * remark 1 . *
two dwd implementations have been published thus far : a matlab software @xcite and an r package ` dwd ` @xcite .
both implementations are based on a matlab socp solver ` sdpt3 ` , which was developed by @xcite .
we notice that the r package ` dwd ` can only compute the linear dwd . * remark 2 . * to our best knowledge ,
the theoretical justification for the kernel dwd in @xcite is still unclear .
the reason is likely due to the fact that the nonlinear extension is purely algorithmic .
in fact , the bayes risk consistency of dwd was proposed as an open research problem in @xcite .
the kernel dwd considered in this paper can be rigorously justified to have a universal bayes risk consistency property ; see details in section [ sec : learningtheory0 ] .
@xcite also attempted to replace the reciprocal in the dwd optimization problem with the @xmath85th power ( @xmath86 ) of the inverse distances , and @xcite also used it as the original definition of dwd .
we name the dwd with this new formulation the generalized dwd : @xmath87,\\ \text { subject to } & \;\;\ ; d_i = y_i(\omega_0 + { \boldsymbol } x_i^t{\boldsymbol } \omega ) + \eta_i \ge 0 , \ \eta_i
\forall i , \text { and } { \boldsymbol } \omega^t { \boldsymbol } \omega=1 , \label{eq : gendwd } \end{aligned}\ ] ] which degenerates to the standard dwd when @xmath88 . the first asymptotic theory for dwd and generalized dwd was given in @xcite who presented a novel geometric representation of the hdlss data .
assuming @xmath89 are the data from the positive class and @xmath90 are from the negative class .
@xcite stated that , when the sample size @xmath0 is fixed and the dimension @xmath12 goes to infinity , under some regularity conditions , there exist two constants @xmath91 and @xmath92 such that for each pair of @xmath59 and @xmath93 , @xmath94 as @xmath10 .
this result was applied the results to study several classifiers including the svm and the generalized dwd . for ease presentation
let us consider the equal subgroup size case , i.e. , @xmath95 .
@xcite assumed that @xmath96 the basic conclusion is that when @xmath97 is greater than a threshold that depends on @xmath98 , the misclassification error converges to zero , and when @xmath97 is less than the same threshold , the misclassification error converges to @xmath99 . for more details , see theorem 1 and theorem 2 in @xcite .
@xcite further relaxed the assumptions thereof . * remark 3 .
* the generalized dwd has not been implemented yet because the socp transformation only works for the standard dwd ( @xmath88 ) , but its extension to handle the general cases is unclear if not impossible .
that is why the current dwd literature only focuses on dwd with @xmath88 .
in fact , the generalized dwd with @xmath100 was proposed as an open research problem in @xcite .
the new algorithm proposed in this paper can easily solve the generalized dwd problem for any @xmath101 ; see section [ sec : computation ] .
@xcite originally solved the standard dwd by transforming into a socp problem .
this algorithm , however , can not compute the generalized dwd with @xmath100 . in this section
, we propose an entirely different algorithm based on the majorization - minimization ( mm ) principle .
our new algorithm offers a unified solution to the standard dwd and the generalized dwd .
our algorithm begins with a @xmath102 formulation of the dwd .
lemma [ lm : dwd_loss ] deploys the result .
note that the loss function also lays the foundation of the kernel dwd learning theory that will be discussed in section [ sec : kerdwd ] .
the generalized dwd classifier in can be written as @xmath103 , where @xmath104 is computed from @xmath105 , \label{eq : dwdlossopt}\ ] ] for some @xmath106 , where @xmath107 [ lm : dwd_loss ] * remark 4 . * the proof of lemma 1 provides the one - to - one mapping between @xmath106 in and @xmath108 in .
write @xmath109 as the solution to .
define @xmath110 considering using @xmath111 , @xmath112,\\ \text { subject to } & \ ; d_i = y_i(\omega_0 + { \boldsymbol } x_i^t{\boldsymbol } \omega ) + \eta_i \ge 0 , \
\eta_i \ge 0 , \
\forall i , \text { and } { \boldsymbol } \omega^t
{ \boldsymbol } \omega=1 , \label{eq : gendwd2 } \end{aligned}\ ] ] we have @xmath113 note that @xmath114 , which means that the generalized dwd classifier defined by is equivalent to the generalized dwd classifier defined by . by lemma [ lm : dwd_loss ] , we call @xmath115 the generalized dwd loss
. it can be visualized in figure [ fig : dwdloss ] .
we observe that the generalized dwd loss decreases as @xmath85 increases and it approaches the svm hinge loss function as @xmath116 .
when @xmath88 , the generalized dwd loss becomes @xmath117 we notice that @xmath118 has appeared in the literature @xcite . in this work
we give a unified treatment of all @xmath85 values , not just @xmath88 .
, and the svm hinge loss . ]
we now show how to develop the new algorithm by using the mm principle @xcite .
some recent successful applications of the mm principle can be seen in @xcite , among others .
the main idea of the mm principle is easy to understand .
suppose @xmath119 and we aim to minimize @xmath120 , defined in .
the mm principle finds a majorization function @xmath121 satisfying @xmath122 for any @xmath123 and @xmath124 , and then we generate a sequence @xmath125 by updating @xmath126 via @xmath127 .
we first expose some properties of the generalized dwd loss functions , which give rise to a quadratic majorization function of @xmath120 .
the generalized dwd loss is differentiable everywhere ; its first - order derivative is given below ,
@xmath128 the generalized dwd loss function
@xmath115 has a lipschitz continuous gradient , @xmath129 which further implies a quadratic majorization function of @xmath115 such that @xmath130 for any @xmath131 and @xmath132 .
[ lm : lps ] denote the current solution by @xmath133 and the updated solution by @xmath119 .
we settle @xmath134 and @xmath135 without abusing notations .
we have that for any @xmath136 , @xmath137 \\ & + \frac{m}{2n } \sum_{i=1}^n \left[y_i(\beta_0 - \tilde\beta_0 ) + y_i { \boldsymbol } x_i^t ( { \boldsymbol } \beta - \tilde{{\boldsymbol } \beta})\right]^2 + \lambda { \boldsymbol } \beta^t { \boldsymbol } \beta \\ \equiv & { \boldsymbol } d(\beta_0 , { \boldsymbol } \beta ) .
\label{eq : cled } \end{split}\ ] ] we now find the minimizer of @xmath138 .
the gradients of @xmath138 are given as follows : @xmath139{\boldsymbol } x_i + 2\lambda { \boldsymbol } \beta \notag\\ = & { \boldsymbol } x^t { \boldsymbol } z + \frac{m}{n}(\beta_0 - \tilde{\beta}_0){\boldsymbol } x^t { \boldsymbol } 1 + \frac{m}{n}\sum_{i=1}^n { \boldsymbol } x_i { \boldsymbol } x_i^t ( { \boldsymbol } \beta - \tilde{{\boldsymbol } \beta } ) + 2\lambda { \boldsymbol } \beta \notag\\ = & { \boldsymbol } x^t { \boldsymbol } z + \frac{m}{n}(\beta_0 - \tilde{\beta}_0){\boldsymbol } x^t { \boldsymbol } 1 + \left(\frac{m}{n } { \boldsymbol } x^t { \boldsymbol } x + 2\lambda{\boldsymbol } i_p\right ) ( { \boldsymbol } \beta - \tilde{{\boldsymbol } \beta } ) + 2\lambda \tilde{{\boldsymbol } \beta}\label{eq : grdnt_alp } , \\ \partial\frac{{\boldsymbol } d(\beta_0 , { \boldsymbol } \beta)}{\partial \beta_0 } = & \frac{1}{n } \sum_{i=1}^n v_q'\left(y_i(\tilde{\beta}_0 + { \boldsymbol } x_i ^t \tilde{{\boldsymbol } \beta})\right ) y_i + \frac{m}{n } \sum_{i=1}^n \left[(\beta_0 - \tilde{\beta}_0 ) + { \boldsymbol } x_i^t({\boldsymbol } \beta - \tilde{{\boldsymbol } \beta})\right ] \notag\\ = & { \boldsymbol } 1^t { \boldsymbol } z + m(\beta_0 - \tilde{\beta}_0 ) + \frac{m}{n } { \boldsymbol } 1^t { \boldsymbol } x ( { \boldsymbol } \beta - \tilde{{\boldsymbol } \beta } ) .
\label{eq : grdnt_b0}\end{aligned}\ ] ] where @xmath69 is the @xmath70 data matrix with the @xmath59th row @xmath71 , @xmath140 is an @xmath141 vector with the @xmath59th element @xmath142 , and @xmath143 is the vector of ones . setting @xmath144 $ ] to be zeros , we obtain the minimizer of @xmath138 : @xmath145 so far we have completed all the steps of the mm algorithm .
details are summarized in algorithm [ alg : linear ] .
initialize @xmath146 compute @xmath147 : @xmath148 compute @xmath149 : @xmath150 compute : @xmath151 set @xmath146 = @xmath152 we have implemented algorithm [ alg : linear ] in an r package ` kerndwd ` , which is publicly available for download on cran . in this section ,
we show the superior computation performance of our r implementation , ` kerndwd ` , over the two existing implementations , the r package ` dwd ` @xcite and the matlab software @xcite . to avoid confusion
, we henceforth use ` ours ` , ` huang ` , and ` marron ` to denote ` kerndwd ` , ` dwd ` , and the matlab implementation , respectively .
since ` huang ` is incapable of non - linear kernels and the generalized dwd with @xmath153 , we only attend to the linear dwd with @xmath85 fixed to be one .
all experiments were conducted on an intel core i5 m560 ( 2.67 ghz ) processor . for a fair comparison ,
we study the four numerical examples used in @xcite , except for different sample sizes and dimensions . in each example
, we generate a data set with sample size @xmath154 and dimension @xmath155 .
the responses are always binary ; one half of the data have responses @xmath156 and the other half have @xmath7 .
data in example 1 are generated from gaussian distribution with means of @xmath157 and an identity covariance for @xmath158 classes respectively .
example 2 has 80% of data drawn as example 1 whereas the other 20% from gaussian distributions with means of @xmath159 for @xmath158 classes . in example 3 , 80% of the data are obtained as example 1 as well , while the means of the remaining 20% have the first coordinate replaced by @xmath160 and one randomly chosen coordinate replaced by @xmath161 for @xmath158 classes .
for example 4 , at the first 25 coordinates , the data from @xmath7 class are standard gaussian and the data from @xmath156 class are @xmath162 times standard gaussian ; for both classes , the last 25 coordinates are just the squares of the first 25 . in each example , we fitted a linear dwd with five different tuning parameter values @xmath163 . after obtaining @xmath164 , we computed @xmath165 ) and the constant @xmath108 in by using remark 4 .
we then used ` huang ` and ` marron ` to compute their solutions .
note that in theory all three implementations should yield identical @xmath166 from table [ tab : simu ] we observe that ` ours ` took remarkably less computation time than ` huang ` and ` marron ` .
in example 1 , for instance , ` ours ` spent only 0.012 second on average to fit a dwd model , while ` huang ` used 14.525 seconds , and ` marron ` took 2.204 seconds , which were 1210 and 183 times larger , respectively . in all four examples , the timings of ` ours ` were 700 times above faster than the existing r implementation ` huang ` , and also more than 70 times faster than the matlab implementation ` marron ` .
the kernel svm can be derived by using the kernel trick or using the view of non - parametric function estimation in a reproducing kernel hilbert space ( rkhs ) .
much of the theoretical work on the kernel svm is based on the rkhs formulation of svms .
the derivation of the kernel svm in a rkhs is given in @xcite .
we take a similar approach to derive the kernel dwd , as our goal is to establish the kernel learning theory for dwd .
consider @xmath167 , a reproducing kernel hilbert space generated by the kernel function @xmath43 .
the mercer s theorem ensures @xmath43 to have an eigen - expansion @xmath168 , with @xmath169 and @xmath170 .
then the hilbert space @xmath167 is defined as the collection of functions @xmath171 , for any @xmath172 such that @xmath173 , and the inner product is @xmath174 .
given @xmath167 , let the non - linear dwd be written as @xmath175 where @xmath176 is the solution of @xmath177 , \label{eq : nonlnr_dwd1}\ ] ] where @xmath115 is the generalized dwd loss .
the representer theorem concludes that the solution of has a finite expansion based on @xmath178 @xcite , @xmath179 and thus @xmath180 consequently , can be paraphrased with matrix notation , @xmath181 , \label{eq : nonlnr_dwd}\ ] ] where @xmath182 is the kernel matrix with the @xmath183th element of @xmath184 and @xmath185 is the @xmath59th column of @xmath182 .
* remark 5 .
* we can compare ( [ eq : nonlnr_dwd ] ) to the kernel svm @xcite @xmath186_+ + \lambda { \boldsymbol } \alpha^t { \boldsymbol } k { \boldsymbol } \alpha \right ] , \label{eq : optmztn}\ ] ] where @xmath187_+$ ] is the hinge loss underlying the svm . as shown in figure 3
, the generalized dwd loss takes the hinge loss as its limit when @xmath188 . in general , the generalized dwd loss and the hinge loss look very similar , which suggests that the kernel dwd and the kernel svm equipped with the same kernel have similar statistical behavior . the procedure for deriving algorithm [ alg : linear ] for the linear dwd
can be directly adopted to derive an efficient algorithm for solving the kernel dwd .
we obtain the majorization function @xmath189 , @xmath190+\lambda { \boldsymbol } \alpha^t { \boldsymbol } k { \boldsymbol } \alpha \\ & & + \frac{m}{2n } \sum_{i=1}^n \left[y_i(\beta_0 - \tilde\beta_0 ) + y_i { \boldsymbol } k_i^t ( { \boldsymbol } \alpha - \tilde{{\boldsymbol } \alpha})\right]^2 + \frac{1}{n } \sum_{i=1}^n v_q\left(y_i(\tilde{\beta}_0 + { \boldsymbol } k_i ^t \tilde{{\boldsymbol } \alpha})\right ) \end{aligned}\ ] ] and then find the minimizer of @xmath189 which has a closed - form expression .
we opt to omit the details here for space consideration .
algorithm [ alg : kernel ] summarizes the entire algorithm for the kernel dwd .
initialize @xmath191 compute @xmath147 : @xmath192 compute @xmath149 : @xmath193 compute : @xmath194 set @xmath191 = @xmath195 @xcite formulated the kernel svm as a non - parametric function estimation problem in a reproducing kernel hilbert space and showed that the population minimizer of the svm loss function is the bayes rule , indicating that the svm directly approximates the optimal bayes classifier .
@xcite further coined a name fisher consistency " to describe such a result .
the vapnik - chervonenkis ( vc ) analysis @xcite and the margin analysis @xcite have been used to bound the expected classification error of the svm .
@xcite used the so - called leave - one - out analysis @xcite to study a class of kernel machines .
the exisiting theoretical work on the kernel svm provides us a nice road map to study the kernel dwd . in this section
we first elucidate the fisher consistency @xcite of the generalized kernel dwd , and we then establish the bayes risk consistency of the kernel dwd when a universal kernel is employed . let @xmath196 denote the conditional probability @xmath197 . under the 0 - 1 loss ,
the theoretical optimal bayes rule is @xmath198 .
assume @xmath196 is a measurable function and @xmath199 throughout .
the population minimizer of the expected generalized dwd loss @xmath200 $ ] is @xmath201 , \label{eq : fisher_min}\ ] ] where @xmath202 is the indicator function .
the population minimizer @xmath203 has the same sign as @xmath204 .
[ lm : fisher ] fisher consistency is a property of the loss function .
the interpretation is that the generalized dwd can approach bayes rule with infinite many samples .
we notice that fisher consistency of @xmath118 has been shown before @xcite . in reality
all classifiers are estimated from a finite sample .
thus , a more refined analysis of the actual dwd classifier is needed , and that is what we achieve in the following . following the convention in the literature , we absorb the intercept into @xmath205 and
present the kernel dwd as follows : @xmath206 .
\label{eq : ker_dwd}\ ] ] the ultimate goal is to show that the misclassification error of the kernel dwd approaches the bayes error rate such that we can say the kernel dwd classifier works as well as the bayes rule ( asymptotically speaking ) .
following @xcite , we derive the following lemma . for a discrimination function @xmath4
, we define @xmath207 .
$ ] assume that @xmath208 is the bayes rule and @xmath209 is the solution of , then @xmath210 where @xmath211 and @xmath212 are defined as follows and @xmath213 is the generalized dwd loss , @xmath214 - e_{{\boldsymbol } xy}\bigg[v_q \left(y\tilde{f}({\boldsymbol } x)\right)\bigg],\\ \varepsilon_e=\varepsilon_e(\hat{f}_n ) & = e_{{\boldsymbol } xy}\bigg[v_q \left(y\hat{f}_n({\boldsymbol } x)\right)\bigg ] - \inf_{f\in{\mathcal{h}}_k } e_{{\boldsymbol } xy}\bigg[v_q(yf({\boldsymbol } x))\bigg ] .
\end{aligned } \label{eq : two_err}\ ] ] [ lm : err_bd ] in the above lemma @xmath215 is the bayes error rate and @xmath216 is the misclassification error of the kernel dwd applied to new data points .
if @xmath217 , we say the classifier is bayes risk consistent .
based on lemma [ lm : err_bd ] , it suffices to show that both @xmath211 and @xmath212 approach zero in order to demonstrate the bayes risk consistency of the kernel dwd .
note that @xmath211 is deterministic and is called the approximation error . if the rkhs is rich enough then the approximation error can be made arbitrarily small . in the literature ,
the notation of _ universal kernel _
@xcite has been proposed and studied .
suppose @xmath218 is the compact input space of @xmath69 and @xmath219 is the space of all continuous functions @xmath220 .
the kernel @xmath43 is said to be _
universal _ if the function space @xmath221 generated by @xmath43 is dense in @xmath222 , that is , for any positive @xmath223 and any function @xmath224 , there exists an @xmath225 such that @xmath226 .
suppose @xmath209 is the solution of , @xmath221 is induced by a universal kernel @xmath43 , and the sample space @xmath227 is compact .
then we have * @xmath228 ; * let @xmath229 .
when @xmath230 and @xmath231 , for any @xmath232 , @xmath233 by ( 1 ) and ( 2 ) and we have @xmath234 in probability .
[ lm : thm ] the gaussian kernel is universal and @xmath235 .
thus theorem [ lm : thm ] says that the kernel dwd using the gaussian kernel is bayes risk consistent .
this offers a theoretical explanation to the numerical results in figure [ fig : svm_dwd ] .
in this section , we investigate the performance of ` kerndwd ` on four benchmark data sets : the bupa liver disorder data , the haberman s survival data , the connectionist bench ( sonar , mines vs. rocks ) data , and the vertebral column data .
all the data sets were obtained from uci machine learning repository @xcite . for comparison purposes , we considered the svm , the standard dwd ( @xmath88 ) and the generalized dwd models with @xmath236 .
we computed all dwd models using our r package ` kerndwd ` and solved the svm using the r package ` kernlab ` @xcite .
we randomly split each data into a training and a test set with a ratio @xmath237 . for each method using the linear kernel
, we conducted a five - folder cross - validation on the training set to tune @xmath106 . for each method using gaussian kernels ,
the pair of @xmath238 was tuned by the five - folder cross - validation .
we then fitted each model with the selected @xmath106 and evaluated its prediction accuracy on the test set .
table [ tab : realdata ] displays the average timing and mis - classification rates .
we do not argue that either svm or dwd outperforms the other ; nevertheless , two models are highly comparable .
svm models work better on sonar and vertebral data , and dwd performs better on bupa and haberman data . for three out of the four data sets
, the best method uses a gaussian kernel , indicating that linear classifiers may not be adequate in such cases . in terms of timing , ` kerndwd ` runs faster than ` kernlab ` in all these examples .
it is also interesting to see that dwd with @xmath239 can work slightly better than dwd with @xmath88 on bupa and haberman data , although the difference is not significant .
in this paper we have developed a new algorithm for solving the linear generalized dwd and the kernel generalized dwd .
compared with the current state - of - the - art algorithm for solving the linear dwd , our new algorithm is easier to understand , more general , and much more efficient .
dwd equipped with the new algorithm can be computationally more efficient than the svm .
we have established the statistical learning theory of the kernel generalized dwd , showing that the kernel dwd and the kernel svm are comparable in theory .
our theoretical analysis and algorithm do not suggest dwd with @xmath88 has any special merit compared to the other members in the generalized dwd family .
numerical examples further support our theoretical conclusions .
dwd with @xmath88 is called the standard dwd purely due to the fact that it , not other generalized dwds , can be solved by socp when the dwd idea was first proposed . now with our new algorithm and theory , practitioners have the option to explore different dwd classifiers . in the present paper
we have considered the standard classification problem under the 0 - 1 loss . in
many applications we may face the so - called non - standard classification problems .
for example , observed data may be collected via biased sampling and/or we need to consider unequal costs for different types of mis - classification . @xcite
introduced a weighted dwd to handle the non - standard classification problem , which follows the treatment of the non - standard svm in @xcite .
@xcite defined the weighted dwd as follows , @xmath240 , \text { subject to } r_i = y_i(\beta_0 + { \boldsymbol } x_i^t{\boldsymbol } \beta)+\xi_i \ge 0 \text { and } { \boldsymbol } \beta^t { \boldsymbol } \beta=1 , \end{aligned } \label{eq : wdwd}\ ] ] which can be further generalized to the weighted kernel dwd : @xmath241 .
\label{eq : kerwtgendwd}\ ] ] @xcite gave the expressions for @xmath242 for various non - standard classification problems .
@xcite solved the weighted dwd with @xmath88 based on the second - order - cone programming .
the mm procedure for algorithm [ alg : linear ] and algorithm [ alg : kernel ] can easily accommodate the weight factors @xmath242 s to solve the weighted dwd and weighted kernel dwd .
we have implemented the weighted dwd in the r package ` kerndwd ` .
* proof of lemma [ lm : dwd_loss ] * write @xmath243 and @xmath244 .
the objective function of can be written as @xmath245 .
we next minimize over @xmath246 for every fixed @xmath59 by computing the first - order and the second - order derivatives of @xmath247 : @xmath248 if @xmath249 , then @xmath250 for all @xmath251 , and @xmath252 is the minimizer .
if @xmath253 , then @xmath254 is the minimizer as @xmath255 and @xmath256 . by plugging in the minimizer @xmath257 into @xmath245
, we obtain @xmath258 where @xmath259 we now simplify .
suppose @xmath260 and @xmath261 .
we define @xmath262 for each @xmath85 , @xmath263 by setting @xmath264 and @xmath265 , we find that becomes @xmath266 which can be further transformed to with @xmath106 and @xmath267 one - to - one correspondent . if both @xmath273 and @xmath278 , @xmath281 .
it is trivial that @xmath282 by , , and , we prove .
we now prove .
let @xmath283 from , it is not hard to show that @xmath284 is strictly increasing . therefore @xmath285 is a strictly convex function , and its first - order condition , @xmath286 verifies directly .
* proof of lemma [ lm : fisher ] * given that @xmath287 , we have that @xmath288 \equiv e_{{\boldsymbol } x } \zeta(f({\boldsymbol } x))$ ] : @xmath289v_q(-f({\boldsymbol } x))\\ & = \begin{cases } \eta({\boldsymbol } x ) \dfrac{1}{f({\boldsymbol } x)^q } \dfrac{q^q}{(q+1)^{q+1 } } + [ 1-\eta({\boldsymbol } x)][1+f({\boldsymbol } x ) ] , & \text { if } f({\boldsymbol } x)>\dfrac{q}{q+1 } , \\ \eta({\boldsymbol } x)[1-f({\boldsymbol } x ) ] + [ 1-\eta({\boldsymbol } x)][1+f({\boldsymbol } x ) ] , & \text { if } -\dfrac{q}{q+1 } \le f({\boldsymbol } x ) \le \dfrac{q}{q+1 } , \\ \eta({\boldsymbol } x)[1-f({\boldsymbol } x ) ] + [ 1-\eta({\boldsymbol } x)]\dfrac{1}{[-f({\boldsymbol } x)]^q}\dfrac{q^q}{(q+1)^{q+1 } } , & \text { if } f({\boldsymbol } x)<-\dfrac{q}{q+1}. \end{cases } \end{aligned}\ ] ] for each given @xmath290 , we take both @xmath291 and @xmath196 as scalars and hereby write them as @xmath4 and @xmath292 respectively .
we then take @xmath293 as a function of @xmath4 and compute the derivative with respect to @xmath4 : @xmath294 we see that ( 1 ) when @xmath295 , @xmath296 only when @xmath297 , and ( 2 ) when @xmath298 , @xmath296 only when @xmath299 . for these two cases
, we also observe that @xmath300 which follows that @xmath301 is the minimizer of @xmath302 . *
proof of lemma [ lm : err_bd ] * as @xmath203 was defined in , we see that for each @xmath290 , @xmath303 v_q\left(-\tilde{f}({\boldsymbol } x)\right)\\ & = \begin{cases } \eta({\boldsymbol } x ) + [ 1-\eta({\boldsymbol } x)]^{\frac{1}{q+1}}\eta({\boldsymbol } x)^{\frac{q}{q+1 } } , & \text { if } \eta({\boldsymbol } x ) \le 1/2,\\ 1 - \eta({\boldsymbol } x ) + \eta({\boldsymbol } x)^{\frac{1}{q+1}}[1-\eta({\boldsymbol } x)]^{\frac{q}{q+1 } } , & \text { if } \eta({\boldsymbol } x ) > 1/2 , \end{cases}\\ & = \dfrac{1}{2}\bigg(1-|2\eta({\boldsymbol } x)-1|\bigg ) + \dfrac{1}{2}\bigg(1+|2\eta({\boldsymbol } x)-1|\bigg)^{\frac{1}{q+1}}\bigg(1-|2\eta({\boldsymbol } x)-1|\bigg)^{\frac{q}{q+1}}. \end{aligned}\ ] ] for @xmath304 $ ] , we define @xmath305 and compute its first - order derivative as follows , @xmath306 + \left[\dfrac{q}{2(q+1 ) } + \dfrac{q}{2(q+1)}\left(\dfrac{1+a}{1-a}\right)^{\frac{1}{q+1 } } - \dfrac{q}{q+1}\right ] \ge 0 .
\end{aligned}\ ] ] hence for each @xmath307 $ ] , @xmath308 . for each @xmath290 , let @xmath309 and we see that @xmath310 by @xmath311 = e_{\{{\boldsymbol } x : f({\boldsymbol } x)\ge 0\}}[1-\eta({\boldsymbol } x ) ] + e_{\{{\boldsymbol } x : f({\boldsymbol } x)\le 0\}}\eta({\boldsymbol } x ) , $ ] we obtain @xmath312 + e_{\{{\boldsymbol } x : \hat{f}_n({\boldsymbol } x ) \le 0,\ f^\star({\boldsymbol } x ) > 0\}}[2\eta({\boldsymbol } x)-1]\\ & \le e_{\{{\boldsymbol } x : \hat{f}_n({\boldsymbol } x)f^\star({\boldsymbol } x)\le 0\}}|2\eta({\boldsymbol } x)-1|\\ & \le \dfrac{q+1}{q } e_{\{{\boldsymbol } x : \hat{f}_n({\boldsymbol } x)f^\star({\boldsymbol } x)\le 0\ } } \left[1 - \zeta\left(\tilde{f}({\boldsymbol } x)\right)\right ] .
\end{aligned } \label{eq : rf}\ ] ] since @xmath313 and @xmath314 share the same sign , @xmath315 implies that @xmath316 .
when @xmath316 , 0 is between @xmath317 and @xmath314 , and thus indicates that @xmath318 . from , we conclude that @xmath319\\ & \le \dfrac{q+1}{q } e_{{\boldsymbol } x } \left[\zeta\left(\hat{f}_n({\boldsymbol } x)\right ) - \zeta\left(\tilde{f}({\boldsymbol } x)\right ) \right]\\ & = \dfrac{q+1}{q } e_{{\boldsymbol } xy } \left[v_q\left(y\hat{f}_n({\boldsymbol } x)\right ) - v_q\left(y\tilde{f}({\boldsymbol } x)\right ) \right]\\ & = \dfrac{q+1}{q}(\varepsilon_a + \varepsilon_e ) .
\end{aligned}\ ] ] * part * ( 1 ) .
we first show that when @xmath167 is induced by a universal kernel , the approximation error @xmath320 . by definition ,
we need to show that for any @xmath232 , there exists @xmath321 such that @xmath322 we first use truncation to consider a truncated version of @xmath323 . for
any given @xmath324 , we define @xmath325 we have that @xmath326 where @xmath327\\ & -e_{{\boldsymbol } x : \eta({\boldsymbol } x)>1-\delta}\left [ \eta({\boldsymbol } x ) v_q \left(\tilde{f}({\boldsymbol } x)\right ) + ( 1-\eta({\boldsymbol } x ) ) v_q\left(-\tilde{f}({\boldsymbol } x)\right ) \right],\\ \kappa_- = & e_{{\boldsymbol } x : \eta({\boldsymbol }
x)<\delta}\left [ \eta({\boldsymbol } x ) v_q(f_\delta({\boldsymbol } x ) ) + ( 1-\eta({\boldsymbol } x ) ) v_q(-f_\delta({\boldsymbol } x ) ) \right]\\ & -e_{{\boldsymbol } x : \eta({\boldsymbol } x)<\delta } \left [ \eta({\boldsymbol } x ) v_q \left(\tilde{f}({\boldsymbol } x)\right ) + ( 1-\eta({\boldsymbol } x ) ) v_q\left(-\tilde{f}({\boldsymbol } x)\right ) \right ] .
\end{aligned}\ ] ] since @xmath328 when @xmath329 , @xmath330\\ & -e_{{\boldsymbol } x : \eta({\boldsymbol } x)>1-\delta}\left [ \eta({\boldsymbol } x ) v_q\left(\tilde{f}({\boldsymbol } x)\right ) + ( 1-\eta({\boldsymbol } x ) ) v_q\left(-\tilde{f}({\boldsymbol } x)\right ) \right]\\ = & \left[\delta + ( 1-\delta)^{\frac{1}{q+1}}\delta^\frac{q}{q+1 } \right ] - e_{{\boldsymbol } x : \eta({\boldsymbol }
x)>1-\delta } \left[1-\eta({\boldsymbol } x ) + \eta({\boldsymbol } x)^{\frac{1}{q+1}}(1-\eta({\boldsymbol } x))^{\frac{q}{q+1 } } \right ] .
\end{aligned}\ ] ] we notice that @xmath331 is a continuous function in terms of @xmath332 . since @xmath329 implies that @xmath333 , we conclude that for any given @xmath334 , there exists a sufficiently small @xmath335 such that @xmath336 .
we can also obtain @xmath337 in the same spirit .
therefore , @xmath338 by lusin s theorem , there exists a continuous function @xmath339 such that @xmath340 .
notice that @xmath341 . define @xmath342 then @xmath343 as well .
hence @xmath344 where the first inequality comes from the fact that @xmath345 is lipschitz continuous , i.e. , @xmath346 notice that @xmath347 is also continuous .
the definition of the universal kernel implies the existence of a function @xmath321 such that @xmath348 by combining , , and we obtain . *
part * ( 2 ) .
in this part we bound the estimation error @xmath349 .
note that rkhs has the following reproducing property @xcite : @xmath350 fix any @xmath232 . by the kkt condition of and the representor theorem , we have @xmath351 we define @xmath352}$ ] as the solution of when the @xmath353th observation is excluded from the training data , i.e. , @xmath354 } = \operatorname*{argmin}_{f \in { \mathcal{h}}_k } \left [ \dfrac{1}{n}\sum_{i=1 , i \neq k}^n v_q \left ( y_i ( f({\boldsymbol } x_i ) \right ) + \lambda_n||f||^2_{{\mathcal{h}}_k } \right ] .
\label{eq : ker_dwd_delete_k}\ ] ] by the definition of @xmath352}$ ] and the convexity of @xmath213 , we have @xmath355}({\boldsymbol } x_i ) \right ) - \lambda_n ||\hat{f}^{[k]}||^2_{\mathcal{h}_k}\\ \le & - \dfrac{1}{n}\sum_{i=1 , i\neq k}^{n}v'_q\left(y_i \hat{f}_n({\boldsymbol } x_i ) \right ) y_i\left(\hat{f}^{[k ] } ( { \boldsymbol } x_i ) - \hat{f}_n({\boldsymbol } x_i)\right ) + \lambda_n ||\hat{f}_n||^2_{\mathcal{h}_k } -\lambda_n ||\hat{f}^{[k]}||^2_{\mathcal{h}_k}. \end{aligned}\ ] ] by the reproducing property , we further have @xmath356 } ( { \boldsymbol } x ) - \hat{f}_n({\boldsymbol } x)\right\rangle_{\mathcal{h}_k } + \lambda_n ||\hat{f}_n||^2_{\mathcal{h}_k } -\lambda_n ||\hat{f}^{[k]}||^2_{\mathcal{h}_k}\\ = & - \dfrac{1}{n}\sum_{i=1 , i\neq k}^{n}v'_q\left(y_i \hat{f}_n({\boldsymbol } x_i ) \right ) y_i\left\langle k({\boldsymbol } x_i , { \boldsymbol } x ) , \hat{f}^{[k ] } ( { \boldsymbol } x ) - \hat{f}_n({\boldsymbol } x)\right\rangle_{\mathcal{h}_k } \\ & - 2 \lambda_n \left\langle \hat{f}_n({\boldsymbol } x ) , \hat{f}^{[k]}({\boldsymbol } x)-\hat{f}_n({\boldsymbol } x ) \right\rangle_{\mathcal{h}_k } - \lambda_n||\hat{f}^{[k]}- \hat{f}_n||^2_{\mathcal{h}_k}\\ = & \dfrac{1}{n } v'_q\left(y_k \hat{f}_n({\boldsymbol } x_k ) \right ) y_k \left\langle k({\boldsymbol } x_k , { \boldsymbol } x ) , \hat{f}^{[k ] } ( { \boldsymbol } x ) - \hat{f}_n({\boldsymbol } x)\right\rangle_{\mathcal{h}_k } - \lambda_n||\hat{f}^{[k]}- \hat{f}_n||^2_{\mathcal{h}_k } , \end{aligned}\ ] ] where the equality in the end holds by .
thus , by cauchy - schwartz inequality , @xmath357}- \hat{f}_n||^2_{\mathcal{h}_k } \le v'_q\left(y_k \hat{f}_n({\boldsymbol } x_k ) \right ) y_k \left\langle k({\boldsymbol } x_k , { \boldsymbol } x ) , \hat{f}^{[k ] } ( { \boldsymbol } x ) - \hat{f}_n({\boldsymbol } x)\right\rangle_{\mathcal{h}_k}\\ \le & \left|v'_q\left(y_k \hat{f}_n({\boldsymbol } x_k ) \right ) \right| ||k({\boldsymbol } x_k , { \boldsymbol } x)||_{\mathcal{h}_k}||\hat{f}^{[k]}- \hat{f}_n||_{\mathcal{h}_k } \le \sqrt{k({\boldsymbol } x_k , { \boldsymbol } x_k ) } \cdot || \hat{f}^{[k]}- \hat{f}_n||_{\mathcal{h}_k } , \end{aligned}\ ] ] which implies @xmath358}- \hat{f}_n||_{\mathcal{h}_k } \le \dfrac{\sqrt{b}}{n\lambda_n},\ ] ] where @xmath359 . by the reproducing property
, we have @xmath360}- \hat{f}_n||^2_{\mathcal{h}_k } \le b \left(\dfrac{\sqrt{b}}{n\lambda_n}\right)^2 .
\end{aligned}\ ] ] by the lipschitz continuity of the dwd loss , we obtain that for each @xmath361 , @xmath362}({\boldsymbol } x_k ) \right ) - v_q\left(y_k\hat{f}_n({\boldsymbol } x_k ) \right ) \le & |\hat{f}^{[k]}({\boldsymbol } x_k ) - \hat{f}_n({\boldsymbol } x_k)| \le \dfrac{b}{n\lambda_n } , \end{aligned}\ ] ] and therefore , @xmath363}({\boldsymbol }
x_k ) \right ) \le \dfrac{1}{n}\sum_{k=1}^n v_q\left(y_k\hat{f}_n({\boldsymbol } x_k ) \right ) + \dfrac{b}{n\lambda_n}. \label{eq : proof_eq100}\ ] ] let @xmath364 such that @xmath365 by definition of @xmath209 , we have @xmath366 since each data point in @xmath367 is drawn from the same distribution , we have @xmath368}({\boldsymbol } x_k ) \right ) \right ] = \dfrac{1}{n}\sum_{k=1}^n e_{{\boldsymbol } t_n}v_q\left(y_k\hat{f}^{[k]}({\boldsymbol } x_k ) \right ) = e_{{\boldsymbol } t_{n-1}}e_{{\boldsymbol } xy } v_q\left(y\hat{f}_{n-1}({\boldsymbol } x ) \right ) .
\quad \label{eq : proof_eq103 } \end{aligned}\ ] ] by combining
we have @xmath369 by the choice of @xmath370 , we see that there exits @xmath371 such that when @xmath372 we have @xmath373 , @xmath374 , and hence @xmath375 \le \inf_{f\in\mathcal{h}_k}e_{{\boldsymbol } xy } v_q\left(yf({\boldsymbol } x ) \right)+\epsilon.\ ] ] because @xmath223 is arbitrary and @xmath376 \ge \inf_{f\in\mathcal{h}_k}e_{{\boldsymbol } xy } v_q\left(yf({\boldsymbol } x ) \right ) $ ] , we have @xmath377 = \inf_{f\in\mathcal{h}_k}e_{{\boldsymbol } xy } v_q\left(yf({\boldsymbol } x ) \right)$ ] , which equivalently indicates that @xmath378 since @xmath379 , then by markov inequality , we prove part ( 2 ) .
fernndez - delgado , m. , cernadas , e. , barro , s. , and amorim , d. ( 2014 ) , `` do we need hundreds of classifiers to solve real world classification problems ? '' _ the journal of machine learning research _ , 15 , 31333181 .
huang , h. , liu , y. , du .
y. , perou , c. , hayes , d. , todd , m. , and marron , j.s .
( 2013 ) , `` multiclass distance - weighted discrimination , '' _ journal of computational and graphical statistics _ , 22(4 ) , 953969 .
huang , h. , lu , x. , liu , y. , haaland , p. , and marron , j.s .
( 2012 ) , `` r / dwd : distance - weighted discrimination for classification , visualization and batch adjustment , '' _ bioinformatics _ , 28(8 ) , 11821183 .
qiao , x. , zhang , h. , liu , y. , todd , m. , marron , j.s .
( 2010 ) , `` weighted distance weighted discrimination and its asymptotic properties , '' _ journal of american statistical association _ , 105(489 ) , 401414 .
wahba , g. , gu , c. , wang , y. , and campbell , r. ( 1994 ) , `` soft classification , aka risk estimation , via penalized log likelihood and smoothing spline analysis of variance , '' in _
santa fe institute studies in the sciences of complexity - proceeding vol _
, 20 , addison - wesley publishing co , 331331 .
wahba , g. ( 1999 ) , `` support vector machines , reproducing kernel hilbert spaces and the randomized gacv , '' _ advances in kernel methods - support vector learning _ , 6 , 6987 .
wang , b. and zou , h. ( 2015 ) , `` sparse distance weighted discrimination , '' _ journal of computational and graphical statistics _
, forthcoming . | distance weighted discrimination ( dwd ) is a margin - based classifier with an interesting geometric motivation .
dwd was originally proposed as a superior alternative to the support vector machine ( svm ) , however dwd is yet to be popular compared with the svm .
the main reasons are twofold .
first , the state - of - the - art algorithm for solving dwd is based on the second - order - cone programming ( socp ) , while the svm is a quadratic programming problem which is much more efficient to solve .
second , the current statistical theory of dwd mainly focuses on the linear dwd for the high - dimension - low - sample - size setting and data - piling , while the learning theory for the svm mainly focuses on the bayes risk consistency of the kernel svm .
in fact , the bayes risk consistency of dwd is presented as an open problem in the original dwd paper . in this work
, we advance the current understanding of dwd from both computational and theoretical perspectives . we propose a novel efficient algorithm for solving dwd , and our algorithm can be several hundred times faster than the existing state - of - the - art algorithm based on the socp .
in addition , our algorithm can handle the generalized dwd , while the socp algorithm only works well for a special dwd but not the generalized dwd .
furthermore , we consider a natural kernel dwd in a reproducing kernel hilbert space and then establish the bayes risk consistency of the kernel dwd .
we compare dwd and the svm on several benchmark data sets and show that the two have comparable classification accuracy , but dwd equipped with our new algorithm can be much faster to compute than the svm . *
key words : * bayes risk consistency , classification , dwd , kernel methods , mm principle , socp . |
the question of whether laboratory spin glasses or the theoretical models used to represent them have a thermodynamic phase transition remains unresolved despite decades of work @xcite .
although the infinite - range sherrington - kirkpatrick ( sk ) ising spin glass is easily shown to possess a phase transition @xcite , the existence of one in the corresponding short - range edwards - anderson ( ea ) @xcite ising model ( on the cubic lattice @xmath1 ) has not been established in _ any _ finite dimension . while some evidence for a transition has been uncovered through high - temperature expansions @xcite , analytical studies of variable long - range @xmath2 models @xcite , and extensive numerical simulations @xcite ,
the issue remains unresolved @xcite .
random graph methods , and in particular the fortuin - kastelyn ( fk ) random cluster ( rc ) representation @xcite , provide a set of useful tools for studying phase transitions ( more specifically , the presence of multiple gibbs states arising from broken spin rotational symmetry ) in discrete spin models . in these representations
spin correlation functions can be expressed through the geometrical properties of associated random graphs .
fk and related models are probably best known in the physics literature for providing the basis for powerful monte carlo methods for studying phase transitions @xcite , but they have also proved important in obtaining rigorous results on phase transitions in discrete - spin ferromagnetic ( including inhomogeneous and randomly diluted ) models ( e.g. , @xcite ) . because of complications due to frustration , however , graphical representations have so far played a less important role in the study of spin glasses . the goal of the studies presented here is to construct a viable approach which uses random graph methods to address the problem of phase transitions and broken symmetry in spin glass models . in earlier papers @xcite , we studied the `` percolation signature '' of spin glass ordering within two different graphical representations the two - replica model of chayes , machta and redner ( cmr ) @xcite and a two - replica version of the fk representation ( trfk ) , as proposed in sec .
4.1 of @xcite .
the purpose of that analysis was to show that fk methods could be utilized to study spin glass phase transitions .
the result of this work was the uncovering of strong evidence that the existence of a spin glass transition coincides with the emergence of _ doubly _ percolating clusters of _ unequal _ densities .
this scenario is more complex than what occurs in ferromagnetic models , where the phase transition coincides with percolation of a single fk cluster . in what follows ,
we first review the fk random cluster representation for ferromagnetic models , and then describe our analytical and numerical results for spin glasses .
in this section we briefly review the fortuin - kasteleyn random cluster representation @xcite which relates the statistical mechanics of ising ( or potts ) models to a dependent percolation problem
. our focus will be on ising models throughout , but the analysis is easily extended to more general potts models .
we start by considering a nearest - neighbor ising ferromagnet , whose couplings @xmath3 are not necesarily identical : @xmath4 where , as already noted , @xmath3 and the sum is over nearest neighbor pairs of sites in @xmath5 .
each such coupling is associated with an edge , or bond , @xmath6 , with the set of all such bonds denoted @xmath7 .
the rc approach introduces parameters @xmath8 through the formula : @xmath9\ , , \ ] ] where @xmath10 is the inverse temperature .
one can then define a probability measure @xmath11 on @xmath12 , that is , on 0- or 1-valued bond occupation variables @xmath13 .
it is one of two marginal distributions ( the other being the ordinary gibbs distribution ) of a joint distribution on @xmath14 of the spins and bonds together ( such a joint distribution will be introduced in sec .
[ subsec : cmr ] ) .
the marginal distribution @xmath15 is given formally by @xmath16 where @xmath17 is a normalization constant , @xmath18 is the number of clusters determined by the realization @xmath19 , @xmath20 is the bernoulli product measure corresponding to independent occupation variables with @xmath21 , and @xmath22 is the indicator function on the event @xmath23 in @xmath12 that there exists a choice of the spins @xmath24 so that @xmath25 for all @xmath6 @xcite .
there are several things to be noted about ( [ eq : mufk ] ) .
the factor 2 in the term @xmath26 arises because we ve confined ourselves to ising models , so that every connected cluster of spins ( each such connected cluster consists of all satisified bonds ) can be in one of two states ( in the ferromagnetic case , all up or all down ) ; in a @xmath27-state potts model , this term would then be replaced by @xmath28 .
more importantly , the indicator function on @xmath23 , which is the event that there is no frustration in the occupied bond configuration , is always one for the ferromagnet ; consequently , this term is superfluous for ferromagnetic models .
we include it , however , because it will be needed when we generalize to models with frustration .
finally , we note that finite - volume versions of the above formulas , with specified boundary conditions , can be similarly constructed . in the case of a ferromagnet ,
there are general theorems @xcite which ensure that when percolation occurs , there is a unique percolating cluster .
it then easily follows that rc percolation within the fk representation ( or `` fk percolation '' for short ) corresponds to the presence of multiple gibbs states ( in the ferromagnet , magnetization up and magnetization down ) , and moreover that the onset of percolation occurs at the ferromagnetic critical temperature . to prove this
, it is sufficient to show that fk percolation is both necessary and sufficient for the breaking of global spin flip symmetry in the ferromagnet . to see that fk percolation is a necessary condition ,
note that the contribution to the expectation of @xmath29 from any finite rc cluster is zero : if a spin configuration @xmath30 is consistent with a given rc bond realization within such a cluster , so is @xmath31 , and both will be equally likely . as a consequence , @xmath32 in infinite volume in the absence of rc percolation . to see that rc percolation is a sufficient condition for the magnetization order parameter to be nonzero , consider a finite volume @xmath33 with fixed boundary conditions , i.e. , a specification @xmath34 for each @xmath35 .
for the ferromagnet , by first choosing all @xmath36 and then all @xmath37 , one can change the sign of the spin @xmath29 at the origin even as @xmath38 .
that is , boundary conditions infinitely far away affect @xmath29 , which is a signature of the existence of multiple gibbs states .
when attention is confined to ferromagnetic models , the mapping of the fk formalism to interesting statistical mechanical quantities is straightforward ( and intuitive ) ; for example @xmath39 where @xmath40 is the usual gibbs two - point correlation function and @xmath41 is the rc probability that @xmath42 and @xmath43 are in the same cluster . similarly ,
with `` wired '' boundary conditions ( i.e. , each boundary spin is connected to its nieghbors ) , one has @xmath44 so for ferromagnets , a phase transition from a unique ( paramagnetic ) phase at low @xmath10 to multiple infinite - volume gibbs states at large @xmath10 is equivalent to a percolation phase transition for the corresponding rc measure . for spin glasses ( or other nonferromagnets with frustration )
the situation is more complicated .
now for two sites @xmath42 and @xmath43 , ( [ eq : ferropath ] ) becomes @xmath45 where @xmath46 is any path of occupied bonds from @xmath42 to @xmath43 . by the definition of @xmath23 , any two such paths @xmath46 and @xmath47 in the _ same
_ cluster will satisfy @xmath48 .
just as in the ferromagnet , if percolation of a random cluster occurs in the fk representation , the percolating cluster is unique ( in each realization of fk spins and bonds ) @xcite . in spite of this ,
rc percolation alone is no longer sufficient to prove broken spin - flip symmetry in the ea spin glass .
this is because even in the presence of rc percolation , it is unclear whether there exist any two sets of boundary conditions that can alter the state of the spin at the origin from arbitrarily far away .
although the infinite cluster in any one rc realization is unique , different rc realizations can have different paths from @xmath49 , and because of frustration this can lead to different signs for @xmath29 .
so percolation might still allow for @xmath50 as @xmath38 , independently of boundary condition .
indeed , it is known that fk bonds percolate well above the spin glass transition temperature .
for the three - dimensional ising spin glass on the cubic lattice , fortuin - kasteleyn bonds percolate at @xmath51 @xcite while the inverse critical temperature is believed to be @xmath52 @xcite . near the spin glass critical temperature ,
the giant fk cluster includes most of the sites of the system . for this reason , the swendsen - wang algorithm , though valid ,
is inefficient for simulating spin glasses .
however , it is an open question as to whether the presence of `` single '' ( see below ) fk percolation would lead to a slower , e.g. power - law , decay of correlation functions even though the gibbs state is unique .
if this were to happen , then the onset of single fk percolation would imply a phase transition in the spin glass , though not multiple gibbs states and hence no broken spin - flip symmetry .
this possibility was suggested in @xcite ; but to date , no evidence exists to support it . nonetheless , single fk percolation remains a _ necessary _
condition for multiple ( symmetry - broken ) gibbs phases in the spin glass , for the same reason as for the ferromagnet .
a slightly stronger version of that argument @xcite proves that the transition temperature for an ea spin glass , if it exists , is bounded from above by the transition temperature in the corresponding ( disordered ) ferromagnet .
the essential difference ( from the point of view of fk percolation ) between the ferromagnet and the spin glass is the factor @xmath22 in ( [ eq : mufk ] ) .
it is somewhat easier to discuss this factor in the context of finite volumes .
let @xmath33 denote the @xmath53 cube centered at the origin , and let @xmath54 denote the set of bonds @xmath55 with both @xmath42 and @xmath43 in @xmath33 . for models containing frustration , @xmath23 is typically _ not _ all of @xmath56 , unlike for the ferromagnet ; it is the set of all _ unfrustrated _ bond configurations ( so that the product of couplings around any closed loop in such a configuration is always positive ) .
so is there a way to extend the above considerations to arrive at a sufficient condition for multiple gibbs states in spin glass models ?
we begin by noting that fk clusters identify magnetization correlations ; but spin glass ordering is manifested by the edwards - anderson ( ea ) order parameter , and not the magnetization , becoming nonzero .
the ea order parameter @xmath57 can be defined with respect to two independent replicas of the system , each with the same couplings @xmath58 . denoting the spins in the two replicas by @xmath24 and @xmath59 , each taking values @xmath60 ,
@xmath57 is defined in terms of the overlap , @xmath61 in the limit as the number of sites @xmath62 . in general
, @xmath63 is a random variable whose maximum possible value is @xmath57 , but in the case where ( in the limit @xmath62 ) @xmath24 and @xmath59 are drawn from a single pure state , @xmath63 takes on only the single value @xmath57 . using this as a guide
, it appears that one possibility , proposed in @xcite , for extending fk methods to spin glasses is to use what might be called _
double _ fk percolation .
here one expands the sample space @xmath64 to include two independent copies of the bond occupation variables ( for a given @xmath65 configuration ) , and defines the variable @xmath66 , where @xmath13 and @xmath67 are taken from the two copies .
one then replaces percolation of @xmath19 in the single rc case with percolation of @xmath68 .
it is not hard to see that this would be a sufficient condition for the existence of multiple gibbs phases ( and consequently , for a phase transition ) .
this is not the only way , however , to use `` double '' percolation of rc clusters in some form to arrive at a condition for spin glass ordering ; the above describes what we denoted in the introduction as the trfk approach . in the next subsection we present a closely related approach , the crm two - replica graphical representation , which requires a more lengthy description .
once this is done , we will discuss how both representations can be used to illuminate the nature of phase transitions in short - range spin glass models . in describing these `` double fk '' representations
, we will find it useful to reformulate our description of the relevant measures slightly using a joint spin - bond distribution introduced by edwards and sokal @xcite .
the statistical weight @xmath69 for the edwards - sokal distribution on a finite lattice @xmath33 is @xmath70 here @xmath71 is the number of occupied bonds and @xmath72 is the total number of bonds on the lattice .
the factor @xmath73 is introduced to require that every occupied bond is satisfied ; it is defined by @xmath74 with @xmath75 defined as in ( [ eq : pxy ] ) and with all @xmath76 , it is easy to verify that the spin and bond marginals of the edwards - sokal distribution are the ferromagnetic ising model with coupling strength @xmath77 and the fortuin - kasteleyn random cluster model ( cf .
( [ eq : mufk ] ) ) , respectively .
we can now adapt the above representation to the @xmath0 ising spin glass .
( with minor modifications , it can also be adapted to gaussian and other distributions for the couplings . )
the corresponding edwards - sokal weight is the same as that given in eq .
( [ eq : es ] ) .
the @xmath78 factor must still enforce the rule that all occupied bonds are satisfied , @xmath79 so ( cf . ( [ eq : mufk ] ) ) @xmath80 .
the spin marginal of the corresponding edwards - sokal distribution is now the ising spin glass with couplings @xmath58 .
but the relationship between spin - spin correlations and bond connectivity is complicated by the presence of antiferromagnetic bonds .
now one has @xmath81 the two - replica cmr graphical representation , introduced in @xcite , incorporates , in addition to the spin variables @xmath24 and @xmath59 , two sets of bond variables @xmath19 and @xmath82 , each taking values in @xmath83 .
now the edwards - sokal weight is @xmath84 where the @xmath85 s are bernoulli factors for the two types of bonds , @xmath86 and the bond occupation probabilities are @xmath87 so @xmath88 .
the @xmath78 and @xmath89 factors constrain where the two types of occupied bonds are allowed , @xmath90 we refer to the @xmath91-occupied bonds as `` blue '' and the @xmath92-occupied bonds as `` red '' .
the @xmath78 constraint says that blue bonds are allowed only if the bond is satisfied in both replicas .
the @xmath89 constraint says that red bonds are allowed only if the bond is satisfied in exactly one replica .
it is straightforward to verify that the spin marginal of the two - replica edwards - sokal weight is that for two independent ising spin glasses with the same couplings , @xmath93\ ] ] connectivity by occupied bonds in the two - replica representation is related to correlations of the local spin glass order parameter , [ sec : pgr ] @xmath94 it is straightforward to verify that @xmath95 as in the case of the fk representation , a minus sign complicates the relationship between correlations and connectivity but in a conceptually different way .
the second term in eq .
( [ eq : trcor ] ) is independent of the underlying couplings in the model and is present for both spin glasses and ferromagnetic models . as noted earlier , for ferromagnets ( in the absence of non - translation - invariant boundary conditions ) , the signature of ordering is a _ single _ percolating fk cluster . for spin glasses ,
the situation is more complicated as there can be more than one percolating cluster .
however , if the cmr graphical representation displays a single percolating blue cluster of largest density , one can similarly show broken symmetry , for either ea or sk models .
this is because one can impose `` agree '' or `` disagree '' boundary conditions between those @xmath96 and @xmath97 boundary spins belonging to the maximum density blue network . in the infinite volume limit , these two boundary conditions give different gibbs states for the @xmath30-spin system ( for fixed @xmath98 ) related to each other by a global spin flip ( of @xmath30 ) .
it is a separate matter , however , to relate in a simple way the ea order parameter to the density difference in blue ( or more generally , doubly percolating ) clusters .
intuitively , it seems that there should be a simple correspondence , and in fact such a relation is easy to show for the sk model : here the overlap @xmath63 is exactly the difference in density between two percolating blue clusters @xcite .
but it is not immediately obvious that a similar density difference can be simply related to the ea order parameter ( although the arguments above imply that here also a density difference implies a nonzero @xmath63 . )
this is because in a two - replica situation , it is not immediately obvious that there is no contribution from finite clusters in short - range models . for the trfk representation ,
similar reasoning shows that the occurrence of exactly two doubly - occupied percolating fk clusters with different densities implies broken symmetry for the spin system @xcite and that @xmath63 should equal ( and once again , does equal in the sk model ) the density difference . in @xcite we presented preliminary numerical evidence that there is such a nonzero density difference below the spin glass transition temperature for the @xmath99 ea @xmath0 spin glass ( for both the trfk and cmr representations ) .
we will discuss these results further in sec .
[ sec : numerical ] .
but now we turn to a review of some rigorous results , particularly for the sk model , that appeared in @xcite and @xcite .
the sk model permits a fairly extensive rigorous analysis of both the cmr and trfk representations which , when combined with other known results , permits a sharp picture to emerge of the connection between double fk percolation ( we use this more general term to refer to any representation that relies on two fk replicas , such as cmr or trfk ) , and a phase transition to a low - temperature spin glass phase with broken spin - flip symmetry . the surprising feature to emerge from this analysis
is that in the cmr representation , there already exist well _ above _ the transition temperature two percolating networks of blue bonds , of equal density and in all respects macroscopically indistinguishable . below the critical temperature @xmath100
, the indistinguishability is lifted : the two infinite clusters assume different densities . on the other hand , there is no double fk percolation at all above @xmath100 in the trfk representation , but below @xmath100 there are again two percolating double clusters of unequal density .
we do not yet know whether this difference in the two pictures above @xmath100 persists in the ea model , but numerical evidence ( to be discussed in the next section ) so far appears to indicate that it does not : in the ea spin glass , both representations seem to behave similarly to cmr in the sk model .
the sk hamiltonian for an @xmath101-spin system is @xmath102 where @xmath103 are vertices on a complete graph .
the couplings @xmath104 are i.i.d .
random variables chosen from a probability measure @xmath105 satisfying the following properties : \1 ) the distribution is symmetric : @xmath106 .
\2 ) @xmath105 has no @xmath107-function at @xmath108 .
\3 ) the moment generating function is finite : @xmath109 for all real @xmath110 .
\4 ) the second moment of @xmath105 is finite ; we normalize it to one : @xmath111 .
so the results of this section hold for gaussian and many other distributions ( although not diluted ones , because of requirement ( 2 ) ) , but the analysis is simplest for the @xmath0 model , to which we confine our attention for the remainder of this section .
we therefore assume that @xmath112 with equal probability ; for this distribution ( or any other satisfying property ( 3 ) ) @xmath113 @xcite .
it is relatively easy to see , using a heuristic argument presented in @xcite , at what temperature a single fk cluster will form .
the sk energy per spin , @xmath114 , is given above the critical temperature by @xmath115 .
therefore , for large @xmath101 the fraction @xmath116 of satisfied edges is @xmath117 from ( [ eq : pxy ] ) it follows that a fraction @xmath118 of satisfied edges are occupied . according to the theory of random graphs ( see @xcite ) , a giant cluster forms in a random graph of @xmath101 vertices when a fraction @xmath119 of edges is occupied with @xmath120 , and there is then a single giant cluster @xcite .
so if edges are satisfied independently ( which of course they re not this is why this argument is heuristic only ) , then single replica fk giant clusters should form when @xmath121 when @xmath120 .
the single replica fk percolation threshold is therefore at @xmath122 and above this threshold , there should be a single giant fk cluster .
this simple argument can be made rigorous @xcite , but the basic ideas are already displayed above .
the rigorous argument obtains upper and lower bounds for the conditional probability that an edge @xmath123 is satisfied , given the satisfaction status of all the other edges .
if these bounds are close to each other ( for large @xmath101 ) then treating the satisfied edges as though chosen independently can be justified .
we now describe that argument . to avoid the problem of non - independence of satisfied edges , we fix all couplings @xmath124 but one , which we denote @xmath125 , and ask for the conditional probability of its sign given the configuration of all other couplings and all spins @xmath96 .
the ratio @xmath126 of the partition functions with @xmath127 and @xmath128 satisfies @xmath129 the conditional probabilities @xmath130 that @xmath131 therefore satisfy @xmath132 let @xmath133 ( @xmath134 ) be the conditional probability for any edge @xmath123 to be satisfied ( unsatisfied ) given the satisfaction status of all other edges .
these must then be bounded as follows : @xmath135 and therefore @xmath136 one now obtains rigorously the same conclusions as before i.e. , ( [ eq : fkperc ] ) is valid with a single giant fk cluster for @xmath137 with any @xmath120 .
essentially the same argument is used for the more interesting case of double percolation .
here there are two spin replicas denoted by @xmath30 and @xmath98 .
we are now interested in percolation of doubly satisfied edges ; the spins at the vertices of each such edge must satisfy @xmath138 ( and then will be satisifed for exactly one of the two signs of @xmath124 ) .
we note an immediate difference between the cmr and trfk representations : for the former , doubly satisfied edges are occupied with probability @xmath139 , while for the latter , @xmath140 ^ 2 \sim 4\beta^2/n$ ] .
it seems likely that this difference occurs only for the sk model ; the various factors of @xmath101 are absent in the ea model .
we can proceed much as in the single - replica case by dividing all @xmath141 configurations into two sectors the _ agree _ ( where @xmath142 ) and the disagree sectors ( where @xmath143 ) .
we also denote by @xmath144 and @xmath145 the numbers of sites in the sectors and denote by @xmath146 and @xmath147 the sector densities ( so that @xmath148 ) .
the spin overlap @xmath63 is then just @xmath149 for @xmath150 , @xmath151 as @xmath152 ( because the ea order parameter is zero in the paramagnetic phase ) while for @xmath153 , @xmath154 as @xmath152 , where @xmath155 denotes an average over couplings .
so it must be that @xmath156 for @xmath157 while @xmath158 for @xmath159 .
the arguments for the single replica case can be repeated separately within the agree and disagree sectors .
letting @xmath160 denote the conditional probabilities that @xmath161 given the other @xmath162 s and all @xmath96 s and @xmath97 s , we have within either of the two sectors that @xmath163 so that the conditional probability _ within a single sector
_ @xmath164 for @xmath165 to be doubly satisfied is @xmath166 . for @xmath167 ,
we have @xmath168 , @xmath169 ( in the limit @xmath152 ) and so in either sector , double fk percolation is approximately a random graph model with @xmath170 sites and bond occupation probability @xmath171 ; thus double fk giant clusters do not occur for @xmath172 in the trfk representation . in the cmr representation
, blue percolation corresponds to bond occupation probability @xmath173 @xmath174 and so the threshold for blue percolation is given by @xmath175 .
but now there are _ two _ giant clusters , one in each of the two sectors , and as noted above they satisfy @xmath156 when @xmath176 . for @xmath177 , @xmath178 . since @xmath179 for @xmath177 ( indeed for any fixed @xmath180 ) , it follows from random graph theory that each giant cluster occupies the entire sector so that @xmath181 and @xmath182 are also the cluster percolation densities of the two giant clusters . in the case of two - replica fk percolation for @xmath177 ,
let us denote by @xmath183 and @xmath184 the larger and smaller of @xmath181 and @xmath182 , so that @xmath185 and @xmath186 .
then for @xmath187 , the bond occupation probability in the larger sector is @xmath188 with @xmath189 and there is a ( single ) giant cluster in that larger sector .
there will be another giant cluster ( of lower density ) in the smaller sector providing @xmath190 .
since @xmath191 , for this to be the case it suffices if for @xmath177 , @xmath192 the estimated behavior of @xmath57 both as @xmath193 and as @xmath194 @xcite suggests that this is always valid . in any case , we have rigorously proved that there is a unique maximal density double fk cluster for @xmath177 .
an important feature of spin glass order is ultrametricity , which is believed to be true at least for the sk model @xcite , although it has not yet been proved rigorously . for
the spin overlaps coming from three replicas put into rank order , @xmath195 , ultrametricity is the property that @xmath196 .
the issue of percolation signatures for ultrametricity in the sk model is treated at length in @xcite .
we only mention here one of those signatures , which concerns the _ four _ percolating clusters that arise in a cmr representation of three replicas when considering bonds that are simultaneously blue _ both _ for replicas one and two as well as for replicas one and three . denoting the four cluster densities in rank order as @xmath197 , the percolation version of the ultrametric property
is that @xmath198 _ and _ @xmath199 . the case @xmath200 ( resp . , @xmath201 ) corresponds to @xmath202 ( resp , @xmath203 ) .
see @xcite for more details .
we have carried out numerical simulations to explore the properties of two - replica fk representations described in the previous sections for the case of the three - dimensional @xmath0 ea spin glass .
the monte carlo algorithm that we use also makes use of the cmr representation in addition to parallel tempering and metropolis sweeps .
similar methods have been previously applied by swendsen and wang @xcite and others @xcite .
the algorithm is described in more detail in @xcite , here we provide some additional details about the cmr component of the algorithm .
the cmr cluster algorithm alternates between `` bond moves '' and `` spin moves . ''
the bond move takes a pair of ea spin configurations @xmath24 and @xmath59 , both in the same realization of disorder , @xmath58 , and populates the bonds of the lattice with red and blue cmr bonds .
each bond that is satisfied in both replicas is occupied with a blue bond with probability @xmath204 and each bond that is satisfied in only one replica is occupied with a red bond with probability @xmath205 , defined in ( [ eqn : blue ] ) and ( [ eqn : red ] ) , respectively .
the bond move is described by the conditional probability , @xmath206 for bond configuration @xmath207 given spin configuration @xmath208 . to simplify the notation
, we omit the lattice indices and the dependence on @xmath124 in the following equations except where explicitly needed . for the bond move
we have , @xmath209 the spin move takes a blue and red cmr bond configuration and produces a pair of spin configurations @xmath24 and @xmath59 .
all spin configurations obeying the constraints that blue bonds are doubly satisfied and red bond are singly satisfied have the same probability .
the spin move is described by the conditional probability , @xmath210 : @xmath211 a bond move followed by a spin move constitututes a single sweep of the cmr cluster algorithm .
we now demonstrate the validity of the cmr algorithm by showing that detailed balance and ergodicity are satisfied .
ergodicity is clearly satisfied since there is a non - vanishing probability that no red or blue bonds will be created during a bond move so that the spin move can then produce any spin configuration .
detailed balance with respect to the equilibrium spin distribution can be stated as @xmath212=\\ \sum_{\{\omega''\},\{\eta''\}}{{\cal t } } ( \{\sigma'\},\{\tau'\ } |\{\omega''\},\{\eta''\ } ) { { \cal t } } ( \{\omega''\},\{\eta''\ } | \{\sigma\},\{\tau\ } ) \ , \times \exp \left [ \beta \sum_{\{xy\ } } j_{xy } ( \sigma_x\sigma_y+\tau_x\tau_y)\right ] .\end{gathered}\ ] ] this equation must hold for all choices of @xmath24 , @xmath59 , @xmath213 and @xmath214 . note that by ( [ eqn : es ] ) and ( [ eqn : spinmarg ] ) the denominator in ( [ eqn : tomsig ] ) is simply the boltzmann weight for two independent ea spin glasses , cancelling the same factors in the numerator of the detailed balance equation .
thus , the lhs of ( [ eqn : detailed ] ) can be written as @xmath215 note that this expression is symmetric under the exchange of @xmath24 , @xmath59 and @xmath213 , @xmath214 and thus equal to the rhs of ( [ eqn : detailed ] ) , demonstrating that detailed balance holds .
the bond configurations observed after the bond move are , in the same fashion , easily shown to be equilibrium cmr bond configurations as described by the bond marginal of the edwards - sokal weight ( [ eqn : es ] ) .
although the cmr algorithm correctly samples equilibrium configurations of ea spin glasses , in 3d it is not efficient . to obtain a more efficient algorithm
, we also make use of parallel tempering and metropolis sweeps .
the cmr algorithm employs two replicas at a single temperature while parallel tempering exchanges replicas at different temperatures .
here we use 20 inverse temperatures equally spaced between @xmath216 to @xmath217 .
the phase transition temperature of the system was recently measured as @xmath52 @xcite .
a single sweep of the full algorithm consists of a cmr cluster sweep for each pair of replicas at each temperature , a parallel tempering exchange between replicas at each pair of neighboring temperatures and a metropolis sweep for every replica .
we simulated the three - dimensional @xmath0 edwards - anderson model on skew periodic cubic lattices for system sizes @xmath218 , @xmath219 , @xmath220 and @xmath221 .
for each size we simulated 100 realizations of disorder for 50,000 monte carlo sweeps of which the first @xmath222 of the sweeps were for equilibration and the remaining @xmath223 for data collection .
the quantities that we measure are the fraction of sites in the largest blue cluster , @xmath224 and second largest blue cluster , @xmath225 and the number of blue cmr wrapping cluster , @xmath226 , and the number of trfk `` wrapping '' clusters , @xmath227 .
a cluster is said to wrap if it is connected around the system in any of the three directions .
figure [ fig : wrap ] shows the average number @xmath228 of cmr blue wrapping clusters as a function of inverse temperature @xmath10 .
the curves are ordered by system size with largest size on the bottom for the small @xmath10 and on top for large @xmath10 .
the data suggests that there is a percolation transition at some @xmath229 . for @xmath230
there are _ two _ wrapping clusters while for @xmath231 there are none . near and above the spin glass transition at @xmath232
the expected number of wrapping clusters falls off but the fall - off diminishes as system size increases .
this figure suggests that in the large size limit there are exactly two spanning clusters near the spin glass transition both above and below the transition temperature . vs. @xmath10 for the 3d ea model .
] figure [ fig : clustersize ] shows the fraction of sites in the largest cmr blue cluster , @xmath224 , second largest cmr blue cluster , @xmath225 and the sum of the two , @xmath233 .
the middle set of four curves is @xmath224 for sizes @xmath218 , @xmath219 , @xmath220 and @xmath221 , ordered from top to the bottom at @xmath234 . the bottom set of curves is @xmath235 with systems sizes ordered from smallest on bottom to largest on top at @xmath234 .
the difference between the fraction of sites in the two largest clusters , @xmath236 is approximately the spin glass order parameter . as the system size increases
, this difference decreases below the transition suggesting that @xmath237 for @xmath238 in the thermodynamic limit . on the other hand ,
the sum of the two largest clusters is quite constant independent of system size . near the transition ,
approximately 96% of the sites are in the two largest clusters .
( middle set ) , @xmath225 ( bottom set ) and @xmath233 ( top set ) vs.@xmath10 for the cmr graphical representation for the 3d ea model . ] the large fraction of sites in the two largest clusters makes the cmr cluster moves inefficient . if all sites were in the two largest clusters then the cluster moves would serve only to flip all spins in one or both clusters or exchange the identity of the two replicas .
equilibration depends on the small fraction of spins that are not part of the two largest clusters .
figure [ fig : fkwrap ] show the average number of wrapping trfk clusters @xmath239 as a function of inverse temperature .
the largest system size is on the bottom for the small @xmath10 and on top for the large @xmath10 .
as for the case of cmr clusters , the data suggests a transition at some @xmath240 from zero to two wrapping trfk clusters . although the number of trfk wrapping clusters is significantly less than two for all @xmath10 and all system sizes , the trend in system size suggests that it might approach two for large systems and @xmath241 . vs. @xmath10 for the 3d ea model .
] the percolation signature for both cmr and trfk clusters is qualitatively similar in three dimensions . in both cases two giant clusters with opposite values of the local order parameter
appear at a temperature substantially above the phase transition temperature . in the high temperature phase ,
the two giant clusters have the same density and the phase transition is marked by the onset of different densities of the two clusters .
the numerical evidence , however , suggests that the transition from equal to unequal giant cluster densities is quite broad for the small system sizes explored here .
we begin by noting that our results , even for the sk infinite - range spin glass , are not used to prove a phase transition ( which has already been proved using other techniques @xcite ) .
rather , we use as an input the fact that such a transition exists and that the ea order parameter is zero above @xmath100 and nonzero below , and then show that the transition coincides exactly with the onset of a density difference in doubly occupied fk clusters .
numerical evidence indicates that something similar is occurring in the ea model .
together these complementary approaches shed light on the nature of the spin glass transition , especially from a geometric viewpoint , and suggest a possible framework through which a spin glass phase transition in realistic spin glass models can be finally proved .
we now summarize our results .
we have introduced a random cluster approach for studying phase transitions and broken symmetry in spin glasses , both short- and infinite - range .
we have shown that , unlike for ferromagnetic models , single fk percolation is a necessary but not sufficient condition for broken spin flip symmetry . however , double fk percolation ( with a unique largest cluster ) is sufficient and probably necessary .
( more precisely , it _ is _ necessary in the sk model , because otherwise the ea order parameter is zero ; and it is probably necessary in the ea model . ) in the sk model , there is a difference above @xmath100 between the cmr and trfk approaches , but not below . in the former , there
already is double percolation below @xmath242 , above an onset inverse temperature of @xmath243 , below which there are no giant clusters . between this inverse temperature and @xmath242 ( which equals one in the class of models we study )
there are exactly two giant clusters of equal density , which become @xmath244 in the @xmath245 limit .
more importantly , there is a second transition occurring at exactly the sk spin glass critical value @xmath246 . above this threshold ,
the two giant clusters take on unequal densities , whose sum is one ( i.e. , every bond and spin belong to one of the two giant clusters )
. it could be the case at some even higher @xmath10 there is only a single giant cluster , but our methods so far are unable to determine whether this is the case . for trfk percolation
there are no giant clusters above @xmath247 . for @xmath248
there are exactly two giant clusters with unequal densities , and the picture then becomes similar to that of the cmr representation .
the numerical simulations of the 3d ea model suggest a scenario similar to what we find in the sk model .
we observe a sharp percolation transition for both the cmr and trfk representations at a temperature well above the spin glass transition temperature . at this percolation transition ,
two giant clusters form with opposite values of the ea order parameter .
these clusters are nearly equal in density and together occupy most of the system .
as the temperature is decreased toward the presumed location of the spin glass transition , the density of the two clusters becomes increasingly unequal .
although this transition in the density of the two largest clusters appears quite rounded for the small systems investigated numerically , we believe it is sharp in the thermodynamic limit . the results for both the sk model and the connection between the ea order parameter and the density difference between giant clusters strongly suggests that the spin glass transition in finite dimensions is marked by the onset of this density difference in both the cmr and trfk representations .
these results provide an interesting geometric avenue for investigating the phase transition in the ea model and related models with frustration such as the random bond ising model or the potts spin glass .
jm was supported in part by nsf dmr-0242402 .
cmn was supported in part by nsf dms-0102587 and dms-0604869 .
dls was supported in part by nsf dms-0102541 and dms-0604869 .
simulations were performed on the courant institute of mathematical sciences computer cluster .
cmn and dls thank the organizers of the 2007 paris summer school `` spin glasses '' , from which these lectures were adapted .
jm and dls thank the aspen center for physics , where some of this work was done .
m. e. fisher and r. r. p. singh .
critical points , large - dimensionality expansions and ising spin glasses . in g.
grimmett and d. j. a. welsh , editors , _ disorder in physical systems _ , pages 87111 .
clarendon press , oxford , 1990 .
j. machta , c. m. newman , and d. l. stein .
percolation in the sherrington - kirkpatrick spin glass . in v.
sidoravicious and m. e. vares , editors , _ progress in probability vol .
60 : in and out of equilibrium 2_. birkhuser , boston , arxiv:0710.1399 . c. m. newman and d. l. stein .
short - range spin glasses : results and speculations . in e.
bolthausen and a. bovier , editors , _ spin glasses ( lecture notes in mathematics , v. 100 ) _ , pages 159175 .
springer , berlin , 2007 . | the magnetically ordered , low temperature phase of ising ferro- magnets is manifested within the associated fortuin - kasteleyn ( fk ) random cluster representation by the occurrence of a single positive density percolating cluster . in this paper , we review our recent work on the percolation signature for ising spin glass ordering both in the short - range edwards - anderson ( ea ) and infinite - range sherrington - kirkpatrick ( sk ) models within a two - replica fk representation and also in the different chayes - machta - redner two - replica graphical representation .
numerical studies of the @xmath0 ea model in dimension three and rigorous results for the sk model are consistent in supporting the conclusion that the signature of spin - glass order in these models is the existence of a single percolating cluster of maximal density normally coexisting with a second percolating cluster of lower density . |
the higgs sector is unknown , because no higgs boson has been discovered yet . in the standard model ( sm ) for elementary particles , only one scalar iso - doublet field
is introduced to break the electroweak gauge symmetry spontaneously . however , as the various models beyond the sm predict the extended higgs sector , there is a possibility of non - minimal higgs sectors .
the non - minimal higgs sectors suffer from the constraints from the rho parameter and the flavor changing neutral current ( fcnc ) in general . in the sm , these constraints are automatically satisfied . it is known that in the higgs sector with only doublets , the rho parameter is predicted to be unity at the tree level .
therefore , two higgs doublet models ( thdms ) would be a simplest viable extension of the sm . however , in the thdm the most general yukawa interaction predicts fcnc at the tree level , because both the doublet couples to a fermion so that the mass matrix and the yukawa matrix can not be diagonalized simultaneously . in order to avoid this
, a discrete symmetry may be introduced under which different properties are assigned to each scalar doublet @xcite . under this symmetry ,
each fermion couples with only one scalar doublet , and hence there is no fcnc at the tree level even in the thdm .
there are four types of yukawa interactions depending on the @xmath1-charge assignments ; i.e. , type - i , ii , x and y. an interesting possibility would be the type - x thdm , where one higgs doublet couples with quarks and the other does with leptons @xcite .
the type - x thdm can appear in the higgs sector of a gauged extension of the type - iii seesaw model @xcite , the model of the three - loop neutrino mass with electroweak baryogenesis @xcite and a model for the positron cosmic ray anomaly @xcite .
the higgs potential of the thdm is defined as @xcite @xmath2 , \label{eq : higgspot } % -------------------------------------------------------------------------------\end{aligned}\ ] ] where @xmath3 are the higgs doublets with hypercharge @xmath4 .
a softly broken @xmath1 symmetry is imposed in the model to forbid fcnc at the tree level , under which the higgs doublets are transformed as @xmath5 and @xmath6 @xcite .
the soft - breaking parameter @xmath7 and the coupling constant @xmath8 are complex in general .
we here take them to be real assuming the cp invariant higgs sector .
the higgs doublets can be written in terms of the component fields as @xmath9 where the vacuum expectation values ( vevs ) @xmath10 and @xmath11 satisfy @xmath12 gev and @xmath13 .
the mass eigenstates are obtained by rotating the component fields as @xmath14 where @xmath15 and @xmath16 are the nambu - goldstone bosons , @xmath17 , @xmath18 , @xmath19 and @xmath20 are respectively two cp - even , one cp - odd and charged higgs bosons , and @xmath21 the eight parameters @xmath22@xmath7 and @xmath23@xmath8 are replaced by the vev @xmath24 , the mixing angles @xmath25 and @xmath26 , the higgs boson masses @xmath27 and @xmath28 , and the soft @xmath1 breaking parameter @xmath29 .
the coupling constants of the cp - even higgs bosons with weak gauge bosons @xmath30 and @xmath31 are proportional to @xmath32 and @xmath33 , respectively .
when @xmath34 , only @xmath17 couples to the gauge bosons while @xmath18 decouples .
we concentrate on this limit ( the sm - like limit ) where @xmath17 behaves as the sm higgs boson @xcite . imposing the transformation under the @xmath1 parity for leptons and quarks as , @xmath35 , @xmath36 , @xmath37 , @xmath38 and @xmath39 ,
we could write down the type - x yukawa interaction @xcite ; @xmath40 in the type - x thdm , more than @xmath41 of @xmath18 and @xmath19 decay into pairs of tau leptons for @xmath42 in the sm - like limit ; @xmath43 @xcite .
the neutral higgs bosons would be produced in pair by @xmath44 process at the large hadron collider ( lhc ) and by @xmath45 process at the international linear collider ( ilc ) .
these higgs bosons predominantly decay into a four-@xmath46 state , @xmath47 , which is the characteristic signal of the type - x thdm . there would be a clear signature in the dimuon channel from the direct decay of the higgs bosons , @xmath48 .
although the number of events is only @xmath49 of the four tau lepton channel , this channel would be important to measure the mass of the higgs bosons at the lhc .
experimental constraints on masses of the neutral higgs bosons @xmath18 ,
@xmath19 in thdms depend on the type of the yukawa interaction . in the type - ii thdm with large @xmath26
, stronger mass bounds can be obtained from these production processes at the tevatron and the lhc @xcite . however , if the yukawa interaction is the lepton specific which is realized in the wide parameter space in the type - x thdm , these higgs bosons are less constrained due to the relatively weak yukawa interaction with quarks . the search for such higgs bosons at the lep experiments
is found in ref . @xcite .
in our analysis , we use the collinear approximation to calculate the four momenta of the tau leptons@xcite .
if tau leptons are energetic , the missing momentum from its decay would be along the direction of the charged track ( either a charged hadron ( hadrons ) or a charged lepton ) , @xmath50 , where @xmath51 , @xmath52 are the momenta of the neutrino and the charged track , respectively
. the proportionality constant @xmath53 can be determined by fixing @xmath54 . accordingly , the momentum of the decaying tau lepton can be approximately reconstructed as @xmath55 , where @xmath16 is the momentum fraction of the charged track from the parent tau lepton . at hadron colliders , the transverse components of the missing momentum @xmath56 can be measured as the negative sum of the visible momenta .
assuming that the missing transverse momentum of the event is accounted solely by the missing particles in the decays of tau leptons , and applying the collinear approximation for two tau leptons , the missing transverse momentum can be expressed by the momenta of charged tracks , as @xmath57 .
unknown parameters @xmath58 and @xmath59 are determined by solving simultaneous equations . using the resulting values of @xmath60 and @xmath61 ,
the invariant mass of the tau lepton pair is related with that of the tau - jet pair as @xmath62 . at @xmath63 colliders ,
neutral higgs boson pair can be produced via @xmath64 , and the four momenta of the four tau leptons are completely solved @xcite .
the signal events are generated by using pythia @xcite , where the decay of tau leptons is simulated by using tauola @xcite .
initial - state - radiation ( isr ) and final - state - radiation ( fsr ) effects are included .
we choose the collision energy to be @xmath66 tev , and use the cteq6l parton distribution functions @xcite .
we set the masses of extra higgs bosons to @xmath67 gev , @xmath68 gev .
the total cross section for @xmath69 is estimated to be @xmath70 fb at the tree level @xcite .
for the lhc study , background events for @xmath71 ( @xmath72 , @xmath73 and @xmath74 ) , @xmath75 processes where the weak bosons decay leptonically and hadronically , and @xmath76jets processes followed by leptonic decays of weak bosons are generated by pythia , where the decays of tau leptons are also handled by tauola .
the total cross sections for these processes are given as @xmath77 pb , @xmath78 pb and @xmath79 nb , respectively for @xmath71 , @xmath75 and @xmath80+jets production processes by pythia .
we identify the tau - jet candidates by the following criteria ; a jet with @xmath81 gev and @xmath82 which contains @xmath83 or @xmath84 charged hadrons in a small cone ( @xmath85 ) centered at the jet momentum direction with the transverse energy deposit to this small cone more than @xmath86 of the jet . in order to evaluate the signal significance
, we use the significance estimator @xmath87 defined as @xcite @xmath88 } , % -------------------------------------------------------------------------------\end{aligned}\ ] ] where @xmath89 and @xmath90 represent the numbers of signal and background events , respectively .
the significance @xmath87 is proportional to the square root of the integrated luminosity .
.table for background reductions in the @xmath65 channel .
listed are the expected number of events for the integrated luminosity of 100 fb@xmath91 at the lhc with @xmath92 tev .
[ cols="^,^,^,^,^,^",options="header " , ] the results of the signal / background reduction are summarized in table [ tab : ilc_4 t ] .
the expected numbers of events are normalized for the integrated luminosity of @xmath93 fb@xmath91 for each process .
we here focus on the hadronic decay mode of all tau leptons . in general , the mixture of the hadronic and the leptonic decay modes can be analysed . and the significance can be improved by combining the all channels . in order to construct the invariant mass of the tau lepton pair from four tau leptons , we choose the combination of the opposite signed tau leptons which gives the highest @xmath94 pair .
the signal events are dominant even at the pre - selection level .
the statistical significance can be further optimized by using the kinematical cuts giving the much better @xmath95 ratio . in order to test the signal with @xmath96 ,
we only need the integrated luminosity of about @xmath97 fb@xmath91 where we only use the @xmath98 channel .
we have presented the simulation study of the tau lepton specific higgs bosons at the lhc and the ilc . in the thdm - x with the sm - like limit
, the additional higgs bosons can be the tau lepton specific .
such scalar bosons can be pair produced by the gauge interaction at the lhc and the ilc , and mainly decay into tau leptons in the wide range of the parameter space . by using the collinear approximation
, we show that multi - tau lepton final state @xmath99 at the lhc and @xmath100 at the ilc can be a clean signal . the tau lepton specific higgs boson can be tested at the lhc with about @xmath101 fb@xmath91 of the integrated luminosity for @xmath96 .
although the huge integrated luminosity is required , the precise mass determination is possible by extracting the primary muon from the higgs boson decay in the @xmath65 channel .
the search potential of the ilc with @xmath98 channel is about @xmath102 times better than that of the lhc with the @xmath65 channel in the sense of the integrated luminosity .
since the @xmath98 channel can be fully reconstructed by the collinear approximation , the mass of higgs bosons can also be measured .
the work of s.k . was supported in part by grant - in - aid for scientific research , japan society for the promotion of science ( jsps ) , nos .
22244031 and 23104006 . the work of k.t .
was supported in part by the national science council of taiwan under grant no .
nsc 100 - 2811-m-002 - 090 .
the work of h.y . was supported in part by the national science council of taiwan under grant no .
nsc 100 - 2119-m-002 - 001 .
v. d. barger , j. l. hewett and r. j. n. phillips , phys .
d * 41 * , 3421 ( 1990 ) ; y. grossman , nucl .
b * 426 * , 355 ( 1994 ) ; a. g. akeroyd , phys .
b * 377 * , 95 ( 1996 ) .
m. aoki , s. kanemura , k. tsumura and k. yagyu , phys . rev . d * 80 * , 015017 ( 2009 ) .
v. barger , h. e. logan and g. shaughnessy , phys .
d * 79 * , 115018 ( 2009 ) ; s. su and b. thomas , phys .
d * 79 * , 095014 ( 2009 ) ; h. e. logan and d. maclennan , phys .
d * 79 * , 115022 ( 2009 ) .
e. ma , mod .
lett . a * 17 * , 535 ( 2002 ) ; e. ma and d. p. roy , nucl .
phys . b * 644 * , 290 ( 2002 ) .
m. aoki , s. kanemura and o. seto , phys .
lett . * 102 * , 051805 ( 2009 ) ; phys .
d * 80 * , 033007 ( 2009 ) ; m. aoki , s. kanemura and k. yagyu , phys .
d * 83 * , 075016 ( 2011 ) .
h. -s .
goh , l. j. hall and p. kumar , jhep * 0905 * , 097 ( 2009 ) .
j. f. gunion , h. e. haber , g. kane and s. dawson , the higgs hunter s guide ( frontiers in physics series , addison - wesley , reading , ma , 1990 ) .
s. chatrchyan _ et al .
_ [ cms collaboration ] , phys .
lett . * 106 * , 231801 ( 2011 ) .
s. schael _ et al . _
[ lep working group for higgs boson searches and aleph and delphi and l3 and opal collaborations ] , eur .
j. c * 47 * , 547 ( 2006 ) d. l. rainwater , d. zeppenfeld , k. hagiwara , phys .
rev . * d59 * , 014037 ( 1998 ) .
j. abdallah _ et al .
_ [ delphi collaboration ] , phys .
lett . b * 552 * , 127 ( 2003 ) .
t. sjostrand , s. mrenna , p. z. skands , jhep * 0605 * , 026 ( 2006 ) .
s. jadach , z. was , r. decker and j. h. kuhn , comput .
commun . * 76 * , 361 ( 1993 ) . | we study the feasibility of the type - x two higgs doublet model ( thdm - x ) at collider experiments . in the thdm - x
, new higgs bosons mostly decay into tau leptons in the wide range of the parameter space .
such scalar bosons are less constrained by current experimental data , because of the suppressed quark yukawa interactions .
we discuss a search strategy of the thdm - x with multi - tau lepton final states at international linear collider and large hadron collider . by using the collinear approximation
, we show that a four tau lepton signature @xmath0 can be a clean signal . |
weak decays of heavy hadrons @xcite present a very rich field of phenomena owing to the complexity of confinement . being the bound states of a heavy - quark and light - constituents ( mesons , singly charmed / bottom baryons ) or even of two heavy quarks and one light constituent ( doubly charmed / bottom baryons ) , heavy hadrons contain soft degrees of freedom which generate nonperturbative power corrections , such as the destructive and/or constructive pauli interference and the w - exchange / annihilation between the quark coming from the heavy - quark decay and one of the light constituents .
inclusive decay rates and lifetimes of charmed mesons were reasonably reliably calculated in the last decade .
the overall picture emerging is qualitatively satisfactory and the lifetime hierarchy has been predicted for singly charmed baryons and found to be in agreement with present experiments @xcite .
also , the difference in lifetimes ( a factor of 2 - 3 ) between @xmath4 and @xmath5 mesons , which is due to the negative pauli interference preasymptotic effect , was explained a long time ago @xcite .
the numerical calculations performed in the middle of the eighties @xcite provided us with the predictions of a lifetime pattern which has recently been confirmed by experiment .
this success is rather surprising since with the advent of the systematic operator product expansion ( ope ) @xcite and heavy quark effective theory ( hqet ) @xcite it has become clear that the charmed quark mass is not heavy enough for the @xmath6 expansion to be trustworthy .
nevertheless , it seems that employing systematically field - theory methods to the very end of the calculation ( up to the hadronic wave function for which we have to rely upon some phenomenological models ) , one is able to make clear predictions to be compared with present and future experiments and possibly to disentangle between various preasymptotic effects . on the other hand
, the inverse bottom quark mass appears to be a good expansion parameter in bottom decays .
however , the role of four quark operators is negligible there ( effects of the @xmath7 ) leaving charmed hadron decays as a playground for studying such effects and as a test of the possible violation of the quark - hadron duality . in this paper
we extend the analysis of singly charmed baryons decays and lifetimes @xcite to the case of doubly charmed baryons .
recently @xcite , a rather phenomenological approach using effective constituent quark masses and fit of singly charmed baryon decays has been employed to study doubly charmed baryon decays .
we , however , have used a systematic field - theory approach to the very end in order to be consistent with the previous treatment of singly charmed hadrons and have also included the preasymptotic effects in semileptonic decay rates of doubly charmed baryons , thus calculating all decay rates at the cabibbo subleading level .
we show that preasymptotic effects dramatically change the simple spectator picture and lead to a very clear pattern of semileptonic branching ratios and lifetimes .
using the optical theorem , the inclusive decay width of a hadron @xmath8 with mass @xmath9 containing two heavy @xmath10 quarks can be written as @xmath11 where @xmath12 is the transition operator @xmath13 in the following we use the operator product expansion which is based on the assumption that the energy release in the decay of a @xmath10 quark is large enough .
this implies that momenta flowing through internal lines are also large and therefore justify the ope .
the general formula for the decay is given by @xcite @xmath14 here @xmath15 and @xmath16 are wilson coefficient functions which are known at one - loop order and tree level , respectively @xcite .
v represents appropriate matrix elements of the ckm matrix .
let us calculate the semileptonic decay rates first .
the main contribution is expected to come from the quark decay - type diagram , which is proportional to @xmath17 and is given , to @xmath18 , as @xmath19 here @xmath20 and @xmath21 parametrize the matrix elements of the kinetic energy and the chromomagnetic operators , respectively .
their determination will be discussed later . throughout the paper
we use the abbreviations @xmath22 and @xmath23 for @xmath24 and @xmath25 ( @xmath26 is the cabibbo angle ) .
the next contribution is coming from the dimension - five operator @xmath27 note that in both ( [ eq : sldec ] ) and ( [ eq : slg ] ) there is an additional factor 2 coming from the decays of _ two _ @xmath10 quarks in the doubly charmed baryon .
the phase - space corrections @xmath28 and @xmath29 are cited explicitly in the appendix .
the radiative qcd correction @xmath30 @xcite is given by @xmath31 where for @xmath32 we have @xmath33 and @xmath34 .
leptons are taken to be massless .
recently , voloshin has noticed @xcite that preasymptotic effects in _ semileptonic _ inclusive decays can be very large owing to the constructive pauli interference , the result up to the ckm matrix element being given by @xmath35 here @xmath36 is a correction due to the hybrid renormalization of the effective lagrangian and it takes care of the evolution of @xmath37 from @xmath38 down to the typical hadronic scale @xmath39 .
the factor 5 in front of @xmath40 reflects the spin structure of doubly charmed baryons .
the baryon wave function @xmath41 will be discussed later . the total semileptonic rate for one lepton species is given by @xmath42 where @xmath43 in view of the significant preasymptotic effects in the sl decay rates of singly charmed baryons
, one can expect a large pauli - interference contribution in the semileptonic decay rate of the @xmath2 baryon ( @xmath44 quark structure ) , where that contribution is present at the leading cabibbo level .
nonleptonic decay rates are slightly more complicated , since in the final state the lepton pair is substituted by a quark pair
. the contributions analogous to ( [ eq : sldec ] ) and ( [ eq : slg ] ) are ( including @xmath45 corrections ) @xmath46 \nonumber \\ & \times & [ 1 -\frac{1}{2}\frac{\mu_{\pi}^2(h_{cc})}{m_c^2 } + \frac{1}{2}\frac{\mu_{g}^2(h_{cc})}{m_c^2 } ] \ , , \end{aligned}\ ] ] @xmath47 \nonumber \\ & + & 2 ( c_{+}^2 - c_{-}^2)[((c^4+s^4)p_{2}(x)+ c^2 s^2)\eta_{nl}(x)+\nonumber \\ & & c^2 s^2 \tilde{p}_{2}(x)\tilde{\eta}_{nl}(x))]\ } ( -2 \frac{\mu_{g}^2(h_{cc})}{m_{c}^{2 } } ) \ , . \ ] ] radiative corrections to the nonleptonic decay , @xmath48 and @xmath49 , are far more complicated than analogous corrections ( [ eq : etasl ] ) and ( [ eq : slmass ] ) to the semileptonic decay and the reader is referred to the original paper where they were first calculated @xcite .
again , the preasymptotic effects are expected to contribute significantly to the total nonleptonic decay rate .
they are given by @xmath50\,5|\psi(0)|^2\ , , \nonumber \\ \gamma^{int}_{- } & = & \frac{g_f^2}{2\pi } m_c^2 [ -\frac{1}{2 } c_{+}(2 c_{-}- c_{+ } ) \nonumber \\ & & - \frac{1}{6}(1 - \sqrt{\kappa})(5 c_{+}^2 + c_{-}^2 - 6 c_{+ } c_{- } ) ] \,5|\psi(0)|^2 \,,\nonumber \\ \gamma^{int}_{+ } & = & \frac{g_f^2}{2\pi } m_c^2 [ \frac{1}{2 } c_{+}(2 c_{-}+ c_{+ } ) \nonumber \\ & & - \frac{1}{6}(1 - \sqrt{\kappa})(5 c_{+}^2 + c_{-}^2 + 6 c_{+ } c_{- } ) ] \,5|\psi(0)|^2 \ , .\end{aligned}\ ] ] an explicit calculation leads to the following nonleptonic decay rates : @xmath51 all corrections @xmath52 and @xmath53 are given explicitly in the appendix .
an important remark to be made here concerns the mass parameters in the calculation of the matrix elements @xmath54 and @xmath55 .
whenever we perform an expansion which is essentially a field - theoretic procedure ( either the ope for the transition operator @xmath12 or the hqet expansion in the case of the @xmath56 operator ) , the expansion parameter is always the current heavy - quark running mass @xmath38 . on the other hand , in the calculation of the matrix elements , which is performed within quark models ,
it is more appropriate to use constituent quark masses @xmath57 . following this procedure , we give the expressions for @xmath54 and @xmath55 . for @xmath54 , we have @xmath58 where @xmath59 is the average heavy - quark velocity in the @xmath60 baryon , @xmath61 and @xmath62 are constituent masses of the heavy and the light quark , respectively , and t is the average kinetic energy of the light quark and the heavy diquark .
the precise description of this calculation is given in @xcite and relies upon some phenomenological features of the meson potential .
the contributions to the @xmath55 operator , connected to the matrix element of the chromomagnetic operator , can be divided into two parts .
the first part includes effects coming from the heavy - light chromomagnetic interaction and these contributions can also be found in the singly charmed baryon @xmath63 .
the second part comprises effects originating within the heavy diquark , i.e. heavy - heavy chromomagnetic interactions .
these effects are new @xcite and characteristic of doubly charmed baryons .
their estimation relies upon the nonrelativistic qcd model calculation @xcite .
the final expression is @xmath64 where the first term describes the heavy diquark - light quark hyperfine interaction , while the second and the third correspond to the interaction of two heavy @xmath10-quarks in a diquark state .
they are of the `` chromomagnetic '' and `` darwin '' type , respectively .
in ( [ eq : mug ] ) , @xmath65 is the mass of the doubly charmed baryon , @xmath66 is the mass of its @xmath67 spin counterpart and @xmath68 is the wavefunction of the @xmath3 pair in the heavy diquark , i.e. @xmath69 is the probability for these two heavy quarks to meet at one point . in the calculations above , up to the hadronic matrix elements we have used field theory only .
the results are expressed in terms of the baryon wave function @xmath70 and the matrix elements of the kinetic and chromomagnetic operators , which are @xmath54 and @xmath55 , respectively .
the use of the usual singly charmed baryon wave function @xmath71 , as given in @xcite , would be premature , since intuitively , one expects a two - heavy - quark system to behave differently from the single - heavy - quark one . in the case of singly charmed baryons ,
the heavy quark is `` sitting '' in the centre of the baryon and the other two light quarks are moving around .
their spin and the color charges are correlated ( in order to have the appropriate spin and color structure of the entire baryon ) , but their spatial motion is not . in this way , one has a three - body picture of the baryon containing a single heavy quark and one should use the baryonic wave function @xmath71 accordingly . in the case of doubly charmed baryons , one assumes that two heavy quarks are strongly bound into a color antitriplet state .
as far as the light quark is concerned , the bound state of two @xmath3-quarks appears as a pointlike diquark object @xcite .
thus , in the heavy - quark limit , which can ( presumably ) still be applied in our case ( @xmath72 ) , a doubly charmed baryon appears to consist of a heavy diquark and a light quark forming a `` meson '' state .
therefore , one expects that the wave function of the doubly charmed baryon ( which has to be considered as the light - quark wave function at the origin of the @xmath3-diquark ) behaves essentially as the mesonic wave function .
we use the derivation of hyperfine splittings of mesons calculated in the constituent nonrelativistic quark model by de rjula et al @xcite to obtain the following relation between the wave functions of the doubly charmed baryon and the d - meson : @xmath73 the factor @xmath74 comes from the different spin content of doubly charmed baryons , i.e. the @xmath3-diquark forms the spin-1 color antitriplet state .
the baryonic wave function squared in ( [ eq : psi ] ) is directly proportional to the d - meson decay constant , @xmath75 , squared .
the factor @xmath76 is the effect of the hybrid renormalization which accounts for the fact that @xmath75 is measured at the scale @xmath77 , and one has to evolve @xmath75 down to the hadronic scale @xmath78 .
the choice of the mesonic wave function @xmath79 , instead of the singly charmed baryon wave function @xmath80 , where @xmath81 is the static value of the d - meson decay constant , also seems to be consistent numerically . in figure 1 .
we have displayed the dependence of the @xmath82 on @xmath83 in the large range of the @xmath75 values . in our numerical calculation
we use @xmath84 as a central value .
this value is consistent both with qcd lattice calculations @xcite and qcd sum rules calculations @xcite .
in the case of @xmath85 there is a negative pauli interference which cancels the contribution coming from the decay - type diagram ( [ eq : nldec ] ) and the chromomagnetic operator ( [ eq : nlg ] ) .
this case is very similar to the role of the negative pauli interference in the @xmath4 decay where it competes with the decay - diagram contribution .
for @xmath86 large enough , the nonleptonic and total rates in both @xmath4 and @xmath0 decays become negative , fig.1 .
however , a reasonable choice of @xmath86 gives positive results . in view of these facts and results
, we may conclude that the phenomenological rule of using @xmath75 in mesonic and its static value @xmath81 in baryonic systems , employed first in singly charmed hadrons @xcite , can be successfully extended to the consideration of doubly charmed baryons .
not doing so , but taking @xmath81 instead of @xmath86 would lead us to the unphysical region .
so , our result is a confirmation of the above mentioned phenomenological rule at the same , qualitative level . taking this rule as postulated for singly charmed hadrons
, we can interpret our results as an extension of the same rule into the doubly charmed sector .
also , we can generalize the rule to some extent .
we see that the use of @xmath86 is required in singly charmed mesons and doubly charmed baryons , while @xmath81 is used in singly charmed baryons @xcite .
so , it is allowed to say that @xmath86 should be used in systems with two - body dynamics ( in the @xmath60 baryon case , a heavy diquark and a light quark ) and @xmath81 in systems with three - body dynamics .
it is important to stress that these considerations and conclusions are of purely phenomenological origin , i.e. they have no direct justification in field theory .
in numerical calculations we use the following set of parameters , which closely follows the set used in @xcite . for @xmath87 , the wilson coefficients are @xmath88 and @xmath89 . the charmed quark mass is taken to be @xmath90 and for the strange quark mass we use @xmath91 .
the value of the average kinetic energy @xmath92 , appearing explicitly in ( [ eq : mupi ] ) , is taken from @xcite to be @xmath93 , and the light and heavy - quark constituent masses are @xmath94 and @xmath95 , respectively .
the numerical value of the diquark wave function is also taken from @xcite to be @xmath96 .
the numerical values for masses of doubly charmed baryons are taken from @xcite .
as far as the @xmath97 dependence is concerned , in the range @xmath98 the lifetimes of @xmath1 and @xmath2 are practically constant and the lifetime of @xmath0 is more sensitive to the value of @xmath97 , being somewhat @xmath99 larger for @xmath100 .
the same is true for the @xmath101-dependence in the reasonable range @xmath102 .
the lifetimes of @xmath1 and @xmath2 stay almost constant with variation of @xmath101 , and the lifetime of @xmath0 grows slowly with @xmath101 ( by @xmath103 ) , fig.2 .
@xmath104 & @xmath105 & @xmath106 + + @xmath107 & 0.345 & 4.158 & 2.859 + + @xmath108 & 0.151 & 0.173 & 0.603 + + @xmath109 & 23.4 & 3.9 & 14.9 + + @xmath110 & 1.55 & 0.22 & 0.25 + from table 1 one can see that the voloshin type of preasymptotic corrections in the semileptonic decay rates of @xmath2 is significant , contributing at the cabibbo leading level , eq .
( [ eq : slvol ] ) .
this contribution brings @xmath111 to be four times larger than @xmath112 , which receives contributions only from ( [ eq : sldec ] ) and ( [ eq : slg ] ) . in the @xmath113
there is the pauli interference effect at the cabibbo suppressed level , but it still makes the rate larger by @xmath114 than that for the @xmath0 baryon . clearly , since both semileptonic and nonleptonic rates are significantly affected by the large preasymptotic effects which are proportional to latexmath:[$|\psi(0 ) branching ratio for @xmath2 depend crucially on the choice of @xmath86 .
the latter is obvious from fig.3 , where
@xmath111 grows linearly with @xmath116 and especially from the second picture in fig .
3 , where it is clear that @xmath117 shows instability for @xmath86 larger than @xmath118 owing to the large cancelation between the negative pauli interference term and the contributions from eqs . ( [ eq : nldec ] ) and ( [ eq : nlg ] ) .
this is a clear signal that one should not take the results for @xmath0 too literally . keeping the above remarks in mind ,
we predict the following pattern for semileptonic branching ratios : @xmath119 and the following pattern for the lifetimes : @xmath120 we can compare our results with the calculations of lifetimes of @xmath0 and @xmath1 , which appeared recently @xcite .
the authors of that paper employed a similar field - theory technique , but had a different approach to the choice of relevant parameters . throughout their paper
they used the constituent heavy - quark mass as an expansion parameter , which is a phenomenological procedure that we do not find fully consistent . in the calculation of semileptonic decay rates
, they did not include large preasymptotic effects which significantly change total semileptonic widths .
comparison shows that our numerical results are significantly different from those of kiselev et .
al . @xcite .
finally , it is worth discussing briefly the decay of the heaviest weakly decaying charmed hadron , the triply charmed baryon @xmath121 .
although its complicated structure and its intrinsic tree - body motion prevent us from applying to this particle the heavy - light picture as described above to other weakly decaying heavy hadrons , it is possible to give some qualitative predictions for the @xmath121 decay rate and the lifetime . in this baryon , preasymptotic effects ( giving large contributions in the singly and the doubly case ) do not exist for lack of light valence quarks .
thus , the dominant contribution comes from the operators of dimensions three and five . since in
doubly charmed decays the contribution of dimension - five operators represents less than @xmath122 of the contribution of the decay ( dimension three ) operator , it seems reasonable to approximate the total decay width of @xmath121 with the triple @xmath10-quark decay contribution and estimate the error of disregarding dimension - five operators at the level of @xmath122 . in this case
, the expression for @xmath123 can be obtained by multiplying the expressions ( [ eq : sldec ] ) , ( [ eq : slg ] ) , ( [ eq : nldec ] ) and ( [ eq : nlg ] ) by a factor of @xmath67 , summing them and taking the limit @xmath124 and @xmath125 .
the numerical value for the lifetime is @xmath126 as the calculation of dimension - five operator contributions in triply charmed baryon decay rates is out of scope of the present paper , this result can be considered only as qualitatively correct .
application of the heavy - quark expansion to the problem of inclusive decays of doubly charmed baryons enables us to give very interesting predictions for their lifetimes and semileptonic branching ratios .
large lifetime differences are present between @xmath0 on the one hand and @xmath1 and @xmath2 on the other .
our numerical results pick out @xmath0 as the longest living charmed particle ( fig.4 ) , although the numerical value for @xmath117 should be taken with certain reserve for reasons already mentioned .
such a large numerical difference within the lifetime hierarchy makes these predictions suitable for testing by the first forthcoming experimental observation of doubly charmed baryons .
a theoretical prediction for semileptonic branching ratios is even clearer and the hierarchy of @xmath109 is unambiguously determined .
the total hierarchy of lifetimes for charmed hadrons is shown in figure 4 .
it is evident that charmed hadrons show a very complex pattern in the @xmath127 plane .
one can note that doubly charmed baryon lifetimes are comparable with those of their singly charmed counterparts .
this result is just opposite to the naive expectation of roughly double widths in the doubly charmed case owing to the decay of two , instead of one @xmath10 quark , or , correspondingly twice smaller lifetimes of @xmath60 baryons .
however , the mesonic nature of doubly charmed baryon wave functions , where one uses the smaller @xmath75 constant instead of its static @xmath128 value , reduces four - quark operator contributions , increasing doubly charmed baryon lifetimes . although the @xmath10 quark is considered not being heavy enough to ensure a reasonable convergence of the heavy - quark expansion series , the numerical results in the singly charmed sector show satisfactory qualitative and even quantitative agreement with experiment @xcite .
if future experiments concerning the doubly ( and triply ) charmed sector should show similar agreement with our theoretical predictions , that would have important implications on the role of four - quark operators and the entire theory of heavy quark expansion . besides , the absence or the presence of agreement might have notable implications on the validity of the quark - hadron duality in the charmed sector .
phase - space corrections can be understood as reduction of particle phase space due to the propagation of the massive particle in the loops of diagrams describing inclusive decays . in our case , we consider the @xmath129 quark to be massive , while the other particles ( @xmath130,@xmath131 quarks and leptons ) are treated as massless .
these corrections can be classified according to the type of the the operator diagram in which they appear and according to the number of massive quarks in the loop ( or in a final state if we consider an inclusive process as a sum of exclusive channels ) .
* one massive quark in the loop : + @xmath133 + @xmath134 + @xmath135 + @xmath136 appears as a correction to the decay - type diagram , while @xmath137 and @xmath138 come as corrections to the chromomagnetic operator .
* two massive quarks in the loop : + using the notation @xmath139 + we have @xmath140 + @xmath141 + @xmath142 @xmath149 appears as a correction to the exchange diagram , @xmath150 corrects for the massive quark in the interference contributions , while @xmath151 is a correction in the case of negative interference when there are two massive quarks in the loop @xcite .
99 b. blok , m. shifman , `` lifetimes of charmed hadrons revised - facts and fancy . '' in proc . of the workshop on the tau - charm factory , p.247 , marbella , spain , 1993 , eds .
j. kirkby and r. kirkby ( editions frontiers , gif - sur - yvette , 1994 ) , tpi - minn-93/55-t , umn - th-1227/93 m. neubert , `` b decays and the heavy - quark expansion '' , preprint cern - th/97 - 24 and hep - ph/9702375 , to appear in `` heavy flavors ii '' , eds .
buras and m. lindner i. bigi , `` the qcd perspective on lifetimes of heavy flavour hadrons '' , preliminary version of the physics report paper , und - hep-95-big02 and hep - ph/9508408 review of particle properties , eur .
j. c*3 * ( 1998 ) b. guberina , s. nussinov , r.d .
peccei , r. rckl , phys .
b*89 * , 111 ( 1979 ) n. bili ' c , b. guberina , j. trampeti ' c , nucl .
b*248 * , 261 ( 1984 ) m. shifman , m. voloshin , yad . phys . * 41 * , 463 ( 1985 ) [ sov . j. nucl . phys .
* 41 * , 120 ( 1985 ) ] b. guberina , r. rckl , j. trampeti ' c , z. phys .
c*33 * , 297 ( 1986 ) m. shifman , m. voloshin , zhetf * 91 * , 1180 ( 1986 ) [ jetp * 64 * , 698 ( 1986 ) m. neubert , phys . rep . *
245 * , 259 ( 1994 ) ; int .
a*11 * , 4173 ( 1996 ) b. guberina , b. meli ' c , eur .
j. c*2 * , 697 ( 1998 ) v.v .
kiselev , a.k .
likhoded , a.i .
onischenko , hep - ph/9807354 c.s .
kim , a.d .
martin , phys .
lett b*225 * , 186 ( 1989 ) y. nir , phys .
b*221 * , 184 ( 1989 ) m.b .
voloshin , phys .
b*385 * , 369 ( 1996 ) e. bagan , p. ball , v.m .
braun , p. gosdzinsky : nucl .
b*432 * , 3 ( 1994 ) m. beneke , g. buchalla , phys . rev . d*53
* , 4991 ( 1996 ) g.p .
lepage , l. magnea , c. nakhleh , u. magnea , k. hornbostel , phys .
rev . d*46 * , 4052 ( 1992 ) w.e .
caswell , g.p .
lepage , phys .
b*167 * , 437 ( 1986 ) m.j .
savage , m.b .
wise , phys .
b*248 * , 177 ( 1990 ) m.j .
white , m.j .
savage , phys .
b*271 * , 410 ( 1991 ) m.j .
savage , r.p .
springer , int .
a*6 * , 1701 ( 1991 ) a.
de rjula , h. georgi , s.l .
glashow , phys .
rev . d*12 * , 147 ( 1975 ) j.l .
cortes , j. sanchez guillen , phys .
d*24 * , 2982 ( 1981 ) h. wittig , int . j. mod .
a*12 * , 4477 ( 1997 ) m. neubert , phys .
d*45 * , 2451 ( 1992 ) s. narrison , phys .
b*322 * , 247 ( 1994 ) d. ebert , r.n .
faustov , v.o .
galkin , a.p .
martynenko , v.a .
saleev , z. phys .
c*76 * , 111 ( 1997 ) j.g .
krner , d. pirjol , m. krmer , prog .
phys.*33 * , 787 ( 1994 ) j.l .
cortes , x.y .
pham , a. tounsi , phys .
d*25 * , 188 ( 1982 ) q. hokim , x.y .
pham , phys .
b*122 * , 297 ( 1983 ) | the analysis of singly charmed hadrons has been extended to the case of doubly charmed baryons , @xmath0 , @xmath1 and @xmath2 .
doubly charmed baryons are described as a system containing a heavy @xmath3-diquark and a light quark , similarly as in a heavy - light meson .
this leads to preasymptotic effects in semileptonic and nonleptonic decays which are essentially proportional to the meson wave function .
interplay between preasymptotic effects in semileptonic and/or nonleptonic decay rates leads to very clear predictions for semileptonic branching ratios and lifetimes of doubly charmed baryons .
pacs numbers : 12.38.lg , 13.30.ce , 13.30.eg , 14.20.lq |
autoregressive moving average ( arma ) models have featured prominently in the analysis of time series .
the versions initially stressed in the theoretical literature ( e.g. , @xcite ) are stationary and invertible . following @xcite , unit root nonstationarity
has frequently been incorporated , while `` overdifferenced '' noninvertible processes have also featured .
stationary arma processes automatically have short memory with `` memory parameter , '' denoted @xmath1 , taking the value zero , implying a huge behavioral gap relative to unit root versions , where @xmath2 .
this has been bridged by `` fractionally - differenced , '' or long memory , models , a leading class being the fractional autoregressive integrated arma ( farima ) . a farima @xmath3 process @xmath4 is given by @xmath5 where @xmath6 is the observable series ; @xmath7 is the lag operator ; @xmath8;@xmath9 with @xmath10 for @xmath11 and by convention @xmath12 ; @xmath13 is the indicator function ; @xmath14 and @xmath15 are real polynomials of degrees @xmath16 and @xmath17 , which share no common zeros , and all of their zeros are outside the unit circle in the complex plane ; and the @xmath18 are serially uncorrelated and homoscedastic with zero mean .
the reason ( [ a ] ) features the truncated process @xmath19 rather than simply @xmath20 is to simultaneously cover @xmath1 falling in both the stationary region @xmath21 and the nonstationary region ( , where otherwise the process would `` blow up '' ) . in the former case
, the truncation implies that @xmath22 is only `` asymptotically stationary . '' in recent years , fractional modeling has found many applications in the sciences and social sciences ; for example , with respect to environmental and financial data . early work on asymptotic statistical theory for fractional models assumed @xmath23 [ and replaced @xmath24 by @xmath20 in ( [ a ] ) ] .
assuming @xmath25 , @xcite and @xcite showed consistency and asymptotic normality of whittle estimates ( of @xmath1 and other parameters , such as the coefficients of @xmath26 and @xmath27 ) , thereby achieving analogous results to those of @xcite for stationary arma processes [ i.e. , ( [ aaaa ] ) with @xmath28 and other short memory models .
more recently , @xcite considered empirical maximum likelihood inference covering this setting .
note that @xcite and @xcite , and much other work , not only excluded @xmath29 but also the short - memory case @xmath30 , as well as negatively dependent processes where @xmath31 . to some degree
, other @xmath1 can be covered , for example , for @xmath32 one can first - difference the data , apply the methods and theory of @xcite and @xcite , and then add 1 to the memory parameter estimate , but this still requires prior knowledge that @xmath1 lies in an interval of length no more than @xmath33 . on the other hand
, @xcite argued that the same desirable properties should hold without so restricting @xmath1 , in case of a conditional - sum - of - squares estimate , and this would be consistent with the classical asymptotic properties established by @xcite for score tests for a unit root and other hypotheses against fractional alternatives , by comparison with the nonstandard behavior of unit root tests against stationary autoregressive alternatives .
however , the proof of asymptotic normality in @xcite appears to assume that the estimate lies in a small neighborhood of @xmath1 , without first proving consistency ( see also @xcite ) . due to a lack of uniform convergence , consistency of this implicitly - defined estimate
is especially difficult to establish when the set of admissible values of @xmath34 is large . in particular , this is the case when @xmath1 is known only to lie in an interval of length greater than @xmath33 .
in the present paper , we establish consistency and asymptotic normality when the interval is arbitrarily large , including ( simultaneously ) stationary , nonstationary , invertible and noninvertible values of @xmath35 .
thus , prior knowledge of which of these phenomena obtains is unnecessary , and this seems especially practically desirable given , for example , that estimates near the @xmath36 or @xmath2 boundaries frequently occur in practice , while empirical interest in autoregressive models with two unit roots suggests allowance for values in the region of @xmath37 also , and ( following @xcite ) antipersistence and the possibility of overdifferencing imply the possibility that @xmath31 . we in fact consider a more general model than ( [ a ] ) , ( [ aaaa ] ) , retaining ( [ a ] ) but generalizing ( [ aaaa ] ) to@xmath38 where @xmath18 is a zero - mean unobservable white noise sequence
, @xmath39 is an unknown @xmath40 vector , @xmath41 , where for all @xmath42 , @xmath43 , @xmath44 is continuous in @xmath45 and @xmath46 , @xmath47 .
more detailed conditions will be imposed below .
the role of @xmath48 in ( [ b ] ) , like @xmath26 and @xmath27 in ( [ aaaa ] ) , is to permit parametric short memory autocorrelation .
we allow for the simplest case farima@xmath49 by taking @xmath39 to be empty . another model covered by ( [ b ] )
is the exponential - spectrum one of bloomfield ( which in conjunction with fractional differencing leads to a relatively neat covariance matrix formula @xcite ) .
semiparametric models ( where @xmath20 has nonparametric autocovariance structure ; see , e.g. , @xcite ) afford still greater flexibility than ( [ b ] ) , but also require larger samples in order for comparable precision to be achieved . in more moderate - sized samples , investment in a parametric model can prove worthwhile , even the simple farima(@xmath50 , @xmath1 , @xmath51 ) employed in the monte carlo simulations reported in the supplementary material @xcite , while model choice procedures can be employed to choose @xmath16 and @xmath17 in the farima(@xmath52 ) , as illustrated in the empirical examples included in the supplementary material @xcite .
we wish to estimate @xmath53 from observations @xmath54 , @xmath55 . for any admissible @xmath56 , define @xmath57 noting that ( [ a ] ) implies @xmath58 , @xmath59 . for a given
user - chosen optimizing set @xmath60 , define as an estimate of @xmath61@xmath62 where @xmath63 and @xmath64 , where @xmath65 for given
@xmath66 , @xmath67 such that @xmath68 , @xmath69 is a compact subset of @xmath70 and @xmath71 .
the estimate @xmath72 is sometimes termed `` conditional sum of squares '' ( though `` truncated sum of squares '' might be more suitable ) . it has the anticipated advantage of having the same limit distribution as the maximum likelihood estimate of @xmath61 under gaussianity , in which case it is asymptotically efficient ( though here we do not assume gaussianity ) .
it was employed by @xcite in estimation of nonfractional arma models ( when @xmath1 is a given integer ) , by @xcite in stationary farima models , where @xmath73 , and by @xcite in nonstationary farima models , allowing @xmath74 .
the following section sets down detailed regularity conditions , a formal statement of asymptotic properties and the main proof details
. section [ sec3 ] provides asymptotically normal estimates in a multivariate extension of ( [ a ] ) , ( [ b ] ) .
joint modeling of related processes is important both for reasons of parsimony and interpretation , and multivariate fractional processes are currently relatively untreated , even in the stationary case .
further possible extensions are discussed in section [ sec4 ] .
useful lemmas are stated in section [ sec5 ] . due to space restrictions
, the proofs of these lemmas , along with an analysis of finite - sample performance of the procedure and an empirical application , are included in the supplementary material @xcite .
our first two assumptions will suffice for consistency . 1 .
a. @xmath75 for all @xmath76 , @xmath77 , on a set @xmath78 of positive lebesgue measure ; b. for all @xmath42 , @xmath79 is differentiable in @xmath80 with derivative in @xmath81 , @xmath82 c. for all @xmath80 , @xmath79 is continuous in @xmath83 d. for all @xmath84 , @xmath85 . condition ( i ) provides identification while ( ii ) and ( iv ) ensure that @xmath20 is an invertible short - memory process ( with spectrum that is bounded and bounded away from zero at all frequencies ) .
further , by ( ii ) the derivative of @xmath86 has fourier coefficients @xmath87 as @xmath88 , for all @xmath42 , from page 46 of @xcite , so that , by compactness of @xmath69 and continuity of @xmath89 in @xmath42 for all @xmath90 , @xmath91 also , writing @xmath92 , we have @xmath93 for all @xmath42 , and ( ii ) , ( iii ) and ( iv ) imply that @xmath94 finally , ( ii ) also implies that@xmath95 assumption a1 is easily satisfied by standard parameterizations of stationary and invertible arma processes ( [ aaaa ] ) in which autoregressive and moving average orders are not both over - specified .
more generally , a1 is similar to conditions employed in asymptotic theory for the estimate @xmath96 and other forms of whittle estimate that restrict to stationarity ( see , e.g. , @xcite ) and not only is it readily verifiable because @xmath48 is a known parametric function , but in practice @xmath48 satisfying a1 are invariably employed by practitioners . 1 .
the @xmath18 in ( [ b ] ) are stationary and ergodic with finite fourth moment , and@xmath97 almost surely , where @xmath98 is the @xmath99-field of events generated by @xmath100 , @xmath101 , and conditional ( on @xmath102 ) third and fourth moments of @xmath18 equal the corresponding unconditional moments .
assumption a2 avoids requiring independence or identity of distribution of @xmath103 , but rules out conditional heteroskedasticity .
it has become fairly standard in the time series asymptotics literature since @xcite .
let ( [ a ] ) , ( [ b ] ) and hold .
then as @xmath104 @xmath105 we give the proof for the most general case where @xmath106 , but our proof trivially covers the @xmath107 situation , for which some of the steps described below are superfluous .
the proof begins standardly . for @xmath108 , define @xmath109 , @xmath110 . for small enough @xmath111,@xmath112 where @xmath113 .
the remainder of the proof reflects the fact that @xmath114 , and thus @xmath115 , converges in probability to a well - behaved function when @xmath116 , and diverges when @xmath117 , while the need to establish uniform convergence , especially in a neighborhood of @xmath118 , requires additional special treatment .
consequently , for arbitrarily small , such that @xmath119 , we define the nonintersecting sets @xmath120 , @xmath121 , @xmath122 , @xmath123 .
correspondingly , define @xmath124 , @xmath125 , so @xmath126 .
thus , from ( [ new1 ] ) it remains to prove @xmath127 each of the four proofs differs , and we describe them in reverse order .
_ proof of _ ( [ new2 ] ) _ for @xmath128 .
_ by a familiar argument , the result follows if for @xmath129 there is a deterministic function @xmath130 ( not depending on @xmath131 ) , such that @xmath132 where @xmath133 @xmath134 throughout denoting a generic arbitrarily small positive constant , and @xmath135 since @xmath58 , @xmath59 , for @xmath129 we set [ cf .
( [ d ] ) ] , @xmath136 , @xmath137 and @xmath138 .
we may write @xmath139 where @xmath140 for all @xmath141 @xmath142 , so by jensen s inequality @xmath143 under a1(i ) , we have strict inequality in ( [ ze ] ) for all @xmath144 , so that by continuity in @xmath141 of the left - hand side of ( [ ze ] ) , ( [ 1 ] ) holds .
next , write @xmath145 where @xmath146 .
because , given a2 , the @xmath147 are stationary martingale differences , @xmath148 then defining @xmath149 , and henceforth writing @xmath150 , ( [ 2 ] ) would hold on showing that @xmath151 \biggr\vert & = & o_{p } ( 1 ) , \\[-2pt ] \label{5 } \sup_{\mathcal{t}_{4}}\biggl\vert\frac{1}{n}\sum_{t=1}^{n}\sum_{j=0}^{t-1}\sum_{k = t}^{\infty}c_{j}c_{k}\gamma _ { j - k}\biggr\vert&=&o_{p } ( 1 ) , \\[-2pt ] \label{5bis } \sup_{\mathcal{t}_{4}}\biggl\vert\frac{1}{n}\sum_{t=1}^{n}\sum_{j = t}^{\infty}\sum_{k = t}^{\infty}c_{j}c_{k}\gamma _ { j - k}\biggr\vert&=&o_{p } ( 1 ) .\end{aligned}\ ] ] we first deal with ( [ 4 ] ) .
the term whose modulus is taken is @xmath152 & & \quad { } + \frac{2}{n}\sum_{j=0}^{n-2}\sum_{k = j+1}^{n-1}c_{j}c_{k}\sum_{l = k - j+1}^{n - j}\bigl\ { u_{l}u_{l- ( k - j ) } -\gamma_{j - k}\bigr\}\\[-2pt ] & & \qquad =( a ) + ( b ) .\nonumber\end{aligned}\ ] ] first,@xmath153 it can be readily shown that , uniformly in @xmath90 , @xmath154 , so @xmath155 by lemma [ lemma1 ] .
next , by summation by parts , @xmath156 is equal to @xmath157 & & \quad{}-\frac{2}{n}\sum_{j=0}^{n-2}c_{j}\sum_{k = j+1}^{n-2 } ( c_{k+1}-c_{k } ) \sum_{r = j+1}^{k}\sum_{l = r - j+1}^{n - j}\bigl\ { u_{l}u_{l- ( r - j ) } -\gamma_{j - r}\bigr\ } \\[-2pt ] & & \qquad= ( b_{1 } ) + ( b_{2 } ) .\end{aligned}\ ] ] it can be easily shown that , uniformly in @xmath90,@xmath158 so we have@xmath159 by lemma [ lemma1 ] , where @xmath160 throughout denotes a generic finite but arbitrarily large positive constant .
similarly , @xmath161 by lemma [ lemma1 ] , where @xmath162 was introduced in a1(ii )
. it can be readily shown that@xmath163 take @xmath164 such that @xmath165 .
then @xmath166 this is bounded by@xmath167 because @xmath168 . for small enough @xmath164 , ( [ new3 ] )
is bounded by @xmath169 , to complete the proof of ( [ 4 ] ) .
next , the term whose modulus is taken in ( [ 5 ] ) is @xmath170 where @xmath171 denotes the spectral density of @xmath20 . by boundedness of @xmath172
( implied by assumption a1 ) and the cauchy inequality , ( [ ab ] ) is bounded by @xmath173 so the left - hand side of ( [ 5 ] ) is bounded by @xmath174 by lemma [ lemma1 ] , to establish ( [ 5 ] ) . finally , by a similar reasoning , the term whose modulus is taken in ( 5bis ) is bounded by @xmath175 to conclude the proof of ( [ 5bis ] ) , and thence of ( [ 2 ] ) .
thus , ( [ new2 ] ) is proved for @xmath128 . with respect to ( [ new2 ] ) for @xmath176 , note from @xmath177 for such @xmath178 , and ( [ 1010 ] ) , that these results follow if @xmath179 _ proof of _ ( [ new2 ] ) _ for @xmath180_. denote , for any sequence @xmath181 , @xmath182 , @xmath183 , the discrete fourier transform and periodogram , respectively , and @xmath184 . for @xmath185 satisfying lemma [ lemma3 ] , setting @xmath186 , @xmath187 where @xmath188
. then @xmath189}}_{{\bolds{\varphi}}\in\psi}\vert\xi ( e^{i\lambda};{\bolds{\varphi } } ) \vert ^{2}\inf_{\delta\in\mathcal{i}_{3}}r_{n } ( { \bolds{\tau}}^{\ast } ) -\sup_{\mathcal{t}_{3}}\frac{1}{n}\vert v_{n } ( { \bolds{\tau } } ) \vert.\ ] ] assumption a1 implies [ see ( [ zf])]@xmath190}}_{{\bolds{\varphi}}\in\psi}\vert\xi ( e^{i\lambda};{\bolds{\varphi } } ) \vert^{2}>\epsilon.\ ] ] thus , @xmath191\\[-10pt ] & & { } -\sup_{\mathcal{t}_{3}}\frac{1}{n}\vert v_{n } ( { \bolds{\tau } } ) \vert-\sup_{\mathcal{i}_{3}}\frac{1}{n}\vert w_{n } ( \delta ) \vert,\nonumber\vspace*{-2pt}\end{aligned}\ ] ] where @xmath192 , and by lemma [ lemma2 ] @xmath193 by lemma [ lemma2 ] and ( 0.6 ) in the proof of lemma [ lemma3 ] in the supplementary material @xcite ( taking @xmath194 there in both cases)@xmath195 and also by lemma [ lemma3 ] ( with @xmath194 there ) @xmath196 next , note that for @xmath197@xmath198 where we introduce the digamma function @xmath199log@xmath200 . from ( [ 51 ] ) and the fact that @xmath201 is strictly increasing in @xmath202 , @xmath203\\[-10pt ] & & { } -\sup_{\mathcal{i}_{3}}\biggl\vert\frac{1}{n}\sum_{t=1}^{n } \mathop{\sum\sum}^{t-1}_{j\neq k}a_{j}a_{k}\varepsilon _ { t - j}\varepsilon _ { t - k}\biggr\vert.\nonumber\vspace*{-2pt}\end{aligned}\ ] ] by a very similar analysis to that of @xmath156 in ( [ ai ] ) , the second term on the right - hand side of ( [ h1 ] ) is bounded by@xmath204 & & \qquad\leq\frac{2}{n}\sup_{\mathcal{i}_{3}}\biggl\vert { \sum}_{j=0}^{n-2}a_{j}{\sum}_{k = j+1}^{n-1}{\sum}_{l = k - j+1}^{n - j}\varepsilon_{l}\varepsilon_{l-(k - j)}\biggr\vert \\[-3pt ] & & \qquad\quad{}+\frac{2}{n}\sup_{\mathcal{i}_{3}}\biggl\vert { \sum}_{j=0}^{n-2}a_{j}{\sum}_{k = j+1}^{n-2}(a_{k+1}-a_{k}){\sum}_{r = j+1}^{k}{\sum}_{l = r - j+1}^{n - j}\varepsilon_{l}\varepsilon _ { l-(k - j)}\biggr\vert,\vspace*{-2pt}\vadjust{\goodbreak}\end{aligned}\ ] ] which has expectation bounded by @xmath205\\[-8pt ] & & \qquad\leq k\biggl ( 1+\frac{1}{n^{{1/2}}}\sum_{j=1}^{n}j^{-{1/2}-a}\sum_{k=1}^{n}k^{-1+a}\biggr ) \leq k\nonumber\end{aligned}\ ] ] for any @xmath206 . therefore , there exists a large enough @xmath160 such that @xmath207 as @xmath104 .
then , noting ( [ cb ] ) , ( [ cb2 ] ) , ( [ cb1 ] ) , ( [ cc ] ) , we deduce ( [ ac ] ) for @xmath180 if@xmath208 now @xmath209 the third term on the right is clearly @xmath210 , whereas , as in the treatment of @xmath211 in ( [ ai ] ) , the second is @xmath212 , so that ( [ 32 ] ) holds as @xmath213 can be made arbitrarily large for small enough @xmath164 .
this proves ( [ ac ] ) , and thus ( [ new2 ] ) , for @xmath180 . _ proof of _ ( [ new2 ] ) _ for @xmath214_. take @xmath215 and note that @xmath216 for @xmath217 .
it follows from lemma [ lemma2 ] and ( 0.6 ) in the proof of lemma [ lemma3 ] ( see supplementary material @xcite ) that @xmath218\\[-8pt ] & = & o_{p } ( n^{2\eta-{1/2 } } ) = o_{p}(1).\nonumber\end{aligned}\ ] ] it follows from lemma [ lemma3 ] that @xmath219 denote @xmath220 . by ( [ ay ] ) , ( [ az ] ) , it followsthat ( [ ac ] ) for @xmath214 holds if for arbitrarily large @xmath160@xmath221 as @xmath104 .
clearly,@xmath222 defining @xmath223 , @xmath224 , the right - hand side of ( h2 ) is bounded below by@xmath225 for @xmath226 , @xmath227 then by ( [ h3 ] ) , using summation by parts as in the analysis of @xmath228 in ( [ ai ] ) , the expectation of the second term in ( [ h4 ] ) is bounded by@xmath229 which , noting ( [ h5 ] ) , is @xmath230 .
next , the first term in ( [ h4 ] ) is bounded below by@xmath231 using ( [ h3 ] ) it can be easily shown that the second term in ( [ h6 ] ) is@xmath232 , whereas the first term is bounded below by @xmath233 _ { 1/n}^{1 } \\ & & \qquad=\frac{\epsilon}{4\eta(2\eta+1)}-o_{p}(n^{-2\eta}).\nonumber\end{aligned}\ ] ] then ( [ ba ] ) holds because the right - hand side of ( [ h7 ] ) can be made arbitrarily large on setting @xmath164 arbitrarily close to zero .
this proves ( [ ac ] ) , and thus ( [ new2 ] ) , for @xmath214 .
_ proof of _ ( [ new2 ] ) _ for @xmath234_. noting that @xmath235 , @xmath236 because @xmath237 .
clearly @xmath238 , where @xmath239 for arbitrarily small @xmath240 , the right - hand side of ( [ q1 ] ) is bounded from below by @xmath241 for @xmath131 large enough , so it suffices to show ( [ q2 ] ) @xmath242 as @xmath104 .
first@xmath243 where @xmath244 , @xmath245 was defined below ( [ h2 ] ) , and@xmath246 for @xmath247 where ( [ ee ] ) is routinely derived , noting that by summation by parts @xmath248 now@xmath249 noting ( [ zf ] ) and that under a1 , @xmath250 , the required result follows on showing that@xmath251 as @xmath104 .
the proof of ( [ qqq2 ] ) is omitted as it is similar to and much easier than the proof of ( [ qqq1 ] ) , which we now give .
let @xmath252 . by the cauchy
inequality@xmath253 so that by ( [ eee ] ) , noting that @xmath254 , @xmath255 because @xmath256 by a1(ii ) . next , by summation by parts@xmath257 so@xmath258\\[-8pt ] & & { } + \frac{1}{n^{{1/2}}}{\sum}_{j=1}^{n-2}\sup_{\mathcal{t}_{1}}\vert s_{j+1,n } ( { \bolds{\tau } } ) -s_{j , n } ( { \bolds{\tau } } ) \vert\biggl\vert { \sum}_{k=1}^{j}u_{n - k}\biggr\vert.\nonumber\end{aligned}\ ] ] given that @xmath259 , @xmath260 so as @xmath261 , noting ( [ eee ] ) and stirling s approximation , the expectation of the first term on the right - hand side of ( [ q11 ] ) is bounded by@xmath262 next , noting that @xmath263 , it can be shown that@xmath264\\[-8pt ] & & { } + \frac{\phi_{j+1 } ( { \bolds{\varphi } } ) } { n^{\delta_{0}-\delta}}{\sum}_{l=1}^{j+1}a_{l } ( \delta_{0}-\delta ) .
\nonumber\end{aligned}\ ] ] thus , noting that , uniformly in @xmath90 , @xmath131 , @xmath265 , by previous arguments the contribution of the last term on the right - hand side of ( [ q12 ] ) to the expectation of the second term on the right - hand side of ( [ q11 ] ) is bounded by@xmath266 by identical arguments , the contribution of the first term on the right - hand side of ( [ q12 ] ) to the expectation of the last term on the right - hand side of ( [ q11 ] ) is bounded by @xmath267\\[-8pt ] & & \qquad\leq\frac{k}{n^{1+\eta}}{\sum}_{j=1}^{n}j^{{1/2}}{\sum}_{k=1}^{j-1}k^{-1-\varsigma}{\sum}_{l = j - k}^{j}l^{-{3/2}+\eta}.\nonumber\end{aligned}\ ] ] given that @xmath268 , the right - hand side of ( [ aa ] ) is bounded by@xmath269 } k^{-\varsigma } ( j - k ) ^{-{3/2}+\eta } \\ & & \qquad\quad{}+\frac{k}{n^{1+\eta}}{\sum}_{j=1}^{n}j^{{1/2}}{\sum}_{k= [ j/2 ] + 1}^{j-1}k^{-\varsigma } ( j - k ) ^{-{3/2}+\eta},\nonumber\end{aligned}\ ] ] where @xmath270 $ ] denotes integer part .
clearly , the right - hand side of ( [ as ] ) is bounded by@xmath271 so @xmath272 because @xmath256 .
next , writing @xmath273 , for @xmath274 , @xmath275 , where , by a1 , a2 , @xmath276 is well defined in the mean square sense , we have@xmath277 in view of previous arguments , it is straightforward to show that@xmath278 , to conclude the proof of ( [ qqq1 ] ) . finally , we prove ( [ qqq3 ] ) . considering @xmath279 as a process indexed by @xmath34 ,
we show first that@xmath280 where @xmath281 is a scalar brownian motion with variance @xmath282 and @xmath283 means weak convergence in the space of continuous functions on @xmath284 .
we give this space the uniform topology .
convergence of the finite - dimensional distributions follows by theorem 1 of @xcite , noting that a2 implies conditions a(i ) , a(ii ) and a(iii ) in @xcite ( in particular a2 implies that the fourth - order cumulant spectral density function of @xmath18 is bounded ) .
next , by theorem 12.3 of @xcite , if for all fixed @xmath285 @xmath279 is a tight sequence , and if for all @xmath286 and for @xmath160 not depending on @xmath287@xmath288 then the process @xmath279 is tight , and ( [ q14 ] ) would follow . first , for fixed @xmath34 , it is straightforward to show that @xmath289 , so @xmath279 is uniformly integrable and therefore tight .
next,@xmath290 by the mean value theorem , where @xmath291 is an intermediate point between @xmath292 and @xmath293 . as in lemma d.1 of @xcite,@xmath294
now ( [ q6 ] ) holds on showing that , for @xmath295,@xmath296 by stirling s approximation , the left - hand side of ( [ q8 ] ) is bounded by@xmath297 regarding ( [ q9 ] ) , it can be shown that uniformly in @xmath284 , @xmath298 ( see , e.g. , @xcite , page 259 ) .
thus , apart from a remainder term of smaller order , the left - hand side of ( [ q9 ] ) is bounded by @xmath299 uniformly in @xmath284 , the right - hand side of ( [ q10 ] ) being bounded by@xmath300 , to conclude the proof of tightness
. then by the continuous mapping theorem@xmath301 this is a.s .
positive because the quantity whose infimum is taken is a @xmath302 random variable times @xmath303 $ ] , which is bounded away from zero on @xmath284 .
thus as @xmath104 @xmath304 and ( [ qqq3 ] ) follows as @xmath134 is arbitrarily small
. then we conclude ( [ ac ] ) , and thus ( [ new2 ] ) , for @xmath234 .
this requires an additional regularity condition . 1 .
a. @xmath305 b. for all @xmath80 , @xmath79 is twice continuously differentiable in @xmath42 on a closed neighborhood @xmath306 of radius @xmath307 about @xmath308 c. the matrix@xmath309 is nonsingular , where @xmath310 . by compactness of @xmath306 and continuity of @xmath311 , @xmath312 , for all @xmath90 , with @xmath313 , where @xmath314 is the @xmath178th element of @xmath42 , a1(ii ) , a1(iv ) and a3(ii ) imply that , as @xmath315 @xmath316 which again is satisfied in the arma case . as with a1
, a3 is similar to conditions employed under stationarity , and can readily be checked in general .
[ theo2.2 ] let ( [ a ] ) , ( [ b ] ) and hold . then as @xmath317@xmath318 the proof standardly involves use of the mean value theorem , approximation of a score function by a martingale so as to apply a martingale convergence theorem , and convergence in probability of a hessian in a neighborhood of @xmath61 . from the mean value theorem , ( [ 213 ] ) follows if we prove that @xmath319 where @xmath320 .
_ proof of _ ( [ x2 ] ) .
it suffices to prove @xmath321 and @xmath322 where @xmath323 . by lemma [ lemma2 ] ,
the left - hand side of ( [ x4 ] ) is the @xmath324 vector @xmath325 , where @xmath326 clearly , @xmath327 , and @xmath328 noting that , by a2 , the @xmath18 and @xmath329 are martingale difference sequences .
thus , @xmath330 .
next , @xmath331 , and @xmath332 equals@xmath333 from ( [ b ] ) and a2 , the expectation is @xmath334 for @xmath335 , and zero otherwise . by a1 , @xmath20 has bounded spectral density .
thus , ( [ x6 ] ) is bounded by @xmath336 now @xmath337 } \frac { ( t - l ) ^{-1-\varsigma}}{l}+\sum_{l= [ t/2 ] + 1}^{t-1}\frac { ( t - l ) ^{-1-\varsigma}}{l}\\ & \leq & k ( t^{-1-\varsigma}\log t+t^{-1 } ) \leq\frac { k}{t}.\end{aligned}\ ] ] then @xmath338 , so @xmath339 next , by lemma [ lemma2 ] @xmath340 also , @xmath341 and @xmath342 since @xmath343 , denoting euclidean norm . finally , by lemmas
[ lemma2 ] and [ lemma4 ] @xmath344 to conclude the proof of ( [ x4 ] ) . next , ( [ x5 ] ) holds by the cramr
wold device and , for example , theorem 1 of @xcite on showing that @xmath345 and @xmath346\\[-8pt ] & & \qquad { } -\frac{1}{n}\sum_{t=2}^{n}e\biggl ( \varepsilon _ { t}^{2}\sum_{j=1}^{\infty}\sum_{k=1}^{\infty}\mathbf{m}_{j } ( { \bolds{\varphi}}_{0 } ) \mathbf{m}_{k}^{\prime } ( { \bolds{\varphi}}_{0 } ) \varepsilon_{t - j}\varepsilon_{t - k}\biggr ) \rightarrow_{p}0,\nonumber\end{aligned}\ ] ] because @xmath347 has expectation @xmath348 , noting that the lindeberg condition is satisfied as @xmath349 is stationary with finite variance .
now ( [ x7 ] ) follows as @xmath350 , @xmath351 , is @xmath102-measurable , whereas the left - hand side of ( [ x56 ] ) is @xmath352 because @xmath353 is stationary ergodic with mean zero .
this completes the proof of ( [ x5 ] ) , and thus ( x2 ) .
_ proof of _ ( [ x3 ] ) .
denote by @xmath354 an open neighborhood of radius @xmath355 about @xmath61 , and @xmath356 trivially,@xmath357 because @xmath358 , it follows that @xmath359 , so the first term in @xmath360 is identically zero . also , as in the proof of ( [ x5 ] ) , the second term of @xmath361 is identically @xmath348 .
thus , given that by slutzky s theorem and continuity of @xmath362 at @xmath61 , @xmath363 , ( [ x3 ] ) holds on showing @xmath364 for some @xmath240 , as @xmath104 .
as @xmath365 , the proof for ( [ x9 ] ) is almost identical to that for ( [ 4 ] ) , noting the orders in lemma [ lemma4 ] . to prove ( [ xxx ] ) , we show that@xmath366 is @xmath367 , the proof for the corresponding result concerning the difference between the second terms in ( [ x12 ] ) , ( [ h12 ] ) being almost identical . by lemma [ lemma4 ] , ( [ h13 ] ) is bounded by@xmath368\\[-8pt ] & & \qquad{}+\frac{k}{n}\sum_{t=1}^{n}\sum_{j = t}^{\infty } \sum_{k = j+1}^{\infty}j^{\epsilon-1}k^{\epsilon-1 } ( k - j ) ^{-1-\varsigma}\log^{2}k,\nonumber\end{aligned}\ ] ] noting that ( [ h14 ] ) implies that @xmath369 .
the first term in ( [ h15 ] ) is bounded by@xmath370 for any @xmath371 .
choosing @xmath372 such that @xmath373 , ( [ h16 ] ) is bounded by@xmath374 similarly , the second term in ( [ h15 ] ) can be easily shown to be @xmath375 , whereas the third term is bounded by@xmath376 for any @xmath371 , so choosing again @xmath372 such that @xmath373 , ( [ h17 ] ) is @xmath377 , to conclude the proof of ( [ x3 ] ) , and thus of the theorem .
when observations on several related time series are available joint modeling can achieve efficiency gains .
we consider a vector @xmath378 given by@xmath379 where @xmath380,@xmath381 in which @xmath382 , @xmath39 is ( as in the univariate case ) a @xmath40 vector of short - memory parameters , @xmath383 , @xmath384 for all @xmath42 , and @xmath385 , where the memory parameters @xmath386 are unknown real numbers . in general
, they can all be distinct but for the sake of parsimony we allow for the possibility that they are known to lie in a set of dimension @xmath387 .
for example , perhaps as a consequence of pre - testing , we might believe some or all the @xmath386 are equal , and imposing this restriction in the estimation could further improve efficiency .
we introduce known functions @xmath388 , @xmath389 , of @xmath390 vector @xmath391 , such that for some @xmath392 we have @xmath393 , @xmath389 .
we denote @xmath394 and define [ cf .
( [ d ] ) ] @xmath395 where @xmath396 .
gaussian likelihood considerations suggest the multivariate analogue to ( [ f ] ) @xmath397 where @xmath398 , assuming that no prior restrictions link @xmath61 with the covariance matrix of @xmath399 .
unfortunately our consistency proof for the univariate case does not straightforwardly extend to an estimate minimizing ( [ 36 ] ) if @xmath400 .
also ( [ 36 ] ) is liable to pose a more severe computational challenge than ( [ f ] ) since @xmath401 is liable to be larger in the multivariate case and @xmath402 may exceed 1 ; it may be difficult to locate an approximate minimum of ( [ 36 ] ) as a preliminary to iteration .
we avoid both these problems by taking a single newton step from an initial @xmath0-consistent estimate @xmath403 . defining @xmath404 we consider the estimate @xmath405 we collect together all the requirements for asymptotic normality of @xmath72 in : 1 . 1 . for all @xmath42
, @xmath406 is differentiable in @xmath80 with derivative in @xmath81 , @xmath82 2 . for all @xmath42 , @xmath407 3 .
the @xmath399 in ( [ zb ] ) are stationary and ergodic with finite fourth moment , @xmath408 , @xmath409 almost surely , where @xmath410 is positive definite , @xmath98 is the @xmath99-field of events generated by @xmath411 , @xmath101 , and conditional ( on @xmath102 ) third and fourth moments and cross - moments of elements of @xmath399 equal the corresponding unconditional moments ; 4 . for all @xmath80 , @xmath406 is twice continuously differentiable in @xmath42 on a closed neighborhood @xmath306 of radius @xmath307 about @xmath308 5 .
the matrix @xmath412 having @xmath413th element@xmath414 is nonsingular , where@xmath415 the @xmath416 being coefficients in the expansion @xmath417 , where @xmath418 is an @xmath419 matrix whose @xmath178th column is the @xmath178th column of @xmath420 and whose other elements are all zero ; 6 .
@xmath421 is twice continuously differentiable in @xmath391 , for @xmath422 7 .
@xmath403 is a @xmath0-consistent estimate of @xmath61 .
the components of a4 are mostly natural extensions of ones in a1 , a2 and a3 , are equally checkable , and require no additional discussion .
the important exception is ( vii ) .
when @xmath423 is a diagonal matrix [ as in the simplest case @xmath424 , when @xmath425 is a farima@xmath426 for @xmath389 ] then @xmath403 can be obtained by first carrying out @xmath428 univariate fits following the approach of section [ sec2 ] , and then if necessary reducing the dimensionality in a common - sense way : for example , if some of the @xmath386 are a priori equal then the common memory parameter might be estimated by the arithmetic mean of estimates from the relevant univariate fits .
notice that in the diagonal-@xmath429 case with no cross - equation parameter restrictions the efficiency improvement afforded by @xmath72 is due solely to cross - correlation in @xmath399 , that is , nondiagonality of @xmath410 . when @xmath423 is not diagonal , it is less clear how to use the @xmath0-consistent outcome of theorem [ theo2.2 ] to form @xmath403 .
we can infer that @xmath430 has spectral density matrix @xmath431 . from the @xmath178th diagonal element of this ( the power spectrum of @xmath432 ) , we can deduce a form for the wold representation of @xmath432 , corresponding to ( [ b ] ) .
however , starting from innovations @xmath399 in ( [ zb ] ) satisfying ( iii ) of a4 , it does not follow in general that the innovations in the wold representation of @xmath432 will satisfy a condition analogous to ( [ 28 ] ) of a2 , indeed it does not help if we simply strengthen a4 such that the @xmath399 are independent and identically distributed .
however , ( [ 28 ] ) certainly holds if @xmath399 is gaussian , which motivates our estimation approach from an efficiency perspective . notice
that if @xmath430 is a vector arma process with nondiagonal @xmath429 , in general all @xmath428 univariate ar operators are identical , and of possibly high degree ; the formation of @xmath403 is liable to be affected by a lack of parsimony , or some ambiguity .
an alternative approach could involve first estimating the @xmath433 by some semiparametric approach , using these estimates to form differenced @xmath434 and then estimating @xmath39 from these proxies for @xmath430 .
this initial estimate will be less - than-@xmath0-consistent , but its rate can be calculated given a rate for the bandwidth used in the semiparametric estimation .
one can then calculate the ( finite ) number of iterations of form ( [ 39 ] ) needed to produce an estimate satisfying ( [ 213 ] ) , following theorem 5 and the discussion on page 539 of @xcite .
let ( [ zc ] ) , ( [ zb ] ) and hold .
then as @xmath317@xmath435 because @xmath72 is explicitly defined in ( [ 39 ] ) , we start , standardly , by approximating @xmath436 by the mean value theorem . then in view of a4(vii ) , ( [ zh ] ) follows on showing @xmath437 for @xmath438 .
we only show ( [ x15 ] ) , as ( [ x16 ] ) , ( [ x17 ] ) follow from similar arguments to those given in the proof of ( [ x3 ] ) . noting that @xmath439 , whereas for @xmath440 , @xmath441 equals @xmath442 by similar arguments to those in the proof of theorem [ theo2.2 ] , it can be shown that the left - hand side of ( [ x15 ] ) equals@xmath443 then by the cramr wold device , ( [ x15 ] ) holds if for any @xmath444-dimensional vector @xmath445 ( with @xmath178th component @xmath446 ) @xmath447 where @xmath448 . as in the proof of ( [ x5 ] ) , ( [ x20 ] )
holds by theorem 1 of @xcite , for example , noting that @xmath449 to conclude the proof .
\(1 ) our univariate and multivariate structures cover a wide range of parametric models for stationary and nonstationary time series , with memory parameters allowed to lie in a set that can be arbitrarily large .
unit root series are a special case , but unlike in the bulk of the large unit root literature , we do not have to assume knowledge that memory parameters are 1 .
indeed , in monte carlo @xcite our method out - performs one which correctly assumes the unit interval in which @xmath1 lies , while in empirical examples our findings conflict with previous , unit root , ones .
\(2 ) as the nondiagonal structure of @xmath450 and @xmath451 suggests , there is efficiency loss in estimating @xmath39 if memory parameters are unknown , but on the other hand if these are misspecified , @xmath39 will in general be inconsistently estimated .
our limit distribution theory can be used to test hypotheses on the memory and other parameters , after straightforwardly forming consistent estimates of @xmath450 or @xmath451 .
\(3 ) our multivariate system ( [ zc ] ) , ( [ zb ] ) does not cover fractionally cointegrated systems because @xmath410 is required to be positive definite .
on the other hand , our theory for univariate estimation should cover estimation of individual memory parameters , so long as assumption a2 , in particular , can be reconciled with the full system specification .
moreover , again on an individual basis , it should be possible to derive analogous properties of estimates of memory parameters of cointegrating errors based on residuals that use simple estimates of cointegrating vectors , such as least squares .
\(4 ) in a more standard regression setting , for example , with deterministic regressors such as polynomial functions of time , it should be possible to extend our theory for univariate and multivariate models to residual - based estimates of memory parameters of errors .
\(5 ) adaptive estimates , which have greater efficiency at distributions of unknown , non - gaussian form , can be obtained by taking one newton step from our estimates ( as in @xcite ) .
\(6 ) our methods of proof should be extendable to cover seasonally and cyclically fractionally differenced processes .
\(7 ) nonstationary fractional series can be defined in many ways . our definition [ ( [ a ] ) and ( [ zc ] ) ] is a leading one in the literature , and has been termed `` type ii . '' another popular one ( `` type i '' )
was used by @xcite for an alternate type of estimate .
that estimate assumes invertibility and is generally less efficient than @xmath72 due to the tapering required to handle nonstationarity .
it seems likely that the asymptotic theory derived in this paper for @xmath72 can also be established in a `` type i '' setting .
the proofs of the following lemmas appear in @xcite .
[ lemma1 ] under @xmath452 with @xmath453 where for any @xmath454 , as @xmath88 , @xmath455\\[-8pt ] { \sup_{{\bolds{\varphi}}\in\psi}}\vert c_{j+1 } ( { \bolds{\tau } } ) -c_{j } ( { \bolds{\tau } } ) \vert&=&o\bigl ( j^{\max ( \delta_{0}-\delta-2,-1-\varsigma ) } \bigr).\nonumber\end{aligned}\ ] ] [ lemma2 ] under @xmath456 where @xmath457 and for any @xmath458@xmath459 and @xmath460 [ lemma3 ] under @xmath461 where for any real number @xmath458 @xmath462 [ lemma4 ] under , given an open neighborhood @xmath354 of radius @xmath355 about @xmath61 , as @xmath88 , @xmath463
we thank the associate editor and two referees for constructive comments that have improved the presentation .
we also thank sren johansen and morten o. nielsen for helpful comments .
some of the second author s work was carried out while visiting universidad carlos iii , madrid , holding a ctedra de excelencia . | we consider the estimation of parametric fractional time series models in which not only is the memory parameter unknown , but one may not know whether it lies in the stationary / invertible region or the nonstationary or noninvertible regions . in these circumstances ,
a proof of consistency ( which is a prerequisite for proving asymptotic normality ) can be difficult owing to nonuniform convergence of the objective function over a large admissible parameter space .
in particular , this is the case for the conditional sum of squares estimate , which can be expected to be asymptotically efficient under gaussianity . without the latter assumption ,
we establish consistency and asymptotic normality for this estimate in case of a quite general univariate model . for a multivariate model
, we establish asymptotic normality of a one - step estimate based on an initial @xmath0-consistent estimate . . |
infrared dark clouds ( irdcs ) are dense , cold molecular clouds seen in silhouette against 8 @xmath3 m galactic plane pah emission .
dust is relatively transparent to 8 @xmath3 m emission , so only the densest parts of molecular clouds ( a@xmath10 @xmath11 ) are opaque at this wavelength @xcite .
since massive stars and their associated clusters form in massive and dense concentrations of dust and gas , it follows that the initial conditions of massive star formation are probably found in the highest density regions of molecular clouds .
the glimpse team imaged about two - thirds of the inner galactic plane @xmath12 using the _ spitzer space telescope ( sst ) _ infrared array camera ( irac ) at 3.6 , 4.5 , 5.8 , and 8 @xmath3 m with better than @xmath13 resolution @xcite .
the glimpse images contain thousands of irdcs @xcite with structure on size scales @xmath14 to as small as the irac resolution ( see figure [ 3col19 ] ) . because they are prominent in silhouette , irdcs
are generally expected to lie at the near kinematic distance .
preliminary studies show that irdcs are cold ( @xmath15 k ) and dense ( @xmath16 @xmath17 ) , and many have masses @xmath18 m@xmath8 @xcite .
@xcite detected high deuteration in irdc clumps , evidence of very low temperatures .
however , the details of the kinematics , chemistry , and star formation within irdcs are not well understood .
high resolution observations of irdcs are needed to provide information on scales small enough to differentiate between cloud properties in quiescent and star forming regions , a distinction crucial to refining models of the earliest stages of massive star formation @xcite . here
we report results of a detailed study of an irdc at radio wavelengths using the very large array ( vla ) of the national radio astronomy observatory .
irdc g19.30 + 0.07 ( @xmath19(j2000)=18@xmath2025@xmath2156@xmath22 , @xmath23(j2000)=@xmath2412@xmath250430 , v@xmath26 km s@xmath27 ) is a filamentary irdc with a length of @xmath28 and a width of @xmath9 . using a flat galactic rotation curve with r@xmath29 kpc and v@xmath30 km s@xmath27 ,
the kinematic distance to g19.30 + 0.07 is @xmath31 kpc .
we have determined the physical and chemical properties of g19.30 + 0.07 using ammonia ( nh@xmath0 ) ( 1,1 ) , ( 2,2 ) , and dicarbon sulfide ( ccs ) ( @xmath1 ) transitions ( upper level energies ( @xmath32 ) of 23.8 k , 65.0 k , and 1.6 k respectively ) with an angular resolution roughly a factor of seven higher than previously reported for this irdc , 6compared to @xmath33 @xcite .
to relate the gas properties to star formation , continuum observations at 8.4 ghz have been carried out to search for ionized gas associated with hiiregions in the irdc .
nh@xmath34 and ccs are important molecules for investigating irdcs : both are typically considered to be high - density tracers ( @xmath35 @xmath17 ) , and the relative strengths of nh@xmath0 inversion lines can be used to estimate the temperature and column density of the gas @xcite . furthermore , [ ccs]/[nh@xmath0 ] relative abundances may provide information about chemical evolution @xcite .
chemical models have mainly been developed from observations of low mass star forming regions ; this study examines whether such models might also apply in irdcs near young star forming clumps .
the main goal of this paper is to examine the physical and chemical properties of the irdc g19.30 + 0.07 and explore the relationship between these properties and signs of star formation .
the high spatial and spectral resolution of this study permits us to distinguish between multiple velocity components in the irdc , map the temperature and density distributions in the irdc at 6 resolution , and examine the relative distribution of two gas species in the irdc .
this is one of the first irdc studies to explore the role of multiple cloud components , distinguished by small offsets in velocity , within an irdc .
furthermore , evidence for star formation and outflows in g19.30 + 0.07 enables us to examine the relationship between irdc properties and star formation .
this paper is structured as follows : in 2 , we present details of our vla observations and data reduction . in 3 , we present results of the nh@xmath0 , ccs , and continuum observations . in
4 we examine the gas kinematics and present the derived temperature , density , and mass distributions , as well as relative abundances of nh@xmath0 and ccs in the irdc . in 5 we discuss the relationship between physical and chemical properties , kinematics , and star formation within the irdc .
conclusions are given in 6 .
table [ obssum ] summarizes the parameters of the vla observations .
nh@xmath0 ( 1,1 ) and ( 2,2 ) and ccs ( @xmath1 ) emission at @xmath2 ghz were observed with the vla in d configuration .
the 3.125 mhz bandwidth contained the main and innermost satellite hyperfine lines of the nh@xmath0 ( 1,1 ) and ( 2,2 ) transitions .
two pointings were used to fully cover the highest opacity regions , as determined from the 8 @xmath3 m images .
all data were calibrated and imaged using aips .
the ccs observations were conducted during the evla upgrade , and eight of the twenty - seven antennas had been converted to evla antennas .
data were obtained using both the vla and evla antennas .
doppler tracking could not be used during the upgrade period , so the aips task cvel was applied during calibration to correct for motion of the earth relative to the local standard of rest during the observations . in the 22 ghz datasets , the uncertainty of the absolute flux densities
is estimated to be 10% .
we estimate the absolute positional accuracy to be better than 1 .
images were made of the two pointings at each frequency using natural weighting of the _ u - v _ data and clean deconvolution .
the aips task flatn was used to mosaic the two pointings and apply a primary beam correction .
the nh@xmath0 ( 1,1 ) and ( 2,2 ) mosaics were cut off at the 50% power level of the primary beam .
the synthesized beam in the full - resolution nh@xmath0 images is @xmath37 ; however , the nh@xmath0 data were later smoothed to 6 resolution to increase the signal - to - noise ratio prior to fitting the emission lines with gaussian profiles as a function of velocity ( see 4 ) .
the multiplication factors needed to convert from jy beam@xmath27 to beam - averaged brightness temperature ( k ) for the various images are listed in table [ obssum ] . because the ccs emission is weak and extended relative to that of nh@xmath0 , a 30 k@xmath38 gaussian _ u - v _ taper was applied while imaging the data , resulting in a resolution of @xmath39 .
the _ u - v _ taper was chosen to maintain the highest resolution possible while still providing a significant ( @xmath40 ) ccs detection .
additionally , a more conservative primary beam cutoff at the 65% power point was used while mosaicing the ccs so that noise peaks at the edges of the primary beam were not amplified into apparently significant detections at the low contour levels .
the 65% power point cutoff was determined by reducing the primary beam cutoff until spurious @xmath41 detections no longer appeared at the image edges .
continuum emission at 8.4 ghz was observed with the vla in a configuration .
these data were also calibrated and imaged with the aips reduction package , using the model for the flux calibrator supplied with the package .
the uncertainty of the absolute flux density scale is estimated to be 10% .
u - v _ weighting and clean deconvolution were used during imaging .
the resolution of the resulting image is @xmath42 .
we have also re - reduced and analyzed the g19.30 + 0.07 vla h@xmath43o maser data published by @xcite ( project aw659 ; the full details of these observations are described by wang et al . ) .
the data for the two pointings toward g19.30 + 0.07 were reduced in the usual manner using aips .
after self - calibration , the rms noise in a single 0.66 km s@xmath27 wide channel is 20 mjy beam@xmath27 toward the ne pointing and 15 mjy beam@xmath27 toward the sw pointing .
these noise levels are about five times better than those reported for the ensemble of sources observed by @xcite .
the synthesized beam for these maser data is approximately @xmath44 .
the nh@xmath0 ( 1,1 ) main line emission , integrated over 24.1 to 30.3 km s@xmath27 , is shown overlaid on an infrared 3-color spitzer image of g19.30 + 0.07 in fig .
[ 3col19 ] .
the three colors in fig .
[ 3col19 ] are glimpse 4.5 @xmath3 m ( blue ) and 8 @xmath3 m ( green ) emission and mipsgal @xcite 24 @xmath3 m emission ( red ) .
[ 3col19 ] shows that the nh@xmath0 ( 1,1 ) emission is spatially well correlated with the 8 @xmath3 m extinction .
nh@xmath0 ( 2,2 ) emission was also detected throughout the irdc , although due to the fainter line emission the ( 2,2 ) emission was not as well defined near the irdc boundaries .
the rms noise in the nh@xmath0 ( 1,1 ) integrated intensity image is 5 mjy beam@xmath27 km s@xmath27 . in our discussions of irdc structure ,
we use the guidelines described by @xcite : _ clouds _ have masses of @xmath45 to 10@xmath46 m@xmath8 and radii of @xmath47 to 10 pc , _ clumps _ have masses of @xmath48 to 10@xmath49 m@xmath8 and radii of @xmath50 to 1 pc , and _ cores _ have masses of @xmath47 to 10@xmath51 m@xmath8 and radii of @xmath52 to 0.1 pc .
clumps produce star clusters , and cores produce individual stars .
a 20 fwhm gaussian unsharp mask was used to identify the compact structure in g19.30 + 0.07 .
the selection criteria for nh@xmath0 clumps were peaks having structure @xmath53 with integrated intensities @xmath54 jy beam@xmath27 km s@xmath27 in the masked image .
g19.30 + 0.07 contains four apparent nh@xmath0 clumps , labeled in fig .
[ 3col19 ] as c1 , c2 , c3 , and c4 .
c1 and c2 coincide with 24 @xmath3 m emission and have previously been studied at mm wavelengths with @xmath55 resolution @xcite .
c3 is not associated with any 24 @xmath3 m emission ; however , a h@xmath43o maser is located nearby ( see 3.4 ) .
c4 is an nh@xmath0 emission peak not associated with any ir emission or masers .
the 24 @xmath3 m source s1 ( fig .
[ 3col19 ] ) is not coincident with a peak in the nh@xmath0 emission .
there is a h@xmath43o maser with velocity 19 km s@xmath27 @xcite spatially coincident with s1 .
this maser s velocity is slightly offset from the velocity of g19.30 + 0.07 , @xmath56 km s@xmath27 . while it is possible that the maser is associated with an unrelated object along the line of sight , the probability of two dense cores capable of forming a h@xmath43o maser along the line of sight to s1 is small .
it is more probable that the velocity offset between the maser and the irdc is caused by motion within the irdc related to star formation activity @xcite .
multiple velocity components and broad line - wing structures are evident in the nh@xmath0 emission , as shown in figure [ profile ] .
double - line profiles indicate locations in the cloud where there are two overlapping kinematic components .
the double - line profile features can not be attributed to self - absorption : as demonstrated in fig .
[ profile]b , the double - line profile structure is seen in the more optically thin ( 2,2 ) emission as well as the ( 1,1 ) emission .
furthermore , the double - line profiles are not correlated with nh@xmath0 high optical depth ( see 4.2 ) .
finally , the individual components of the double - line profiles can be traced continuously from regions with single - line profiles at the same velocity ; it is in overlap regions that they appear double , as illustrated by the profiles shown in fig .
[ profile ] .
there are three distinct velocity components over the entire g19.30 + 0.07 dark cloud , which we will henceforth refer to as subclouds . "
the spatial extents of these subclouds are derived from fits to the nh@xmath0 line profiles , as described in 4.1 .
the subcloud positions relative to the 8 @xmath3memission are shown in fig . [ profile ] . where both ccs and nh@xmath0 emission is detected their velocities match very well .
however , in many locations there is emission from one molecule and not the other .
figure [ ccs ] shows the distribution of the ccs and the nh@xmath0 ( 1,1 ) main line emission along with representative ccs profiles ; peaks in the ccs emission have been labeled as p1-p5 .
the ccs in fig .
[ ccs ] is integrated over four channels ( 1.2 km s@xmath27 ) centered on the velocities of the subclouds as determined from the nh@xmath0 emission .
the nh@xmath0 emission in figure [ ccs ] has been smoothed to match the resolution of the ccs emission ( @xmath39 ) and integrated over the same velocity ranges .
the rms noise in the ccs integrated intensity images is 1.3 mjy beam@xmath27 km s@xmath27 .
no 8.4 ghz continuum emission is detected from g19.30 + 0.07 .
the rms noise in the field is 0.025 mjy beam@xmath27 .
a 5@xmath57 ( 0.125 mjy ) upper limit on any free - free continuum emission implies an upper limit on the uv photon flux of @xmath58 s@xmath27 @xcite .
however , it is unlikely that all of the emitted uv photons go into ionization of h and he
. the uv flux could be reduced by absorption by an accretion disk or dust surrounding the protostar .
thus , the uv flux limit is likely an underestimate of @xmath59 .
furthermore , early - stage hii regions can be small , dense , and optically thick , and consequently have low radio continuum flux densities . further discussion of the protostellar mass limit implied by the uv flux limit is given in 5.3 . from our re - reduction and analysis of the @xcite h@xmath43o maser data we have detected two regions of maser emission .
a previously unreported maser with a peak flux density of 1.58 jy was detected at @xmath19(j2000)=18@[email protected]@xmath22 , @xmath23(j2000)=@xmath2412@xmath6005140 at an lsr velocity of 28.8 km s@xmath27 .
this maser is located within the c3 ammonia clump ( fig .
[ 3col19 ] ) , and is the first reported evidence of star formation associated with this clump .
the second h@xmath43o maser , first reported in @xcite , has a peak flux density and velocity of 0.51 jy and 19.0 km s@xmath27 , respectively , at a position of @xmath19(j2000)=18@[email protected]@xmath22 , @xmath23(j2000)=@xmath2412@xmath6004215 .
this maser is located close to the 24 @xmath3 m source s1 .
the reported peak flux densities have been corrected for primary beam attenuation .
the line profiles from the spatially smoothed @xmath61 resolution nh@xmath0 ( 1,1 ) and ( 2,2 ) images have been fitted using a non - linear least squares gaussian fitting routine , providing central velocities , linewidths , and amplitudes .
fitting errors for the nh@xmath0 central velocity and linewidths are typically @xmath62 km s@xmath27 , and @xmath63% for the amplitude , with the uncertainty reduced by approximately a factor of two in areas with a high signal - to - noise ratio .
the ( 1,1 ) line fitting was done in two steps .
first , the main ( 1,1 ) line was fitted .
second , the main line and two satellite hyperfine lines were fit simultaneously with a triple - gaussian profile , using the results from the first ( main - line only ) fit to set initial guesses for , and constraints on , the fits to the satellite lines .
the initial guesses and constraints were set as follows : ( 1 ) the initial guesses for the central velocities of the satellite lines were set at the velocity of the main line @xmath64 km s@xmath27 , constrained to @xmath65 km / s of this initial guess ; ( 2 ) the initial guess for the width of the satellite lines was set to the width of the main line , constrained to @xmath65 km s@xmath27 .
the fit to the ( 2,2 ) line was not constrained by the results of the ( 1,1 ) fit . in areas with double - line profiles ( see fig . [ profile ] ) , the fitting included additional steps .
one of the ( 1,1 ) main lines and its associated satellite lines were fitted as described above .
then the second ( 1,1 ) main and satellite lines were fitted as described above .
finally , the results of these fits were used to set the initial guesses and constraints for a simultaneous ( six - gaussian ) fit that separated the various emission contributions .
the double - profile fits of the satellite lines used the same main - line based constraints as described for single - profile fits .
the fits to ( 2,2 ) double - line profiles were done simultaneously using a two - gaussian profile ; as in the single - profile case , the fits were not constrained by the ( 1,1 ) line fits .
c1 and c2 exhibit an emission component with broad line - wings , possibly indicating outflows ; in these areas the line - wing component was modeled as an additional gaussian during simultaneous line fitting .
three subclouds were identified based on fits of the nh@xmath0 ( 1,1 ) profiles .
their median lsr velocities are centered at 25.7 km s@xmath27 , 26.7 km s@xmath27 , and 28.4 km s@xmath27 .
the spatial distributions of the velocities and linewidths for each of the subclouds are shown in figure [ velcombo ] .
internal motion is indicated within each subcloud by central velocity changes of @xmath66 km s@xmath27 , shown in figs .
[ velcombo]a , [ velcombo]b , and [ velcombo]c .
typical fwhm linewidths are @xmath67 km s@xmath27 , but the linewidths are @xmath68 km s@xmath27 at the subcloud boundaries and increase to @xmath69 km s@xmath27 near c1 and c4 , shown in figs .
[ velcombo]d , [ velcombo]e , and [ velcombo]f . since c4 contains no evidence of star formation , the increased velocity dispersion at c4 may be due to a nearby outflow originating from c2 .
as discussed in 3.2 , ccs emission is found in all of the subclouds .
the ccs emission lines could not be fitted reliably because of their low signal - to - noise ratio , but based on the profiles shown in fig .
[ ccs ] we estimate the ccs fwhm to be @xmath68 km s@xmath27 , and therefore assume the ccs emission from each subcloud is contained within @xmath67 km s@xmath27centered at that subcloud s velocity . where both ccs and nh@xmath0 are detected , the linewidths and emission line structure from both molecules are in good agreement , even in areas with multiple velocity components .
thus , ccs velocity components were separated by averaging over 1.2 km s@xmath27centered at the subcloud velocities determined from nh@xmath0 .
nh@xmath0 ( 2,2 ) line profiles were used to examine the kinematics of the line - wing features , as the greater separation between the main line and satellite hyperfine components for the ( 2,2 ) transition simplified simultaneous fitting of narrow and broad components .
the line - wings toward c1 and c2 had typical fwhm linewidths of @xmath70 km s@xmath27 .
a residual image cube was made of the nh@xmath0 emission with the fitted narrow component subtracted , so that only the broad component remained .
the integrated blue- and red - shifted emission was obtained from the residual data by integrating over the velocity ranges 14.6 to 26.6 km s@xmath27 and 26.6 to 38.7 km s@xmath27 .
representative profiles and the red- and blue - shifted integrated nh@xmath0 ( 2,2 ) emission are shown in figure [ outflow ] .
the red and blue line - wing emission from c1 is spatially coincident with , and centered on , a 24 @xmath3 m source , and can be interpreted as an outflow oriented close to the line of sight .
the infall velocity derived from the mass and radius of c1 ( see 4.4 ) is @xmath28 km s@xmath27 , insufficient to account for the broadening observed in the line - wing emission .
the age of c1 s outflow can not be determined from the flow velocities because of the line - of - sight orientation .
the line - wing emission at c2 can be interpreted as an outflow originating at the 24 @xmath3 m source , with the red shifted outflow extending @xmath71 pc ne and the blue shifted outflow extending @xmath72 pc sw of the source .
the outflow from c2 appears well collimated , with a collimation factor ( length / width ) @xmath73 .
the outflow s inclination angle is unknown , so we assume an angle of 45@xmath25 in estimating the measured length of the outflow .
the central velocity derived from a gaussian fit to the broad component at each pixel within the c2 outflow gives an average outflow velocity of @xmath74 km s@xmath27 , corrected for an assumed inclination angle of 45@xmath25 .
the outflow s kinematic age was determined by dividing the extent of the outflow by the average outflow velocity , resulting in an age of @xmath75 yr .
nh@xmath0 inversion transitions have hyperfine splitting due to magnetic dipole interactions and the interaction of the nitrogen electric quadrupole moment and the electron electric field @xcite .
the optical depth of the nh@xmath0 ( 1,1 ) main line , @xmath76 , can be derived from the ratio of the main and satellite hyperfine line strengths . assuming the excitation temperature is the same for the main and satellite lines , the ratio of the brightness temperatures @xmath77 and @xmath78 can be expressed as @xmath79}{1-\exp[-\alpha\tau_{(1,1,m)}]}\ ] ] where ` @xmath80 ' and ` @xmath81 ' indicate the main and satellite hyperfine components , @xmath82 is the measured brightness temperature
, @xmath76 is the optical depth of the main quadrupole hyperfine component of the ( 1,1 ) line , and @xmath19 is the theoretical intensity ratio of the satellite line to the main line if the hyperfine levels are populated according to their statistical weights .
for the innermost satellite hyperfine transition , @xmath83 @xcite . the total optical depth of the ( 1,1 ) line , @xmath84 , can be determined from @xmath76 and the relative intensities of the hyperfine components , such that @xmath85 , where @xmath86 is the ratio of the total ( 1,1 ) line optical depth to the optical depth of the main hyperfine component . if the nh@xmath0 ( 1,1 ) hyperfine levels are populated according to their statistical weights , @xmath87 @xcite .
the optical depth @xmath84 ranges from @xmath47 to 4 at the boundaries of the subclouds and increases towards the cloud centers . near c1 , c2 , and
c3 @xmath84 is @xmath74 to 10 .
the uncertainty in @xmath84 is @xmath88% where the nh@xmath0 signal to noise is highest . at the subcloud boundaries , the uncertainty in @xmath84 increases , and
is on the order of a factor of two . at positions where @xmath89
, @xmath78 approaches @xmath77 too closely to yield accurate results ; at these positions a @xmath84 value of 20 was assumed .
the optical depth distribution is shown in figure [ tau ] .
it is notable that areas where nh@xmath0 double - line profiles are detected , shown as subcloud overlap in fig .
[ profile ] , are not coincident with the highest optical depth gas , ruling out self - absorption as a cause of the double peaked line profile .
the ratio of the brightness temperatures of the nh@xmath0 ( 1,1,@xmath80 ) and ( 2,2,@xmath80 ) lines can be used to determine the rotation temperature @xmath90 that characterizes the population distribution among the ( 1,1 ) and ( 2,2 ) states @xcite : @xmath91 \bigr ) \bigr ] } \biggr ] } \right)^{-1}.\ ] ] the distribution of @xmath92 for locations where the statistical uncertainty in @xmath92 is 10% or less is shown in figure [ temperature ] .
the @xmath92 gradient is steep near the nh@xmath0 clumps , and consequently @xmath92 , in particular , suffers from the effects of beam dilution caused by smoothing the data .
thus , we created temperature maps at the original @xmath37 resolution of c1 and c2 , where the nh@xmath0 emission is strong enough to obtain good line fits .
the higher resolution images are shown outlined in black in fig .
[ temperature](a ) .
values of @xmath92 range from 10 k up to @xmath9 k , above which @xmath93 approaches @xmath77 too closely to yield reliable temperature estimates .
a value of 35 k was assumed at points where @xmath94 to indicate points with uncertain temperatures .
the higher resolution images of the clumps show that @xmath92 reaches this maximum at c1 and s1 , and @xmath95 k in c2 , indicating heat sources at these locations .
heating is consistent with the presence of 24 @xmath3 m emission in c1 , c2 , and s1 .
an area with @xmath96 k approximately 10 east of c1 has no corresponding peak in nh@xmath0 , and the nature of this peak in @xmath92 remains unknown .
using earlier results from @xcite , @xcite empirically find a relationship between irdc gas kinetic temperature , @xmath97 , and @xmath92 , @xmath98}.\ ] ] this relationship is good to better than 5% when @xmath92 is 5 to 20 k. above 20 k , the relationship is not as accurate .
since the majority of the gas in g19.30 + 0.07 has @xmath99 k , we apply this relationship and find a mean @xmath97 in the irdc of @xmath100 k compared to a mean @xmath92 of @xmath101 k. the column density of the nh@xmath0 ( 1,1 ) transition was calculated from the total optical depth and brightness temperature of the ( 1,1 ) transition , @xmath84 and @xmath102 , which include all hyperfine components . in the optically thin limit , @xmath102=@xmath103 , where @xmath87 and is defined as above .
the column density of the nh@xmath0 ( 1,1 ) transition was determined using @xmath104\right ) } \frac{8 k \pi^{1.5 } \nu_{(1,1)}^2}{h c^3 } \frac{1}{a_{(1,1)}},\ ] ] where @xmath105 is the fwhm linewidth ( in velocity units ) , @xmath106 is the frequency of the ( 1,1 ) transition , and @xmath107 is the einstein coefficient for the sum of all hyperfine transitions . equation 4 contains an optical depth correction @xcite that corrects for the opacity of the ( 1,1 ) emission . the total column density is obtained using the nh@xmath0 partition function , @xmath108 $ ] : @xmath109}{g_{1,1 } } \exp\biggl[\frac{e_{(1,1)}}{k t_{r}}\biggr].\ ] ] equation 5 assumes the same @xmath92 among all nh@xmath0 metastable levels , with negligible contribution from non - metastable levels @xcite .
@xmath110 is the energy above the ground state and @xmath111 is the total statistical weight of the ( 1,1 ) inversion state .
values for the partition function were derived from jpl line catalog values calculated at 9.375 k , 18.75 k , and 37.50 k @xcite using a linear interpolation .
equations 4 and 5 assume that the emission fills the synthesized beam , and neglect the microwave background . in areas where the excitation temperatures approach the cosmic microwave background temperature ( 2.7 k ) , equation 5 may underestimate the column density by roughly a factor of two @xcite .
corrections for the opacity of the ( 1,1 ) emission were included in the calculation of the nh@xmath0 column densities ; however , the column densities will be lower limits in areas of very high optical depth ( @xmath112 ) , since the optical depth can not be measured accurately in these areas .
fortunately , such high - opacity regions cover a very small area of the irdc , and are not associated with the clumps .
the nh@xmath0 column density distribution is shown in figure [ density ] .
the column densities lie in the range @xmath113 to @xmath114 @xmath6 . in all three subclouds
, there appears to be a lower density envelope ( @xmath5 @xmath6 ) around higher density clumps ( @xmath115 to @xmath116 @xmath6 ) .
the uncertainty in column density is @xmath88% within the clumps , but at the cloud boundaries the values are only good to approximately a factor of two .
c1 , c2 , and c3 are coincident with column density peaks .
@xcite found the [ nh@xmath0]/[h@xmath43 ] abundance ratio in irdcs ranged from @xmath118 to @xmath119 .
we assume a [ nh@xmath0]/[h@xmath43 ] abundance ratio of @xmath120 @xcite .
the h@xmath43 column densities imply peak @xmath121 @xcite , in good agreement with the observed 8 @xmath3 m absorption .
to convert from column density to volume density , the line of sight depth must be determined .
we assume a cylindrical geometry for the filamentary irdc , with the semi - major axis ( length " ) directed from the northeast corner to the southwest corner of the irdc .
we approximate the semi - minor axis ( `` width '' ) of the irdc , aligned perpendicular to the semi - major axis , to be @xmath122 pc .
the assumption of cylindrical geometry was then used to calculate the depth at each point in the irdc .
the calculated h@xmath43 number densities then range from @xmath123 to @xmath124 @xmath17 , values consistent with previous number densities reported for irdcs @xcite .
the ccs ( @xmath1 ) column density , @xmath125 , is determined using @xmath126 where the symbols are defined as in equations 4 and 5 .
equation 6 assumes the ccs ( @xmath1 ) emission is optically thin and fills the synthesized beam .
we assume @xmath127 to be the integrated emission of the ccs over 1.2 km s@xmath27 . using the partition function
@xmath128 $ ] , where @xmath129 is the temperature describing the relative level populations , the total column density is obtained : @xmath130}{g_{1 } } \exp\biggl[\frac{e}{k t_{ex}}\biggr].\ ] ] where @xmath131 is the energy of the @xmath132 level above the ground state .
values for the partition function were interpolated from jpl line catalog values in the same manner as for nh@xmath0 @xcite .
we assume a ccs @xmath129 of 10 k based on typical values of nh@xmath0 @xmath92 in areas where ccs was detected .
typical ccs column densities ranged from @xmath133 to @xmath134 @xmath6 , roughly two orders of magnitude lower than the nh@xmath0 column densities .
unfortunately , the unknown ccs @xmath129 introduces significant uncertainty in the column density measurements ; using an excitation temperature of 5 k reduces the column densities to @xmath135 to @xmath133 @xmath6 , while 15 k produces column densities ranging from @xmath136 to @xmath137 @xmath6 . due to the large uncertainties in the ccs excitation temperature and linewidth , these ccs column densities are likely accurate to only an order of magnitude . the ccs @xmath129 can be constrained using ccs ( 4@xmath03@xmath43 ) observations of @xcite ( hpbw @xmath138 ) .
@xcite used a non - detection of ccs ( 4@xmath03@xmath43 ) emission to report an upper limit on the total ccs column density of @xmath139 @xmath6 at c1 and @xmath140 @xmath6 at c2 .
we used these results to derive an upper limit for the ccs ( 4@xmath03@xmath43 ) column density , @xmath141 , at c1 and c2 . to compare our observations with those of @xcite we smoothed our ccs @xmath125 column density image to match the nobeyama radio observatory resolution , fwhm 37 , and measured @xmath125 at the @xcite pointings . by comparing our measured @xmath125 values and the @xmath141 upper limits and assuming local thermodynamic equilibrium , we derived a ccs @xmath129 upper limit of @xmath142 k. irdc and clump masses were derived from nh@xmath0 column densities .
the mass of the irdc was determined by integrating the nh@xmath0 column density ( shown in fig . [ density ] ) over the entire source area ( clumps as well as diffuse gas ) , giving the total number of nh@xmath0 molecules in the irdc .
the number of molecules was converted to total mass by assuming the mass is dominated by h@xmath43 and using the assumed [ nh@xmath0]/[h@xmath43 ] abundance ratio of @xmath120 @xcite and an h@xmath43 mass of @xmath143 g. the irdc mass is @xmath144 m@xmath8 , consistent with the @xmath145 m@xmath8 reported by @xcite .
@xcite determined a total irdc mass of @xmath146 m@xmath8 using 1.2 mm dust emission .
the apparent discrepancy between the masses determined from nh@xmath0 observations and dust observations may be attributed to the large systematic uncertainties in the dust emissivity , dust to gas ratio , and [ nh@xmath0]/[h@xmath43 ] ratio .
the individual masses of c1 , c2 , c3 , and c4 were calculated by integrating nh@xmath0 column density over only the clumps spatial extents , and assuming an [ nh@xmath0]/[h@xmath43 ] abundance ratio of @xmath120 .
the clump radii were determined by fitting the nh@xmath0 emission distribution in c1 , c2 , c3 and c4 with a 2-d gaussian using aips task jmfit and using the fwhms as the radii of the clumps .
clumps c1 , c2 , c3 , and c4 have masses 140 , 160 , 30 , and 30 m@xmath8 , respectively .
the mass uncertainties given in table [ clumps ] are based on uncertainty in the column densities and clump radii ; it should be noted that uncertainties in the distance to the irdc and the [ nh@xmath0]/[h@xmath43 ] abundance ratio introduce additional systematic uncertainty . while the mass uncertainty may be quite large , @xcite determined masses of 113 m@xmath8 for c1 and 114 m@xmath8 for c2 using 1.2 mm dust emission .
@xcite assumed a dust temperature of 15 k , comparable to the temperatures we have measured in this irdc ( mean @xmath147 k , mean @xmath148 k ) .
the clump radii determined by @xcite were 0.15 pc for c1 and 0.21 pc for c2 , comparable to the @xmath50 pc determined for c1 and c2 in this study .
the @xcite masses are in good agreement with those found in this study , despite the fact that the overall irdc mass determined by @xcite is a factor of three lower than that determined in this study .
the fact that the masses agree on the scale of the clumps but not on the scale of the cloud may be due to sensitivity and/or resolution differences between the two studies .
we examined the stability of c1 , c2 , c3 , and c4 by comparing their gas mass to their virial masses .
the virial mass ( @xmath149 ) is the mass at which thermal plus turbulent gas pressure balances gravitational forces in the clumps , i.e. , the gravitational and kinetic energy are in equipartition . the ratio of virial mass and clump mass ( @xmath150 ) is a basic indicator of whether the clumps are gravitationally bound and likely to collapse : if @xmath151 , the clump contains enough mass to collapse under gravity .
this balance analysis assumes negligible contributions to clump stability from additional forces such as magnetic fields and clump rotation , neither of which are known for the clumps in g19.30 + 0.07 .
additionally , clump confinement by external pressure @xcite has not been included .
we calculated @xmath149 from the clump radius ( @xmath152 , determined in 4.4 ) and the nh@xmath0 fwhm linewidth ( @xmath105 ) using @xmath153 the nh@xmath0 ( 1,1 ) fwhm linewidth was measured at the density peak of the clumps .
values of @xmath105 , @xmath149 , and @xmath150 are listed in table [ clumps ] .
the clumps all have @xmath154 , inconsistent with indications of star formation that imply the clumps are undergoing gravitational collapse
. however , equation 8 assumes virial equilibrium . if the clumps have already collapsed and harbor outflows the conditions of virial equilibrium are violated , and the linewidths will not accurately predict the thermal and turbulent motion supporting the clumps . indeed , clumps c1 , c2 , and c3 show evidence of outflow through either broad nh@xmath0 linewidths or h@xmath43o masers .
clump c4 has @xmath155 ; this high ratio may be a result of increased linewidths toward c4 caused by the outflow originating from c2 ( see 4.1 ) .
it is also worth noting that where the cores are optically thick the @xmath105 may be overestimated , which would lead to a greater overestimate of virial mass ( which is proportional to @xmath156 ) than of clump mass ( which is proportional to @xmath105 ) .
the virial masses of c1 and c2 reported by @xcite are 893 and 823 @xmath157 , significantly larger than our @xmath149 values .
@xcite used larger clump radii and higher linewidths in their virial mass calculations because of the lower resolution of their study ( 40 ) .
our higher resolution shows most of the mass is contained within a smaller radius than previously thought .
furthermore , we were able to resolve velocity gradients in the emission near the clumps ( fig .
[ velcombo ] ) .
the larger beam used by @xcite would have confused this emission , contributing to their larger linewidth measurements .
chemical evolution models comparing ccs and nh@xmath0 distributions have mainly been developed from observations of low mass star forming regions and the environs of hii regions ; we now examine whether such models might also apply in irdcs .
chemical models such as those of @xcite , @xcite , and @xcite predict that ccs traces chemically young gas in lower density , quiescent envelopes of molecular clouds , while nh@xmath0 traces dense , evolved clumps .
time - dependent gas - phase modeling suggests that for low - mass star - forming clumps the [ ccs]/[nh@xmath0 ] abundance ratio is a measure of the age of clumps , i.e. , a `` chemical clock '' : ccs abundance peaks in chemically young gas and decreases as the ccs components are reprocessed into molecules with stronger binding energies , such as co. @xcite and @xcite propose ccs is reprocessed on timescales of @xmath158 years , while de gregorio - monsalvo et al .
( 2005 , 2006 ) concluded that ccs has a lifetime of @xmath159 years after the onset of star formation .
freeze - out onto dust grains likely plays a significant role in chemical changes during clump evolution .
dust grain chemistry is particularly important for explaining the high abundances of nh@xmath0 observed in dense clumps @xcite .
when co freezes out onto dust grains the abundance of nh@xmath0 is enhanced , because gas - phase co destroys nh@xmath0 parent molecules @xcite .
furthermore , the abundance of hydrogenated molecules , including nh@xmath0 , may be enhanced as ices evaporate off dust grains @xcite near hot clumps in star - forming regions .
freeze - out of ccs onto dust grains may also be responsible for low ccs abundances in gas clumps where @xmath160 @xmath17 @xcite .
dust grain chemistry also explains how the chemical clock may be reset .
@xcite propose collisions in gas with properties similar to irdc subclouds can trigger mhd waves and/or grain impacts that provide sufficient energy for thermal desorption of molecules from grain surfaces , leading to the presence of early - time molecules like ccs .
the critical densities of the nh@xmath0 ( 1,1 ) and ccs ( @xmath1 ) emission lines differ by almost two orders of magnitude , @xmath161 @xmath17for nh@xmath0 @xcite and @xmath75 @xmath17 for ccs @xcite , which can complicate the interpretation of molecular distribution . however ,
despite its higher critical density , we detect ccs emission in the lower density envelopes around the high - density nh@xmath0 clumps rather than toward the center of the dense clumps , suggesting their spatial distributions are the result of chemical evolution in the irdc rather than a density effect .
previous studies of the chemistry in g19.30 + 0.07 have been limited by poor resolution . prior to this investigation
, @xcite observed g19.30 + 0.07 in nh@xmath0 emission using the effelsberg 100 m telescope with 40 resolution , but lacked the resolution needed to comment on small - scale structure . @xcite
observed n@xmath43h@xmath162(10 ) , hc@xmath0n(54 ) , ccs(4@xmath03@xmath43 ) , ch@xmath0oh(76 ) , and nh@xmath0 ( 1,1 ) , ( 2,2 ) , and ( 3,3 ) emission towards the clumps in g19.30 + 0.07 with the nro 45-m telescope ( hpbw @xmath138 for ccs and @xmath163for nh@xmath0 ) .
@xcite report abundances suggesting the clumps in g19.30 + 0.07 are more chemically evolved than low - mass starless cores .
the high spatial resolution data we present here improve upon these findings by showing how morphology , temperature , and density structure in the irdc impact the distributions of the nh@xmath0 and ccs .
[ ccs ] shows the integrated ccs emission in each of the subclouds and compares it to the nh@xmath0 integrated emission . although ccs emission is associated with the subclouds at 25.7 km s@xmath27 , 26.7 km s@xmath27 , and 28.4 km s@xmath27 , the spatial distribution of the ccs and nh@xmath0 emission are not coincident .
enhancements in the ccs with clump - like morphology are labeled as p1 to p5 in fig .
[ ccs ] ; no evidence of nh@xmath0 clumps exist at these locations .
typical ccs column densities ranged from @xmath133 @xmath6 to @xmath164 @xmath6 .
we examined the [ ccs]/[nh@xmath0 ] column density ratio in areas with detectable emission from at least one of the molecules . in ccs peaks lacking detectable nh@xmath0 emission we assign an nh@xmath0 column density upper limit of @xmath113 @xmath6 . where nh@xmath0 but not ccs was detected we assign a ccs column density upper limit of @xmath133 @xmath6 . at clumps
c1 , c2 , c3 , and c4 the [ ccs]/[nh@xmath0 ] ratio is @xmath165 , slightly higher than the upper limit determined by @xcite , and consistent with active , star forming clumps @xcite . at ccs emission peaks , [ ccs]/[nh@xmath0 ] ratios are @xmath166 , consistent with results from @xcite , who found [ ccs]/[nh@xmath0 ] @xmath167 in low - mass starless clumps . a map of the column density ratios is shown in figure [ ratio ] . comparing fig .
[ ratio ] with fig .
[ ccs ] shows that p1 is associated with the [ ccs]/[nh@xmath0 ] peak and located in the subcloud located at 25.7 km s@xmath27 , the only subcloud lacking evidence of star formation .
the ccs morphology and [ ccs]/[nh@xmath0 ] ratio in this subcloud are consistent with a chemically young clump in a stage too early to present evidence of star formation @xcite .
the ccs peaks p2 , p4 , and p5 are located at subcloud interfaces , and p3 is spatially coincident with the outflow from c2 .
the spatial coincidence of these ccs peaks with subcloud - subcloud and outflow - subcloud interfaces indicates collisions may be resetting the chemical clock of the gas at these locations , consistent with the @xcite model .
one major caveat must be considered in the analysis of ccs emission .
ccs ( @xmath1 ) is a low - lying energy state transition ( upper energy level only 1.6 k above ground ) , so it is difficult to know whether enhanced ccs emission is due to excitation effects or increased abundance . using the results of @xcite we were able to roughly constrain the ccs excitation temperature to @xmath142 k.
however , the ccs ( @xmath168 ) observations were low resolution and only provided a column density upper limit . additional transitions of ccs must be measured at comparable resolution to this study to more accurately determine the ccs excitation temperature .
sensitive interferometric observations at millimeter wavelengths could explore whether the abundance measurements in the ground state are consistent with abundances measured from higher energy transitions .
we draw the following conclusions from the ccs detections : ( 1 ) ccs and nh@xmath0 are present in each of the three subclouds comprising g19.30 + 0.07 , but each molecule traces different parts of the subclouds ; ( 2 ) ccs appears to trace the boundaries of the irdc while nh@xmath0 peaks in the inner parts of the irdc , consistent with ccs enhancement in low - density envelopes and nh@xmath0 enhancement in denser clumps , in spite of the fact that the ccs transition has the higher critical density .
this emphasizes that nh@xmath0 does not give the full picture of the dense gas in star - forming regions ; ( 3 ) in some cases , ccs appears to have clump - like morphology .
the [ ccs]/[nh@xmath0 ] peak at p1 is located in the only subcloud lacking evidence of star formation .
this supports the hypothesis that ccs cores may represent the very earliest stages of core formation ; ( 4 ) ccs peaks at subcloud - subcloud or outflow - subcloud interfaces may indicate collisions are resetting the chemical clock in some parts of the irdc ; and ( 5 ) further investigations at higher - energy transitions are needed to distinguish excitation and abundance effects in ccs emission .
although early studies of g19.30 + 0.07 did not find any evidence of star formation @xcite , recent investigations , including this one , indicate substantial active star formation in this irdc .
the observed evidence for star formation is : ( 1 ) archival mipsgal 24 @xmath3mobservations showing areas of hot dust associated with c1 , c2 , and s1 ; ( 2 ) structure in the nh@xmath0 spectral line observations consistent with outflows originating from c1 and c2 ; ( 3 ) h@xmath43o masers coincident with c3 and s1 ( fig .
[ 3col19 ] ) .
no 8.4 ghz continuum emission was detected in the irdc .
a 5@xmath57 detection limit of 0.125 mjy beam@xmath27 sets the upper limit of the uv photon flux at @xmath169 s@xmath27 , implying a stellar type of b3 or later @xcite .
however , if the protostar is accreting material , much of the stellar uv radiation will be absorbed close to the star @xcite . also , in an irdc it is likely that some uv photons are absorbed by dust .
thus , stars with types hotter than b3 could be present in g19.30 + 0.07 but be undetected by our measurements .
in fact , orion source i , an early - b star with mass @xmath170 m@xmath8 @xcite , has an 8.4 ghz flux density of @xmath171 mjy @xcite at a distance of @xmath146 pc @xcite .
if orion source i were located at the distance of g19.30 + 0.07 , 2.4 pc , its flux density would be only @xmath172 mjy , below the detection limit of this study .
clump masses of @xmath173 m@xmath8 and the presence of outflows suggest that c1 and c2 may be protoclusters . using the initial mass function from @xcite and assuming a 50% star forming efficiency
, these clumps would form stars up to masses of @xmath174 m@xmath8 .
the smaller nh@xmath0 clump and h@xmath43o maser without accompanying 24 @xmath3memission at c3 suggest c3 likely contains protostars that are either lower - mass or younger than those in c1 and c2 .
the 24 @xmath3 m emission toward s1 suggests the presence of at least one protostar , consistent with the h@xmath43o maser coincident with s1 .
it may be s1 is more evolved than c1 and c2 , and has destroyed or dispersed any dense gas associated with its formation .
however , no shell of dispersed nh@xmath0 is detected around s1 .
assuming the linewing structure discussed in 4.2 indicates outflows , the outflow masses provide insight into the masses of the stars contained in c1 and c2 .
the column densities in the outflows were determined as in equations 6 and 7 , using the integrated emission from the nh@xmath0 ( 2,2 ) line - wings and assuming a @xmath129 of 10 k for nh@xmath0 .
the column densities were summed over the outflow area to determine the masses of the outflows .
the red and blue outflow components of c1 each contain @xmath175 m@xmath8 and the components in c2 each contain @xmath176 m@xmath8 , suggesting intermediate or high mass protostar progenitors .
however , we note that nh@xmath0 only traces high density gas , so observations of the low - density gas are needed to better constrain the total outflow masses .
various models attempt to explain the formation of massive stars , but recent work suggests accretion in turbulent clumps is the likely creation mechanism @xcite .
@xcite present an evolutionary sequence of massive stars starting from high mass starless clumps : low- and intermediate - mass protostars form within the massive starless clumps and accrete matter until becoming a massive protostellar object , which develops into a massive star and clears the natal gas from its environment .
c1 and c2 may be examples of clumps containing the embedded accreting protostars proposed in this model . based on the results from this single irdc ,
the role that subclouds play in massive protostar and protocluster development is still unclear .
they may indicate an early stage of fragmentation , cloud formation / collision , or a superposition along the line of sight .
the subclouds appear distinct , but if less dense gas below our detection limits connects the subclouds , they may all be part of a larger structure .
further work is needed to understand the role of subclouds in irdcs .
understanding the role of irdcs in star formation requires a survey of their properties .
we have examined one irdc in great detail , but is g19.30 + 0.07 a typical irdc , and how general are our findings ?
our results are consistent with previous studies that indicate g19.30 + 0.07 has temperatures and densities typical for an irdc @xcite .
the high resolution of this study compared to previous work has revealed new irdc features .
notably , the kinematics have been analyzed in detail .
we find that the irdc is composed of three subclouds , each of which has a slightly different velocity .
the universality of multiple subclouds in irdcs and the role they play in protocluster evolution requires further investigation .
massive star formation continues to be poorly understood , largely because early stages are difficult to observe due to the short timescales involved and the embedded nature of massive protostars . because of their large masses , low temperatures , and high densities , irdcs are currently the most likely candidates to host the first stages of massive star formation .
understanding their properties may shed light on the processes that lead to the creation of massive protostars . although irdcs are identified using ir observations , their high opacity makes radio wavelengths an important probe of their structure .
we have examined irdc g19.30 + 0.07 using vla spectral line observations of the nh@xmath0 ( 1,1 ) , ( 2,2 ) , and ccs ( @xmath1 ) transitions . from this work , we draw the following conclusions : 1 .
g19.30 + 0.07 has structure down to the resolution limit of our observations ( 4 to 6 ) . at a distance of 2.4 kpc ,
the structure is @xmath177 au in size .
this level of detail is seen in the ir images as well as the nh@xmath0 and ccs observations .
2 . g19.30 + 0.07 is composed of three subclouds at different velocities .
star formation and [ ccs]/[nh@xmath0 ] levels may indicate that one of the subclouds is at an earlier evolutionary stage than the other two .
further investigation is needed to determine if irdcs are typically composed of multiple subclouds , and the role subclouds play in irdc evolution .
typical values of @xmath92 in g19.30 + 0.07 range from 10 to 30 k with a mean value of @xmath178 k. increases in @xmath92 at c1 , c2 , and s1 indicate heat sources at these locations .
4 . nh@xmath0 column densities in g19.30
+ 0.07 are @xmath5 to @xmath179 @xmath6 , with lower density envelopes around higher density clumps .
the nh@xmath0-derived masses of c1 , c2 , c3 , and c4 are @xmath180 , @xmath181 , @xmath182 , and @xmath182 m@xmath8 .
the clumps have masses significantly lower than their virial masses ; however , signs of star formation indicate c1 , c2 , and c3 are gravitationally bound and collapsing . in clumps that are collapsing or contain outflows ,
@xmath183 is not a reliable metric for testing gravitational boundedness .
ccs and nh@xmath0 are present in each of the three subclouds comprising g19.30 + 0.07 , and appear to trace different parts of the subclouds .
the nh@xmath0 emission closely traces the ir extinction seen in the glimpse 8 @xmath3 m images , and peaks where clumps are forming and evolving in the irdc .
conversely , the ccs tends to trace the boundaries of the irdc , where nh@xmath0 column densities indicate the irdc is more diffuse .
given the high critical density of ccs ( @xmath75 ) @xcite , the ccs detections at the irdc boundaries demonstrate that nh@xmath0 does not give the full picture of the dense gas in star - forming regions .
7 . in some places
, the ccs appears to have clump - like morphology .
the ccs peaks may show the very earliest stages of clump formation and/or result from collisions that reset the chemical clock of the gas .
stars are forming in g19.30 + 0.07 , most obviously in clumps c1 and c2 .
younger or lower mass clumps may look like c3 , where a h@xmath43o maser but no ir emission is present , or c4 , where no signs of star formation have yet been detected .
s1 , where a h@xmath43o maser coincides with a 24 @xmath3 m source but no nh@xmath0 clump is detected , may be a more evolved protostar , or associated with nh@xmath0 at a velocity outside our observed bandwidth .
we did not detect 8.4 ghz emission from the irdc , although the 8.4 ghz emission from the early - b star orion source i would be below our detection limits if it were at a distance of 2.4 kpc .
kd thanks the nrao pre - doctoral research program for support and funding . ec and kd
were also supported during this work by the national science foundation under grants ast-0303689 and ast-0808119 . during this investigation ri
was supported by a spitzer fellowship to the university of virginia and by jpl / spitzer grants rsa 1276990 and rsa 127546 .
the authors thank the anonymous referee for his or her insightful comments and careful reading of the text .
lllll observation date & 2005 dec 10 , 11 & 2005 dec 10 , 11 & 2007 feb 19 & 2006 jan 30 + nrao proposal i d & ad516 & ad516 & ad556 & ad524 + rest frequency ( ghz ) & 23.6944955 & 23.7226333 & 22.3440330 & 8.4351 + vla configuration & d & d & d & a + total bandwidth ( mhz ) & 3.125 & 3.125 & 3.125 & 50 + number of channels & 128 & 128 & 128 & 1 + spectral resolution ( khz / km s@xmath27 ) & 24.4/0.3 & 24.4/0.3 & 24.4/0.3 & + primary beam fwhm & 1.9 & 1.9 & 1.9 & 5.3 + synthesized beam & 45@xmath18434 & 46@xmath18434 & 77@xmath18457 & 031@xmath184019 + rms noise ( mjy beam@xmath27 channel@xmath27 ) & 4.5 & 4.5 & 3 & 0.003 + beam position angle & 4.8@xmath25 & @xmath240.6@xmath25 & 6.7@xmath25 & 20.1@xmath25 + conversion factor ( k / jy ) & 144 & 138 & 55.7 & 2.9 @xmath185 + smoothed synthesized beam & 60@xmath18460 & 60@xmath18460 & & + smoothed beam position angle & 0@xmath25 & 0@xmath25 & & + rms noise ( smoothed data , mjy beam@xmath27 channel@xmath27 ) & 6 & 6 & & + smoothed conversion factor ( k / jy ) & 60.5 & 60.3 & & + largest detectable angular scale & 94 & 94 & 100 & 16 + phase and amplitude calibrator & j1832@xmath24105 & j1832@xmath24105 & j1832@xmath24105 & j1832@xmath24105 + flux density calibrator & 3c48 & 3c48 & 3c286 & 3c48 + bandpass calibrator & 3c273 & 3c273 & 3c345 & + lcccc @xmath186 ( h@xmath43 ) ( m@xmath8 ) & @xmath180 & @xmath181 & @xmath182 & @xmath182 + @xmath186 ( dust ) ( m@xmath8 ) & 113 & 114 & & + @xmath187 ( pc ) & @xmath188 & @xmath189 & @xmath190 & @xmath191 + @xmath105 ( km s@xmath27 ) & @xmath192 & @xmath193 & @xmath194 & @xmath195 + @xmath196 ( m@xmath8 ) & @xmath197 & @xmath198 & @xmath199 & @xmath200 + @xmath196/@xmath186 & @xmath201 & @xmath202 & @xmath203 & @xmath204 + | we present very large array observations of ammonia ( nh@xmath0 ) ( 1,1 ) , ( 2,2 ) , and ccs ( @xmath1 ) emission toward the infrared dark cloud ( irdc ) g19.30 + 0.07 at @xmath2 ghz .
the nh@xmath0 emission closely follows the 8 @xmath3 m extinction .
the nh@xmath0 ( 1,1 ) and ( 2,2 ) lines provide diagnostics of the temperature and density structure within the irdc , with typical rotation temperatures of @xmath4 to 20 k and nh@xmath0 column densities of @xmath5 @xmath6 . the estimated total mass of g19.30 + 0.07 is @xmath71130 m@xmath8 .
the cloud comprises four compact nh@xmath0 clumps of mass @xmath9 to 160 m@xmath8 .
two coincide with 24 @xmath3memission , indicating heating by protostars , and show evidence of outflow in the nh@xmath0 emission .
we report a water maser associated with a third clump ; the fourth clump is apparently starless .
a non - detection of 8.4 ghz emission suggests that the irdc contains no bright hiiregions , and places a limit on the spectral type of an embedded zams star to early - b or later . from the nh@xmath0 emission we find g19.30 + 0.07 is composed of three distinct velocity components , or subclouds . "
one velocity component contains the two 24 @xmath3 m sources and the starless clump , another contains the clump with the water maser , while the third velocity component is diffuse , with no significant high - density peaks .
the spatial distribution of nh@xmath0 and ccs emission from g19.30 + 0.07 is highly anti - correlated , with the nh@xmath0 predominantly in the high - density clumps , and the ccs tracing lower - density envelopes around those clumps .
this spatial distribution is consistent with theories of evolution for chemically young low - mass cores , in which ccs has not yet been processed to other species and/or depleted in high - density regions . |
during the last two decades high precision observations have established the existence of two component of the universe .
these are known as dark energy and dark matter .
what are their fundamental nature ?
this one of the questions we need to answer@xcite .
one of the most succesful models to explain observations is the @xmath1cdm but it still has some problems . in the past
we have proposed a dark matter model in the framework of the newtonian limit of a scalar - tensor theory@xcite .
we have done @xmath0-body simulations using a numerical code developed to take into account the scalar field ( sf ) force contribution which is of yukawa type in its newtonian limit and we have use it to study several situations , like collision of protogalaxies@xcite , dynamics and collision of galaxies@xcite , and to study cosmological simulations@xcite .
recently we have used it to address a possible solution to two of the main problems the @xmath1cdm has , i.e. , the tendency to over populate the formed halos of galaxies with satellites and the flatness problem@xcite .
several authors have also been using similar dark matter models to study large - scale structure formation , see for instance@xcite .
one of the tools to characterize the clustering properties of a universe model is the power spectrum .
constraints on cosmological models and on cosmological parameters have been derived from measuring power spectrum .
therefore , in this work we consider a dark matter model build from the newtonian limit of a scalar - tensor theory@xcite and use it to study the large - scale structure formation of the universe by analyzing the power - spectrum versus the parameters of the model .
giving that we are interested in the study of the large - scale structure formation of the universe we will use the newtonian limit of the scalar - tensor theory .
this limit applies in a region small compared to the hubble length @xmath2 , with @xmath3 the speed of light and @xmath4 the hubble factor , and large compared to the schwarzschild radii of any collapsed objects . in this case
the vlasov - poisson equation in an expanding universe describes the evolution of the six - dimensional , one - particle distribution function , @xmath5@xcite .
the vlasov equation is , @xmath6 where @xmath7 is the comoving coordinate , @xmath8 , @xmath9 is the particle mass , and @xmath10 is the self - consistent gravitational potential given by the poisson equation , @xmath11 \right]\ ] ] where @xmath12 is the background mass density .
( [ vlasov_eq ] ) and ( [ poisson_eq ] ) form the vlasov - poisson equation , constitutes a collisionless , mean - field approximation to the evolution of the full @xmath0-body distribution .
an @xmath0-body code attempts to solve eqs .
( [ vlasov_eq ] ) and ( [ poisson_eq ] ) by representing the one - particle distribution function as @xmath13 substitution of ( [ discrete_distribution_eq ] ) in the vlasov - poisson system of equations yields the exact newton s equations for a system of @xmath0 gravitating particles . in the newtonian limit of stt of gravity , the newtonian motion equation for a particle @xmath14
is written as@xcite @xmath15 where the sum includes all periodic images of particle @xmath16 , and @xmath17 is @xmath18 which , for small distances compared to @xmath19 , is @xmath20 and , for long distances , is @xmath21 , as in newtonian physics . to simulate cosmological systems ,
the expansion of the universe has to be taken into account . also , to determine the nature of the cosmological model we need to determine the composition of the universe , i. e. , we need to give the values of @xmath22 for each component @xmath14 , taking into account in this way all forms of energy densities that exist at present . in this work
we will consider a model with only two energy density contribution .
one which is a pressureless and nonbaryonic dark matter with @xmath23 that does not couple with radiation .
other , that will be a cosmological constant contribution @xmath24 .
also , the model must be consistent with a static sf according to the newtonian limit we are using in this work .
thus , the scale factor , @xmath25 , is given by the following friedman model@xcite , @xmath26\ ] ] where @xmath27 , @xmath28 and @xmath29 are the matter and energy density evaluated at present , respectively .
we have also demanded that the universe be a flat one . to be consistent with the cmb spectrum and structure formation numerical experiments ,
cosmological constraints must be applied on @xmath30 in order for it to be within the range @xmath31 @xcite .
in this section , we present results of cosmological simulations of a @xmath1cdm universe with and without sf contribution .
we start our simulations with an initial distribution of @xmath32 particles in a box with sides of @xmath33 mpc at @xmath34 .
this case is similar to the one that comes with gadget 1 @xcite . at present epoch , @xmath35 , @xmath36 , @xmath37 km / s / mpc , @xmath38 .
we restrict the values of @xmath30 to the interval @xmath31 @xcite and use @xmath39 mpc , since this scale turns out to be an intermediate scale between the size of the clump groups and the separation of the formed groups .
however , we have analyzed elsewhere several values of @xmath19@xcite .
other universe models have been analyzed in @xcite .
because the visible component is the smaller one and given our interest to test the consequences of including a sf contribution to the evolution equations , our model excludes gas particles , but all its mass has been added to the dark matter . in fig .
1 we show the power spectrum at redshift @xmath40 of our @xmath1cdm model for several values of @xmath30 . the line without symbols shows the case of a @xmath1cdm model without the contribution of the scalar field . the line with ( @xmath41 ) is the case of a @xmath1cdm model with scalar field with @xmath42 , the line with ( @xmath43 ) uses @xmath44 , and the line with ( o ) uses @xmath45 .
one notes clearly how the sf modifies the clustering of matter of the system .
the most dramatic cases are the models with @xmath44 and @xmath42 .
we now analyze the general effect the parameter @xmath30 has on the clustering . for a given @xmath19
the role of @xmath30 in our approach can be seen in equations ( [ eq_motion])-([eq_f ] ) and is contained in the factor @xmath17 .
the factor @xmath46 augments ( diminishes ) for positive ( negative ) values of @xmath30 for small distances compared to @xmath19 , resulting in more ( less ) structure formation for positive ( negative ) values of @xmath30 compared to the @xmath1cdm model . in the case of the line with ( @xmath41 ) , for @xmath47
, the effective gravitational pull has been augmented by a factor of @xmath48 , in contrast to case shown with line and ( @xmath43 ) where it has diminished by a factor of 1/2 ; in the model with the line and ( o ) the pull diminishes only by a factor of 3/4 .
that is why one observes for @xmath49 more structure formation in case with line and ( @xmath41 ) , less in case with line and ( @xmath43 ) , and lesser in model with line and ( o ) .
the effect is then , for a growing positive @xmath30 , to speed up the growth of perturbations , then of halos and then of clusters , whereas negative @xmath30 values ( @xmath50 ) tend to slow down the growth .
the general gravitational effect of the scalar - tensor dark matter model is that the interaction changes by a factor @xmath51 for @xmath49 in comparison with the newtonian case . for negative values of @xmath30
the effect is to diminish the formation of structures and the opposite ocurrs for positive values of @xmath30 .
for @xmath52 the dynamics is essentially newtonian . by computing the power spectrum we were able to determine the effects of the scalar field contribution to the galaxy clustering
of course we will need to do a systematic and with greater resolution study and make detailed comparisons with observations to support or to rule out a dark model based on scalar - tensor theories .
10 m. a. rodrguez - meza , j. klapp , j.l .
cervantes - cota , and h. dehnen in _ exact solutions and scalar fields in gravity : recent developments _ , edited by a. macias , j.l .
cervantes - cota , and c. lmmerzahl , kluwer academic / plenum publishers , ny , 2001 , p. 213 . | we study the large - scale structure formation in the universe in the frame of scalar - tensor theories as an alternative to general relativity .
we review briefly the newtonian limit of non - minimally coupled scalar - tensor theories and the evolution equations of the @xmath0-body system that is appropriate to study large - scale structure formation in the universe .
we compute the power - spectrum of the universe at present epoch and show how the large - scale structure depends on the scalar field contribution . |
in this paper , we consider the system of nonlocal continuity equations @xmath0 \rho^i(0)=\bar \rho^i\,,&\qquad i \in \{1,\ldots , k\}\ , , \end{array } \right.\ ] ] where the unknown @xmath1 is a vector of measures , @xmath2 is a vector of convolution kernels and we set @xmath3 . for any time @xmath4 , if @xmath5 is a bounded measure on @xmath6 and @xmath7 is a bounded function on @xmath6 , then the convolution is taken with respect to space only and is defined as usually as @xmath8 .
for example in @xcite , the authors consider the scalar conservation law where @xmath9 is a nonlocal functional with respect to @xmath10 .
this equation stands for various models such as a sedimentation model , a supply - chain model , a pedestrian traffic model . for physical reason , in the following
we are looking for positive solutions possibly with concentration , i.e. for any time @xmath11 the solution has to be in @xmath12 . our goal here is to improve the results of @xcite , not only by considering a system , but also lightening the hypotheses on @xmath9 and @xmath13 .
we prove here existence and uniqueness of weak measure solutions to ( [ eq : system ] ) .
let us introduce the following sets of hypotheses : ( v ) : : the vector field @xmath14 is uniformly bounded and it is lipschitz in @xmath15 uniformly in time : @xmath16 ( @xmath17 ) : : the convolution kernel @xmath18 is uniformly bounded and it is lipschitz in @xmath19 uniformly in time : @xmath20 the main result of this paper is the following theorem .
[ thm : main ] let @xmath21 .
let us assume that @xmath9 satisfies * ( v ) * and @xmath22 satisfies * ( @xmath17)*. then there exists a unique solution @xmath23 to ( [ eq : system ] ) with initial condition @xmath24 .
we refer to section [ sec : solution ] for precise notations and definitions , in particular for the notion of solution .
[ rk : propsol ] assume @xmath9 satisfies * ( v ) * and @xmath22 satisfies * ( @xmath17)*. then theorem [ thm : main ] is completed by the following properties : * if @xmath25 then @xmath26 , up to redefinition on a negligible set of times , and for all time @xmath4 , for all @xmath27 we have @xmath28 . *
if @xmath29 then @xmath30 and for all time @xmath4 , we have @xmath31 with @xmath32 a constant dependent on @xmath33 , @xmath34 , @xmath35 and @xmath36 .
* let @xmath37 such that for any @xmath38 , @xmath39 .
let @xmath10 and @xmath40 be the solutions of ( [ eq : system ] ) associated to the initial conditions @xmath24 and @xmath41 , then we have the estimate : @xmath42 where @xmath43 and @xmath44 is the wasserstein distance of order one between @xmath45 and @xmath46 .
these properties are described in corollary [ cor : l1 ] and in proposition [ prop : stabgeneral ] .
the wasserstein distance of order one is rigorously defined in section [ sec : tools ] . in theorem [ thm : main ] as well as in the other results of this papers ,
it is in fact sufficient to require that @xmath47 is @xmath48 in @xmath49 and @xmath50 in @xmath51 .
indeed , @xmath52 is uniformly bounded by @xmath53 . consequently , denoting @xmath54 the closed ball of center 0 and radius @xmath55 in @xmath56 , it is sufficient to have @xmath57 .
note also that , restricting the definition of @xmath9 and @xmath13 to the time interval @xmath58 $ ] , we obtain a solution defined on the same time interval .
consequently , we can as well ask only @xmath9 and @xmath13 to be @xmath50 in time instead of @xmath48 .
the system ( [ eq : system ] ) stands for a variety of models .
let us present first a macroscopic model of pedestrian traffic . in a macroscopic pedestrian crowd model ,
@xmath10 is the density of the crowd at time @xmath11 and position @xmath59 and @xmath9 is a vector field giving the speed of the pedestrian . according to the choice of @xmath9 , various behaviors can be observed .
several authors already studied pedestrian traffic in two dimensions space ( @xmath60 ) .
some of these models are local in @xmath10 ( see @xcite ) ; other models use not only the local density @xmath61 but the entire distribution of @xmath10 , typically they depend on the convolution product @xmath62 ( see @xcite ) which represents the spatial average of the density . within the framework of ( [ eq : system ] )
, we can study the models presented in @xcite . in @xcite ,
the authors considered for @xmath9 the expression @xmath63 where @xmath64 is a scalar function giving the speed of the pedestrians ; @xmath13 is a convolution kernel averaging the density ; and @xmath65 is a bounded vector field giving the direction the pedestrian located in @xmath59 will follow .
this model is more adapted to the case of panic in which pedestrians will not deviate from their trajectory and will adapt their velocity to the averaged density . indeed
, even if the density is maximal on a given trajectory , if the averaged density is not maximal , the pedestrians will push , trying all the same to reach their goal . this behavior can be associated with rush phenomena in which people can even die due to overcompression ( e.g. on jamarat bridge in saudi arabia , see @xcite ) .
a similar model was introduced in piccoli & tosin @xcite , where the authors instead of an isotropic convolution kernel , consider a nonlocal functional taking into account the direction in which the pedestrians are looking .
in @xcite , the authors study the scalar case in the framework of krukov entropy solutions .
they obtained existence and uniqueness of weak entropy solutions under the hypotheses @xmath66 , @xmath67 , and @xmath68 .
this result was slightly improved in @xcite where , under the same set of hypotheses on @xmath64 , @xmath69 and @xmath13 , the authors consider a system instead of a scalar equation and obtain global in time existence and uniqueness of entropy solutions .
we recover these results with lighter hypotheses . indeed , although we consider weak measure solutions , these in fact are unique and consequently coincide with entropy solutions when the initial condition is in @xmath70 .
another model of crowd dynamics that we recover consists in the coupling of a group of density @xmath61 with an isolated agent located in @xmath71 .
this can modelize for example the interaction between groups of preys of densities @xmath10 and an isolated predator located in @xmath72 .
such a model was introduced in @xcite where the authors obtained existence and uniqueness of weak entropy solutions under very strong hypotheses .
we recover here partially the results concerning the coupling pde / ode of @xcite .
indeed , the measure framework allows us also to introduce particles / individuals through dirac measures .
for instance , let us assume that @xmath73 such that @xmath74 are in fact functions belonging to @xmath75 and that @xmath76 are dirac measures located in @xmath77 .
let us denote @xmath78 and @xmath79 .
we also denote with @xmath80 ( resp .
@xmath81 ) the vector fields ( resp .
kernels ) associated to @xmath10 , and with @xmath82 ( resp .
@xmath83 ) the vector fields ( resp .
kernels ) associated to @xmath72 .
note that @xmath84 . by definition of weak measure solution ( see definition [ def : weaksol ] ) , if @xmath85 , { \mathbb{r}}^d)$ ] , the dirac measures are satisfying , for any @xmath86 @xmath87 \,,\end{aligned}\ ] ] which can be rewritten @xmath88 consequently , in this case , the system ( [ eq : system ] ) becomes @xmath89 \dot p^j(t)=\phi^j\left(t , \,p(t),\ , \rho_t{{\boldsymbol { * } } } \eta_t^j ( p^j(t ) ) \right)\ , , & j \in \{1,\ldots , k_1\}\,.\\ \end{array } \right.\ ] ] so we are coupling ode with conservation laws .
system ( [ eq : system ] ) can also stands for models of aggregation , studied in @xcite under weaker hypotheses admitting singular kernels .
the system ( [ eq : system ] ) comprised also a model of particles sedimentation @xmath90 which has been introduced @xcite and studied in @xcite , where the author proved existence and uniquness of weak solutions with initial condition in @xmath48 .
finally , a similar nonlocal model is the one the supply - chain model @xcite , in which we consider the nonlocal term @xmath91 instead of a convolution product .
this last model was studied for example in @xcite with furthermore boundary conditions in @xmath92 and @xmath93 .
the proof of theorem [ thm : main ] is divided into two main steps .
first , we prove some a priori properties of the solutions ( see section [ sec : solution ] ) : mainly , we prove that the weak measure solutions of ( [ eq : system ] ) coincide with the lagrangian solutions of this system .
important consequences are the conservation of the regularity of the initial condition and the strong continuity in time in the case the solution is a function , as stated in remark [ rk : propsol ] .
second , we prove the existence and uniqueness of lagrangian solutions thanks to a fixed point argument ( see section [ sec : proofmaintheorem ] ) . indeed , introducing the set of probability measures endowed with the wasserstein distance of order one , we are able to prove a stability estimate with respect to the nonlocal term ( see section [ sec : stability ] ) .
the technique used there is quite similar to the one of loeper @xcite , who studied the vlasov - poisson equation and the euler equation in vorticity formulation .
this article is organized as follows : in section [ sec : solution ] we define the two different notions of solution and prove that they coincide . in section
[ sec : tools ] we give some useful tools on optimal tranport ; in section [ sec : stability ] we prove an important lemma giving a stability estimate and in section [ sec : proofmaintheorem ] , we give the proof of theorem [ thm : main ] .
let @xmath94 be the space dimension and @xmath95 be the size of the system . in the following , @xmath96 is the set of matrices of size @xmath97 with real values and @xmath98 is the set of matrices of size @xmath99 with real values .
we denote by @xmath100 ( resp .
@xmath101 ) the set of bounded ( resp . bounded and positive ) measure on @xmath6 and by @xmath102 the set of probability measures on @xmath6 , that is the set
of bounded positive measures with total mass 1 . in the following the lipschitz norms with respect to @xmath59 or @xmath51
are taken uniformly with respect to the other variables .
that is to say , for example : @xmath103 let us also underline that in the computations , we considered the norm 1 on the vectors in @xmath56 .
when considering another norm , a constant depending on @xmath104 appears in the various estimates .
similarly , if @xmath105 we define the total measure of @xmath10 as @xmath106 .
the space @xmath107 , { \mathcal{m}}^+({\mathbb{r}}^d))$ ] consists of the parametrized measures @xmath108}$ ] such that , for any @xmath109 , the application @xmath110 is measurable and such that @xmath111 } { { \left\|\mu_t\right\|}}_{{\mathcal{m}}}<\infty$ ] .
we say that @xmath112,{\mathcal{m}}^+({\mathbb{r}}^d)^k)$ ] is a _ weak measure solution _ of ( [ eq : system ] ) with initial condition @xmath113 if , for any @xmath114 and for any test - function @xmath115-\infty , t[\times { \mathbb{r}}^d , { \mathbb{r}})$ ] we have @xmath116 { \mathinner{\mathrm{d}{\rho^i_t ( x)}}}{\mathinner{\mathrm{d}{t } } } + \int_{{\mathbb{r}}^d } \phi(0,x){\mathinner{\mathrm{d}{\bar \rho^i(x)}}}=0\,.\ ] ] a priori for weak measure solutions of the continuity equation @xmath117 , with a given vector field @xmath118 , we have only continuity in time for the weak topology ( see @xcite ) , that is to say , for all @xmath114 , for all @xmath119 , the application @xmath120 is continuous , up to redefinition of @xmath45 on a negligible set of times . in the case of the system ( [ eq : system ] ) , we have a gain of regularity in time when the initial condition is a function in @xmath121 ( see corollary [ cor : l1 ] ) .
when @xmath122 is a measure on @xmath123 and @xmath124 a measurable map , we denote @xmath125 the push - forward of @xmath126 , that is the measure on @xmath127 such that , for every @xmath128 , @xmath129 if we assume that @xmath126 and @xmath130 are absolutely continuous with respect to the lebesgue measure so that there exist @xmath131 such that @xmath132 and @xmath133 , and that @xmath134 is a @xmath135-diffeomorphism , then we have the _ change of variable formula _ @xmath136 besides , we denote @xmath137 the _ projection on the first coordinate _ ; that is , for any @xmath138 , @xmath139 . in a similar way , @xmath140 is the _ projection on the second coordinate _ ; that is , for any @xmath138 , @xmath141 .
we say that @xmath142 , { \mathcal{m}}^+({\mathbb{r}}^d)^k)$ ] is a lagrangian solution of ( [ eq : system ] ) with initial condition @xmath113 if , for any @xmath114 , there exists an ode flow @xmath143\times { \mathbb{r}}^d \to { \mathbb{r}}^d$ ] , that is a solution of @xmath144 { \displaystyle}x^i(0,x)=&x\ , ; \end{array } \right.\ ] ] and such that @xmath145 where @xmath146 is the map defined as @xmath147 for any @xmath148\times{\mathbb{r}}^d$ ] .
[ rk : blip ] assume @xmath9 satisfies * ( v ) * and @xmath13 satisfies * ( @xmath17)*. then , for any @xmath149 , { \mathcal{m}}^+({\mathbb{r}}^d)^k)$ ] , the vector fields @xmath150 lipschitz in @xmath59 and @xmath151 consequently , if @xmath152 is uniformly bounded , the ode flow @xmath153 above is always well - defined , for a fixed @xmath10 . if @xmath154 , then the push - forward formula ( [ eq : changeofvariable ] ) becomes , for a.e .
@xmath155 , @xmath156 we now show that the two notions of solution in fact coincide . if @xmath10 is a lagrangian solution of ( [ eq : system ] ) , then @xmath10 is also a weak measure solution of ( [ eq : system ] ) .
conversely , if @xmath10 is a weak measure solution of ( [ eq : system ] ) , then @xmath10 is also a lagrangian solution of ( [ eq : system ] ) .
let @xmath10 be a lagrangian solution of ( [ eq : system ] ) .
let us denote @xmath157 and let @xmath153 be the ode flow associated to @xmath158 .
then , for any @xmath159-\infty , t[\times { \mathbb{r}}^d , { \mathbb{r}})$ ] , we have @xmath160 which proves that @xmath10 is also a weak measure solution .
* let @xmath10 be a weak measure solution of ( [ eq : system ] ) .
for any @xmath114 , let us denote @xmath161 .
let @xmath40 be the lagrangian solution of the equation @xmath162 which exists and is unique since @xmath158 is lipschitz as noted in remark [ rk : blip ] . then , arguing similarly as in point 1 , @xmath40 is also a weak measure solution to ( [ eq : lag ] ) . denoting @xmath163 , we obtain that @xmath164 is a weak measure solution of the equation @xmath165 with initial condition @xmath166 .
consequently , for any @xmath159-\infty , t[\times
{ \mathbb{r}}^d , { \mathbb{r}})$ ] , @xmath167 let @xmath168-\infty , t[\times{\mathbb{r}}^d , { \mathbb{r}})$ ] .
since @xmath169\times { \mathbb{r}}^d , { \mathbb{r}}^d)$ ] is lipschitz in @xmath59 , by computation along the characteristics , we can find @xmath170-\infty , t[\times{\mathbb{r}}^d , { \mathbb{r}})$ ] so that @xmath171 .
hence , for any @xmath172-\infty , t[\times{\mathbb{r}}^d , { \mathbb{r}})$ ] , we have @xmath173 , which implies @xmath174 a.e . and @xmath175 a.e
. consequently , we have also @xmath176 , and @xmath177 is finally a lagrangian solution of ( [ eq : system ] ) .
[ def : weaklag ] as a consequence of the previous theorem , in the following we simply call _ solution _ of ( [ eq : system ] ) a weak measure solution or a lagrangian solution of ( [ eq : system ] ) , that in fact coincide .
it is now possible to prove some of the properties given in remark [ rk : propsol ] .
[ cor : l1 ] assume that @xmath9 satisfies * ( v ) * and @xmath22 satisfies * ( @xmath17)*. let @xmath149 , { \mathcal{m}}^+({\mathbb{r}}^d)^k)$ ] be a solution to ( [ eq : system ] ) with initial condition @xmath178 .
* if @xmath179 .
then we have @xmath180 , { { \mathbf{l}^1}}({\mathbb{r}}^d , ( { \mathbb{r}}^+)^k))$ ] and for all time @xmath181
$ ] , all @xmath182 , @xmath28 . *
if furthermore @xmath183 , then for all @xmath181 $ ] we have @xmath184 and we have the estimate @xmath185 where @xmath32 depends on @xmath186 , @xmath9 and @xmath13 .
let @xmath10 be a solution of ( [ eq : system ] ) with initial condition @xmath187 . according to definition [ def : weaklag ] , @xmath10 is a lagrangian solution associated to a flow @xmath188 and we have immediately that @xmath189 .
besides , as @xmath190\times { \mathbb{r}}^d , { \mathbb{r}}^d)$ ] is bounded in @xmath11 and lipschitz in @xmath59 , then @xmath191 and we can use the change of variable formula ( [ eq : cov ] ) . if @xmath192 , with @xmath193 we obtain the desired @xmath48 bound and @xmath194 for all @xmath181 $ ] .
the continuity in time can be proved directly by estimating @xmath195 using egorov theorem .
this computation is straightforward although a bit long so we prefer to omit the details . besides , note that the continuity in time is also ensured by the results of diperna & lions ( * ? ? ?
* section 2.ii ) and the notion of renormalized solutions .
let us remind the definition of the wasserstein distance of order 1 .
[ def : wass ] let @xmath126 , @xmath196 be two borel probability measures on @xmath6 .
we denote @xmath197 the set of _ plans _ , that is the set of probability measures @xmath198 such that @xmath199 and @xmath200 .
we define the _
wasserstein distance of order one _ between @xmath126 and @xmath196 by @xmath201 let @xmath202 , @xmath203 be two vectors such that @xmath204 and @xmath205 are borel probability measures on @xmath6 . we define
the _ wasserstein distance of order one _ between @xmath10 and @xmath40 , denoted @xmath206 , as @xmath207 [ rk : existmin ] by ( * ? ? ?
* theorem 1.3 ) , for any @xmath208 , there exist a plan @xmath209 realizing the minimum in the wasserstein distance so that @xmath210 [ rk : inf ] let @xmath113 be a probability measure ; and let @xmath211 be mappings such that @xmath212 and @xmath213 . then , the probability measure @xmath214 satisfies @xmath215 , @xmath216 and so @xmath217 [ prop : villani ] let @xmath218 be two probability measures .
the wasserstein distance of order one between @xmath126 and @xmath196 satisfies @xmath219
in the following we consider probability measures instead of bounded positive measures . this is not a real restriction since we pass from one case to the other just by a rescaling . before giving a stability estimate in proposition [ prop : stabgeneral ] , we prove a technical lemma . [
lem : stability ] let @xmath9 satisfy * ( v ) * and @xmath22 satisfy * ( @xmath17)*. let @xmath220 . for any @xmath182 , we have the following estimate @xmath221 in the previous lemma , the quantity @xmath222 could be infinite .
if we restrict ourselves to bounded positive measures with first moment finite , then the quantity above is always finite .
the proof follows from proposition [ prop : villani ] on the wasserstein distance .
note first that in the case @xmath223 then @xmath224 is constant and we have @xmath225 .
now , in the case @xmath226 , thanks to proposition [ prop : villani ] , we have @xmath227 as we obtain the same estimate for @xmath228 , we can conclude that @xmath229 let @xmath230 , \mathcal{p}({\mathbb{r}}^d)^k)$ ] .
we want to compare the following equations , in which the nonlocal has been replaced by fixed applications , so that the system is made of decoupled equations .
@xmath231 \textrm{for all } i \in \{1,\ldots , k\}\qquad \quad { \partial}_t \sigma^i + { { \rm div}\,}\left(\sigma^i u^i(t , x , \nu^i { { \boldsymbol { * } } } s)\right)=0\,,\qquad \quad \sigma^i(0,\cdot)=\bar\sigma^i \ , .
\end{array}\ ] ] [ prop : stabgeneral ] assume @xmath232 satisfy * ( v ) * and @xmath233 satisfy * ( @xmath17)*. let @xmath24 , @xmath41 be two probability measures such that for any @xmath38 , @xmath39 .
let @xmath230,\mathcal{p}({\mathbb{r}}^d)^k)$ ] .
if @xmath10 and @xmath40 are lagrangian solutions of ( [ eq : system2 ] ) associated to the initial conditions @xmath24 and @xmath41 , then we have the estimate : @xmath234 } \mathscr{w}_1 ( r_t , s_t ) + { { \left\|\eta-\nu\right\|}}_{{{\mathbf{l}^\infty}}}+ { { \left\|v - u\right\|}}_{{{\mathbf{l}^\infty } } } \right]\label{eq : estimate}\end{aligned}\ ] ] where @xmath32 is a constant depending on @xmath33 , @xmath34 , @xmath35 and @xmath186 .
furthermore , in the special case @xmath235 and @xmath236 , we get : @xmath237\,,\label{eq : stab}\end{aligned}\ ] ] where @xmath238is a constant depending on @xmath33 , @xmath34 , @xmath35 and @xmath186 .
note that the estimate above comprises the case @xmath239 .
let @xmath10 , @xmath40 be two lagrangian solutions to the cauchy problem for ( [ eq : system ] ) with initial conditions @xmath24 and @xmath41 respectively .
let @xmath188 , @xmath240 be the associated ode flows . for any @xmath181
$ ] , we define the map @xmath241 by @xmath242 let @xmath243 so that @xmath244 and @xmath245 .
let us define the probability measure @xmath246 .
then , @xmath247 and @xmath248 so that @xmath249 .
let @xmath250 , we define , for @xmath4 @xmath251 note first that @xmath252 is lipschitz . indeed , let @xmath253 , then we have @xmath254 let us assume that @xmath255 , otherwise the thesis is trivial . then
, by remark [ rk : existmin ] , for all @xmath256 , we can find a bounded positive measure @xmath243 so that @xmath257 consequently we have , for any @xmath258 , @xmath259 .
hence , using ( [ eq : qlip ] ) , for any @xmath4 , we have @xmath260 thus , for any @xmath4 , @xmath261 remains finite when @xmath262 and since @xmath263 is increasing with respect to @xmath264 , we can define @xmath265 .
let us now consider @xmath266 .
the same computation as in ( [ eq : qlip ] ) ensures that @xmath266 is lipschitz so we can differentiate for almost every @xmath11 and obtain @xmath267 & & \qquad+ { { \left|v^i\left(t , y^i_t(y ) , r_t{{\boldsymbol { * } } } \eta_t^i(y^i_t(y))\right)-v^i\left(t , y^i_t(y ) , s_t { { \boldsymbol { * } } } \eta_t^i(y^i_t(y))\right)\right|}}\\[2pt ] & & \qquad+ { { \left|v^i\left(t , y^i_t(y ) , s_t{{\boldsymbol { * } } } \eta_t^i(y^i_t(y))\right)-v^i\left(t , y^i_t(y ) , s_t { { \boldsymbol { * } } } \nu_t^i(y^i_t(y))\right)\right|}}\\ & & \qquad+ { { \left|v^i\left(t , y^i_t(y ) , s_t{{\boldsymbol { * } } } \nu_t^i(y^i_t(y))\right)-u^i\left(t , y^i_t(y ) , s_t { { \boldsymbol { * } } } \nu_t^i(y^i_t(y))\right)\right| } } \bigg){\mathinner{\mathrm{d}{\gamma^i_0 ( x , y)}}}\,.\end{aligned}\ ] ] note that @xmath268 using lemma [ lem : stability ] we obtain @xmath269\,.\label{eq : stima}\end{aligned}\ ] ] taking the @xmath270 in time of @xmath271 on the right - hand side and applying gronwall lemma , we get @xmath272}\mathscr{w}_1(r_\tau , s_\tau)+ { { \left\| \eta-\nu\right\|}}_{{{\mathbf{l}^\infty}}}+ { { \left\|u - v\right\|}}_{{{\mathbf{l}^\infty}}}\right)\,.\end{aligned}\ ] ] note now that , thanks to remark ( [ rk : inf ] ) , for any @xmath4 @xmath273 furthermore , we have chosen @xmath274 in an optimal way thanks to remark [ rk : existmin ] so that @xmath275 . hence we obtain , for any @xmath181 $ ] : @xmath276}\mathscr{w}_1(r_\tau , s_\tau)+ { { \left\| \eta-\nu\right\|}}_{{{\mathbf{l}^\infty}}}+ { { \left\|u - v\right\|}}_{{{\mathbf{l}^\infty}}}\right)\,,\end{aligned}\ ] ] which is the expected result ( [ eq : estimate ] ) . in the particular case
@xmath235 and @xmath236 , applying ( [ eq : wq ] ) to ( [ eq : stima ] ) we obtain @xmath277 applying gronwall lemma , we finally obtain @xmath278 , which is ( [ eq : stab ] ) .
the proof of theorem [ thm : main ] is based on the following idea : let us fix the nonlocal term and , instead of ( [ eq : system ] ) , we study the cauchy problem @xmath279 where @xmath51 is a given application .
we consider here probability measures . in the more general case of positive measures with the same total mass , by rescaling we are back to the case of probability measures .
( a ) : : the space @xmath282 is equipped with a distance @xmath283 that makes @xmath282 complete : for @xmath284 , we define @xmath285 } \mathscr{w}_1(\mu_t , \nu_t)\,.\ ] ] ( b ) : : the application @xmath286 is well - defined : the lagrangian solution @xmath287 to ( [ eq : fix ] ) exists and is unique ( for a fixed @xmath51 ) .
indeed , let @xmath288 be the ode flow associated to @xmath289 , then we can define @xmath290 . since @xmath291 is a positive measure , then
so is @xmath45 .
( c ) : : the application @xmath286 is a contraction : this is given by proposition [ prop : stabgeneral ] .
indeed , let @xmath292 in @xmath293 , { \mathcal{m}}^+({\mathbb{r}}^d)^k)$ ] and denote @xmath294 , @xmath295 the associated solutions to ( [ eq : fix ] ) .
note that @xmath10 and @xmath40 have the same initial condition .
thanks to proposition [ prop : stabgeneral ] , we obtain the contraction estimates @xmath296}\mathscr{w}_1(\rho_t , \sigma_t){\leqslant}c t e^{ct } \sup_{[0,t]}\mathscr{w}_1(r_t , s_t)\,,\ ] ] where @xmath32 depends only on @xmath33 , @xmath34 , @xmath35 and @xmath186 . hence , for @xmath134 small enough , by the banach fixed point theorem we obtain existence and uniqueness in @xmath282 of a lagrangian solution to ( [ eq : system ] ) for @xmath181 $ ] . as @xmath297 the coefficient @xmath32 does not depend on time and we can iterate the procedure .
thus we have existence and uniqueness on @xmath298 . | in this paper , we prove existence and uniqueness of measure solutions for the cauchy problem associated to the ( vectorial ) continuity equation with a non - local flow . we also give a stability result with respect to various parameters .
keywords : system of conservation laws , continuity equations , wasserstein distance , nonlocal flow , pedestrian traffic .
2010 msc : 35l65 |
in the previous investigations the significant influence of the relevant ion - atom absorption processes on the solar photosphere opacity was already demonstrated .
so , in @xcite and @xcite ) have been studied such symmetric ion - atom processes , as the molecular ion h@xmath3 photo - dissociation @xmath8 and the absorption charge exchange in ( h@xmath1 + h)-collisions @xmath9 where h = h(1s ) , h@xmath3 is the hydrogen molecular ion in the ground electronic state , and @xmath10 - the energy of a photon with the wavelength @xmath11 . the significance of these processes was established within the solar photosphere models from @xcite and @xcite in the optical , and from @xcite in far uv and euv regions of @xmath11 .
later , the symmetric processes ( [ eq : sim1 ] ) - ( [ eq : sim2 ] ) were included _
ab initio _ in one of the new solar photosphere models @xcite .
then , in @xcite was undertaken the investigation of some non - symmetric ion - atom absorption processes , namely the photo - dissociation and photo - association of the the molecular ions @xmath12 @xmath13 and the absorption charge - exchange in the ion - atom collisions @xmath14 where @xmath5 is the ground state atom of one of metals , relevant for the used solar photosphere model , whose ionization potential @xmath15 is smaller than the hydrogen atom ionization potential @xmath16 , @xmath4 - the corresponding atomic ion in its ground state , h@xmath4 and ( @xmath17 - the molecular ion in the electronic states which are adiabatically correlated ( at the infinite internuclear distance ) with the states of the ion - atom systems h + @xmath4 and h@xmath1 + @xmath5 respectively .
these processes were examined within the same solar photosphere model as in @xcite , i.e. the model c from @xcite , with @xmath5 = mg , si and al .
also , in accordance with the composition of the solar atmosphere , the processes of the type ( [ eq : nonsim1 ] ) - ( [ eq : nonsim2 ] ) , but with atom he(1s@xmath18 ) and ion h@xmath1 instead h and @xmath4 , were included in the consideration .
however , it was established that in the case of this atmosphere ( for the difference of some helium reach stellar atmospheres considered in @xcite ) such processes can be practically neglected .
since the ion - atom systems with the mentioned @xmath5 are strongly non - symmetric , the examined processes generate the quasi - molecular absorption bands in the neighborhoods of @xmath11 which correspond to the energies @xmath19 . according to the values of @xmath20 with the mentioned x these absorption bands lie in the part of the far uv region .
we should note here that the photo - dissociation of the molecular ion hsi@xmath1 was considered first time from the astrophysical aspect ( interstellar clouds and the atmospheres of red giant stars ) in @xcite .
in @xcite was shown that , in the case of the quiet sun , ion - atom processes ( [ eq : sim1 ] ) - ( [ eq : sim2 ] ) and ( [ eq : nonsim1 ] ) - ( [ eq : nonsim2 ] ) together become seriously concurrent to some other relevant absorbtion processes in far uv and euv regions within the whole solar photosphere .
this result is especially important , since among all possible ion - atom non - symmetric processes only those of them , for which the needed data about the corresponding molecular ions were known , were taken into account . because of that , it was natural to conclude that the non - symmetric processes ( [ eq : nonsim1 ] ) - ( [ eq : nonsim2 ] ) should be also included _ ab initio _ in the corresponding solar photosphere models .
one can see that the previous investigations of the ion - atom absorption processes were performed only in the case of the quiet sun .
however , it is well known how significant role for the solar atmosphere the sunspots play , and certainly it was interesting to see what is the situation with these processes in such objects . because of that , this investigation , whose some preliminary results were referred recently on a corresponding astrophysical conference @xcite , was undertaken .
all considerations were within the sunspot model m from @xcite .
such choice was caused by the fact that only this model , among other models mentioned in the literature ( see e.g. @xcite ) , provided all data needed for the calculations of the absorption coefficients which characterize the considered absorption processes .
certainly , the ion - atom absorption processes of the type ( [ eq : nonsim1 ] ) - ( [ eq : nonsim2 ] ) were included in the considerations .
however , here we take into account also additional processes of the molecular ion photo - dissociation and photo - association @xmath21 @xmath22 where @xmath23 is the ion in excited state with the excitation energy @xmath24 and h@xmath23 - the molecular ion in the electronic state adiabatically correlated with the state of the h + @xmath23 system , as well as the corresponding collisional excitation @xmath25 which realizes over creation of the quasimolecular complex ( h + @xmath26 in the excited electronic state adiabatically correlated with the state of the same system h + @xmath23 as in the process ( [ eq : sat3 ] ) .
depending of the values of @xmath27 the absorption bands generated by the processes ( [ eq : sat1 ] ) - ( [ eq : sat2 ] ) can lie not only in far uv region of @xmath11 , but also in near uv and visible regions .
the reason for the consideration of these processes is the fact that the absorption bands generated by some of them overlap with the bands generated by the processes ( [ eq : sim1 ] ) - ( [ eq : sim2 ] ) and ( [ eq : nonsim1 ] ) - ( [ eq : nonsim2 ] ) .
if the radiative transition @xmath28 is allowed by the dipole selection rules the corresponding absorption band can be treated , as in the case of the similar phenomena caused by atom - atom collisions ( @xcite ) , i.e. as the satellite of the ion spectral line connected with the mentioned transition .
the basic task of this investigation is to estimate the significance of the symmetric and non - symmetric ion - atom absorption processes in the case of the sunspot with respect to the processes of the negative hydrogen ion h@xmath29 photo - detachment and inverse `` bremsstrahlung '' in ( @xmath30 + h)-collisions , namely @xmath31 @xmath32 where @xmath30 and @xmath33 denote the free electron in initial and final channel , which , similarly to the previous papers , are treated here as the referent processes .
it is connected with the concept of this paper which stays the same as in all previous papers .
namely , the aim of these papers ( see e.g. @xcite ) was to pay attention on the considered ion - atom radiative processes as the factors of the influence on the opacity of the solar atmosphere .
for that purpose it was needed ( and enough in the same time ) to show that the efficiency of these processes in the considered spectral region is close to the efficiency of some known radiative processes whose significance for the solar atmosphere is accepted in literature .
it is clear that in the case of this atmosphere just the electron - atom processes ( [ eq : eat1 ] ) and ( [ eq : eat2 ] ) can be taken as the referent ones .
because of that in these previous papers many other radiative processes have not been considered , including here the certainly very important processes of the metal atom photo - ionization , which were already discussed in the literature in connection with the quiet sun atmosphere in @xcite .
we mean the processes @xmath34 where @xmath35 denotes the given metal atom in the ground state , i.e. @xmath5 , or in any possible ( under the considered conditions ) excited state , i.e. @xmath36 . however , within the sunspot we have a significantly smaller temperature than in the quiet sun photosphere ( see figure about model ) and consequently it is possible to expect there more larger efficiency of these photo - ionization processes , so that the position of the mentioned electron - atom processes as the referent ones is not so clear . because of that the processes of the metal atom photo - ionization were taken into account from the beginning of this investigation ( some of them were considered already in @xcite ) . in accordance with above mentioned the following absorption processes are included in the consideration in this work : + - the symmetric ion - atom processes ( [ eq : sim1 ] ) - ( [ eq : sim2 ] ) ; + - the non - symmetric processes ( [ eq : nonsim1 ] ) - ( [ eq : nonsim2 ] ) with @xmath5 = na , ca , mg , si and al , for which the needed data about the corresponding molecular ions h@xmath4 are known ; + - the non - symmetric processes ( [ eq : sat1 ] ) - ( [ eq : sat2 ] ) with @xmath23 = ca@xmath37 , ca@xmath38 , ca@xmath39 , ca@xmath40 , ca@xmath41 and al@xmath42 , + - the electron - atom processes ( [ eq : eat1 ] ) and ( [ eq : eat2 ] ) ; + - the photo - ionization processes ( [ eq : phion ] ) with all metal atoms relevant for the used sunspots model , i.e. including the case @xmath5 = fe . + let us note that as in @xcite only such non - symmetric processes are taken into account here for which all needed data about the corresponding molecular ions are known from the literature . it is well known that inside sunspot is needed to take into account the presence of its magnetic field . in connection with this we have to note that according to the existing data ( see e.g. @xcite ) this field is always not larger than 4000 gs . this is very important for us since it can be shown that in our further considerations the presence of such magnetic field can be completely neglected and all needed calculations can be performed as in the case of the quiet sun . the aims of this work request determination of the corresponding spectral absorption coefficients for all mentioned ion - atom processes , as well as for the concurrent absorption processes ( [ eq : eat1 ] ) - ( [ eq : eat2 ] ) and ( [ eq : phion ] ) , as functions of @xmath11 and the height @xmath43 above the referent solar atmosphere layer .
in this context the processes ( [ eq : sim1 ] ) - ( [ eq : sim2 ] ) , ( [ eq : nonsim1 ] ) - ( [ eq : nonsim2 ] ) and ( [ eq : sat1 ] ) - ( [ eq : sat2 ] ) are treated as the processes from the groups `` 1 '' , `` 2 '' and `` 3 '' respectively , which is denotes by the corresponding index : @xmath44 = 1 , 2 or 3 .
let us note that here the spectral characteristics of the non - symmetric processes of the type ( [ eq : nonsim1 ] ) - ( [ eq : nonsim3 ] ) , but with @xmath5 = li , are also determined in rather wide regions of the temperatures and wavelengths .
namely , such processes could be of interest for lithium rich stellar atmospheres ( `` li stars '' , @xcite ) as an additional canal for the creation of the neutral lithium atoms .
all relevant matter is distributed below in five sections and two appendices .
so , the expressions of the considered absorbtion processes are given in the sections 2 and 3 , together with the needed comments about the methods of their determination .
the section 4 contains the needed comments about used calculation methods .
the results of the calculations of the spectral coefficients and other quantities , which characterize the relative efficiencies of the considered processes , are presented ( with the corresponding discussions ) in the section 4 , and in the last section 5 are given some conclusions and are indicated directions of the further investigations .
then , the potential curves and dipole matrix elements of the molecular ions h@xmath4 with @xmath5 = na and li , which are determined within this work , are given in appendix a. finally , some of the spectral characteristics of the processes ( [ eq : nonsim1 ] ) - ( [ eq : nonsim2 ] ) and ( [ eq : sat1 ] ) - ( [ eq : sat2 ] ) , which can be used in some other applications , are presented in appendix b.
the above mentioned ion - atom absorption processes , namely : + - the photo - dissociation ( bound - free ) processes ( [ eq : sim1 ] ) , ( [ eq : nonsim1 ] ) and ( [ eq : sat1 ] ) , + - photo - association ( free - bound ) processes ( [ eq : nonsim3 ] ) and ( [ eq : sat3 ] ) , + - absorption charge - exchange and collisional excitation ( free - free ) processes ( [ eq : sim2 ] ) , ( [ eq : nonsim2 ] ) and ( [ eq : sat2 ] ) + are illustrated by fig . [
fig : schem ] .
the efficiencies of these processes are characterized in this work by the corresponding partial spectral absorption coefficients denoted with : + - @xmath45 , @xmath46 and @xmath47 , + - @xmath48 and @xmath49 - @xmath50 , @xmath51 and @xmath52 respectively . in accordance with the previous papers these coefficients
are given by the similar expressions , namely @xmath53 @xmath54 @xmath55 where @xmath56 , @xmath57 and @xmath58 , where @xmath44 = 2 and 3 , are the corresponding rate coefficients .
with @xmath59 , @xmath60 , @xmath61 and @xmath62 are denoted here the local temperature and densities of the hydrogen atoms and ions h@xmath1 and @xmath4 respectively , whose values are taken ( for each @xmath43 ) from the used sunspot model of @xcite . here , as in the previous papers
, it is understood that the photo - dissociation rate coefficients are defined by the known relations
@xmath63 @xmath64 @xmath65_{t},\ ] ] @xmath66_{t},\ ] ] where @xmath67 and @xmath68 , where @xmath44 = 2 and 3 , are the mean thermal photo - dissociation cross - sections : for the molecular ions h@xmath3 in the process ( [ eq : sim1 ] ) , and for ions h@xmath4 in the processes ( [ eq : nonsim1 ] ) and ( [ eq : sat1 ] ) respectively . with @xmath69 and @xmath70
are denoted the densities of the molecular ions @xmath71 and @xmath72 , and the designation @xmath73_{t}$ ] denotes that the factors @xmath74 and @xmath75 are determined under the condition of the local thermodynamical equilibrium ( lte ) with given @xmath59 .
the total efficiency of the symmetric ion - atom processes ( [ eq : sim1 ] ) - ( [ eq : sim2 ] ) , i.e. group `` 1 '' , is characterized here by the spectral absorption coefficient @xmath76 , given by @xmath77 in connection with the non - symmetric processes we will introduce firstly the total efficiencies of the processes ( [ eq : nonsim1 ] ) - ( [ eq : nonsim2 ] ) and ( [ eq : sat1 ] ) - ( [ eq : sat2 ] ) , i.e. groups `` 2 '' and `` 3 '' , for given @xmath5 which are characterized by the spectral absorption coefficients @xmath78 and @xmath79 .
they are given by @xmath80 @xmath81 where @xmath44 = 2 and 3 , and @xmath5 is the metal atom relevant in these cases .
than , the total efficiencies of the whole groups `` 2 '' and `` 3 '' , i.e. the non - symmetric processes ( [ eq : nonsim1 ] ) - ( [ eq : nonsim2 ] ) and ( [ eq : sat1 ] ) - ( [ eq : sat2 ] ) with all relevant @xmath5 , are characterized by the spectral absorption coefficients @xmath82 and @xmath83 given by @xmath84 where @xmath85 denotes that the summation is performed over all metal atoms relevant for the processes from the corresponding groups . as the consequence
, the efficiency of symmetric and non - symmetric processes ( [ eq : nonsim1 ] ) - ( [ eq : nonsim2 ] ) is characterized by the spectral absorption coefficient @xmath86 given by @xmath87 and the total efficiency of all considered ion - atom absorption processes - by the corresponding spectral absorption coefficient @xmath88 , namely @xmath89 where the coefficients @xmath90 is given by eq .
( [ eq : kapansim2 - 3 ] ) .
the efficiency of the electron - atom absorption processes ( [ eq : eat1 ] ) and ( [ eq : eat2 ] ) are characterized here by the corresponding bound - free ( bf ) and free - free ( ff ) spectral absorption coefficients @xmath91 and @xmath92 .
they are taken in the usual form @xmath93 @xmath94 from where it follows that the efficiency of these electron - atom processes together is characterized here by the total absorption coefficient @xmath95 given by @xmath96 where @xmath97 and @xmath98 are the corresponding partial rate coefficients .
the first of them is defined by the known relations @xmath99 , \end{split}\ ] ] where @xmath100 is the spectral cross - section for the ion h@xmath29 photo - detachment , the designation @xmath73_{t}$ ] - is already defined above , and @xmath101 and the free electron density @xmath102 are taken from the used sunspot model .
than , the photo - ionization processes ( [ eq : phion ] ) , which are connected with the metal atoms of the given kinds , are characterized here by the corresponding effective spectral absorption coefficient @xmath103}(\lambda;h)$ ] , namely @xmath104}(\lambda;t)\cdot n([x];h),\ ] ] where @xmath105 $ ] marks the kind of the given metal atoms , @xmath106;h)$ ] - the mean local density of all these atoms , i.e. the atoms in the ground state and in all excited states atoms which are realized with existing @xmath107 and @xmath59 , and @xmath108}(\lambda;)$ ] is the corresponding mean ( for given @xmath59 ) photo - ionization cross - sections .
as it is known the spectral rate coefficients introduced in eqs .
( [ eq : kapabf ] ) - ( [ eq : kapafb ] ) are calculated using the data about the relevant molecular ions characteristics .
we mean the potential curves of the initial and final electronic states of the considered molecular ions and the corresponding transition dipole moments .
the behavior of these characteristics as the functions of the internuclear distance @xmath109 was already shown in the cases of the symmetric and non - symmetric ( @xmath5 = mg and si ) ion - atom processes in @xcite and @xcite respectively .
also , it is illustrated here by figure in the appendix a in the case @xmath5 = na .
the potential curves and transition dipole moments needed for the non - symmetric processes ( [ eq : nonsim1 ] ) - ( [ eq : nonsim2 ] ) with @xmath6 na and li are determined within this work and presented in appendix a. for other above described non - symmetric ion - atom processes they are taken from @xcite , @xcite , @xcite and @xcite , and for symmetric processes - from @xcite . in this work all mentioned ion - atom processes , excluding ( [ eq : sim2 ] ) and ( [ eq : sat2 ] ) with @xmath23 = ca@xmath1(@xmath110 ) , have completely quantum - mechanical treatment , and the corresponding partial rate coefficients are determined in the same way as in @xcite and @xcite , where the whole procedure was described in details . since the transition dipole moment of the system h + h@xmath1 unlimitedly increases ( with increasing of @xmath109 ) , the partial rate coefficient for the free - free process ( [ eq : sim2 ] ) is determined here semi - classically , as it was described in @xcite and @xcite .
then , in the case of the non - symmetric processes ( [ eq : sat1 ] ) - ( [ eq : sat2 ] ) with @xmath23 = ca@xmath111 the radiative transition ca@xmath112 ca@xmath111 is allowed by the dipole selection rules and , consequently , quasi - molecular band generated by these processes represent a satellite of the corresponding ion spectral line . in this case
the transition dipole moment approaches ( when @xmath109 increases ) to _ constant _ @xmath113 . because of that the rate coefficient for this free - free process can be determined by now only semi - classically .
however , in this case it is needed to use the semi - classical procedure which is described in @xcite in connection with the ion - atom processes in helium - rich stellar atmospheres .
let us note that the spectral rate coefficients of the considered non - symmetric absorption processes , including the processes with @xmath5 = li , are presented here in the appendix b. the electron - atom absorption processes ( [ eq : eat1 ] ) - ( [ eq : eat2 ] ) are described in this work similarly to @xcite and @xcite .
because of that the photo - detachment cross - section @xmath100 and the inverse `` bremsstrahlung '' rate coefficient @xmath98 introduced in eqs .
( [ eq : kapaea12 ] ) - ( [ eq : kapaea1 ] ) , which cause the efficiencies of these processes , are determined here using the data presented in @xcite and @xcite . according to eq .
( [ eq : kapaphi ] ) the efficiencies of the processes ( [ eq : phion ] ) of the considered metal atoms photo - ionization depend on the densities @xmath106;h)$ ] of these atoms and on the mean photo - ionization cross - sections @xmath108}(\lambda;t)$ ] .
these cross - sections for the atoms of all relevant metals ( na , ca , mg , si , al , fe ) are calculated here in the wide region of @xmath11 and @xmath59 on the basis of the data from @xcite . the densities @xmath106;h)$ ] are determined from saha s equation , with given @xmath59 , @xmath102 and the ion densities @xmath114 . for that purpose
is needed to know the corresponding partition functions @xmath115}(t;\delta_{x})$ ] , where @xmath116 is the lowering of the ionization potential of the given ground state atom @xmath5 .
the calculation of the partition functions , which is the key element of the whole procedure of @xmath106;h)$ ] determination , is performed here on the basis of the data from @xcite .
however , because of this source of the data about the partition functions it is needed to stay for a moment on this point .
the reason for this is the fact that the partition functions are calculated in @xcite for the values of @xmath117 from the region which is limited from below with 0.1 ev , while in our case ( @xmath118 k ) we have that @xmath119 ev . as it is known in the saha s equation @xmath120 appears also in the argument of its exponent , but in the considered case @xmath121 and the influence of @xmath120 on the values of that exponent can be completely neglected . however , the fact that existing @xmath122 0.1 ev has to be examined from the aspect of its influence on the partition functions .
namely , the moving of the value of @xmath120 from 0.1 ev to about 10@xmath123 ev causes the increasing of the principal quantum number of the last realized excited atomic state from about 11 to about 200 .
it was shown that such increasing of the largest principal quantum number causes the increasing of the values of the partition functions only for about 0.5 percentage .
this result means that the partition functions calculated by using the data from @xcite , which relate to @xmath124 0.1 ev , were applicable for determination of the densities @xmath106;h)$ ] .
the sunspot model m from @xcite , as well as its differences with the respect to the quiet sun referent model , also from @xcite , are illustrated by fig .
[ fig : abund ] .
so , the part of this figure ( `` a '' ) presents the densities of the hydrogen ions h@xmath1 and h@xmath29 and relevant metal ions ( na@xmath1 , ca@xmath1 , etc . ) as the functions of @xmath43 .
the part ( `` b '' ) presents the local temperature and the densities of the free electrons and hydrogen atoms ( @xmath59 , @xmath107 and @xmath101 ) from the model m together with the corresponding quantities ( @xmath125 , @xmath126 @xmath127 ) from the referent model , also as the functions of @xmath43 . in accordance with the mentioned in the introduction
we should obtain firstly the picture which reflects the relative efficiencies within the sunspot of all symmetric and non - symmetric ion - atom processes ( [ eq : sim1 ] ) - ( [ eq : sim2 ] ) , ( [ eq : nonsim1 ] ) - ( [ eq : nonsim2 ] ) and ( [ eq : sat1 ] ) - ( [ eq : sat2 ] ) , as well as the electron - atom absorption processes ( [ eq : eat1 ] ) - ( [ eq : eat2 ] ) and metal atom photo - ionization processes ( [ eq : phion ] ) .
for that purpose the several plots of spectral absorption coefficients of all these processes were performed for the values of log @xmath128 between @xmath129 and @xmath130 in the region 70 nm @xmath131 800 nm .
we mean on the ion - atom coefficients @xmath76 , @xmath86 and @xmath132 , which are given by eqs .
( [ eq : kapasimtot ] ) - ( [ eq : kapaiatot ] ) , as well as on the electron - atom and photo - ionization coefficients , namely @xmath95 and @xmath103}(\lambda;h)$ ] , which are given by ( [ eq : kapaea ] ) and ( [ eq : kapaphi ] ) respectively .
it was established that in the short wave part of this spectral region , generally we have the domination of the photo - ionization processes ( [ eq : phion ] ) in respect to all other considered absorption processes .
so , for @xmath133 log @xmath134 the mentioned domination exists in the region @xmath135 250 nm , but with the increasing of the value of log@xmath128 this region decreases and for log @xmath136 the region of this domination is @xmath135 175 nm . in the same time
it was established that the electron - atom absorption processes ( [ eq : eat1 ] ) - ( [ eq : eat2 ] ) can be treated now as the referent processes not only in the case of the quiet sun photosphere , but also in the case of a sunspot . moreover , for any value of log @xmath128 just these processes dominate in the best part of the whole considered spectral region in respect to all other absorption processes , including the processes ( [ eq : phion ] ) .
all mentioned is illustrated by fig s .
[ fig : plot0 ] and [ fig : plot1 ] which show the plots of all considered absorption processes for log @xmath128 = 0.5 and 0.0 respectively .
let us note that the plots for log @xmath128 = 0.5 were needed since just in the region log @xmath137 0.0 the density of the h@xmath1 ions becomes very fast increasing . in fig s .
[ fig : plot0 ] and [ fig : plot1 ] the plot of each of all considered photo - ionization processes ( [ eq : phion ] ) is shown separately .
these plots are marked by designations of the corresponding atoms ( `` mg '' , `` si '' , etc . ) .
we draw attention that here was not planed to obtain the plot describing the total efficiency of all photo - ionization processes .
namely to obtain such a plot would be certainly useful in the context of the synthesis of the corresponding sunspot spectrum .
however , such task is out of the frame of the present work . the electron - atom processes ( [ eq : eat1 ] ) - ( [ eq : eat2 ] ) , which are treated sometimes as the h@xmath29-continuum , are represented here by their common plot which in these figures is marked by @xmath138 .
however , in accordance with the aims of this work , the examined ion - atom absorption processes are characterized by three corresponding plots , namely : one for the whole group `` 1 '' , i.e. the symmetric processes ( [ eq : sim1 ] ) and ( [ eq : sim2 ] ) , which are known also as the h@xmath3-continuum ; second for these first processes together with the non - symmetric processes of the group `` 3 '' , i.e. ( [ eq : nonsim1 ] ) - ( [ eq : nonsim2 ] ) ; third for all considered ion - atom processes together , including the non - symmetric processes of the group `` 3 '' , i.e. ( [ eq : sim1 ] ) - ( [ eq : sim2 ] ) .
these plots are marked here by `` @xmath139 '' , `` @xmath140 '' and `` @xmath141 '' respectively .
the presented ion - atom plots show that it was indeed necessary to take into account additional non - symmetric processes of the group `` 3 '' .
namely , one can see that the inclusion of these processes significantly increases the total efficiency of the non - symmetric ion - atom absorption processes .
the figures [ fig : plot0 ] and [ fig : plot1 ] demonstrate that the total efficiencies of all considered symmetric and non - symmetric ion - atom absorption processes in the region 110 nm @xmath7 230 nm is of the same order of the magnitude as the one of the electron - atom processes ( [ eq : eat1 ] ) - ( [ eq : eat2 ] ) together ( h@xmath29-continuum ) .
however , here is needed to emphasis that in the significant part of this region these total efficiencies are close , or the ion - atom efficiency is even larger than electron - atom one . from fig .
[ fig : abund ] one can see that the plasma temperature in sunspot ( @xmath59 ) is significantly smaller than in the case of the quiet sun ( @xmath125 ) in the larger part of that region of log @xmath128 or @xmath43 which is shown in this figure .
consequently , it was possible to expect that in this region the relative importance of the non - symmetric absorption processes ( in respect to the symmetric ones ) , at least for 110 nm @xmath7 230 nm , will be larger than in the case of the quiet sun . in order to test this assumption and to establish what is the contribution of
all considered non - symmetric processes to the total ion - atom efficiencies in far uv region of @xmath11 we have determined the adequately defined quantity @xmath142 , namely @xmath143 where @xmath144 , with @xmath44 = 2 and 3 , and @xmath88 are given by eqs .
[ eq : kapansim2 - 3 ] and [ eq : kapaiatot ] respectively .
the behavior of this quantity ( for several values of @xmath11 ) is illustrated by fig . [
fig : gqs ] in the interval of heights : -125 km @xmath145 200 km . from this figure
one can see that in far uv region the non - symmetric processes are dominant in respect to the symmetric ones for @xmath146 50 km , i.e. in the largest part of the region of @xmath43 shown in fig .
[ fig : abund ] , but for -75 km @xmath147 50 km the efficiencies of the symmetric and non - symmetric processes are close .
this fact shows that in the case of the sunspot ( as in the case of the quiet sun ) it is useful to treat the symmetric and all non - symmetric ion - atom absorption processes together .
finally , we will have to compare within the same part of the sunspot ( which is shown in fig .
[ fig : abund ] ) the total efficiencies of the ion - atom and referent electron - atom absorption processes ( [ eq : eat1 ] ) and ( [ eq : eat2 ] ) together ( h@xmath29-continuum ) in the considered part of the far uv region of @xmath11 .
already this figure gives the possibility to expect that in the sunspot case these total efficiencies could be close at least in a part of the mentioned spectral region .
it can be tested on the basis of the behavior of the quantity @xmath148 , defined by @xmath149 where @xmath95 is given by eq .
[ eq : kapaea ] .
the behavior of this quantity ( also for several values of @xmath11 from the far uv region ) is illustrated by fig .
[ fig : f ] in the same interval of heights : -125 km @xmath150 200 km .
this figure shows that : for any of the taken @xmath11 the total efficiency of all ion - atom absorption processes is close , or even larger than the efficiency of the referent electron - atom processes ( h@xmath29-continuum ) in the largest part of the mentioned interval of @xmath43 .
this means that in the case of the sunspot it is needed to consider the discussed ion - atom absorption process in the mentioned part of the far uv region ( 110 nm @xmath7 230 nm ) always together with the referent electron - atom processes ( h@xmath29-continuum ) .
consequently , the considered ion - atom processes should be also included _ ab initio _ in the corresponding models of sunspots of solar - type and near solar - type stars .
due to possible future investigations of these ion - atom processes , the needed spectral rate coefficients are determined and presented in appendix b. since the ion - atom processes of the type ( [ eq : nonsim1 ] ) - ( [ eq : nonsim2 ] ) with @xmath5 = li could be interesting in connection with the so called lithium stars ( mentioned in the introduction ) , the corresponding spectral rate coefficients are also presented in this appendix .
from the presented material it follows that the considered symmetric and non - symmetric ion - atom absorption processes influence on the opacity of sunspots in the considered spectral region ( 110 nm @xmath7 230 nm ) not less and in some parts even larger than the referent electron - atom processes .
the presented results show that further investigations of the non - symmetric ion - atom absorption processes promise that their efficiency could be increased considerably .
namely , the processes ( [ eq : nonsim1 ] ) - ( [ eq : nonsim2 ] ) with @xmath6 fe were not considered here because of the absence of the data about the needed characteristics of the corresponding molecular ion .
however the fe component , according to fig .
[ fig : abund ] , gives the significant contribution to the electron density .
also , some of the possible processes of the ( satellite ) type ( [ eq : sat1 ] ) - ( [ eq : sat2 ] ) could be very efficient .
it means that the inclusion in the consideration of all possible relevant non - symmetric ion - atom absorption processes would surely increase their total efficiency . because of that
, we take such inclusion as the task for the investigations in the nearest future . for this purpose , here are presented the spectral characteristics of the considered ion - atom absorption processes which can be used in some further applications .
finally , the presented results show that in far uv region the significance of the considered ion - atom absorption processes is sufficiently large that they should be included _
ab initio _ in the corresponding models of the sunspots .
apart of that , obtained results could be useful also in the case of different solar like atmospheres .
the authors wish to thank to profs .
obridko , a.a .
nusinov and n. s. polosukhina for the shown attention to this work . also , the authors are thankful to the ministry of education , science and technological development of the republic of serbia for the support of this work within the projects 176002 , iii4402 .
aymar , m. , & dulieu , o. 2012 , journal of physics b atomic molecular physics , 45 , 215103 drawin , h .- w . , & felenbok , p. 1965 , data for plasmas in local thermodynamic equilibrium , by drawin , hans - werner . ; felenbok , paul .
paris , gauthier - villars , 1965 .
fontenla , j. m. , avrett , e. , thuillier , g. , & harder , j. 2006 , apj , 639 , 441 , j. m. , curdt , w. , haberreiter , m. , harder , j. , & tian , h. 2009 , apj , 707 , 482 fontenla , j. m. , harder , j. , livingston , w. , snow , m. , & woods , t. 2011 , journal of geophysical research ( atmospheres ) , 116 , 20108 habli , h. , ghalla , h. , oujia , b. , & gada , f. x. 2011 , european physical journal d , 64 , 5 hack , m. , polosukhina , n. s. , malanushenko , v. p. , & castelli , f. 1997 , a & a , 319 , 637 heine v. 1970 , solid state physics 24 , 1 - 36 .
ignjatovi , l. m. , & mihajlov , a. a. 2005 , phys.rev.a , 72 , 022715 ignjatovi , l. m. , mihajlov , a. a. , & klyucharev , a. n. 2008 , journal of physics b atomic molecular physics , 41 , 025203 , l. m. , mihajlov , a. a. , sakan , n. m. , dimitrijevi , m. s. , & metropoulos , a. 2009 , mnras , 396 , 2201 , l. m. , mihajlov , a. a. , srekovi , v. a. , & dimitrijevi , m. s. 2014 , mnras , 439 ( 3 ) , 2342 maltby , p. , avrett , e. h. , carlsson , m. , kjeldseth - moe , o. , kurucz , r. l. , loeser , r. 1986 , apj , 306 , 284 mihajlov , a. a. , & dimitrijevi ' c , m. s. 1986 , a&a , 155 , 319 mihajlov , a. a. , dimitrijevi ' c , m. s. , & ignjatovi ' c , l. m. 1993 , a&a , 276 , 187 mihajlov , a. a. , dimitrijevi ' c , m. s. , ignjatovi ' c , l. m. , & djuri ' c , z. 1994 , a&as , 103 , 57 , a. a. , & ignjatovi , l. m. 1996 , diam dynamique des ions , des atomes et des molecules , contrib .
papers , p. 157 mihajlov , a. a. , ignjatovi , l. m. , sakan , n. m. , & dimitrijevi , m. s. 2007 , a&a , 437 , 1023 mihajlov , a. a. , ignjatovi , l. m. , srekovi , v. a. , dimitrijevi , m. s. & metropoulos , a. 2013 , mnras , 431 , 589 nguyen , j. h. v. , viteri , c. r. , hohenstein , e. g. , sherill , c. d. , brown , k. p. , & odom , b. 2011 , new journal of physics , 13 , 063023 penn , m. j. , & livingston , w. 2011 , iau symposium , 273 , 126 shavrina , a. v. , polosukhina , n. s. , zverko , j. , et al .
2001 , a & a , 372 , 571 shavrina , a. v. , polosukhina , n. s. , pavlenko , y. v. , et al.2003 , a & a , 409 , 707 , h. , beuc , r. , ban , t. , & pichler , g. 2002 , european physical journal d , 19 , 49 srekovi , v. a. , mihajlov , a. a. , ignjatovi , l. m. & dimitrijevi , m. s. 2013 , adv . space .
, doi:10.1016/j.asr.2013.11.017 .
stancil , p. c. , kirby , k. , sannigrahi , a. b. , buenker , r. j. , hirsch , g. , gu , j .- p .
1997 , apj , 486 , 574 stilley , j. l. , & callaway , j. 1970 , apj , 160 , 245 travis , l. d. , & matsushima , s. 1968 , apj , 154 , 689 vernazza , j. , avrett , e. , & loser , r. 1981 , apjs , 45 , 635 , d. , beuc , r. , miloevi ' c , s. , & pichler , g. 1998 , european physical journal d , 2 , 45 wishart , a. w. 1979 , mnras , 187 , 59
the potential curves of the ground and several low lying excited electronic states of the molecular ions hna@xmath1 and hli@xmath1 , as well as the corresponding transition dipole matrix elements , are calculated here by means of the method which was described in details in @xcite .
the calculated characteristics of these molecular ions are shown as the functions of the internuclear distance @xmath109 in fig s .
[ fig : napotdip ] and [ fig : lipotdip ] , where the zero of the energy is chosen in such a way that the potential energy @xmath151 of the ground electronic state is equal to zero at @xmath152 . in these figures
are given the characteristics not only of the ground and first excited electronic state with the energy @xmath153 , but also of the several excited states whose energies are larger then @xmath153 .
namely , we keep in mined that such excited electronic state of the ions hna@xmath1 and hli@xmath1 could be needed in some further applications .
the mentioned calculation method was developed for the molecular ions @xmath154 , @xmath155 and h@xmath156 , where @xmath157 and @xmath158 are the alkali metal atoms ( li , na , etc . ) . within this method
the wave functions of the adiabatic electronic states of the considered molecular ion are described in the single - electron approximation , under the condition which are analogous to the orthogonality conditions in the known pseudopotential method of @xcite . using this method , in @xcite
were successfully determined the potential curves and the dipole matrix elements of the ions na@xmath3 and li@xmath3 , and later in @xcite - of the ion lina@xmath1 .
let us note that the first version of this method was used in @xcite just in the connection with the radiative processes in ( h + li@xmath1)-collisions .
the spectral rate coefficients @xmath159 and @xmath160 for the non - symmetric processes ( [ eq : nonsim1 ] ) - ( [ eq : nonsim2 ] ) and ( [ eq : sat1 ] ) - ( [ eq : sat2 ] ) respectively , which can be used in some further applications , are presented here in the relevant regions of @xmath11 and @xmath59 . since the properties of the behavior of the similar spectral rate coefficients were already discussed ( with quite a lot of details ) in @xcite , here
they are shown in fig s .
[ fig : kmgsi ] - [ fig : klica ] without additional comments . | as a continuation of the previous investigations of the symmetric and strongly non - symmetric ion - atom absorption processes in the far uv region within the models of the quiet sun photosphere , these processes are studied here within a model of the sunspot . here
we mean the absorption processes in the h@xmath0+h@xmath1 and h@xmath2 collisions and the processes of the photo - dissociation of the h@xmath3 and h@xmath4 molecular ions , where @xmath5 is one of the metal atoms : @xmath6na , ca , mg , si and al . obtained results show that the influence of the considered ion - atom absorption processes on the opacity of sunspots in the considered spectral region ( 110 nm @xmath7 230 nm ) is not less and in some parts even larger than the influence of the referent electron - atom processes .
in such a way , it is shown that the considered ion - atom absorption processes should be included _
ab initio _ in the corresponding models of sunspots of solar - type and near solar - type stars .
apart of that , the spectral characteristics of the considered non - symmetric ion - atom absorption processes ( including here the case @xmath5 = li ) , which can be used in some further applications , have been determined and presented within this work .
[ firstpage ] stars : atmospheres sunspots : general radiative transfer atomic processes molecular processes |
the luminosity function ( lf ) of galaxies within clusters of galaxies is a key tool for understanding the role of the environment on galaxy formation and evolution .
the shape of the cluster lf as a function of the galaxy colors and morphologies , as well as a function of the cluster radius or local density , can provide strong observational constraints on the theories of galaxy formation . for example , springel et al .
( 2001 ) recently showed that semi analytical models of hierarchical structure formation could now explain both the shape of the composite cluster lf ( @xmath9band lf of trentham 1998 ) and the morphology radius relationship of whitmore et al .
( 1993 ) using just a simple prescription for the properties of galaxies in clusters based on their merger and cooling rates ( see also okamoto , nagashima 2001 ) . empirically , there is also growing evidence for a correlation between the shape of the cluster lf and the underlying cluster properties .
phillipps et al .
( 1998 ) and driver et al .
( 1998 ) show that more evolved clusters , based on either their density profile or the presence of a cd galaxy , have flatter faint - end slopes , which they attribute to the disruption of faint galaxies in the cores of such evolved systems [ see the earlier theoretical work on galaxy cannibalism by hausman & ostriker ( 1978 ) ] . in summary , the lf of galaxies in clusters as a function of both the galaxy and cluster properties is a powerful observational test for theories of galaxy formation and evolution .
the reader is referred to the seminal review by binggeli , sandage , and tammann ( 1988 ) , which is still relevant today . in this paper , we present an analysis of the composite cluster lf based on the commissioning data of the sloan digital sky survey ( sdss ; see gunn et al . 1998 ; york et al . 2000 ; stoughton et al .
this analysis has several key advantages over previous studies of the composite cluster lf , including accurate multi - color ccd photometry for all galaxies ( in optical passbands @xmath0 and @xmath1 ; fukugita et al .
1996 ) , large aerial coverage , thus enabling us to make a local correction for the projected field lf , and finally , the availability of several objectively - measured galaxy properties like morphology .
furthermore , we selected our clusters from the sdss cut and enhance ( ce ) cluster catalog of goto et al .
( 2002 ) , which has two major benefits over previous cluster samples used for lf studies .
first , the ce catalog was objectively constructed using the latest cluster - finding algorithms , and therefore has a well - determined selection function ( see goto et al .
secondly , ce has obtained an accurate photometric redshift for each cluster based on the observed color of the e / s0 ridge - line using the method of j. annis et al .
( in preparation ) .
the error on this cluster photometric redshift is only @xmath10 for @xmath11 clusters
( see figure .
14 of goto et al . 2002 ) and , as we show herein , is accurate enough to allow us to determine the composite lf for a large sample of ce clusters without the need for spectroscopic redshifts .
thus , our analysis of the composite cluster lf is based on one of the largest sample of clusters to date .
we present this work now to provide a low - redshift benchmark for on - going multi - color photometric studies of high redshift clusters of galaxies . with the advent of large - area ccd imagers on large telescopes
, the number of distant clusters with such data will increase rapidly over the next few years ; e.g. kodama et al . ( 2001 ) recently presented large - area multi - color ccd photometry for the distant cluster a 851 ( @xmath12 ) using suprime - cam on the subaru telescope .
gladders and yee .
( 2000 ) searched distant clusters over 100 deg@xmath13 of ccd data .
this paper is organized as follows : in section 2 we describe the methods used to construct the composite lf of ce clusters and show our results as a function of passband and morphology . in section 3
we test the robustness of the analysis , and in section 4 we summarize our work . throughout this paper , we use @xmath14 = 0.7 , @xmath15 = 0.3 and @xmath16 = 0.7 .
in this section , we outline the data used in this paper .
the photometric data used herein was taken from the sdss commissioning data , as discussed by york et al .
our analysis focuses on the 150 deg@xmath13 contiguous area made up from the overlap of sdss photometric runs 752 & 756 i.e. @xmath17r.a.@xmath17 and @xmath17dec.@xmath17 .
this is a subset of the sdss early data release , as discussed in stoughton et al .
( 2002 ) , and similar to the data used by scranton et al .
( 2002 ) for studying the angular clustering of sdss galaxies .
this photometric data reach 5@xmath18 detection limits for point sources of 22.3 , 23.3 , 23.1 , 22.3 and 20.8 mag in the @xmath0 and @xmath1 passbands , respectively ( for an airmass of 1.4 and @xmath19 seeing ) , and @xmath20 to stress the preliminary nature of the calibration . ] .
the photometric uniformity of the data across the whole area is less than 3% [ see hogg et al .
( 2001 ) and smith et al .
( 2002 ) for photometric calibration ] , while the star - galaxy separation is robust to @xmath2121.0 ( scranton et al .
this is significantly better than previous photographic surveys , which suffer from larger plate - to - plate photometric fluctuations and a lower dynamic range [ see lumsden et al .
1997 for the problems associated with photographic studies of the cluster composite lf ] .
for each galaxy , we used the model magnitude computed by the photo data analysis pipeline , which has been shown by lupton et al .
( 2001 ) and stoughton et al .
( 2002 ) to be the optimal magnitude for faint sdss galaxies .
it is also close to the total magnitude for the fainter sdss galaxies . for a full discussion of the photometric data , and the galaxy parameters derived from that data , we refer the reader to lupton et al .
( 2001 ) and stoughton et al .
( 2002 ) .
the clusters used herein were drawn from the large sample of ce clusters presented in goto et al .
( 2002 ) , which were selected over the same photometric runs of 752 & 756 .
we only selected the richer systems which were determined by the number of galaxies brighter than @xmath418th magnitude , ( @xmath22 ) .
the ce clusters used here satisfy the following conditions : \1 ) number of galaxies brighter than @xmath418th magnitude in the @xmath23 band ( @xmath22 ) @xmath24 20 , \2 ) [email protected] .
condition 1 was used to select richer systems .
@xmath22 is defined as the number of galaxies brighter than @xmath418th magnitude in the @xmath23 band after subtracting the background using the method described in section [ method ] to construct composite lfs .
galaxies within 0.75 mpc from a cluster center were used .
condition 1 was used to avoid letting small groups with only a few very bright galaxies dominate the composite lfs in the weighting scheme .
( the weighting scheme is explained in detail in section [ method ] . )
even though we used @xmath2620 as a criteria to select our clusters , we show in section [ discussion ] that our composite lfs were not affected by this richness criteria .
since the high redshift clusters ( @xmath270.3 ) are not imaged to the fainter galaxies , we restricted our clusters to be in the range [email protected] . in total , 204 clusters satisfy these criteria .
we discuss here the construction of the composite luminosity function of galaxies within the subsample of the ce clusters discussed above .
the first critical step in such an analysis is the subtracting the background and foreground contamination . ideally , one would wish to do this via spectroscopic observations , but since the ce cluster catalog contains @xmath282000 galaxies in the region used , it is not feasible to observe all clusters spectroscopically . therefore , we must make a statistical correction based on the expected contamination from projected field galaxies .
one of the main advantages of the sdss data is that such a correction can be estimated locally ( i.e. free from any galaxy number count variances due to the large - scale structure ) for each cluster since we possess all the photometric data , to the same depth and in the same filter set , well outside of the cluster .
indeed , such local background subtraction was thought to be ideal in previous works , but was not possible due to the small coverage of the sky . for the composite cluster lf
, we only used galaxies within 0.75 mpc of the cluster centroid .
this radius was determined empirically so as not to lose statistics by using a too small radius , and not to lose the contrast of clusters against the background by using a too large radius .
foreground and background contamination were corrected for using an annulus around each cluster with an inner radius of 1.5 mpc and an outer radius of 1.68 mpc .
these radii represent a compromise between having as large an aperture as possible to avoid removing legitimate cluster galaxies , while still providing an accurate estimate of the local projected field population . since the background / foreground galaxies are themselves highly clustered , it is important to obtain as local an estimate as possible .
the photometric redshift of each cluster was used to convert these metric apertures into angular apertures .
the center of each cluster was taken from the ce catalog , and estimated from the position of the peak in the enhanced density map of goto et al .
the cluster centroids were expected to be determined with an accuracy better than @xmath2840 arcsec through monte - carlo simulation .
when an annulus touches the boundary of the sdss data , we corrected for contamination using the number - magnitude relationship of the whole data set instead ( this only affected a few of the clusters used herein ) .
since each sample cluster has a different redshift , each cluster reaches the sdss apparent magnitude limit at different absolute magnitudes . also ,
because they have various richnesses , the number of galaxies in each cluster is different . to take these different degrees of completeness into account
, we followed the methodology of colless ( 1989 ) to construct the composite cluster lf .
the individual cluster lfs are weighted according to the cluster richness and the number of clusters which contribute to a given bin .
this is written as @xmath29 where @xmath30 is the number of galaxies in the @xmath31th bin of the composite lf , @xmath32 is the number in the @xmath31th bin of the @xmath33th cluster lf , @xmath34 is the normalization of the @xmath33th cluster lf , and was measured to be the field - corrected number of galaxies brighter than @xmath35 , @xmath36 is the number of clusters contributing the @xmath31th bin and , finally , @xmath37 .
the formal errors on the composite lf were computed using @xmath38^{1/2 } , \ ] ] where @xmath39 and @xmath40 are the errors on the @xmath31th bin for the composite and @xmath33th cluster , respectively . in this way we can take into account the different degrees of completeness . like other authors , we discarded the brightest cluster galaxy ( bcg ) within 0.75 mpc of the cluster centroid when constructing the composite lf , since such bcgs tend not to follow the cluster lf .
we only used sdss galaxies brighter than @xmath41=21.0 , since this is the limit of the sdss star
galaxy separation ( scranton et al . 2002 ; lupton et al . 2001 ) . this magnitude limit and
weighting scheme combined with our cosmology enabled us to dig lf down to @[email protected] .
when converting apparent to absolute magnitudes , we assumed a k - correction for the early - type galaxy given by fukugita et al .
( 1995 ) . in figure
[ fig : all.eps ] , we show the composite lf of the subset of ce clusters discussed above .
we present one composite lf for each of the five sdss passbands .
we also present in table [ tab:5color ] the best - fit parameters from a joint fit of a schechter function to these data . for a comparison
, we also show the field values as derived by blanton et al .
( 2001 ) ( corrected for @xmath14 = 0.7 ) . in figure
[ fig : all.eps ] , field lfs normalized to cluster lfs are shown by dotted lines .
as expected , the @xmath2 for our cluster lfs is significantly brighter ( by 1 1.5 magnitudes depending on the bands ) than those seen for the field lfs in all five bands .
furthermore , the faint end slopes ( @xmath3 ) of the cluster composite lfs are much flatter that those seen for the field lfs .
this is especially noticeable for the redder passbands ( @xmath33 and @xmath1 ) while the slope of the cluster lf systematically flattens from the @xmath43 passband to the @xmath1 passband .
these results are consistent with the hypothesis that the cluster lf has two distinct underlying populations i.e. the bright end of the lf is dominated by bright early types that follow a gaussian - like luminosity distribution , while the faint - end of the cluster lf is a steep power - law - like function dominated by star - forming ( bluer ) galaxies .
binggeli et al . (
1988 ) originally suggested this hypothesis , while the recent work of adami et al .
( 2000 ) , rakos et al .
( 2000 ) and dressler et al .
1999 supported this idea .
particularly , boyce et al .
( 2001 ) showed lf of abell 868 is made up of three different populations of galaxies ; luminous red and two fainter blue populations .
the idea is illustrated by the fact that the cluster lfs in the redder passbands , which are presumably dominated by the old stellar populations of the early types , have much brighter @xmath2 s and significantly shallower slopes than those measured in the bluer passbands .
those results can also be interpreted as showing that bright elliptical galaxies are more populated in dense regions , like inside of clusters . they are consistent with the morphology - density relation advocated by dressler et al .
( 1980 , 1997 ) .
one of the key aspects of the sdss photometric data is the opportunity to statistically study the distribution of galaxies as a function of their morphology . in this subsection
, we discuss the composite cluster lf as a function of morphology using three complementary methods for determining the morphological type of each galaxy .
these include : i ) the best - fit de vaucouleur or exponential model profile ; ii ) the inverse of concentration index and iii ) the @xmath8 color of the galaxies .
we present all three methods , since at present it is unclear which method is the most successful at separating the different morphological galaxy types .
also , each method suffers from different levels of contamination , and the differences in the methods can be used to gauge the possible systematic uncertainties in the morphological classifications .
we discuss the three methods used in detail below .
the first method we consider here is using the de vaucouleur and exponential model fits of the galaxy light profiles measured by sdss photometric pipeline ( @xmath44 r.h .
lupton et al . , in preparation ) to broadly separate galaxies into the late and early - type .
if the likelihood of a de vaucouleur model fit to the data is higher than the that of an exponential model fit , the galaxy is called a late - type , and vice versa . galaxies that have the same likelihoods for both model fits are discarded . in figure
[ fig : exp_dev.ps ] , we present the composite cluster lf of late - type and early - type galaxies ( as defined using the model fits above ) for all five sdss passbands . in table
[ tab : dev_exp ] , we present the best - fit schechter function parameters to these data , and show the fits in figure [ fig : exp_dev.ps ] .
the second method uses the inverse of the concentration index , which is defined as @xmath45 = @xmath46 , where @xmath47 is the radius that contains 50% of the petrosian flux and @xmath48 is the radius that contains 90% of the petrosian flux ( see r.h .
lupton et al . , in preparation ) .
both of these parameters are measured by the sdss photo analysis pipeline for each galaxy .
the concentration parameter used here ( @xmath45 ) is just the inverse of the commonly used concentration parameter , and thus early - type galaxies have a lower @xmath45 parameter than late - type galaxies .
the correlation of @xmath45 with visually - classified morphologies has been studied in detail by shimasaku et al .
( 2001 ) and strateva et al .
they found that galaxies with @xmath490.4 are regarded as early - type galaxies , while galaxies with @xmath500.4 are regarded as late - type galaxies .
therefore , in figure [ fig : cin.ps ] , we show the composite cluster lf of late - type and early - type galaxies as defined using this second method for all five sdss passbands . in table
[ tab:5color_cin ] , we present the best - fit schechter function parameters to these data .
the third method used herein for morphological classification was to use the observed @xmath8 color of the galaxy which has been proposed by strateva et al .
( 2001 ) . using the fact that k - correction for @xmath8 is almost constant
until @xmath1 = 0.4 , they showed that galaxies shows a clear bimodal distribution in their @xmath8 color and @xmath8 = 2.2 serves as a good classifier of morphology until @xmath27 0.4 by correlating @xmath8 classification with visual classifications .
therefore , we have classified galaxies with @xmath51 2.2 as early - type and galaxies with @xmath52 2.2 as late - type .
figure [ fig : ur.ps ] shows the composite cluster lf for both types of galaxies along with their best - fit schechter functions ( in all five passbands ) .
the best - fit schechter parameters are summarized in table [ tab:5color_ur ] .
as expected , there are noticeable differences in these three morphological classifications , as portrayed by the differences in their composite lfs ( see figures .
[ fig : exp_dev.ps ] , [ fig : cin.ps ] , and [ fig : ur.ps ] )
. however , it is worth stressing here the similarities between the methods .
for example , the faint end slope of the lf is always shallower for early - type galaxies than late - type , regardless of the passband and methodology .
also , the faint end slope for early - type galaxies decreases steadily toward the redder passbands , while the faint - end slope for the late - type galaxies is nearly always above @xmath53 and consistent ( or steeper ) than the field lf in most passbands .
these observations are again qualitatively in agreement with the hypothesis that the bright end of the cluster lf is dominated by bright , old early - types , while the faint - end of the cluster lf represents late - type galaxies maybe in greater numbers than the average field .
this model is in agreement with hierarchical models of structure formation and the model for the tidal disruption of dwarf galaxies by the dominant early types .
in this section , we discuss various tests which we have performed on our measurement and results . to test the robustness of our methods , we performed monte - carlo simulations which involved adding artificial clusters to the sdss data and computing their composite lf using the same algorithms and software as used on the real data . our model for the artificial clusters was constructed using the sdss data on abell 1577 ( at @xmath54 , richness@xmath55 ) .
we used the method described in goto et al .
( 2002 ) to make artificial clusters . the radial profile for the artificial clusters was taken to be a king profile ( ichikawa 1986 ) with a concentration index of 1.5 and a cut - off radius of 1.4 mpc , which is the size of abell 1577 ( struble , rood 1987 ) .
the color - magnitude distributions for the artificial clusters were set to be the observed , field - corrected , color - magnitude distributions of abell 1577 binned into 0.2 magnitude bins in both color and magnitude . from this model
, we then constructed artificial clusters as a function of the redshift and overall richness .
for redshift , we created clusters at @xmath1 = 0.2 , 0.3 , 0.4 and 0.5 , ensuring that we properly accounted for the cosmological effects ; i.e. the clusters became smaller , redder and dimmer with redshift .
we used the k - corrections for an early type spectrum . for richness
, we change the number of galaxies within each cluster randomly between 10 and 50 . for each redshift
, we created 100 clusters ( 400 clusters in total ) .
the galaxies within these artificial cluster were distributed randomly in accordance with the radial and color - magnitude distributions discussed above .
we made no attempt to simulate the density - morphology relation nor the luminosity segregation in clusters .
the artificial clusters were randomly distributed within the real sdss imaging data and we constructed a composite lf for these clusters using exactly the same software as for the real clusters . since the artificial clusters were all made from the same luminosity distribution , the composite lf should therefore look very similar in shape as the original input lf .
figure [ fig : tfake_1577_ab.ps ] shows the result of our monte - carlo simulations .
the histogram shows the original absolute magnitude distribution of abell 1577 after a field correction , while the symbols show the composite luminosity functions which we constructed as a function of the input redshift .
one of the most innovative parts of this analysis is the use of photometric redshifts to determine the composite luminosity function of clusters . as demonstrated in goto et al .
( 2002 ) , the accuracy of photometric redshift is excellent ( @xmath56 = @xmath50.015 for @xmath570.3 ) and this method will certainly be used in the near future as the number of clusters with photometric redshifts will increase rapidly , far quicker than the number of clusters with spectroscopically confirmed redshifts . to justify our use of photometric redshifts ,
since all previous composite cluster lf s used spectroscopic redshifts , we constructed a composite lf using only the clusters with spectroscopically confirmed redshifts .
we derived our spectroscopic redshift for ce clusters by matching the sdss spectroscopic galaxy data with our ce clusters .
this was achieved by searching the sdss spectroscopic galaxy sample for any galaxies within the ce cluster radius and within @xmath56 = @xmath50.01 of the photometric redshift of the cluster .
the radius used here was from goto et al .
( 2002 ) . if multiple galaxies satisfied this criteria , the closest spectroscopic redshift to the photometric redshift was adopted .
the number of clusters with spectroscopic redshifts was 75 out of 204 at the date of this writing .
the results of this test are shown in figure [ fig : all.eps ] ( in the bottom right - hand panel ) . also the parameters for the best - fit schechter functions
are given in table [ tab:5color ] . and referred to as @xmath41(spec ) .
we only performed this test for the @xmath23 passband . the slope and characteristic magnitude of the best - fit schechter function for the spectroscopically determined lf is in good agreement with that derived using photometric redshifts .
as can be seen in table [ tab:5color ] , both @xmath58 and the slope agree within the error .
this test shows that we can truly construct composite lfs using photometric redshift of clusters .
one key aspect of measuring the composite cluster lf is the choice of the cluster centroid . to test the effect of different cluster centroids on the composite lf
, we constructed a composite cluster lf using the position of brightest cluster galaxies ( bcgs ) as a centroid instead of the peak in the enhanced density map , as discussed in goto et al .
the bcgs have been determined to be the brightest galaxy among galaxies fainter than @xmath424th magnitude within 0.75 mpc .
galaxies brighter than @xmath424th magnitude are regarded as being foreground galaxies .
the mean offset between the bcg position and the centroid previously used is 1.02 arcmin .
table [ tab : bcg_center ] lists the parameters of the best - fit schechter functions to the five sdss passbands , which should be compared to the values obtained using the optical centroid given in table [ tab:5color ] . in all five bands ,
the characteristic magnitudes and slopes agree very well within the error .
this test shows that our composite lfs are not dependent on a center determination .
since we constructed composite lfs from 2-dimensional , projected sky image , subtraction of fore / background galaxies played an important role in this work .
we test here the effect of making a global background subtraction for all clusters instead of the local background subtraction discussed above .
we use the number - magnitude relation of all the galaxies in the entire 150 deg@xmath13 region as the global background .
table [ tab : global ] gives the best - fit schechter parameters of composite lfs constructed using global background subtraction . compared with table [ tab:5color ] , again
, every schechter parameter agrees very well within 1 @xmath18 .
although we use annuli around clusters to subtract the background to avoid the large - scale structure disturbing the measurement of composite lfs , this test shows that our composite lfs are not dependent on background subtraction .
valotto et al .
( 2001 ) showed that a statistical background subtraction can not re - produce composite lfs using a mock galaxy catalog constructed from a large @xmath59-body simulation .
our result , however , combined with the fact that we derived the same lf as input through a monte - carlos simulation ( in subsection [ sec : monte ] ) , supports that our composite lfs are not subject to background subtraction .
another aspect we were concerned about was our choice of cluster richness criteria . to test this
, we constructed composite lfs of different subsample with @xmath2620 and @xmath2640 , given in table [ tab : richness ] .
@xmath22 here is defined as the number of galaxies brighter than @xmath418th magnitude after subtracting the background in the way we construct composite lfs .
@xmath2620 was used to construct composite lfs , as mentioned in section [ method ] . in table
[ tab : richness ] , even though @xmath2 is slightly brighter and the slope is slightly steeper for the richer sample , they agree within 1 @xmath18 .
the steepening of the slopes can be interpreted as a bias in selecting richer systems using @xmath22 , i.e. clusters with steeper tails tend to have a larger value of @xmath22 .
this , however , confirms that our composite lfs are not dependent on the richness criteria , which we chose . as the final test of our composite cluster lf , we compare here our composite lfs with previous works .
first , we must be careful to match the different cosmologies used by various authors as well as the different photometric passbands .
to facilitate such a comparison , therefore , we present in table [ tab : previous ] the best - fit schechter function parameters for our composite lf , but calculated for each author s cosmology and passband using the color corrections of fukugita et al .
( 1995 ) and lumsden et al .
( 1992 ) . in the case of the three @xmath60 photographic surveys of colless ( 1989 ) , valotto et al .
( 1997 ) , and lumsden et al .
( 1997 ) , we find a significantly brighter @xmath2 than these studies as well as a much shallower slope .
we also tried to fit a schechter function using their @xmath3 value for the slope , but @xmath2 s become even brighter .
the fits are not good when fixed @xmath3 s are used .
lugger ( 1989 ) found @[email protected]@xmath50.13 and @[email protected]@xmath50.09 by re - analyzing nine clusters presented in lugger ( 1986 ) .
the slope is steeper and @xmath2 is slightly brighter than our results .
when we fix the slope with her value at @[email protected] , the two lfs agree well .
garilli et al .
( 1999 ) studied 65 abell and x - ray selected samples of galaxies in the magnitude range of -23.0@xmath62 - 17.5 and found that @[email protected]@xmath50.15 and @[email protected]@xmath50.07 ( in isophotal magnitudes ) .
this slope is steeper than ours .
a possible difference with ours is that they used the color condition to select cluster galaxies .
@xmath2 is in agreement with our results within the error .
we also tried to fit a schechter function with a fixed value of @xmath3 = 0.84 .
@xmath2 became brighter by 0.18 mag although the fit was poor .
paolillo et al .
( 2001 ) studied composite lf of 39 abell clusters using the digitized poss - ii plates .
they obtained @[email protected]@xmath50.16 in @xmath23 .
the slope is @[email protected]@xmath63 . although the slope differs significantly , @xmath2 agrees well compared with our composite lf .
yagi et al . ( 2002a , b ) observed 10 nearby clusters with their mosaic ccd camera to derive composite lf .
their best - fit schechter parameters are @[email protected]@xmath50.2 and @[email protected]@xmath50.05 in @xmath64 .
they also studied type - specific lf using exponential and @xmath65 profile fits to classify the galaxy types .
they derived @[email protected] and @[email protected] for exponential galaxies and @[email protected] and @[email protected] for @xmath65 galaxies .
considering that they derived composite lfs using the data taken with different instruments analyzed in a different way , it is reassuring that they reached the same conclusion as our results discussed in subsection [ sec : morph ] .
i.e. exponential galaxies have the steeper faint end tail than @xmath65 galaxies , while their @xmath2 are almost the same .
concerning the disagreement of our lfs with previous studies , various differences in measuring composite lfs may be the reason .
the possible sources of differences are different ways of weighting , different ways of background subtraction , and different depths of the luminosity function .
the sample clusters , themselves , should have , to some extent , different richness distributions .
for @xmath2 , although we tried to transform our magnitude into their magnitude , the color conversion between sdss bands and others might not be accurate enough .
thus , the difference with the previous studies is not necessarily a mistake in the analysis , but rather it represents a different way of analysis . throughout our analysis discussed in section [ method ] we carefully used exactly the same way to construct the composite lfs . we thus keep our composite lfs internally consistent .
we studied the composite lf of 204 the sdss ce galaxy clusters .
the over - all composite lf is compared with other composite lfs . comparing it to the field luminosity function
, a tendency of a brighter @xmath2 and a flatter slope is seen .
this is consistent with our understanding that cluster regions are dominated by brighter galaxies than field galaxies .
we divided the composite lf by galaxy morphology in three ways . in all three cases
, we found that early - type galaxies have flatter slopes than late - type galaxies .
these observations are in agreement with the hypothesis that the bright end of the cluster lf is dominated by bright , old early - types , while the faint - end of the cluster lf represents late - type galaxies .
this is also consistent with the morphology density relation originally advocated by dressler et al .
( 1980 , 1997 ) .
we also studied these composite lfs in five sdss color bands .
the slopes become flatter and flatter toward the redder color bands .
this again suggests that cluster regions are dominated by elliptical galaxies with old stellar population .
these composite lfs provide a good low redshift benchmark to study higher - redshift clusters in the future .
since the data used in this work came from 2% of the sdss data , further studies with large sdss data will increase the statistical significance on these topics as the sdss proceeds .
we would like to thank the referee , steven phillipps , for a detailed revision and useful recommendations provided for this work .
we are grateful to michael crouch and christopher j. miller for valuable comments , which contributed to improve the paper . t.g . wishes to thank robert c. nichol for hospitality during the stay at carnegie mellon university .
t. g. acknowledges financial support from the japan society for the promotion of science ( jsps ) through jsps research fellowships for young scientists .
the sloan digital sky survey ( sdss ) is a joint project of the university of chicago , fermilab , the institute for advanced study , the japan participation group , the johns hopkins university , the max - planck - institute for astronomy , new mexico state university , princeton university , the united states naval observatory , and the university of washington .
apache point observatory , site of the sdss telescopes , is operated by the astrophysical research consortium ( arc ) .
funding for the project has been provided by the alfred p. sloan foundation , the sdss member institutions , the national aeronautics and space administration , the national science foundation , the u. s. department of energy , japanese monbukagakusho , and the max planck society .
the sdss web site is @xmath66http://www.sdss.org/@xmath67 .
fukugita , m. , shimasaku , k. , & ichikawa , t. 1995 , , 107 , 945 fukugita , m. , ichikawa , t. , gunn , j. e. , doi , m. , shimasaku , k. , & schneider , d. p. 1996
, , 111 , 1748 .
garilli , b. , maccagni , d. , & andreon , s. 1999 , , 342 , 408 gladders , m. d. , & yee , h. k. c. 2000 , , 120 , 2148 goto , t. , sekiguchi , m. , nichol , r. c. , bahcall , n. a. , kim , r. s. j. , annis , j. , ivezi ' c , v zeljko , brinkmann , j. , et al .
2002 , , 123 , 1807 okamoto , t. , & nagashima , m. 2001 , , 547 , 109 paolillo , m. , andreon , s. , longo , g. , puddu , e. , gal .
, r. r. , scaramella r. , djorgovski , s. g. , de carvalho , r. 2001 , a&a , 367 , 59 phillipps , s. , driver , s. p. , couch , w. j. , & smith , r. m. 1998 , , 498 , l119 rakos , k. d. , schombert , j. m. , odell , a. p. , & steindling , s. 2000 , , 540 , 715 sandage , a. , binggeli , b. , & tammann , g. a. 1985 , , 90 , 1759 scranton , r. , johnston , d. , dodelson , s. , frieman , j. a. , connolly , a. , eisenstein , d. j. , gunn , j. e. , hui , l. , et al .
2002 , submitted to apj .
composite lf of galaxy clusters from the sdss ce galaxy cluster catalog in five sdss bands .
the solid line is the best - fit schechter functions .
the y - axis is arbitrary .
the dotted line is the field lfs from blanton et al .
( 2001 ) re - scaled to our cosmology .
the normalization of field lfs was adjusted to match the cluster best - fit schechter functions .
the lower - right panel is for clusters with spectroscopic redshifts in @xmath23 .
the best - fit schechter parameters are summarized in table [ tab:5color ] .
, title="fig : " ] composite lf of galaxy clusters from the sdss ce galaxy cluster catalog in five sdss bands .
the solid line is the best - fit schechter functions .
the y - axis is arbitrary .
the dotted line is the field lfs from blanton et al .
( 2001 ) re - scaled to our cosmology .
the normalization of field lfs was adjusted to match the cluster best - fit schechter functions .
the lower - right panel is for clusters with spectroscopic redshifts in @xmath23 .
the best - fit schechter parameters are summarized in table [ tab:5color ] .
, title="fig : " ] composite lf of galaxy clusters from the sdss ce galaxy cluster catalog in five sdss bands .
the solid line is the best - fit schechter functions .
the y - axis is arbitrary .
the dotted line is the field lfs from blanton et al .
( 2001 ) re - scaled to our cosmology .
the normalization of field lfs was adjusted to match the cluster best - fit schechter functions .
the lower - right panel is for clusters with spectroscopic redshifts in @xmath23 .
the best - fit schechter parameters are summarized in table [ tab:5color ] .
, title="fig : " ] composite lf of galaxy clusters from the sdss ce galaxy cluster catalog in five sdss bands .
the solid line is the best - fit schechter functions .
the y - axis is arbitrary .
the dotted line is the field lfs from blanton et al .
( 2001 ) re - scaled to our cosmology .
the normalization of field lfs was adjusted to match the cluster best - fit schechter functions .
the lower - right panel is for clusters with spectroscopic redshifts in @xmath23 .
the best - fit schechter parameters are summarized in table [ tab:5color ] .
, title="fig : " ] composite lf of galaxy clusters from the sdss ce galaxy cluster catalog in five sdss bands .
the solid line is the best - fit schechter functions .
the y - axis is arbitrary .
the dotted line is the field lfs from blanton et al .
( 2001 ) re - scaled to our cosmology .
the normalization of field lfs was adjusted to match the cluster best - fit schechter functions .
the lower - right panel is for clusters with spectroscopic redshifts in @xmath23 .
the best - fit schechter parameters are summarized in table [ tab:5color ] .
, title="fig : " ] composite lf of galaxy clusters from the sdss ce galaxy cluster catalog in five sdss bands .
the solid line is the best - fit schechter functions .
the y - axis is arbitrary .
the dotted line is the field lfs from blanton et al .
( 2001 ) re - scaled to our cosmology .
the normalization of field lfs was adjusted to match the cluster best - fit schechter functions .
the lower - right panel is for clusters with spectroscopic redshifts in @xmath23 .
the best - fit schechter parameters are summarized in table [ tab:5color ] .
, title="fig : " ] composite luminosity functions of de vaucouleur galaxies and exponential galaxies .
the galaxies are divided into two subsamples using profile fitting .
the lines show the best - fit schechter functions ( solid for de vaucouleur galaxies , dotted for exponential galaxies ) .
the y - axis is arbitrary
. de vaucouleur galaxies always have a brighter @xmath2 and a flatter faint end tail .
the best - fit schechter parameters are summarized in table [ tab : dev_exp ] .
, title="fig : " ] composite luminosity functions of de vaucouleur galaxies and exponential galaxies .
the galaxies are divided into two subsamples using profile fitting .
the lines show the best - fit schechter functions ( solid for de vaucouleur galaxies , dotted for exponential galaxies ) .
the y - axis is arbitrary
. de vaucouleur galaxies always have a brighter @xmath2 and a flatter faint end tail .
the best - fit schechter parameters are summarized in table [ tab : dev_exp ] .
, title="fig : " ] composite luminosity functions of de vaucouleur galaxies and exponential galaxies .
the galaxies are divided into two subsamples using profile fitting .
the lines show the best - fit schechter functions ( solid for de vaucouleur galaxies , dotted for exponential galaxies ) .
the y - axis is arbitrary
. de vaucouleur galaxies always have a brighter @xmath2 and a flatter faint end tail .
the best - fit schechter parameters are summarized in table [ tab : dev_exp ] .
, title="fig : " ] composite luminosity functions of de vaucouleur galaxies and exponential galaxies .
the galaxies are divided into two subsamples using profile fitting .
the lines show the best - fit schechter functions ( solid for de vaucouleur galaxies , dotted for exponential galaxies ) .
the y - axis is arbitrary
. de vaucouleur galaxies always have a brighter @xmath2 and a flatter faint end tail .
the best - fit schechter parameters are summarized in table [ tab : dev_exp ] .
, title="fig : " ] composite luminosity functions of de vaucouleur galaxies and exponential galaxies .
the galaxies are divided into two subsamples using profile fitting .
the lines show the best - fit schechter functions ( solid for de vaucouleur galaxies , dotted for exponential galaxies ) .
the y - axis is arbitrary
. de vaucouleur galaxies always have a brighter @xmath2 and a flatter faint end tail .
the best - fit schechter parameters are summarized in table [ tab : dev_exp ] .
, title="fig : " ] composite luminosity functions of high - concentration and low - concentration galaxies .
the concentration index ( @xmath45 ) used here is the ratio of the 50% petrosian flux radius to the 90% petrosian flux radius . in this figure ,
early - type galaxies have @xmath490.4 , and late - type galaxies have @xmath500.4 .
early - type galaxies have flatter faint end tails in all five bands .
lines are the best - fit schechter functions ( solid for @xmath490.4 , dotted for @xmath500.4 ) .
the y - axis is arbitrary .
the best - fit schechter parameters are summarized in table [ tab:5color_cin ] .
, title="fig : " ] composite luminosity functions of high - concentration and low - concentration galaxies .
the concentration index ( @xmath45 ) used here is the ratio of the 50% petrosian flux radius to the 90% petrosian flux radius . in this figure ,
early - type galaxies have @xmath490.4 , and late - type galaxies have @xmath500.4 .
early - type galaxies have flatter faint end tails in all five bands .
lines are the best - fit schechter functions ( solid for @xmath490.4 , dotted for @xmath500.4 ) .
the y - axis is arbitrary .
the best - fit schechter parameters are summarized in table [ tab:5color_cin ] .
, title="fig : " ] composite luminosity functions of high - concentration and low - concentration galaxies .
the concentration index ( @xmath45 ) used here is the ratio of the 50% petrosian flux radius to the 90% petrosian flux radius . in this figure ,
early - type galaxies have @xmath490.4 , and late - type galaxies have @xmath500.4 .
early - type galaxies have flatter faint end tails in all five bands .
lines are the best - fit schechter functions ( solid for @xmath490.4 , dotted for @xmath500.4 ) .
the y - axis is arbitrary .
the best - fit schechter parameters are summarized in table [ tab:5color_cin ] .
, title="fig : " ] composite luminosity functions of high - concentration and low - concentration galaxies .
the concentration index ( @xmath45 ) used here is the ratio of the 50% petrosian flux radius to the 90% petrosian flux radius . in this figure , early - type galaxies have @xmath490.4 , and late - type galaxies have @xmath500.4 .
early - type galaxies have flatter faint end tails in all five bands .
lines are the best - fit schechter functions ( solid for @xmath490.4 , dotted for @xmath500.4 ) .
the y - axis is arbitrary .
the best - fit schechter parameters are summarized in table [ tab:5color_cin ] .
, title="fig : " ] composite luminosity functions of high - concentration and low - concentration galaxies .
the concentration index ( @xmath45 ) used here is the ratio of the 50% petrosian flux radius to the 90% petrosian flux radius . in this figure ,
early - type galaxies have @xmath490.4 , and late - type galaxies have @xmath500.4 .
early - type galaxies have flatter faint end tails in all five bands .
lines are the best - fit schechter functions ( solid for @xmath490.4 , dotted for @xmath500.4 ) .
the y - axis is arbitrary .
the best - fit schechter parameters are summarized in table [ tab:5color_cin ] .
, title="fig : " ] composite luminosity functions of @xmath51 2.2 ( late - type ) and @xmath52 2.2 ( early - type ) galaxies .
early - type galaxies have flatter faint end tails in all five bands .
the lines are the best - fit schechter functions ( solid for @xmath51 2.2 , dotted for @xmath52 2.2 ) .
the y - axis is arbitrary .
the best - fit schechter parameters are summarized in table [ tab:5color_ur ] .
, title="fig : " ] composite luminosity functions of @xmath51 2.2 ( late - type ) and @xmath52 2.2 ( early - type ) galaxies .
early - type galaxies have flatter faint end tails in all five bands .
the lines are the best - fit schechter functions ( solid for @xmath51 2.2 , dotted for @xmath52 2.2 ) .
the y - axis is arbitrary .
the best - fit schechter parameters are summarized in table [ tab:5color_ur ] .
, title="fig : " ] composite luminosity functions of @xmath51 2.2 ( late - type ) and @xmath52 2.2 ( early - type ) galaxies .
early - type galaxies have flatter faint end tails in all five bands .
the lines are the best - fit schechter functions ( solid for @xmath51 2.2 , dotted for @xmath52 2.2 ) .
the y - axis is arbitrary .
the best - fit schechter parameters are summarized in table [ tab:5color_ur ] .
, title="fig : " ] composite luminosity functions of @xmath51 2.2 ( late - type ) and @xmath52 2.2 ( early - type ) galaxies .
early - type galaxies have flatter faint end tails in all five bands .
the lines are the best - fit schechter functions ( solid for @xmath51 2.2 , dotted for @xmath52 2.2 ) .
the y - axis is arbitrary .
the best - fit schechter parameters are summarized in table [ tab:5color_ur ] . ,
title="fig : " ] composite luminosity functions of @xmath51 2.2 ( late - type ) and @xmath52 2.2 ( early - type ) galaxies .
early - type galaxies have flatter faint end tails in all five bands .
the lines are the best - fit schechter functions ( solid for @xmath51 2.2 , dotted for @xmath52 2.2 ) .
the y - axis is arbitrary .
the best - fit schechter parameters are summarized in table [ tab:5color_ur ] .
, title="fig : " ] results of a monte - carlo simulation to test the robustness of the weighting scheme .
the histogram shows the luminosity function of model cluster a 1577 .
the circles , triangles , squares and pentagons represent the composite luminosity function at each redshift ( @xmath1 = 0.2 , 0.3 .
0.4 and 0.5 , respectively ) constructed with 100 fake clusters at each redshift .
the hexagonals show the composite luminosity function from all 400 fake clusters distributed on the real sdss data . ] .
best - fit schechter parameters of the composite luminosity function of five sdss bands .
the field values are from blanton et al .
( 2001 ) , whose paramters were shifted to match our cosmology .
galaxies within 0.75 mpc are used . [ cols="<,<,<,<,<",options="header " , ] | we present here results on the composite luminosity functions ( lf ) of galaxies in the clusters of galaxies selected from the cut and enhance cluster catalog ( ce ) of the sloan digital sky survey .
we constructed composite lfs in the five sdss bands , @xmath0 and @xmath1 , using 204 ce clusters ranging from @xmath1 = 0.02 to @xmath1 = 0.25 .
background and foreground galaxies were subtracted from the lf using an annular region around clusters to take large - scale , galaxy - number - count variances into consideration .
a lf of each cluster was weighted according to the richness and number of contributing galaxies to construct the composite lf .
taking advantage of accurate photometry of sdss , we used photometric redshifts to construct composite luminosity functions and thus study a large number of clusters .
the robustness of the weighting scheme was tested using a monte - carlo simulation .
the best - fit schechter parameters are ( @xmath2,@xmath3)=(@[email protected] , @[email protected]),(@[email protected] , @[email protected]),(@[email protected] , @[email protected] ) , ( @[email protected] , @[email protected] ) and ( @[email protected] , @[email protected] ) in @xmath6 and @xmath1 , respectively . we find that the slope of composite lfs becomes flatter toward a redder color band . compared with the field lfs of the sdss , the cluster lfs have brighter characteristic magnitude and flatter slopes in the @xmath7 and @xmath1 bands .
these results are consistent with the hypothesis that the cluster lf has two distinct underlying populations i.e. the bright end of the lf is dominated by bright early types that follow a gaussian - like luminosity distribution , while the faint - end of the cluster lf is a steep power - law like function dominated by star - forming ( bluer ) galaxies .
we also studied the composite lfs for early - type and late - type galaxies using profile fits , a concentration parameter and @xmath8 color to classify galaxy morphology .
a strong dependence of lf on galaxy morphology was found .
the faint end slope of the lf is always flatter for early - type galaxies than late - type , regardless of the passband and methodology .
these results are consistent with the hypothesis that the cluster regions are dominated by bright elliptical galaxies .
this work also provides a good low - redshift benchmark for on - going multi - color photometric studies of high redshift clusters of galaxies using 48 m class telescopes . |
population iii ( pop iii ) stars formed from metal - free gas clouds where h@xmath5 molecule is the main coolant of gas are considered to be more massive than 10@xmath6 @xmath7 ( bromm & larson 2004 and references therein ) , though there are still numerous uncertainties regarding the mass range of the first stars due to a poor understanding of the relevant radiative feedback in the late accretion phase ( see for example tan & mckee 2004 ) . once the primordial gas is enriched by dust grains , dust
have great influences on the subsequent formation and evolution history of stars and galaxies ; the star formation rate ( sfr ) in the metal - poor star - forming clouds is enhanced via the formation of h@xmath5 molecules on the surface of dust grain ( hirashita & ferrara 2002 ; cazaux & spaans 2004 ) .
in addition , the cooling of gas by thermal radiation from dust themselves makes even the gas clouds with metallicity as low as @xmath8 @xmath9 fragment into low - mass gas clumps ( schneider et al .
2002 , 2003 ; omukai et al .
thus , dust grains in the early universe may cause the characteristic mass scale of stars to shift from very high mass to low mass of @xmath10 @xmath7 observed at present time and strongly affect the evolution of the initial mass function ( imf ) at the early epoch of the universe . also , dust grains in the early universe play crucial roles in revealing the structure and evolution of the universe from the relevant observations because they absorb stellar light and reemit it at infrared ( ir)submillimeter ( submm ) wavelengths .
the thermal emission from dust at high redshifts can distort the cosmic microwave background ( cmb ) radiation ( loeb & haiman 1997 ) , and the obscuration and reddening of starlight by dust residing in the early interstellar and/or intergalactic space lead to the serious underestimate in evaluating the cosmic sfr from the observations directed toward the higher redshifts ( e.g. , hauser & dwek 2001 ) . in fact , the recent observations have confirmed the presence of large amounts of dust grains at high redshifts of @xmath11 .
the submm ( priddey et al . 2003 ; robson et al . 2004 ) and millimeter ( bertoldi et al . 2003 ) observations reported continuum thermal emission of dust grains from some quasars at @xmath12 , and have suggested a large amount of cold dust reaching up to @xmath13@xmath14 @xmath7 in the host galaxies of quasars .
maiolino et al .
( 2004a ) discovered the substantial extinction of stellar light by dust in the broad absorption line ( bal ) quasars at @xmath15 .
they have shown that their extinction curves could be different from those of low - redshift quasars at @xmath16 , and have suggested that the difference may reflect different formation and evolution mechanisms of dust grains at @xmath11 .
the main sources of dust grains in the early universe at @xmath12 are considered to be supernovae ( sne ) evolving from the massive stars because of their short lifetime .
dust grains formed in the ejecta of sne are injected into the interstellar medium ( ism ) , and are subjected to destructive processes in the interstellar shocks induced by the ambient sne ( e.g. , jones et al . 1994 ; dwek , foster , & vancura 1996 ) .
thus , the size distribution and amount of dust in the early interstellar space are determined by the balance between their production in sne and their destruction in shocks , and change with time according to the star formation activity ( e.g. , dwek 1998 ) .
how much amount of dust grains are destructed in shocks as well as how much dust grains absorb stellar light and reemit it by thermal emission heavily depends on their chemical composition , size and amount .
thus , it is essential to clarify the evolution of dust in the early universe by treating the formation and destruction processes of dust grains in a consistent manner , in order to elucidate the sfr and the imf during the early evolution of the universe from the future observations .
nozawa et al . ( 2003 ) investigated the chemical composition , size distribution and amount of dust grains formed in the ejecta of pop iii sne . in this paper , as the second step to reveal the evolution of dust grains in the early universe , we explore dust destruction in the high - velocity interstellar shock driven by the sn explosion , adopting the dust model by nozawa et al .
( 2003 ) for the initial dust residing in the early ism .
we focus on dust destruction in nonradiative shock where dust grains are considered to be efficiently destructed by non - thermal and thermal sputtering ( e.g. , mckee 1989 ) because the nonradiative shock has a high shock velocity ( @xmath17 km s@xmath18 ) and a high gas temperature ( @xmath19 k ) .
however , only a few set of experimental data on sputtering yield are available for the projectile - target combinations of astrophysical interest .
hence , we first evaluate the sputtering yields for the combinations of dust species produced in pop iii sne and ion species of interest to us , collecting a large amount of sputtering data and applying the universal relation for sputtering yield proposed by bohdansky ( 1984 ) with a slight modification so as to reproduce well the experimental data .
then , we investigate the dependence of the efficiency of dust destruction for each grain species not only on the gas density @xmath0 in the ism but also on the explosion energy @xmath2 and progenitor mass @xmath1 of sn driving the interstellar shocks by applying the hydrodynamical models of pop iii sne by umeda & nomoto ( 2002 ) . finally , we evaluate the time - scale of destruction for each grain species . the efficiency of dust destruction by sputtering depends on the temperature and density of gas and the velocity of dust relative to gas in shock as well as the sputtering yield and the dust size .
thus , in the calculation , we solve the time evolution of temperature and density of gas in spherically symmetric shocks , including the cooling of gas by the atomic process , the inverse compton scattering and the thermal emission of dust . in coupled with the hydrodynamical calculations for gas
, we carefully treat the erosion of dust by thermal and non - thermal sputtering caused by motion of dust relative to gas , taking into account the size distribution of each dust species .
this paper is organized as follows : we first define the efficiency of dust destruction in 2 . in 3 , we describe the properties of the ism in the early universe together with the dust model adopted in the calculations , and in 4 , we present the basic equations of the hydrodynamic calculations for gas and the method of the simulation for dust destruction in shock
. the sputtering yield of each grain species necessary for the calculations of dust destruction is evaluated in 5 , and the grain physics in shock is described in 6 .
the results of calculations of dust destruction are given in details for the model of interstellar shock that is driven by a sn with @xmath20 and @xmath21 @xmath22 and is propagating into the ism with @xmath23 @xmath24 , and the approximation formula for the time - scale of dust destruction as a function of @xmath2 and @xmath0 is derived in 7 , where the effects of the cooling of gas on the evolution of nonradiative shock and the destruction of dust are briefly discussed .
the summary is presented in 8 .
our main goal is to reveal the time - scale of dust destruction by interstellar shocks in the early universe . in general ,
the time - scale of destruction @xmath25 for grain species @xmath26 is defined as ( e.g. , mckee 1989 ) @xmath27 where @xmath28 is the efficiency of destruction of grain species @xmath26 , @xmath29 is the mass of gas swept up by shock until the shock velocity @xmath30 decelerates below 100 km s@xmath18 , @xmath31 is the mass of gas and dust in the ism , and @xmath32 is the effective sn rate . mckee ( 1989 ) has estimated @xmath28 to be nearly constant and considered that @xmath29 is proportional to the sn explosion energy @xmath33 in units of @xmath34 erg .
however , it is expected that both @xmath28 and @xmath29 depend on not only @xmath33 but also the progenitor mass @xmath1 of sn and the number density of gas @xmath35 in the ism . hence , we aim at deriving the dependences of @xmath28 and @xmath36 on @xmath33 , @xmath1 and @xmath35 .
the truncation time @xmath37 being defined as the time at which @xmath30 drops down to 100 km s@xmath18 , the efficiency of destruction @xmath28 for each dust species is calculated by @xmath38 where @xmath39 and @xmath40 are , respectively , the mass of dust destructed and the mass of dust swept up by shock , and the mass of gas @xmath29 swept up by spherically symmetric shock is given by @xmath41 with the gas density @xmath42 in the ism and the travel distance of the shock front @xmath43 .
we assume that the ism in the early universe is homogeneous and stationary ( @xmath44 ) , referring to the results of the simulations on the formation of the early h ii region by the first stars ( kitayama et al .
2004 ) , which have shown that the radiative feedback from the massive stars causes the ambient ism to be homogenized with the mean gas density of @xmath45 1 @xmath24 . in the calculations , we consider the hydrogen number density in the ism to be in the range of 0.03 @xmath24 @xmath46 @xmath24 to reveal the dependence of the efficiency of dust destruction on the gas density in the ism .
the ism surrounding the sn explosion is considered to be ionized by the radiation of the massive progenitor stars and heated to by @xmath47 k ( kitayama et al .
2004 ; machida et al . 2005 ) , and then @xmath48 is expected to decrease below @xmath49 k where the recombination rate can be high even for the metallicity as low as @xmath50 @xmath4 ( wolfire et al .
1995 ) . however , the evolution of gas temperature in postshock flow is almost independent of @xmath48 until the end of the sedov taylor phase of high - velocity shock considered in this paper , and the temperature of gas @xmath51 at the shock front does not decrease below several of @xmath49 k at the truncation time , as is shown in the result of calculations ( see 7.1 ) .
thus , we take @xmath52 k as the representative gas temperature in the ism , independently of the density of gas . the truncation time as well as
the time evolution of gas temperature and density structures in postshock flow is considered to be affected by the cooling of gas depending on the metallicity of the ism .
especially , when @xmath53 k , the efficient line coolings of h and he ions cause the shock velocity to decelerate rapidly , and resultingly regulate the truncation time .
however , the gas cooling function is almost independent of the metallicity for @xmath3 @xmath9 corresponding to [ fe / h ] @xmath54 since the metal lines do not contribute to the cooling of gas at @xmath55 k ( sutherland & dopita 1993 ) , and we adopt the gas cooling function for the zero metal case given in sutherland & dopita ( 1993 ) ( see 4.1 ) .
note that the efficiency of dust destruction , which is defined by the mass ratio of dust destructed to dust swept up by shock , does not depend on the metallicity as long as @xmath3 @xmath9 , despite the fact that the amount of dust in the ism is proportional to the metallicity , and in the calculations we assume that the metallicity of the ism in the early universe is @xmath50 @xmath9 to investigate the destruction of dust in the ism .
the elemental composition of gas in the ism is given in table 1 where the c , n , and o metals are included so as to adjust the metallicity to @xmath50 @xmath9 and their abundances are simply evaluated by multiplying the metallicity and the primordial ratios tabulated in sutherland & dopita ( 1993 ) ; note that these metals play no role in the physical processes relevant to dust destruction because of the much lower abundances than h and he .
the electron number density @xmath56 is fixed to 1.128 @xmath57 given by sutherland & dopita ( 1993 ) .
dust grains in the ism at the early epoch of the universe considered in this paper are expected to be predominantly produced in the ejecta of sne . to reveal what kind of dust grains form in the early universe , todini & ferrara ( 2001 ) calculated dust formation in primordial core - collapse supernovae ( ccsne ) , assuming the uniform elemental composition and gas density within the he core .
they have shown that the newly formed grains are amorphous carbon with size around 300 @xmath58 , and al@xmath5o@xmath59 , mgsio@xmath59 , mg@xmath5sio@xmath60 and fe@xmath59o@xmath60 whose sizes are around 1020 @xmath58 . by extending the model of dust formation by todini & ferrara ( 2001 ) to pair - instability supernovae ( pisne )
whose progenitors evolve from metal - free stars with @xmath1= 140260 @xmath7 ( umeda & nomoto 2002 ; heger & woosley 2002 ) , schneider , ferrara , & salvattera ( 2004 ) found that the typical sizes of dust formed in pisne range from 0.001 to 0.3 @xmath61 m depending on grain species .
the dust models by todini & ferrara ( 2001 ) can successfully reproduce the extinction curve observed toward a bal quasar sdss1048 + 46 at @xmath62 ( maiolino et al .
2004b ) and the ir spectral energy distribution ( sed ) of blue compact dwarf galaxy sbs 0335052 ( takeuchi et al . 2003 ) except for the slight overestimate of the far ir continuum ; sbs 0335052 is a good target in the local universe suitable for scrutinizing the properties of dust formed in sne at high redshifts because of its young age ( @xmath63 yr ) and low metallicity ( 1/41 @xmath9 ) ( vanzi et al .
2000 ) .
however , as discussed by kozasa , hasegawa , & nomoto ( 1989 ) , the species of dust formed in the ejecta largely depend on the elemental composition in the he core , and their sizes are affected by the time evolution of gas density and temperature .
thus , nozawa et al . ( 2003 ) extensively investigated dust production in pop iii ccsne and pisne , by considering two extreme cases for the elemental composition in the ejecta ; the unmixed case with the original onion - like structure and the uniformly mixed case within the he core . in the calculations of dust formation
, the radial profile of gas density is properly treated and the time evolution of gas temperature is calculated by solving the radiative transfer equation including the energy deposition of radioactive elements , which are not taken into account by todini & ferrara ( 2001 ) .
nozawa et al . ( 2003 ) found that a variety of grain species ( c , si , fe , fes , al@xmath5o@xmath59 , mgsio@xmath59 , mg@xmath5sio@xmath60 , sio@xmath5 and mgo ) condense in the unmixed ejecta reflecting the difference of the elemental composition at the formation site , and that only oxide and silicate grains ( al@xmath5o@xmath59 , mgsio@xmath59 , mg@xmath5sio@xmath60 , sio@xmath5 and fe@xmath59o@xmath60 ) form in the mixed ejecta with c / o @xmath64 . the average sizes of these grains span the range of 0.0011 @xmath61 m depending on the elemental composition and gas density at the location of formation in the ejecta .
the species of dust formed in pop iii sne and the behaviors of their size distributions are almost independent of the progenitor mass , but the relative mass fraction of each grain species is dependent on . based on the dust models by nozawa et al .
( 2003 ) , hirashita et al .
( 2005 ) have shown that the extinction curve of a quasar sdss1048 + 46 at @xmath62 can be better explained by the dust grains produced in the unmixed pop iii ccsne than those in the mixed pop iii ccsne .
takeuchi et al .
( 2005 ) have also shown by using the dust models for @xmath65 @xmath22 that only dust grains formed in the unmixed ejecta but not in the mixed ejecta can give an excellent fit to the observed sed of sbs 0335052 over near to far ir wavelengths .
these works strongly suggest that the comparisons with the observations relevant to dust in the early universe prefer dust grains produced in the unmixed pop iii ccsne rather than those in the mixed pop iii ccsne .
therefore , as the representative model of the initial dust grains residing in the early ism , we adopt the size distribution and abundance of each grain species obtained by the calculations of dust formation in the unmixed ejecta of pop iii sn with @xmath65 @xmath22 by nozawa et al .
( 2003 ) . for comparison
, we also consider the model of dust formed in the mixed ejecta of the same sn .
hereafter , the models of dust formed in the unmixed and mixed ejecta are referred to as the unmixed grain model and the mixed grain model , respectively .
the initial values of the depletion factor defined by the ratio of the total dust mass to the total metal mass produced in the sn is 0.224 for the unmixed grain model and 0.297 for the mixed grain model . the bulk density @xmath66 , mass fraction @xmath67 and
average size @xmath68 of each grain species are summarized in table 2 , and their size spectrums are given in figure 6a for the unmixed grain model and in figure 7a for the mixed grain model , where the bin width is 0.1 dex .
it is useful for the following discussion to address here the behavior of size distribution of each dust species produced in the unmixed ejecta .
c , fe and sio@xmath5 grains have a lognormal - like size distribution with the relatively large average radius of @xmath69 @xmath61 m , while the size distribution functions of fes , mg@xmath5sio@xmath60 and mgo grains are approximately power - law . the average radii of al@xmath5o@xmath59 and mgsio@xmath59 grains are very small ( @xmath70 @xmath61 m ) .
note that si grain shows a bimodal size distribution ; the grains with the size less than 0.03 @xmath61 m occupy almost 100 % in number but only 0.2 % in mass .
thus , as the average radius of si grain , we take 0.25 @xmath61 m calculated for the grains larger than 0.03 @xmath61 m including 99.8 % in mass .
the size distribution function summed up over all grain species is well fitted with a broken power - law formula whose index is @xmath71 for the radius larger than 0.06 @xmath61 m and @xmath72 for the smaller one .
the deceleration rate due to drag force and the erosion rate due to sputtering of dust grains depend on the time evolution of gas temperature and density in postshock flow . in this section
, we present the basic equations and initial conditions for the hydrodynamical calculation of the time evolution of shock and the method of the simulation for dust destruction in shock .
the hydrodynamic equations for gas to calculate the time evolution of spherically symmetric interstellar shock driven by a sn explosion , using conventional physical variables , are given by @xmath73 @xmath74 @xmath75 = -\lambda(n_{\rm h},t),\end{aligned}\ ] ] where @xmath76 is the gas pressure with the mean molecular weight @xmath61 and the hydrogen mass @xmath77 , and the sum of kinetic energy and thermal energy per unit volume is @xmath78 with adiabatic index @xmath79 , and we adopt @xmath80 in the calculation .
the cooling rate of gas @xmath81 in units of erg s@xmath18 @xmath24 is expressed by @xmath82 where each term represents the rate of cooling by the atomic process , by the inverse compton scattering , and by the thermal emission of dust . in the calculations ,
we adopt the gas cooling function @xmath83 for the zero metal case given in sutherland & dopita ( 1993 ) and ignore the contribution of cooling by metal ions ejected from dust by sputtering , because as mentioned in 3.1 , the metals do not contribute to the cooling of gas for the metallicity of @xmath50 @xmath9 ( sutherland & dopita 1993 ) . for the inverse compton scattering caused by collisions of hot electrons in shock with cmb photons , the cooling rate @xmath84 is given by ( ikeuchi & ostriker 1986 ) @xmath85 where @xmath86 is the gas temperature in units of 10@xmath87 k , and we take @xmath88 in the calculation .
the cooling function @xmath89 through the thermal emission from dust heated by collisions with gas particles in shock is calculated by ( dwek 1987 ) @xmath90 where @xmath91 and @xmath92 are , respectively , the number density and size distribution function of dust species @xmath26 at a given @xmath93 and @xmath94 .
the collisional heating rate @xmath95 of grains species @xmath26 is given in 6.3 .
the time evolution of gas temperature and density in shock driven by sn is characterized by the explosion energy @xmath33 and the progenitor mass @xmath1 which specify the structures of density and velocity of gas in the ejecta .
thus , as the initial condition for interstellar shocks , we adopt the hydrodynamical models of the ejecta of pop iii sne with various @xmath33 and @xmath1 by umeda & nomoto ( 2002 ) to investigate the dependence of the efficiency of dust destruction on @xmath33 and @xmath1 .
note that in the early universe with @xmath50 @xmath4 , the progenitors of sne are not only pop iii stars but also massive stars of the subsequent generation .
however , the structures of density and velocity in the sn with a given @xmath33 and @xmath1 do not so much depend on the initial metallicity for the progenitor star whose metallicity is less than @xmath96 @xmath4 ( umeda , h. 2002 in private communication ) . in the calculations , we consider ordinary ccsne with the explosion energy of @xmath33=1 , hypernovae ( hne ) whose explosion energy is more than ten times that of ccsne , and pisne whose progenitor is very massive ( @xmath97 150 , 170 and 200 @xmath22 ) , and hereafter ccsne and hne with progenitor mass of @xmath97 1330 @xmath22 are referred to as type ii sne ( sne ii ) .
the ejecta models used in the calculations are given in table 3 , where the labels c , h and p represent ccsne , hne and pisne , respectively , and the numerical value denotes the mass of the progenitor in units of solar mass .
equations ( 4)(6 ) are solved by using the flux - splitting method with second - order accuracy in space and first - order accuracy in time ( van albada , van leer , & roberts 1982 ; mair et al .
this algorithm is one of the upwind schemes for the euler equations and is well suited to solve problems involving a shock .
the calculations for the time evolution of shock and dust destruction are performed via the following two steps . in the first step
, we assume that the sn ejecta collides with the ism in ten years after the explosion and calculate the time evolution of gas density and temperature without including dust , based on the ejecta models of umeda & nomoto ( 2002 ) .
the spatial mesh number is 5005000 depending on the gas density in the ism , and the inner 20 meshes are assigned to the ejecta .
note that dust destruction in this first step does not contributes to the evaluation of the destruction efficiency because the mass of gas swept up by shock is less than 0.1 % of that by @xmath98 .
also , the cooling of gas by dust has no effects on the subsequent structure of shock . in the second step , fixing the spatial mesh number to 430 , we rearrange the density , temperature and velocity of postshock gas obtained from the first step into the inner 30 linear meshes .
then , including dust , we calculate the time evolution of shock as well as dynamics and destruction of dust . by treating dust as a test particle ,
the physical processes relevant to dust are calculated as follows .
consider a stationary ( @xmath99 ) dust grain characterized by its composition and number density with radius between @xmath100 and @xmath101 at a given position @xmath102 in the ism .
once the grain enters into the shock front , the relative velocity to gas @xmath103 is assigned .
then , for the temperature @xmath104 and density @xmath57 of gas at the shock front , we calculate the deceleration rate @xmath105 due to drag force , the erosion rate @xmath106 by sputtering and the heating rate @xmath107 by collisions with gas . by updating the relative velocity by @xmath108 and the grain radius by @xmath109 for a given small time step @xmath110 , the velocity and position of the dust at time @xmath111 are calculated by @xmath112 and @xmath113 .
the procedure described above is repeated for the gas temperature @xmath104 and density @xmath57 at the position @xmath114 of the grain .
note that the dust grain is considered to be completely destroyed if the grain radius is smaller than the nominal monomer radius of condensate given in nozawa et al .
sputtering is divided into two categories reflecting the difference in the underlying mechanisms ; physical sputtering and chemical sputtering .
physical sputtering invokes a transfer of kinetic energy from the incident particle to target atoms and subsequent ejection of those atoms which have acquired kinetic energy enough to overcome the binding forces exerted by the target .
chemical sputtering invokes a chemical reaction induced by the impinging particles which produces an unstable compound at the target surface ( sigmund 1981 ) . in the high - velocity shocks considered in this paper , dust grains are predominantly destroyed by physical sputtering because dust temperature is at most 100 k in the postshock region ( dwek 1987 ) ; chemical sputtering is of no importance since this process is realized at the temperature as high as 800 k ( roth 1983 ) .
thus , in what follows , we simply refer to physical sputtering as `` sputtering '' .
the efficiency of destruction by sputtering is characterized by the sputtering yield defined as the mean number of emitted atoms per incident particle .
the sputtering yield depends on the impact energy of a projectile and the incident angle to the target surface as well as on the projectile - target combination . however , only a few set of experimental data on sputtering yield are available for the combinations of dust materials and ion species of interest to us .
one method evaluating the sputtering yields for the projectile - target combinations of astrophysical interest is to apply the monte - carlo code for the _ transport of ions in matter _
( trim , ziegler , biersack , & littmark 1985 ) that simulates the energy loss of a projectile ion with given impact energy and incident angle in a solid ( field et al .
1997 ; may et al .
2000 ; bianchi & ferrara 2005 ) .
the trim code requires as input parameters not only the surface binding energy of the target , but also the displacement energy @xmath115 and the lattice binding energy @xmath116 .
particularly , uncertainties in the value of @xmath115 are a significant source of error in the calculated sputtering yield ( may et al .
2000 ; bianchi & ferrara 2005 ) .
in fact , the sputtering yields computed by this simulation code have been known to significantly differ from the experimental data at low impact energies ( bianchi & ferrara 2005 ) .
furthermore , it is time - consuming to combine the monte - carlo simulation with the hydrodynamical calculation for dust destruction .
another one is to apply the universal relation which is the analytic formula derived by bohdansky ( 1984 ) for the energy dependence of sputtering yield at normal incidence for a monatomic solid . to investigate the destruction of interstellar dust such as c , fe , sic , sio@xmath5 and h@xmath5o ice , tielens et al .
( 1994 ) have applied the universal relation by treating a constant @xmath117 in this formula as a parameter , and determined it by comparing the available experimental data with the sputtering yield calculated by the universal relation .
they have shown that the universal relation can be applied to not only monatomic solids but also compound solids by taking the appropriate average values for the atomic and mass numbers and the surface binding energy under the assumption that each of element consisting of the target is sputtered off with the same rate . however , the agreement of the calculated sputtering yield with the experimental data is not so much good for the ion - target combinations with the mass ratio of target atom to incident ion ranging from 0.7 and 2 .
therefore , we apply the universal relation for sputtering yield with a slight modification so as to reproduce well the experimental data , collecting and examining minutely a large amount of sputtering data , in order to evaluate the sputtering yields for the combinations of dust and ion species of interest to us . following tielens
( 1994 ) , we treat @xmath117 as a free parameter and determine it by fitting to the experimental data for the grain species with the yield data covering a wide range of impact energy of the projectiles .
for the grain species for which little or no sputtering data are available , we deduce the value of @xmath117 by fitting to the sputtering yield calculated by means of the monte - carlo code for the _ erosion and deposition based on dynamic model _ ( eddy , hirooka 2001 ) .
the eddy code requires only the surface binding energy as an input parameter and allows us to predict the sputtering yield for specific impact energy and incident angle more easily than the trim code , though the energy range applicable to the calculations is limited to 0.110 kev . in the following subsection ,
we at first introduce the universal relation for sputtering yield , and then determine the value of @xmath117 for each grain species by fitting to the available experimental data and/or the results of the eddy simulation . for backward sputtering and with the neglect of inelastic energy losses , the sputtering yield at normal incidence @xmath118 by the projectile @xmath119 impacting with energy @xmath120
is given by ( bohdansky 1984 ) @xmath121 \left(1- \frac{e_{\rm th}}{e } \right)^2,\end{aligned}\ ] ] where @xmath122 is the surface binding energy in units of ev , @xmath123 is the ratio of the mass number of the target atom to the incident ion , @xmath124 is the energy - independent function of @xmath125 , and @xmath117 is considered to be a free parameter to be adjusted to reproduce the experimental data ( tielens et al .
the threshold energies @xmath126 have been obtained by fitting the yield data for low - energy sputtering and are approximately given by ( bohdansky et al .
1980 ; andersen & bay 1981 ) @xmath127 where @xmath128 is the maximum fractional energy transfer possible in a head - on elastic collision .
the function @xmath129 is the nuclear stopping cross section in units of ergs cm@xmath6 and can be expressed by the following universal relation ( sigmund 1981 ) @xmath131 where @xmath132 is the elementary charge , and @xmath133 and @xmath134 are the atomic numbers of the projectile and the target , respectively .
the screening length @xmath135 for the interaction potential between the nuclei is given by @xmath136 with the bohr radius @xmath137 .
the function @xmath138 depends on the detailed form adopted for the screened coulomb interaction and can be approximated by ( matsunami et al .
1980 ) @xmath139 where the reduced energy @xmath140 is given by @xmath141 the value of @xmath142 in equation ( 11 ) depends on how the distribution of energy deposited in the target is approximated .
sigmund ( 1969 ) suggested the three types of approximations according to the adopted distribution function ( gaussian , corrected gaussian , and non - gaussian ) .
the values of @xmath142 obtained by these approximations are very similar ( sigmund 1969 ) and is approximately expressed for @xmath143 as follows ( e.g. , bohdansky 1984 ) ; @xmath144 however , equation ( 17 ) for @xmath124 slightly overestimates the sputtering yields for the projectile - target combinations with the mass ratio of @xmath145 .
hence , from a comparison with a large amount of sputtering data , we apply the following formula for @xmath124 ; @xmath146 as we shall show below , this modified function @xmath124 can give an excellent agreement with the sputtering data for any values of @xmath125 considered here ( @xmath147 ) . by being @xmath117 as a parameter
, the surface binding energy @xmath122 must be specified to evaluate the normal - incidence sputtering yield @xmath148 from equation ( 11 ) .
we assume @xmath122 to be equal to the sublimation energy and evaluate it from the janaf thermochemical tables ( chase et al . 1985 ) except for c and al@xmath5o@xmath59 whose values of @xmath122
are discussed later .
the values of @xmath122 for each grain species are summarized in table 4 along with the values of @xmath134 and @xmath149 . in order to evaluate the sputtering yield for the projectile - target combinations of interest to us ,
we determine the value of @xmath117 by fitting to the experimental data for c , si , fe , sio@xmath5 and al@xmath5o@xmath59 grains for which a sufficient amount of sputtering data are available .
for mgo , fes and fe@xmath59o@xmath60 grains for which little or no yield data are available , we evaluate the value of @xmath117 by fitting to the results of the eddy simulation .
the experimental and simulated sputtering data at normal incidence , for which the references are summarized in table 4 , are given in figure 1 along with the best - fitting theoretical yield calculated by the universal relation ( solid lines ) . in figures 1a1e ,
the experimental data are shown for the following ion species ; h@xmath150 ( open circles ) , d@xmath150 ( asterisks ) , he@xmath150 ( open squares ) , ne@xmath150 ( crosses ) and ar@xmath150 ( open triangles ) . even in the case that the experimental data can be sufficiently available , for comparison , we also plot the results of the eddy calculations for the projectiles h@xmath150 ( filled circles ) , he@xmath150 ( filled squares ) and ar@xmath150 ( filled triangles ) . as for c ,
the sublimation energy estimated to be @xmath151 ev from the thermodynamical data can not reproduce the experimental data .
tielens et al .
( 1994 ) have shown that the ion fluences efficiently amorphize the surface layer of carbon material and reduce the surface binding energy .
thus , we take @xmath152 ev following tielens et al . (
1994 ) , and this value gives a good fit to the measured yields with @xmath153 ( fig .
1a ) being different from @xmath154 by tielens et al .
( 1994 ) ; the difference stems from the adopted formula for @xmath124 . the sputtering yield calculated by the eddy code
is enhanced for h@xmath150 by a couple of factor compared to the experimental data , but gives a good agreement for he@xmath150 .
the sputtering data for si and fe targets are sufficiently available .
the modified universal relation employing equation ( 18 ) for @xmath124 shows the excellent agreement with the data points for all projectiles considered here ( figs .
1b and 1c ) , by adopting @xmath155 and @xmath156 , respectively . the eddy calculations for si and fe targets also present a good agreement with their measured data .
the dotted lines indicate the sputtering yields by ar@xmath150 projectile calculated by using the values of @xmath117 derived above but adopting @xmath124 given by equation ( 17 ) .
it can be seen from these figures that equation ( 17 ) overestimates the sputtering yields for the projectile - target combinations with the mass ratio of @xmath145 as mentioned in 5.1 .
it should be emphasized that the previous formula can not reproduce the experimental data better than the modified universal relation whatever values of @xmath117 are selected .
a very good fit of the universal relation to the experimental data is obtained for sio@xmath5 target by taking @xmath157 ( fig .
this justifies the application of the universal relation to compound materials as shown by tielens et al .
( 1994 ) . the results by the eddy simulations are consistent with the experimental data , but a comparison with the theoretical curves of sputtering yield shows a little enhancement for h@xmath150 and ar@xmath150 projectile at @xmath158 ev . for mgsio@xmath59 and mg@xmath5sio@xmath60 materials for which the sputtering data can not be found in the literatures , as is the same as tielens et al .
( 1994 ) , we assume @xmath117 to be 0.1 , considering sio@xmath5 as a representative of silicates .
in fact , this could be reasonable since @xmath159 for the compounds including oxygen atoms ( see table 4 ) .
given the surface binding energy of @xmath160 ev evaluated from the thermodynamical data , the yield data of al@xmath5o@xmath59 can not be reproduced by the universal relation for any values of @xmath117 .
the disagreement of the theoretical prediction with the yield data is expected to be due to an unsuitable estimate of @xmath122 , as is the same as c. indeed , roth , bohdansky , & ottenberger ( 1979 ) have suggested @xmath161 ev for al@xmath5o@xmath59 from the behavior of sputtering yield at impact energies below the maximum yield by he@xmath150 . adopting this surface binding energy , we can get a better fit with @xmath162 ( fig .
1e ) , although the agreement of the universal relation with the yield data is the worst among grain materials considered in this paper . to our knowledge ,
little or no experimental data on sputtering yield exist for mgo , fes and fe@xmath59o@xmath60 .
thus , we employ the results by the eddy calculations to extract the values of @xmath117 for these materials . in the figures from 1f to 1h
( 1f for mgo , 1 g for fes , and 1h for fe@xmath59o@xmath60 ) , we show the eddy results and the best - fitting theoretical curves . the agreements are not always good for h@xmath150 and ar@xmath150 at low energies , which is also true for sio@xmath5 , but the universal relation well reproduces the yield data of sio@xmath5 .
thus , we consider that the fits are reasonable and satisfactory for these materials .
the values of @xmath117 determined are 0.06 , 0.18 and 0.15 for mgo , fes and fe@xmath59o@xmath60 , respectively .
table 4 summarizes the value of @xmath117 for each grain species evaluated by fitting to the available experimental data and/or the results of the eddy simulations .
we confirm that the derived value of @xmath117 can well reproduce the experimental data of compounds as well as single - element materials , by adopting appropriate average values for the atomic and mass numbers as proposed by tielens et al .
( 1994 ) . by slightly modifying the function @xmath142 in the universal relation
, we realized that the calculated sputtering yields well agree with the available experimental data than before .
the modified universal relation derived here could be applicable to evaluate the sputtering yield for any combinations of targets and projectiles of astrophysical interest as long as the experimental data are available .
the collisional interactions between dust and a hot gas in shock efficiently transfers charge , momentum and energy between the two phases . as a result
, dust grains acquire the electric charge , undergo the resistance to their motion through the gas , are eroded by sputtering , and are heated by collisions with gas . also , heated dust grains cool the gas through their thermal emission . electric charge on a grain plays an important role in grain physics in hot plasmas through the coulomb interaction and the lorentz force in the presence of magnetic field . here , we concentrate on the destruction of dust grains in fast ( @xmath163 km s@xmath18 ) and hot ( @xmath164 k ) nonradiative shocks . for @xmath165
k , the effect of the electric charge of dust grain can be neglected because the dimensionless potential parameter @xmath166 defined by the ratio of the electric potential acquired by dust to the kinetic energy of gas is approximated by @xmath167 and is significantly smaller than unity ( draine & salpeter 1979 ; mckee et al .
also , the erosion rate of refractory grains by thermal sputtering quickly decreases for @xmath168 k ( draine & salpeter 1979 ) .
correspondingly , the erosion rate of dust grans by non - thermal sputtering decreases quickly for the relative velocity of dust to gas @xmath169 km s@xmath18 .
therefore , it can be expected that the electric charge does not significantly affect the destruction of dust grains in nonradiative shocks considered in this paper .
also , it is assumed that the early ism is not pervaded by a magnetic field because the magnetic field in the early universe is considered to be very weak ( gnedine , ferrara , & zweibel 2000 ) .
thus , we formulate the basic equations describing the motion of dust grains and their erosion by sputtering in postshock flow , neglecting the effect of charge on dust grain . once dust grains being at rest in the preshock ism encounter the interstellar shock , they move ballistically behind the shock front with the initial velocity relative to gas @xmath170 where @xmath30 is the shock velocity , and are eroded by non - thermal sputtering . a grain streaming through the ionized gas is decelerated by a drag force due to the direct collisions with gas particles , and thus the velocity of dust relative to gas decreases .
the drag force acting on a grain results from the momentum transfer from gas to dust . under the assumption
that a gas particle reflects specularly from grain surface and that the kinetic temperature of gas is the same for all gas species , the deceleration rate of a spherical grain of radius @xmath100 with the velocity relative to gas @xmath171 is given by ( e.g. , draine & salpeter 1979 ) @xmath172 where @xmath57 is the number density of hydrogen atom , @xmath173 is the boltzmann constant , @xmath104 is the gas temperature , @xmath174 is the bulk density of grain and @xmath175 with the mass @xmath176 of gas species @xmath119 .
the summation is taken over all gas species , and @xmath177 is the number abundance of gas species @xmath119 relative to that of hydrogen atom . for the function @xmath178
whose exact formula is given by baines , williams , & asebiomo ( 1965 ) , the following analytical approximation has been proposed by draine & salpeter ( 1979 ) @xmath179 with the accuracy within 1 % for @xmath180 . in the calculations , we adopt this formula .
it can be seen from equation ( 19 ) , grains with smaller size and/or higher relative velocity can be effectively decelerated by drag force and come to comove with gas in shock .
the destruction of dust grains in high - velocity and high - temperature nonradiative shock is dominated by sputtering .
actually , we can neglect the thermal evaporation of dust , because the temperature of dust can not be raised as high as their sublimation temperature ( @xmath181 10002000 k ) for the preshock gas density considered in this paper ( e.g. , dwek 1987 ) .
furthermore , we neglect the shattering process and partial vaporization caused by grain - grain collisions , since the number density of dust particles is very small in postshock flow and collisions between dust grains are extremely rare without a magnetic field .
the erosion rate of grain by sputtering is calculated by taking the angle - averaged sputtering yield @xmath182 ( draine & salpeter 1979 ) and is given by ( e.g. , dwek et al .
1996 ) @xmath183 where @xmath184 and @xmath185 is the average mass of the sputtered atoms .
equation ( 21 ) can be reduced to the following equations for the two extreme cases of @xmath186 ( dwek & arendt 1992 ; tielens et al .
1994 ) ; one is for @xmath187 where a stationary grain suffers the thermal sputtering caused by collisions originating from thermal motion of gas , and the erosion rate is written as @xmath188 the other is in the limit of @xmath189 where the non - thermal sputtering erodes a hypersonic grain with the rate given by @xmath190 figure 2 shows the erosion rate of each dust species by sputtering in the gas with the primordial elemental composition of metallicity @xmath191 @xmath4 given in table 1 ; figure 2a for thermal sputtering versus the gas temperature @xmath104 from equation ( 22 ) and figure 2b for non - thermal sputtering versus the velocity of dust relative to gas @xmath171 from equation ( 23 ) .
the erosion rate by thermal ( non - thermal ) sputtering steeply increases from @xmath192 k ( @xmath193 km s@xmath18 ) , reaches a peak at @xmath194 420 @xmath195 k ( @xmath1961300 km s@xmath18 ) , and then slowly decreases with increasing @xmath104 ( @xmath171 ) .
this behavior of the erosion rate as a function of @xmath104 and @xmath171 reflects the dependence of sputtering yield on impact energy ( see 5.2 ) . among dust species considered in this paper ,
c grain has the lowest erosion rate at @xmath197 k ( @xmath198 km s@xmath18 ) , which is about one order of magnitude lower than that of fes grain with the highest rate at @xmath199 k ( @xmath200 km s@xmath18 ) . for thermal sputtering , the erosion rate @xmath106 for the other dust species
is @xmath201 @xmath61 m yr@xmath18 @xmath202 within a factor of three at @xmath197 k. for non - thermal sputtering at @xmath203 km s@xmath18 , the destruction by he@xmath150 is dominant , and the erosion rate of al@xmath5o@xmath59 is the lowest . the collisions with gas can also heat dust grains in postshock flow . the collisional heating rate @xmath107 of dust grain with radius @xmath100 is presented by dwek & arendt ( 1992 ) as @xmath204 where @xmath205 is the fraction of kinetic energy of gas species @xmath119 deposited into the dust grain . the values of @xmath205 are calculated by comparing an effective grain thickness @xmath206 ( @xmath207 for a spherical grain ) with the stopping range of the incident particle @xmath208 . when @xmath209 , the incident particle deposits almost all of the energy into dust grains , otherwise , the difference between @xmath206 and @xmath208 is exploited to determine the fraction of deposited energy .
the approximations for the stopping range of ion species considered here are given in dwek ( 1987 ) .
for the electron stopping range , we adopt the approximation derived by tabata , ito , & okabe ( 1972 ) and iskef , cunningham , & watt ( 1983 ) .
this electron stopping range reduces the heating rate by @xmath181 30 % at @xmath210 k @xmath211 k compared with that given by dwek ( 1987 ) .
the values and detailed derivation of @xmath212 will be presented elsewhere ( nozawa et al . in preparation ) .
in this section , we show the results of numerical simulations for dust destruction in nonradiative shock , focusing on the results calculated for the ism - dust specified by the unmixed grain model described in 3 .
in 7.1 , we demonstrate the time evolution of temperature and density of gas and the destruction of dust grains in the interstellar shock that is driven by the sn model of c20 with @xmath20 and @xmath213 @xmath22 and is propagating into the ism with @xmath23 @xmath24 . in 7.2
, we investigate the dependences of the destruction efficiency and the mass of gas swept up by shock on @xmath33 , @xmath1 and @xmath35 , and derive the approximation formula for the time - scale of destruction for each dust species in the early universe .
the effects of the cooling processes of gas on the evolution of nonradiative shock and the destruction of dust are briefly discussed in 7.3 .
figure 3 shows the structures of density ( upper panel ) and temperature of gas ( lower panel ) in nonradiative shock at given times , and the solid lines in figure 4 give the time evolution of the shock velocity @xmath30 and the gas temperature @xmath51 at the shock front .
as shown in figure 3 , the gas density increases to 4 times that in the ism and the gas temperature steeply raises up at the shock front which is indicated by the arrow for a given time step . with the initial shock velocity @xmath214 6000 km s@xmath18 , the gas temperature at the shock front remains above 10@xmath215 k until @xmath216 years , and both @xmath30 and @xmath51 decrease with time . at @xmath217 years after the explosion
, the density of gas in the inner postshock region is about one order of magnitude lower than that in the unshocked region , and the low - density hot bubble with temperature of several times 10@xmath218 k is formed ( fig .
3 ) . in this model ,
@xmath30 decelerates below 100 km s@xmath18 at @xmath219 years when the shock front travels the distance of @xmath220 pc and @xmath51 drops down to several 10@xmath87 k. after then , a dense thin - shell forms at the shock front because of the effective line cooling , and the snr enters into radiative phase .
the motion of dust grains ( upper panel ) and the evolution of their sizes ( lower panel ) in postshock flow as a function of time are depicted in figure 5 , as an example for c grains with the initial size of 0.01 @xmath61 m ( dashed ) , 0.1 @xmath61 m ( dotted ) , and 1 @xmath61 m ( solid ) .
the thick solid curve in the upper panel indicates the position of the shock front .
as dust grains initially at rest in the ism intrude into the blast wave , they depict the different trajectories depending on their initial sizes , which clearly demonstrates that the dust grains are segregated and subjected to the different sputtering processes in the postshock flow ; small grains with initial radius of 0.01 @xmath61 m are quickly decelerated by drag force , are trapped into gas near the shock front , and are completely destroyed by thermal sputtering .
the 0.1 @xmath61m - sized grains are gradually decelerated to comove with gas in @xmath221 years after entering into the shock .
large - sized grains with radius of 1 @xmath61 m are not so much suffered from deceleration by drag force and continue to keep the high velocity relative to gas . as a result
, they are subjected to non - thermal sputtering , but their sizes are reduced very little , partly because they stay in the inner region of postshock flow where the gas density is lower than that near the shock front ( see fig .
the modification of size distribution of each dust species due to destruction by sputtering is illustrated in figure 6 ; figure 6a for the initial size distribution before destruction and figure 6b for the size distribution after destruction .
since the erosion rate by sputtering does not strongly depend on grain size ( see 6.2 ) , small grains are predominantly destroyed regardless of grain species ; the number of small - sized grain such as al@xmath5o@xmath59 is greatly reduced .
c , sio@xmath5 and fe grains whose initial size distributions are lognormal with the relatively large average size are eroded but not completely destructed , and increase the numbers of smaller ones .
grains with the size larger than 0.1 @xmath61 m are little affected by erosion .
note that the size distribution summed up over all grains species gets flatter for the smaller radius with time compared with the corresponding initial size distribution approximated by a power - law formula with index of @xmath72 , while that for the larger radius remains almost unchanged .
the destruction efficiency @xmath28 of each grain species is given in table 5 ( see unmixed grain model ) along with the initial average radius .
it is expected that the efficiency of dust destruction are higher for grain species with the smaller average size because the smaller grains are predominantly destructed by sputtering .
in fact , the efficiency of destruction of al@xmath5o@xmath59 grain with the smallest average radius is 0.667 and is the highest among dust species considered here .
si and fe grains with initial average radii larger than 0.1 @xmath61 m have the destruction efficiencies less than 0.2 ; 0.13 for si grain and 0.15 for fe grain .
however , the destruction efficiency of c grain with the smaller average radius is smaller than that of sio@xmath5 grain because of the lowest erosion rate among all grain species at @xmath222 k ( see fig .
2 ) . also , fes grain has higher destruction efficiency ( 0.578 ) than those of mg@xmath5sio@xmath60 ( 0.451 ) and mgo grains ( 0.505 ) despite of the larger initial average radius , which reflects not only the higher erosion rate at @xmath223 k but also the mass distribution much more weighted toward the smaller grains .
likewise , the destruction efficiency of mgsio@xmath59 grain with average size comparable to that of mg@xmath5sio@xmath60 and mgo is high ( 0.637 ) because of the lack of large grains .
these facts indicate that the efficiency of dust destruction depends on the initial size distribution of dust grains as well as on the sputtering yield . in order to clarify the effect of the initial size distribution on the efficiency of dust destruction , we present the results of dust destruction for the mixed grain model calculated with the same sn model and value of @xmath0 as those for the unmixed grain model .
figure 7 shows the size distribution of each grain species ( al@xmath5o@xmath59 , mgsio@xmath59 , mg@xmath5sio@xmath60 , sio@xmath5 and fe@xmath59o@xmath60 ) before destruction ( fig .
7a ) and after destruction ( fig .
7b ) , and the destruction efficiencies of these grains are tabulated in table 5 ( see mixed grain model ) . as is the same as the unmixed grain model ,
the numbers of al@xmath5o@xmath59 and fe@xmath59o@xmath60 grains with the small average size ( @xmath224 a few tens ) are considerably reduced , which results in high efficiencies of destruction ( @xmath225 ) . for mgsio@xmath59 , mg@xmath5sio@xmath60 and sio@xmath5 grains with lognormal size distributions
, the erosion of large grains leads to the increase of the number of smaller ones , and their destruction efficiencies span the range of @xmath226 .
note that although the average size is twice larger than that in the unmixed case , the destruction efficiency ( 0.59 ) of mg@xmath5sio@xmath60 grain in the mixed case is significantly larger than that ( 0.45 ) in the unmixed case .
the reason is that mg@xmath5sio@xmath60 grains formed in the unmixed ejecta have power - law - like size distribution and include the large - sized grains of @xmath227 @xmath61 m almost undestructed in shock ; the average size is not always suitable for assessing the feasibility of dust destruction .
thus , we conclude that the dust destruction efficiency is very sensitive to the initial size distribution .
the mass fraction of dust destructed reaches up to 34 % for the unmixed grain model and 50 % for the mixed grain model . in this subsection
, we investigate the dependences of the efficiency of dust destruction @xmath28 and the mass of gas swept up by shock @xmath29 on @xmath33 , @xmath1 and @xmath35 , and derive an analytic formula describing the time - scale of dust destruction for the unmixed grain model .
figure 8 shows the destruction efficiency of each grain species versus sn explosion energy calculated for @xmath228 @xmath24 ; figure 8a is for al@xmath5o@xmath59 , fes , mg@xmath5sio@xmath60 and fe grains , and figure 8b is for mgsio@xmath59 , mgo , sio@xmath5 , c and si grains .
also , the overall efficiency of dust destruction , which is defined as the ratio of the total mass of dust destructed to the total mass of dust swept up by shock , is plotted in figure 8a .
the sn models used for the calculations are distinguished by open circles ( ccsne ) , open squares ( hne ) and filled triangles ( pisne ) .
the destruction efficiencies for each grain species are almost the same for ccsne and a pisn with the explosion energy of @xmath229 irrespective of the progenitor mass @xmath1 , and increase with increasing @xmath33 .
note that high explosion energy with @xmath230 causes the temperature of gas in postshock flow to raise up as high as @xmath14 k , but this does not directly influence on the efficiency of dust destruction because the increase of gas temperature does not always lead to the enhancement of the erosion rate by sputtering ( see fig .
the reason for the increased efficiency with increasing @xmath33 is considered as follows .
the high - velocity shock ( the initial shock velocity @xmath231 km s@xmath18 ) generated by the energetic sn explosion induces high velocity of dust relative to gas . then
, dust grains are efficiently decelerated by drag force ( see equation ( 19 ) ) , are trapped in the high - density region near the shock front , and are significantly eroded by thermal sputtering .
furthermore , since the high shock velocity takes much longer time to drop down to 100 km s@xmath18 ( see fig .
9 ) , dust grains are immersed in a hot plasma for a long time , which causes even larger grains to be more eroded .
it should be pointed out here that the efficiencies of destruction for the models of p170 with @xmath232 and p200 with @xmath233 are a little higher than that for h30 with @xmath234 .
the reason is considered that the ejecta mass of pisne is more than 6 times larger than that of hne , and the resulting longer duration of the free expansion phase causes the longer truncation time than hne ( fig . 9 ) . on the other hand , the efficiency of destruction for p150 with @xmath235 @xmath22 and @xmath236 is almost the same as that for ccsne with @xmath236 because the initial shock velocity ( @xmath181 3000 km s@xmath18 ) much lower than that of ccsne makes the truncation time comparable to ccsne . therefore , although the destruction efficiency @xmath28 is almost independent of @xmath1 as long as @xmath229 , @xmath28 for pisne higher than that for hne at @xmath230 reflects the difference in the explosion mechanism depending on the progenitor mass . to examine the dependence of @xmath28 on @xmath33 for each type of sn
, we calculate the coefficients @xmath237 and @xmath238 in equation ( a1 ) given in appendix for sne ii ( c13 , c20 , c25 , c30 , h25 and h30 ) and pisne ( p150 , p170 and p200 ) separately , and the calculated overall efficiencies of dust destruction are indicated by dotted ( sne ii ) and dashed lines ( pisne ) in figure 8a . although the difference in the efficiency of dust destruction between sne ii and pisne increases with increasing @xmath33 , the deviations from the values calculated by @xmath237 and @xmath238 for all sn models are at most about 10 % at @xmath239 .
thus , we consider that @xmath28 is almost independent of @xmath240 as long as @xmath2 is the same . for reference , we tabulate the values of @xmath237 and @xmath238 for sne ii and pisne in table 6 .
the mass of gas swept up by shock @xmath29 and the truncation time @xmath37 calculated for @xmath228 @xmath24 are presented in figure 9 as a function of sn explosion energy . as is the same as @xmath28 , @xmath29 and @xmath37
are almost the same for @xmath229 regardless of @xmath1 , and increase with increasing @xmath33 .
next , we show the dependences of @xmath28 and @xmath29 on the preshock gas density @xmath35 .
figure 10a plots the overall efficiency of dust destruction versus @xmath35 for c20 ( crosses ) , h25 ( open circles ) and h30 models ( filled triangles ) with @xmath241 and 30 , respectively . the efficiency of dust destruction increases with increasing @xmath242 , since higher gas density causes more frequently collisions between dust and gas to erode efficiently the surface of dust grains by sputtering ; for example , the mass fraction of dust destroyed for h30 model is 78 % for @xmath243 @xmath24 , but only 23 % for @xmath244 @xmath24 . in figure 10b , we present the mass of gas swept up by shock @xmath29 as a function of @xmath245 for c20 , h25 and h30 models .
note that @xmath29 decreases with increasing @xmath35 , because shock wave more quickly decelerates and travels only small distance .
the approximation formulae presented in appendix being combined , the time - scale of destruction for each dust species by the interstellar shock in the early universe is presented by @xmath246 for all sn models , where the dependences of @xmath28 on @xmath33 and @xmath245 are given by equations ( a1 ) and ( a3 ) , respectively .
we derived the time - scale of dust destruction as a function of not only the explosion energy of sn but also the gas density in the ism .
the swept - up gas mass is proportional not to @xmath33 but to @xmath247 , being different from the formula proposed by mckee ( 1989 ) .
the time - scale of dust destruction derived here could be applicable to investigate the time evolution of dust grains in the early universe .
especially , the difference of the efficiency of dust destruction for each grain species may have great influences on the amount and size distribution of dust grains residing in the early interstellar space .
figure 11 shows the cumulative energy lost by the atomic process ( dashed ) , the inverse compton scattering ( dotted ) and the thermal emission from dust ( solid ) calculated for the sn model of c20 and for the ism with parameters of @xmath248 and @xmath50 @xmath4 at the redshift of @xmath88 . the total energy lost by these cooling processes
is drawn by the thick solid curve .
the inverse compton cooling is comparable to that by the atomic process in the early phase of the snr ( @xmath249 years ) .
as the gas temperature decreases , the atomic process by h and he line coolings becomes dominant .
compared with the above two cooling processes , the thermal emission from dust is extremely low and contributes only less than 0.1 % to the total energy loss . the transition of nonradiative shock to radiative shock occurs when the cumulative energy loss reaches to 0.01 % of the explosion energy .
thus , only the h and he line coolings affect the time evolution of nonradiative shock and the dust destruction efficiencies in the ism with metallicity less than @xmath50 @xmath4 corresponding to the dust - to - gas mass ratio of @xmath250 in this paper . however , it is expected that the effects of cooling of gas by dust come to be important for the ism with higher dust - to - gas mass ratio .
the result of the calculation for dust destruction in the ism with @xmath50 @xmath4 and the dust - to - gas mass ratio @xmath251 shows that the truncation time and the overall efficiency of dust destruction decrease by 4.4 % and 2.6 % , respectively , compared with the results for dust - to - gas mass ratio of @xmath250 , though the gas temperature in postshock gas decreases by about 20% .
even if the dust - to - gas mass ratio is raised up to by four orders of magnitude , the cooling of gas by dust in nonradiative shock does not have the significant effects on the efficiency of dust destruction as well as the evolution of nonradiative shock .
furthermore , the inverse compton cooling has influences on the evolution of nonradiative shock in the early universe . because the cooling rate by the inverse compton scattering is proportional to @xmath252 and that by atomic process is proportional to @xmath253 , the contribution of the inverse compton cooling is enhanced in postshock flow with the lower gas density and/or the higher gas temperature at the higher redshift . at the redshift of @xmath254 , for the sn model c20 with the preshock gas density of @xmath23 @xmath24 ,
the truncation time is only 2 % shorter than that calculated at @xmath255 .
thus , in this case , the inverse compton cooling seems not to affect the destruction efficiency of dust grains . however , more systematic studies covering wide ranges of the sn explosion energy and the gas density in the ism are necessary to reveal the effect of the inverse compton cooling .
note that the cooling function of gas for @xmath256 @xmath257 is almost the same as that for the zero metal case ( sutherland & dopita 1993 ) .
thus , the parameters of the ism considered in this paper ( @xmath50 @xmath4 and @xmath88 ) is competent for investigating dust destruction in the early universe , and the time - scale of dust destruction derived in 7.2 is applicable over the wide ranges of metallicity of gas in the ism ( @xmath3 @xmath4 ) and redshift ( @xmath258 ) .
we investigate the destruction of dust grains in the interstellar shocks driven by sne as the second step to reveal the evolution of dust in the early universe , based on the dust models obtained by nozawa et al .
we focus on dust destruction in nonradiative shock where dust grains are predominantly destroyed by non - thermal and thermal sputtering because of high temperature and high velocity of gas . the sputtering yields for the combinations of dust and ion species of interest to us
are evaluated by applying the universal relation for sputtering yield with a slight modification and by determining the value of @xmath117 by fitting to the available experimental data and/or the results of the eddy simulations .
the modified universal relation derived here can present the better fits to the available experimental data than the previous one . in the calculations of dust destruction ,
we solve the time evolution of gas temperature and density in spherically symmetric shocks , adopting the hydrodynamical models by umeda & nomoto ( 2002 ) as the initial condition for interstellar shock and including the cooling of gas by the atomic process , the inverse compton scattering and the thermal emission of dust .
the erosion of dust by thermal and non - thermal sputtering caused by motion of dust relative to gas is carefully treated by taking into account the size distribution of each dust species .
the results of calculations are summarized as follows .
\(1 ) because the sputtering predominantly destroys the small grain , the number of small - sized grain such as al@xmath5o@xmath59 and mgsio@xmath59 is greatly reduced .
the erosion of c , sio@xmath5 and fe grains whose size distributions are lognormal - like with the average size larger than @xmath259 @xmath61 m increases the numbers of smaller ones .
the size distribution summed up over all grains species becomes flatter for the small radius compared with the initial size distribution , while that for the radius larger than 0.2 @xmath61 m remains almost unchanged .
\(2 ) the efficiency of dust destruction is higher for the grains with the small average size such as al@xmath5o@xmath59 and mgsio@xmath59 and with the power - law - like size distribution such as fes , mg@xmath5sio@xmath60 and mgo . on the other hand ,
si , fe , c and sio@xmath5 grains , which have the lognormal - like size distribution and the relatively large average radius , have lower efficiency of destruction .
the detailed analysis of the behavior of dust destruction efficiency for each grain species indicates that not only sputtering yields but also the initial size distribution plays a crutial role in the efficiency of dust destruction by sputtering .
\(3 ) the efficiency of destruction @xmath28 for each dust species increases with increasing the explosion energy @xmath33 and/or with increasing the preshock gas density @xmath245 , but is almost independent of the progenitor mass @xmath1 of sn as long as @xmath33 is the same .
the destruction efficiency @xmath28 as a function of @xmath33 is well reproduced by a power - law formula given by equation ( a1 ) .
the dependence of @xmath28 on @xmath245 is well expressed by the quadratic equation ( a3 ) in terms of @xmath260 .
\(4 ) as is the same as @xmath28 , the mass of gas swept up by shock wave @xmath29 is the increasing function of @xmath33 and is approximated by a power - law formula given by equation ( a2 ) .
however , @xmath29 decreases with increasing the gas density in the ism , and its dependence is also reproduced by a power - law formula whose index is given by equation ( a4 ) as a function of @xmath33 .
finally , by combining these results , we present the analytic formula for the time - scale of destruction for each grain species in the early universe as a function of @xmath33 and @xmath245 .
the derived time - scale of dust destruction can be employed to investigate the time evolution of the amount of dust grains in the early universe .
\(5 ) in the early universe , only the h and he line coolings affect the time evolution of nonradiative shock and the efficiencies of dust destruction .
the thermal emission from dust grains is not so much important even for the ism with the dust - to - gas mass ratio of @xmath261 as long as the metallicity in the ism is @xmath3 @xmath4 .
also , it is expected that the inverse compton cooling does not have the great effects on the evolution of shock at high redshifts .
thus , the time - scale of dust destruction derived in this paper is applicable for ranges of metallicity of gas in the ism of @xmath3 @xmath4 and redshift of @xmath262 . in this study , we focused on dust destruction in nonradiative shock where dust grains in the ism are considered to be predominantly destroyed .
we should be mentioned here that a part of dust grains formed in the ejecta of sne are expected to be destroyed by reverse shock penetrating into the ejecta . because sputtering is the most dominant destruction process of dust grains in reverse shock as well , the simulation code of dust destruction constructed in this paper
can be applied to explore the destruction of dust grains by reverse shocks .
this work is now in progress .
furthermore , in radiative shock with the shock velocities below 100 km s@xmath18 , shattering and partial vaporization by grain - grain collisions become important in the presence of a magnetic field . a magnetic field frozen into the gas
causes charged grains to gyrate and accelerate behind the shock front and results in the enhancement of not only the sputtering rate but also grain - grain collision frequency .
the shattering by grain - grain collision mainly leads to the redistribution of grain size ( tielens et al . 1994 ; jones , tielens , & hollenbach 1996 ) . the sophisticated study on this subject
is left for the future work .
the authors are grateful to the anonymous referee for critical comments which have improved the manuscript .
the authors thank dr . h. umeda and prof .
k. nomoto for making the ejecta model of pop iii supernovae available .
this work has been supported in part by a grant - in - aid for scientific research from the japan society for the promotion of sciences ( 13640229 ) .
in this appendix , we present the approximate formulae for the efficiency of dust destruction @xmath28 and the mass of gas swept up by shock @xmath29 as function of the sn explosion energy @xmath33 and the gas density @xmath263 in the ism . the energy dependence of the destruction efficiency for each grain species can be well reproduced by a power - law formula given by @xmath264 the solid lines in figure 8 indicate the least - squares fittings by this power - law formula for all sne whose numerical coefficients @xmath237 and @xmath238 are tabulated in table 6 .
the dependence of @xmath29 on @xmath33 is well fitted by a power - law formula @xmath265 using the coefficients @xmath266 and @xmath267 given in table 6 , and is depicted in figure 9 for all sne ( solid ) , sne ii ( dotted ) , and pisne ( dashed ) .
the destruction efficiency @xmath28 for each grains species @xmath26 as well as the overall destruction efficiency as a function of @xmath268 is expressed by @xmath269 and the coefficients @xmath270 , @xmath271 and @xmath272 derived by the least squares fittings are given in table 7 for the models c20 , h25 and h30 , respectively , along with the coefficients for overall destruction efficiency .
the solid curves in figure 10a depict the overall destruction efficiencies calculated by equation ( a3 ) for the models c20 , h25 and h30 .
this formula reproduces the destruction efficiencies calculated by the simulations with the accuracy less than 10 % for @xmath273 @xmath24 .
the calculated @xmath29 can be reproduced by a power - law formula @xmath274 where the index @xmath275 is weakly dependent on @xmath33 and is approximated by @xmath276 this formula gives good agreements with the results of simulations , as represented by the solid lines in figure 10b .
andersen , h. h. , & bay , h. l. 1981 , in sputtering by particle bombardment i , ed .
r. behrisch ( berlin ; springer - verlag ) , 145 bach , h. 1970 , nucl .
meth . , 84 , 4 bach , h. , kitzmann , i. , & schroder , h. 1974 , radiation effects , 21 , 31 baines , m. j. , williams , i. p. , & asebiomo , a. s. 1965 , , 130 , 63 bertoldi , f. , carilli , c. l. , cox , p. , fan , x. , strauss , m. a. , beelen , a. , omont , a. , & zylka , r. 2003 , , 406 , l55 bianchi , s. , & ferrara , a. 2005 , , 358 , 379 blank , p. , & wittmaack , k. 1979 , j. appl .
, 50 , 1519 bohdansky , j. 1984 , nucl .
b2 , 587 bohdansky , j. , bay , h. l. , & ottenberger , w. 1978 , j. nucl . mat .
, 7677 , 163 bohdansky , j. , bay , h. l. , & roth , j. 1977 , in proc .
. vacuum congress and 3rd int .
conf . on solid surfaces , ed . by dobrozemsky , r. , rudenauer , f. , viehbock , f. p. & breth , a. ( private publisher , wien ) , p. 1509
bohdansky , j. , roth , j. , & sinha , m. k. 1976 , in proc .
on fusion technology ( pergamon ; london ) , p. 541 bohdansky , j. , roth , j. , & bay , h. l. 1980 , j. appl .
phys . , 51 , 2861 bromm , v. , & larson , r. b. 2004 , , 42 , 79 cazaux , s. , & spaans , m. 2004 , , 611 , 40 chase , m. w. , davies , c. a. , downey , j. r. , frurip , d. j. , mcdonald , r. a. , & syverud , a. n. 1985 , janaf thermochemical tables 3d ed .
( j. phys .
ch em . ref .
data , 14 , suppl 1 ) coburn , j. w. , winters , h. f. , & chuang , t. j. 1977 , j. appl .
, 48 , 3532 draine , b. t. , & salpeter , e. e. 1979 , , 231 , 77 dwek , e. 1987 , , 322 , 812 dwek , e. 1998 , , 501 , 643 dwek , e. , & arendt , r. g. 1992 , , 30 , 11 dwek , e. , foster , s. m. , & vancura , o. 1996 , , 457 , 244 edwin , r. p. 1973
, j. phys .
d , 6 , 833 eernisse , e. p. 1971
, j. appl .
, 42 , 480 field , d. , may , p. w. , pineau des forets , g. , & flower , d. r. 1997 , , 285 , 839 gnedin , n. y. , ferrara , a. , & zweibel , e. g. 2000 , , 539 , 505 hauser , m. g. , & dwek , e. 2001 , , 39 , 249 hechtl , e. , bohdansky , j. , & roth , j. 1981 , j. nucl . mat .
, 103104 , 333 heger , a. , & woosley , s. e. 2002 , , 567 , 532 hirashita , h. , & ferrara , a. 2002 , , 337 , 921 hirashita , h. , nozawa , t. , kozasa , t. , ishii , t. t. , & takeuchi , t. t. 2005 , 357 , 1077 hirooka , y. 2001 , http://dppwi.nifs.ac.jp/ ikeuchi , s. , & ostriker , j. p. 1986
, , 301 , 522 iskef , h. , cunningham , j. w. , & watt , d. e. 1983 , phys
, 28 , 535 jones , a. p. , tielens , a. g. g. m. , hollenbach , d. j. , & mckee , c. f. 1994 , , 433 , 797 jones , a. p. , tielens , a. g. g. m. , & hollenbach , d. j. 1996 , , 469 , 740 kitayama , t. , yoshida , n. , susa , h. , & umemura , m. 2004 , , 613 , 631 kozasa , t. , hasegawa , h. , & nomoto , k. 1989 , , 344 , 325 laegreid , n. , & wehner , g. k. 1961 , j. appl . phys . , 32 , 365 loeb , a. , & haiman , z. 1997 , , 490 , 571 machida , m. n. , tomisaka , k. , nakamura , f. , & fujimoto , m. y. 2005 , , 622 , 39 maiolino , r. , oliva , e. , ghinassi , f. , pedani , m. , mannucci , f. , mujica , r. , & juarez , y. 2004a , , 420 , 889 maiolino , r. , schneider , r. , oliva , e. , bianchi , s. , ferrara , a. , mannucci , f. , pedani , m. , & roca sogorb , m. 2004b , , 431 , 533 mair , g. , mllar , e. , hillebrandt , w. , & arnold , c. n. 1988 , , 199 , 114 matsunami , n. , yamamura , y. , itakawa , w. , itoh , n. , kazumata , y. , myagawa , s. , morita , k. , & shimizu , r. 1980 , radiation effects lett . , 57 , 15 may , p. w. , pineau des forets , g. , flower , d. r. , field , d. , allan , n. l. , & purton , j. a. 2000 , , 318 , 809 mckee , c. f. 1989 , in interstellar dust , ed . l. j. allamandora & a. g. g. m. tielens ( dordrechent : kluwer ) , 431 mckee , c. f. , hollenbach , d. j. , seab , c. g. , & tielens , a. g. g. m. 1987 , , 318 , 674 nenadovic , t. , perraillon , b. , bogdanov , z. , djordjevic , z. , & milic , m. 1990 , nucl .
instr . meth .
res . , b48 , 538 nozawa , t. , kozasa , t. , umeda , h. , maeda , k. , & nomoto , k. 2003 , , 598 , 785 omukai , k. , tsuribe , t. , schneider , r. , & ferrara , a. 2005 , , 626 , 627 priddey , r. s. , isaak , k. g. , mcmahon , r. g. , robson , e. i. , & pearson , c. p. 2003 , , 344 , l74 robson , i. , priddey , r. s. , isaak , k. g. , & mcmahon , r. g. 2004 , , 351 , l29 rosenberg , d. , & wehner , g. k. 1962 , j. appl .
, 33 , 1842 roth , j. 1983 , in sputtering by particle bombardment ii , ed .
r. behrisch ( berlin ; springer - verlag ) , 91 roth , j. , bohdansky , j. , & ottenberger , w. 1979 , rpt . ipp .
9/26 ( mpi plasmaphysik , garching ) schneider , r. , ferrara , a. , natarajan , p. , & omukai , k. 2002 , , 571 , 30 schneider , r. , ferrara , a. , salvaterra , r. , omukai , k. , & bromm , v. 2003 , , 422 , 869 schneider , r. , ferrara , a. , & salvaterra , r. 2004 , , 351 , 1379 sigmund , p. 1969
, 184 , 383 sigmund , p. 1981
, in sputtering by particle bombardment i , ed .
r. behrisch ( berlin ; springer - verlag ) , 9 sommerfeldt , h. , mashkova , e. s. , & molchanov , v. a. 1972 , phys .
lett . , 38a , 237 southern , a. l. , willis , w. r. , & robinson , m. t. 1963 , j. appl .
34 , 153 sutherland , r. s. , & dopita , m. a. 1993 , , 88 , 253 tabata , t. , ito , r. , & okabe , s. 1972 , nucl .
, 103 , 85 takeuchi , t. t. , hirashita , h. , ishii , t. t. , hunt , l. k. , & ferrara , a. 2003 , , 343 , 839 takeuchi , t. t. , ishii , t. t. , nozawa , t. , kozasa , t. , & hirashita , h. 2005 , , 362 , 592 tan , j. c. , & mckee , c. f. 2004 , , 603 , 383 tielens , a. g. g. m. , mckee , c. f. , seab , c. g. , & hollenbach , d. j. 1994 , , 431 , 321 todini , p. , & ferrara , a. 2001 , , 325 , 726 umeda , h. , & nomoto , k. 2002 , , 565 , 385 van albada , g. d. , van leer , b. , & roberts , w. w. jr . 1982 , , 108 , 76 vanzi , l. , hunt , l. k. , thuan , t. x. , & izotov , y. i. , 363 , 493 von seefeld , h. , schmidl , h. , behrisch , r. , & scherzer , b. m. u. 1976 , j. nucl .
, 63 , 215 wolfire , m. g. , hollenbach , d. , mckee , c. f. , tielens , a. g. g. m. , & bakes , e. l. o. 1995 , , 443 , 152 ziegler , j. f. , biersack , j. p. , & littmark , u. 1985 , the stopping and range of ions in solids .
pergamon , new york lccccc c & 2.26 & 0.103 & 0.029 & & + si & 2.32 & 0.304 & 0.250 & & + fe & 7.89 & 0.050 & 0.148 & & + fes & 4.84 & 0.156 & 0.0088 & & + al@xmath5o@xmath59 & 3.98 & 0.002 & 0.0007 & 0.001 & 0.0003 + mgsio@xmath59 & 3.18 & 0.010 & 0.0045 & 0.102 & 0.011 + mg@xmath5sio@xmath60 & 3.20 & 0.219 & 0.0066 & 0.205 & 0.010 + sio@xmath5 & 2.64 & 0.100 & 0.047 & 0.588 & 0.033 + mgo & 3.56 & 0.056 & 0.0058 & & + fe@xmath59o@xmath60 & 5.21 & & & 0.103 & 0.0022 + lccccl c & 4.0 & 6 & 12 & 0.61 & 1 , 2 , 3 , 4 , 5 + si & 4.66 & 14 & 28 & 0.43 & 3 , 4 , 6 , 7 , 8 , 9 , 10 , 11 + fe & 4.31 & 26 & 56 & 0.23 & 3 , 4 , 5 , 7 , 8 , 12 , 13 + fes & 4.12 & 21 & 44 & 0.18 & eddy + al@xmath5o@xmath59 & 8.5 & 10 & 20.4 & 0.08 & 3 , 14 , 15 + mgsio@xmath59 & 6.0 & 10 & 20 & 0.1 & 16 + mg@xmath5sio@xmath60 & 5.8 & 10 & 20 & 0.1 & 16 + sio@xmath5 & 6.42 & 10 & 20 & 0.1 & 3 , 14 , 15 , 17 , 18 + mgo & 5.17 & 10 & 20 & 0.06 & 14 , 15 , eddy + fe@xmath59o@xmath60 & 4.98 & 15.7 & 33.1 & 0.15 & eddy + lcccc c & 0.247 & 0.029 & & + si & 0.134 & 0.250 & & + fe & 0.154 & 0.148 & & + fes & 0.578 & 0.0088 & & + al@xmath5o@xmath59 & 0.667 & 0.0007 & 0.794 & 0.0003 + mgsio@xmath59 & 0.637 & 0.0045 & 0.631 & 0.011 + mg@xmath5sio@xmath60 & 0.451 & 0.0066 & 0.586 & 0.010 + sio@xmath5 & 0.411 & 0.047 & 0.399 & 0.033 + mgo & 0.505 & 0.0058 & & + fe@xmath59o@xmath60 & & & 0.741 & 0.0022
+ overall & 0.340 & & 0.497 & + lcccccc
c & 2.532e-01 & 1.660e-01 & 2.526e-01 & 1.407e-01 & 2.580e-01 & 1.820e-01 + si & 1.380e-01 & 2.371e-01 & 1.377e-01 & 1.979e-01 & 1.414e-01 & 2.631e-01 + fe & 1.583e-01 & 2.402e-01 & 1.586e-01 & 2.005e-01 & 1.594e-01 & 2.707e-01 + fes & 5.842e-01 & 7.897e-02 & 5.850e-01 & 6.758e-02 & 5.834e-01 & 8.857e-02 + al@xmath5o@xmath59 & 6.724e-01 & 5.649e-02 & 6.733e-01 & 4.803e-02 & 6.709e-01 & 6.393e-02 + mgsio@xmath59 & 6.429e-01 & 6.358e-02 & 6.438e-01 & 5.421e-02 & 6.417e-01 & 7.165e-02 + mg@xmath5sio@xmath60 & 4.585e-01 & 1.146e-01 & 4.590e-01 & 9.827e-02 & 4.595e-01 & 1.275e-01 + sio@xmath5 & 4.183e-01 & 1.238e-01 & 4.189e-01 & 1.054e-01 & 4.19e-01 & 1.383e-01 + mgo & 5.117e-01 & 1.046e-01 & 5.122e-01 & 9.020e-02 & 5.124e-01 & 1.160e-01 + overall & 3.458e-01 & 1.307e-01 & 3.460e-01 & 1.099e-01 & 3.476e-01 & 1.463e-01 + & & & & & & + & @xmath266 & @xmath267 & @xmath266 & @xmath267 & @xmath266 & @xmath267 + @xmath29 & 4.144e+03 & 7.967e-01 & 4.118e+03 & 6.627e-01 & 4.481e+03 & 8.864e-01 + lccccccccc c & @xmath281 & 0.41 & @xmath282 & @xmath283 & 0.36 & @xmath284 & @xmath285 & 0.33 & @xmath286 + si & @xmath287 & 0.54 & @xmath288 & @xmath289 & 0.50 & @xmath290 & @xmath291 & 0.47 & @xmath292 + fe & @xmath293 & 0.55 & @xmath294 & @xmath295 & 0.49 & @xmath282 & @xmath296 & 0.46 & @xmath297 + fes & @xmath298 & 0.23 & @xmath299 & @xmath300 & 0.18 & @xmath301 & @xmath302 & 0.16 & @xmath303 + al@xmath5o@xmath59 & @xmath304 & 0.14 & @xmath305 & @xmath306 & 0.12 & @xmath303 & @xmath307 & 0.11 & @xmath308 + mgsio@xmath59 & @xmath309 & 0.18 & @xmath310 & @xmath293 & 0.14 & @xmath303 & @xmath311 & 0.12 & @xmath308 + mg@xmath5sio@xmath60 & @xmath312 & 0.31 & @xmath313 & @xmath314 & 0.25 & @xmath315 & @xmath316 & 0.22 & @xmath310 + sio@xmath5 & @xmath317 & 0.33 & @xmath286 & @xmath318 & 0.27 & @xmath319 & @xmath316 & 0.24 & @xmath320 + mgo & @xmath321 & 0.29 & @xmath322 & @xmath317 & 0.23 & @xmath323 & @xmath324 & 0.20 & @xmath301 + overall & @xmath325 & 0.33 & @xmath326 & @xmath325 & 0.29 & @xmath313 & @xmath291 & 0.26 & @xmath322 + | we investigate the destruction of dust grains by sputtering in the high - velocity interstellar shocks driven by supernovae ( sne ) in the early universe to reveal the dependence of the time - scale of dust destruction on the gas density @xmath0 in the interstellar medium ( ism ) as well as on the progenitor mass @xmath1 and explosion energy @xmath2 of sn . the sputtering yields for the combinations of dust and ion species of interest to us
are evaluated by applying the so - called universal relation with a slight modification .
the dynamics of dust grains and their destruction by sputtering in shock are calculated by taking into account the size distribution of each dust species , together with the time evolution of temperature and density of gas in spherically symmetric shocks .
the results of calculations show that the efficiency of dust destruction depends not only on the sputtering yield but also on the initial size distribution of each grain species .
the efficiency of dust destruction increases with increasing @xmath2 and/or increasing @xmath0 , but is almost independent of @xmath1 as long as @xmath2 is the same .
the mass of gas swept up by shock is the increasing function of @xmath2 and the decreasing function of @xmath0 .
combining these results , we present the approximation formula for the time - scale of destruction for each grain species in the early universe as a function of @xmath2 and @xmath0 .
this formula is applicable for investigating the evolution of dust grains at the early epoch of the universe with the metallicity of @xmath3 @xmath4 .
the effects of the cooling processes of gas on the destruction of dust are briefly discussed . |
link:\doibase 10.1524/zkri.219.7.391.35643 [ * * , ( ) ] link:\doibase 10.1039/c2cs35063 g [ * * , ( ) ] link:\doibase 10.1103/physrevlett.53.1951 [ * * , ( ) ] link:\doibase 10.1016/j.cocis.2004.12.003 [ * * , ( ) ] link:\doibase 10.1073/pnas.1008695108 [ * * , ( ) ] link:\doibase 10.1002/ijch.201100146 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.58.706 [ * * , ( ) ] link:\doibase 10.1103/physrevb.39.446 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.70.2924 [ * * , ( ) ] link:\doibase 10.1103/physreve.60.2664 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.98.225505 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.99.235503 [ * * , ( ) ] link:\doibase 10.1021/jz101706k [ * * , ( ) ] link:\doibase 10.1063/1.3499323 [ * * , ( ) ] link:\doibase 10.1088/0256 - 307x/29/5/050204 [ * * , ( ) ] link:\doibase 10.1002/polb.22395 [ * * , ( ) ] link:\doibase 10.1038/nature08641 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.215702 [ * * , ( ) ] link:\doibase 10.1063/1.3651370 [ * * , ( ) ] link:\doibase 10.1073/pnas.1019763108 [ * * , ( ) ] link:\doibase 10.1063/1.3679653 [ * * , ( ) ] link:\doibase 10.1021/ar010013x [ * * , ( ) ] link:\doibase 10.1088/0953 - 8984/20/15/153101 [ * * , ( ) ] link:\doibase 10.1063/1.448024 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.225502 [ * * , ( ) ] link:\doibase 10.1038/nmat1949 [ * * , ( ) ] link:\doibase 10.1002/marc.200900614 [ * * , ( ) ] link:\doibase 10.1063/1.2759922 [ * * , ( ) ] link:\doibase 10.1063/1.2752155 [ * * , ( ) ] link:\doibase 10.1063/1.3454907 [ * * , ( ) ] link:\doibase 10.1039/c0sm01377c [ * * , ( ) ] link:\doibase 10.1088/0953 - 8984/24/28/284124 [ * * , ( ) ] link:\doibase 10.1039/b614955c [ * * , ( ) ] link:\doibase 10.1088/0953 - 8984/22/10/104105 [ * * , ( ) ] link:\doibase 10.1126/science.1133162 [ * * , ( ) ] link:\doibase 10.1021/cm070051w [ * * , ( ) ] link:\doibase 10.1039/b716393b [ * * , ( ) ] link:\doibase 10.1021/ja8079803 [ * * , ( ) ] link:\doibase 10.1002/anie.200801061 [ * * , ( ) ] link:\doibase 10.1002/adfm.200902076 [ * * , ( ) ] link:\doibase 10.1039/c0cs00048e [ * * , ( ) ] link:\doibase 10.1038/nature11564 [ * * , ( ) ] link:\doibase 10.1038/nature09713 [ * * , ( ) ] link:\doibase 10.1063/1.1569473 [ * * , ( ) ] link:\doibase 10.1021/nl0493500 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.97.168301 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.103.237801 [ * * , ( ) ] link:\doibase 10.1039/c0sm01494j [ * * , ( ) ] link:\doibase 10.1039/c0cp02296a [ * * , ( ) ] link:\doibase 10.1107/s0365110x58000487 [ * * , ( ) ] link:\doibase 10.1107/s0365110x59001499 [ * * , ( ) ] link:\doibase 10.1103/physrevb.48.6966 [ * * , ( ) ] link:\doibase 10.1080/00268979300100881 [ * * , ( ) ] link:\doibase 10.1063/1.465023 [ * * , ( ) ] link:\doibase 10.1063/1.467526 [ * * , ( ) ] link:\doibase 10.1016/s0378 - 4371(98)00515 - 9 [ * * , ( ) ] link:\doibase 10.1126/science.271.5248.484 [ * * , ( ) ] link:\doibase 10.1126/science.1089389 [ * * , ( ) ] link:\doibase 10.1021/ja0541938 [ * * , ( ) ] link:\doibase 10.1021/ja0665141 [ * * , ( ) ] link:\doibase 10.1073/pnas.0803841105 [ * * , ( ) ] link:\doibase 10.1063/1.1699114 [ * * , ( ) ] @noop _ _ , ed .
( , , ) link:\doibase 10.1080/00268978400101951 [ * * , ( ) ] link:\doibase 10.1016/0009 - 2614(77)85375-x [ * * , ( ) ] link:\doibase 10.1063/1.481457 [ * * , ( ) ] * supplementary material *
the simulations described in our work used the metropolis monte carlo approach @xcite in the isobaric - isothermal ensemble @xcite with periodic boundary conditions .
the potential as discussed in the manuscript is given more formally by @xmath36 where @xmath11 is the interparticle vector connecting the centres of the two particles @xmath10 and @xmath37 , @xmath38 is its magnitude , @xmath39 and @xmath40 are the orientations of the particles @xmath10 and @xmath37 , the generalised lennard - jones potential is @xmath41,\label{quasi - eqn - lj}\ ] ] where @xmath4 for the standard lennard - jones potential , and @xmath42\exp\left[\frac{-\theta_{l_\text{min}ji}^2}{2\,\sigma_\text{pw}^2}\right],\ ] ] where @xmath7 is the patch width , @xmath8 is the angle between the patch vector of patch @xmath9 on particle @xmath10 and the interparticle vector @xmath11 , and @xmath43 is the patch that minimises the magnitude of this angle .
two particles therefore interact only through a single pair of patches . in the simulations reported here
, we use a potential cutoff of @xmath44 , and the potential in eq .
is shifted so that it equals zero at @xmath45 .
the crossover to including angular modulation in the potential ( eq . ) is likewise adjusted so that it still occurs when the potential energy is identically zero .
most simulations reported used relatively small system sizes ( of the order of 500 particles ) , but the phase diagram in fig .
[ fig - quasicrystal - phasediags - all](a ) and the free energies in fig .
[ fig - quasi - free - energies - bp1.5 ] were also calculated using a system of the order of 2500 particles , and certain transitions were calculated with systems of 5000 and 10000 particles , with no detectable differences in the behaviour of the systems , confirming that reasonably small simulations were sufficient to calculate phase diagrams .
here , we summarise the methods we used in the manuscript in somewhat more detail . the approaches described here
are all standard and are provided here as a quick summary for the benefit of the reader only .
direct coexistence simulations @xcite are simulations in which an explicit interface is set up between the two phases in question , and the thermodynamically stable phase is expected to grow into the thermodynamically unstable phase such that the interface between the two phases moves . in brute force simulations ,
no interface is set up ; instead , a single phase is simulated and it spontaneously converts into the thermodynamically stable phase . typically , this brute force simulation process is prone to hysteresis effects , where for example the quasicrystal will form at some temperature @xmath46 , but when the quasicrystal is heated in a reverse brute force simulation , it will melt at temperature @xmath47 , where @xmath48 , because nucleation involves a free energy barrier that must be overcome . in direct coexistence simulations , this hysteresis does not occur , because both phases are present from the start , and the interface between them is pre - formed . in our simulations
, we established that there is essentially no hysteresis between the quasicrystal and the fluid under the conditions described in this paragraph , and so direct coexistence simulations offer no advantage over brute force simulations . indeed , because nucleation is so facile , numerous phase interfaces are spontaneously formed in a direct coexistence simulation , and so direct coexistence simulations were not useful in phase diagram determination at this stage .
however , we have used direct coexistence simulations in certain cases as a check in confirming that the frenkel
ladd simulations yielded the correct coexistence points .
thermodynamic integration @xcite refers to a series of integration schemes of thermodynamic potentials . for integration along
iso-(@xmath49 ) curves , where @xmath50 , we start from the product rule @xmath51 from where an application of simple thermodynamics allows us to write @xmath52 this can be integrated to give @xmath53 and this is the result we require for thermodynamic integration .
provided we know the gibbs energy at some @xmath46 , then we can find the gibbs energy at @xmath47 along the iso-(@xmath49 ) curve by running several @xmath14 simulations for temperatures between @xmath46 and @xmath47 ( at fixed @xmath49 ) , and finding a polynomial fit to the integrand , which can then by integrated analytically .
analogous results arise for integration along isobars @xcite , @xmath54 and isotherms @xcite , @xmath55 it is also possible to extend the scheme by coupling to a different hamiltonian , such that the potential energy is given by @xmath56 , and then integrating along @xmath57 ; we find that @xcite @xmath58 this is known as hamiltonian thermodynamic integration @xcite and is one component of the einstein crystal frenkel
ladd approach @xcite .
the free energy of a fluid phase evaluated from the ideal gas by decreasing the pressure at constant temperature is given by @xcite @xmath59 \,{\ensuremath{\mathrm{d}}}\rho,\ ] ] where @xmath60 is the de broglie thermal wavelength .
it is usually beneficial to fit the entire integrand to a single polynomial to avoid diverging terms .
as the density tends to zero , the integrand tends to the second virial coefficient @xcite ; it is helpful to evaluate this explicitly to benchmark the simulation results . in gibbs
duhem integration @xcite , we numerically integrate the clapeyron equation , @xmath61 or an analogous expression , in order to obtain new coexistence points from already known coexistence points . in our simulations
, we used the fourth - order runge
kutta algorithm to perform this integration .
the einstein crystal approach @xcite is a fairly involved way of calculating the free energy of a crystal from the einstein crystal , for which the free energy is known .
we direct the interested reader to frenkel and smit s book ( ref .
@xcite ) and the review paper by vega and co - workers ( ref .
@xcite ) for an in - depth discussion of the method . in the frenkel
ladd scheme as implemented for our model , we take the rotational energy to be @xmath62,\ ] ] where @xmath63 is a constant that we vary and @xmath64 is the number of patches .
we thus ensure that the particle s symmetry is accounted for : rotations into degenerate positions give the same frenkel
ladd energy .
computing absolute free energies from simulations requires a combination of numerical techniques , each of which can introduce errors , both systematic and random . since it is difficult to keep track of how the errors propagate ,
the approach which is usually undertaken is to perform _ a posteriori _ checks on the whole process . as discussed by vega and co - workers @xcite ,
` consistency checks ' can be applied to simulations to ensure that the data obtained in free energy calculations are self - consistent and correspond to reality .
there are several such checks that can be performed .
for example , we can calculate the free energy at a particular point using several routes ( _ e.g. _ by integrating along different isotherms , isobars or iso-(@xmath49 ) curves to the same point ) .
furthermore , direct coexistence simulations can be used to ensure that the transition temperature does in fact occur where free energy curves intersect . finally , in simulations involving gibbs
duhem integration , we can integrate first in the ` forward ' and then in the ` reverse ' direction , ensuring that the method reproduces the starting point .
we have performed all these checks to ensure that the results we have obtained are robust .
gao and co - workers compared the results of debye crystal reference state calculations with different force constants to obtain an estimate of the error in the free energy of hexagonal ice @xcite .
we do something similar to estimate the error in free energy in fig .
[ fig - quasi - free - energies - bp1.5 ] in the main paper : we calculate the frenkel ladd free energy at several points ( @xmath65 , @xmath66 , @xmath67 and @xmath68 ) along the iso-(@xmath49 ) curve , and then use thermodynamic integration to obtain free energies from each of the starting frenkel ladd free energies . the largest difference in free energy between any pair of calculations along the entire curve
can then serve as an estimate of the error in free energy in our simulations .
largest _ such deviation was @xmath69 , which is sufficiently small not to affect any conclusion , and the error in practice is likely to be significantly smaller , as we were able to confirm phase transition temperatures in independent direct coexistence simulations . | we introduce an approach to computing the free energy of quasicrystals , which we use to calculate phase diagrams for systems of two - dimensional patchy particles with five regularly arranged patches that have previously been shown to form dodecagonal quasicrystals .
we find that the quasicrystal is a thermodynamically stable phase for a wide range of conditions and remains a robust feature of the system as the potential s parameters are varied .
we also demonstrate that the quasicrystal is entropically stabilised over its crystalline approximants .
quasicrystals are a type of aperiodic crystal structure : they have well - ordered structures , but can not be described by a periodic lattice , not even one with incommensurate lattice parameters @xcite .
they were first reported by shechtman _
et al . _
@xcite , in whose work a rapidly cooled al - mn alloy was found to result in a metastable quasicrystal .
most quasicrystals discovered so far are metallic alloys @xcite ; however , recently , an increasing number of examples have been reported in the field of soft condensed matter @xcite .
quasicrystals have also been observed to form in computer simulations , in which not only binary , but also one - component quasicrystals have been seen @xcite . while certain quasicrystals have been shown to be stable in experiment @xcite , in simulations , the relatively short timescales accessible mean that it is not necessarily clear whether quasicrystals are truly the stable phase , or rather a metastable phase that is more kinetically accessible . in this work
, we address this important question of assessing the stability of quasicrystals and apply it to a soft matter system that has the potential to be realised experimentally .
to prove that quasicrystals are thermodynamically stable , their free energy must be computed .
however , such a calculation is not straightforward : the principal problem is that there is no obvious reference state whose free energy is known and from which thermodynamic integration @xcite could be used to calculate the free energy of the quasicrystalline phase .
a phase transition intervenes when integrating from an ideal gas , which is normally used as a reference state for fluid phases , and an einstein crystal , used as a reference state for crystalline phases , would also not be suitable because it would fail to capture the configurational entropy associated with the quasicrystal s many possible structures .
one approach to assessing quasicrystal stability is to compute the free energy of an approximant crystalline phase .
however , it is not immediately obvious whether quasicrystals are more or less stable than their approximants @xcite ; whilst the enthalpy and vibrational properties of the quasicrystal and its approximant phases are likely to be similar , the free energy of a quasicrystal has a contribution from its configurational entropy that is not present for the approximant @xcite .
nevertheless , such calculations have provided some insight into the phase behaviour of systems such as hard tetrahedra and bipyramids @xcite and spherical micelles @xcite . by contrast , kiselev _ et al .
_ recently estimated the free energy of the quasicrystal itself by combining a phonon contribution for a particular quasicrystal configuration from thermodynamic integration and a configurational contribution based on an approximation of uncorrelated phason flips @xcite .
they confirmed that , within their approximation , the quasicrystal s free energy is in certain conditions lower than the approximant s .
our solution to this conundrum is to note that thermodynamic integration is not the only way the melting point of a solid can be computed .
another method is to simulate the direct coexistence of two phases with an interface @xcite . by performing such simulations at a range of temperatures
, we can bracket the regions where the quasicrystal melts ( @xmath0 ) and grows ( @xmath1 ) . at @xmath2 , the chemical potential of
the quasicrystal is equal to that of the fluid , which we can calculate using thermodynamic integration .
such an equilibrium approach by construction accounts for the quasicrystal s configurational entropy .
once the quasicrystal s free energy is known at one point , we can use thermodynamic integration to reach other points of interest on the phase diagram . here
, we apply this approach to a two - dimensional system of patchy particles @xcite with five regularly arranged attractive patches , which we model using a simple potential that has previously been used in simulations of self - assembly and crystallisation @xcite .
the potential is based on the lennard - jones ( lj ) form , @xmath3 $ ] , where @xmath4 for the standard lj potential , where for @xmath5 , this potential is multiplied by an angular modulation factor @xmath6 \exp\left[\frac{-\theta_{l_\text{min}ji}^2}{2\,\sigma_\text{pw}^2}\right],\ ] ] where @xmath7 is the ` patch width ' ( in radians ) and @xmath8 is the angle between the patch vector of patch @xmath9 ( chosen to minimise this angle ) on particle @xmath10 and the interparticle vector @xmath11 .
this interaction potential provides a simple model for the patchy colloidal particles that many experimental groups have , with progressively increasing success , been seeking to develop @xcite .
the enhanced range of structural behaviour that such patchy interactions could facilitate has been extensively studied in computer simulation @xcite . phase .
( b ) approximant crystal .
( c ) plastic hexagonal ( z ) phase .
( d ) quasicrystal , @xmath12 .
( e ) neighbour classification of @xmath13 , h and z environments .
the numbers of common neighbours are given for each of the central particle s neighbours .
these local environments are used to colour the particles in ( a)(d ) , where in addition , particles that do not adopt one of these environments are coloured green .
( e ) diffraction pattern corresponding to the configuration shown in ( a ) .
( f ) dodecagonal motif characteristic of the quasicrystal . ]
dodecagonal quasicrystals were previously found to form on cooling systems of these five - patch particles in certain conditions , and structures with the lowest enthalpy were also identified @xcite .
some of the relevant phases studied are shown in fig .
[ fig - quasicrystal](a d ) ; particles are classified based on a common neighbour analysis @xcite into @xmath13 , h and z environments ( shown in fig .
[ fig - quasicrystal](e ) ) , by analogy to the frank
kasper phases @xcite .
the hexagonal ( z ) phase ( fig .
[ fig - quasicrystal](c ) ) was found to form at high pressures ( as it is densest ) and for wide patches ( where the potential is closer to being isotropic ) . at low pressures and at reasonably narrow patch widths , the structure best satisfying the attractive patches is the enthalpically favoured phase .
a crystal can not exist with five - fold symmetry , and so no crystal with five perfectly aligned patches exists . as a compromise
, each particle has five neighbours , but the angles between the neighbours do not perfectly match the five - fold symmetry of the particle itself .
local environments satisfying this requirement are the @xmath13 and h environments shown in fig .
[ fig - quasicrystal](e ) ; the lowest enthalpy structure at low pressure and reasonably narrow patch widths is the @xmath13 crystal shown in fig .
[ fig - quasicrystal](a ) @xcite .
quasicrystals were observed in some cooling runs in the region where the @xmath13 phase is the lowest in enthalpy , but near to the hexagonal ` boundary ' .
the structure of a typical quasicrystal is shown in fig .
[ fig - quasicrystal](d ) ; its diffraction pattern ( fig .
[ fig - quasicrystal](f ) ) exhibits a characteristic twelve - fold symmetry .
the local environments in the quasicrystal are predominantly of the @xmath13 type , but a significant fraction of z environments can be seen : these are typically located at the centre of a dodecagonal motif ( fig .
[ fig - quasicrystal](g ) ) .
much of the structure can be analysed in terms of packing of such dodecagons into triangular , square and rectangular arrangements @xcite , which can readily be seen in fig .
[ fig - quasicrystal](d ) ; the various possible ways of arranging these dodecagons are likely to lead to substantial entropy . there is no translational order , but bonds can be orientated in any one of twelve directions , resulting in long - range orientational order of dodecagonal character .
we perform monte carlo simulations in the @xmath14 ensemble using a rectangular box with periodic boundaries . as a starting point in the determination of phase diagrams for this system , we chose a patch width , temperature and pressure at which the quasicrystal is known to form on cooling , and thus locate the fluid - quasicrystal ( f - qc ) equilibrium transition .
this transition is mostly rapid and facile , with essentially no hysteresis , which allows us to determine the coexistence points directly by performing simulations in relatively large boxes starting from both phases . whilst we initially performed direct coexistence simulations to determine the melting point of the quasicrystal , these simulations
mostly did not prove to be any more efficient than direct brute - force simulations , in which no initial interface was introduced , as the growth of the phases was not restricted to the initial interface .
this lack of hysteresis is most likely indicative of a particularly low f - qc interfacial free energy .
phase transitions can be observed from a kink in the potential energy or density plotted against the variable we are changing ( such as the temperature or the pressure ) .
however , we can further test for the nature of the phase by calculating appropriate diffraction patterns : the quasicrystal exhibits a distinctive 12-fold diffraction pattern ( fig .
[ fig - quasicrystal](f ) ) , the z phase exhibits a six - fold pattern , and the fluid phase does not exhibit a well - defined pattern .
once a f - qc coexistence point is known , we can equate the chemical potential of the fluid ( as calculated by thermodynamic integration ) with that of the quasicrystal .
from here , we can use thermodynamic integration to calculate the chemical potential of the quasicrystal at other temperatures and pressures . by also performing thermodynamic integration to find the chemical potential of the @xmath13 phase
, we can locate the putative @xmath13-qc transition .
we plot in fig .
[ fig - quasi - free - energies - bp1.5 ] the free energies of the @xmath13 , qc and f phases as a function of @xmath15 at a constant @xmath16 . from this figure , we can identify two clear transitions ( @xmath13-qc and qc - f ) , and a temperature window in which the quasicrystal is thermodynamically stable . in the same figure , we also show the free energy of an approximant crystal phase that is based on the most common packing of dodecagons in the quasicrystal ; it is never thermodynamically stable . to rationalise why the quasicrystal is stable ,
we also calculate the enthalpy ( obtained directly from simulations ) and hence obtain the entropy as a function of the temperature .
these functions are also depicted in fig .
[ fig - quasi - free - energies - bp1.5 ] , and it is clear that in the region in which the quasicrystal is stable , it does indeed have a higher enthalpy than the @xmath13 phase , but its significantly higher entropy nevertheless allows it to become more stable .
the origin of this entropy is a combination of the differing vibrational properties and the configurational disorder present in the quasicrystalline phase .
if we assume the vibrational properties of the quasicrystal and its approximant are identical , the configurational component of this entropy can be estimated from the entropy difference between the quasicrystal and its approximant , @xmath17 at @xmath18 .
is significantly larger than the tiling entropy of a random stampfli tiling , but somewhat smaller than that of a maximally random square - triangle tiling , where both have been evaluated at zero phason strain @xcite . ]
phase , as a function of @xmath15 at @xmath19 , @xmath20 and @xmath4 .
lines end at the limit of ( meta)stability of the phases .
dashed lines denote coexistence points , @xmath21 and @xmath22 . the stable phase in each region is explicitly marked .
the error in @xmath23 is @xmath24 @xcite . ] from fig .
[ fig - quasi - free - energies - bp1.5 ] , we can determine two coexistence points on the phase diagram .
we construct the remainder of the phase diagram partly through the same procedure as above at , say , other pressures , but also through the use of gibbs duhem integration @xcite .
this method allows us to integrate clapeyron equation analogues to obtain new coexistence points from already known coexistence points , but requires metastability and hysteresis in the transition so that the simulated phases retain their identity when simulated at and around the coexistence point . as a result
, the method can not be applied to the qc - f and qc - z transitions for much of the phase diagram due to their relative reversibility .
-@xmath15 phase diagram ; @xmath20 , @xmath4 .
( b ) @xmath15-@xmath7 phase diagram ; @xmath25 , @xmath4 .
( c ) @xmath15-@xmath26 phase diagram ; @xmath27 , @xmath20 . ] the resulting phase diagram is shown in fig .
[ fig - quasicrystal - phasediags - all](a ) .
we note that the quasicrystal is stable for a wide range of pressures . at very high pressures ,
the z phase becomes stable due to its greater density .
the z phase is a plastic crystal , meaning that particles are able to rotate relatively freely : this is understandable , as there is no obvious preferred way to orientate a five - fold particle in a six - fold environment . no orientational ordering ensues in the z phase even at low temperatures .
whether it would form an orientationally - ordered crystal or an orientational glass at sufficiently low @xmath15 is not clear .
it is noteworthy that for the qc - z transition , @xmath28 ; since @xmath29 , the clapeyron equation implies that @xmath30 : this is likely to stem from the orientational entropy of the plastic crystal .
interestingly , in addition to the two triple points ( @xmath13-qc - z and qc - z - f ) , there is a range of pressures where the stable thermodynamic phase changes from f to z to qc to @xmath13 as the system is cooled .
we also look at the effect of modifying potential parameters on the phase diagram , thus allowing us to assess the extent of quasicrystal stability and how finely - tuned the particle properties would need to be to observe a quasicrystalline phase in experiment . to determine the phase diagram as the potential parameters change , we additionally use hamiltonian gibbs
duhem integration , whereby we numerically integrate a clapeyron equation that has been generalised to allow for changes to the potential itself @xcite .
we first investigate the behaviour of the system as the patch width is varied . from the phase diagram in fig .
[ fig - quasicrystal - phasediags - all](b ) , we see that the quasicrystal is only stable for a limited range of @xmath7 , although the range may perhaps be wider at other pressures . as the patch width narrows , the quasicrystal becomes increasingly enthalpically destabilised with respect to the @xmath13 phase , as the many six - fold environments can not satisfactorily fulfil patch - patch interactions .
by contrast , as the patch width increases , the quasicrystal becomes more stable with respect to the @xmath13 phase , but the hexagonal phase is stabilised even more : as the patches are wider and closer to the isotropic case , six - fold environments are enthalpically preferred .
another noteworthy feature of this phase diagram is that below @xmath31 , @xmath32 begins to decrease : as the patches become narrower , there is an increasing energetic penalty for patches that do not point directly at each other .
for very narrow patches , the @xmath13 phase transforms into a distorted @xmath13 phase ( d@xmath13 ) @xcite , in which three of the five patches point directly at the patches of neighbouring particles , thus breaking the square symmetry of the lattice .
the reason for this behaviour is that at narrow patch widths , only ` perfect ' connections count , as a slight misalignment of the patches rapidly reduces the interparticle attraction , and three perfect interactions become favoured over five imperfect ones .
because colloidal interactions are often quite short - ranged , we also wish to investigate what happens to the system when we change the lj exponent ( @xmath26 ) , which changes the range of the potential and the width of the potential well .
we show the phase diagram as a function of @xmath26 in fig .
[ fig - quasicrystal - phasediags - all](c ) .
the effects of increasing @xmath26 on isotropic potentials are well understood @xcite : the liquid phase is energetically destabilised because the intrinsic disorder in interparticle neighbour distances is penalised by the narrower potential @xcite . in fig .
[ fig - quasicrystal - phasediags - all](c ) , the range of stability of the quasicrystal initially increases .
the enthalpies of the qc and @xmath13 phases increase only slightly as @xmath26 increases , because their relative order means that most interparticle neighbour distances can be close to optimal , but do so more for the quasicrystal , as some disorder is typically present in the quasicrystal configurations ( for example , particles with unclassified environments in fig . [ fig - quasicrystal](d ) ) ; the @xmath13-qc coexistence temperature does therefore increase slightly .
by contrast , the fluid is initially much more enthalpically penalised : the fluid enthalpy increases rapidly , and the fluid density decreases as @xmath26 increases . at the maximum in @xmath33 ( @xmath34 ) ,
the size of the temperature window of quasicrystal stability is three times that of the @xmath4 lj potential .
however , a narrower potential well associated with increasing @xmath26 also means that there is less vibrational entropy associated with the ordered structures , and beyond @xmath34 , this effect becomes dominant .
the qc - f coexistence temperature decreases until the quasicrystal loses stability at @xmath35 .
the potential at the higher values of @xmath26 shown in fig .
[ fig - quasicrystal - phasediags - all](c ) is very short - ranged indeed .
the fact that the quasicrystal maintains and even increases its window of stability for ranges of attraction typical of colloidal particles suggests that this is a robust , general result and that stable quasicrystals can be expected to form in experimental realisations of this system .
there are several potential approaches to such experimental realisations .
if patchy colloidal particles similar to the ones studied here could be created , it would be reasonable to expect that a quasicrystal may form when they are confined in two dimensions .
for example , chen _ et al .
_ observed the formation of a 2d kagom lattice from tri - block janus particles by introducing a density mismatch with the solvent to confine their colloidal system into two dimensions @xcite .
a possible alternative to using colloidal patchy particles might involve the use of dna multi - arm motifs @xcite , for which a variety of 2d crystalline arrays have been observed for motifs with different numbers of arms , including a @xmath13 phase for five - arm motifs that is analogous to what we see in the current model when patches are narrow .
however , such dna motifs have a well - defined ` valence ' , and there is no equivalent of the patch width that could be tuned to make two co - ordination numbers compete .
therefore , a possible approach to forming a dna quasicrystal might be to use a two - component mixture of five - arm and six - arm motifs of the appropriate composition . in summary
, we have shown how we can compute if and where a quasicrystalline phase is thermodynamically stable . to the best of our knowledge ,
this is the first such calculation of the chemical potential of a quasicrystal and the associated phase diagrams that has been obtained directly from simulations with no approximations . for our patchy particle system
, we found that the quasicrystalline phase is stable over a significant portion of the phase diagram and is stabilised primarily by its configurational entropy .
it is robust to parameter changes in the model , which inspires confidence that such a thermodynamically stable quasicrystal might be experimentally realised .
we wish to thank the engineering and physical sciences research council for financial support .
69ifxundefined [ 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 |
the presence of the vortex core does not only introduce the gyrotropic mode , but also modifies the high - frequency spin wave normal mode spectrum @xcite by lifting the degeneracy of the azimuthal modes with opposite rotation sense @xcite .
the out - of - plane component of the magnetization @xmath3 of these azimuthal spin waves can be described in polar coordinates @xmath4 as @xcite : @xmath5 these modes are characterised by a radial mode number @xmath6 ( positive integer values starting with 1 ) denoting the number of nodes of the radial part within the disk and an azimuthal mode number @xmath7 ( integer values ) , with @xmath8 being the number of nodes in the azimuthal part .
the sign of @xmath7 denotes the rotation sense of the azimuthal wave ( ccw : @xmath7 positive , cw : @xmath7 negative ) .
the symmetry in the frequency splitting of the azimuthal mode can be written as @xmath9 .
a general concept of resonant vortex core reversal by rotating magnetic fields is sketched in fig .
[ fig1 ] . on the left - hand side ,
the gyrotropic mode ( @xmath10 , @xmath11 ) , is shown with a ccw sense of rotation for a core up state , and a cw sense of rotation for a core down state .
the vortex core can only be reversed resonantly by applying an external rotating magnetic field if its frequency and sense of rotation correspond to the mode , as already shown experimentally in @xcite . on the right hand side of fig .
[ fig1 ] , the frequency - split azimuthal modes are shown . for a vortex up
this means that the mode with a cw sense of rotation , denoted by ( @xmath6 , @xmath12 ) , has a lower frequency and the mode with a ccw sense of rotation , denoted by ( @xmath6 , @xmath13 ) , has a higher frequency . for a vortex core down the sign of the azimuthal mode number @xmath7 , indicating the sense of rotation , is reversed for the higher and lower frequency modes compared to vortex core up , as sketched at the bottom part of fig .
these spin wave modes with ( @xmath6 , @xmath14 ) can be selectively excited by the application of rotating in - plane magnetic fields with the corresponding frequency and rotation sense .
thus at a fixed ghz frequency and core polarity the excitation is polarization selective in the same way as for the sub - ghz gyrotropic mode : only rotating fields with one rotation sense will pump the system .
this suggests that a similar polarization selective switching mechanism for the vortex core may exist when any of the azimuthal spinwave modes are excited , albeit at much higher frequencies .
we excited ferromagnetic permalloy discs , 1.6@xmath15 m in diameter , 50 nm thick , in the vortex magnetic ground state with bursts of ghz rotating magnetic fields of variable frequency and amplitude .
the fields were applied with either a clockwise ( cw ) or counterclockwise ( ccw ) rotation sense @xcite .
magnetic scanning soft x - ray microscopy at the maxymus endstation , helmholtz zentrum berlin , bessy ii was used to determine the out - of - plane magnetization of the vortex core . the experimental verification of the concept as described before is presented in fig . [ fig2 ] ( upper part ) .
it shows the vortex core switching events from the initial vortex core up state to the down state as a function of excitation frequency , amplitude @xmath16 and sense of rotation of the external field . here
, the burst length was fixed at 24 periods of the excitation frequency .
three regions with a minimum in amplitude for vortex core reversal can be observed in the experimental data .
dependent on the sense of rotation of the external magnetic field they are marked with blue or red backgrounds .
the regions for cw excitation at about 4.5ghz and 6.5ghz , ccw excitation at about 5.5ghz can be assigned to the ( @xmath17 , @xmath12 ) , ( @xmath18 , @xmath12 ) and ( @xmath17 , @xmath19 ) spin wave modes , respectively ( cf . fig .
[ fig1 ] and the middle part of fig .
[ fig2 ] ) . in order to further investigate the core reversal at these different modes
, we performed micromagnetic simulations @xcite for the same sample size and experimental conditions .
the results are shown in fig .
[ fig2 ] , lower part and are in good agreement with our experimental results .
the small differences between the experimental and simulated results can be attributed to the uncertainty in the thickness , saturation magnetization and damping coefficient of the material , as well as to possible structural imperfections in the samples .
the phases of a local fourier analysis @xcite of the dynamic magnetization let us clearly identify the azimuthal modes ( cf .
[ fig2 ] , lower right part ) .
the simulations with an extended frequency range up to 10.5ghz reveal even more spin wave eigenmodes capable of vortex core reversal : the ( @xmath18 , @xmath19 ) mode at about 9ghz and the two ( @xmath20 ) spin wave modes at about 10 and 10.3ghz .
both the experiments and the micromagnetic simulations show that resonant switching of the vortex core from the down state back to the up state at the same frequency can be achieved only if the sense of rotation of the external field is reversed .
this allows selective and unidirectional switching at each resonance frequency . detailed analysis of the micromagnetic simulations reveal that the excitation of all spinwave modes also leads to the creation of a dip near the vortex core , resulting in the same core reversal mechanism as found for gyrotropic excitation @xcite .
it is based on the creation and annihilation of a vortex - antivortex pair .
but the dip creation itself differs dependent on which spin wave mode is excited .
this dip creation for a vortex up ( @xmath21 ) is illustrated by snapshots of the simulations in fig .
[ fig7 ] for the modes ( @xmath17 , @xmath22 ) . in the first frame ( 0 )
, the unperturbed out - of - plane magnetization of the vortex can be seen .
after a short time of excitation ( 135 phase angle ) the azimuthal spin wave with the same rotation sense as the external field develops .
the interaction of the spin waves with the gyrotropic mode leads to a displacement of the vortex core towards the oppositely polarized region of the bipolar mode , resulting in a gyration with the same frequency as the external field ( 225 phase angle ) . at 405 ,
a negatively polarized out - of - plane region can be observed next to the core , to the outside with respect to the centre .
it is gyrating around the centre almost in phase with the vortex core .
in the ( @xmath12 ) case ( upper panels ) , this region develops into a dip which finally leads to core reversal .
but surprisingly , for the ( @xmath19 ) in the lower panels , there is an additional negatively polarized region .
this region is situated towards the structure centre where the azimuthal mode has a positive polarization and gyrates with a phase shift of about 180 with respect to the vortex core . as can be seen from the last frame at 630 , it is this region which evolves into a fully out - of - plane dip , creating the vortex - antivortex pair for switching .
the typical magnetization configurations are shown in fig .
[ fig3 ] including the higher radial mode numbers .
the trajectories for the vortex core as well as for the dips and their relative phases are illustrated by arrows .
the gyrofield , as used by guslienko _
_ to describe the dip formation in the gyrotropic mode @xcite , can also be taken to explain the difference in dip formation for the ( @xmath19 ) and ( @xmath12 ) modes , as is sketched in fig .
this effective field originates from the movement of the vortex and results in out - of - plane contributions to the left and right side of the core ( in respect to the direction of motion ) which are respectively negative and positive ( middle sketch in fig .
[ fig8 ] ) .
this velocity dependent behaviour breaks the symmetry between counter rotating vortices .
if we now combine the spin wave amplitude with the effect of the gyrofield , we can understand the dip formation for any of the counter rotating modes ( right sketch in fig .
[ fig8 ] ) . for a vortex up and
the mode ( @xmath12 ) with a cw vortex motion , the gyrofield acts constructively on the spin wave amplitude .
for the ( @xmath19 ) mode , it acts destructively , causing the double dip configuration as seen in fig .
alternatively , these dip structures can be regarded as a hybridisation between the gyrotropic mode and the azimuthal modes @xcite .
_ @xcite also presented a dip formation mechanism for spin wave excitation under the condition of an artificially fixed vortex core .
however , our results show that this condition can not be used to predict the reversal in free vortices , as in realistic structures . according to their analysis ,
the polarization direction of the dip is given by the rotation sense of the external field only .
this would mean that reversal of a given core polarization is only possible for one of the two azimuthal modes .
e.g. the reversal at the ( @xmath19 ) as presented in fig .
[ fig2 ] can not be explained with this assumption , as the dip polarization would be the same as the core polarization .
the dynamic origin of the dip structure clearly explains this contradiction . for a free vortex ,
the dip formation seems to be dominated by the gyrofield .
vortex core reversal by the low - frequency gyromode excitation results in a typical gyration radius of about 300 nm for this kind of sample .
in contrast , the gyration radius of the vortex core at ghz frequencies is only in the order of 10 nm ( cf .
[ fig3 ] ) .
but due to the high - frequency rotation , this small radius still results in a very high velocity of the vortex core . the top part of fig .
[ fig5 ] shows the velocities of the vortex core directly before the reversal process as calculated from the simulations .
it varies between about 100 m / s ( @xmath12 ) and up to more than 600 m / s ( @xmath19 ) .
this demonstrates that for ghz spin wave excitation , no well - defined critical velocity is observed as it was calculated to be about 320 m / s for the gyromode of a permalloy nanodot @xcite .
the reason can be found in our explanation of the dip formation as a combination of the gyrofield with the spin wave amplitude .
although in first order , the gyrofield is proportional to the vortex velocity , the spin wave amplitude is not and thus the critical velocity before switching is higher or lower than the expected value , depending if their combination is constructive or destructive .
similar arguments were used to explain the changes in critical velocity when perpendicular bias fields are applied @xcite .
the large fluctuations in the velocities can be understood when the random contribution of short wavelength spin waves are also taken into account .
consequently , the velocity itself is not a switching criterion for the spin wave mode induced vortex core reversal . the bottom part of fig .
[ fig5 ] shows how long it takes for the vortex to switch after the burst is applied .
it is no surprise that the larger the amplitude , the shorter the time for vortex core reversal .
the minima in the excitation amplitude show a clear resonant behaviour . at high amplitudes
the switching times become smaller than one period of the excitation and the frequency dependence is washed out , indicating the transition to non - resonant excitation @xcite .
however , the feature of unidirectionality gets lost in the intersection of two oppositely rotating modes if too short bursts are applied .
because both the vortex and the created vortex - antivortex pair remain close to the centre of the structure during the excitation , the relaxation after the spin wave induced reversal process is much faster than in the gyrotropic case , where it has to spiral back from its orbit far away to the centre . in combination with the very small gyration radius
, this fast relaxation is accelerating the whole reversal process by up to two orders of magnitude , while the total energy pumped into the system is roughly the same compared to reversal at the gyrotropic mode .
this fact makes this new reversal process very beneficial for fast technological applications . in conclusion ,
our experiments show vortex core reversal at ghz frequencies . by taking advantage of the vortex polarity dependent frequency splitting of the azimuthal modes ( @xmath6 , @xmath14 ) we present a concept for unidirectional reversal by exciting any of these modes up to 10 ghz .
although the dynamics depend significantly on the different combinations of core polarity and azimuthal mode number , they can be explained by the superposition of the gyrofield with the spin wave amplitude .
consequently , immediately before the switching event , different vortex core velocities are observed , which excludes the velocity as a critical condition for core reversal at ghz excitation .
this finding highlights the importance of spin wave vortex interaction and boosts vortex core reversal to much higher frequencies , which may offer new routes for spintronics applications .
permalloy discs , 1.6@xmath15 m in diameter and 50 nm thickness , are grown on top of 2.5@xmath15 m wide and 170 nm thick copper striplines patterned on a 100 nm thick silicon nitride membrane .
all structures are prepared by electron - beam lithography , thermal evaporation and lift - off processes .
ac currents with a phase shift of @xmath2390 degree flowing through orthogonal striplines are generating rotating in - plane magnetic fields @xcite .
the out - of - plane magnetization of the permalloy disc was imaged by scanning transmission x - ray microscopy at the maxymus end station at hzb bessy ii , berlin , by taking advantage of the x - ray magnetic circular dichroism ( xmcd ) @xcite .
the instrument has a resolution of about 25 nm , which is sufficient to resolve the vortex core and its polarization .
micromagnetic simulations were performed with the object oriented micromagnetic framework ( oommf ) @xcite .
a disc , 1.62@xmath24 in diameter and 50 nm thick , was simulated with a cell size of 3nm@xmath253nm@xmath2550 nm .
typical material parameters for permalloy @xcite were used : @xmath26=750@xmath2510@xmath27a / m , exchange constant @xmath28=13@xmath2510@xmath29j / m , damping parameter @xmath30=0.006 , anisotropy constant @xmath31 . for all simulations shown , the initial state was a vortex core up .
the symmetry of the results were verified by also simulating with the opposite initial state which is a vortex down .
the sample was excited by homogeneous in - plane rotating magnetic field bursts with duration of 24 periods .
the frequency of the bursts was scanned from 3ghz to 11ghz in 0.4ghz steps for both , cw and ccw sense of rotation .
we would like to thank manfred fhnle , mpi stuttgart , franz mertens , bayreuth university , yuri gaididei and volodymyr kravchuk , bogolyubov institute for theoretical physics , kiev , and denis sheka , kiev university , for fruitful discussions .
we would also like to thank all people involved in the construction and operation of the maxymus scanning x - ray microscope at hzb bessy ii in berlin , in particular brigitte baretzky , michael bechtel , rolf follath , corinne grvent and eberhard goering . cooperation with rolf heidemann , ulm university , in the stripline design for microwave application is gratefully acknowledged .
project planning : m.k .
, g.s . ; sample preparation : g.w .
, c.h.b . ;
experiments : m.k .
micromagnetic simulations : m.k . ; analyse data : m.k .
, b.v.w . , h.s .
; writing the paper : m.k .
, b.v.w . ,
h.s . , g.w .
, g.s .
the authors declare that they have no competing interests as defined by nature publishing group , or other interests that might be perceived to influence the results and/or discussion reported in this paper . and the azimuthal mode number ( @xmath32 ) , denoting the sense of rotation of the eigenmode . in vortex structures
the symmetry is broken by the out - of - plane component of the core and thus a frequency splitting is observed between ( @xmath33 ) and ( @xmath13 ) modes . ,
scaledwidth=50.0% ] , a frequency @xmath34 and a duration of 24 periods have been applied .
like indicated in the legend of the top panel with the recorded x - ray images ( c ) , the blue triangles with a dot in the middle indicate vortex core switching only after a cw rotating field burst , while red triangles indicate switching only after a ccw field burst .
black dots indicate no switching for either rotation sense .
the minima in the switching threshold with differing sense of rotation correspond to the resonance frequencies of the excited azimuthal spin wave modes with the same sense of rotation as sketched in the middle part of fig .
the modes are identified with the help of the phases derived from a local fast fourier transform of the simulated out - of - plane magnetization of the sample with @xmath35 in diameter as shown in the inset to the right of the bottom panel ( d ) .
, scaledwidth=40.0% ] m disc before the core reversal .
( a ) a section of 300 nm in diameter showing the excited azimuthal spin wave modes as well as the dip. the dynamics of modes with the same rotation sense are similar in the number of dips and in their phase relation to the vortex core .
this is indicated by the arrows below the snapshots .
the difference between oppositely rotating modes are the result of the symmetry breaking due to the gyrofield .
( b ) for comparison the central part of the same structure with 500 nm in diameter is given , excited at the gyrotropic resonance .
this results in a much larger trajectory of the core and the dip. the black bars correspond to a lateral size of 100 nm in both cases ( a , b ) .
, scaledwidth=50.0% ] | micron - sized magnetic platelets in the flux closed vortex state are characterized by an in - plane curling magnetization and a nanometer - sized perpendicularly magnetized vortex core .
having the simplest non - trivial configuration , these objects are of general interest to micromagnetics and may offer new routes for spintronics applications .
essential progress in the understanding of nonlinear vortex dynamics was achieved when low - field core toggling by excitation of the gyrotropic eigenmode at sub - ghz frequencies was established . at frequencies more than an order of magnitude higher vortex state structures possess spin wave eigenmodes arising from the magneto - static interaction . here
we demonstrate experimentally that the unidirectional vortex core reversal process also occurs when such azimuthal modes are excited .
these results are confirmed by micromagnetic simulations which clearly show the selection rules for this novel reversal mechanism .
our analysis reveals that for spin wave excitation the concept of a critical velocity as the switching condition has to be modified .
= 1 thin film soft magnetic platelets of suited micron or submicron size are characterized by an in - plane closed flux magnetization , minimizing the dipolar energy . however , at the centre , the exchange energy forces the magnetization out - of - plane in a small area of only a few exchange lengths in diameter @xcite creating the vortex core with a distinct polarization @xmath0 , either up ( @xmath1 ) or down ( @xmath2 ) .
the vortex introduces an important sub - ghz mode to the excitation spectrum of the magnetization @xcite , which corresponds to the gyrotropic motion of the vortex .
recently , dynamic vortex core reversal was discovered by application of pulsed magnetic fields @xcite , alternating magnetic fields @xcite or spin polarized currents @xcite .
these discoveries did not only open up the possibility of using magnetic vortices as memory bits and further spintronics applications @xcite , but also initiated wide investigation of the physics behind this reversal mechanism .
the reversal process appeared to be universal and independent of the type of excitation : it proceeds through the creation and subsequent annihilation of a vortex - antivortex pair @xcite .
the source of this pair is the dynamical distortion of the otherwise stable vortex : it forms a dip an oppositely polarized region close to the moving vortex core @xcite .
guslienko _
et al .
_ @xcite explained this dip-formation by an effective field , the gyrofield , which is proportional to the vortex core velocity .
this leads to the concept of a critical velocity for switching , which is about 320 m / s in permalloy @xcite .
furthermore , the sense of rotation of a freely gyrating vortex is determined by the vortex core polarization @xcite , a property which manifests itself in a high asymmetry of the gyrotropic resonance excited by rotating magnetic fields with different rotation senses @xcite .
this can be utilized to selectively reverse the core polarization , as shown experimentally by curcic _
et al .
_ @xcite .
micromagnetic simulations from kravchuk _ et al .
_ @xcite have indicated that rotating magnetic fields can even induce vortex core reversal at ghz frequencies as a result of the vortex spin wave interaction . here
we present experiments which verify that polarisation selective vortex core reversal can indeed be achieved by spin wave excitation with in - plane rotating ghz magnetic fields .
however , compared to the results of kravchuk et al .
@xcite where only one spin wave mode was consiered , we will present selection rules for different modes . and
in contrast to the calculations given by kravchuk et al .
@xcite where the vortex core was artificially fixed during spin wave excitation , we will demonstrate by micromagnetic simulations that the core motion is essential and explains the differences observed in the vortex core switching triggered by distinct spin waves . |
one of the outstanding problems in particle physics is the determination of the structure of hadrons such as the proton and neutron in terms of their fundamental quark and gluon degrees of freedom . over the past twenty years
two fundamentally different pictures have developed .
one , the constituent quark model is closely related to experimental observation . the other ,
quantum chromodynamics is based on a covariant non - abelian quantum field theory .
the front form ( also known as light - cone quantization ) appears to be the only hope of reconciling these two .
this elegant approach to quantum field theory is a hamiltonian gauge - fixed formulation that avoids many of the most difficult problems in the conventional equal - time formulation of the theory .
the natural gauge for light - cone hamiltonian theories is the light - cone gauge @xmath0 . in this physical gauge
the gluons have only the two physical transverse degrees of freedom .
one imagines that there is an expansion in multi - particle occupation number fock states .
but even in the case of the simpler abelian quantum theory of electrodynamics very little is known about the nature of the bound state solutions in the strong - coupling domain . in the non - abelian quantum theory of chromodynamics a calculation of bound - state structure
has to deal with many difficult aspects simultaneously .
confinement , vacuum structure and chiral symmetry inter - twine with the difficulties of describing a ( relativistic ) many - body system and the non - perturbative renormalization of a hamiltonian . in the conventional approach based on equal - time quantization
the fock state expansion becomes quickly intractable because of the complexity of the vacuum .
furthermore , boosting such a wavefunction from the hadron s rest frame to a moving frame is as complex a problem as solving the bound state problem itself .
the presence of the square root operator in the equal - time hamiltonian approach presents severe mathematical difficulties .
fortunately ` light - cone quantization ' offers an elegant avenue of escape .
it can be formulated independent of the lorentz frame .
the square root operator does not appear , and the vacuum structure is relatively simple .
there is no spontaneous creation of massive fermions in the light - cone quantized vacuum .
in fact , there are many reasons to quantize relativistic field theories at fixed light - cone time .
dirac @xcite showed , in 1949 , that in this so called ` front form ' of hamiltonian dynamics a maximum number of poincar generators become independent of the interaction , including certain lorentz boosts .
in fact , unlike the traditional equal - time hamiltonian formalism , quantization on a plane tangential to the light - cone , on the ` null plane ' , can be formulated without reference to a specific frame .
one can construct an operator whose eigenvalues are the invariant masses of the composite physical particles .
the eigenvectors describe bound states of arbitrary four - momentum and invariant mass , and allow the computation of scattering amplitudes and other dynamical quantities . in many field theories
the vacuum state of the free hamiltonian is also an eigenstate of the light - cone hamiltonian .
the fock expansion built on this vacuum state provides a complete relativistic many - particle basis for diagonalizing the full theory .
the main thrust of these lectures will be to discuss the complexities that are unique to this formulation of qcd , in varying degrees of detail .
the goal is to present a self - consistent framework rather than trying to cover the subject exhaustively .
a review all of the successes or applications will , however , not be undertaken .
one of the reasons is , that the subject was reviewed recently @xcite .
another is that other lecturers in this school emphasize complementary aspects .
stan brodsky shows how the knowledge of the the light - cone wavefunction has impact on hadronic physics and exclusive processes .
steve pinsky demonstrates how the method of discretized light - cone quantization is constructive for analyzing super - symmetric string and m(atrix ) theories .
last not least , simon dalley expands on the transverse - lattice calculations within the light - cone approach .
comparatively little space will be devoted to canonical field theory , just so much as to plausibilize that a light - cone hamiltonian exists .
this so called ` naive hamiltonian ' will be derived and written down explicitly as a fock - space operator in the light - cone gauge and disregarding zero modes . for historical and paedagogical reasons
these notes will be rather outspoken for one space and one time dimension , mostly to show that periodic boundary conditions ( discretized light - cone quantization ) are helpful , indeed , for solving the bound - state problem in a relativistic theory .
the attempt is made to be complementary to the review @xcite and some new material on the schwinger model is included . the bulk of these notes deals with the many - body aspects of a gauge field theory in the real world of 3 + 1 dimensions .
a hadron not only contains the valence quarks but also an infinite amount of gluons and sea - quarks .
progress often comes with new technologies .
the presentation of the method of iterated resolvents therefore takes broad room .
it allows to derive a well - defined effective interaction . some thus far unpublished work
is included , in particular more instructive examples .
the presentation is separated into two parts . in the first considerations are essentially exact . in the second
some approximations and simplifications are admitted and well marked in the text .
one arrives at the effective interaction in the form of a tractable and solvable integral equation .
its solutions allow to construct explicitly the many - body amplitudes corresponding to see - quarks and gluons by comparatively simple quadratures , without the need of solving another bound - state problem .
explicit equations for that are given .
moreover , some new research work will be included in the section on renormalization . what is not included in these notes
, however , is a complete survey of the literature .
mostly for the reasons of space , it is refered to the some 469 items of @xcite .
i apologize with my colleagues whose work is not mentioned .
but i will be careful to cite the work which i need for the present discussion and presentation .
some selected monographies which i found useful to consult are @xcite , some selected review articles or conference proceedings might be @xcite .
dirac defines the hamiltonian @xmath1 of a closed system as that operator whose action on the state vector @xmath2 has the same effect as taking the partial derivative with respect to time @xmath3 , _ i.e. _ @xmath4 the concept of an hamiltonian is applicable irrespective of whether one deals with the motion of a non - relativistic particle in classical mechanics or with a non - relativistic wave function in the schrdinger equation , and it generalizes almost unchanged to a relativistic and covariant field theory . in a covariant theory ,
the very notion of ` time ' becomes , however , questionable since the time is only one component of four - dimensional space - time .
but the concept of usual space and of usual time can be generalized .
one can define ` space ' as that hypersphere in four - space on which one chooses the initial conditions .
the remaining fourth coordinate can be understood as ` time ' .
more formally , one conveniently introduces generalized coordinates @xmath5 . starting from a baseline parametrization of space - time like the instant form in figure [ fig : bir2 ] , one parametrizes space - time by a coordinate transformation @xmath6 .
the metric tensors for the two parametrizations are then related by @xmath7 the physical content of the theory can not depend on such a re - parametrization .
but the raising and the lowering of lorentz indices are then non - trivial operations . as an example consider the front form parametrization in figure [ fig : bir2 ] .
interpret @xmath8 as the new space coordinate and denote it by @xmath9 .
then @xmath10 must be interpreted as the new time coordinate denoted by @xmath11 , or by the ` light - cone time ' @xmath12 .
of course , one also could exchange the two . since the lowering operation is @xmath13 , both @xmath14 are space , and @xmath15 are time coordinates .
the new space derivative is therefore @xmath16 , while @xmath17 is a time - derivative .
the lorentz indices ` + ' and ` - ' have a different physical meaning , depending on whether they occur up- or down - stairs .
co - variant and contra - variant vectors are different objects . the hamiltonian is only one component of a four - vector @xmath18 , particularly its time - like component .
taking the partial time derivative of the state vector @xmath19 defines then @xmath20 as the hamiltonian in the transformed coordinates , in line with eq.([eq:1 ] ) , and @xmath21 as the longitudinal momentum .
following dirac @xcite there are no more than three different parametrisations of space - time .
they are illustrated in figure [ fig : bir2 ] , and can not be mapped onto each other by a lorentz transform .
they differ by the hypersphere on which the fields are initialized .
they have thus different ` times ' and different ` hamiltonians ' .
dirac @xcite speaks of the three _ forms of hamiltonian dynamics _ : the _ instant form _ is the familiar one , with its hypersphere given by @xmath22 . in the _ front form _
the hypersphere is a tangent plane to the light cone . in the _ point form _
the time - like coordinate is identified with the eigentime of a physical system and the hypersphere has a shape of a hyperboloid .
which of the three forms should be prefered , is an ill - posed question , since all three forms must yield the same physical results .
comparatively little work has been done in the point form .
the bulk of research on field theory implicitly uses the instant form . although it is the conventional choice for quantizing field theory , it has many practical disadvantages .
for example , given the wavefunctions of an @xmath23-electron atom at an initial time @xmath22 , @xmath24 , one can use the hamiltonian @xmath25 to evolve @xmath26 to later times @xmath27 .
however , an experiment which specifies the initial wave function would require the simultaneous measurement of the positions of all of the bounded electrons .
in contrast , determining the initial wave function at fixed light - cone time @xmath28 only requires an experiment which scatters one plane - wave laser beam , since the signal reaching each of the @xmath23 electrons , along the light front , at the same light - cone time @xmath29 .
dirac s legacy had been forgotten and was re - invented several times .
the front form approach carries therefore names as different as _ infinite - momentum frame _ , _ null - plane quantization _
, _ light - cone quantization _ , or most recently _ light - front quantization_. in the essence they are all the same .
the infinite - momentum frame is a misnomer since the total momentum is finite and since the front form is _ frame - independent _ and covariant .
light - cone quantization is also unfortunate , since the inital data are set one a plane tangential to but not on the light cone , and since both equal - time or equal light - cone - time quantization stand both for the same quantization , for a quantum as opposed to a classical theory .
we propose to stay with dirac s different ` forms of hamiltonians ' .
the prototype of a field theory is faraday s and maxwell s electrodynamics .
every field theory has its own canonical hamiltonian and is governed by the action density .
the lagrangian is the subject of the canonical calculus of variation , given in many text books @xcite .
its essentials shall be recalled briefly .
the lagrangian , in general , is a function of a finite number of fields @xmath30 and their first space - time derivatives @xmath31 , thus @xmath32 $ ] .
independent variation of @xmath33 with respect to @xmath34 and @xmath35 results in the equations of motion , @xmath36 \equiv { \delta { \cal l } \over \delta \left(\partial _
\kappa \phi _
r \right ) } , \nonumber % \label { eq:2.9}\ ] ] canonically refered to as the euler equations .
the canonical formalism is particularly suited for discussing the symmetries of a field theory .
every continuous symmetry of the lagrangian is associated with a vanishing four - divergence of a current and a conserved charge . since @xmath33 does not explicitely depend on the coordinates , every field theory in 3 + 1 dimensions
has ten conserved four - currents .
the four - divergences of the energy - momentum tensor @xmath37 and of the boost - angular - momentum stress tensor @xmath38 vanish , @xmath39 as a consequence the lorentz group has ten ` conserved charges ' , the 4 components total momentum @xmath40 and the 6 components of boost - angular momentum @xmath41 , @xmath42 with the totally antisymmetric tensor in four dimensions @xmath43 ( @xmath44 ) , the three - dimensional surface elements of a hypersphere are @xmath45 .
a finite volume is thus @xmath46 , and correspondingly in the front form @xmath47 .
the time - like components of @xmath48 is the hamiltonian , _
i.e. _ @xmath49 for the instant form and @xmath50 for the front form .
the transition from instant to front form is thus simple : substitute ` 0 ' by ` + ' .
working out the canonical procedure for qed with its lagrangian , @xmath51 , \label{eq:2.7}\ ] ] where @xmath52 is the electro - magnetic field tensor and @xmath53 the covariant derivative , @xmath54 one ends up with a manifestly gauge - invariant total momentum , @xmath55 \biggr ) \ , \nonumber \\ p_\nu & = & \int_\omega \ !
+ \ \biggl ( f^ { + \kappa } f _ { \kappa\nu } + { 1\over4}g^+_\nu f^ { \kappa\lambda } f_{\kappa\lambda } + { 1\over2}\bigl[i\overline\psi \gamma^+ d _
\nu \psi +
\ { \rm h.c . } \bigr ] \biggr ) , \nonumber % \label { eq:2.53 } \end{aligned}\ ] ] in both the instant and the front form , respectively . the boost angular momenta will not be used explicitly .
the gauge invariant lagrangian density for qcd is @xmath56 , \nonumber \\ & = & - { 1\over4 } f^{\mu\nu}_a f_{\mu \nu}^a + { 1\over2 } \bigl [ \overline \psi \bigl(i\gamma^\mu { \bf d } _ \mu - { \bf m } \bigr ) \psi + \ { \rm h.c . }
\bigr ] , \nonumber % \end{aligned}\ ] ] the color - electro - magnetic fields and the covariant derivative are now @xmath57 & = & { \bf t}^a\left ( \partial^\mu a^\nu_a - \partial^\nu a^\mu_a - g f^{ars } a^\mu_r a^\nu_s \right ) , \nonumber\\ { \bf d } ^\mu _ { cc^\prime } = \delta_{cc^\prime } \partial^\mu + i g { \bf a}^\mu _ { cc^\prime } & = & \delta_{cc^\prime } \partial^\mu + i g t^a_{cc^\prime } a^\mu _ { a } .\nonumber % \label { eq:2.31}\end{aligned}\ ] ] as compared to qed , each local gauge field @xmath58 is replaced by the @xmath59 matrix @xmath60 .
all such matrices can be parametrized @xmath61 . more generally for su(n ) , the vector potentials @xmath62 are hermitian and traceless @xmath63 matrices .
the color index @xmath64 ( or @xmath65 ) runs now from 1 to @xmath66 , and correspondingly the gluon index @xmath67 ( or @xmath68 ) from 1 to @xmath69 .
no distinction will be made between raising or lowering them . in order to make sense of expressions like @xmath70 the quark fields @xmath71 must carry a color index @xmath64 .
they , as well as the dirac indices , are usually suppressed .
the color matrices @xmath72 obey @xmath73 _ { cc ^\prime } = i f ^{rsa } t _ { cc ^\prime } ^a \qquad\quad { \rm and } \qquad { \rm tr}\ \bigl(t ^r t ^s \bigr ) = { 1\over2}\delta _ r ^s .\label{eq:2.26}\ ] ] for su(2 ) the color matrices are @xmath74 , with @xmath75 being the pauli matrices .
the structure constants @xmath76 are therefore the totally antisymmetric tensor @xmath77 . for su(3 ) , @xmath78 with the gell - mann matrices @xmath79 , with the corresponding structure constants tabulated in the literature @xcite .
everything proceeds in analogy with qed .
the energy - momentum vector , @xmath80 \biggr ) , \nonumber \\
p_\nu & = & \int_\omega \ !
+ \ \biggl(f^{+\kappa}_a f ^a_{\kappa\nu } + { 1\over4 } g^+_\nu f^{\kappa\lambda}_a f _ { \kappa\lambda}^a + { 1\over2}\bigl[i\overline\psi\gamma^+ t^a d^a_\nu\psi + \ { \rm h.c . } \bigr ] \biggr ) , \label { eq:2.60 } \end{aligned}\ ] ] is manifestly gauge - invariant both in the instant and the front form .
the ten constants of motion @xmath81 and @xmath82 are observables with real eigenvalues .
it is advantageous to construct representations in which the constants of motion are diagonal .
but one can not diagonalize all ten constants of motion simultaneously because they do not commute .
the algebra of the four - energy - momentum @xmath83 and four - angular - momentum @xmath84 for free particles with the basic commutator @xmath85 = \delta ^\mu _ \nu $ ] is @xmath86 & = & g ^{\rho \mu } p ^\nu - g ^{\rho \nu } p ^\mu \ , \quad { 1 \over i \hbar } \left [ p ^\rho , p ^\mu \right ] = 0 \ , \nonumber \\ \quad { \rm and }
\ \quad { 1 \over
i \hbar } \left [ m ^{\rho \sigma } , m ^{\mu \nu } \right ] & = & g ^{\rho \nu } m ^{\sigma \mu } + g ^{\sigma \mu } m ^{\rho \nu } - g ^{\rho \mu } m ^{\sigma \nu } - g ^{\sigma \nu } m ^{\rho \mu } .\nonumber % \label{eq : group}\end{aligned}\ ] ] it is postulated that the generalized momentum operators satisfy the same commutator relations as a single particle .
they form thus a group , the poincar group .
it is convenient to discuss the structure of the poincar group in terms of the pauli - lubansky vector @xmath87 .
@xmath88 is orthogonal to the generalized momenta , @xmath89 .
the two group invariants are the operator for the invariant mass - squared @xmath90 and the operator for intrinsic spin - squared @xmath91 .
they are lorentz scalars and commute with all generators @xmath92 and @xmath93 , as well as with all @xmath94 .
a convenient choice of six mutually commuting operators is : the invariant mass squared @xmath95 , the three space - like momenta @xmath96 and @xmath97 , the total spin squared @xmath98 , and one component of @xmath99 , say @xmath100 . inspecting the definition of boost - angular momentum @xmath101 in eq.([eq:2.17 ] )
one identifies which components are dependent on the interaction and which are not .
dirac @xcite calls them complicated and simple , or dynamic and kinematic , or hamiltonians and momenta , respectively .
in the instant form , the three components of the boost vector @xmath102 are dynamic , and the three components of angular momentum @xmath103 are kinematic .
as noted already by dirac , the front form is special in having four kinematic components of @xmath104 ( @xmath105 ) and only two dynamic ones ( @xmath106 and @xmath107 ) . in the front form one deals thus with seven mutually commuting operators @xmath108 instead of the six in the instant form .
these symmetries imply the very important aspect of the front form that both the hamiltonian and all amplitudes obtained in light - cone perturbation theory are manifestly invariant under a large class of lorentz transformations : @xmath109 _
i.e. _ for parallel boosts , transverse boosts , and rotations , respectively .
all of these hold for every single particle momentum @xmath110 , and for any set of dimensionless c - numbers @xmath111 and @xmath112 . the four - vector of energy - momentum for gauge theory in eq.([eq:2.60 ] ) contains time - derivatives and other constraint field components .
they will be eliminated in this section using the equations of motion with the goal to express @xmath40 in terms of the free fields and to isolate the dependence on the coupling constant .
the color - maxwell equations are four equations for determining the four functions @xmath113 .
one of the equations of motion is identically fullfilled by choosing the light - cone gauge @xmath114 .
two of the equations give expressions the time derivatives of the two transversal components @xmath115 .
the fourth is the gauss law in the front form , @xmath116 , or explicitly @xmath117 it contains only space - derivatives and is a constrained equation for the components of @xmath113 . for the free case ( @xmath118 )
, @xmath113 reduces to @xmath119 and therefore to @xmath120 \right ) .\nonumber % \ ] ] as a consequence , @xmath121 is purely transverse .
the formal inversion of eq.([eq:2.56 ] ) is therefore @xmath122 which must be substituted everywhere .
the inverse space derivatives @xmath123 and @xmath124 , used here and below , are actually green s functions .
since they depend only on @xmath125 , they are comparatively simple . the color - dirac equations @xmath126 can be used to express the time derivatives @xmath127 as function of the other fields .
after multiplication with @xmath128 one gets first @xmath129 with the usual dirac matrices @xmath130 and @xmath131 . in order to isolate the time derivative one introduces the projectors @xmath132 and projected spinors @xmath133 by @xmath134 multiplying the color - dirac equation once with @xmath135 and once with @xmath136 gives @xmath137 a coupled set of spinor equations .
the first of them contains a time derivative .
the second contains a space derivative and is a constraint equation .
the component @xmath138 must therefore be substituted everywhere .
the time derivative becomes then @xmath139 in analogy to the color - maxwell case one defines free spinors by @xmath140 contrary to the full spinor , @xmath141 is independent of the interaction . inserting the above expressions into eq.([eq:2.60 ] ) ,
the space - like components of @xmath40 become @xmath142 inserting them into @xmath50 gives rather lengthy expressions , which are conveniently written as a sum of five terms @xmath143 only the first term survives the limit @xmath144 , and therefore is called the free part of the hamiltonian , or its ` kinetic energy ' @xmath145 the vertex interaction @xmath146 is linear in the coupling constant and is the light - cone analogue of the conventional @xmath147-structures in the instant form .
note that the current @xmath148 has contributions from both quarks and gluons .
the four - point gluon interaction @xmath149 describes the four - point gluon - vertices which is quadratic in @xmath150 .
the remainders are the ` instantaneous interactions ' .
they are characterized by the inverse derivatives .
the instantaneous gluon interaction arises from the gauss equation , @xmath151 and is the light - cone analogue of the coulomb energy .
the instantaneous fermion interaction originates from the light - cone specific decomposition of dirac s equation @xmath152 it has no analogue in the instant form . most remarkable , however , is that the relativistic hamiltonian is additive in the ` kinetic ' and the ` potential ' energy , very much like a non - relativistic hamiltonian @xmath153 . in this respect
the front form is distinctly different from the conventional instant form .
the free solutions of the dirac and the maxwell equations are in the front form @xmath154 the properties of the dirac spinors @xmath155 and @xmath156 , and of the polarization vectors @xmath157 , are given for example in @xcite .
the single particle state are specified by the quantum numbers @xmath158 .
their creation and destruction operators are subject to the usual relations @xmath159 = \left\{b(q ) , b^\dagger(q^\prime)\right\ } = \left\{d(q),\ d^\dagger(q^\prime)\right\ } = \delta ( p^+-p^{+\,\prime } ) \delta ^{(2)}(\vec p_{\!\bot}-\vec p_{\!\bot } ^{\,\prime } ) \delta _
\lambda ^{\lambda ^\prime } \delta _ c^{c^\prime } \delta _ f^{f^\prime } , \nonumber % \ ] ] which carry the operator structure of the theory .
when inserting the free fields into @xmath48 one can integrate out the dependence on @xmath160 , producing essentially dirac delta - functions in the single particle momenta , which reflect momentum conservation : @xmath161 in detail this can be quite laborious , as shown by the example with the fermionic contribution to the vertex interaction @xmath162 \nonumber\\ & \times & \phantom{\gamma ^\mu _ { \alpha\beta } t^{a_3}_{c_1,c_2 } d^\dagger q } \left.\left ( a^\dagger(q_3)\epsilon_\mu^\star ( p_3,\lambda_3 ) e^{+ip_3x } + a(q_3)\epsilon_\mu(p_3,\lambda_3 ) e^{-ip_3x } \right ) \right .
.\nonumber % \end{aligned}\ ] ] note that the sum of these single particle momenta is essentially the sum of the particle momenta minus the sum of the hole momenta . consequently ,
if a particular term has only creation or only destruction operators as in @xmath163 its contribution vanishes since the light - cone longitudinal momenta @xmath164 are all positive and can not add to zero . as a consequence ,
all energy diagrams which generate the vacuum fluctuations in the usual formulation of quantum field theory are absent in the front form .
[ t ] .[tab : verspi ] _ the vertex interaction in terms of dirac spinors .
the matrix elements @xmath165 are displayed on the right , the corresponding ( energy ) graphs on the left .
all matrix elements are proportional to @xmath166 , with @xmath167 . for the periodic boundary conditions one uses @xmath168 . _
[ cols= " < " , ] + the above results are are quite generally applicable : they hold for arbitrary non - abelian gauge theory @xmath169 .
they hold for abelian gauge theory ( qed ) , formally by replacing the color - matrices @xmath170 with the unit matrix and by setting to zero the structure constants @xmath171 , thus @xmath172 and @xmath173 .
they hold for 1 time dimension and arbitrary @xmath174 space dimensions , with @xmath175 .
all what has to be adjusted is the volume integral @xmath176 .
one has to find a language in which one can represent hadrons in terms of relativistic confined quarks and gluons .
as reviewed in @xcite , the bethe - salpeter formalism has been the central method for analyzing hydrogenic atoms in qed and provides a completely covariant procedure for obtaining bound state solutions .
however , calculations using this method are extremely complex and appear to be intractable much beyond the ladder approximation
. it also appears impractical to extend this method to systems with more than a few constituent particles .
an intuitive approach for solving relativistic bound - state problems would be to solve the instant form hamiltonian eigenvalue problem @xmath177 for the hadron s mass and wave function .
here , one imagines that @xmath178 is an expansion in multi - particle occupation number fock states , and that the operators @xmath25 and @xmath179 are second - quantized heisenberg operators .
unfortunately , this method is complicated by its non - covariant structure and the necessity to first understand its complicated vacuum eigenstate over all space and time .
the presence of the square root operator presents severe mathematical difficulties . even if these problems could be solved , the eigensolution is only determined in its rest system ( @xmath180 ) ; determining the boosted wave function is as complicated as diagonalizing @xmath25 itself .
this is why instant form wave function can not be applied in practice to scattering problems .
structure functions for example can not be calculated . in principle
, the front form approach works in the same way .
one aims at solving the hamiltonian eigenvalue problem @xmath181 which for several reasons is easier : contrary to @xmath182 the operator @xmath183 is positive , having only positive eigenvalues .
the square - root operator is absent .
the boost operators are kinematic .
having determined the wave function in a particular frame with fixed total momenta @xmath96 and @xmath184 the kinematic boost operators allow to covariantly transcribe to any other frame .
in fact , as discussed below , one can formulate the theory frame - independently .
the ket @xmath185 can be calculated in terms of a complete set of functions @xmath186 or @xmath187 , @xmath188 \
\left|\mu\right\rangle\left\langle\mu\right| = \sum\limits_{n}\int\!d[\mu_n ] \
\left|\mu_{n}\right\rangle\left\langle\mu_{n}\right| = \mathbf { 1 } .\ ] ] the transformation between the complete set of eigenstates @xmath189 and the complete set of basis states @xmath190 are then @xmath191 and usually called the _ wavefunctions _ @xmath192 .
in addition to the quantum numbers of the lorentz group , the eigenfunction is labeled by quantum numbers like charge , parity , or baryon number which specify a particular hadron @xmath193 , thus @xmath194 \
\left|\mu_n\right\rangle \psi_{n / h } ( \mu ) .\ ] ] one constructs the complete basis of fock states @xmath190 in the usual way by applying products of free field creation operators to the vacuum state @xmath195 : @xmath196 the operators @xmath197 , @xmath198 and @xmath199 create bare leptons ( electrons or quarks ) , bare anti - leptons ( positrons or antiquarks ) and bare vector bosons ( photons or gluons ) .
all of these particles are ` on - shell ' , @xmath200 .
the various fock - space classes are conveniently labeled with a running index @xmath23 .
each fock state @xmath201 is an eigenstate of @xmath202 and @xmath184 and the free part of the energy @xmath203 , with eigenvalues @xmath204 the free invariant mass square of a fock - state is @xmath205 , thus @xmath206 the fock and the physical vacuum have eigenvalues @xmath207 .
the restriction to @xmath208 is a key difference between light - cone quantization and ordinary equal - time quantization .
in equal - time quantization , the state of a parton is specified by its ordinary three - momentum @xmath209 .
since each component of @xmath210 can be either positive or negative , there exist zero total momentum fock states of arbitrary particle number , and these will mix with the zero - particle state to build up the ground state , the physical vacuum .
however , in light - cone quantization each of the particles forming a zero - momentum state must have vanishingly small @xmath211 .
the free or fock space vacuum @xmath212 is then an exact eigenstate of the full front form hamiltonian @xmath25 , in stark contrast to the quantization at equal usual - time .
however , the vacuum in qcd is undoubtedly more complicated due to the possibility of color - singlet states with @xmath213 built on zero - mode massless gluon quanta , but the physical vacuum in the front form is _ still far simpler _ than in the usual instant form . since @xmath214 and @xmath215
, one can define boost - invariant longitudinal momentum fractions @xmath216 and boost - invariant intrinsic transverse momenta @xmath217 their values are constrained , @xmath218 corresponding to the intrinsic frame @xmath219 . all particles in a fock state @xmath220 have a boosted four - momentum @xmath221 the free invariant mass square of the fock state , eq.([eq : freemass ] ) , is therefore @xmath222 as a direct consequence of the transverse boost properties . the phase - space differential @xmath223 $ ] depends on how one normalizes the single particle states . in the convention where commutators are normalized to a dirac - delta function , the phase space integration is @xmath224 \dots & = & \sum _
{ \lambda_i \in n } \int \left[dx_i d^2 k_{\!\perp i } \right ] \dots \ , \qquad{\rm with } \nonumber \\
\left[dx_i d^2 k_{\!\perp i } \right ] & = & \delta \big(1-\sum_{j\in n } x_j\big ) \delta^{(2 ) } \big(\sum _ { j\in n } \vec k_{\perp j}\big ) \ dx_1 \dots dx_{n_n } \
d^2 k _ { \!\perp 1 } \dots d^2 k _ { \!\perp n_n } , \nonumber\end{aligned}\ ] ] where @xmath225 is the number of particles in fock state @xmath226 .
the additional dirac @xmath227-functions account for the constraints ( [ eq:4.12 ] ) .
the eigenvalue equation ( [ eq:4.2 ] ) stands then for an infinite set of coupled integral equations @xmath228 \ \langle
n : x_i , \vec k _ { \!\perp i } , \lambda_i \vert h \vert n^\prime : x_i^\prime , \vec k _ { \!\perp i}^{\,\prime } , \lambda_i^\prime \rangle \,\psi_{n^\prime / h } ( x_i^\prime , \vec k _ { \!\perp i}^{\,\prime},\lambda_i^\prime ) \nonumber\\ & = & { m^2 + \vec p _ { \!\perp}^{\,2}\over p ^+ } \psi _ { n / h}(x_i , \vec k _ { \!\perp } , \lambda_i ) , \qquad{\rm for}\quad n=1,\dots,\infty .\end{aligned}\ ] ] since @xmath96 and @xmath184 are diagonal operators one can rewrite this equation as @xmath228 \
n : x_i , \vec k _ { \!\perp i } , \lambda_i \vert hp ^+ - \vec p _ { \!\perp}^{\,2 } \vert n^\prime : x_i^\prime , \vec k _ { \!\perp i}^{\,\prime } , \lambda_i^\prime \rangle \,\psi_{n^\prime / h } ( x_i^\prime , \vec k _ { \!\perp i}^{\,\prime},\lambda_i^\prime ) \nonumber\\ & = & m^2 \psi _ { n / h}(x_i , \vec k _ { \!\perp } , \lambda_i ) .\label{eq:4.17}\end{aligned}\ ] ] it is therefore possible to define a ` light - cone hamiltonian ' as the operator @xmath229 so that its eigenvalues correspond to the invariant mass spectrum @xmath230 of the theory . eq.([eq:4.17 ] ) thus stands for @xmath231 in analogy to eq.([eq:4.2 ] ) .
the lorentz invariance of @xmath232 and the boost invariance of the wave functions reflects the fact that the boost operators are kinematical .
in fact one can boost the system to an ` intrinsic frame ' in which the transversal momentum vanishes @xmath233 , thus @xmath234 . the transformation to an arbitrary frame with finite values of @xmath235
is then trivially performed .
consider a pion in qcd with momentum @xmath236 as an example .
it is described by @xmath237 \left|{n : x_i p^+ , \vec k _ { \!\perp i } + x_i \vec p _
{ \!\perp } , \lambda_i}\right\rangle\ , \psi_{n/\pi } ( x_i,\vec k _
{ \!\perp i},\lambda_i ) , \ ] ] where the sum is over all fock space sectors of eq.([eq:4.7 ] ) .
the ability to specify wavefunctions simultaneously in any frame is a special feature of light - cone quantization .
the light - cone wavefunctions @xmath238 do not depend on the total momentum , since @xmath239 is the longitudinal momentum fraction carried by the @xmath240 parton and @xmath217 is its momentum `` transverse '' to the direction of the meson ; both of these are frame - independent quantities .
they are the probability amplitudes to find a fock state of bare particles in the physical pion . but given these light - cone wavefunctions @xmath241 , one can compute any hadronic quantity by convolution with the appropriate quark and gluon matrix elements , see for example @xcite .
[ t ] in addressing to solve eq.([eq : lc - hamiltonianii ] ) one faces several major difficulties , among them that the above equations are ill - defined for very large values of the transversal momenta ( ` ultraviolet singularities ' ) and for values of the longitudinal momenta close to the endpoints @xmath242 or @xmath243 ( ` endpoint singularities ' ) .
one has to introduce cut - offs @xmath244 to regulate the theory in some convenient way .
subsequently one has to remove the cut - off dependence by renormalization group analysis .
renormalization theory is known however only for perturbation theory , for example when calculating feynman scattering amplitudes in a certain order .
renormalization theory is not available for the bound - state problem . but even if one has found a convenient regularization scheme , one faces the large ( infinite ) number of coupled integral equations .
their nature resides in the complicated structure of the kernel @xmath245 in analyzing the structure of this very complicated many - body problem , as done in figure [ fig : holy-1 ] , one realizes that most of its matrix elements are zero by nature of the operator structure of the hamiltonian .
the hamiltonian is zero for all sectors whose particle number difference is larger than 2 .
as an example , consider the fork interaction given in eq.([eq : fork_int ] ) particularly its matrix element @xmath246 .
it scatters a quark into an other momentum state and distroys two gluons . in the block matrix element @xmath247 one
has many ways to realize that , without that the anti - quark @xmath248 changes its momentum . in fig .
[ fig : holy-1 ] , all matrix elements in the block @xmath249 are therefore represented diagrammatically by the same energy diagram as in table [ tab : forspi ] .
usually , the typical seagull matrix elements occur for the diagonal blocs of this matrix , with one exception : in the graph @xmath250 of table [ tab : seaspi ] a @xmath251-pair is annihilated and scatteredg instantaneously into a @xmath252-pair .
correspondingly , not all entries of the block matrix element @xmath253 can vanish . in order to cope with the formidable many - body problem imposed by eq.([eq : lc - hamiltonianii ] ) , the method of discretized light - cone quantization
is very useful .
let us devote a section to the perturbative treatment of front - form gauge theory .
light - cone perturbation theory is really hamiltonian perturbation theory , and we give the complete set of rules which are the analogues of the feynman rules .
we shall demonstrate in a selected example , that one gets the same covariant and gauge - invariant scattering amplitude as in feynman theory , see also @xcite .
the green s functions @xmath254 are the probability amplitudes that a state starting in fock state @xmath255 ends up in fock state @xmath256 at a later time @xmath257 @xmath258 its fourier transform @xmath259 is called the resolvent of the hamiltonian @xmath25 , _
i.e. _ @xmath260 separating the hamiltonian @xmath261 into a free part @xmath262 and an interaction @xmath263 , one can expand the resolvent into the series @xmath264 the rules for @xmath257-ordered perturbation theory follow when the resolvent of the free hamiltonian @xmath265 is replaced by its spectral decomposition @xmath266
\left| n : k^+_i,\vec k_{\!\perp i},\lambda_i\right\rangle \ , { 1\over \delta _ n } \left\langle n : k^+_i,\vec k_{\!\perp i},\lambda_i\right| , \label{eq : spec - decom}\ ] ] with the energy denominator @xmath267 .
the sum runs over all fock states @xmath23 intermediate between two interactions @xmath268 . to calculate then @xmath269 perturbatively ,
all @xmath257-ordered diagrams must be considered , the contribution from each graph computed according to the rules of old - fashioned hamiltonian perturbation theory @xcite : 1 .
draw all topologically distinct @xmath257-ordered diagrams .
assign to each line a single particle momentum @xmath270 , a helicity @xmath271 , as well as color and flavor . with fermions ( electrons or quark )
associate a spinor @xmath272 , with antifermions @xmath273 , and with vector bosons ( photons or gluons ) a polarization vector @xmath274 .
3 . for each vertex include the matrix element @xmath275 between fock state @xmath23 and @xmath276 as given in table [ tab : verspi ] .
4 . for each intermediate state
there is an energy denominator @xmath277 in which @xmath278 is the incident free light - cone energy .
5 . to account for three - momentum conservation include for each intermediate state the delta - functions @xmath279 and @xmath280 .
sum over all internal helicities ( and colors for gauge theories ) and integrate over each internal @xmath281 with the weight @xmath282 . 7 .
include a factor @xmath283 for each closed fermion loop , for each fermion line that both begins and ends in the initial state , and for each diagram in which fermion lines are interchanged in either of the initial or final states .
imagine that every internal line is a sum of a ` dynamic ' and an ` instantaneous ' line , and draw all diagrams with @xmath284 instantaneous lines .
two consecutive instantaneous interactions give a vanishing contribution .
10 . for each instantaneous line insert a factor @xmath285 , with the matrix element given in table [ tab : forspi ] and [ tab : seaspi ] . the light - cone fock state representation can thus be used advantageously in perturbation theory .
the sum over intermediate fock states is equivalent to summing all @xmath286ordered diagrams and integrating over the transverse momentum and light - cone fractions @xmath287 . because of the restriction to positive @xmath288 diagrams corresponding to vacuum fluctuations or those containing backward - moving lines
are eliminated .
the simplest application of the above rules is the calculation of the electron - muon scattering amplitude to lowest non - trivial order .
but the quark - antiquark scattering is only marginally more difficult .
we thus imagine an initial @xmath290-pair with different flavors @xmath291 to be scattered off each other by exchanging a gluon .
let us treat this problem as a pedagogical example to demonstrate the rules .
rule 1 : there are two time - ordered diagrams associated with this process . in the first one
the gluon is emitted by the quark and absorbed by the antiquark , and in the second it is emitted by the antiquark and absorbed by the quark .
for the first diagram , we assign the momenta required in rule 2 by giving explicitly the initial fock state @xmath292 .
note that it is invariant under @xmath293 .
the final fock state is @xmath294 .
the intermediate state @xmath295 has ` a gluon in flight ' .
since the gluons longitudinal momentum is positive , the diagram allows only for @xmath296 .
rule 3 requires at each vertex the factors @xmath297 \over \sqrt{2k^+_{q } } \sqrt{2 k^+ _ { g } } \sqrt{2k'^+_{q } } } \ , \\
\langle q^\prime , \bar q , g \vert \,v\ , \vert q ^\prime , \bar q ^\prime \rangle & = & { g\over ( 2\pi)^{{3\over2 } } } \sqrt{{n_c^2 - 1\over 2n_c } } \ { \left [ \overline v(k_{\bar q}^\prime,\lambda_{\bar q}^\prime ) \,\gamma^\nu\epsilon^\star _ \nu(k_g,\lambda_g)\ , v ( k_{\bar q},\lambda_{\bar q } ) \right ] \over \sqrt{2k^+_{\bar q } } \sqrt{2k^+_{g } } \sqrt{2k'^+_{\bar q } } } , \end{aligned}\ ] ] respectively , wich are found by means of table [ tab : verspi ] .
working with color neutral fock states , all color structure reduces to an overall factor @xmath298 , with @xmath299 .
rule 4 requires the energy denominator @xmath300 .
it is useful to work the four - momentum transfers of the quark @xmath301 .
the anti - quark has @xmath302 . with the initial energy @xmath303 ,
the energy denominator becomes then @xmath304 rule 5 requires two dirac - delta functions , one at each vertex , to account for conservation of three - momentum .
one of them is removed by the requirement of rule 6 , namely to integrate over all intermediate internal momenta and the other remains in the final equation ( [ eq:4s.15 ] ) .
the gluon momentum is thus fixed by the external legs of the graph .
the polarization sum over the gluon helicity gives @xmath305 the null vector has the components @xmath306 and thus the properties @xmath307 and @xmath308 . as shown explicitly in @xcite one gets for the second order diagram @xmath309 after some non - trivial steps @xmath310 \over \sqrt{4k^+_q k'^+_q } } { 1\over q^2 } { \left [ \overline u ( k_{\bar q},\lambda_{\bar q } ) \gamma_\mu u(k'_{\bar q},\lambda'_{\bar q } ) \right]\over \sqrt{4k^+_{\bar q } k'^+_{\bar q } } } \nonumber\\ -{g^2 c^2\over ( 2\pi)^3 } \ { \left [ \overline u ( k_q,\lambda_q ) \gamma^+ u(k'_q,\lambda'_q)\right ] \over \sqrt{4k^+_q k'^+_q } } { 1\over \big ( k^+_g \big ) ^2 } { \left [ \overline u ( k_{\bar q},\lambda_{\bar q } ) \gamma^+ u(k'_{\bar q},\lambda'_{\bar q } ) \right]\over \sqrt{4k^+_{\bar q } k'^+_{\bar q } } } .\end{aligned}\ ] ] the delta - functions and a step function @xmath311 are omitted for simplicity .
one proceeds with rule 8 , by including the instantaneous lines .
table [ tab : seaspi ] gives @xmath312_q \over \sqrt{4k^+_q k'^+_q } } { 1\over \big ( k^+_{q } - k'^+_{q}\big ) ^2 } { \left [ \overline u ( k,\lambda ) \gamma^+ u(k',\lambda ' ) \right]_{\bar q } \over \sqrt{4k^+_{\bar q } k'^+_{\bar q } } } .\nonumber\end{aligned}\ ] ] the @xmath289-scattering amplitude , the sum @xmath313 has then the correct gauge - invariant result known from feynman theory up to second order @xmath314_{q } \left [ \overline u ( k,\lambda ) \gamma_\mu u(k^\prime,\lambda^\prime)\right]_{\bar q } .\label{eq:4s.15}\end{aligned}\ ] ] the instantaneous diagram is cancelled exactly .
the infinitely many coupled integral equations in the preceeding section a very difficult to cope with in practice . because of the many integrations and summations it is very difficult to even write them down .
but field theory becomes much more transparent when one works with periodic boundary conditions .
the coupled integral equations become then coupled matrix equations . since rows and colums of a matrix can be denumerated , one can keep track of the necessary manipulations much easier . in fact , working with periodic boundary conditions , or with ` discretized light - cone quantization ' ( dlcq ) , one has obtained the first total solutions to non - trivial quantum field theories in 1 + 1 dimensions . in 3 + 1 dimensions the method has the ambitious goal to calculate the spectra and wavefunctions of physical hadrons from a covariant gauge field theory .
the ingredients of the method shall be reviewed in short in this section .
let us first briefly review the difficulties for a conventional non - relativistic many - body theory .
one starts with a many - body hamiltonian @xmath315 .
the kinetic energy @xmath316 is usually a one - body operator and thus simple .
the potential energy @xmath317 is at least a two - body operator and thus complicated .
one has solved the problem if one has found the eigenvalues and eigenfunctions of the hamiltonian equation , @xmath318 .
one always can expand the eigenstates in terms of products of single particle states @xmath319 belonging to a complete set of ortho - normal functions .
when antisymmetrized , one refers to them as ` slater - determinants ' .
all slater - determinants with a fixed particle number form a complete set .
one can proceed as follows . in the first step one
chooses a complete set of single particle wave functions .
these single particle wave functions are solutions of an arbitrary single particle hamiltonian . in a second step
, one selects one slater determinat as a reference state .
all slater determinants can be classified relative to this reference state as 1-particle-1-hole ( 1-ph ) states , 2-particle-2-hole ( 2-ph ) states , and so on .
the hilbert space is truncated at some level . in a third step
, one calculates the hamiltonian matrix within this hilbert space . in figure
[ fig - kyf-1 ] , the hamiltonian matrix for a two - body interaction is displayed schematically .
most of the matrix - elements vanish , since a 2-body hamiltonian changes the state by up to 2 particles .
therefore the structure of the hamiltonian is a finite penta - diagonal block matrix .
the dimension within a block is made finite by an artificial cut - off on the kinetic energy , _
i.e. _ on the single particle quantum numbers @xmath320
. a finite matrix can be diagonalized on a computer . at the end
one must verify that the physical results are reasonably insensitive to the cut - off(s ) and other formal parameters .
this procedure was actually carried out in one space dimension @xcite with two different sets of single - particle functions , @xmath321 the two sets are the eigenfunctions of the harmonic oscillator ( @xmath322 ) with its hermite polynomials @xmath323 , and the eigenfunctions of the momentum of a free particle with periodic boundary conditions .
both are suitably normalized ( @xmath324 ) , and both depend parametrically on a characteristic length parameter @xmath325 .
the calculations are particularly easy for particle number @xmath326 , and for a harmonic two - body interaction .
the results are displayed in figure [ fig - kyf-1 ] , and surprisingly different .
for the plane waves , the results converge rapidly to the exact eigenvalues @xmath327 , as shown in the right part of the figure .
opposed to this , the results with the oscillator states converge extremely slowly .
obviously , the larger part of the slater determinants is wasted on building up the plane wave states of center of mass motion from the slater determinants of oscillator wave functions .
it is obvious , that the plane waves are superior , since they account for the symmetry of the problem , namely galilean covariance .
the approach was successfully applied for getting the exact eigenvalues and eigenfunctions for up to 30 particles . from these calculations
, one should conclude that discretized plane waves are a useful tool for many - body problems , and that the generate good wavefunctions even for a ` confining ' potential like the harmonic oscillator .
dlcq had been applied first to yukawa theory @xcite in 1-space and 1-time dimensions followed by an application to qed @xcite and to qcd @xcite , but the advantages of working with periodic boundary conditions in the front form , particularly when discussing the ` zero modes ' in the @xmath328-theory , had been noted also by maskawa and yamawaki @xcite in 1976 . in one space dimension
are no rotations hence no spin .
the dirac spinor has two components and the dirac matrices are 2 by 2 matrices .
the gauge field @xmath329 field has two components .
one of them is eliminated by fixing the gauge , and the other determined by gauss law .
the gauge field carries no dynamical degree of freedom .
the theory confines quarks ( or electrons ) because the poisson equation in 1 space dimension gives rise to a linearly rising potential .
quantum electrodynamics in 1 + 1 dimension has played an important role in field theory because its massless version , the schwinger model , is analytically soluable .
the fundamental solution corresponds to a composite @xmath330-state
the schwinger - boson with invariant mass @xmath331 .
note that the coupling constant @xmath150 in 1 + 1 dimension has the dimension of a mass .
v ' '' '' sector & n & 1 & 2 & 3 & 4 & 5 & 6 & 7 & 8 + @xmath332 & 1 & & & & & & & & + @xmath333 & 2 & & & & & & & & + @xmath334 & 2 & & & & & & & & + @xmath335 & 4 & & & & & & & & + 5 @xmath336-pairs & 5 & & & & & & & & + 6 @xmath336-pairs & 6 & & & & & & & & + 7 @xmath336-pairs & 7 & & & & & & & & + 8 @xmath336-pairs & 8 & & & & & & & & + consider first the massive schwinger model .
the lagrangian for the theory takes the same form as in eq.([eq:2.7 ] ) .
again one works in the light - cone gauge @xmath0 , and uses the same projection operators @xmath337 .
the dirac equation decomposes again into a time - derivative @xmath338 and a constraint @xmath339 .
one is left with only one independent field , @xmath340 , which is canonically quantized at @xmath341 , @xmath342 the formalism gets somewhat simplified if one uses the chiral representation @xcite @xmath343 the @xmath344 are then diagonal and project on the chiral components @xmath345 the ( light - cone ) momentum and energy operators become @xmath346 one can expand @xmath347 ( or @xmath340 ) with periodic boundary conditions @xcite , but anti - periodic boundary conditions @xcite avoid the zero mode : @xmath348 the creation and destruction operators obey @xmath349 , consistent with eq.([ceq:19a ] ) .
one redefines units by @xmath350 the length @xmath325 drops out in @xmath351 .
inserting the above fields gives @xmath352 the symbols @xmath353 are kronecker delta s and @xmath354 . in 1 + 1 dimensions
it is very important to keep the ` self - induced inertias ' @xmath355 from the normal ordering .
they are needed to cancel the infrared singularity in the interaction term in the continuum limit . ' '' '' + 2 & @xmath356 + 2 & @xmath357 + 4 & @xmath358 + 5 & @xmath359 ' '' ''
+ [ t ] 75 mm ' '' '' k & 1 @xmath289&2 @xmath289 & 3 @xmath289 & 4 @xmath289 & total + 1 & 1 & - & - & - & 1 + 4 & 4 & 1 & - & - & 5 + 9 & 9 & 20 & 1 & - & 30 + 16 & 16 & 140 & 74 & 1 & 231 + the next step is to actually solve the equations of motions in the discretized space .
typically one proceeds as follows : one constructs the fock space @xmath360 in the same way as above in eq.([eq:4.7 ] ) , and arranges it in denumerated fock space sectors , as illustrated in table [ tab : qed1 + 1 ]
. each fock state must be an eigenstate to @xmath202 and thus of the _ harmonic resolution _
@xmath361 @xcite , with eigenvalue @xmath362 .
one selects now one value for @xmath361 and constructs all fock states .
both the number of fock space sectors and the number of fock states within each sector are finite for every finite @xmath361 , due to the positivity condition on the light - cone momenta . for @xmath363
one has only one fock - space class with one fock state @xmath364 , for @xmath365 one has the 5 fock states given in table [ tab : five_states ] .
the numbers of fock states increase with given @xmath361 , as shown in table [ tab : number_states ] , but less than exponentially due to the exclusion principle .
next , one calculates the matrix elements of @xmath366 . in the last step one
diagonalizes @xmath25 .
any of its eigenvalues @xmath367 depends on @xmath361 and corresponds to an invariant mass @xmath368 .
notice that one gets a spectrum of invariant mass - squares for any value of @xmath361 .
the eigenvalue spectrum of qed@xmath369 was given first by eller @xcite , for periodic boundary conditions on the fermion fields .
the plot likeon the left side in figure [ fig - kyf-4 ] was calculated @xcite with anti - periodic boundary conditions .
it shows the full mass spectrum of qed in the charge zero sector for all values of the coupling constant and the fermion mass , parametrized by @xmath370 .
the eigenvalues @xmath230 are plotted in units where the mass of the lowest ` positronium ' state has the numerical value 1 .
all states with @xmath371 are unbound .
the plot includes the free case @xmath372 ( @xmath118 ) and the the schwinger limit @xmath373 for @xmath374 ( @xmath375 ) , where dlcq generates the exact eigenvalue .
the lower left part of the figure illustrates the following point .
the rich complexity of the spectrum allows for multi - particle fock states _ at the same invariant mass _ as the ` simple @xmath289-states ' shown in the figure as the ` 2 particle sector ' .
the spectrum includes not only the simple bound state spectrum , but also the associated discretized continuum of the same particles in relative motion .
one can identify the simple bound states as two quarks connected by a confining string as displayed in the figure .
the smallest residual interaction mixes the simple configuration with the large number of ` continuum states ' at the same mass .
the few simple states have a much smaller statistical weight , and it looks as if the long string ` breaks ' into several pieces of smaller strings .
loosely speaking one can interpret such a process as the decay of an excited pion into multi - pion configurations @xmath376 .
in @xmath369 dimensions quantum electro dynamics @xcite and quantum chromo dynamics @xcite show many similarities , both from the technical and from the phenomenological point of view . in the right part of figure [ fig - kyf-4 ] some of the results of hornbostel @xcite on the spectrum and the wavefunctions for qcd are displayed .
fock states in non - abelian gauge theory su(n ) can be made color singlets for any order of the gauge group and thus one can calculate mass spectra for mesons and baryons for almost arbitrary values of n. in the upper right part of the figure the lowest mass eigenvalue of a meson is given for @xmath377 .
lattice gauge calculations are available only for @xmath378 and for the lowest two eigenstates @xcite . in general
the agreement is very good . in the left lower part of the figure
the structure function of a baryon is plotted versus ( bjrken-)@xmath287 for @xmath379 .
with dlcq it is possible to calculate also higher fock space components . as an example
, the figure includes the probability distribution to find a quark in a @xmath380-state .
based on the low energy theorems from times prior to qcd , it is believed that the square of the pion mass is linear in the quark mass @xmath320 for sufficiently small @xmath320 .
the proportionality constant has to have a dimension of mass , and since there is no other scale in the problem except the _ quark condensates in the vacuum _
@xmath381 , one believes that the square of the pion mass is approximatively given by @xmath382 , a theorem which was succesful in many phenomenological applications .
the schwinger model has played an important role as a paradigm for our understanding of hadronic physics . among other aspects
it has the desired feature that the invariant mass square of the @xmath330-boson is linear in the electron mass @xmath320 @xmath383 where @xmath331 is the invariant mass in the schwinger limit .
constant @xmath384 was understood as a signal for non - perturbative physics . in his analysis of the schwinger model ,
bergknoff @xcite showed that the take - off from the schwinger limit obeys @xmath385 the result of this ` chiral perturbation theory ' to first order , @xmath386 , is numerically very close to @xmath387 .
[ t ] its is actually quite easy to derive the bergknoff equation from dlcq @xcite . for a fixed resolution @xmath361 ,
a @xmath289-state is fixed by the quantum number @xmath23 of the electron .
the eigenvalue equation is then given by @xmath388 } \langle n \vert h_{lc } \vert n ' \rangle\ , \langle n ' \vert \psi \rangle , \quad{\rm for}\quad n=\frac{1}{2},\frac{3}{2 } , \dots,\ [ k ] , \ ] ] with @xmath389 \equiv k-\frac{1}{2}$ ] . using the hamiltonian matrix elements
are given above , one introduces @xmath390 and goes to the continuum limit .
after a few steps @xcite one gets the integral equation of bergknoff @xcite , @xmath391 for @xmath375 , the solution @xmath392 has the eigenvalue @xmath393 , the schwinger boson .
eqs.([eq : schwinger - mass ] ) and ( [ eq : schwinger - bergknoff - mass ] ) state that the mass - squared of the schwinger boson is _ linear in the fermion mass _ @xmath320 , in the limit when @xmath320 goes to zero . for a long time , this result was in conflict with the explicit dlcq - calulations @xcite .
the shortcoming was taken as a hint @xcite that light - cone quantization was in failure because of the trivial vacuum structure which does not allow for condensates .
the most recent results by vllinger @xcite for dlcq are displayed in figure [ fig : vol-1 ] .
they are compared with the lattice calculations of crewther and hamer @xcite and with chiral perturbation theory up to second order @xcite .
the lattice results and dlcq show some discrepancy for very small @xmath320 which however fades away for larger @xmath320 . on the other hand ,
the lattice results and chiral perturbation theory agree perfectly at small and deviate for large @xmath320 , where chiral perturbation theory is not suposed to work .
the figure show aslo that a correct inclusion of the ( gauge field ) zero modes has no significant impact .
the re - solution of this puzzle came by van de sande @xcite .
he realized that conventional dlcq has difficulties to reproduce the wave function @xmath394 near the endpoints @xmath395 and @xmath396 very close to the schwinger limit .
his ` improved dlcq ' accounts for that , and indeed when properly included as shown in figure [ fig : vol-2 ] all discrepancies fade away .
what should be learned from this exercise is that not everything what is called a ` vacuum condensate ' in the literature deserves this name in a physical sense : in naive light - cone quantization particularly dlcq the vacuum is trivial and can not have condensates .
the naive front - form vacuum is simple .
however , one commonly associates important long range properties of a field theory with the vacuum like spontaneous symmetry breaking , the goldstone pion , or color confinement .
if one can not associate long range phenomena with the vacuum state itself , then the only alternative is the zero momentum components of the field , the ` zero modes ' . in some cases ,
the zero mode operator is not an independent degree of freedom but obeys a constraint equation .
consequently , it is a complicated operator - valued function of all the other modes of the field .
zero modes of this type have been investigated first by maskawa and yamawaki as early as in 1976 @xcite .
an analysis of the zero mode constraint equation for ( 1 + 1)dimensional @xmath328 field theory , by van de sande and pinsky @xcite , shows how spontaneous symmetry breaking occurs within the context of this model .
[ t ] the model represents a new paradigm for spontaneous symmetry breaking and shall be reviewed shortly .
the lagrangian in one space and one time dimension is @xmath398 .
imposing periodic boundary conditions with a length parameter @xmath399 gives @xmath400 .
the field integral @xmath401 is convenient to discuss the problem and becomes @xmath402 also the canonical hamiltonian can be disentangled into @xmath403 and the zero mode @xmath404 @xmath405 one can apply canonical quantization . following the dirac - bergman prescription , described in @xcite , one identifies first - class constraints which define the conjugate momenta @xmath406 , where @xmath407 = { { \delta_{n , m}}/ 2}$ ] and @xmath408 .
the secondary constraint @xmath409 determines the zero mode @xmath404 .
this result can also be obtained by integrating the equations of motion , @xmath410 . to quantize the system
one replaces the dirac bracket by a commutator .
one must choose a regularization and an operator - ordering prescription in order to make the system well - defined .
one begins by defining creation and annihilation operators @xmath411 and @xmath412 , @xmath413 which satisfy the usual commutation relations @xmath414
= \delta_{k , l } $ ] .
likewise , one defines the zero mode operator @xmath415 . in the quantum case , one normal orders the operator @xmath403 .
one redefines this the fields in terms of operators @xmath416 and notes that they obey the canonical commutation relations @xmath417 _ { x^+=y^+=0 } = \delta(x^--y^- ) , \ ] ] for a boson field in the front form , see also eq.([ceq:19a ] ) .
the solution of the constraint eq.([eq : eom_phi ] ) is very difficult .
van de sande and pinsky @xcite have shown that the zero mode @xmath418 could acquire finite values , @xmath419 depending on the coupling constant .
one finds the following general behavior : for small coupling ( large @xmath150 , where @xmath420 ) the constraint equation has a single solution and the field has no vacuum expectation value ( vev ) . as one increases the coupling ( decreases @xmath150 ) to the `` critical coupling '' @xmath421 , two additional solutions which give the field a nonzero vev
these solutions differ only infinitesimally from the first solution near the critical coupling , indicating the presence of a second order phase transition . above the critical coupling ( @xmath422 )
, there are three solutions : one with zero vev , the `` unbroken phase , '' and two with nonzero vev , the `` broken phase '' .
the `` critical curves '' shown in figure [ ff1 ] , is a plot of the vev as a function of @xmath150 .
periodic boundary conditions on @xmath423 can be realized by periodic boundary conditions on the vector potentials @xmath424 and anti - periodic boundary conditions on the spinor fields , since @xmath423 is bilinear in the @xmath425 . in momentum representation one
expands these fields into plane wave states @xmath426 , and satisfies the boundary conditions by _
discretized momenta _
@xmath427 as an expense , one has to introduce two artificial length parameters , @xmath428 and @xmath429 .
they also define the normalization volume @xmath430 .
more explicitly , the free fields are expanded as @xmath431 particularly for the two transverse vector potentials @xmath432 , ( @xmath433 ) . as above , each single particle state
` @xmath434 ' is specified by six quantum numbers , the three discrete momenta @xmath435 , helicity , color and flavor . the creation and destruction operators like @xmath436 and @xmath437
create and destroy single particle states @xmath434 , and obey ( anti- ) commutation relations like @xmath438 = \big\ { b _ q , b^\dagger _ { q^\prime } \big\ } = \big\ { d _ q , d^\dagger _
{ q^\prime } \big\ } = \delta_{q , q^\prime } .\ ] ] the kronecker symbol is unity only if all six quantum numbers coincide . inserting the free fields of eq.([enexdi ] ) into the hamiltonian ,
one performs the space - like integrations and ends up with the light - cone energy - momenta @xmath439 as operators acting in fock space .
the spatial components of @xmath440 are simple and diagonal , the temporal component @xmath441 is complicated and off - diagonal .
the integrals over the coordinates @xmath125 are conveiently expressed in kronecker delta functions @xmath442 and correspondingly for the transversal integrations . in the tables given above
they appear typically in the overall factor @xmath443 the continuum formulation of the hamiltonian problem in gauge field theory with its endless multiple integrals is usually cumbersome and untransparent . in dlcq
, the continuum limit corresponds to harmonic resolution @xmath444 .
the compactified formulation with its simple multiple sums is straightforward .
the key relation is the connection between sums and integrals @xmath445 combined they yield @xmath446 similarly , dirac delta and kronecker delta functions are related by @xmath447 because of that , in order to satisfy the respective commutation relations , one must modify also the creation and destruction operators . denoting the single boson operators in the continuum by @xmath448 and in the discretized case by @xmath67 , they must be related by @xmath449 and correspondingly for fermion operators .
of course , one has formally to replace sums by integrals , kronecker delta by dirac delta functions , and single particle operators by their tilded versions . in practice
, it suffices to replace the tilded coupling constant @xmath450 in order to convert the discretized expressions in tables [ tab : verspi]-[tab : seaspi ] to the continuum formulation .
the dlcq method can be considered a general framework for solving problems such as relativistic many - body theories or approximate models .
the general procedure is : ( 1 ) phrase the physics problem in dlcq ; ( 2 ) apply approximation and simplifications ; ( 3 ) derive the final result ; ( 4 ) at the end convert th so obtained expressions to the continuum .
the finite number of fock - space sectors is a consequence of the positivity of the longitudinal light - cone momentum @xmath451 .
the transversal momenta @xmath452 can take either sign , and the number of fock states within each sector can be arbitrarily large . in order to face a finite dimensional hamiltonian matrix one must have a finite number of fock states , and this is achieved by _ fock space regularization _ : following lepage and brodsky @xcite , a fock state with @xmath23 particles is included only if its free invariant mass does not exceed a certain threshold @xmath453 the sum extends over all @xmath23 particles in a fock state .
the lowest possible value of @xmath454 is taken when all particles are at rest relative to each other , _ i.e. _ @xmath455 .
this frozen invariant mass should be removed from the cut - off the mass scale @xmath456 is a lorentz scalar and one of the parameters of the theory .
however , it was not realized in the past @xcite , that fock - space regularization is almost irrelevant in the continuum theory . _
vertex regularization _ seem to be a better alternative . at each vertex
, a particle with four - momentum @xmath110 is scattered into two particles with respective four - momentum @xmath457 and @xmath458 . in order to avoid potential singularities one can _ regulate the interaction _ by setting the matrix element to zero if the off - shell mass @xmath459 exceeds a certain scale @xmath244 .
the condition @xmath460 will be referred to as the sharp cut - off for _ vertex regularization_. the dependence on the regularization parameter @xmath244 must be removed by the renormalization group .
renormalization looks like a terrible problem in the context of non - perturbative theory , but with the above regularization scheme it could be simple in principle : the eigenvalues may not depend on the regulator scale(s ) @xmath244 . to require
this is easier than to find a practical realization .
in principle one proceeds in 3 + 1 like in 1 + 1 dimensions : one selects a particular value of the harmonic resolution @xmath361 and the cut - off @xmath244 and diagonalizes the finite dimensional hamiltonian matrix by numerical methods .
but the bottle neck of any hamiltonian approach is that the dimension of the hamiltonian matrix increases exponentially fast with the cut - off . as a concrete example consider the matrix structure as given for @xmath365 in fig .
[ fig : holy-1 ] .
suppose the regularization procedure allows for 10 discrete momentum states in each direction .
for every single particle one has about @xmath461 possibilities to define a momentum state .
a fock - space sector with @xmath23 particles and fixed total momentum has then roughly @xmath462 different fock states .
sector 13 in fig .
[ fig : holy-1 ] alone , with its 8 particles , has thus about @xmath463 fock states .
chemists are able to handle matrices with some @xmath464 dimensions , but @xmath463 dimensions exceeds the calculational capacity of any computer in the foreseeable future . for 3 + 1 dimensions one
is thus confronted with a a similar problem as in conventional many - body physics , displayed in fig .
[ fig - kyf-1 ] .
one has to diagonalize finite matrices with exponentially large dimensions ( typically @xmath465 ) .
in fact , the problem in quantum field theory is even more difficult since the particle number is unlimited .
one needs an effective interaction which acts in smaller matrix spaces and which has a well defined relation to the full interaction .
one needs it also for the physical understanding .
the effective interaction between two electrons , for example , is the coulomb interaction , at least to lowest order of approximation .
the goal is therefore to develop an exact effective interaction between a quark and an anti - quark in a meson . to achieve this goal ,
the hamiltonian dlcq - matrix is discussed in terms of block matrices , since the fock - space sectors appear quite naturally in a gauge theory , see also eq.([eq:4.17 ] ) .
each fock - space sector has a finite number of fock states , which are kept track of collectively .
eq.([eq : lc - hamiltonianii ] ) is therefore rewritten as a block matrix equation , with @xmath466 and @xmath467 , @xmath468 rows and columns are denumerated in the same convention as in table [ tab : qcd_blocks ] .
as to be shown , it can be mapped identically on a matrix equation which acts only in sector @xmath469 .
once this is achieved , one can go to the continuum limit .
effective interactions are a well known tool in many - body physics @xcite . in field
theory the method is known as the tamm - dancoff - approach , applied first by tamm @xcite and by dancoff @xcite , which shall be reviewed it in short .
the rows and columns of a matrix are split into the @xmath470- and the @xmath471-space . in terms of the sector numbers of eq.([aeq:319 ] ) , @xmath472 and @xmath473 , where @xmath474 .
eq.([aeq:319 ] ) can thus be written as a 2 by 2 block matrix equation @xmath475 if one can invert the quadratic matrix @xmath476 one could express the q - space in terms of the @xmath470-space wavefunction . but here is a problem : the eigenvalue @xmath477 is unknown at this point .
but one can replace it by another number , _ the starting point energy _
@xmath478 , which is at first a free parameter .
the matrix inverse to @xmath479 is called the resolvent of the hamiltonian in the @xmath471-space and is denoted by @xmath480 .
the @xmath471-space wave function becomes then @xmath481 substituting it in eq.([aeq:321 ] ) produces an eigenvalue equation in the @xmath482-space , @xmath483 and defines the effective @xmath482-space hamiltonian @xmath484 varying @xmath478 one generates a set of _ energy functions _ @xmath485 .
every solution of the _ fix - point equation _ @xmath486 generates one of the eigenvalues @xmath25 , in fact , it generates all of them .
v ' '' '' sector & n & 1 & 2 & 3 & 4 & 5 & 6 & 7 & 8 & 9 & 10 & 11 & 12 & 13 + @xmath332 & 1 & & & v & & & & & & & & & & + @xmath487 & 2 & & & v&&v & & & & & & & & + @xmath488 & 3 & v&v&&v&&v & & & & & & & + @xmath333 & 4 & & & v&&&&v & & & & & & + @xmath489 & 5 & & v&&&&v&&&v & & & & + @xmath490 & 6 & & & v&&v&&v&&&v & & & + @xmath491 & 7 & & & & v&&v&&v&&&v & & + @xmath334 & 8 & & & & & & & v&&&&&v & + @xmath492 & 9 & & & & & v&&&&&v & & & + @xmath493 & 10 & & & & & & v&&&v&&v & & + @xmath494 & 11 & & & & & & & v&&&v&&v & + @xmath495 & 12 & & & & & & & & v&&&v&&v + @xmath335 & 13 & & & & & & & & & & & & v & + if one identifies the @xmath482- with the @xmath289-space one seems to have found the effective interaction which acts in the fock space of a single quark and a single anti - quark .
it looks as if one has mapped a difficult problem , the diagonalization of a big matrix onto a simpler problem , the diagonalization of a small matrix .
but the price to pay is to invert a matrix .
matrix inversion takes about the same numerical effort as its diagonalization . in view of having to vary @xmath478 , the numerical work is therefore rather larger than smaller as compared to a direct diagonalization .
the advantage of working with a resolvent is of analytical nature to the extent that resolvents can be approximated systematically .
the two resolvents @xmath496 defined once with and once without the off - diagonal interaction @xmath268 , are identically related by @xmath497 , or by the infinite series of ( tamm - dancoff ) perturbation theory @xmath498 see also eq.([eq : spec - decom ] ) .
the free resolvent @xmath499 can be obtained trivially since the kinetic energy @xmath500 is diagonal .
conceptually , it is the same object as the one in eq.([eq : spec - decom ] ) , except for two aspects : the present @xmath499 acts only in the @xmath471-space , and the starting point energy @xmath478 is a _ constant _ ( one of the eigenvalues ) ; the free energy @xmath501 in ( [ eq : spec - decom ] ) , however , is a _ function of the incoming momenta_. the starting point energy not being a kinetic energy creates problems allover the place , since non - integrable singularities appear in every order of ( tamm - dancoff ) perturbation theory .
essentially two conclusions are possible : either gauge theory has no bound - state solution , or the series in eq.([aeq : m24 ] ) has to be resumed to all orders of perturbation theory before the singularities begin cancel eachother . in practice , tamm and
dancoff @xcite have restricted themselves to the first non - trivial order . in order to make things work ,
they have replaced the energy denominator with the eigenvalue @xmath478 by the energy denominator with the function @xmath501 , as it appears in the perturbative scattering amplitudes .
the same trick was applied in the later work with the front form @xcite .
perturbation theory within a bound state problem is known to be a very difficult question .
it has motivated the formal work in the next few sections .
what we are after is the impossible , some kind of non - perturbative perturbation theory ! the tamm - dancoff approach can be interpreted as the reduction of a block matrix dimension from 2 to 1 . but having a matrix with block matrix dimension @xmath502 as in table [ tab : qcd_blocks ] , it could be interpreted as the reduction from @xmath503 , simply by choosing the @xmath471-space appropriately .
but then the procedure can be iterated , one can reduce the block matrix dimension from @xmath504 , and so on until one arrives at @xmath505 .
this method of ` iterated of resolvents ' has certain advantages , which will be discussed as the formalism develops .
first , one needs a reasonable and compact notation .
one of them is to denumerate the fock space sectors as in table [ tab : qcd_blocks ] .
the full hamiltonian will be denoted by @xmath506 , since by the definition in eq.([aeq:319 ] ) it has @xmath507 blocks .
suppose , during the reduction one has arrived at block matrix dimension @xmath23 , with @xmath508 .
the eigenvalue problem corresponding to eq.([aeq:345 ] ) reads then @xmath509 in analogy to eq.([aeq:332 ] ) , define the @xmath23-space resolvent , and get @xmath510 the effective interaction in the ( @xmath511)-space becomes then @xmath512 for every block matrix element .
it is unpleasant but unavoidable , that the symbol @xmath23 denotes both the block matrix dimension and the number of the last sector .
else one proceeds like for tamm - dancoff , including the fixed point equation @xmath513 .
but one has achieved much more : eq.([aeq:414 ] ) is a _ recursion relation _ !
a few comments seem to be in order .
the method of iterated resolvents @xcite is particularly suited for gauge theory with its many zero block matrices , see table [ tab:3 ] .
the zero matrices remove many of the multiplications in eq.([aeq:410 ] ) .
the tamm - dancoff procedure can not make use of them .
both the tamm - dancoff procedure and the iterated resolvents can be put on a computer in model studies .
the iterated resolvents require to invert several smaller matrices insted of one big one .
since matrix diagonalization ( and inversion ) grows with power 3 in the dimension , the technique of iterated resolvents might thus even be faster . the iterated and the tamm - dancoff resolvents are distinctly different in the following aspect : @xmath514 conserves particle number , but @xmath515 does not .
both however conserve the incoming momentum . the notation in eq.([aeq:414 ] ) is very compact and will be explained further below .
the method applies also equally well to conventional many - body problems since a pair - interaction generates for example the bloc matrix structure shown in figure [ fig - kyf-1 ] . finally , it should be emphasized that the higher sector wavefunctions can be retrieved by matrix multiplications from the eigenfunction in the lowest sector , from @xmath516 .
no additional matrix diagonalizations or inversions are required . to show this , consider eq.([aeq:410 ] ) for @xmath517 ,
_ i.e. _ @xmath518 both @xmath519 and @xmath520 were calculated as a matrix for getting down to the effective interaction in the @xmath521-space . required is thus one additional matrix multiplication .
next get @xmath522 .
substituting @xmath523 gives @xmath524 the general case , @xmath525 can be proven by induction .
it might be interesting to study a hamiltonian with a tridigonal band structure , such as was given for example in table [ tab : qed1 + 1 ] : @xmath526 according to the rules it develops the structure a of continued fraction , since with @xmath527 one has explicitly @xmath528 here and above a very compact notation is used , which is explained by the example @xmath529 by reasons of space , the abbrevation @xmath530 is used sometimes in the diagonal blocks .
the kinetic energies @xmath500 are diagonal matrix elements , and the seagulls do not change particle number by definition . even that the above is very compact .
here is what it means in terms of matrix operations : @xmath531 within each sector @xmath255 , the number of basis or fock states is finite , _
i.e. _ @xmath532 with @xmath533 .
the steps to be taken are then : take the reactangular block matrix @xmath534 ; matrix - multiply it from the left with the inverse matrix of @xmath535 which is @xmath536 ; matrix - multiply the result with the transpose ( and complex conjugated ) matrix @xmath537 ; the result is the quadratic matrix @xmath538 ; add this to the quadratic matrix @xmath539 . as a net result , the old matrix @xmath539 in eq.([eq:4x4_matrix ] )
is replaced the effective ( and @xmath478-dependent ) matrix @xmath540 . to play the game a little further
, one can reduce the number of states within a block to 1 .
eq.([eq : example - matrix ] ) stands then for a tri - diagonal matrix and eq.([eq : cont - fraction ] ) stands literally for a continued fraction .
such one is analyzed in figure [ fig:4_1 ] , with a perfect agreement between the method of iterated resolvents and a direct numerical diagonalization .
[ t ] in the front form , the fock - space for harmonic resolution @xmath541 has only 4 fock - space sectors , see table [ tab : qcd_blocks ] .
@xmath541 provides thus a good example for all 4 by 4 hermitean bloc matrices , in which the block matrix element @xmath542 vanishes .
the case is considered as an exercise and a non - trivial example . since @xmath543
one starts out with the block matrix @xmath544 the block matrix element @xmath545 is a rectangular zero - matrix , since the light - cone operator has no matrix elements between two gluons @xmath252 and two @xmath289-pairs .
the notation keeps track of the field theoretic property , that each block matrix has only one type of interactions , namely instaneneous seagull ( s ) and fork ( f ) , or dynamic vertex ( v ) interactions .
the hamiltonian in the @xmath546-sector is quadratic and has the resolvent @xmath547 using eq.([aeq:414 ] ) to reduce the block matrix dimension from 4 to 3 gives @xmath548 almost every block matrix element is replaced by an ` effective ' element .
now continue to reduce from 3 to 2 : the resolvent of @xmath549 is now @xmath550 with @xmath551 one gets @xmath552 finally , evaluate the last resolvent @xmath553 to end up with the 1 by 1 block matrix @xmath554 here the procedure stops . for gauge theory one can order the result with the power of the coupling constant @xmath555 with the abbreviations @xmath556 finally , for completeness , the higher sector wave function in eq.([eq : sector - wave ] )
are written out explicitely @xmath557 \vert 1 \rangle\,\langle 1 \vert \psi\rangle , \\
\langle 3\vert\psi\rangle & = & g_3 \left[v + \overline { v g_2 v g_3 v } + v g_2 v g_3 v g_4 f \right ] \vert 1 \rangle\,\langle 1 \vert \psi\rangle , \\ \langle
4\vert\psi\rangle & = & g_4 \left [ \overline{\overline { v g_3 v } } + \overline{\overline { v g_3 v g_2 v g_3 v } } + v g_3 v g_2 v g_3 v g_4 f \right ] \vert 1 \rangle\,\langle 1 \vert \psi\rangle , \label{eq : wavefunction}\end{aligned}\ ] ] with the additional abbreviations @xmath558 these abbreviations are not very relevant for the general case . but for gauge theory they have a deeper physical meaning as to be discussed further below .
[ t ] 10em all effective interactions have a number of strings like @xmath559 .
this is a colossal simplification as compared infinite number of terms of tamm - dancoff perturbation theory .
iterated resolvents resume them in a systematic way .
both kinds of series can be identified term by term if one expands the iterated resolvents also about the kinetic energies . denoting the free resolvent by @xmath560 it is related to the tamm - dancoff resolvent @xmath561 by @xmath562
if one then expands @xmath563 using @xmath564 one gets @xmath565 thus infinite series of infinite series , which after reordering become identical with the tamm - dancoff series .
a look back to eq.([eq:4-interaction ] ) is helpful .
the various term have been arranged there by order of the coupling constant . in the first line one
finds the combination @xmath566 , defined in eq.([eq : overline - v ] ) . in the caption to figure [ fig:6_6 ]
it is explained how one can simplify the formal procedures , if one associates with each intrisic line a dynamic plus an instantaneous line .
doing that replaces the above sum of six terms into a single one namely @xmath567 , and similarily one replaces the term in the first line by @xmath568 .
there is a physical reason behind that : when calculating scattering diagrams to a fixed order in perturbation theory , one obtains the gauge invariant feynman amplitudes only , if all instantaneous graphs of the same order are included .
an example for that was given above , when calculating the scattering graph up to second order . obviously , the method of iterated resolvents accounts for this aspect automatically , in a formal way . since the remaining terms in eq.([eq:4-interaction ] )
are of order 6 or 8 in the coupling constant , one concludes that the restriction to the phase space for @xmath541 satisfies gauge invariance only up to terms of order 6 in the coupling constant .
one can verify , that the order of violating gauge invariance is pushed up if one includes more fock - space sectors .
this rule was checked explicitly in many rather laborious calculations in @xcite .
an easy trick to achieve gauge invariance is the following : set to zero formally all instantaneous interactions , perform the steps required by the resolvents . at the end ,
when having calculated the effective interaction in terms of the vertex interaction @xmath569 , one replaces each internal line by the sum of a dynamical and a instantaneous line , according to the general rules of light - cone perturbation theory .
if one follows thes rules for the 4 by 4 matrix given in eq.([eq:4x4_matrix ] ) one obtains for the sector hamiltonians and wavefunctions simply @xmath570 these expressions are more transparent than the full equations ; they are also gauge invariant .
when one substitutes formally all instantaneous interactions by block matrices with only zeros , keeping only the vertex interactions @xmath569 , one gets a block matrix structure as displayed in table [ tab:3 ] . because of the many zero matrices , the construction of the the sector hamiltonians is now rather straightforward and simple . here
they are for the first 12 sectors : @xmath571 actually , for @xmath365 , the exact expressions corresponding to table [ tab:3 ] are obtained by setting @xmath572 and @xmath573 for @xmath574 .
but it is easy to write down the expressions in eqs.([eq : u12])-([aeq:622 ] ) valid for a higher value of @xmath361 .
one also notes that they give all sector hamiltonians for @xmath541 in eqs.([aeq : m36 ] ) , by setting formally @xmath573 for @xmath575 . for sufficiently large harmonic resolution @xmath361 the formal expressions for effective hamiltonian in sufficiently low sectors become independent of @xmath361 .
one therefore can go to the limit @xmath444 and thus to the continuum limit .
v ' '' '' sector & n & 1 & 2 & 3 & 4 & 5 & 6 & 7 & 8 & 9 & 10 & 11 & 12 & 13 + @xmath332 & 1 & & & v & & & & & & & & & & + @xmath487 & 2 & & & v&&v & & & & & & & & + @xmath488 & 3 & v&v&&v&&v & & & & & & & + @xmath333 & 4 & & & v&&&&v & & & & & & + @xmath489 & 5 & & v&&&&v&&&v & & & & + @xmath490 & 6 & & & v&&v&&v&&&v & & & + @xmath491 & 7 & & & & v&&v&&v&&&v & & + @xmath334 & 8 & & & & & & & v&&&&&v & + @xmath492 & 9 & & & & & v&&&&&v & & & + @xmath493 & 10 & & & & & & v&&&v&&v & & + @xmath494 & 11 & & & & & & & v&&&v&&v & + @xmath495 & 12 & & & & & & & & v&&&v&&v + @xmath335 & 13 & & & & & & & & & & & & v & + here seems to be a problem : for calculating @xmath576 one needs @xmath577 , @xmath578 and @xmath579 , for calculating @xmath577 one needs @xmath580 , @xmath581 and @xmath582 , and so on .
in the next section will be shown how the hierarchy can be broken in a rather effective way . that final step will be comparatively simple if one has analyzed the propagators for the sectors with one @xmath289-pair and arbitrarily many gluons , as follows next .
consider first the the effective interaction in the space of one @xmath289-pair and one gluon as given by eq.([aeq:620 ] ) .
the corresponding diagrams can be grouped into two topologically distinct classes , displayed in figs .
[ fig:6_2 ] and [ fig:6_3 ] .
the the diagrams in fig .
[ fig:6_2 ] are obtained by adding a non - interacting gluon line to the former diagrams in fig .
[ fig:6_1 ] .
the gluon does not change quantum numbers under impact of the interaction and acts like a spectator .
therefore , the graphs in fig . [ fig:6_2 ] will be refered to as the ` spectator interaction ' @xmath583 . in the graphs of fig .
[ fig:6_3 ] the gluons are scattered by the interaction , and correspondingly these graphs will be refered to as the ` participant interaction ' @xmath584 .
thus , @xmath585 .
the separation into a graph with only one interacting quark - pair fock - space sectors @xmath586 except in those with only gluons .
more explicitly , the spectator and participant interactions in the lowest sectors with one quark - pair become @xmath587 the same operators can appear in both @xmath588 and @xmath589 , but they refer to different graphs . since the hamiltonian is additive in spectator and participant interactions , @xmath590 can be associated with its own resolvent @xmath591 the relation of @xmath592 to the full resolvent is @xmath593 , or @xmath594 writing it as an infinite series .
note that the propagator @xmath592 contains the interaction @xmath595 in contrast to the free tamm - dancoff propagator @xmath561 in eq.([aeq:352 ] ) .
one deals here therefore with ` perturbation theory in medium ' .
but since it is driven to all orders one better speaks of ` propagation in medium '
. the series can be identically written as @xmath596 one can verify this order by order in perturbation theory or by the identity @xmath597 the inverse gives @xmath598 , and with the identity @xmath599 , one gets eq.([aeq:6520 ] ) .
this remarkable property is peculiar to the method of iterated resolvents .
whenever a quark - pair - glue resolvent is sandwiched in between two vertex interactions , it allows to define a modified vertex interaction @xmath600 , since @xmath601 one can thus rewrite the sector hamiltonians like for example @xmath602 now they have all essentially the same structure , contrary to eqs.([aeq:610 ] ) where they were different .
note that @xmath519 , @xmath578 and @xmath581 are pure gluon propagators .
generally spoken , the rectangular block matrix @xmath569 is multiplied with the square matrix @xmath603 .
below it will be shown that @xmath604 is approximately diagonal and independent of the spin .
each vertex matrix element is multiplied with a simple function which depends on the momentum transfer @xmath471 across the vertex .
equivalently one replaces the coupling constant @xmath150 by @xmath605 , such that @xmath604 can be interpreted as a _ coupling function_. the most important result of this section is that gauge theory particularly qcd has only two structurally different contributions to the effective interaction in the @xmath289-space , see eq.([aeq:664 ] ) .
it scatters a quark with helicity @xmath606 and four - momentum @xmath607 into a state with @xmath608 and four - momentum @xmath609 , and correspondingly the antiquark . in the _ continuum limit _ , the resolvents are replaced by propagators and the matrix eigenvalue problem @xmath610 becomes an integral equation with a rather transparent structure : @xmath611 \langle x,\vec k_{\!\perp } ; \lambda_{q } , \lambda_{\bar q } \vert \psi _
i\rangle \label{aeq:445}\\ & + & \sum _ { \lambda_q^\prime,\lambda_{\bar q}^\prime } \!\int\!dx^\prime d^2 \vec k_{\!\perp}^\prime\ , \ , r ( x^\prime , k_{\perp}^\prime ) \,\langle x,\vec k_{\!\perp } ; \lambda_{q } , \lambda_{\bar q } \vert u_{\rm oge } \vert x^\prime,\vec k_{\!\perp}^\prime
; \lambda_{q}^\prime , \lambda_{\bar q}^\prime\rangle\ , \langle x^\prime,\vec k_{\!\perp}^\prime ; \lambda_{q}^\prime,\lambda_{\bar q}^\prime \vert \psi _
i\rangle \nonumber\\ & + & \sum _ { \lambda_q^\prime,\lambda_{\bar q}^\prime } \!\int\!dx^\prime d^2 \vec k_{\!\perp}^\prime\ , \ , r ( x^\prime , k_{\perp}^\prime ) \,\langle x,\vec k_{\!\perp } ; \lambda_{q } , \lambda_{\bar q } \vert u_{\rm tga } \vert x^\prime,\vec k_{\!\perp}^\prime ; \lambda_{q}^\prime , \lambda_{\bar q}^\prime\rangle\ , \langle x^\prime,\vec k_{\!\perp}^\prime ; \lambda_{q}^\prime,\lambda_{\bar q}^\prime \vert \psi _ i\rangle .\nonumber\end{aligned}\ ] ] the domain of integration is set by the cut - off function @xmath612 given in eq.([aeq : i1 ] ) . the effective one - gluon exchange @xmath613 conserves the flavor along the quark line . as illustrated in fig .
[ fig:6_1 ] the vertex interaction @xmath569 creates a gluon and scatters the system virtually into the @xmath614-space . as indicated in the figure by the vertical line with subscript ` 3 ' , the three particles propagate there under impact of the full hamiltonian before the gluon is absorbed .
the gluon can be absorbed either by the antiquark or by the quark .
if it is absorbed by the quark , it contributes to the effective quark mass @xmath615 . the second term in eq.([aeq:664 ] )
is the effective two - gluon - annihilation interaction , @xmath616 as represented by the graph @xmath617 in fig .
[ fig:6_1 ] .
the virtual annihilation of the @xmath289-pair into two gluons can generate an interaction between different quark flavors .
the eigenvalue @xmath618 is one of the invaraint - mass squared eigenvalues of the full light - cone hamiltonian , and the wavefunction @xmath619 gives the probability amplitudes for finding in the @xmath289-state a flavored quark with momentum fraction @xmath287 , intrinsic transverse momentum @xmath620 and helicity @xmath621 , and correspondingly an anti - quark with @xmath622 , @xmath623 and @xmath624 .
the content of the preceeding sections is exact but rather formal . in the sequel
, rigor will be given up by the aim to obtain a solvable equation . in order to spot easier the essential approximations
, they will be dressed as ` theorems ' , which can or can not be proven later .
all sector hamiltonians can be diagonalized on their own merit , for example @xmath625 the spectra might be continuous , but the eigenvalues are denumerated for simplicity by @xmath626 . by construction , one knows the relation between these two sets of eigenvalues , by the following reason : because of the separation into spectators and participants the gluon in the @xmath614-equation is a non - interacting particle .
it moves freely relative to a @xmath289-bound state .
if its four - momentum is parametrized as @xmath627 one has @xmath628 every @xmath289 bound state @xmath629 is a band - head in the @xmath614 spectrum . with @xmath630 one
gets @xmath631 knowing the eigenvalues and the eigenfunctions of the @xmath289 bound state one alos know the spectrum and the eigenfunctions of @xmath632 . knowing the eigenfunctions , the exact resolvent can be calculated .
the exact propagators can be approximated by closure .
the substitution of the exact propagators by closure is a widely used approximation in many - body theory and in chemistry . in principle , the exact propagator is non - diagonal .
performing closure , it becomes diagonal , which is reflected by the dirac - delta function @xmath633 in the single particle momenta and helicities @xmath634 the dirac - delta function is multiplied with the function @xmath635 the in - medium propagator @xmath636 ceases to be a functional of the higher resolvents . _
the hierarchy of the iterated resolvents is broken_. in the solution , the interacting particles propagate free particles .
the in - medium propagators can be replaced by the free propagators .
the free propagator in the @xmath637-space can be written @xmath638 where @xmath639 is the four - momentum transfer along the quark line , in the notation of fig .
[ fig : g_1 ] . for sufficiently small @xmath640 holds
@xmath641 $ ] .
if one substitutes @xmath642 , which holds to rather good approximation , the momentum transfer in eq.([aeq : m72 ] ) is the same as in eq.([aeq:6.68 ] ) .
similar considerations also hold all other propagators with at least one @xmath289-pair .
having replaced the in - medium - propagators @xmath643 by the free propagators , one can calculate the graphs like in the usual ( light - cone- ) time - ordered perturbation theory . the diagram @xmath644 in fig .
[ fig:6_1 ] yields the effective quark mass @xmath645 correspondingly one can calculate the coupling function @xmath646 as defined by eq.([aeq:6520 ] ) . with @xmath647 from eq.([aeq : n50 ] ) , and thus @xmath648 from eq.([aeq:652 ] ) , the first two terms in the expansion are @xmath649 in the last step @xmath650 was set . in general , they contribute terms of higher orders in @xmath150 , which must be suppressed for consistency . figs .
[ fig : g_1 ] and [ fig : g_2 ] give only 2 representative graphs . according to the rules of time ordered perturbation theory ,
there are about 22 different time - ordered and instantaneous diagrams .
all of them were re - calculated by raufeisen @xcite , for all values of @xmath471 .
as shown to some detail in @xcite one gets for sufficiently large @xmath244 @xmath651 with an arbitrary @xmath652 and a @xmath653 independent of @xmath244 given in eq.([aeq : m96 ] ) .
one conculdes that the coupling function @xmath654 is identical with the familiar vertex correction . in the limit where @xmath471 is larger than all mass scales
, this result agrees with the calculations of both , thorn @xcite and perry @xcite , having been done previously .
what happens if one substitutes the free propagators in eq.([aeq : m74 ] ) by the non - perturbative propagators @xmath655 and @xmath656 , at least in an approximate fashion ? there are additional graphs . in the fermion loop of vacuum polarization
appear two graphs in addition to fig .
[ fig : g_1 ] . in one of them ,
a gluon is emitted and absorbed on the same quark line which changes the bare quark mass @xmath657 into the physical quark mass @xmath658 . in the other graph , the gluon is emitted from the quark and absorbed by the anti - quark which represents an interaction . in consequence one has a bound state with a physical mass scale @xmath659 .
we replace therefore @xmath660 .
similar considerations hold for the gluon loop in fig .
[ fig : g_2 ] and lead to the substitution @xmath661 . both @xmath662 and @xmath659
are interpreted as physical mass scales .
the physical gluon mass @xmath663 vanishes of course due to gauge invariance .
this is not in conflict with _ f.e .
_ cornwall s suggestion of a finite effective gluon mass @xcite since one can define @xmath664 . as a consequence of the approximation in eq.([aeq:6.69 ] ) , the propagators @xmath665 and @xmath666 are diagonal in fock space , which in turn leads to diagonal products @xmath667
they can be resumed therefore to all orders according to eq.([aeq : m74 ] ) . with @xmath668 ,
one gets @xmath669 the effective fine structure constant depends on the momentum transfer @xmath471 across the vertex and on the cut - off @xmath244 .
now , all the pieces are together which are needed for a further discussion of the effective interaction as defined in eq.([aeq:445 ] ) .
finally , the the instantaneous interaction in the effective one - gluon - exchange interaction @xmath670 is restored according to @xmath671 , with the expression given in eq.([eq:4s.15 ] ) .
what have we reached ? the general expression for the effective hamiltonian in the @xmath289-sector was given in eq.([aeq:445 ] ) .
if one restricts to consider mesons in which the quark and the anti - quark have different flavors , @xmath672 , the two - gluon annihilation diagram can not get active and one remains with @xmath673 \langle x,\vec k_{\!\perp } ; \lambda_{q } , \lambda_{\bar q } \vert \psi _ n\rangle & & \nonumber\\ + \sum _ { \lambda_q^\prime,\lambda_{\bar q}^\prime } \!\int\!dx^\prime d^2 \vec k_{\!\perp}^\prime\ , \,\langle x,\vec k_{\!\perp } ; \lambda_{q } , \lambda_{\bar q } \vert u_{\rm oge } \vert x^\prime,\vec k_{\!\perp}^\prime ; \lambda_{q}^\prime , \lambda_{\bar q}^\prime\rangle \
\langle x^\prime,\vec k_{\!\perp}^\prime ; \lambda_{q}^\prime,\lambda_{\bar q}^\prime \vert \psi _ n\rangle & . &
\label{new - aeq:445}\end{aligned}\ ] ] the essential achievement of the two preceeding sections is that the kernel of this integral equation can be written down in the very explicit form @xmath674 all many - body effects reside in the effective mass @xmath675 and in the effective coupling constant @xmath676 . they are given in eqs.([aeq : i64 ] ) and ( [ aeq : m96 ] ) , respectively .
the mean of the four - momentum transfers of quark and anti - quark , @xmath677 and @xmath678 , respectively , is denoted by @xmath679 .
the spinor factor is @xmath680_q \ , \left [ \overline u ( k,\lambda ) \gamma_\mu u(k^\prime,\lambda^\prime)\right ] _
{ \bar q } .\ ] ] the cut - off function @xmath612 sets the domain of integration and is defined in eq.([aeq : i1 ] ) .
the approximations made have been carefully enumerated in section 7 .
one has thus reached the goal of reducing the field theoretical many - body problem of @xmath681 to a one - body problem with an effective interaction . for any value of the lagrangian coupling constant @xmath150 , the flavor masses @xmath657 , and the cut - off @xmath244
, one can generate the eigenvalues @xmath682 on a computer , very much in analogy to qed @xcite .
with the known eigenfunction in the @xmath289-sector one an generate all higher fock - space amplitudes , according to the explicit formulas in section 6 .
it looks as if one has solved the problem .
but one has solved the problem only in a superficial way , since the solutions depend on the unphysical parameter @xmath244 .
one of the central issues in gauge - field theory is to remove this dependence by a renormalization group analysis , see below .
the eigenvalues of the integral equation are the invariant masses squared .
very often it is more convenient to think in terms of an analogue of a non - relativistic hamiltonian @xmath683 , @xmath684 whose eigenvalues @xmath685 are the more convential binding energies , while the eigenfunctions are the same . as shown in @xcite ,
the so defined ` schrdinger operator ' has much in common with a non - relativistic hamiltonian for the coulomb problem .
all eigenvalues @xmath686 depend thus on the cut - off @xmath244 , the bare masses @xmath320 and the bare @xmath687 . in line with modern interpretation of quantum field theory @xcite , they are unphysical parameters and can be replaced by ( @xmath688 ) , by functions of @xmath244 , such that the eigenvalues are cut - off independent , _
i.e. _ @xmath689 this fundamental equation of the renormalization group must hold for all eigenvalues .
the effective hamiltonian @xmath690 depends on @xmath244 only through @xmath691 , @xmath692 and the regulator function @xmath693 .
the renormalization group equations can be written therefore in the operator form @xmath694 the simultaneous variation of all three terms is a very difficult problem .
but one can replace this equation by three independent ones , @xmath695 this is possible since one still solves eq.([ceq:9 ] ) .
eqs.([ceq:10 ] ) and ( [ ceq:11 ] ) are two equations for the two unknown functions @xmath696 and @xmath697 .
the first equation says that the effective flavor masses @xmath698 are renormalization group invariants .
the numerical value of @xmath691 must be fixed by experiment .
eq.([ceq:11 ] ) is then considered as an equation for @xmath697 at fixed values of @xmath691 .
using the expression for @xmath676 given in eq.([aeq : m96 ] ) yields then @xcite @xmath699 with @xmath653 given in eq.([aeq : m96 ] ) .
all @xmath244-dependence cancels exactly in favor of the renormalization group invariant @xmath652 , which is sometimes called the qcd - scale @xmath700 .
the scale @xmath652 must be determind from experiment .
once @xmath697 is known , eq.([ceq:10 ] ) is an equation for @xmath696 .
this function can be determined from eq.([aeq : i64 ] ) , but we renounce to do that here , since @xmath696 is needed nowhere in the present context .
having fixed the functions @xmath697 and @xmath696 one has exhausted all freedom provided by conventional renormalization anaysis of field theory @xcite .
since the cutoff - dependence was removed by renormalization , the cut - off can be driven to infinity , thus @xmath701 formally , this solves eq.([ceq:12 ] ) .
after renormalization , the kernel of the one - body equation is independent of @xmath244 .
let us discuss its relevant part @xmath702 at very small momentum transfers @xmath703 , where @xmath704 and @xmath705 , the effective coupling constant locks into the finite value @xmath706 , see eq.([aeq : m96 ] ) .
the spinor function also goes to a constant , @xmath707 , see @xcite or @xcite .
the remaining ` coulomb singularity ' @xmath708 is square - integrable and can be dealt with by convenient numerical means @xcite . for very large momentum transfers @xmath709 ,
the behaviour is quite different .
both @xmath710 and @xmath711 tend to diverge , @xmath712 and @xmath713 , but such that the ratio @xmath714 tends to a _ finite constant_. disregarding the very slow logarythmic increase of @xmath653 , the kernel of the integral equation is therefore essentially a dimensionless constant @xmath715 one has sufficient evidence @xcite that this behaviour creates all kinds of problems , among them a diverging eigenvalue . one can separate the kernel of the integral equation identically into two pieces @xmath716 , _ i.e. _ @xmath717 with an arbitrary mass parameter @xmath718 . the first part @xmath719 is well behaved and vanishes at least like @xmath720 in the limit of large @xmath471 .
the second part @xmath721 is a strange object : it is a constant almost everywhere , in particular @xmath722 one should emphasize the following aspect . the typical field theoretical divergences like the divergence of the effective coupling constant lead to a _ divergence _ of the kernel , but adding a _ finite constant _ to the kernel leads also to a divergence .
thus far one does not understand the reason for the latter .
c|rrrrrr ' '' '' & @xmath723 & @xmath724 & @xmath725 & @xmath726 & @xmath727 + ' '' '' @xmath728 & & 768 & 892 & 2007 & 5325 + @xmath729 & 140 & & 896 & 2010 & 5325 + @xmath730 & 494 & 498 & & 2110 & + @xmath64 & 1865 & 1869 & 1969 & & + @xmath731 & 5278 & 5279 & 5375 & & + c|cccccc ' '' '' & @xmath723 & @xmath724 & @xmath725 & @xmath726 & @xmath727
+ ' '' '' @xmath728 & & @xmath732&@xmath733&@xmath734&@xmath735 + @xmath729 & @xmath736 & & @xmath737&@xmath738&@xmath739 + @xmath730 & @xmath740 & @xmath741 & & @xmath742&@xmath743 + @xmath64 & @xmath744 & @xmath745&@xmath746 & & @xmath747 + @xmath731 & @xmath748 & @xmath749&@xmath750&@xmath751 & + c|rrrrrr ' '' '' & @xmath723 & @xmath724 & @xmath725 & @xmath726 & @xmath727 + ' '' '' @xmath728 & & @xmath752768 & 902 & 2012 & 5331 + @xmath729 & @xmath752140 & & 902 & 2012 & 5331 + @xmath730 & @xmath752494 & 494 & & 2155 & 5476 + @xmath64 & @xmath7521865 & 1865 & 2108 & & 6617 + @xmath731 & @xmath7525278 & 5278 & 5423 & 6573 & c|rrrrrr ' '' '' & @xmath723 & @xmath724 & @xmath725 & @xmath726 & @xmath727 + ' '' '' @xmath728 & & @xmath752768 & 1002 & 2301 & 5696 + @xmath729 & @xmath752140 & & 1002 & 2301 & 5696 + @xmath730 & @xmath752494 & 494 & & 2535 & 5829 + @xmath64 & @xmath7521865 & 1865 & 2102 & & 7227 + @xmath731 & @xmath7525278 & 5278 & 5512 & 6811 & + perhaps one can get further insight by studying the following eigenvalue in ( usual ) three - momentum space ( @xmath210 ) @xmath753 \psi(\vec k ' ) .\label{peq:106}\ ] ] in the spirit of eq.([peq:105 ] ) , the right - hand side is a conventional schrdinger equation for the coulomb problem with the reduced mass @xmath754 , but where an arbitrary constant has been added in the kernel . in heidelberg
, we are presently working on the question whether its eigenvalue @xmath755 diverges .
preliminary studies with the fourier transform of eq.([peq:106 ] ) , where the constant works its way into a weird dirac - delta function @xmath756 as a potential , @xmath757 \phi(\vec x ' ) , \ ] ] indicate indeed that the eigenvalue of the @xmath758-state in particular diverges . if these preliminary results materialize , one must regularize and subsequently renormalize eq.([peq:106 ] ) .
the energy of the @xmath758-state will then be essentially a renormalization constant @xmath759 . given that to be true , all of us have been riding the wrong horse .
it is not the regular part of the interaction kernel @xmath719 , but it is the ` constant ' @xmath721 which determines the mass in the first place .
since @xmath721 involves a renormalization and since the singlet can have an energy shift ( @xmath760 ) different from the triplet ( @xmath761 ) , one has at most two additional renormalization constants , subject to be determined by experiment .
is there some physical evidence for the above considerations ?
inspired by eq.([peq:105 ] ) , one can seek the masses of the physical mesons in the form @xcite @xmath762 and fit the energy shifts @xmath763 and the physical quark masses @xmath764 to the physical meson masses , compiled in table [ tab : meson - masses ] from the data of the particle data group @xcite which do not include yet the topped mesons .
the following procedure was applied .
first , the up and the down mass were chosen equal .
then the empirical masses of @xmath765 and @xmath732 were used to determine the mass shifts for the singlet and the triplet .
the remaining quark masses are obtained from the pseudo - scalar mesons with an up quark , which exhausts all freedom in determining physical parameters .
the remaining 13 off - diagonal pseudo - scalar and vector meson masses are calculated straightforwardly and compiled in table [ tab : mesons - mass - calc ] . in comparing the so obtained numbers with the experiment , one notes with great surprise that the discrepancy exceeds only rarely an estimated error of 10% .
the resulting quark masses are given in table [ tab : quarks ] . actually , there is no reason why one should stick to eq.([peq:103 ] ) .
equivalently , one can replace @xmath766 by @xmath767 .
choosing @xmath768 mev , one gets @xmath769 a form like this was suggested to me by dae sung hwang during the lectures . redoing the fit for the heavy quarks gives the results compiled in table [ tab : mesons - mass - calc ] .
the discrepancy with the experimental numbers is now on the level of 1% .
this excellent agreement should not be over - emphasized . in any case
it does not falsify the above considerations .
c||rrrrr||rr ' '' '' method & @xmath770 & @xmath771 & @xmath772 & @xmath773 & @xmath774 & @xmath760 & @xmath761 + ' '' '' eq .
( ) & 350 & 350 & 583 & 1881 & 5275 & -336 & 71 + ' '' '' eq .
( ) & 350 & 350 & 495 & 1637 & 4972 & -336 & 71
in these lecture notes the attempt was made to phrase and discuss a relativistic quantum theory such as qcd from the point of view of dirac s front form of hamiltonian dynamics
. sometimes refered to as light - cone quantization , it has the goal to describe and understand the bound - state structure of hadrons in terms of their constituents from a covariant theory . as the lectures show , this goal has not been reached yet .
as reviewed in @xcite , there might be alternative roads to reach the same goal , like for example the bethe - salpeter equations , lattice gauge theory , hamiltonian flow equations , just to name a few .
each of these approaches have their own inherent adventages or disadventages . by reasons of space their discussion
is left out in these notes . by the same reason the light - front renormalizations approach of wilson and collaborators @xcite is omitted .
the many articles have thus far not provided conclusive evidence for success , apart from the fact that virtually every langrangian symmetry has been violated in the approach .
it is much too early to draw definitive conclusions .
the front form is however useful to study the structure of arbitrary field theories .
the theories considered are often not physical , but are selected to help in the understanding of a particular non - perturbative phenomenon .
the relatively simple vacuum properties of light - front field theories underly many of these ` analytical ' approaches .
the relative simplicity of the light - cone vacuum provides a firm starting point to attack many non perturbative issues .
as we have seen , the problems in two dimensions , in one space and one time dimension , not only are they tractable from the outset but in many cases like the schwinger model can they be solved numerically by ` discretized light - cone quantization ( dlcq ) ' to almost arbitrary accuracy .
this solution gives a unique insight and understanding .
the schwinger particle indeed has the simple parton structure that one hopes to see in qcd .
unfortunately the same can not be said for the physical world in 3 space and 1 time dimensions .
the essential problem is that the number of degrees of freedom needed to specify each fock state even in a discrete basis grows much too quickly .
as discussed in this review , the basic procedure is to diagonalize the full light - cone hamiltonian in the free light - cone hamiltonian basis .
the eigenvalues are the invariant mass squared of the discrete and continuum eigenstates of the spectrum .
the projection of the eigenstate on the free fock basis are the light - cone wavefunctions and provide a rigorous relativistic many - body representation in terms of its particle degrees of freedom . given the light - cone wavefunction one can compute the structure functions and distribution amplitudes .
more generally , the light - cone wavefunctions provide the interpolation between hadron scattering amplitudes and the underlying parton subprocesses .
the method of iterated resolvents can be a useful intermediary step to generate these wavefunctions .
the unique property of light - cone quantization that makes the calculations of light cone wavefunctions particularly useful is that they are independent of the reference frame and that the same wavefunction can be use in many different problems . *
the light - cone wavefunctions are independent of the momentum of the bound state only relative momentum coordinates appear . *
the vacuum state is simple and in many cases trivial . *
fermions and fermion derivatives are treated exactly ; there is no fermion doubling problem .
* the minimum number of physical degrees of freedom are used because of the light - cone gauge .
no gupta - bleuler or faddeev - popov ghosts occur and unitarity is explicit . * the output is the full color - singlet spectrum of the theory , both bound states and continuum , together with their respective wavefunctions .
pauli , in : _ neutrino mass , monopole condensation , dark matter , gravitational waves and light - cone quantization _ , b.n .
kursunoglu , s. mintz and a. perlmutter , eds . , ( plenum press , new york , 1996 ) p.183 - 204 ; hep - th/9608035 , hep - th/9707361 , heidelberg preprint mpih - v9 - 1998 , mar . | light - cone quantization of gauge theories is discussed from two perspectives : as a calculational tool for representing hadrons as qcd bound - states of relativistic quarks and gluons , and as a novel method for simulating quantum field theory on a computer . a general non - perturbative method for numerically solving quantum field theories , ` discretized light - cone quantization ' ,
is outlined .
both the bound - state spectrum and the corresponding relativistic wavefunctions can be obtained by matrix diagonalization and related techniques .
emphasis is put on the construction of the light - cone fock basis and on how to reduce the many - body problem to an effective hamiltonian .
the usual divergences are avoided by cut - offs and subsequently removed by the renormalization group .
for the first time , this programme is carried out within a hamiltonian approach , from the beginning to the end . starting with the qcd - lagrangian
, a regularized effective interaction is derived and renormalized , ending up with an almost solvable integral equation .
its eigenvalues yield the mass spectrum of physical mesons , its eigenfunctions yield their wavefunctions including the higher fock - space components . an approximate but analytic mass formula is derived for all physical mesons .
= 0 mm = 5 mm = -10 mm preprint mpih - v19 - 1999 + 60 pages , 13 figures , 11 tables , 41 references + pacs - index : 11.10ef , 11.15tk , 12.38lg , 12.40yx + better figure quality is obtained from .uu and .ps files from + ftp [email protected]/pub/pauli/99korea/ + ' '' '' 1ex _ to be published in : _ + _ new directions in quantum chromodynamics _ , c.r .
ji , _ ed . ,
_ + american institute of physics , new york , 1999 .
+ ( proceedings of a summer school in seoul , korea , 26 may - 18 june 1999 . )
+ ' '' '' |
magnetars are a small class of isolated x - ray pulsars , believed to have the highest magnetic fields known to date ( @xmath0@xmath12@xmath13 g ; @xcite ) .
this class comprises the anomalous x - ray pulsars ( axps ) and the soft gamma - ray repeaters ( sgrs ) , observationally very similar in many respects ( see @xcite for a recent review ) : a spin period in the 212s range , large period derivatives ( @xmath14@xmath15 ) , unpredictable bursting activity on different timescales ( from ms to hundreds of seconds ) and luminosities ( @xmath16@xmath17 ) . until not long ago axps were thought as persistent and stable x - ray sources . only in 2003
the first transient axp was discovered by , namely , which displayed a factor of @xmath4100 flux enhancement with respect to the pre - outburst luminosity level as seen by the _ rosat _ and _ einstein _ missions ( @xmath0@xmath18 ; @xcite ) .
even more surprising was the discovery of a highly variable pulsed radio emission which followed the outburst @xcite , never observed before in any other magnetar @xcite .
moreover , the transient nature of this axp provided the first hint that a relatively large number of members of this class has not been discovered yet , and suggested that others would manifest themselves in the future through a phenomenology ( outburst ) similar to that displayed by .
indeed , after this first discovery , two more transient axps have been detected showing similar outbursts , , @xcite and .
the latter axp was first suggested as a candidate magnetar in the supernova remnant g327.240.13 through x - ray observations @xcite , and subsequently recognized as a radio transient magnetar through the discovery of radio pulsations @xcite at a period of @xmath02.1s , and of an x - ray outburst @xcite . in particular
, /x - ray telescope ( xrt ) observations taken in summer 2007 revealed at an x - ray flux level of @xmath0@xmath19 erg @xmath2 s@xmath3 ( more than one order of magnitude brighter than in quiescence ) , with characteristics similar to those of in outburst .
no x - ray bursts were observed during the 2007 outburst of , possibly due to a sparse x - ray coverage .
more generally , the first phases of the 2003 and 2007 x - ray outbursts from and , respectively , were missed .
relatively deep pointings were obtained in august 2006 and 2007 , the first during quiescence , and the second during the outburst decay .
pulsed fractions of about @xmath015% and @xmath07% were inferred for the 2006 and 2007 observations , respectively @xcite . the 2006 quiescent spectrum ( @xmath0@xmath20 erg @xmath2 s@xmath3 in the 110 kev range ) can be fitted well with the usual thermal ( blackbody ) plus non - thermal ( power - law ) model ( @xmath21 kev and @xmath22 ; @xcite ) , with a 1d resonant cyclotron scattering model ( with a surface temperature @xmath23=0.33 kev , an electron velocity @xmath24=0.32 and an optical depth @xmath25=1.0 ; @xcite ) , or with 3d monte carlo model ( with a surface temperature of @xmath23=0.33 kev , an electron bulk velocity of @xmath26=0.15 , and a twist angle of @xmath27=1.14 ; @xcite ) . on the other hand , the 2007 spectrum ( @xmath0@xmath28 erg @xmath2 s@xmath3 in the 110 kev range ) is characterized by a harder emission with @xmath29 kev and @xmath30 ( by using the bb plus pl model ) .
deep infrared observations taken from eso
vlt during the 2007 outburst have revealed four objects consistent with the radio axp position , although none of them showed variability @xcite .
recently , during the giant outburst detected from on 2009 january 22 , a relatively bright transient ir source ( @xmath31 mag ) was discovered within the radio positional uncertainty region of the source @xcite . within the magnetar scenario ,
axp outbursts are thought to be caused by large scale rearrangements of the external magnetic field , either accompanied or triggered by fracturing of the neutron - star crust .
these events may result in renewed magnetospheric activity through the interaction between thermal surface photons and charges flowing along the ( closed ) field lines .
in addition , hot spots on the neutron - star surface may appear where the currents impact on the star .
repeated resonant cyclotron scatterings onto the magnetospheric particles result in a modification of the seed thermal spectrum with the appearance of a non - thermal , high - energy tail @xcite . during the first month since the onset of an outburst , the source experience a large flux variations ( approximately by one or two orders of magnitudes ) . in this respect ,
monitoring campaigns , obtained just after the outburst onset , are giving an important opportunity to track the evolution of the main physical parameters and , therefore , to check the goodness of the ( empirical and/or more physical ) components used to model the spectra .
here we report on the discovery of bursting activity and a new outburst from , as observed by ( [ swift ] ) , which was promptly triggered by the onset of the bursting activity of this transient axp on 2008 october 3 .
the monitoring of the source in the 100s22days time interval allowed us to carry out a detailed timing and spectral study , the first ever for this source , of the initial phases of its outburst on a daily basis .
we also report on the results of an target of opportunity observation performed during the x - ray outburst ( [ integral ] ) , and discuss our findings in comparison with other magnetar outbursts ( [ disc ] ) .
the satellite @xcite is an efficient observatory for the discovery and the multiwavelength monitoring of axp bursting activity .
whenever the wide - field coded - aperture mask burst alert telescope ( bat ; @xcite ) triggers on a burst from an interesting hard x - ray transient , the system on - board can rapidly re - point the satellite in order to bring the source in the narrower field of view of the other two telescopes on - board , the x - ray telescope ( xrt ; @xcite ) and the ultraviolet / optical telescope ( uvot ; @xcite ) . in this section
we present the results obtained from our analysis of the bat and xrt observations of axp performed during its bursting activity on 2008 october , being the source too absorbed to be detected with the uvot .
on 2008 october 3 at 09:28:08 ut bat triggered on and localized a short burst from a position consistent with that of the axp ( trigger 330353 ; @xcite ) which was followed by several other bursts extending to at least 09:38:24 ut .
about 2 hours later , at 11:16:13 ut , bat detected one last , bright burst which reached a peak count rate of @xmath020000 counts s@xmath3 in the 15350 kev band ( trigger 330367 ; @xcite ) .
the bat time coverage of the source was discontinuous along the bursting period , totalling an effective exposure time of 4 ks ( see figure [ bat_timeline ] ) .
datasets containing information for individual photons ( event data ) and data composed of 80-channel histograms with a typical integration time of @xmath05 minutes ( survey data ) were distributed . in the following , we present the results of our analysis for the bursts detected in the event data .
@lcccccccc burst time ( t t@xmath32 ) & @xmath25 & net counts & duration & @xmath23 & r & l & fluence & @xmath33(d.o.f . ) + & ( s ) & & ( s ) & ( kev ) & ( km ) & erg s@xmath3 & erg @xmath2 & + + -2.45 & 0.007@xmath34 & 358 & 0.016 & 10.1@xmath35 & 1.5@xmath36 & 6.0 @xmath37 10@xmath38 & 4.4 @xmath37 10@xmath39 & 0.93(38 ) + 0 & 0.018@xmath40 & 1492 & 0.070 & 10.9@xmath41 & 1.3@xmath42 & 6.4 @xmath37 10@xmath38 & 2.1 @xmath37 10@xmath43 & 1.04(38 ) + 111.6 & 0.036@xmath44 & 2964 & 0.098 & 12.4@xmath45 & 1.2@xmath46 & 9.4 @xmath37 10@xmath38 & 4.3 @xmath37 10@xmath43 & 1.18(38 ) + 282.45 & 0.015@xmath47 & 1451 & 0.035 & 11.3@xmath48 & 1.6@xmath42 & 1.1 @xmath37 10@xmath49 & 1.9 @xmath37 10@xmath43 & 1.44(38 ) + 427.9 & 0.038@xmath50 & 988 & 0.100 & 9.3@xmath51 & 1.2@xmath52 & 2.8 @xmath37 10@xmath38 & 1.3 @xmath37 10@xmath43 & 0.88(38 ) + 524.2 & 0.038@xmath53 & 8081 & 0.254 & 11.8@xmath36 & 1.4@xmath46 & 1.0 @xmath37 10@xmath49 & 1.2 @xmath37 10@xmath54 & 0.80(38 ) + & 0.133@xmath55 & & & & & & & + 524.2 rise & & 310 & 0.010 & 11.1@xmath51 & 1.5@xmath36 & 8.6 @xmath37 10@xmath38 & 4.0 @xmath37 10@xmath39 & 1.34(38 ) + 524.2 peak & & 742 & 0.007 & 8.3@xmath56 & 4.6@xmath57 & 2.6 @xmath37 10@xmath49 & 8.0 @xmath37 10@xmath39 & 1.01(38 ) + 524.2 decay & & 7029 & 0.237 & 11.9@xmath36 & 1.3@xmath46 & 9.4 @xmath37 10@xmath38 & 1.0 @xmath37 10@xmath54 & 0.94(38 ) + 612.55 & 0.012@xmath58 & 757 & 0.100 & 11.1@xmath59 & 0.7@xmath52 & 2.0 @xmath37 10@xmath38 & 9.5 @xmath37 10@xmath39 & 0.92(38 ) + 6485 & 0.098@xmath60 & 14317 & 0.730 & 10.8@xmath36 & 1.3@xmath46 & 5.8 @xmath37 10@xmath38 & 2.0 @xmath37 10@xmath54 & 1.17(38 ) + & 0.283@xmath61 & & & & & & & + 6485 rise & & 676 & 0.020 & 9.6@xmath62 & 2.0@xmath63 & 8.6 @xmath37 10@xmath38 & 8.0 @xmath37 10@xmath39 & 0.82(38 ) + 6485 peak & & 806 & 0.010 & 7.4@xmath36 & 5.0@xmath57 & 2.0 @xmath37 10@xmath49 & 8.2 @xmath37 10@xmath39 & 1.03(38 ) + 6485 decay & & 12835 & 0.700 & 10.9@xmath36 & 1.2@xmath46 & 5.6 @xmath37 10@xmath38 & 1.8 @xmath37 10@xmath54 & 1.08(38 ) + the data reduction was performed using version 2.9 of the standard /bat
software distributed within ftools under the heasoft package ( version 6.5.1 ) . a weight map , taking into account the contribution from the source to the counts in each bat detector ( mask - weighting technique ) ,
was applied in order to obtain the background - subtracted counts of the source . in figure [ bat_lcurve ] the 1550 kev
light curves of bursts with more than 350 counts are plotted . for each burst we extracted a spectrum .
all spectra are fitted well by a single blackbody model ( bb ) with a temperature of about 11 kev and an emitting radius of about 1 km ( see table [ table_bat_spec_fit ] ) .
the limited statistics of bat spectra did not allow us to check whether an additional bb component is present similar to the case of sgr1900 + 14 @xcite .
we also tried to fit the 15 - 100 kev spectra with a power law and an optically thin thermal bremsstrahlung obtaining worse reduced @xmath64 values ( in the [email protected] and [email protected] range for the faintest and brightest bursts , respectively , and for 38 degree of freedom ) .
the characteristic time - scales @xmath25 of each burst decay was also inferred by fitting the corresponding light curves with one or two ( for the longest events ) exponential laws and are reported in table [ table_bat_spec_fit ] .
+ whenever the bursts had sufficient statistics , time - resolved spectra were generated corresponding to the rise , peak , and decay phases of the bursts .
the results of the fitting procedure is reported in table[table_bat_spec_fit ] .
no significant variations were found in the spectral parameters , except for the normalization factor due to the luminosity evolution ( decay ) during the burst . on the other hand ,
time - resolved spectra show a moderate spectral evolution with the peak bb temperature being systematically softer than that during decay ( similar results where obtained for and : @xcite ) .
+ the presence of persistent emission during the non - bursting time intervals was also searched for in the event data .
an image was generated from all the events of each file covering time intervals where no bursts were present and the standard source detection was run ( batcelldetect ) .
two time intervals were investigated , the first one ranging from @xmath66 s to @xmath67 s ( net exposure time 1001 s ) and the second one from @xmath68 s to @xmath69 s ( net exposure time 1050 s ) .
no significant emission in the direction of axp were found .
the 3@xmath70 upper limits on the flux in the 1450 kev band for the two intervals described above are @xmath71 and @xmath72 , respectively .
[ cols="^,^,^,^ , > , > " , ] the exposure time is usually spread over several snapshots ( single continuous pointings at the target ) during each observation .
during this observation xrt repeatedly switched between pc and wt modes ; since the exposure in wt is short ( about 39s ) , we did not include it in our analysis
. executed a prompt slew on and the first xrt observation started only 99 s after the bat trigger .
table [ log ] reports the log of the observations that were used for this work .
the xrt uses a ccd detector sensitive to photons with energies between 0.2 and 10 kev .
we considered windowed timing ( wt ) and photon counting ( pc ) mode data . in pc mode
the entire ccd is read every 2.507s , while in wt mode only the central 200 columns are read and only one - dimensional imaging is preserved , achieving a time resolution of 1.766 ms ( see @xcite for a detailed description of xrt modes ) .
+ the data were processed with standard procedures using the ftools task xrtpipeline ( version 0.12.0 ) and events with grades 012 and 02 were selected for the pc and wt data , respectively ( see @xcite ) .
for the timing and spectral analysis , we extracted the pc source events in the 0.5 - 10 kev range within a circle with a radius of 30 pixels ( @xmath0@xmath73 ) .
the wt data ( in the 0.5 - 10 kev range ) were extracted in a rectangular region 40 pixels long ( and 20 pixels wide ) along the image strip . to estimate the background , we extracted pc events within an annular region ( radii of 50 and 75 pixels ) centered on the source and wt events within a rectangular box ( @xmath74 pixels ) far from .
+ the 0.5 - 10 kev events were used to study the timing properties of the pulsar after having corrected the photon arrival times to the barycentre of the solar system with the barycorr task ( ra=15@xmath75 50@xmath76 54@xmath7712 , dec=-54@xmath78 18@xmath79 24@xmath8019 and j2000 was assumed for the source position ; israel et al .
we started by inferring an accurate period measurement by folding the data from the first xrt pointing ( in wt mode ) at the period reported by @xcite and studying the phase evolution within the observation by means of a phase - fitting technique ( details on this technique are given in @xcite ) . due to the high variability of the pulse shape
we decided not to use a pulse template in the cross - correlation ( which might artificially affect the phase shift ) , but to fit each individual folded light curve with the fundamental plus two higher harmonics . in the following we also implicitly assumed that the pulsation period ( and its derivatives ) is a reliable estimate of the spin period ( and its derivatives ) , an assumption which is usually considered correct for isolated neutron stars . the resulting best - fit period ( @xmath64=1.1 for 2 degree of freedom , hereafter d.o.f . )
is @xmath81s ( 1@xmath70 c.l . ; epoch 54742.0 mjd ; see @xcite ) . the above period accuracy of 3 @xmath82s is enough to phase - connect coherently the later pointings carried out on a daily baseline ( see table[log ] ) .
therefore , the procedure was repeated by adding , each time , a further pointing .
the relative phases were such that the signal phase evolution could be followed unambiguously for the whole visibility window ( see figure [ figure : phase ] ) .
a quadratic term was clearly present in the phase time evolution since the very beginning .
the resulting phase - coherent solution had a best - fit period @xmath83s and period derivative @xmath84s s@xmath3 ( mjd 54742.0 was used as reference epoch ; 1@xmath70 c.l . ) .
these values are in agreement ( within @xmath03@xmath70 ) with those , @xmath85 s and @xmath86s s@xmath3 , reported by @xcite and based on a reduced dataset .
the fit of the phases with a linear and a quadratic term remained statistically acceptable ( reduced @xmath87 for @xmath88 d.o.f . ) for the first two weeks since the bat trigger .
however , during the latest week the pulse phases increasingly deviate from the extrapolation of the above @xmath5@xmath6 solution ( see figure[figure : phase ] central panel , empty triangles ) , resulting in an unacceptable fit ( @xmath89 for 12 d.o.f . ) .
therefore , we added a higher order component in order to take into account for the possible presence of a temporary or secular @xmath7 term .
the resulting new phase - coherent solution is @xmath90 s , @xmath91 s s@xmath3 , and @xmath92 s s@xmath93 ( mjd 54742.0 was used as reference epoch ; 1@xmath70 c.l . ;
@xmath94 for 11 d.o.f . ) , or
@xmath95 hz , @xmath96 hz s@xmath3 , and @xmath97 hz s@xmath93 .
the time residuals with respect to the new timing solution are reported in figure[figure : phase ] ( central panel ; empty squares ) .
the significance level for the inclusion of the cubic component is 5.1@xmath70 .
moreover , the new timing solution implies a r.m.s .
variability of only 0.08 s , corresponding to a timing noise level of about 3% , usually observed in isolated neutron stars .
it is worth emphasizing that the second period derivative we found is very unlike to be related to a random change of the pulse profiles which is expected to introduce only a random distribution of the phase residuals , rather then a cubic term .
recent studies on a sample of 366 radio pulsars showed that cubic terms on phase residuals are possible though smaller by several orders of magnitudes than that detected in and recorded on time - scales longer ( years ) than those we are sampling in our dataset @xcite .
more in general the ( long - term ) timing noise of young radio pulsars ( @xmath98yr ) can be explained as being caused by the recovery from previous glitch events @xcite .
the latter results make unlike that the period second derivative is due to random noise .
finally , we note that in the above analysis we assumed that the emitting region is fixed in time with respect to the observer , as suggested by studies on other transient magnetars ( see @xcite ) .
correspondingly , we can also reasonably exclude that the cubic term is introduced by a random motion of the hot - spot .
+ based on the above phase coherent solution we also studied the pulse shape and pulse fraction evolution .
these are changing as shown in figure[figure : phase ] ( lower panel ) and figure[figure : efold ] as a function of time .
in particular the pulsed fraction ( semi - amplitude of the sinusoid divided by the mean count rate ) has increased smoothly from 20% just after the outburst onset , up to almost 50% at the end of the visibility window , 22 days after the bat trigger .
the pulse shape is variable too , from a multi - peak profile at the beginning of the outburst , to a less structured sinusoid during the latest xrt observations , though the lower statistics of the latter datasets might have hidden a more complex behavior . to assess the significance of the observed pulse shape variations as a function of time , we compared the folded pairs of contiguous lightcurves reported in figure[figure : efold ] by using a two - dimensional kolmogorov - smirnov test @xcite .
the results show that the probability that they do not come from the same underlying distribution is @xmath49@xmath70 ( t1-t2 ) , @xmath02.2@xmath70 ( t2-t3 ) , @xmath03.3@xmath70 ( t3-t4 ) , @xmath02.8@xmath70 ( t4-t5 ) , @xmath03.8@xmath70 ( t5-t6 ) , and @xmath02.9@xmath70 ( t6-t7 ) . for the spectral fitting ( with xspec version 12.4 )
, the data were grouped so as to have at least 20 counts per energy bin .
the ancillary response files were generated with xrtmkarf , and they account for different extraction regions , vignetting and point - spread function corrections .
we used the latest available spectral redistribution matrix ( v011 ) in caldb .
spectral modelling was performed by fitting together all the datasets in the 110 kev energy range ( the high absorption value made the data below 1.5 kev almost useless ) .
we first fit the spectra to simple models : a power law and a blackbody , both modified for interstellar absorption .
all the parameters were left free to vary , except for the absorption column density , that was left free but with the request to be the same for all observations .
both models provide acceptable fits .
the power - law fit ( @xmath99 for 1292 d.o.f . ) yields a column density @xmath100 @xmath2 and photon indices that increase from @xmath101 to 4 ( see figure [ spec_evol ] , lower panel ) . in the data taken immediately after the first bat trigger ,
was found to be in a record high state : the observed flux was @xmath0@xmath102 in the 210 kev band , about 100 times greater than that observed in 2008 july
september @xcite , and @xmath010 times greater than the previous highest recorded flux @xcite .
the observed flux declined by 70% in three weeks and the fading trend ( see figure [ spec_evol ] , upper panel ) can be described by a power - law with decay index @xmath103 ( using the bat trigger as origin of time ) .
+ from the blackbody fit ( @xmath104 for 1292 d.o.f . )
we derived a lower column density of @xmath105 @xmath2 .
again there is a clear indication of spectral softening during the decay , with the blackbody temperature changing from @xmath106 to 0.7 kev ( while the radius increased from @xmath01 to 3 km ) . for those observations for which the statistics was good enough ( those obtained during the first day since the bat trigger ) , the @xmath107 was approximately in the 1.21.3 range ( for 549 d.o.f . ) , suggesting a more complex spectrum . in order to check this issue ,
we re - analyzed the first four xrt / pc observations ( see table[log ] ) fitting a bb+pl model .
the results are shown in figure[spec ] and can be summarized as follows : ( a ) the first spectrum ( 99s after the bat trigger ) can be totally ascribed to a hard power - law ( @xmath108 , 90% c.l . ) , ( b ) during the second and the third observations ( between 0.05 and 0.2 days after the trigger ) the bb component becomes dominant ( @xmath109 kev in both cases ) , though a pl component is still needed ( but not statistically significant ) above 46 kev , ( c ) after one day since the bat trigger ( fourth observation ) the pl component is not any more detectable in the bb+pl model , while in the single pl model the photon index becomes @xmath44 ( mimicking a wien tail of a bb component ) . given the relatively low number of counts we did not attempt to fit more sophisticated and physical based models .
following the flux increase detected by /xrt we triggered our target of opportunity observation program .
@xcite observed the source from 2008 october 8 at 21:35 ut until october 10 at 04:53 ut for a total exposure time of 95.8 ks .
we analyzed the ibis @xcite data , the coded mask imager on board .
in particular we selected the data derived from its low energy ( 15 kev1 mev ) detector plane isgri @xcite , which is the most sensitive instrument below 100 kev .
we analyzed the 48 pointings that composed our observation , and derived the corresponding mosaicked images in two energy bands : 1860 kev and 18100 kev .
the source is not detected during our observations , and the 3@xmath70 upper limits at the source position are 0.36 and 0.42 counts s@xmath3 in the two energy bands respectively .
this corresponds , assuming a power law spectral shape with photon index @xmath110 , to a flux upper limit of @xmath111 in the 1860 kev band , and @xmath112 in the 18100 kev band . assuming a photon index @xmath113 the upper limits are @xmath11 in the 1860 kev band and @xmath114 in the 18100 kev energy band . in order to search for short bursts , we extracted event lists from each pointing .
the events were then selected in energy ( 1860 kev ) , and extracted from the pixels that were illuminated by the source direction for at least 60% of their surface . from these event lists , we derived light curves with a binning of 0.05 s. no burst was detected above the background level .
thanks to the capability of detecting high energy impulsive events with the bat , and of rapidly pointing to the bat position with the narrow field instrument xrt , we had the chance of studying the short burst and long - term outbursting properties of one of the most interesting objects of the magnetar class , , which entered a renewed phase of activity on 2008 october 3 . in particular , given the relatively large period derivative of the source @xcite , we set up an observational strategy so that daily pointings were performed during a three week baseline ( during which the source was observable by ) .
we studied the pulsed emission and inferred a rather accurate phase - connected timing solution which also includes a second derivative component .
the spectral evolution was also studied as a function of the decreasing flux level .
the connection between the emission of short x - ray bursts and the increase of the persistent x - ray emission of axps and sgrs , has been reliably assessed in the last few years , especially thanks to the prompt _ swift _ x - ray follow - up observations after the detection of short bursts in , , and ( @xcite ) . furthermore , magnetar outbursts have been found to be characterized by an intensity - hardness correlation , the harder the more intense ( and viceversa ) , as also observed in . both the connection with the occurrence of the bursts , and the hardness intensity correlation point toward a scenario in which these events are related to the evolution of a twisted magnetic field , which , once implanted by sudden crustal motions , necessarily decays @xcite .
particularly interesting are the burst energetics recorded by bat .
in fact , the 15150kev fluence of the bursts reported in table[table_bat_spec_fit ] is in the @xmath115ergs range ( inferred for a distance of 4kpc ; @xcite ) .
these estimates are conservative lower limits to the total energy emitted in the bursts by the source . on the other hand ,
the upper limit on the persistent emission of the source , as seen by xrt , is of the order of @xmath11610@xmath38ergs ( by integrating the exponential flux decay law from the beginning of october 2008 to mid - january 2009 ; see also first panel of figure[spec_evol ] ) , corresponding to a ratio between energy released in the bursts and in the persistent tail @xmath42 .
the latter value is in between those observed for sgrs , i.e. @xmath410 , and those of axps where the typical ratio is of the order of unity or less .
the x - ray timing properties of in october 2008 are such that a strong second derivative component , beside the usual @xmath5 and @xmath6 ones , is needed in order to account for the evolution of the pulse phases during the about three weeks covered by observations .
the latter component is @xmath92 s s@xmath93 or @xmath97 hz s@xmath93 , therefore with the same sign as that of the first derivative .
second period derivatives have been detected only for a few axps / sgrs and , in all cases , occurred in response to glitches , such as in the case of @xcite and @xcite , and/or outbursts , @xcite and @xcite . in nearly all cases
the second derivatives were relatively small , while the sign of these components was such that it acts in decelerating the spin down rates ; typical @xmath117 values are approximately in the 0.05@xmath655 @xmath3710@xmath118hzs@xmath93 range . only for one axp ,
namely , negative second period derivatives in the [email protected] @xmath3710@xmath119hzs@xmath93 range has been detected just after a glitch @xcite .
the timing properties of were already known to be unusual during the first 6 months of radio monitoring ( june
november 2007 ; @xcite ) . even in that case
, observations were taken after a period of activity and also at that time both the first and second frequency derivatives were negative : @xmath120 hz s@xmath3 and @xmath121 hz s@xmath93 .
moreover , during the same time span in which the @xmath122 ( and torque ) increased , the x ray flux decreased , a property which is difficult to account for with simple assumptions . according to @xcite , a possibility is that an extra torque ( either than the magnetic spin down ) originated from a particle wind that twists the magnetic field and , at the same time , transfers some angular momentum . in any case ,
the observed phenomenology calls for a mechanism in which x - ray luminosity and spin - down rate are not caused by the same mechanism or , at least , not by the same mechanism and also at the same time .
even more dramatic is the case of october 2008 reported in this paper , with showing a @xmath117 which is a factor of about 50 larger than in 2007 , while the x ray flux is a factor of 430 higher than during 2007 .
it is very likely that the @xmath117 components related with glitch - only events have a different origin than those related to outburst events .
for another transient axp , , a @xmath117 component has been detected during an outburst , although after two years since the onset @xcite .
correspondingly , a straightforward comparison with the properties is not feasible ( similar arguments also hold for the radio properties of the two sources ) . in the case of ,
the second derivative component was inferred during the first two months since the outburst onset and has a sign opposite to that of .
a plausible scenario is that the dissimilar behavior seen in the various sources is related to the difference between the outburst @xmath123 and the pulsar secular one @xmath124 .
regardless of the mechanism responsible for the outburst and for the net timing effects at the outburst onset , it is reasonable to imagine the pulsar as a system which tends to recover its ( almost ) initial timing properties , i.e. same pre - outburst and post - outburst @xmath124 @xcite .
correspondingly , a negative or positive second period derivatives might be ascribed to whether the @xmath123 over @xmath124 ratio is larger or smaller than unity .
unfortunately , for all the four magnetars for which an outburst was detected in the latest few years ( , , and ) the @xmath124 is unknown . in the case of the positive second derivative detected in 2007 and 2008
would imply a @xmath123 smaller than @xmath124 : the latter being the value to which the source is approaching asymptotically . within the classical scenario of isolated neutron stars ,
@xmath124 is the spin - down value to be used when inferring the pulsar dipolar magnetic field strength .
correspondingly , the inferred value of @xmath125 gauss should be considered as a lower limit on the intrinsic dipolar magnetic field strength .
taken at face values , the increase of @xmath126 during the outburst decay appears difficult to reconcile with the predictions of the twisted magnetosphere model . in such model ,
in fact , the spin - down rate is dictated by the radial variation of the external field .
a globally twisted dipolar field has radial dependence @xmath127 , where the index @xmath128 decreases with increasing twist angle ( @xmath129 ) .
if , as it seems natural to assume , the twist decays as the source returns to quiescence , one expects a decrease in the spin - down rate , i.e. a negative @xmath130 .
as pointed out by @xcite , however , in realistic situations the twist is likely to be confined to a rather narrow bundle of current - carrying field lines . in this case , and if the twist is suddenly implanted but its strength is moderate ( twist angle less than @xmath01 rad ) , the twist may still grow for a while in spite of the luminosity released by dissipation monotonically decreasing . only after the twist angle has reached its maximum value
, it will start to decrease , together with the torque .
this implies that the spin - down rate may be non - monotonic during the outburst decay , with a first phase characterized by a positive @xmath130 followed by a `` recovery '' during which @xmath131 and @xmath126 goes back to its secular value .
the typical duration of the increasing spin - down stage is difficult to quantify , but it can last @xmath0 months ( see eq . 55 in @xcite ) .
indeed , this effect has been already proposed by @xcite to explain the anti - correlation between x - ray luminosity and torque observed in 1e1048.1 - 5937 @xcite . in @xmath126
has been observed to increase monotonically during summer / fall 2007 ( during the decay of an outburst the peak of which has been missed ; * ? ? ?
the value of @xmath132 corresponding to the latest radio observation of @xcite is @xmath133 , and that in october 3 , at the beginning of the x ray outburst , @xmath134 s@xmath93 .
a continuous decrease in @xmath132 seems therefore hardly compatible with observations , and , since a constant @xmath132 has never been observed in axps , it may well be that @xmath132 started increasing at some epoch between 2007 november 22 ( latest radio observations ) and 2008 october 3 ( x - ray outburst onset )
. this behavior would not be in contradiction with the untwisting magnetosphere model mentioned above .
finally , it is worth mentioning the possibility that the unusual @xmath126 behaviour after the october 2008 outburst might be one of the reasons for the extremely intense outburst displayed by in january 2009 . in other words , the positive @xmath130 would be acted in accelerating the onset of stresses on the ns surface .
the pulsed fraction of clearly increases during october 2008 , a behavior already observed during the initial phases of axp outbursts , such as the case of where the quiescent pulsed fraction ( @xmath080% ) dropped to @xmath010% at the outburst onset and started increasing thereafter towards the quiescent value ( @xcite ; similar results have been recently obtained for ; @xcite ) . however , in the case of the pulsed fraction at the outburst onset ( @xmath020% ) is already larger than the quiescent value ( as seen by ; @xmath07% ) and it reached @xmath050% at the end of the october 2008 monitoring window .
this finding , together with the very variable pulse profile as a function of time , suggests that the emission and/or magnetospheric geometry is complex and variable during the initial phases ( on month timescales ) of the outburst , and likely not easily related to the geometry in quiescence .
pulsations in the x - ray flux may be a consequence of emission from a limited area of the star surface , of non - isotropic magnetospheric emission , or a combination of both effects .
again , if the star surface is heated by the returning currents which appear when a magnetospheric twist is implanted , even considering the simple evolution of the twist dictated by ohmic dissipation , an increase of the pulsed fraction in time is expected because , as time elapses , the bundle of current - carrying field lines shrinks and the heated area decreases . on the other hand ,
an increase in both pulsed fraction and emitting area , which is not excluded by the present data , would point toward a complete disassociation between the hot spot and the timing properties , with the latter perhaps of magnetospheric origin .
the xrt statistics is not enough to unambiguously assess the evolution of the emitting area , nor even the presence of two distinct spectral components ( bb and pl ) detected by in 2006 and 2007 , although there are several hints suggesting that also in the spectra observed during october 2008 two components are needed .
overall , the spectral properties of are also very similar to those of other transient magnetars , with an evident decay of the flux and a softening of the emitted photons ( at least in the 110 kev range ) as the outburst evolves in time .
the spectral softening is apparent while using both the bb and the pl spectral models .
referring , for concreteness , to the power - law model , the photon index switches from @xmath135 at the outburst onset to @xmath136 after @xmath137 d. a softening is expected if the spectrum originates in a globally twisted magnetosphere by resonant up - scattering .
as the twist angle decreases the charge density decreases making scattering less efficient .
this implies that the high - energy tail filled up by up - scattered thermal photons becomes less and less populated , producing a softening of the spectrum .
this has been verified by detailed monte carlo simulations in the case of a globally twisted dipolar field ( e.g. @xcite ) .
although no detailed model has been presented as yet , it seems likely that the same picture holds if a localized twist is considered instead .
as it was mentioned earlier , the twist angle may still increase for a while following the onset of the outburst , which would be consistent with the negative @xmath7 , while its spatial extent decreases @xcite .
this latter effect may compensate for the first one and be the main effect responsible for the spectral evolution : in particular , if the region in which photons can be efficiently scattered by the currents is reduced , a flux softening and decrease in luminosity are expected .
this research is based on observations with the nasa / uk / asi mission .
we thank the duty scientists and science planners for making these observations possible .
the italian authors acknowledge the partial support from asi ( asi / inaf contracts i/088/06/0 , i/011/07/0 , aae th-058 , aae da-044 , and aae da-006 ) .
pe thanks the osio sotto city council for support with a g. petrocchi fellowship .
sz acknowledges support from stfc .
nr is supported by a ramn y cajal fellowship .
dg acknowledges the cnes for financial support .
s. d. , barbier l. m. , cummings j. r. , fenimore e. e. , gehrels n. , hullinger d. , krimm h. a. , markwardt c. b. , palmer d. m. , parsons a. , sato g. , suzuki m. , takahashi t. , tashiro m. , tueller j. , 2005 , space science reviews , 120 , 143 j. e. , burrows d. n. , nousek j. a. , abbey a. f. , ambrosi r. m. , bruninger h. w. , burkert w. , campana s. , et al .
2004 , in flanagan k. a. , siegmund o. h. w. , eds , x - ray and gamma - ray instrumentation for astronomy xiii .
5165 of spie conference series , bellingham wa , readout modes and automated operation of the swift x - ray telescope .
pp 217231 g. , stella l. , covino s. , campana s. , angelini l. , mignani r. , mereghetti s. , marconi g. , perna r. , 2004 , in iau symposium unveiling the multi - wavelength phenomenology of anomalous x - ray pulsars . | we report on the detailed study of the 2008 october outburst from the anomalous x - ray pulsar ( axp ) discovered through the /burst alert telescope ( bat ) detection of sgr - like short x - ray bursts on 2008 october 3 .
the /x - ray telescope ( xrt ) started observing the source after less than 100s since the bat trigger , when the flux ( @xmath0@xmath1 erg @xmath2 s@xmath3 in the 210 kev range ) was @xmath450 times higher than its quiescent level .
monitored the outbursting activity of on a daily basis for approximately three weeks .
this strategy allowed us to find a phase - coherent solution for the source pulsations after the burst , which , besides @xmath5 and @xmath6 , requires a positive @xmath7 term ( spin - down increase ) .
the time evolution of the pulse shape is complex and variable , with the pulsed fraction increasing from 20% to 50% within the observational window .
the xrt spectra can be fitted well by means of a single component , either a power - law ( pl ) or a blackbody ( bb ) . during the very initial phases of the outburst
the spectrum is hard , with a pl photon index @xmath8 2 ( or @xmath9kev ) which steepens to @xmath84 ( or @xmath10kev ) within one day from the bat trigger , though the two components are likely present simultaneously during the first day spectra .
an observation carried out five days after the trigger provided an upper limit of @xmath0@xmath11 to the emission of in the 1860kev band .
[ firstpage ] stars : neutron x - rays : bursts x - rays : individual : . |
first - principles descriptions of quasiparticle excitations in extended and large confined molecular systems are prerequisite for understanding , developing and controlling molecular electronic , optoelectronic and light - harvesting devices . in search of reliable theoretical frameworks , it is tempting to use kohn - sham density functional theory ( dft),@xcite which provides accurate predictions of the structure and properties of molecular , nanocrystal and solid state systems .
however , kohn - sham dft ( ks - dft ) approximations predict poorly quasiparticle excitation energies both in confined and in extended systems,@xcite even for the frontier occupied orbital energy , for which ks - dft is expected to be exact.@xcite this has led to the development of two main first - principles alternative frameworks for quasiparticle excitations : many - body perturbation theory , mainly within the so - called gw approximation@xcite on top of dft , @xcite and generalized
ks dft.@xcite recently , range - separated hybrid ( rsh ) functionals @xcite combined with an optimally - tuned range parameter @xcite were shown to very successfully predict quasiparticle band gaps , band edge energies and excitation energies for a range of interesting _ small _ molecular systems , well matching both experimental results and gw predictions.@xcite the key element of the range parameter tuning is the minimization of the deviation between the highest occupied orbital energy and the ionization energy@xcite or the direct minimization of the energy curvature.@xcite the use of gw and the optimally - tuned rsh ( ot - rsh ) approaches for describing quasiparticle excitations in _ extended _ systems is hampered by high computational scaling .
the computational bottleneck in gw is in the calculation of the screened potential within the random phase approximation ( rpa ) while in ot - rsh it is the application of non - local exchange to each of the molecular orbitals . ot - rsh
is a self - consistent method and should therefore be compared to self - consistent gw calculations ; however , the latter are extremely expensive as the self - energy operator must be applied to all dyson orbitals .
recently , we proposed a stochastic formulation limited to the @xmath1 approach , where the computational complexity was reduced by combining stochastic decomposition techniques and real - time propagation to obtain the expectation value of the self - energy within the @xmath2 approximation.@xcite the stochastic gw ( _ _ s__gw ) was used to describe charge excitations in very large silicon nanocrystals ( ncs ) with @xmath3 ( @xmath4 is the number of electrons ) , with computational complexity scaling nearly linearly with the system size .
similar stochastic techniques have been developed by us for dft,@xcite for embedded dft,@xcite and for other electronic structure problems.@xcite here we develop a stochastic formalism suitable for applying the ot - rsh functionals for studying quasiparticle excitations in _ extended _ systems .
the approach builds on our previous experience with the exchange operator,@xcite but several new necessary concepts are developed here for the first time .
we start with a brief review of the ot - rsh approach , then move on to describe the specific elements of the stochastic approach , and finally present results .
we dedicate this paper to prof .
ronnie kosloff from the hebrew university to acknowledge his important contributions to the field of computational / theoretical chemistry .
kosloff has been our teacher and mentor for many years and his methods , such as the chebyshev expansions and fourier grids,@xcite are used extensively in our present work as well .
for a systems of @xmath4 electrons in an external one - electron potential @xmath5 having a total spin magnetization @xmath6 in the @xmath7 direction , the ot - rsh energy is a functional of the spin - dependent density matrix ( dm ) @xmath8 given in atomic units as : @xmath9 & = & tr\left[\rho\left(-\frac{1}{2}\hat{\mathbf{\nabla}}^{2}+v_{ext}\left(\hat{\mathbf{r}}\right)\right)\right]\\ & + & e_{h}\left[n\right]+e_{xc}^{\gamma}\left[n\right]+k_{x}^{\gamma}\left[\rho_{\uparrow},\rho_{\downarrow}\right],\nonumber \end{aligned}\ ] ] where @xmath10 is the range - parameter , discussed below , while @xmath11=\frac{1}{2}\iint u_{c}\left(\left|\mathbf{r}-\mathbf{r}'\right|\right)n\left(\mathbf{r}\right)n\left(\mathbf{r}'\right)d\mathbf{r}d\mathbf{r}'\ ] ] is the hartree energy functional of the density @xmath12 and @xmath13 is the coulomb potential energy .
@xmath14 $ ] is the unknown @xmath10dependent exchange - correlation energy functional which in practical applications is approximated .
the non - local exchange energy functional is given by @xmath15=-\frac{1}{2}\sum_{\sigma=\uparrow,\downarrow}\iint u_{c}^{\gamma}\left(\left|\mathbf{r}-\mathbf{r}'\right|\right)\left|\rho_{\sigma}\left(\mathbf{r},\mathbf{r}'\right)\right|^{2}d\mathbf{r}d\mathbf{r}',\ ] ] where @xmath16 .
this choice of @xmath17 accounts for long - range contributions to the non - local exchange energy and thus dictates a complementary cutoff in the _ local _ exchange - correlation energy , @xmath14 $ ] , to avoid over - counting the exchange energy.@xcite when the exact @xmath14 $ ] functional is used , minimizing @xmath18 $ ] with respect to @xmath19 under the constraints specified below leads to the exact ground - state energy and electron density @xmath20 . for approximate @xmath14 $ ] approximate estimates of these quantities are obtained . to express the constraints we first require the spin - dependent dm to be hermitian and thus expressible as : @xmath21 where @xmath22 and @xmath23 are its eigenvalues and orthonormal eigenfunctions .
the constraints are then given in terms of the eigenvalues @xmath24 as : @xmath25 @xmath26 @xmath27 the necessary conditions for a minimum of @xmath18 $ ] is that @xmath23 obey the generalized ks equations : @xmath28 where @xmath29 are the spin - dependent eigenvalues of the generalized ks hamiltonian ( @xmath30 and @xmath31 ) given by : @xmath32 note that the dm and its eigenstates minimizing the energy functional @xmath18 $ ] are themselves @xmath10dependent and are thus denoted by @xmath33 , @xmath34 ; the dm eigenvalues are not @xmath10dependent , as shown below .
the one - electron hamiltonian @xmath35 contains the kinetic energy , a local potential in _ @xmath36_space @xmath37 and a non - local exchange operator @xmath38 .
the local @xmath36space potential is further decomposed into three contributions : @xmath39 where @xmath40}{\delta n\left(\mathbf{r}\right)}=\int n\left(\mathbf{r}\right)u_{c}\left(\left|\mathbf{r}-\mathbf{r}'\right|\right)d\mathbf{r}'$ ] is the hartree potential and @xmath41}{\delta\rho_{\sigma}\left(\mathbf{r},\mathbf{r}\right)}$ ] is the short - range exchange - correlation potential .
the non - local exchange operator @xmath42}$ ] is expressed by its operation on a wave function @xmath43 of the same spin as : @xmath44 in this work we consider closed shell systems where @xmath45 and @xmath46 where @xmath47 is the number of electron pairs , i.e. , the level number of the highest occupied orbital . in this case , as in hartree fock theory and dft , the dm eigenvalues @xmath24 which minimize @xmath18 $ ] are @xmath48 if @xmath49 and 0 otherwise.@xcite hence , these conditions are used _ a - priori _ as constraints during the minimization of @xmath18 $ ] .
however , for the tuning process the ensemble partial ionization of an up - spin ( or down - spin ) electron needs to be considered .
thus , these values for @xmath22 are still used except for @xmath50 and @xmath51 where @xmath52 is fixed to be a _ _ positive fraction ( i.e. , the negative of the overall charge of the system , @xmath53 ) during the minimization of the gks ensemble energy @xmath18 $ ] ( for clarity , we abbreviate @xmath54 for the frontier orbital energy ( @xmath55 ) and occupation ( @xmath56 ) ) .
we note in passing that tuning is often done by combining a linearity condition from the @xmath57 electron system.@xcite we leave this for future work , and state that it can be done along the same lines as described here for the @xmath58 electron system .
the optimally - tuned range - parameter @xmath10 is determined from the requirement that the highest occupied generalized ks orbital energy @xmath59 is independent of its occupancy @xmath60 : @xmath61 through janak s theorem @xcite this equation implies that the energy curvature @xmath62 is zero . in practical terms , eq .
is solved by a graphical root search as shown in fig .
[ fig : gamma ] and discussed below .
in real - space or plane - waves implementations the application of the hamiltonian @xmath63 on a single particle wave function involves a pair of fast fourier transforms ( fft ) to switch the wave function between @xmath64space where the kinetic energy is applied and @xmath36space for applying the potential energy.@xcite therefore , for a grid of @xmath65 grid - points the operational cost is @xmath66 .
the ks hamiltonian operation scales quasi linearly with system size .
the scaling is much steeper for the rsh hamiltonian because the non - local exchange operator @xmath38 applies @xmath4 coulomb convolution integrals , each of which is done using an fft of its own thus involving @xmath67 operations .
therefore , the gks hamiltonian operation which scales quasi - quadratically is much more time consuming than the ks hamiltonian .
our approach , described next , reduces significantly the operation cost and even lowers the scaling due to the reduction of @xmath68 as the system size grows .
we first express the occupations in the dm in eq . as a combination of a occupations of a closed - shell density matrix and a remnant due to the overall charge of the molecule , @xmath69 ( assuming @xmath70 ; this separation reduces the stochastic error later when the charge of the system is continuously varied , as needed for the optimal tuning .
thus : @xmath71 where @xmath72 is the frontier orbital being charged and @xmath69 is the amount of charge .
when tuning the neutral system @xmath73 is the homo and it is being positively charged ( electrons removed from homo ) so @xmath74 .
when tuning for the anion @xmath75 is the lumo and the system being negatively charged ( electrons are added to the lumo ) so @xmath76 .
we assume without loss of generality that the spin of the charge frontier orbital is up .
next we evaluate the first term on the rhs of eq . using stochastic orbitals : @xmath77
where @xmath78 is a projected - stochastic orbital described in terms of the eigenstates of @xmath79 ( which can be alternatively obtained using a chebyshev expansion of the relevant projection operator @xcite ) : @xmath80 and @xmath81 is a stochastic orbital with a random sign ( @xmath82 ) at each grid - point .
( note that application of eq .
is itself a quadratic step however it is a `` cheap '' step as it is done only once in each scf iteration . ) with this , eq .
is rewritten as : @xmath83 @xmath84 next , we address the convolution in the random part of the above expression , by rewriting the range - separated coulomb potential as @xmath85 where @xmath86 , @xmath87 is the fourier transform of @xmath88 , and @xmath89 is a random phase between @xmath90 and @xmath91 at each @xmath64space grid point .
this can be seen by inserting the definition of @xmath92 into eq . and using the identity @xmath93}\right\rangle _ { \varphi}=\left(2\pi\right)^{3}\delta\left(\mathbf{k}-\mathbf{k}'\right)$ ] .
( see appendix [ sec : the - k=00003d0-term ] for the treatment of the @xmath94 term ) .
the non - local exchange operation is finally written as : @xmath95 @xmath96 in actual applications we use a finite number @xmath97 of pairs of stochastic orbitals @xmath98 and thus : @xmath99 the @xmath92 s are calculated once and stored in memory while the @xmath78 s are generated on the fly .
the computational scaling of the non - local exchange operation on @xmath43 is thus @xmath100 ( vs. @xmath67 for the deterministic case ) .
typically , @xmath101 and @xmath102 and thus , the operation of the stochastic non - local exchange becomes comparable in terms of computational effort to that of operating with the kinetic energy , so the computational cost of applying the gks hamiltonian is similar to that of the ks hamiltonian .
the new method has been implemented using the bnl functional @xcite for a series of hydrogen passivated silicon nanocrystals of varying sizes : @xmath103 , @xmath104 @xmath105 , @xmath106 and @xmath107 with real - space grids of @xmath108 , @xmath109 , @xmath110 , @xmath111 and @xmath112 grid - points , respectively .
we solve the generalized ks equations fully self - consistently using the chebyshev filtered subspace acceleration @xcite to obtain the occupied and low lying unoccupied eigenfunctions and eigenvalues .
the curvature for the different ncs , estimated from a forward difference formula @xmath113 with @xmath114 , is plotted as a function of @xmath10 in fig .
[ fig : gamma ] .
the curvature is a decreasing function of @xmath10 and has a node at the optimal value of the range parameter @xmath68 . for each
nc the curvature results are shown for several values of the number of stochastic orbitals @xmath97 .
we find that the statistical fluctuations near @xmath68 become smaller as the system grows and can be reduced with proper choice of @xmath97 .
for example , for the larger system the results near @xmath68 can be converged with only @xmath115 compared to the total number of occupied states for this system which is @xmath116 .
the reduction of these fluctuations is partially due to the decrease of @xmath68 itself as the nc size increases ( this decrease is shown in the right panel of fig . [
fig : gamma ] ) , leading to a smaller contribution of the non - local exchange to the orbital energies .
the results in the right panel of the figure also show that @xmath68 closely follows a linear function of @xmath117 .
we expect that for larger ncs with @xmath118 , this linear relation will break down and the optimal range parameter will converge to the bulk value , which through reverse engineering @xcite can be estimated as @xmath119 ( shown as a horizontal dotted line ) . such a localization - induced by the exchange has been seen for 1d conjugated polymers @xcite but not for bulk solids like silicon , likely due to the enormity of the calculation . in fig .
[ fig : nchi ] we plot the highest occupied molecular orbital ( homo , left panel ) and lowest unoccupied molecular orbital ( lumo , right panel ) energies obtained from the relations @xcite @xmath120}{\partial c}\right|_{c\to0^{+}}\nonumber \\ \varepsilon_{l,\uparrow } & = & -\left.\frac{\partial e_{rsh}^{\gamma}\left[\rho_{\uparrow},\rho_{\downarrow}\right]}{\partial c}\right|_{c\to0^{-}}\label{eq : orbital - energies}\end{aligned}\ ] ] respectively , as a function of @xmath97 at @xmath68 .
we find that determining the homo and lumo energies using the above first derivative relations reduces the noise compared to obtaining their values directly from the eigenvalues .
clearly @xmath121 and @xmath122 converge as @xmath97increases
. moreover , as the system size increases the fluctuations in @xmath121 and @xmath122 decreases for a given value of @xmath97 , consistent with the discussion above .
this is evident from the plot of the differences between the frontier orbital energies at adjacent values of @xmath97 .
= lower panel : comparison of the homo and lumo energies obtained using the _ _
s__gw approach ( black triangles ) and the stochastic rsh within the bnl functional ( red asterisk ) for a series of silicon nanocrystals .
the cyan diamond represent the deterministic rsh within the bnl functional .
upper panel : the corresponding quasiparticle band gaps .
also shown is the dft result within the lda ( blue circles ) .
, title="fig:",width=302 ] in the lower panel of fig .
[ fig : qp ] we plot the converged ( with respect to @xmath97 ) homo and lumo energies at @xmath68 for the series of silicon ncs . for the smallest system ( @xmath123 )
we compare the stochastic approach developed here with a deterministic rsh calculation using a non - local exchange with all occupied orbitals and obtaining the coulomb convolution integrals with ffts , thereby eliminating any source of statistical error . the purpose it to show that when the stochastic results are converged the agreement with a deterministic calculations is perfect on a relevant magnitude of energy .
we find that the homo energy increases and the lumo energy decreases with the size of the nc .
this is consistent with our recent calculations on silicon ncs using the stochastic gw approach , albeit the fact that there is a small shift in the quasi - particle energies obtained from the stochastic rsh approach compared to the _ _ s__gw .
indeed , a similar shift has been reported previously for much smaller silicon ncs.@xcite however , the source of this discrepancy is not clear , particularly , since the gw calculations were done within the so called _
@xmath124 _ limit , and
the ot - rsh often provides better quasiparticle energies in comparison to experiments.@xcite in the upper panel of fig .
[ fig : qp ] we plot the fundamental ( quasiparticle ) gaps . here , the agreement with the _ _
s__gw approach is rather remarkable , especially compared to the lda results which significantly underestimate the quasiparticle gap across all sizes studied .
.[tab : convergence - with-]optimally - tuned bnl frontier orbital energies and computational times @xmath125 vs. number of stochastic orbitals @xmath97 for two ( medium and large ) silicon clusters .
values for lda are also given for comparison .
as the system size grows , @xmath125 for the optimally - tuned bnl decreases relative to the lda timings due to decrease of @xmath68 . [ cols="^,^,^,^,^,^,^",options="header " , ] \(a ) @xmath126 \(b ) @xmath127 \(c ) in cpu - hrs in table [ tab : convergence - with- ] we provide numerical details of the calculations for the smallest and largest nc studied .
we report the results for the homo and lumo orbital energies for two different choices of @xmath128 comparing these two values we can conclude that the statistical errors for the lumo are very small ( @xmath129 ev ) for the largest nc and even the homo has small errors of around @xmath130 ev .
moreover , similar or even larger statistical errors are observed for the smaller nc for much larger values of @xmath97 , indicating that for a given accuracy the number of stochastic orbitals decreases with the system size .
this is partially correlated with the reduction of @xmath68 with the system size , as discussed above .
we have developed a stochastic representation for the non - local exchange operator in order to combine real - space / plane - waves methods with optimally - tuned range - separated hybrid functionals within the generalized kohn - sham scheme .
our formalism uses two principles , one is a stochastic decomposition the coulomb convolution integrals and the other is the representation of the density matrix using stochastic orbitals . combining these two ideas leads to a significant reduction of the computational effort and , for the systems studied in this work , to a reduction of the computational scaling of the non - local exchange operator , at the price of introducing a statistical error .
the statistical error is controlled by increasing the number of stochastic orbitals and is also found to reduce with the system size applications to silicon ncs of varying sizes show relatively good agreement for the band - edge quasiparticle excitations in comparison to a many - body perturbation approach within the _ _ s__gw approximation and excellent agreement for the fundamental band gap .
the stochastic approach has a major advantage over the @xmath131gw by providing a self - consistent hamiltonian which is central for post - processing , for example in conjunction with a real - time bethe
salpeter approach.@xcite the results shown here for @xmath3 and @xmath132 are the largest reported so far for the optimally - tuned range - separated generalized kohn - sham approach .
r. b. and e. r. gratefully thank the israel science foundation
first program ( grant no .
1700/14 ) .
r.b . gratefully acknowledges support for his sabbatical visit by the pitzer center and the kavli institute of the university of california , berkeley .
d. n. and e. r. acknowledge support by the nsf , grants che-1112500 and che-1465064 , respectively .
for accelerating convergence , it turns out to be better to remove the @xmath133 term from the the random vector expression representing the interaction , i.e. , @xmath134 this is because in practice the @xmath135term is very large .
analytically , this term is easily shown to commute with the fock hamiltonian and simply contribute a constant ( times the occupation ) to the eigenvalues and to the total energy , so it can be added a - posteriori : @xmath136 | we develop a stochastic formulation of the optimally - tuned range - separated hybrid density functional theory which enables significant reduction of the computational effort and scaling of the non - local exchange operator at the price of introducing a controllable statistical error .
our method is based on stochastic representations of the coulomb convolution integral and of the generalized kohn - sham density matrix .
the computational cost of the approach is similar to that of usual kohn - sham density functional theory , yet it provides much more accurate description of the quasiparticle energies for the frontier orbitals .
this is illustrated for a series of silicon nanocrystals up to sizes exceeding @xmath0 electrons .
comparison with the stochastic gw many - body perturbation technique indicates excellent agreement for the fundamental band gap energies , good agreement for the band - edge quasiparticle excitations , and very low statistical errors in the total energy for large systems .
the present approach has a major advantage over one - shot gw by providing a self - consistent hamiltonian which is central for additional post - processing , for example in the stochastic bethe - salpeter approach . |
recently , the cdf collaboration has reported @xcite some evidence for the production of heavy quarks with mass of order 174 gev at the tevatron .
the most likely explanation is , of course , the standard model ( sm ) top quark .
however , at present , in principle it is still possible @xcite that the signal is due to some other heavy quark , whereas the actual top quark is hiding below @xmath5 .
this is because the top quark semileptonic branching ratio ( br ) has not yet been measured .
if , for some reason , @xmath20 , the sm expected value , the top quark may have evaded detection .
this can arise basically only through scalar induced interactions @xcite .
one such scenario was proposed @xcite earlier by mukhopadhyaya and nandi ( mn ) . following a suggestion @xcite by barbieri and hall ( bh ) ,
mn considered the existence of @xmath21 singlet charge @xmath22 quark @xmath6 alongside sm fermions .
since gim is broken , the mixing of @xmath6 with up - type quarks induce tree level flavor changing neutral couplings of the sm higgs boson .
if @xmath23 , @xmath4 transitions may trigger the aforementioned mechanism of suppressing @xmath24 . in a subsequent paper , facing criticisms of naturalness " @xcite ( @xmath25 suppressions of heavy @xmath6 effects ) , mn retracted , and considered @xmath4 dominance to be not very likely @xcite . in this letter we study the precise conditions that @xmath4 dominance can be realised .
we find that this requires @xmath6 to be strongly mixed with _ both _ charm and top , which can occur even with a small @xmath7 eigenvalue .
however , we find that although @xmath4 can be dominant , it is unlikely to be overwhelmingly dominant
. thus , @xmath26 should occur at reduced but still substantial fraction .
this offers hope that , even if @xmath27 , the top quark can be uncovered at the tevatron by a renewed study with existing data
. the heavy quark @xmath6 can decay both via @xmath28 and @xmath11 bosons @xcite , hence it could be the heavy quark observed by cdf .
although one could not explain the larger than expected cross section for 174 gev quarks , one could plausibly account for the @xmath12-tagged @xmath29 4 jets events @xcite .
eventually , events with @xmath30 2 jets should start to emerge with increased luminosity @xcite .
another point of great phenomenological interest is @xmath15 decays , first stressed in this context by bh @xcite
. these occur at tree level again because of gim violation .
although widely known , the possibility apparently has not been studied with actual lep data because of the standard expectation of a very heavy top .
we estimate @xcite that @xmath15 could occur at the @xmath16 level , but _ with @xmath31 of order a few percent_. this results in a signal branching ratio of @xmath32 at the few @xmath33 level , and each lep experiment could have a few tens of events at present .
the background level can probably be managed , and lep experiments are strongly urged to conduct such a search .
besides the standard @xmath34-type quarks @xmath35 , @xmath36 ( @xmath37 ) , we add a left - right singlet charge @xmath2 quark @xmath38 , @xmath39 .
the left - handed singlet field @xmath38 can pair up with the four right - handed fields to form gauge invariant singlet masses , which we denote as @xmath40 and @xmath41 respectively .
the right - handed singlet field @xmath39 introduces three extra yukawa couplings , resulting in off - diagonal masses which we denote as @xmath42 .
thus , the @xmath34-type quark mass matrix is @xmath43,\ ] ] where @xmath44,\;\;\ ; \mbox{\boldmath $ s$ } = \left [ \begin{array}{rr } 0 & 0 \cr m^\prime & m \end{array } \right],\ ] ] are yukawa and singlet masses .
is diagonalised by a biunitary transform , @xmath45 where , departing from the notation of mn @xcite , @xmath46,\ ] ] and @xmath47 is a @xmath48 matrix .
the yukawa matrix is not simultaneously diagonalised , @xmath49 and the off - diagonal term controls fcnc @xmath50 and @xmath51 couplings . the apparent freedom due to the presence of @xmath52 rotation matrix on right - handed fields led mn originally to conclude that @xmath53 coupling could easily be rather large .
however , from eq .
( [ eq : mdiag ] ) , simple algebra gives @xmath54,\ ] ] where @xmath55 is the diagonal @xmath48 mass matrix ( see eq .
( [ eq : mdiag ] ) ) .
we see that no reference to @xmath52 is left , and the off - diagonal couplings depend only on mass eigenvalues and @xmath6-related mixing elements of @xmath56 @xcite .
the relevant flavor changing higgs couplings are @xcite ( @xmath57 ) @xmath58 the fcnc @xmath11 couplings are @xcite @xmath59 which is simply related to the higgs couplings . the charged current becomes @xmath60 where @xmath61 the @xmath48 km matrix @xmath62 is no longer unitary .
both @xmath62 and @xmath63 depend on the @xmath48 left - handed down quark rotation matrix @xmath64 .
we wish to explore the range of parameter space where @xmath65 coupling could be sizable . to this end
we make a special choice of basis to focus on the problem .
first , we choose @xmath66 fields such that @xmath67 in .
second , we choose @xmath35 , @xmath68 , such that the matrix @xmath69 is diagonal , hence the km matrix largely comes from the down - type quark sector ( we have checked that it is not possible to generate the observed km matrix structure just by introducing @xmath34-type singlet quarks ) .
only the charged current is affected by the @xmath70-type quark sector , the fcnc higgs and @xmath11 couplings depend only on @xmath71 and @xmath72 .
the @xmath34-type quark mass matrix is now in the form @xmath73 .
\label{eq : ansatz}\ ] ] the relevant freedom introduced by the singlet quark @xmath6 is parametrized as 3 new off - diagonal yukawa terms , plus the diagonal , gauge invariant dirac mass @xmath41 .
the parameters @xmath71 , @xmath74 and @xmath72 can be found by diagonalizing @xmath75 . without loss of generality , we set @xmath76 so @xmath34 quark decouples from our discussion . to illustrate the correlation between @xmath77 and @xmath78 , we set @xmath79 and plot , in fig .
1 , the mass eigenvalues @xmath80 , @xmath81 vs. @xmath82 for different @xmath83 values .
level repulsion is evident : @xmath84 and @xmath85 for @xmath86 , @xmath87 . for larger @xmath88 values
, we adopt the convention that , if @xmath89 , the heavier state is defined as the top quark . thus , fig .
1 depicts both the mass eigenvalues and the label for @xmath1 and @xmath6 .
we are more interested in the effect of @xmath90 . with finite @xmath90 , but negligible @xmath91 , @xmath92 , @xmath93 , the heavy mass eigenvalues are @xmath94 where @xmath95 . for sake of discussion , we consider the case where @xmath96 ( top lighter ) .
note that in fig . 1 when @xmath97 is not too large , the top mass eigenvalue is close to the diagonal term @xmath86 .
this is a generic feature .
when other @xmath98 s can be ignored and @xmath99 is not too big , the mass eigenvalue and mixing are roughly @xmath100 these relations become affected only when there are _ two _ @xmath101 values that are sizable , which follows largely as a consequence of unitarity of the @xmath102 matrix @xmath56 . in fig .
2 we plot @xmath103 , @xmath104 as a function of @xmath97 for @xmath105 and @xmath106 @xmath107 , @xmath108 , @xmath109 .
note the remarkable feature that @xmath104 is almost independent of @xmath103 , but @xmath103 is suppressed by large @xmath104 through unitarity . the physical reason for this can be traced back to the fact that @xmath110 is very small compared to 1 , and that @xmath7 is small .
thus , the eigenvalue @xmath111 could be made small by choosing a small value for @xmath93 , but this does not forbid @xmath112 from being sizable .
this is precisely _
the hierarchy principle " @xcite advocated in ref . @xcite . however , given that @xmath86 , @xmath113 are large , other than being a prejudice , there is really no reason why @xmath90 can not be large , since it is an independent parameter .
of course , if @xmath114 , then the conclusions of ref .
@xcite would hold .
inspection of eq . ( [ eq : hu ] ) suggests that @xmath115 transitions are suppressed by @xmath116 @xcite , but @xmath117 transitions have the effective coupling @xmath118 . since @xmath119 is not small , so long that @xmath120 is not too suppressed , the @xmath4 mode has good probability to be dominant over @xmath26 @xcite .
the necessary condition is therefore that both @xmath104 and @xmath103 are sizable and neither are suppressed .
hence , _ q , t and c all become rather arbitarily mixed _ although the charm mass is fixed by @xmath93 .
such an unusual situation is bound to have unusual consequences beyond @xmath121 being sizable . to illustrate the possibility of @xmath4 dominance , _ and _ at same time account for cdf s apparent observation of a heavy quark of mass 174 gev , we demand that the heavier quark ( whether dominantly doublet or singlet ) mass to be pinned to the cdf value .
we then choose @xmath122 , @xmath123 gev , vary @xmath86 ( to get @xmath124 , @xmath125 etc . ) and plot , in fig .
3 , @xmath126 vs. @xmath124 ( the physical mass ) up to 90 gev , for @xmath127 5075 gev .
we allow for @xmath128 below the present lep bound in case there are more than one higgs doublet @xcite . in producing fig .
3 , we compute the @xmath4 and @xmath129 decay width using the couplings of eqs .
( [ eq : hu ] ) and ( [ eq : wu ] ) .
we assume that @xmath64 amounts to a small " rotation close to the standard " km matrix , ignoring all phases .
we have also ignored @xmath130 decay as this is a three body process subdominant compared to @xmath26 .
it is clear that , if the higgs boson mass is sufficiently light , @xmath4 can be dominant .
however , the combined demand of @xmath131 174 gev , and @xmath9 , dictates that the @xmath4 mode can not be overwhelmingly dominant .
thus , although suppressed , @xmath132 should not be vanishingly small . for larger @xmath128 ,
@xmath4 dominance quickly fades , and the possibility is ruled out by cdf since @xmath26 is not drastically suppressed . in the following , we shall assume that one works in the domain where @xmath24 for the top " ( it could be the dominantly singlet quark , since we do not know the scale for @xmath41 ) is suppressed by @xmath133 or more .
that is , @xmath134 .
3 corresponds to @xmath135 @xmath136 , with @xmath137 gev .
the corresponding @xmath138 branching ratios are 0.51 , 0.24 , 0.04 , 0.1 , 0.02 , 0.09 , respectively .
note that , as a consequence of large @xmath104 , @xmath139 decay has a sizable rate ! the modes @xmath140 and @xmath141 are suppressed by phase space , while @xmath142 and @xmath143 are suppressed by an extra power of @xmath144 .
thus , @xmath28 induced decays are still dominant , but the @xmath12 content in final state is diluted slightly by the @xmath14 mode .
although one can not account for the large production cross section for the heavy quark ( _ one could always add another singlet @xmath34-type quark for this purpose _ ) , other features reported by cdf can be accounted for @xcite , in particular , the appearance of @xmath12-tagged @xmath145 4 jet events .
the @xmath11 boson comes from @xmath146 , @xmath141 , while a @xmath12-tag could come from @xmath147 , or from @xmath26 or @xmath148 , etc .
, in subsequent decays . this could be at @xmath149 of the @xmath150 cross section , hence consistent to what is observed .
on the other hand , single-@xmath28 with @xmath12-tag is slightly depressed ( @xmath151 ) compared to the standard top .
we therefore conclude that the heavy quark observed by cdf may well be a doublet - singlet mixed state @xmath6 .
the scenario offers many signatures and can be checked experimentally .
the light top with not too suppressed @xmath24 can perhaps be probed with existing tevatron data @xcite .
the scenario has a consequence that may be studied at lep . as first pointed out by barbieri and hall @xcite
, @xmath15 can be quite sizable with existence of charge 2/3 singlet quarks.using eq .
( [ eq : zu ] ) and @xmath104 , @xmath103 values for the example above , we estimate that @xmath152 is of order a few @xmath153 , which is consistent with ref . @xcite . other phenomenological constraints are not particularly stringent , and can be found in ref .
for example , @xmath154@xmath155 mixing constraint can be satisfied with small @xmath92 .
since @xmath156 does not vanish , we estimate that the potentially observable signal of @xmath157 2 jet ( where the jets contain @xmath12 and @xmath0 ) could have a branching ratio of order a few @xmath33 . since the lepton and neutrino should be well isolated with sizable ( @xmath158 gev ) @xmath18 or missing energy ( they are _ bona fide _ virtual @xmath28 decay events ! ) , and that @xmath159 , @xmath160 and one jet should pair up to be the top mass , there should be sufficient handles for the suppression of background .
the latter presumably comes from events with @xmath161 plus gluon bremsstrahlung .
this work is supported in part by the national science council of the republic of china under grant nsc-83 - 0208-m-002 - 023 .
wsh wishes to thank the cern theory division for hospitality and a stimulating environment , and t. ferguson for discussion .
cdf collaboration : f. abe _
et al . _ , fermilab - pub-94/097-e ; and phys . rev .
73 * , 225 ( 1994 ) .
hou , phys .
* 72 * , 3945 ( 1994 ) .
b. mukhopadhyaya and s. nandi , phys .
lett . * 66 * , 285 ( 1991 ) .
r. barbieri and l.j .
hall , nucl .
phys.*b319 * , 1 ( 1989 ) .
cheng and l .- f .
li , phys .
* d45 * , 1708 ( 1992 ) .
hou , phys . rev
69 * , 3587 ( 1992 ) .
b. mukhopadhyaya and s. nandi , phys .
rev . * d46 * , 5098 ( 1992 ) .
see , for example , v. barger and r.j.n .
phillips , preprint mad / phys/830 ( 1994 ) ; and references therein .
p. agrawal , s.d .
ellis and w.s .
hou , phys . lett . *
b256 * , 289 ( 1991 ) .
f. del aguila and m.j .
bowick , nucl . phys . *
b224 * , 107 ( 1983 ) .
v. barger _
_ , preprint mad / ph/844 and ral-94 - 070
. k. hikasa _
_ , the particle data group , phys . rev .
* d45 * ( 1992 ) s1 . | if @xmath0 and @xmath1 quarks are strongly mixed with a weak singlet charge @xmath2 quark , @xmath3 could be suppressed via the @xmath4 mode , thereby the top quark could still hide below @xmath5 , whereas the heavy quark signal observed at the tevatron is due to the dominantly singlet quark @xmath6 .
this may occur without affecting the small @xmath7 value .
demanding @xmath8 gev and @xmath9 , we find that @xmath3 can not be too suppressed .
the heavy quark @xmath6 decays via @xmath10 , and @xmath11 bosons .
the latter can lead to @xmath12-tagged @xmath13 jet events , while the strong @xmath0@xmath6 mixing is reflected in sizable @xmath14 fraction .
@xmath15 decay occurs at tree level and may be at the @xmath16 order , leading to the signature of @xmath17 , all isolated and with large @xmath18 , at @xmath19 order . |
heavy ion collision experiments at rhic typically create a medium where the net baryon density is non - vanishing @xcite . as we enter a new era of precision measurements , it is therefore important to consider the effect of a non - vanishing baryon chemical potential on the thermal photon yield of the quark gluon plasma . previously , the effect of non - zero chemical potential in @xmath2 processes was studied in @xcite .
the complete leading order calculation of photon production including the effect of collinear enhancement in the @xmath3 and @xmath4 cases was first carried out by arnold , moore , and yaffe ( amy ) starting from first principles within the framework of quantum field theory at finite temperature @xcite .
the analysis of @xcite , however , is mostly carried out at zero chemical potential .
the effect of adding a chemical potential was mentioned , but a detailed analysis was not fully performed .
the chemical potential will modify the quark and gluon self - energies as well as change the statistical factors , thereby potentially modifying the power counting analysis of amy . in this paper
, we will precisely determine under which circumstances the power counting must be modified to account for the presence of a chemical potential .
we also explore numerically the consequences of including a chemical potential in the photon production . as in the previous cases ,
the most convenient basis to work in when analyzing the parametric sizes of diagrams is the keldysh or @xmath5 basis whereas the rates are most conveniently written in the usual @xmath6 basis .
the addition of a finite chemical potential also changes the way in which we switch from the @xmath6 basis to the @xmath5 basis , and we will carefully explain the required changes .
the organization of this paper is as follows . in section [
sect : g1122 ] , we briefly recall the structure of perturbation theory in the real time formalism and outline the basic formula giving photon production in terms of feynman diagrams . in section [
sect : heinz_and_wang ] , we show how to generalize a result of heinz and wang @xcite that allows one to express green functions in the @xmath6 basis in terms of a reduced set of green functions in the @xmath5 basis . in section [
sect : hard_thermal_loops ] , we show how a finite chemical potential alters quark and gluon thermal masses .
then , in section [ sect : power_counting ] , we perform a power counting analysis to determine which diagrams contribute at leading order in the strong coupling @xmath7 and resum these in section [ sect : resummation ] .
finally , some numerical results are shown in section [ sect : numerical_results ] .
we start by briefly outlining the structure of real time perturbation theory at finite chemical potential .
since we are interested in a qcd plasma , our starting point is the qcd lagrangian : @xmath8 where the sum is over the @xmath9 fermion flavors and the gauge group is @xmath10 with @xmath11 .. ] this lagrangian has a conserved charge @xmath12 which is equal to the net fermion ( quark ) number .
the density operator describing the grand - canonical ensemble is therefore @xmath13 . in the imaginary time formalism
, one can show that this changes the matsubara frequencies from @xmath14 to @xmath15 @xcite .
the chemical potential here is hence the _ quark _ chemical potential which is 1/3 of the baryon chemical potential .
to connect the imaginary and real time formalisms , one defines the retarded and advanced propagators : @xmath16 the anticommutators above are replaced with commutators for gauge fields . by using the spectral representation of imaginary time propagators ( for instance , see @xcite ) one can show that the retarded propagator is obtained by analytically continuing the matsubara propagator through the prescription @xmath17 and the advanced propagator from the continuation @xmath18 , where @xmath19 is an arbitrary real number . in the real time formalism , one is forced to double the degrees of freedom because when constructing a generating functional from the partition function , the time integration contour has to traverse the real time axis once forward and backwards ( schwinger - keldysh closed time path ) @xcite .
the first set of fields @xmath20 , @xmath21 , and @xmath22 corresponds to fields with a time argument on the forward directed part of the contour and conversely the set of fields @xmath23 , @xmath24 , and @xmath25 corresponds to fields with time arguments on the backwards directed part . for computational purposes ,
it is sometimes more convenient to use another basis than the @xmath6 basis above .
one such basis is the @xmath5 or keldysh basis @xcite , defined as follows : @xmath26 where @xmath27 denotes any of our fields . by using the above algebraic relation and the spectral representation of propagators in the @xmath5 basis
, one can show that for fermions : @xmath28 the above allows us to compute real time propagators by analytically continuing imaginary time ones . from now on ,
we use capital letters for 4-momenta and lower case letters to denote the magnitude of the 3-momenta .
for instance , @xmath29 in the above expressions and @xmath30 . to tie this formalism to the problem of photon production , we recall the standard formula that relates the emissivity of the plasma to a wightman current - current correlator @xcite @xmath31 where we have defined the wightman current - current correlator by : @xmath32 wightman functions are most naturally expressed in the @xmath6 formalism of the real time theory : @xmath33 where @xmath34 is a fermionic four - point function where two external vertices are of the ` 1 ' types and the others are of the ` 2 ' types .
the `` parallel '' component of @xmath35 is always defined relative to the fixed direction of the emitted photon momentum @xmath36 . a schematic diagram for @xmath37 is shown in figure [ fig : momenta_assignments ] .
our convention is that fermion momenta flow into the shaded box .
arrows going into the shaded box correspond to insertions of the particle operator @xmath38 , and arrows coming out of it correspond to antiparticle insertions @xmath39 . for power counting and actual computations ,
the @xmath5 basis is more convenient than the @xmath6 basis . by using the algebraic relation between fields in the @xmath6 and @xmath5 bases ( c.f .
eq.([eq : def_ra_basis ] ) ) , we can express @xmath34 as a linear combination of the sixteen possible four - point functions in the @xmath5 basis .
the result is : @xmath40 one can show that in general , @xmath41 , and hence we can immediately eliminate one of the sixteen terms in the above
see for example @xcite . in ref.@xcite ,
wang and heinz showed that the remaining 15 terms can be re - expressed in terms of seven four - point functions and their complex conjugates .
the authors of ref.@xcite only considered the case of a real scalar field .
however , their derivation works equally well for fermions see appendix a. hence the following results from the reference still holds @xmath42 with the coefficients @xmath43 and @xmath44 composed of fermi - dirac distribution functions instead of bose - einstein distribution functions and @xmath45 denoting the charge conjugate ( @xmath46 ) of @xmath47 . here
all @xmath48 are functions of 4 momenta .
our convention is that the momenta @xmath49 and @xmath50 correspond to the insertion of @xmath38 and @xmath51 and @xmath52 correspond to the insertion of @xmath53 .
this expression greatly reduces the number of diagrams we need to estimate . as a matter of fact , what emerges from our power counting analysis is that only @xmath54 contributes to photon production at leading order .
this is because only this labeling gives rise to pinching poles as we will explain shortly .
thus , we only need to know the coefficients @xmath55 and @xmath56 .
these are given by @xmath57 this is proven in appendix a. one can intuitively understand the signs of @xmath58 in the above by recalling that @xmath59 is the distribution function for particles while @xmath60 is the distribution function for antiparticles . since @xmath51 and @xmath52 correspond to antiparticle insertions , they must be associated with the distribution @xmath61 , and conversely for @xmath49 and @xmath50 . as we will discuss in section
[ sect : power_counting ] , all gluon exchange momenta must be soft at leading order in the strong coupling @xmath7 . therefore
, we have that @xmath62 in figure [ fig : momenta_assignments ] .
hence , @xmath63 and @xmath64 and consequently , @xmath65 .
further , one can verify that @xmath66 see appendix b. therefore , only the real part of @xmath54 is relevant to photon production at leading order .
when we compute the wightman correlator to derive the photon production , we need to do a power counting analysis to identify all leading order diagrams . in this analysis ,
the appearance of `` pinching poles '' makes it crucial to resum the thermal self - energies into the quark and gluon propagators .
therefore , we need to know how the presence of a chemical potential affects the self - energies of quarks and gluons .
this is well - known and dates back to the original paper by braaten and pisarski on hard thermal loops @xcite . in order to keep our work self -
contained , we review the derivation of self - energies with a full inclusion of a chemical potential in appendix c. in this section , we simply quote the final results . the gluon polarization tensor at finite chemical potential @xmath58 takes the form : @xmath67 where in the above , @xmath68 is the debye mass , and @xmath69 with @xmath70 .
similarly , the quark self - energy at finite chemical potential is : @xmath71 the above two results show that the only effect of the chemical potential is to shift the dependence of self - energies on @xmath72 by a @xmath73 term .
as we will see , the consequence of this is that in carrying out our power counting analysis , we need not alter the arguments of arnold , moore , and yaffe as long as @xmath74 .
when evaluating the wightman function , there are two regions of the spatial part of the loop momentum integration that are of interest .
* the non - collinear region : @xmath75 and @xmath76 are both @xmath77 . * the near - collinear region :
@xmath75 or @xmath76 is @xmath78 or less .
the non - collinear region corresponds to the contributions of the basic @xmath79 processes treated by kapusta et al .
@xcite and baier et al . @xcite .
the near - collinear region will have both pinching pole and near - collinear enhancements .
it corresponds to the contribution of the bremsstrahlung and inelastic pair annihilation processes modified by the landau - pomeranchuk - migdal ( lpm ) effect .
we briefly review how pinching pole enhancements arise . as will be shortly shown ,
the leading order diagrams for photon radiation all contain a pair of fermionic propagators as shown in figure [ fig : pinching_poles ] .
the spatial part @xmath35 of the loop momentum @xmath80 is assumed to be nearly collinear with @xmath36 . in other words ,
@xmath81 . .
we assume that @xmath82 is nearly collinear with @xmath80 and that the other parts of the diagram are essentially constant over the region of width @xmath83 centered around the pinching poles of the integrand.,scaledwidth=40.0% ] given that we impose the collinearity of @xmath35 and @xmath36 , we may focus just on the frequency integral @xmath84\left[\left(p^0+k^0-\frac{i\gamma_{{\mathbf{p}}+{\mathbf{k}}}}{2}\right)^2-e_{{\mathbf{p}}+{\mathbf{k}}}^2\right ] } \label{eq : freq_int}\end{aligned}\ ] ] where we have absorbed the dirac matrix structures into the vertices . in eq.([eq : freq_int ] ) , @xmath85 is the decay width of the quark generated by the imaginary part of the self energy . to leading order in @xmath7 ,
it can be replaced by its asymptotic value @xmath86 ( @xcite ) .
the poles of the integrand are at the following locations : @xmath87 and @xmath88 .
it is straightforward to see that @xmath89 when @xmath90 are @xmath77 and collinear .
hence the two pole positions @xmath91 and @xmath92 almost coincide at @xmath93 although they are located on opposite halves of the contour ( c.f .
figure [ fig : contour_pinch_aligned ] ) . and
@xmath36 are nearly aligned .
the poles above the real line belong to @xmath94 and the ones below belong to @xmath95 .
the integration contour is along the real axis .
we see that two of the poles nearly coincide at @xmath96 .
these poles are said to `` pinch '' the integration contour.,scaledwidth=70.0% ] we use the residue theorem to close the contour and pick up the contributions of the poles located below the real axis .
for the case where @xmath97 , this gives : @xmath98 when @xmath99 , the factors in the denominators should be replaced by the general expression @xmath100 therefore , the frequency integral is approximately : @xmath101 but @xmath102 and @xmath103 are both of order @xmath104 , while @xmath35 and @xmath36 are hard .
hence , we get a @xmath105 enhancement from the frequency integral .
these enhancements make a large class of diagrams contribute to leading order even though we would naively expect them to be subleading .
suppose that in the above demonstration we had not taken the integrand to be the product of an advanced and a retarded propagator but rather @xmath106 or @xmath107 .
then all poles would have been on one side of the frequency integration contour only and we could have closed it on the side with no poles . since the contribution from great circles at infinity vanishes , this means that the integral would vanish . hence pinching
pole enhancements only arise when we have a retarded and an advanced propagator .
this is the reason why only @xmath54 contributes to the leading order photon production .
due to the fact that there is no @xmath108 propagator ( c.f .
eq.([eq : gaa ] ) ) and that an interacting vertex must contain an odd number of @xmath109-fields @xcite , any other labelling leads to reduction of the number of pinching poles .
we will illustrate general power counting arguments by estimating the size of the ladder diagram shown in figure [ fig : ladder_diagrams ] .
we start with the ladder diagram that has only a single rung ( figure [ fig : one_rung_ladder ] ) .
notice that the assignment of the @xmath5 indices ensures we have a retarded and an advanced propagator in each pair of quark propagators .
propagator.,scaledwidth=70.0% ] to get a pinching pole enhancement , we need that the quark with momentum @xmath80 be nearly collinear with the emitted photon .
further , if the gluon carries a soft exchange momentum @xmath110 , then it can not disturb the collinearity of the quark with the emitted photon .
therefore , the second pair of propagators also gives a pinching pole enhancement .
so we have a @xmath111 enhancement .
a soft gluon @xmath112 propagator is of order @xmath113 because of the bose - einstein factor , _
provided @xmath114_. however , a soft gluon also brings in a phase space suppression of order @xmath115 which cancels this enhancement .
gathering all powers , we have for suppressions : * @xmath116 from the integral over @xmath117 which is @xmath118 since this is the near collinear regime . * @xmath115 from the soft @xmath119 spatial integral over @xmath119 . * @xmath116 from two gluon exchange vertices . * @xmath116 from the @xmath120 factor arizing from the contraction with the external photons . and for enhancements : * @xmath111 from two pinching pole enhancements . * @xmath121 from the soft gluon propagator .
adding up all powers of @xmath7 , we get that this two - loop diagram is of order @xmath116 , the same as the non - collinear one - loop diagram .
the above analysis easily extends to ladder diagrams with an arbitrary number of rungs .
each additional rung brings with it one more pinching pole enhancement , giving a @xmath122 , and a soft gluon propagator of order @xmath121 .
it also brings in a @xmath115 suppression from the spatial integral over the new soft gluon momentum and a @xmath116 suppression from the two additional gauge boson exchange vertices . adding these up
, we get a net contribution of @xmath123 , so that at least all ladder diagrams contribute to leading order .
one ought to make an important remark at this point .
since the thermal mass of the gluon goes roughly as @xmath124 , the spectral function can be of higher order than @xmath122 if @xmath58 is of order @xmath125 .
in this case , the size of the soft @xmath112 gluon propagator fails to cancel the @xmath115 phase space suppression that we get from a soft gluon loop momentum .
consequently , all ladder diagrams are subleading when @xmath126 , and it becomes unnecessary to resum the contributions of all ladder diagrams .
the rest of the power counting analysis aims at identifying which combinations of external @xmath5 indices contribute at leading order , showing that all other diagram topologies are subleading , and also proving that gauge boson momenta of order higher or less than @xmath127 also give subleading contributions .
as we have argued , as long as @xmath128 , the size of the propagators comprising each diagram is parametrically the same as in the analysis of @xcite .
we will not pursue it here in more detail .
in the regime where @xmath114 , the power counting analysis of @xcite is unchanged . therefore , the conclusion that one needs to resum ladder diagrams ( and only these ) to get the photon emissivity to leading order is also valid .
hence , we can apply the same resummation procedure as @xcite . here
we briefly outline that resummation procedure .
graphically , the sum we need to evaluate is illustrated in figure [ fig : graph_opt_d ] .
sums all ladder diagrams with one photon vertex removed .
all external legs are amputated.,scaledwidth=95.0% ] we have defined the photon resummed vertex @xmath129 as the sum of all ladder diagrams with the leftmost photon vertex and pair of quark propagators removed . from figure
[ fig : graph_opt_d ] , one sees that each ladder is constructed from the previous one by concatenating a gluon rung attached to two quark - gluon vertices and a pair of quark propagators to the previous one , where the quark propagators must be evaluated at the location of pinching poles .
the ladder diagrams form a geometric series .
@xmath130 the graphical operator @xmath131 adds the pinching pole pair of propagators . in the above equation
, each @xmath132 contributes a pair of pinching poles : @xmath133\end{aligned}\ ] ] the operator @xmath134 adds the rung as illustrated in figure [ fig : graph_opt_m ] @xmath135 adds a pair of propagators evaluated at the location of the pinching poles .
the rung @xmath134 adds a soft gauge boson propagator . ]
adds a pair of propagators evaluated at the location of the pinching poles .
the rung @xmath134 adds a soft gauge boson propagator . ] finally , @xmath136 is the bare photon vertex on the far end of the ladder diagram as illustrated in figure [ fig : graph_opt_i ] .
.,scaledwidth=30.0% ] the resummed vertex @xmath137 then satisfies the following integral equation : @xmath138 after defining the variable @xmath139 , the integral equation ( [ eq : int ] ) becomes @xcite : @xmath140\end{aligned}\ ] ] after applying a sum rule , the collision kernel @xmath141 is found to have the simple form @xcite @xmath142 with @xmath143 and @xmath144 strictly colinear .
the term @xmath103 is the difference between the locations of the pinching poles of the quark propagators , as seen in eq.([eq : pinch_difference ] ) . through the energies of the quarks , it includes their asymptotic thermal masses @xmath145 .
therefore , the chemical potential enters through both @xmath146 and @xmath145 . recall that @xmath147 finally , the contribution of the @xmath0 and @xmath1 processes to the wightman correlator ( and thus to the photon emissivity ) is given by : @xmath148\times\nonumber\\ & \quad\times\mathcal{i}^{\mu}(p , k){\mathrm{re}}[\mathcal{f}(p , k)\mathcal{d}^\nu(p , k)]\end{aligned}\ ] ]
following a method developed by aurenche et al .
@xcite , we can solve equation numerically by solving the differential equation in impact parameter space . as in ref.@xcite
, we decompose the photon emission rate as follows @xmath149\end{aligned}\ ] ] where @xmath150 here , @xmath151 is the magnitude of the emitted photon s momentum and @xmath152 is the fermi - dirac factor . since the @xmath2 part of the spectrum has been already calculated @xcite , we only plot @xmath153 in figure [ fig : photon_plot ] .
the plots show the contribution of the @xmath0 and @xmath1 processes with the full treatment of the lpm effect .
the temperature of the plasma was taken to be @xmath154 .
and @xmath155 bremsstrahlung contribution to the photon emission .
the ratio against the @xmath156 case _ increases _ with increasing @xmath157 .
, scaledwidth=75.0% ] case _ decreases _ with increasing @xmath157 . , scaledwidth=75.0% ] we see from figure [ fig : photon_plot ] that as the ratio @xmath157 increases , the photon production rate also increases .
this finding is consistent with the behavior of the @xmath2 emission rates .
one of the reasons for this enhancement turned out to be just the statistical factors in eq.([eq : wightman ] ) .
these factors represent 3 different processes depending on the sign of @xmath19 and the relative sizes of @xmath158 and @xmath159 . to illustrate the effect of non - zero @xmath58 in each of the three processes , consider the following ratio of the statistical factors in eq.([eq : wightman ] ) and the same statistical factors with @xmath156 .
@xmath160 when @xmath161 , the underlying physical process is the bremsstrahlung from the quarks . in this case
, the ratio is mostly greater than 1 .
hence the rate is enhanced . this reflects the fact that a positive chemical potential enhances the number of quarks more than the pauli - blocking factor reduces the emission rate .
when @xmath162 , @xmath163 becomes the anti - quark phase space density . for @xmath162 and @xmath164 , the factor @xmath165 represents the pauli - blocking factor for the anti - quark bremsstrahlung . in this case
, a positive chemical potential reduces the phase density of anti - quarks but enhances the pauli - blocking factor .
but since pauli - blocking can never be enhanced above 1 , the effect is to reduce the photon emission rate .
the combined effect of quark bremsstrahlung and anti - quark bremsstrahlung is a net enhancement as shown in figure [ fig : brem ] . for @xmath162 and @xmath166 ,
the factor @xmath167 represents the density of quarks that can annihilate with anti - quarks to produce a photon with energy @xmath151 . in this case
@xmath168 is always smaller than @xmath169 , reflecting the fact that the annihilation process is necessarily controlled by the lesser number of anti - quarks . in figure
[ fig : annih ] , we show the contribution of the annihilation process to the photon emission rate for various values of @xmath58 .
among these three effects , the enhancement of the quark - bremsstrahlung dominates because it increases much faster than the reductions in the annihilation contribution and the anti - quark bremsstrahlung .
therefore , overall , the effect of having @xmath170 is to enhance the photon production at the same temperature as shown in figures [ fig : photon_plot_restricted ] and [ fig : photon_plot ] .
this trend was also observed in studies of @xmath2 processes @xcite .
in one previous study of @xmath171 processes @xcite , it was found that increasing the chemical potential _ decreases _ photon production .
however , this was for constant _ energy density _ instead of constant temperature . of the order of the temperature @xmath172 of the plasma , the photon emission rate increases by about @xmath173 relative to the @xmath156 value .
, scaledwidth=75.0% ] for @xmath174 . as @xmath157 increases , so does the ratio .
, scaledwidth=75.0% ] phenomenologically , the baryon chemical potential at rhic is about 30mev and at sps it is about 240mev@xcite .
the quark chemical potential is therefore about @xmath175 and @xmath176 for rhic and sps , respectively . with @xmath177 ,
the enhancement at rhic is negligibly small and it will be modest at sps , no more than 5% .
in this paper , we have explored the effect of non - zero baryon chemical potential in thermal photon production .
following the zero chemical potential study @xcite , we have computed photon production from non - abelian plasmas at leading order by resumming ladder diagrams to fully incorporate the lpm effect . after a careful analysis , we have found that as long as @xmath74 , the formulation in @xcite is still valid with appropriate changes in the statistical factors , thermal masses and the debye mass .
however , when @xmath178 , resummation of ladder diagrams is no longer necessary because thermal masses become @xmath77 instead of @xmath78 . hence , inverse powers of thermal masses no longer provide enhancements .
numerically , it is found that the inclusion of a chemical potential up to @xmath179 enhances the photon emission rate moderately , up to @xmath180 when @xmath181 .
this trend is also valid for hard photons from the @xmath2 process@xcite .
since the quark chemical potential is relatively small compared to typical qgp temperatures at rhic and the lhc , we expect a relatively small effect from a finite @xmath58 on thermal photon production at rhic and the lhc although it could be significant at sps energies and also for the lower energy runs at rhic .
h.g and s.j . are supported in part by the natural sciences and engineering research council of canada .
h.g is also supported in part by le fonds nature et technologies of qubec .
we gratefully acknowledge many discussions with c.gale , g.d.moore and j.cline .
also would like to thank j.f .
paquet for his help in numerical calculations .
in this appendix , our goal is to prove equation ( [ eq : change_basis ] ) .
this equation was proven in reference @xcite for the case of scalar fields .
their argument exploits the kms condition to derive a system of equations that allows them to solve for @xmath182 , @xmath183 , @xmath184 , @xmath185 , @xmath186 , @xmath187 , @xmath188 , and @xmath189 in terms of @xmath54 , @xmath190 , @xmath191 , @xmath192 , @xmath193 , @xmath194 , @xmath195 and their complex conjugates .
their procedure also goes through for dirac fermions in the grand canonical ensemble provided we modify the usual kms condition to include a chemical potential .
consider first the commutation relation between dirac field operators and the conserved charge @xmath196 : @xmath197=-\psi & \rightarrow e^{\beta\mu\hat{q}}\psi = e^{-\beta\mu}\psi e^{\beta\mu\hat{q}}\nonumber\\ \label{eq : commutator_q_psi}[\hat{q},{\bar{\psi}}]={\bar{\psi}}&\rightarrow e^{\beta\mu\hat{q}}{\bar{\psi}}=e^{\beta\mu}{\bar{\psi}}e^{\beta\mu\hat{q}}\end{aligned}\ ] ] for green functions in the @xmath6 basis , these commutation relations allow us to relate green functions whose @xmath6 index assignments are opposite each other . indeed , consider : @xmath198 note that the @xmath199 index refers to the space - time points @xmath200 and the @xmath169 index refers to the points @xmath201 .
we have also omitted spinor indices in the above for simplicity .
using the fact that @xmath202 is a time evolution operator in imaginary time ( @xmath203 ) , we obtain : @xmath204 the operators @xmath205 can be pulled outside of the reversed time - ordering symbol .
therefore , when writing out the thermal average explicitly in terms of the density operator @xmath13 , we obtain : @xmath206 we want to commute the operators inside the reverse time - ordering operator past @xmath207 and use the cyclicity of the trace to take them to the right of the operators inside the time - ordering symbol .
equation shows that in doing this , we pick up a factor of @xmath208 when we are commuting a @xmath209 operator and a factor of @xmath210 for a @xmath38 operator .
let us define the symbol @xmath211 to be equal to @xmath169 if the space - time point @xmath212 has a @xmath209 operator insertion at it , or @xmath213 if it has a @xmath38 insertion . with this notation
, we obtain the following generalization of the kms boundary condition to @xmath214-point functions : @xmath215 finally , we use the well - known formula @xmath216 to obtain : @xmath217 we are allowed to take the @xmath218 operator outside of the time - ordering symbol provided we take time - ordering to be given by the @xmath219 prescription .
that is , the time - ordered @xmath214-point function is _ defined _ to be the path integral of the product of the @xmath214 operators weighted by the exponential of @xmath220 times the action . for a discussion of this prescription , see for example @xcite .
we define the `` tilde conjugate '' @xmath221 of a green function @xmath47 to have same @xmath6 index assignment but with the time - ordering symbols reversed .
therefore , we can express our previous result as : @xmath222 in @xcite , wang and heinz use the fact that in momentum space , tilde conjugation is equivalent to complex conjugation .
as this is a key fact , we proceed to extend it to fermion fields at finite chemical potential . for notational simplicity
, we will focus on the case where there are four fermion operators two of which bear `` @xmath169 '' indices , and the other two of which bear `` @xmath199 '' indices .
although this is not necessary for this discussion , we have also included the @xmath223 matrices which appear in the current - current correlator [ eq : current_current ] .
extensions to other cases should be obvious .
@xmath224 taking a complex conjugation and changing @xmath225 , we obtain : @xmath226 next , we make use of @xmath227 invariance . recall that the dirac bilinear @xmath228 transforms as follows under the anti - unitary @xmath227 transformation @xmath229 : @xmath230 then , using anti - unitarity , thermal expectations of a product of fields can be related as follows : @xmath231 we have used @xmath227 invariance in commuting @xmath229
past @xmath232 and relabeling the eigenstates @xmath233 .
further , note that the sign of @xmath58 changes as we commute @xmath229 past @xmath207 . combining the above with our previous result , we obtain :
@xmath234 where @xmath45 denotes the charge conjugate ( @xmath46 ) of @xmath47 .
recapitulating , we have the two equations : @xmath235 the above can be taken as the starting point to reproduce the derivation of heinz and wang
. however , the latter also uses the energy conservation condition @xmath236 and the identity @xmath237 . in order for the derivation of heinz and wang to proceed in the same manner
, we need suitable generalizations of these to treat fermions at finite chemical potential . by charge conservation
, we must have that @xmath238 in other words , the green function in question must have an equal number of @xmath209 and @xmath38 insertions to not vanish .
therefore , we can define @xmath239 and maintain `` energy conservation '' @xmath240 .
also , by performing the customary replacement @xmath241 , we preserve the relation @xmath237 since @xmath242 in general . the only relations among the @xmath243 s that heinz and wang use is energy conservation , and the only property of the bose - einstein distributions they use is @xmath237 .
this is because after they have used the kms condition like we have done above , their work consists of algebraic manipulations such as solving large systems of equations involving the bose - einstein distribution . in their work , it is possible to treat @xmath47 and @xmath244 as independent variables rather than considering the real and imaginary parts of @xmath47 .
hence , it causes no harm to their derivation to replace @xmath244 by @xmath245 as is required by .
we conclude that we could in principle perform the exact same manipulations as heinz and wang by relabeling every @xmath246 as @xmath247 .
the net result is that we make the same replacements in their final results , bearing in mind that @xmath209 insertions come with a @xmath248 and @xmath38 insertions come with @xmath249 . in reference
@xcite , heinz and wang obtain @xmath250 and @xmath251 by making the replacements we have prescribed above , we get : @xmath252 and @xmath253 as required .
in this appendix , we want to show that @xmath254 .
it is sufficient to prove the corresponding identity in position space . @xmath255
in this appendix , we closely follow the zero @xmath58 analysis in @xcite to compute self - energies in the hard thermal loop approximation with a finite chemical potential .
consider first the gluon self - energy @xmath256 .
the color structure of this tensor is trivial , so that @xmath257 .
we have four diagrams to consider : the quark loop , the gluon loop , the gluon tadpole , and the gluon ghost .
only the quark loop is affected by the presence of a chemical potential and hence we make it the focus of our attention . with all momenta labeled , this diagram is shown in figure [ eq : quark_loop ] .
our strategy is to evaluate this diagram in the imaginary - time , invert it to get the gluon propagator , and then analytically continue the result to retarded frequencies to get the real - time @xmath258 propagator from which all other propagators in the @xmath5 basis may be obtained . proceeding ,
we first get : @xmath259 in the above , the integration measure is @xmath260 , where @xmath261 is the spatial momentum integration measure .
since we have factored out the color structure of the gluon self - energy , there is no sum over @xmath109 in the @xmath262 factor .
further , @xmath263 is the fermionic loop momentum and @xmath264 is the difference between the momentum of the gluon ( @xmath265 ) and the virtual quark .
the frequency @xmath266 is bosonic , and the frequency @xmath267 is fermionic .
the color matrices @xmath268 are the generators of @xmath269 in the fundamental representation . in this representation ,
the group factor @xmath262 is simply @xmath270 .
@xmath271 denotes the fermion propagator time evolved using the `` thermodynamics hamiltonian '' @xmath272 since this is what satisfies the periodic boundary condition ( see appendix a for kms conditions at finite @xmath58 ) .
the cleanest way to compute the above is to use the spectral representation @xmath273 indeed , this expansion of the propagator makes the frequency sum trivial . after performing the spin trace and employing standard identities involving the fermi - dirac distribution ,
we obtain : @xmath274 for the spatial component of the self - energy tensor , the result is @xmath275\label{eq : spatial_gluon__pol_quark_loop}\end{aligned}\ ] ] using @xmath276 .
the above expression holds for arbitrary external momenta .
henceforth , we only consider the hard thermal loop approximation so that @xmath277 after the analytic continuation @xmath278 has been performed .
it turns out that the self - energies in the hard thermal loop approximation are also valid for arbitrary external momenta @xmath80 at the one loop order @xcite .
hence , we will freely use our results as the self - energies of the gluons and quarks when estimating the size of their propagators .
using the fact that @xmath277 , we obtain : @xmath279\end{aligned}\ ] ] where @xmath280 . after an integration by parts and an application of the result @xmath281 we get : @xmath282\end{aligned}\ ] ] by symmetry
, we have @xmath283 .
thus , we finally obtain the hard thermal loop part of the gluon polarization tensor .
@xmath284 by going through the same analysis , we can evaluate the other components of the gluon polarization tensor . if we include the contributions of the three other loops , we finally have : @xmath67 where in the above
, @xmath68 is the debye mass , and @xmath69 .
the contribution of the quark loop to the gluon self - energy corresponds to the screening of the strong interaction by quarks in the medium .
it is therefore natural to expect that the chemical potential of the quarks will have an influence on the repeated scattering events that occur during the photon emission process .
the chemical potential also appears explicitly in the quark propagator and thus affects the thermal mass of the quarks .
this will affect the integral equation that we will derive for photon production , so we turn to computing the self - energy of the quarks .
the relevant diagram is shown in figure [ fig : quark_self_energy ] . in the above
, @xmath286 denotes a sum over bosonic frequencies followed by an integration over the spatial part of @xmath119 . as usual , @xmath287 where @xmath288 is the identity matrix in @xmath269 . in the coulomb gauge
, we find : @xmath289 where we have used the spectral representation of the longitudinal and transverse part of the coulomb gauge gluon propagator following @xcite .
here @xmath290 .
the above shows explicitly that we have three contributions to the quark self energy : @xmath291 .
the term @xmath292 is the contribution of the @xmath293 term in the longitudinal piece of the gluon propagator .
this corresponds to the instantaneous coulomb interaction term .
explicitly , it takes the form : @xmath294 once again , using the spectral representation of the fermion propagator , we obtain : @xmath295 this has no term proportional to @xmath296 .
thus , we will not pursue its computation any further , but we will rather focus on the transverse part . as for the longitudinal part , @xmath297 at tree level , so that it gives a vanishing contribution this is not true for dressed propagators , and in particular , @xmath298 does contribute to the quark damping rate . after employing the spectral representation of the fermionic propagators , we obtain for the transverse self - energy : @xmath299
retarded boundary conditions are obtained from the above by taking the analytic continuation @xmath300 . simplifying the dirac structure and applying the hard thermal loop approximation as in the gluon self - energy calculation , we get : @xmath71 with @xmath301 .
23ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1016/j.nuclphysa.2006.03.012 [ * * , ( ) ] , @noop ( ) , link:\doibase 10.1142/s021773239300101x [ * * , ( ) ] link:\doibase 10.1016/0370 - 2693(95)00004 - 5 [ * * , ( ) ] , link:\doibase 10.1103/physrevc.61.064911 [ * * , ( ) ] , link:\doibase 10.1016/j.physletb.2005.09.030 [ * * , ( ) ] @noop * * , ( ) , link:\doibase 10.1103/physrevd.66.025008 [ * * , ( ) ] , @noop _ _ ( , , ) @noop _ _ ( , , ) @noop * * , ( ) link:\doibase 10.1016/0370 - 1573(85)90136-x [ * * , ( ) ] link:\doibase 10.1016/0550 - 3213(90)90508-b [ * * , ( ) ] link:\doibase 10.1103/physrevd.44.2774 , 10.1103/physrevd.47.4171 , 10.1103/physrevd.44.2774 , 10.1103/physrevd.47.4171 [ * * , ( ) ] link:\doibase 10.1007/bf01625902 [ * * , ( ) ] link:\doibase 10.1016/0550 - 3213(96)00023 - 5 [ * * , ( ) ] , link:\doibase 10.1016/0003 - 4916(90)90040-u [ * * , ( ) ] @noop * * , ( ) , @noop * * , ( ) , @noop * * , ( ) , link:\doibase 10.1016/s0375 - 9474(01)01257-x [ * * , ( ) ] , @noop _ _ ( , , ) link:\doibase 10.1016/s0370 - 1573(01)00061 - 8 [ * * , ( ) ] , | we compute the photon production of a qcd plasma at leading order in the strong coupling with a finite baryon chemical potential .
our approach starts from the real time formalism of finite temperature field theory .
we identify the class of diagrams contributing at leading order when a finite chemical potential is added and resum them to perform a full treatment of the landau - pomeranchuk - migdal ( lpm ) effect similar to the one performed by arnold , moore , and yaffe at zero chemical potential .
our results show that the contribution of @xmath0 and @xmath1 processes grows as the chemical potential grows . |
combinatorial spiders can be used to describe the space of invariants of tensor products of irreducible representations for lie algebras and their quantum analogues .
kuperberg @xcite gave a description of the spiders for types @xmath14 , @xmath15 using certain combinatorial graphs called _ webs_. in this paper , we will focus on webs of type @xmath0 .
an @xmath0-web is a directed graph such that ( 1 ) vertices on a boundary line have degree one ; ( 2 ) internal vertices are trivalent ; and ( 3 ) each vertex is either a sink ( all its edges are directed inward ) or a source ( all its edges are directed outward ) .
a boundary vertex that is a sink corresponds to the natural three - dimensional module @xmath8 for @xmath6 or @xmath16 and carries a label `` @xmath2 '' ; while a boundary vertex that is a source corresponds to the dual module @xmath17 and is labeled `` @xmath4 '' .
an @xmath0-web is said to be _ nonelliptic _ if each interior face is bounded by at least six edges .
let @xmath18 be a partition of the integer @xmath19 .
the corresponding young diagram is a left - justified array with @xmath20 boxes in the first row , @xmath21 boxes in the second row , and so forth .
a _ standard tableau _ of shape @xmath22 is a filling of the young diagram with @xmath19 distinct positive integers in such a way that the entries in the boxes increase left to right across each row and from top to bottom down each column .
let @xmath23 be positive integers .
a _ semistandard tableau _ of shape @xmath22 and _ type _
@xmath24 such that @xmath25 is a filling of the boxes of @xmath22 with @xmath26 entries @xmath27 , @xmath28 entries @xmath29 etc .
, so that the entries in the boxes weakly increase left to right across the rows and strictly increase from top to bottom down each column .
petersen - pylyavskyy - rhoades @xcite re - interpreted kuperberg s result ( * ? ? ?
6.1 ) on the dimension of the space of invariants ( i.e. the number of @xmath0-webs ) in terms of semistandard tableaux of a particular shape and type .
[ thm : prp ] let @xmath30 be a fixed boundary with @xmath1 @xmath2"s and @xmath3 @xmath4"s .
the number of semistandard tableaux of shape @xmath31 and type @xmath32 is equal to the number of nonelliptic @xmath0-webs with boundary @xmath30 . in @xcite , tymoczko gave a bijective proof of theorem [ thm : prp ] for the case @xmath33 by exhibiting an explicit bijection between the set of standard tableaux of shape @xmath9 and the set of nonelliptic @xmath0-webs which have @xmath34
@xmath4"s on the boundary .
her approach introduces the notion of an _
@xmath35-diagram_. an @xmath35-diagram on @xmath36 vertices consists of a boundary line labeled with the numbers @xmath37 , together with a collection of arcs drawn above it . for a given standard tableau @xmath38 of shape @xmath9 , tymoczko associates an @xmath35-diagram to @xmath38 by the following algorithm : [ alg : tym ] 1 .
take the conjugate of @xmath38 ( transpose the rows and columns ) to obtain a standard tableau @xmath39 of shape @xmath40 .
2 . starting with the bottom row of @xmath39 , read the entries of @xmath39 from left to right across the rows .
if @xmath41 is an entry in this reading , connect vertex @xmath41 to vertex @xmath42 on the boundary line of the @xmath35-diagram with an arc , where @xmath42 is the largest number less than @xmath41 in the row immediately above the row with @xmath41 that has not already been connected with an arc . after finishing with the bottom row , move to the entries in the second row of @xmath39 and repeat the same procedure .
[ exam : first ] if @xmath43{\includegraphics{grimfiles / st1.pdf } } } } } } \,$ ] , then @xmath44{\includegraphics{grimfiles / st2.pdf } } } } } } \,$ ] .
starting with @xmath45 in the bottom row , connect it to @xmath46 , and then connect @xmath47 to @xmath7 . moving to the second row ,
begin with @xmath7 and connect it to @xmath48 , since @xmath48 is the largest entry less than @xmath7 in the first row . finally , connect @xmath46 and @xmath49 .
the resulting @xmath35-diagram associated to @xmath38 is @xmath50{\includegraphics{grimfiles / arc.pdf } } } } } } .\ ] ] the next step of tymoczko s procedure is to replace vertices in the @xmath35-diagram with certain configurations as pictured below .
@xmath51{\includegraphics{grimfiles / biarc1.pdf } } } } } } \longrightarrow { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.55}{\includegraphics{grimfiles / biarc2.pdf } } } } } { \vcenter{\hbox{\scalebox{0.55}[]{\includegraphics{grimfiles / biarc2.pdf } } } } } } , \
\quad { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.55}{\includegraphics{grimfiles / biarc11.pdf } } } } } { \vcenter{\hbox{\scalebox{0.55}[]{\includegraphics{grimfiles / biarc11.pdf } } } } } } \longrightarrow { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.55}{\includegraphics{grimfiles / biarc22.pdf } } } } } { \vcenter{\hbox{\scalebox{0.55}[]{\includegraphics{grimfiles / biarc22.pdf } } } } } } , \ \quad { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.55}{\includegraphics{grimfiles / biarc3n.pdf } } } } } { \vcenter{\hbox{\scalebox{0.55}[]{\includegraphics{grimfiles / biarc3n.pdf } } } } } } \longrightarrow { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.55}{\includegraphics{grimfiles / biarc4n.pdf } } } } } { \vcenter{\hbox{\scalebox{0.55}[]{\includegraphics{grimfiles / biarc4n.pdf } } } } } } .\ ] ] this is called _ trivalizing _ in @xcite , because vertices off the boundary become _ trivalent _ by this process .
for example [ exam : first ] above , the corresponding @xmath0-web is @xmath52{\includegraphics{grimfiles / web.pdf } } } } } } .\ ] ] in @xcite , russell exhibited a bijection between the set of semistandard tableaux of shape @xmath31 and type @xmath32 and the nonelliptic @xmath0-webs with one boundary line having @xmath1 `` @xmath4 '' vertices and @xmath53 `` @xmath2 '' vertices for arbitrary @xmath1 such that @xmath54 . in this paper
, we demonstrate an explicit bijection between the set of such semistandard tableaux and the nonelliptic @xmath13-webs having @xmath1 `` @xmath2 '' vertices on the top boundary line and @xmath3 `` @xmath4 '' vertices on the bottom boundary line .
such webs combinatorially model the space @xmath55 .
several noteworthy special cases are when @xmath33 , which corresponds to the space @xmath56 of @xmath6-invariants in @xmath57 , and when @xmath58 , which corresponds to the centralizer algebra @xmath59 of the @xmath6-action on the tensor space @xmath60 .
our approach uses classical results from the representation theory of the symmetric group @xmath61 such as the well - known robinson - schensted correspondence ( rs correspondence ) .
that the representation theory of @xmath61 plays an essential role is no surprise , as @xmath61 is closely related with the representation theory and invariant theory of the general linear and special linear groups via schur - weyl duality .
another example of the connection with the rs correspondence can be found in @xcite .
the rs correspondence ( see for example , ( * ? ? ?
3.1 ) or @xcite ) is a bijection between permutations and ordered pairs of standard tableaux of the same shape , consisting of an insertion tableau and a recording tableau . here
, we reverse the order of the standard tableaux in the pair , so that a permutation @xmath62 corresponds to @xmath63 , where @xmath64 is the recording tableau and @xmath65 is the insertion tableau .
this switch is simply a notational convenience for when we show in what follows that the pair of standard tableaux constructed using the notion of path depth is the same as the pair @xmath63 .
for example , suppose @xmath66 .
the tableau @xmath65 is created by inserting the numbers @xmath67 in succession , and @xmath64 is created by recording where the new box has been added to @xmath65 at each step . in the insertion process
, a number @xmath68 is adjoined to the right - hand end of first row of @xmath65 unless that violates the `` standard '' property .
when a violation occurs , the smallest @xmath69 in the row such that @xmath70 is bumped to the next row , and @xmath68 is placed in the position vacated by @xmath69 .
the insertion - bumping process then continues with @xmath69 until a number is adjoined to the end of a row in the tableau . for @xmath62 ,
the steps to construct @xmath63 are as follows : @xmath71{\includegraphics{grimfiles / tabp1.pdf } } } } } } \ , , \ , { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / tabq1.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / tabq1.pdf } } } } } } \right ) \longrightarrow \left({\ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / tabp2.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / tabp2.pdf } } } } } } \ , , \ , { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / tabq2.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / tabq2.pdf } } } } } } \right ) \longrightarrow \left({\ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / tabp3.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / tabp3.pdf } } } } } } \ , , \ , { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / tabq3.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / tabq3.pdf } } } } } } \right ) \longrightarrow \left({\ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / tabp4.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / tabp4.pdf } } } } } } \ , , \ , { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / tabq4.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / tabq4.pdf } } } } } } \right ) \longrightarrow \left({\ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / tabp5.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / tabp5.pdf } } } } } } \ , , \ , { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / tabq5.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / tabq5.pdf } } } } } } \right)\ ] ] in section [ sec : central ] , we review basic definitions related to @xmath13-webs and their connections with centralizer algebras of the @xmath6-action on tensor products . in section [ sec : bijection ] , we construct a map @xmath72 from semistandard tableaux to @xmath13-webs , and then prove in section [ sec : proof ] that the webs are nonelliptic .
the results on @xmath12-avoiding permutations in section [ sec : combinatorics ] are used in section [ sec : inverse ] to show that @xmath72 is indeed a bijection . in section [ sec :
flow ] , we illustrate how westbury s flow diagrams provide a different bijection , which avoids using the rs correspondence .
we obtained this alternate bijection after submitting the first version of this paper , and thank heather russell for bringing @xcite to our attention .
we note that the result on shuffle and join obtained in ( * ? ? ?
3 ) as an application of russell s bijection can be proved using our bijection also , although we do not include that proof here .
the @xmath13-spider category introduced by kuperberg in @xcite describes the invariants of tensor products of copies of the natural 3-dimensional representation @xmath73 of the complex lie algebra @xmath6 and its dual representation @xmath74 .
a tensor product of copies of @xmath75 and @xmath76 corresponds to a string @xmath77 of symbols @xmath2 and @xmath4 , and we write @xmath78 .
let @xmath79 for all @xmath80 denote the space of @xmath6-invariants in @xmath81 .
the objects in the @xmath13-spider category are finite strings of @xmath2 and @xmath4 signs , including the empty string .
the morphisms of the @xmath13-spider category are @xmath13-webs .
an @xmath13-web is defined combinatorially as a directed graph with a boundary labeled by the sign string @xmath77 so that 1 .
vertices on the boundary have degree one and are labeled by the components of @xmath82 in a counterclockwise fashion ; 2 .
internal vertices are trivalent ; and 3 .
each vertex is either a sink ( all its edges are directed inward ) or a source ( all its edges are directed outward ) .
such an @xmath0-web is said to be _ nonelliptic _ if each interior face is bounded by at least six edges .
the following is an example of nonelliptic @xmath13-web with boundary @xmath83 : @xmath84{\includegraphics{grimfiles / webexam.pdf } } } } } } \ ] ] no distinction is made between two diagrams that are the same up to rotation .
let @xmath85 denote the set of nonelliptic @xmath13-webs with boundary condition @xmath82 .
the vector space of formal linear combinations of @xmath85 is the @xmath13-web space @xmath86 .
[ t : inv ] ( ( * ? ? ?
* thm . 6.1 ) ) assume @xmath82 is a sign string .
then the vector spaces @xmath86 and @xmath87 have the same dimensions .
consider the special case that @xmath82 consists of @xmath1 minuses followed by @xmath1 pluses , so that @xmath88 .
in this situation , @xmath89 , since the @xmath6-action on @xmath90 is given by @xmath91 .
the above theorem implies that we may label a basis of the _ centralizer algebra _
@xmath92 by nonelliptic @xmath13-webs that have a bottom row of @xmath1 minus signs and a top row of @xmath1 plus signs .
for example , @xmath93{\includegraphics{grimfiles / webexample.pdf } } } } } } \;$ ] is an element in @xmath94 .
multiplication of @xmath13-webs is governed by the following rules : ( m1 ) @xmath95{\includegraphics{grimfiles / circrule.pdf } } } } } } = 3 $ ] ( m2 ) @xmath96{\includegraphics{grimfiles / bigonrule.pdf } } } } } } = 2 \ ; { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.7}{\includegraphics{grimfiles / linerule.pdf } } } } } { \vcenter{\hbox{\scalebox{0.7}[]{\includegraphics{grimfiles / linerule.pdf } } } } } } $ ] ( m3 ) @xmath97{\includegraphics{grimfiles
/ fourgonrule.pdf } } } } } } = { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.7}{\includegraphics{grimfiles / fourgonpass1.pdf } } } } } { \vcenter{\hbox{\scalebox{0.7}[]{\includegraphics{grimfiles / fourgonpass1.pdf } } } } } } + { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.7}{\includegraphics{grimfiles / fourgonpass2.pdf } } } } } { \vcenter{\hbox{\scalebox{0.7}[]{\includegraphics{grimfiles / fourgonpass2.pdf } } } } } } $ ] . for example
, @xmath98{\includegraphics{grimfiles / g1.pdf } } } } } } \ ; \right)^2 \ ; = \ ; { \ifthenelse{\isempty{0.8 } } { \vcenter{\hbox{\scalebox{0.7}{\includegraphics{grimfiles / mul1.pdf } } } } } { \vcenter{\hbox{\scalebox{0.7}[0.8]{\includegraphics{grimfiles / mul1.pdf } } } } } } \ ; \overset{\text{(m2)}}{=}\ ; 2\ ; { \ifthenelse{\isempty{0.8 } } { \vcenter{\hbox{\scalebox{0.7}{\includegraphics{grimfiles / mul2.pdf } } } } } { \vcenter{\hbox{\scalebox{0.7}[0.8]{\includegraphics{grimfiles / mul2.pdf } } } } } } \ ; \overset{\text{(m1)}}{=}\ ; 6\ ; { \ifthenelse{\isempty{0.8 } } { \vcenter{\hbox{\scalebox{0.7}{\includegraphics{grimfiles / g1.pdf } } } } } { \vcenter{\hbox{\scalebox{0.7}[0.8]{\includegraphics{grimfiles / g1.pdf } } } } } } \ ; .\ ] ] the product of @xmath99{\includegraphics{grimfiles / mulexam1.pdf } } } } } } \;$ ] and @xmath100{\includegraphics{grimfiles / f1.pdf } } } } } } \;$ ] is computed by placing the first graph above the second and concatenating the diagrams using the rules above : @xmath101{\includegraphics{grimfiles / mulexam2.pdf } } } } } } \ ; \ ; \overset{\text{(m3)}}{= } \ ; { \ifthenelse{\isempty{0.8 } } { \vcenter{\hbox{\scalebox{0.7}{\includegraphics{grimfiles / mulexam3.pdf } } } } } { \vcenter{\hbox{\scalebox{0.7}[0.8]{\includegraphics{grimfiles / mulexam3.pdf } } } } } } \ ; \ ; + \ ; \ ; { \ifthenelse{\isempty{0.8 } } { \vcenter{\hbox{\scalebox{0.7}{\includegraphics{grimfiles / mulexam4.pdf } } } } } { \vcenter{\hbox{\scalebox{0.7}[0.8]{\includegraphics{grimfiles / mulexam4.pdf } } } } } } \ ; .\ ] ] the thick edge indicates the portion of the web to which the rule is applied .
the symmetric group @xmath102 acts on @xmath60 by permuting the tensor factors , and this action commutes with the action of @xmath6 .
classical schur - weyl duality says that there is an epimorphism @xmath103 and that the kernel of this map is the ideal generated by the young symmetrizer @xmath104 , where the sum is over the permutations @xmath62 of @xmath105 .
the transpositions @xmath106 for @xmath107 generate @xmath102 .
we let @xmath108 for @xmath109 , denote the image of @xmath110 in @xmath111 .
then , the elements @xmath112 , generate @xmath111 , and they satisfy the following relations : * @xmath113 ; * @xmath114 if @xmath115 ; * @xmath116 for all @xmath117 ; * @xmath118 , where @xmath119 for all @xmath120 . relation ( r4 ) comes from the fact that if @xmath121 and @xmath122 in @xmath123 , then @xmath124 , as @xmath125 ranges over the permutations of @xmath126 , and @xmath62 ranges over the permutations of @xmath105 . here
@xmath127 is the element of @xmath128 that interchanges @xmath41 and @xmath129 for @xmath130 and acts as the identity on the other values .
the following relations can be deduced from ( r1)-(r4 ) : * @xmath131 for all @xmath132 , @xmath133 ; * @xmath134 ; * @xmath135 ; * @xmath136 .
each of the generators @xmath108 can be identified with a nonelliptic @xmath13-web as shown below , and one can verify that the multiplication above corresponds to web multiplication .
@xmath137{\includegraphics{grimfiles / generator1.pdf } } } } } } \\ { \mathsf{g}}_i \ \ \leftrightarrow&\ \
{ \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / gf1.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / gf1.pdf } } } } } } .
\end{aligned}\ ] ] the following figure illustrates @xmath138 ( equivalently , @xmath139 ) : @xmath140{\includegraphics{grimfiles / f1f2f1.pdf } } } } } } \ ; \overset{\text{(m3)}}{= } \ ; { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.7}{\includegraphics{grimfiles / g1.pdf } } } } } { \vcenter{\hbox{\scalebox{0.7}[]{\includegraphics{grimfiles / g1.pdf } } } } } } \ ; + \ ; { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.7}{\includegraphics{grimfiles / f1l.pdf } } } } } { \vcenter{\hbox{\scalebox{0.7}[]{\includegraphics{grimfiles / f1l.pdf } } } } } } \;.\ ] ]
let @xmath141 be a partition which sits inside the partition @xmath142 having @xmath1 rows of @xmath36 boxes each . the _ complementary partition _ @xmath143 also fits inside @xmath142 in the lower right - hand corner by rotating @xmath144 by @xmath145 degrees .
[ thm : cute ] there is a bijection @xmath146 between the set of semistandard tableaux of shape @xmath147 and type @xmath148 and the set of standard tableaux with entries @xmath149 and shape @xmath150 . for a semistandard tableau @xmath38 of shape
@xmath151 with type @xmath148 , the standard tableau @xmath152 is obtained by the following procedure : ( i ) place @xmath38 inside the rectangle of shape @xmath142 ; ( ii ) fill in the rest of the boxes in each column with the missing numbers in @xmath153 in ascending order from bottom to top ; ( iii ) rotate the resulting tableau 180 degrees and then take its conjugate .
for example , when @xmath154 , @xmath155 , and @xmath156{\includegraphics{grimfiles / exam1.pdf } } } } } } \;,\ ] ] we obtain @xmath157{\includegraphics{grimfiles / exam2.pdf } } } } } } \ ; , \ ] ] so that @xmath158{\includegraphics{grimfiles / exam3.pdf } } } } } } ^{\,\,\mathsf t } = \
{ \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / exam4.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / exam4.pdf } } } } } } \;,\ ] ] which has shape @xmath150 .
the map @xmath146 is clearly a bijection .
the theorem above has a number of applications ; for example , it can be used to prove results about representations .
we illustrate this with one quick example .
recall that the finite - dimensional irreducible representations @xmath159 of the complex special linear lie algebra @xmath160 are indexed by partitions @xmath22 with at most @xmath1 parts ( i.e. @xmath1 rows ) .
the irreducible representations @xmath161 of the symmetric group @xmath162 are labeled by partitions @xmath163 of @xmath1 and have a basis indexed by the standard tableaux of shape @xmath163 with entries @xmath153 .
now @xmath159 decomposes into weight spaces ( common eigenspaces ) relative to the diagonal matrices in @xmath160 , and the weight space corresponding to @xmath164 ( where @xmath165 projects a diagonal matrix onto its @xmath166-entry ) has a basis indexed by the semistandard tableaux of shape @xmath22 and type @xmath167 .
because matrices in @xmath160 have zero trace , @xmath168 on @xmath160 , and the weight 0 corresponds to the semistandard tableaux of type @xmath169 , where each entry occurs with the same multiplicity .
thus , if @xmath170 is a weight of @xmath159 , then @xmath171 is a multiple of @xmath1 .
the converse holds as well , which can be seen by filling the boxes of @xmath172 with @xmath173 in order , starting with the first row of @xmath22 and moving left to right across the rows from top to bottom .
in particular , if @xmath22 is a partition of @xmath174 , for some @xmath175 , and @xmath176 , then @xmath177 is a partition of @xmath1 .
the zero weight space @xmath178 of @xmath159 has a basis indexed by the semistandard tableaux of type @xmath179 . since the standard tableaux of shape @xmath180 are in one - to - one correspondence with the standard tableaux of shape @xmath177 , which give a basis for the irreducible @xmath102-module @xmath181 , we have the following consequence of the bijection in theorem [ thm : cute ] under these assumptions : @xmath182 , where @xmath183 is the product of the hook - lengths of the boxes in @xmath177 .
the last equality comes from the well - known hook - length formula that gives dimensions of the irreducible @xmath128-modules ( see e.g. ( * ? ? ?
3.10.2 ) ) .
the bijection we define consists of a sequence of elementary steps , which we illustrate by a running example . *
start with a semistandard tableau @xmath184 of shape @xmath9 and type @xmath185 : * for example , let @xmath186 , and @xmath187{\includegraphics{grimfiles / semi1.pdf } } } } } } \ , .\ ] ] * step 1 .
let @xmath188 be the subtableau of @xmath184 having the entries @xmath189 , and let @xmath190 be the shape of @xmath191 : * @xmath192{\includegraphics{grimfiles / exam1.pdf } } } } } } \quad \hbox{\rm with } \ \ \ \nu = ( 3 ^ 3 , 2 ^ 2 , 1).\ ] ] * step 2 .
let @xmath193 be the standard tableau @xmath194 obtained from @xmath188 using theorem [ thm : cute ] : * @xmath195{\includegraphics{grimfiles / exam4.pdf } } } } } } \ ] ] * step 3 .
let @xmath196 be the skew - subtableau of @xmath184 formed by the entries @xmath197 : * @xmath198{\includegraphics{grimfiles / exam7.pdf } } } } } } \ ] ] * step 4 .
replace each entry @xmath199 of @xmath196 with @xmath200 , rotate the result by 180 degrees , and take its conjugate to obtain a standard tableau @xmath201 : * @xmath202{\includegraphics{grimfiles / exam8.pdf } } } } } } \ ] ] * _ @xmath203 has shape @xmath204 . _ * _ @xmath201 has shape @xmath205 _ in our example , @xmath206 and @xmath207 . * step 5 . assign a pair @xmath208 to @xmath203 by the following procedure * : let @xmath209 be the entries in the third row of @xmath203 .
do the following steps in order to create the tableaux @xmath210 and
@xmath211 : 1 . place @xmath212 from left to right across the third row of @xmath213 .
2 . for @xmath214 ,
let @xmath215 be the largest entry in the second row of @xmath193 less than @xmath216 , such that @xmath215 is different from @xmath217 when @xmath218 .
place @xmath219 in the second row of @xmath213 so that they are increasing from left to right across the row .
3 . for @xmath214 ,
let @xmath220 be the largest entry in the first row of @xmath193 less than @xmath215 , such that @xmath220 is different from @xmath221 when @xmath218 .
place @xmath222 in the first row of @xmath213 so that they are increasing from left to right across the row .
4 . shift the remaining entries in each row of the tableau @xmath193 to the left to form the tableau @xmath211 .
the shape of @xmath213 is @xmath223 , and the shape of @xmath211 is the two - rowed partition @xmath224 .
@xmath225{\includegraphics{grimfiles / exam5.pdf } } } } } } \ ; , \qquad \mathsf{t}^+_\diamond={\ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / exam6.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / exam6.pdf } } } } } } \ ; .\ ] ] by its construction , @xmath213 is a standard tableau . to verify that @xmath226 is standard , it suffices to show that if @xmath227 is the @xmath228-entry and @xmath229 is the @xmath230-entry of @xmath193 , and they are in the same column of @xmath226 , then @xmath231 .
assume the contrary , that @xmath232 .
since @xmath227 is immediately above @xmath229 in the same column of @xmath226 and @xmath233 , there must be a pair @xmath234 from * step 5 * above with @xmath220 to the left of @xmath227 and @xmath215 to the right of @xmath229 in @xmath193 , such that @xmath235
. however , this contradicts the way @xmath220 was chosen in * step 5*.
so , we conclude that @xmath226 is a standard tableau .
perform ( 2 ) of algorithm [ alg : tym ] with @xmath213 to get an @xmath35-diagram @xmath236 and then trivalize the vertices to obtain an @xmath13-web @xmath237 whose boundary has @xmath1
@xmath2"s on the top : * @xmath238{\includegraphics{grimfiles / arc1.pdf } } } } } } \ ] ] * step 7 .
apply step 5 to @xmath239 , to obtain @xmath240 and @xmath241 : * @xmath242{\includegraphics{grimfiles / exam9.pdf } } } } } } \ ; , \qquad { \mathsf{t}}^-_\diamond={\ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / exam10.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / exam10.pdf } } } } } } \ ] ] _ the shape of @xmath240 is @xmath243 and of @xmath241 is @xmath244 which is @xmath245 .
hence , @xmath211 and @xmath241 have the same shape .
_ * step 8 .
apply step 6 to @xmath240 to obtain an @xmath35-diagram @xmath246 and an @xmath13-web @xmath247 whose boundary has @xmath3
@xmath4"s on the bottom : * @xmath248{\includegraphics{grimfiles / arc2.pdf } } } } } } \ ] ] * step 9 .
if @xmath249 are the entries of @xmath211 , and @xmath250 are the entries of @xmath241 , replace each @xmath251 and each @xmath252 by @xmath199 to produce a pair of standard tableaux with entries in @xmath253 . then perform the inverse of the robinson - schensted correspondence on the pair to get a permutation @xmath62 in @xmath254 : * @xmath255{\includegraphics{grimfiles / exam6.pdf } } } } } } \ ; , \quad \mathsf{t}^-_\diamond={\ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / exam10.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / exam10.pdf } } } } } } \right ) \longrightarrow \left({\ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / exam11.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / exam11.pdf } } } } } } \ ; , \quad { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / exam12.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / exam12.pdf } } } } } } \right)\ ] ] @xmath256{\includegraphics{grimfiles / exam11.pdf } } } } } } \ ; , \quad { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / exam12.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / exam12.pdf } } } } } } \right ) \longrightarrow \sigma= \begin{pmatrix } 1 & 2 & 3 & 4 \\ 3 & 1 & 2 & 4 \end{pmatrix}.\ ] ] * step 10 . for each @xmath257 ,
connect vertex @xmath251 on the top row to vertex @xmath258 on bottom row ( which we refer to as _ isolated vertices _ ) according to the following rule : * _ the path from @xmath251 to @xmath258 should stay within the open region @xmath259 between the two webs @xmath237 and @xmath247 and cross the smallest number of arcs in @xmath237 and @xmath247 .
after these paths are drawn , let @xmath260 be the resulting diagram .
_ @xmath261{\includegraphics{grimfiles / arcf0.pdf } } } } } } \ ] ] in the example , @xmath262 , so we connect the isolated vertex @xmath263 on the top with the isolated vertex @xmath264 on the bottom , @xmath265 on the top with @xmath266 on the bottom , @xmath267 on the top with @xmath268 on the bottom and @xmath269 on the top with @xmath270 on the bottom .
@xmath271{\includegraphics{grimfiles / arcf.pdf } } } } } } \ ] ] * step 11 .
trivalize the vertices as described in section [ sec : intro ] to produce an @xmath13-web @xmath272 : * @xmath273{\includegraphics{grimfiles / final.pdf } } } } } } \ ] ] our aim is to show the following : [ thm : main ] the map @xmath274 is a bijection between the set of semistandard tableaux @xmath38 of shape @xmath9 and type @xmath275 and the set of nonelliptic @xmath13-webs with @xmath1 `` @xmath2 ' 's on the top boundary and @xmath3 `` @xmath4 ' 's on the bottom boundary .
we close this section with some comments and a result we will use later . recall that a permutation @xmath276 is @xmath12-avoiding if there do not exist @xmath277 such that @xmath278 . it is well known that the rs correspondence gives a bijection between the set of @xmath12-avoiding permutations in @xmath61 and the set of pairs of tableaux of shape @xmath22 for some partition @xmath22 of @xmath41 , whose length is less than or equal to @xmath48 ( see for example , @xcite ) . because the tableaux @xmath211 and @xmath241 above are of the same shape and have at most 2 rows , the permutation obtained in step 9 will be @xmath12-avoiding .
the steps above construct an @xmath13-web having @xmath1 `` @xmath2 ' 's on the top boundary and @xmath53 `` @xmath4 ' 's on the bottom boundary .
the top and bottom boundaries have the same number of isolated vertices , and the @xmath12-avoiding permutation @xmath62 prescribes how to connect the isolated vertices on the top boundary to those on the bottom boundary .
therefore , we have the following corollary : assume @xmath38 is a semistandard tableau as in theorem [ thm : main ]
. then steps 1 - 10 above associate to @xmath38 a triple @xmath279 , where @xmath236 is an @xmath35-diagram with @xmath1 `` @xmath2 ' 's on the top boundary , @xmath246 is an @xmath35-diagram with @xmath3 `` @xmath4 ' 's on the bottom boundary , and @xmath62 is a @xmath12-avoiding permutation .
the notions of circle depth and path depth from @xcite will enable us to prove theorem [ thm : main ] .
the faces ( regions ) determined by an @xmath35-diagram @xmath280 and its trivalization @xmath281 are the same .
[ def : circle ] suppose @xmath280 is an @xmath35-diagram that lies above a horizontal boundary line , and let @xmath281 be its trivalization .
let @xmath68 be a point lying on a face of @xmath280 ( or @xmath281 ) .
* the _ circle depth _ of @xmath68 , denoted @xmath282 , is the number of semicircles in @xmath280 that lie above @xmath68 .
* the _ path depth _ of @xmath68 , denoted @xmath283 , is the minimal number of edges in @xmath281 crossed by any path from @xmath68 to the unbounded region @xmath284 that lies above the boundary line .
in particular , any point in @xmath284 has circle depth and path depth equal to @xmath170
. the following result can be found in @xcite .
[ depth ] ( ( * ? ? ?
* lemma 4.5 ) ) let @xmath285 be a standard tableau of shape @xmath9 , and let @xmath280 be the @xmath35-diagram and @xmath281 the @xmath13-web corresponding to @xmath285 by algorithm [ alg : tym ] . then , for a point @xmath68 in a face of @xmath280 ( or @xmath281 ) , the path depth and the circle depth of @xmath68 agree . in the following example , the numbers indicate path ( circle ) depths , which are the same for all interior points in a given face .
@xmath286{\includegraphics{grimfiles / webwithdepth.pdf } } } } } } \ ] ] because of this result , the only time we will need to make a distinction between an @xmath35-diagram and its trivalization is when we show the nonelliptic property .
we fix two positive integers @xmath1 and @xmath36 such that @xmath288 for all arguments in what follows .
let @xmath38 be a semistandard tableau of shape @xmath9 and type @xmath289 , and assume @xmath287 is the @xmath13-web obtained by applying the algorithm given in the previous section .
although the process certainly defines an @xmath13-web , it is not readily apparent that @xmath287 is _
nonelliptic_. we give a proof that @xmath287 is nonelliptic in this section .
we will use the notation from the steps in section [ sec : bijection ] that created @xmath287 .
in particular , @xmath236 and @xmath246 are the @xmath35-diagrams produced from 3-rowed standard tableaux in steps 6 and 8 , respectively . in algorithm
[ alg : tym ] , an entry @xmath41 in the third row of a tableau is connected to an entry @xmath42 in the second row , and @xmath42 is connected to an entry @xmath199 in the first row .
we refer to the union of the arcs connecting @xmath41 with @xmath42 and @xmath42 with @xmath199 as an _
@xmath35-configuration_. thus , the @xmath35-diagrams considered here are collections of @xmath35-configurations .
[ lemma : relative](@xcite ) if @xmath290 and @xmath291 are two @xmath35-configurations in an @xmath35-diagram obtained from a standard tableau with three rows , then there are only five possible relative positions for @xmath290 and @xmath291 as depicted in the following : @xmath292{\includegraphics{grimfiles / marc1.pdf } } } } } } \ ] ] @xmath293{\includegraphics{grimfiles / marc2.pdf } } } } } } \ ] ] @xmath294{\includegraphics{grimfiles / marc3.pdf } } } } } } \ ] ] @xmath295{\includegraphics{grimfiles / marc4.pdf } } } } } } \ ] ] @xmath296{\includegraphics{grimfiles / marc5.pdf } } } } } } \ ] ] there are five standard tableaux of shape @xmath297 , and each standard tableau corresponds to one of the relative positions of two @xmath35-configurations .
any @xmath35-configuration @xmath298 consists of a left and a right arc ( semicircle ) , which we denote @xmath299 and @xmath300 .
now assume @xmath193 and @xmath239 are the tableaux constructed in steps 2 and 4 and @xmath236 and @xmath246 are the corresponding @xmath35-diagrams from steps 6 and 8 , respectively .
[ lemma : isopos ] assume @xmath301 is an isolated vertex in @xmath246 ( resp . in @xmath236 ) , and @xmath301 has label @xmath227 .
let @xmath35 be an @xmath35-configuration in @xmath246 ( resp . in @xmath236 ) .
if @xmath301 is in @xmath300 , then @xmath227 lies in the first row of @xmath239 ( resp . of @xmath193 ) .
if @xmath301 is in @xmath299 , then @xmath227 lies in the second row of @xmath239 ( resp . of @xmath193 ) .
assume @xmath35 has vertices labeled @xmath302 . since @xmath301 is an isolated vertex , @xmath227 lies in either the first or second row of @xmath239 .
if @xmath301 is in @xmath300 , then @xmath227 is greater than @xmath42 , so it must lie in the first row of @xmath239 by the way @xmath42 was chosen .
similarly , if @xmath301 is in @xmath299 , then @xmath227 is greater than @xmath199 and less than @xmath42 , so it must lie in the second row of @xmath239 by the way @xmath199 was chosen .
[ lemma : lr ] let @xmath301 be an isolated vertex in @xmath246 ( resp . in @xmath236 ) , then it is impossible for @xmath301 to be in both @xmath303 and @xmath304 for two @xmath35-configurations @xmath290 and @xmath291 of @xmath246 ( resp . of @xmath236 ) .
there exists a unique shortest path from an isolated vertex @xmath301 on the bottom boundary to the unbounded region @xmath259 of @xmath305 .
due to proposition [ depth ] , there exists a shortest path from @xmath301 to @xmath259 , and such a path is determined by the circle depth of @xmath301 .
the path moves from one face of @xmath246 to another so that the circle depth is strictly decreased .
therefore , if there is only one choice of direction for moving , the shortest path must be unique .
there is only one choice by corollary [ lemma : lr ] , because there can be no isolated vertices in the shaded regions of figures iv and v of lemma [ lemma : relative ] , since those regions lie inside both a right semicircle and a left semicircle .
recall from * step 9 * in section [ sec : bijection ] that @xmath249 are the labels of the isolated vertices in @xmath236 , @xmath250 are the labels of the isolated vertices in @xmath246 , and @xmath62 is a @xmath12-avoiding permutation on @xmath306 .
let @xmath307 and @xmath308 be order - preserving bijections .
[ lemma : nocrossing1 ] let @xmath309 and @xmath310 be two isolated vertices labeled by @xmath227 and @xmath229 respectively , where @xmath311 . if @xmath309 and @xmath310 satisfy either of the following conditions , * @xmath309 and @xmath310 are contained in a semicircle in @xmath236 ( resp .
@xmath246 ) , * @xmath309 is inside @xmath312 and @xmath310 is inside @xmath313 for some @xmath35-configurations @xmath290 and @xmath291 of @xmath236 ( resp .
@xmath246 ) , then @xmath314 ( resp .
@xmath315 ) . for both ( i ) and
( ii ) , we present the proof only for the case @xmath236 .
\(i ) assume @xmath309 and @xmath310 lie inside one semicircle
. then they are in either @xmath299 or @xmath300 of an @xmath35-configuration of @xmath236
. therefore , @xmath227 and @xmath229 must be on the same row of @xmath226 .
suppose that @xmath227 and @xmath229 are in the second row of @xmath226 , and suppose on the contrary that @xmath316 .
consider the rs correspondence applied to @xmath317 . since @xmath227 is in the second row of @xmath226
, there is a bump when we insert @xmath318 .
so there exists an @xmath68 with @xmath319 , @xmath320 .
thus , we have @xmath321 and @xmath322 .
this contradicts the fact that @xmath62 is @xmath12-avoiding .
suppose that @xmath227 and @xmath229 are in the first row of @xmath226 , that is , @xmath309 and @xmath310 are in the right semicircle @xmath300 .
let @xmath323 be the three vertex labels of @xmath35 , so that @xmath324 .
we again assume that @xmath316 .
note that , when we insert @xmath325 , there is no bump , since @xmath227 is in the first row of @xmath226 . then , since @xmath229 is in the first row , there is no bump when we insert @xmath326 also . but @xmath316 .
hence , @xmath325 must be bumped down to second row before inserting @xmath326 .
therefore , there exists @xmath68 with @xmath327 and @xmath328 is in the second row of @xmath226 .
this is impossible by the way @xmath42 was chosen since @xmath329 .
\(ii ) if @xmath309 is inside @xmath303 and @xmath310 is inside @xmath304 , then @xmath227 is in the second row of @xmath226 and @xmath229 is in the first row of @xmath226 .
suppose on the contrary that @xmath330 .
since @xmath227 is in the second row , there is a bump when we insert @xmath325 while applying the rs correspondence to @xmath62 .
so there exists an @xmath68 with @xmath319 and @xmath320 .
then , @xmath331 and @xmath332 .
this contradicts the @xmath12-avoiding property of @xmath62 .
[ lemma : trivalizing ] let @xmath333 be a bounded region in @xmath334 , where @xmath335 is as in * step 10 * of section [ sec : bijection ] .
suppose that there is a point of crossing ( which has not yet been trivalized ) on the boundary of @xmath333 such that a directed edge on the boundary is moving toward the point , and the other directed edge on the boundary is moving away from the point .
then when we trivalize the crossing , there will be one more boundary edge added to the boundary of @xmath333 .
this is clear from the definition of trivalizing : @xmath336{\includegraphics{grimfiles / bound1.pdf } } } } } } \ ; \longrightarrow \ ; { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.55}{\includegraphics{grimfiles / bound2.pdf } } } } } { \vcenter{\hbox{\scalebox{0.55}[]{\includegraphics{grimfiles / bound2.pdf } } } } } } \,.\ ] ] the following proposition completes the proof that @xmath287 in theorem [ thm : main ] is a nonelliptic @xmath13-web . for a semistandard tableau @xmath38 of shape @xmath9 and type @xmath32
, the @xmath13-web @xmath287 is _
nonelliptic_. let @xmath333 be a bounded , internal region of @xmath334 , and let @xmath259 be the open region between the @xmath13-webs @xmath237 and @xmath247 of @xmath334 , as in * step 10 * of section [ sec : bijection ] .
after trivalization has taken place , internal regions are bounded by an even number of edges , since _
sink _ and _ source _ vertices alternate along the boundary of internal faces . hence ,
if @xmath333 has more than four boundary edges in @xmath334 , then after trivalization , @xmath333 is bounded by at least six edges .
so we may restrict our attention to the case that @xmath333 is bounded by either three or four edges .
we know from the work of tymoczko @xcite , that @xmath237 and @xmath247 ( as @xmath13-webs ) have no internal regions with four edges .
therefore , we may suppose that @xmath333 has at least one edge coming from a vertical line .
at this juncture , we break our considerations into several cases .
` case 1 : ` _ @xmath333 is not contained in @xmath259 _ we treat only the @xmath247 case , since every argument can be translated to @xmath237 by reversing the arrows on the directed edges .
since there are no crossings between vertical lines outside @xmath259 , there must be a part of the vertical line passing two semicircular arcs of @xmath247 ; let them be @xmath337 and @xmath338 so that the vertical line passes @xmath290 first in going from bottom to top , where @xmath339 is either @xmath340 or @xmath341 . by corollary [ lemma : lr ] and lemma [ lemma : relative ] , we see that @xmath337 and @xmath338 do not intersect each other .
@xmath342{\includegraphics{grimfiles / region1.pdf } } } } } } \qquad { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{.6}{\includegraphics{grimfiles / region2.pdf } } } } } { \vcenter{\hbox{\scalebox{.6}[]{\includegraphics{grimfiles / region2.pdf } } } } } } \ ] ] hence , there must be another arc ( part of a vertical line or a semicircle ) . since the direction of every arc in @xmath247 is upward , before trivalization there will be at least one point ( of crossing ) such that one edge on the boundary of @xmath333 is moving from the point while the other is toward the point .
this crossing point is indicated by @xmath343 in the illustrations above . therefore , as in , at least one more boundary edge will be added to @xmath333 upon trivalizing , so there will be at least six edges on the boundary of @xmath333 in @xmath287 . `
case 2 : ` _ @xmath333 is contained in @xmath259 .
_ @xmath339 the boundary edges of @xmath333 are all from vertical lines : + since the permutation @xmath62 in * step 9 * is @xmath12-avoiding , there can not be a bounded region with three sides .
but when @xmath333 is bounded by four edges , then since every vertical line moves upward , trivalizing crossings will add at least one more edge ( see ) .
@xmath339 there are boundary edges of @xmath333 from just one of the webs @xmath237 or @xmath247 : + assume that @xmath333 has an boundary edge @xmath344 from @xmath237 . then @xmath345 must be a connected arc .
if @xmath345 is a part of a semicircle , then there must be two vertical lines on the boundary of @xmath333 that pass through the semicircle . by lemma [ lemma : nocrossing1 ] , these two vertical lines do not make a crossing , and there must be at least one more vertical line on the boundary of @xmath333 .
if there are at least two more vertical lines , then @xmath333 will clearly have at least five edges ( four from the four vertical lines and the edge @xmath345 ) , and after trivalizing the crossings , @xmath333 has at least six edges .
@xmath346{\includegraphics{grimfiles / case2.pdf } } } } } } \ ] ] if there is only one more vertical line on the boundary of @xmath333 , then since every line is moving upward , the crossings of two lines on the boundary of @xmath333 will make another edge when we trivalize the crossings , due to lemma [ lemma : trivalizing ] .
@xmath347{\includegraphics{grimfiles / case21.pdf } } } } } } \ ] ] now we consider the case that there are at least two semicircles whose subarcs are in @xmath345 .
assume @xmath348 ( resp .
@xmath349 ) is a vertical line on the boundary of @xmath333 which is connected to a top vertex , and let @xmath350 ( resp .
@xmath338 ) be the semicircle that @xmath348 ( resp .
@xmath349 ) passes through .
if the pair @xmath351 is @xmath352 , then lemma [ lemma : nocrossing1 ] implies that @xmath348 and @xmath349 do not cross each other .
therefore , there must be another vertical line on the boundary of @xmath333 , and by lemma [ lemma : trivalizing ] , there will be at least six edges on the boundary of @xmath333 after trivalizing the crossings .
@xmath353{\includegraphics{grimfiles / lemma1.pdf } } } } } } \ ] ] when @xmath351 is @xmath354 as depicted in ( i ) below , trivalizing the intersection point of @xmath355 and @xmath348 will produce one more edge on the boundary of @xmath333 by lemma [ lemma : trivalizing ] . when @xmath351 is @xmath356 as in ( ii ) ,
trivalizing the crossing of @xmath355 and @xmath348 and of @xmath357 and @xmath349 will produce two more edges on the boundary of @xmath333 .
finally , when @xmath351 is @xmath358 as in ( iii ) , trivalizing the crossing point of @xmath357 and @xmath349 will produce one more edge on the boundary of @xmath333 .
0.33 @xmath359{\includegraphics{grimfiles / lemma3.pdf } } } } } } $ ] 0.33 @xmath360{\includegraphics{grimfiles / lemma2.pdf } } } } } } $ ] 0.33 @xmath361{\includegraphics{grimfiles / lemma4.pdf } } } } } } $ ] @xmath339 there are boundary edges of @xmath333 from both @xmath237 and @xmath247 : + there must be two noncrossing vertical lines on the boundary of @xmath333 .
therefore , if at least two semicircles appear on the boundary on @xmath237 or @xmath247 , then @xmath333 is not a quadrilateral .
hence , we only have to treat the case when just one semicircle from each of the webs @xmath237 and @xmath247 appears on the boundary : it is easy to check that there are at least two crossing points between vertical lines and semicircles where we can apply lemma [ lemma : trivalizing ] , so that at least two more edges will be added after trivalization .
in section [ sec : inverse ] , we will use the path depth of definition [ def : circle ] to create a pair of standard tableaux from an @xmath13-web , and this will enable us to show that the map @xmath72 in section [ sec : bijection ] is a bijection . the algorithm that constructed the @xmath13-web @xmath287 from a semistandard tableau @xmath38 produced a @xmath12-avoiding permutation via the rs correspondence in step 9 . in this section , we construct a pair of standard tableaux from a @xmath12-avoiding permutation using the path depth of a permutation and show that this pair is exactly the same as the one coming from the rs correspondence . the @xmath12-avoiding permutations in @xmath61 and the temperley - lieb diagrams with @xmath362 vertices are both counted by the catalan number @xmath363 ( see for example , ( * ? ? ? * chap .
6 ) ) . in remark
[ rem : tlbijection ] , we show how the considerations here give a direct bijection between the set of @xmath12-avoiding permutations and the set of temperley - lieb diagrams .
bijections using the reduced expressions for temperley - lieb diagrams as words in the generators are known and can be found for example in @xcite .
_ for the remainder of the section , we assume that @xmath276 is a @xmath12-avoiding permutation . _
[ lem : s5l1 ] assume @xmath364 . * if @xmath365 and @xmath366 , then @xmath367 . * if @xmath368 and @xmath369 , then @xmath367 . * if @xmath370 , then @xmath371 for all @xmath372 and @xmath373 for all @xmath374 . for ( i ) , assume the contrary that @xmath375 . if for all @xmath376 , @xmath377 , then @xmath378 since @xmath379 .
then @xmath62 is not bijective , since @xmath375 and @xmath380 .
thus , for some @xmath381 , we have @xmath382 , which contradicts the fact that @xmath62 is @xmath12-avoiding . part ( ii )
can be done by a similar argument , and ( iii ) follows from ( i ) and ( ii ) .
we associate to @xmath62 a diagram with vertices numbered @xmath383 from left to right on the top and on the bottom boundary lines , and with vertex @xmath199 on top connected to vertex @xmath384 on the bottom by a vertical line .
the vertical lines divide the region between the top and bottom boundaries into _ faces_. next we introduce an analog of the path depth given in definition [ def : circle ] for the diagram of @xmath62 .
[ def : path2](see @xcite . )
let @xmath284 be the unbounded face open to left of the diagram of @xmath62 .
for any point @xmath68 in the interior of a face of the diagram , the _ path depth _ of @xmath68 , denoted @xmath385 , is the minimal number of edges crossed by any path from @xmath68 to @xmath284 that does not intersect the top or bottom boundaries .
note that when a path from @xmath68 to @xmath284 crosses horizontally at the intersection of two lines in the diagram , it is counted as a single edge crossing .
all points in the interior of a face @xmath386 have the same path depth , so we set @xmath387 for any point @xmath68 in @xmath386 . for a @xmath12-avoiding permutation @xmath62
, we define the path depth difference @xmath388 of a boundary vertex @xmath301 to be @xmath389 where @xmath390 is the face to the left of @xmath301 and @xmath391 is the face to the right of @xmath301 . in below
, we will show that @xmath388 is always either @xmath170 and @xmath49 for every boundary vertex @xmath301 .
[ exam : perm ] for @xmath392 , the path depths of the internal faces are displayed here . also when @xmath393 ( resp .
@xmath394 ) for a boundary vertex @xmath301 we write @xmath395 ( resp .
@xmath396 ) .
@xmath397{\includegraphics{grimfiles / perm321pavoid.pdf } } } } } } .\ ] ] to distinguish the top and bottom vertices , we denote the @xmath199th vertex on the top ( _ resp .
bottom _ ) boundary of the diagram of @xmath62 by @xmath398 ( _ resp .
@xmath399 _ ) . using the values @xmath400 and @xmath401 ,
we create standard tableaux @xmath402 and @xmath403 of at most two rows by assigning the position of @xmath199 as follows : [ def : inverse ] for @xmath404 , * put @xmath199 in the first row of @xmath402 ( _ of @xmath403 _ ) if @xmath405 ( _ if @xmath406 _ ) ; * put @xmath199 in the second row of @xmath402 ( _ of @xmath403 _ ) if @xmath407 ( _ if @xmath408 _ ) ; * let @xmath409 .
now we assume @xmath62 corresponds to a pair of standard tableaux @xmath63 by the rs correspondence , where @xmath64 is the recording tableau and @xmath65 is the insertion tableau .
note that the permutation @xmath62 in example [ exam : perm ] corresponds to @xmath410{\includegraphics{grimfiles / tabr.pdf } } } } } } \ ; \ ; , \;\ ; { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / tabq.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / tabq.pdf } } } } } } \ ; \right)$ ] by the rs correspondence .
the following is the main theorem of this section , which we prove after a sequence of results .
[ thm : rspairs ] let @xmath62 be a @xmath12-avoiding permutation , and assume @xmath409 as in definition [ def : inverse ] .
suppose @xmath62 corresponds to @xmath63 under the rs correspondence .
then @xmath411 and @xmath412 .
[ def : lrfsets ] let @xmath413 , @xmath414,@xmath415 be the subsets of all @xmath416 such that @xmath417 [ lem : s5rs ] * if @xmath418 ( or if @xmath419 ) with @xmath420 , then @xmath421 . * if @xmath422 , then @xmath423 is in the first row of the recording tableau @xmath64 of @xmath62 . part ( i )
is a consequence of lemma [ lem : s5l1 ] . for ( ii ) ,
let @xmath422 and @xmath372 .
if @xmath424 , then we have @xmath425 by ( i ) . if @xmath426 , then @xmath427 .
hence for any @xmath428 , we have @xmath429 .
so in the robinson - schensted procedure , we adjoin @xmath430 to the end of the first row of the insertion tableau , and therefore , @xmath423 is in the first row of recording tableau . for an intersection of lines in the diagram of @xmath62
, the path depths of the faces around the crossing must be as in one of the following configurations : @xmath431{\includegraphics{grimfiles / depthconf1.pdf } } } } } } \ ; , \qquad { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / depthconf2.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / depthconf2.pdf } } } } } } \ ; , \qquad { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / depthconf3.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / depthconf3.pdf } } } } } } \ ; .\ ] ] note that the vertices @xmath432 in these pictures represent either a vertex on a boundary or a crossing point of two vertical lines . if @xmath227 is a vertex on the top boundary , then the second configuration is not possible , since the shortest path from a point in @xmath433 to @xmath284 must cross an edge to get to @xmath434 as shown below . likewise ,
if @xmath435 is a bottom vertex , then the third configuration is impossible .
@xmath436{\includegraphics{grimfiles / pathtof0.pdf } } } } } } \ ; , \qquad \qquad { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / pathtof1.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / pathtof1.pdf } } } } } } \ ; .\ ] ] [ lem : ellcond1 ] * if @xmath437 , then @xmath438 . * if @xmath439 , then @xmath440 . *
if @xmath441 , then @xmath442 . only the first and third configurations in can occur when
@xmath443 on the top boundary of the diagram of @xmath62 , so ( i ) is apparent .
similarly , if @xmath444 for some @xmath445 , then only the first two configurations are possible and we have ( ii ) . part ( iii ) is clear .
we have seen in lemma [ lem : s5rs ] that each @xmath446 lies in the first row of the recording tableau @xmath64 .
to determine conditions for @xmath439 to lie in the first row of @xmath64 will require some additional concepts .
assume @xmath447 and start moving along the edge from vertex @xmath448 on top .
whenever we reach a crossing , we make one of the following allowable turns : @xmath449{\includegraphics{grimfiles / turn1.pdf } } } } } } \ ; \qquad \qquad { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / turn2.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / turn2.pdf } } } } } } \;.\ ] ] we keep traveling until we arrive at a vertex on the top or bottom boundary and then stop
. the following illustrates what happens in example [ exam : perm ] .
@xmath450{\includegraphics{grimfiles / turnexample.pdf } } } } } } \;.\ ] ] let @xmath451 be the number of turns we made to arrive at a vertex in top or bottom starting from @xmath448 .
the following lemma is clear . for @xmath439 ,
starting from vertex @xmath448 , we arrive at a vertex @xmath452 in the top if and only if @xmath451 is odd , and in this case , @xmath453 . we arrive at a vertex @xmath454 in the bottom if and only if @xmath451 is even , and in this case , @xmath455 .
as we traverse an edge , there are faces @xmath456 and @xmath457 on the two sides of that edge .
the absolute value of the difference @xmath458 is either @xmath170 or @xmath49 by ( [ eq : depthconf ] ) .
when the edge is connected to a vertex @xmath301 on either the top or bottom boundary , the path depth difference equals @xmath388 .
the differences switch back and forth between @xmath170 and @xmath49 as we make a turn ( see and ) .
therefore , starting from vertex @xmath448 , @xmath447 , if we make an odd number of turns and arrive at vertex @xmath452 , @xmath453 , then @xmath459 and @xmath460 are different . if instead we make an even number of turns and arrive at vertex @xmath454 , then @xmath461 , and @xmath462 . now from lemma [ lem : ellcond1 ] , we have the following corollary .
[ c : delsig ] for @xmath439 , when we start from the vertex @xmath448 , we end at a vertex on the top if and only if @xmath463 .
let @xmath447 .
assume @xmath423 is in the second row of the recording tableau @xmath64 .
hence , @xmath464 is bumped out of the first row of @xmath65 by @xmath465 for some @xmath420 while applying the robinson - schensted process . and by lemma [ lem : s5l1 ] , @xmath453 .
note any @xmath464 can be bumped out at most once .
thus , each @xmath447 in the second row of @xmath64 is paired with a unique @xmath453 with @xmath420 such that @xmath466 .
we call such a pair @xmath467 an _ rs pair_. for @xmath468 the crossings in the diagram of @xmath62 made by the rs pairs are indicated by squares in the following example .
@xmath469{\includegraphics{grimfiles / perm321pasqure.pdf } } } } } } \;.\ ] ] recall @xmath62 corresponds to @xmath470{\includegraphics{grimfiles / tabr.pdf } } } } } } \ ; \ ; , \;\ ; { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / tabq.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / tabq.pdf } } } } } } \ ; \right)$ ] .
assume @xmath447 , and there is an rs pair @xmath467 .
we will show that a turn ii is always followed by a turn i during the trip starting from @xmath448 , and hence @xmath471 by corollary [ c : delsig ] .
the following illustrates the possibilities that occur when a turn ii occurs .
@xmath472{\includegraphics{grimfiles / travela1.pdf } } } } } } \ ; , \qquad { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{1}{\includegraphics{grimfiles / travelb1.pdf } } } } } { \vcenter{\hbox{\scalebox{1}[]{\includegraphics{grimfiles / travelb1.pdf } } } } } } \;.\ ] ] in either case , since @xmath473 and @xmath474 is not paired with @xmath475 , there must be an @xmath476 such that @xmath477 , @xmath478 , @xmath479 is an rs pair , and a turn i follows : @xmath480{\includegraphics{grimfiles / travela2.pdf } } } } } } \ ; , \qquad { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{1}{\includegraphics{grimfiles / travelb2.pdf } } } } } { \vcenter{\hbox{\scalebox{1}[]{\includegraphics{grimfiles / travelb2.pdf } } } } } } \;.\ ] ] we repeat this procedure until we arrive at a vertex on the top boundary in this case . on the other hand , assume that @xmath447 and @xmath423 is in the first row of the recording tableau @xmath64 , i.e. , there is no rs pair of @xmath423 .
we show that if a turn i happens , then always a turn ii follows during the trip starting from @xmath448 , and hence @xmath481 by corollary [ c : delsig ] . in order for a turn to occur ,
the line traveled is crossed by a line starting from a vertex @xmath452 with @xmath453 , @xmath420 , and @xmath482 . since @xmath423 is not paired , @xmath475 must be paired with another @xmath483 with @xmath473 .
if there is no @xmath484 such that @xmath485 , then we make a turn ii when the line from @xmath486 crosses the line from @xmath452 .
@xmath487{\includegraphics{grimfiles / travelc2.pdf } } } } } } \ ] ] if there is @xmath484 such that @xmath485 , then we make a turn ii at the crossing of the line from @xmath488 with the line from @xmath452 .
if @xmath489 is paired with a @xmath478 , then @xmath490 since @xmath491 is paired with @xmath475 .
@xmath492{\includegraphics{grimfiles / travelc3.pdf } } } } } } \ ] ] then , the line from @xmath493 must meet with the line from @xmath494 , and we get a contradiction , since we made a turn at the crossing of the line from @xmath495 and the line from @xmath494 . hence , @xmath489 is not part of an rs pair , and we make a turn at the crossing of the line from @xmath495 and the line from @xmath488 .
@xmath496{\includegraphics{grimfiles / travelc4.pdf } } } } } } \ ] ] we continue this procedure until we arrive at a vertex on the bottom boundary in this case .
we now have the following result , which summarizes the relationship we have determined between the path depth differences and the positions in the recording tableau @xmath64 of @xmath62 .
[ lem : c2c2 ] assume @xmath64 is the recording tableau of @xmath62 and @xmath413 is as in definition [ def : lrfsets ] . 1 . if @xmath497 , then @xmath438 and @xmath423 is in the first row of @xmath64 .
2 . if @xmath439 and @xmath438 , then @xmath423 is in the first row of @xmath64 .
3 . if @xmath439 and @xmath463 , then @xmath423 is in the second row of @xmath64 .
now we are ready to prove the main theorem in this section .
( * proof of theorem [ thm : rspairs ] * ) . from a @xmath12-avoiding permutation @xmath62
, we obtained a pair of standard tableaux @xmath498 using the path depth differences in definition [ def : inverse ] and a pair of standard tableaux @xmath63 by the rs correspondence .
lemma [ lem : c2c2 ] shows that @xmath499 .
since @xmath500 and @xmath501 corresponds to @xmath502 , we have @xmath412 .
[ rem : tlbijection ] starting a trip from @xmath448 , @xmath503 , we only need turn i and turn ii . however , beginning the trip from any vertex on the top ( or bottom )
, also requires using the following turns .
@xmath504{\includegraphics{grimfiles / turn3.pdf } } } } } } \ ; , \qquad { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / turn4.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / turn4.pdf } } } } } } \ ; .\ ] ] connecting the initial and terminal vertices of each path traveled results in a planar diagram ( no edges cross ) with each of the @xmath362 vertices connected to a unique vertex .
this determines a bijection between the set of @xmath12-avoiding permutations and temperley - lieb diagrams .
in particular , example [ exam : perm ] gives @xmath505{\includegraphics{grimfiles / turnexamplecmp.pdf } } } } } } \ ; ; \ ] ] hence , the following temperley - lieb diagram @xmath506{\includegraphics{grimfiles / perm321patl.pdf } } } } } } \ ; .\ ] ]
in section [ sec : bijection ] , we constructed a map @xmath72 from the set of semistandard tableaux of shape @xmath9 and type @xmath32 to the set of nonelliptic @xmath13-webs with a boundary of @xmath1 pluses on the top and a boundary of @xmath3 minuses on the bottom . in this section , we show that @xmath72 is a bijection by using khovanov - kuperberg s definition ( definition [ def : circle ] ) of path depth .
we connect the rightmost vertex on the top boundary of an @xmath13-web to the rightmost vertex on the bottom boundary to form one boundary line , and consider the unbounded region of the web that opens to left as @xmath284 .
then , we assign a path depth as in definition [ def : circle ] from each internal face of the web to @xmath284
. we will use this web path depth to construct two standard tableaux from the web , providing the inverse of @xmath72 .
this is analogous to what we did in section [ sec : combinatorics ] , where we used the path depth of a @xmath12-avoiding permutation to construct two standard tableaux that were exactly the same tableaux produced from the inverse of the rs correspondence by theorem [ thm : rspairs ] .
recall from that @xmath388 is the path depth difference for a boundary vertex @xmath301 of a @xmath12-avoiding permutation @xmath62 , and @xmath507 or @xmath49 ( see ) . in an analogous manner
, we define @xmath508 for a boundary vertex of a nonelliptic @xmath13-web .
the only difference is that @xmath508 may have values @xmath509 and @xmath510 .
( we will see below that @xmath508 has only these values . ) using the values @xmath511 and @xmath512 , we create two standard tableaux @xmath402 and @xmath403 of at most three rows by inserting @xmath199 as follows : * put @xmath199 on the first row of @xmath402 ( _ of @xmath403 _ ) if @xmath513 ( _ resp . if @xmath514 _ ) ; * put @xmath199 on the second row of @xmath402 ( _ of @xmath403 _ ) if @xmath515 ( _ resp .
if @xmath516 _ ) ; * put @xmath199 on the third row of @xmath402 ( _ of @xmath403 _ ) if @xmath517 ( _ resp .
if @xmath518 _ ) .
note that this is an extension of definition [ def : inverse ] . from @xcite and @xcite , we know that the set of semistandard tableaux of shape @xmath9 and type @xmath519 and set of nonelliptic @xmath13-webs with @xmath1 pluses on the top boundary and @xmath3 minuses on the bottom boundary have the same cardinalities . hence , to show that @xmath72 is bijective , it is enough to prove that @xmath72 is injective , which amounts to showing that we can recover @xmath38 from @xmath287 .
observe that if we can recover @xmath193 and @xmath239 , then getting @xmath38 is straightforward by applying @xmath520 ( for @xmath146 as given in theorem [ thm : cute ] ) , adjusting entries , rotating tableaux and joining them together .
so , we will only show that we can determine @xmath193 and @xmath239 from @xmath287 using the path depth differences @xmath508 .
we start by noting that path depth differences @xmath508 of boundary vertices @xmath301 on @xmath287 are same as those on @xmath305 given in step 10 of section . the diagram @xmath305 is formed by the diagrams of the two webs @xmath237 and @xmath247 and the diagram of a @xmath12-avoiding permutation @xmath62 on the isolated vertices .
let @xmath301 be an isolated vertex on the top boundary of @xmath305 and consider its path depth difference @xmath521 as a vertex in the diagram of the permutation @xmath62 as in section [ sec : combinatorics ] .
( an isolated vertex on the bottom boundary can be treated by a similar argument . )
corollary [ lemma : lr ] implies that only the following two configurations can occur when an @xmath35-configuration is added to the diagram of @xmath62 .
the value of the difference does not change by the addition of the @xmath35-configuration ; hence @xmath522 for isolated vertices .
this implies that @xmath211 and @xmath241 can be recovered by theorem [ thm : rspairs ] from the path depth @xmath508 of the isolated vertices : @xmath523{\includegraphics{grimfiles / config1.pdf } } } } } } \longrightarrow { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.6}{\includegraphics{grimfiles / config11.pdf } } } } } { \vcenter{\hbox{\scalebox{0.6}[]{\includegraphics{grimfiles / config11.pdf } } } } } } \ ] ] @xmath524{\includegraphics{grimfiles / config2.pdf } } } } } } \longrightarrow { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.6}{\includegraphics{grimfiles / config22.pdf } } } } } { \vcenter{\hbox{\scalebox{0.6}[]{\includegraphics{grimfiles / config22.pdf } } } } } } \ ] ] ( path depths are indicated as @xmath227 , @xmath525 , or @xmath526 . )
now assume that we have an @xmath35-configuration in @xmath237 ( or @xmath247 ) . adding a vertical line connecting an isolated vertex on the top boundary with an isolated vertex on bottom boundary to the @xmath35-configuration results in no change to @xmath508 for a vertex @xmath301 in the @xmath35-configuration , since the shortest path from @xmath527 and @xmath528 to @xmath284 either will cross the vertical line if @xmath301 is to the right of the line or will not cross the vertical line if @xmath301 is to the left of the line .
hence , the path depth difference of a vertex in @xmath237 ( resp .
@xmath247 ) is the same as it is in @xmath305 .
thus , by tymoczko s result ( see proposition [ depth ] ) , the path depths in @xmath237 and the circle depths in @xmath236 are the same ( resp
. of @xmath247 and of @xmath246 are the same )
. then we can recover @xmath529 from @xmath236 ( @xmath240 from @xmath246 ) by @xmath530 ( or @xmath531 ) if and only if @xmath199 in the first , second or third row of @xmath213 ( or @xmath240 ) , respectively .
once @xmath213 and @xmath226 are determined , then @xmath193 can be obtained by putting @xmath199 in row @xmath475 of @xmath193 if and only if @xmath199 is in row @xmath475 of @xmath213 or @xmath226 .
similarly , we can recover @xmath239 from @xmath240 and @xmath532 . in summary , we uniquely determined @xmath38 from @xmath287 using @xmath508 , and hence @xmath72 is injective ( and bijective , too ) .
the following is our example from section [ sec : bijection ] .
the path depths are indicated on @xmath305 and @xmath287 .
@xmath533{\includegraphics{grimfiles / arcfwithdepth.pdf } } } } } } \quad { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.45}{\includegraphics{grimfiles / finalinverse.pdf } } } } } { \vcenter{\hbox{\scalebox{0.45}[]{\includegraphics{grimfiles / finalinverse.pdf } } } } } } \ ] ] then , using the path depths , we recover the standard tableaux @xmath534{\includegraphics{grimfiles / exam4.pdf } } } } } } \ \text { and } \ { \mathsf{t}}^-={\ifthenelse{\isempty { } }
{ \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / exam8.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / exam8.pdf } } } } } } \;.
\hspace{.8 truein } \square\ ] ]
the algorithm in section [ sec : bijection ] produces an @xmath13-web from a pair @xmath535 of standard tableaux obtained from a semistandard tableau @xmath38 .
westbury @xcite has given an alternative to @xmath536-webs connected with @xmath537 using flow diagrams in triangular shapes .
in this final section , we describe how to obtain an @xmath13 flow diagram from such a pair @xmath535 of standard tableaux .
the results described here for @xmath13 can be extended to @xmath536 for arbitrary @xmath475 , and more details on this will appear elsewhere .
our aim here is to describe the relationship between the results of the previous sections to the flow diagram approach .
first , we briefly review westbury s result . to @xmath538
, we associate two equilateral triangles whose sides are one unit in length : @xmath539{\includegraphics{grimfiles / tria.pdf } } } } } } , \qquad { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.7}{\includegraphics{grimfiles / trima.pdf } } } } } { \vcenter{\hbox{\scalebox{0.7}[]{\includegraphics{grimfiles / trima.pdf } } } } } } \ ; .\ ] ] for compatibility with the conventions in previous sections , our arrows are the reverse of what is in @xcite , so they go from @xmath4 to @xmath2 rather than from @xmath2 to @xmath4 .
the edge labels @xmath170 and @xmath475 are omitted in the diagrams , and the triangles associated to @xmath540 are @xmath541{\includegraphics{grimfiles / tri1.pdf } } } } } } , \quad { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.7}{\includegraphics{grimfiles / tri1ma.pdf } } } } } { \vcenter{\hbox{\scalebox{0.7}[]{\includegraphics{grimfiles / tri1ma.pdf } } } } } } , \quad { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.7}{\includegraphics{grimfiles / trin.pdf } } } } } { \vcenter{\hbox{\scalebox{0.7}[]{\includegraphics{grimfiles / trin.pdf } } } } } } , \quad { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.7}{\includegraphics{grimfiles / trinma.pdf } } } } } { \vcenter{\hbox{\scalebox{0.7}[]{\includegraphics{grimfiles / trinma.pdf } } } } } } \ ; .\ ] ] now to each pair @xmath542 , @xmath543 , we associate a rhombus such that when @xmath544 , @xmath545{\includegraphics{grimfiles / rhombst.pdf } } } } } } \quad \text{if $ s\ne t$ , } \qquad \quad { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.7}{\includegraphics{grimfiles / rhombss.pdf } } } } } { \vcenter{\hbox{\scalebox{0.7}[]{\includegraphics{grimfiles / rhombss.pdf } } } } } } \quad \text{if $ s= t$.}\ ] ] if @xmath546 or @xmath547 , the corresponding rhombi are @xmath548{\includegraphics{grimfiles / rhombs0.pdf } } } } } } \ ; \qquad \text{or } \qquad { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.7}{\includegraphics{grimfiles / rhombt0.pdf } } } } } { \vcenter{\hbox{\scalebox{0.7}[]{\includegraphics{grimfiles / rhombt0.pdf } } } } } } \;.\ ] ] reversing the orientation of an edge changes the sign of the label : @xmath549{\includegraphics{grimfiles / rhombst.pdf } } } } } } \ = \
{ \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.7}{\includegraphics{grimfiles / rhombts.pdf } } } } } { \vcenter{\hbox{\scalebox{0.7}[]{\includegraphics{grimfiles / rhombts.pdf } } } } } } .\ ] ] also for @xmath536 , an edge labeled with @xmath550 , @xmath551 is identified with an edge labeled with @xmath552 .
for example , in the @xmath13 case we have @xmath553{\includegraphics{grimfiles / rhomb12.pdf } } } } } } = { \ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.7}{\includegraphics{grimfiles / rhomb11.pdf } } } } } { \vcenter{\hbox{\scalebox{0.7}[]{\includegraphics{grimfiles / rhomb11.pdf } } } } } } .\ ] ] we associate to a word @xmath554 of @xmath36 symbols from @xmath555 an equilateral triangle of length @xmath36 . along the top boundary of this triangle ,
we place triangles of unit length as in , so that the numbers on the top of these triangles , when read from left to right , give the word @xmath554 .
then , we fill in the rest of the triangle with rhombi as described above , so that the labels on the edges match on the boundaries of the rhombi .
the triangular diagram obtained in this way is the _ flow diagram _ determined by @xmath554 ( see @xcite for more details ) .
the following is an example of @xmath13 flow diagram associated to the word @xmath556 .
( note we use @xmath557 instead of @xmath558 when writing a word to simplify the notation .
) @xmath559{\includegraphics{grimfiles / flowexam.pdf } } } } } } .\ ] ] we say a word @xmath554 ( or its associated flow diagram ) is _ admissible _ if the associated flow diagram of @xmath554 has no contact with the boundaries of the large triangle other than with the top boundary of triangle .
( the diagram above is not admissible . )
let @xmath73 be the natural @xmath475-dimensional @xmath537-module and @xmath74 be the dual module .
a tensor product of copies of @xmath75 and @xmath76 corresponds to a string @xmath77 of symbols @xmath2 and @xmath4 and is denoted @xmath78 .
each word @xmath560 in @xmath555 of length @xmath36 determines a sign string @xmath561 given by @xmath562 if @xmath563 and @xmath564 if @xmath565 . for a given sign string @xmath82 of length @xmath36 , we let @xmath566 be the set of admissible words @xmath554 in @xmath555 of length @xmath36 whose sign string @xmath567 . then the following analog of theorem [ t : inv ]
holds in this setting : let @xmath82 be a sign string and @xmath566 be the set of admissible words in @xmath555 whose sign string is @xmath82 .
then , there is a bijection between @xmath566 and a basis for the space @xmath568 of @xmath569-invariants in @xmath570 .
any interior face of a flow diagram is bounded by at least six edges , since the following rhombus configuration is not allowable : @xmath571{\includegraphics{grimfiles / rhombssimpo.pdf } } } } } } .\ ] ] therefore , the nonelliptic @xmath13-webs discussed earlier in this paper are indeed @xmath13 flow diagrams .
thus , in the @xmath13 case , what we want is a bijection ( in fact an injection ) mapping the pair @xmath535 in section [ sec : bijection ] to an admissible flow diagram . to achieve this ,
we apply the following algorithm , which we illustrate with the running example from section [ sec : bijection ] , where @xmath572{\includegraphics{grimfiles / exam4.pdf } } } } } } $ ] has shape @xmath206 and @xmath573{\includegraphics{grimfiles / exam8.pdf } } } } } } $ ] has shape @xmath574 . 1 . for @xmath214 ,
let @xmath577 be the largest entry in the second row of @xmath193 less than @xmath578 , such that @xmath577 is different from @xmath579 , @xmath580 , @xmath581 when @xmath218 .
associate @xmath577 with @xmath578 .
2 . for @xmath582 ,
let @xmath583 be the largest entry in the first row of @xmath193 less than @xmath584 such that @xmath583 is different from @xmath585 , @xmath580 , @xmath586 when @xmath218 .
we associate @xmath583 with @xmath584 .
now assemble associated entries of @xmath193 to form one - column tableaux .
there will be @xmath587 tableaux of shape @xmath588 , @xmath589 tableaux of shape @xmath590 , and @xmath591 tableaux of shape @xmath592 . for @xmath193 as above , we obtain @xmath593{\includegraphics{grimfiles / tplus1.pdf } } } } } } $ ] , @xmath594{\includegraphics{grimfiles / tplus2.pdf } } } } } } $ ] , @xmath595{\includegraphics{grimfiles / tplus3.pdf } } } } } } $ ] , @xmath596{\includegraphics{grimfiles / tplus4.pdf } } } } } } $ ] .
the index @xmath41 in @xmath597 indicates that the shape of the tableau is @xmath598 . to determine the second index , we compare the bottom entries in the tableaux of shape @xmath598 and number the tableaux so the one with the smallest last row entry is @xmath599 , the next one is @xmath600 , and so forth until we reach the one with the largest last row entry . thus , if @xmath573{\includegraphics{grimfiles / exam8.pdf } } }
} } } $ ] , then @xmath601{\includegraphics{grimfiles / tmin1.pdf } } } } } } $ ] , @xmath602{\includegraphics{grimfiles / tmin2.pdf } } } } } } $ ] , @xmath603{\includegraphics{grimfiles / tmin3.pdf } } } } } } $ ] , @xmath604{\includegraphics{grimfiles / tmin4.pdf } } } } } } $ ] , @xmath605{\includegraphics{grimfiles / tmin5.pdf } } } } } } $ ] .
recall that @xmath193 has @xmath1 entries and @xmath239 has @xmath3 entries , where @xmath155 and @xmath607 in this particular example . from @xmath601{\includegraphics{grimfiles / tmin1.pdf } } } } } } $ ] , we get the word @xmath608 of length @xmath609 ( which is 17 here ) where @xmath610 @xmath611 where @xmath612 is in position 2 , @xmath613 in position 7 , and @xmath614 in position 8 . a wall @xmath615 separates the @xmath616 from @xmath239 on the left from the @xmath616 on the right from @xmath193 .
@xmath339 * repeat the same procedure for the other one - column tableaux of length 3 coming from @xmath239 and @xmath193 .
* @xmath619{\includegraphics{grimfiles / tmin2.pdf } } } } } } : \ \ \bullet \bullet \bullet \bullet \,\bar{1}\bar{2 } \bullet \bullet \bullet \bar{3}\ , \big | \bullet \bullet \bullet \bullet \bullet \bullet \bullet \ ; \null \\ & & { \mathsf{t}}^+_{3,1}={\ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / tplus1.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / tplus1.pdf } } } } } } : \
\ \bullet\bullet\bullet\bullet\bullet\bullet\bullet\bullet\bullet\bullet \big | \ , 3 \bullet \bullet \ , 2 \bullet \bullet \,1 \ ; , \end{aligned}\ ] ] where the word @xmath608 from @xmath620 is constructed from setting @xmath621 the corresponding flow diagram is @xmath622{\includegraphics{grimfiles / runexam3.pdf } } } } } } .\ ] ] now for @xmath623 , we pair @xmath624 and @xmath625 and to get an _ admissible _ flow diagram
. the number of such tableaux will always be the same for each @xmath626 .
( recall from step 7 of section [ sec : bijection ] that @xmath226 and @xmath532 have the same shape . ) from @xmath627{\includegraphics{grimfiles / tmin3.pdf } } } } } } \ ; , \ ; { \mathsf{t}}^+_{2,1}={\ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / tplus2.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / tplus2.pdf } } } } } } \right)$ ] , we get the word @xmath628 and the flow diagram @xmath629{\includegraphics{grimfiles / runexam4.pdf } } } } } } .\ ] ] applying a similar process to the pair @xmath630{\includegraphics{grimfiles / tmin4.pdf } } } } } } \ ; , \ ; { \mathsf{t}}^+_{1,1}={\ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / tplus3.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / tplus3.pdf } } } } } } \right)$ ] , we get the word @xmath631 and to the pair @xmath632{\includegraphics{grimfiles / tmin5.pdf } } } } } } \ ; , \ ; { \mathsf{t}}^+_{1,2}={\ifthenelse{\isempty { } } { \vcenter{\hbox{\scalebox{0.8}{\includegraphics{grimfiles / tplus4.pdf } } } } } { \vcenter{\hbox{\scalebox{0.8}[]{\includegraphics{grimfiles / tplus4.pdf } } } } } } \right)$ ] , the word @xmath633 the result of superimposing all the flow diagrams is the following flow diagram : @xmath634{\includegraphics{grimfiles / runexam5.pdf } } } } } } .\ ] ] note by our choice of index @xmath199 for the tableaux @xmath624 and @xmath625 above , any two edges that cross have different labels . by replacing the crossings with rhombi as in , which is equivalent to _ trivalizing _ the crossings that was used earlier
, we have the final flow diagram associated to @xmath535 , which is admissible . | the nonelliptic @xmath0-webs with @xmath1 `` @xmath2 ' 's on the top boundary and @xmath3 `` @xmath4 ' 's on the bottom boundary combinatorially model the space @xmath5 of @xmath6-module maps on tensor powers of the natural @xmath7-dimensional @xmath6-module @xmath8 , and they have connections with the combinatorics of springer varieties .
petersen , pylyavskyy , and rhodes showed that the set of such @xmath0-webs and the set of semistandard tableaux of shape @xmath9 and type @xmath10 have the same cardinalities . in this work ,
we use the @xmath11-diagrams introduced by tymoczko and the robinson - schensted correspondence to construct an explicit bijection , different from the one given by russell , between these two sets . in establishing our result ,
we show that the pair of standard tableaux constructed using the notion of path depth is the same as the pair constructed from applying the robinson - schensted correspondence to a @xmath12-avoiding permutation .
we also obtain a bijection between such pairs of standard tableaux and westbury s @xmath13 flow diagrams . |
in recent years , the possible measurement of the absolute mass scale of neutrinos through @xmath9 decay has become of considerable interest .
this decay takes place only when the neutrino is a majorana particle with finite mass .
its occurrence has not been confirmed yet and is at the present time the subject of many experimental investigations .
concomitant with the @xmath9 there is the @xmath4 decay mode .
this mode is allowed by the standard model and it has now been observed in several nuclei @xcite . while the @xmath9 mode can be safely calculated in the closure approximation , since the average virtual neutrino momentum is of order 100 mev / c and thus well above the scale of nuclear excitations , the closure approximation is not expected to be good for @xmath4 where the neutrino momentum is of order of few mev / c and thus of the same scale of nuclear excitations .
@xmath4 decay without the closure approximation has been calculated within the framework of qrpa @xcite and lssm @xcite . in this paper
, we initiate a new approach to calculate @xmath4 without the closure approximation within the framework of the interacting boson model ( ibm-2 ) and its extensions ( ibfm-2 and ibffm-2 ) @xcite .
this latter model has been used extensively to calculate spectra of odd - even medium mass and heavy mass nuclei ( ibfm-2 ) and of odd - odd nuclei ( ibffm-2 ) @xcite , which are crucial for the calculation of @xmath4 decay . after a description of the ibfm formalism , we proceed to do a calculation of two - neutrino double-@xmath3 decays for @xmath10te@xmath11xe and @xmath12te@xmath13xe .
the aim of the paper is two - fold : ( i ) first and foremost we want to understand what is the mechanism of @xmath4 in te , that is , what intermediate states in the odd - odd nucleus contribute to the decay and ( ii ) from a comparison of our calculated matrix elements with experimental single-@xmath3 and @xmath4 decay , extract the value of the effective axial vector coupling constant , @xmath14 .
the inverse half - life @xmath15 for @xmath16 double-@xmath3 decay has been derived by several authors @xcite . we use here the formulation of tomoda @xcite as adapted in @xcite . for transitions
@xmath17 , @xmath15 can be factorized to a good approximation as @xmath18^{-1 } = g^{(0)}_{2\nu } { g_{a}}^{4 } |m_{2\nu}|^{2 } , \label{eq : halflife}\ ] ] where @xmath19 is the lepton phase - space integral , @xmath20 is the axial vector coupling constant and @xmath21 .
\label{eq : m2nu}\ ] ] the gamow - teller ( gt ) matrix elements @xmath22 are calculated by @xmath23 where @xmath24 is the isospin increasing / decreasing operator , @xmath25 is the pauli spin matrix , while @xmath26 is the @xmath27 value of the double-@xmath3 decay , and @xmath28 and @xmath29 are the energies of the initial and the intermediate states , respectively .
the coefficient @xmath19 is the lepton phase - space integral .
its values are given in refs .
@xcite and @xcite .
the fermi ( f ) matrix elements @xmath30 are calculated by @xmath31 the inverse half - life of the decay @xmath32 @xmath33^{-1 } = g^{(0 ) 0^{+ } \to 2^{+}}_{2\nu } { g_{a}}^{4 } @xmath34 with @xmath35 in this case , there is no fermi contribution .
the aim of this paper is the calculation of the matrix elements @xmath22 , @xmath30 and @xmath36 .
the ingredients in the calculation are the matrix elements from even - even to odd - odd nuclei , @xmath37 , @xmath38 , and from odd - odd to even - even , @xmath39 , @xmath40 , which we now proceed to evaluate . a formulation of @xmath3-decay in the proton - neutron interacting boson - fermion model ( ibfm-2 )
was given years ago [ 4 , chap .
7 ] and @xcite .
the microscopic theory gives the images of the fermi and gamow - teller transition operators as @xmath41^{(0 ) } , \label{eq : fermi } \\
& t^{\pm}\sigma & & \longrightarrow & o^{\rm gt } & = \sum_{j'j } \eta_{j'j } \bigl [ p^{(j')}_{\pi } p^{(j)}_{\nu}\bigr]^{(1 ) } , \label{eq : gt}\end{aligned}\ ] ] where @xmath42 the operator @xmath43 stands for the boson - fermion image of the particle transfer operator . for the transitions from an even - even nucleus to an odd - odd nucleus
, it can be either of the two operators : @xmath44^{(j)}_{m } , \label{eq : adagger } \\
\tilde{b}^{(j)}_{m } & = -\theta_{j}^ { * } s a_{jm}^{\dagger } - \sum_{j ' } \theta_{jj'}^ { * } [ \tilde{d } \times a_{j'}^{\dagger } ] ^{(j)}_{m } , \label{eq : btilde}\end{aligned}\ ] ] where @xmath45 is the fermion creation operator , @xmath46 is the @xmath47-boson creation operator , and the @xmath48-component of @xmath49 is related to the @xmath50-boson annihilation operator by @xmath51 . in these operators , the distinction between the proton ( @xmath52 ) and the neutron ( @xmath53 ) will be made later when necessary .
the operator from an odd - odd nucleus to an even - even nucleus is @xmath54^{(j)}_{m } , \label{eq : atilde } \\
b^{\dagger(j)}_{m } & = \theta_{j } s^{\dagger } \tilde{a}_{jm } + \sum_{j ' } \theta_{jj ' } [ d^{\dagger } \times \tilde{a}_{j ' } ] ^{(j)}_{m } , \label{eq : bdagger}\end{aligned}\ ] ] where @xmath55 is related to the fermion annihilation operator @xmath56 by @xmath57 , @xmath47 is the @xmath47-boson annihilation operator , and @xmath58 is the @xmath50-boson creation operator .
the coefficients of the transfer operators are @xcite @xmath59 where @xmath60 and @xmath61 are bcs unoccupation and occupation amplitudes , and the quantities @xmath62 , @xmath63 , @xmath64 , @xmath65 are calculated from the expectation values of the @xmath47-boson and @xmath50-boson numbers , @xmath66 , @xmath67 , and @xmath68 if the odd fermion is a hole , then @xmath60 and @xmath61 are interchanged , and the sign of @xmath69 is reversed in eqs . to
in order to calculate @xmath4 decay , we use the wave functions of the initial , intermediate and final nuclei , in the present case the wave functions of @xmath0te , @xmath0i and @xmath0xe .
they are obtained from a calculation of the energy levels and electromagnetic properties . for purposes of checking the accuracy of our approach
, we also calculate energy levels and electromagnetic properties of the odd - even nuclei @xmath71i , @xmath72te , @xmath73i , @xmath71te .
the entire set of nuclei we calculate is shown in table [ ta : nuclei ] .
.nuclei related to the double-@xmath3 decay from @xmath74te . [ cols="^,^,^,^",options="header " , ] the calculations reported in the previous subsections are based on contributions of states in the intermediate odd - odd nucleus with @xmath75 mev .
it is of interest to investigate how likely or unlikely is that states above 3 mev could contribute significantly to two - neutrino decay .
the @xmath76 and @xmath77 states with @xmath78 mev are built from two contributions : ( 1 ) states constructed from single - particle orbitals within the model space of table [ ta : spe ] ; ( 2 ) states constructed from single - particle orbitals above the shell gap at 82 or below the shell gap at 50 . to investigate the contribution of states with @xmath78 mev within the model space of table [ ta : spe ] , one can simply extend the calculation from the current lowest @xmath79 @xmath76 states and @xmath80 @xmath77 states to larger numbers .
it appears that the properties of the strength distributions remain the same and that therefore states of this type will not contribute significantly to two - neutrino decay .
states constructed from single - particle orbits outside the model space of table [ ta : spe ] will appear in the spectrum at energies above the shell gaps , @xmath81 mev .
we have in fact investigated their contributions by including the proton orbit @xmath82 and the neutron orbit @xmath83 , as in table [ ta : spefi ] , and concluded that also this type of states will not contribute significantly to two - neutrino decay .
the argument is as follows .
inclusion of excitations across major shells will give major contributions to the strength distribution for the gt@xmath84 and f@xmath84 `` legs '' , te @xmath5 i , especially in the region @xmath85 mev , where the giant gt resonance ( @xmath86 mev ) and the isobaric analogue state , ias ( @xmath87 mev ) are located .
however , because of their composition in terms of single - particle states , we expect its contribution to the gt@xmath88 ( or f@xmath88 ) `` leg '' , i @xmath5 xe , which will still remain concentrated in few low - lying states , figs .
[ fig : gt128fi ] and [ fig : gt130fi ] , center panel .
it is therefore , in our opinion , quite unlikely that states above 3 mev will contribute significantly to two - neutrino decay .
in this article , a detailed investigation of the ten nuclei : @xmath0te , @xmath0i , @xmath0xe , @xmath1i and @xmath2te , within the framework of the interacting boson model-2 , ibm-2 , and its generalizations ibfm-2 and ibffm-2 has been done .
the parameters needed in this investigation have been obtained as much as possible from the available experimental information .
the wave functions so obtained have been used to calculate single-@xmath3 and @xmath4 matrix elements .
the main results of our investigation are : 1 .
the mechanism of @xmath4 in these nuclei appears to be single - state dominance ( ssd ) .
2 . the gt strength in @xmath0te ( @xmath89 ) @xmath5 @xmath0i ( @xmath90 )
is evenly distributed .
3 . the gt strength in @xmath0i ( @xmath90 ) @xmath5 @xmath0xe ( @xmath89 )
is concentrated in one state .
4 . use of a single value @xmath91 appears to describe well both single-@xmath3 and @xmath4 decay .
the results are very sensitive to the choice of single - particle energies , most notably the weak branch @xmath92 . however
, when the renormalization of @xmath20 is taken into account through fitting the single-@xmath3 decay , different choices of s. p. e. give similar results for @xmath4 , but with varying degree of accuracy .
the best choice appears to be that of the s. p. e. of @xcite which describes all observed quantities fairly : energies , electromagnetic transitions and moments , single-@xmath3 matrix elements , ( @xmath93he , @xmath94 ) strength distributions , and double-@xmath3 matrix elements , in the quenched approximation .
our best estimates of @xmath4 matrix elements are therefore those given in table [ ta : meff ] , @xmath95 for @xmath96 and @xmath97 for @xmath98 , and of the @xmath4 half - lives @xmath99 ( @xmath100 yr , @xmath99 ( @xmath101 yr
. our extracted values @xmath102 and @xmath103 for the single - particle levels of table [ ta : spe ] and [ ta : spefi ] , respectively , are rather low .
the values of @xmath14 depend on mass number , @xmath104 , and on the nuclear model used in their extraction . a preliminary study of @xmath14 in the interacting boson model ( ibm-2 ) in the closure approximation for @xmath4 decay and the interacting shell model ( ism ) has been done in @xcite .
it has been found that @xmath14 has a smooth dependence that can be parametrized as @xmath105 plus shell effects .
the extracted values of @xmath106 in the mass region , @xmath107 , are @xmath108 for ibm-2 and @xmath109 for ism .
a similar analysis has been done within the framework of qrpa @xcite with similar results .
we intend to continue the study of @xmath14 within the framework described in the present article to understand how general is the result presented here , and also to study the related question of the extent to which @xmath110 is quenched in heavy nuclei , if at all .
the question of the impact of the small value of @xmath14 found in @xmath3-decay and @xmath4-decay to @xmath9-decay is the subject of much debate .
while only gt ( @xmath76 ) and f ( @xmath77 ) multipoles contribute to allowed @xmath3- and @xmath4-decay , all multipoles ( @xmath111 ... ) , ( @xmath112 ... ) contribute to @xmath9 decay .
it is not clear whether or not the higher multipoles are quenched .
nonetheless , since @xmath76 and @xmath77 still provide the largest contributions , we expect the quenching of @xmath20 and @xmath110 to play an important role in @xmath9 decay . since , in view of the fact that @xmath20 appears to the fourth power in the decay rate , the quenching of @xmath20 has major repercussions on @xmath9 experiments , we plan to investigate this problem in depth in subsequent papers .
this work was performed in part under the us doe grant de - fg-02 - 91er-40608 .
we want to thank d. frekers for stimulating this work and providing ref .
@xcite prior to publication , and h. ejiri for many useful discussions .
99 for a review , see , a. s. barabash , phys .
c * 81*,035501 ( 2010 ) . for reviews , see , j. suhonen and o. civitarese , phys . rep . , * 300 * , 123 ( 1998 ) ; a. faessler and f. imkovic , j. phys .
c * 24*,2139,(1998 ) .
e. caurier , f. nowacki , and a. poves , int .
e , * 16 * , 552 ( 2007 ) .
f. iachello and p. van isacker , _ the interacting boson - fermion model _ , ( cambridge university press , cambridge , 1991 ) . s. brant , n. yoshida , and l. zuffi , phys . rev .
c , * 24 * , 024303 ( 2006 ) ; n. yoshida , l. zuffi and s. brant , int .
e , * 15 * , 1933 ( 2006 ) .
m. doi , t. kotani , h. nishiura , k. okuda , and e. takasugi , prog .
phys . * 66 * , 1739 ( 1982 ) .
m. doi , t. kotani , h.nishiura , and e. takasugi , prog .
* 69 * , 602 ( 1983 ) .
f. boehm and p. vogel , _ physics of massive neutrinos _ , ( cambridge university press , cambridge , 1987 ) .
t. tomoda , rep .
* 54 * , 53 ( 1991 ) .
j. kotila and f. iachello , phys .
c , * 85 * , 034316 ( 2012 ) .
f. dellagiacoma , ph.d . thesis , yale university , 1988 .
f. dellagiacoma and f. iachello , phys .
b , * 218 * , 399 ( 1989 ) .
g. puddu , o. scholten , and t. otsuka , nucl .
a * 348 * , 109 ( 1980 ) .
o. scholten , ph .
d. thesis , university of groningen , the netherlands , 1980 .
s. brant and v. paar , z. phys .
, * 329 * , 151 ( 1988 ) .
p. j. brussaard and p. w. m. glaudemans , _ shell - model applications in nuclear spectroscopy _ , ( north - holland publishing , amsterdam , 1977 ) .
k. nakamura et al . ( particle data group ) j. phys .
g , * 37 * , 07502 , ( 2010 ) .
m. kanbe , k. kitao , nucl .
data sheets , * 94 * , 227 ( 2001 ) .
balraj singh , nucl .
data sheets , * 93 * , 33 ( 2001 ) .
fujita and k. ikeda , nucl .
67 * , 145 , ( 1965 ) .
p. puppe et al.phys .
c , * 86 * , 044603 ( 201 ) . h. miyahara , h. matumoto , g. wurdiyanto , k. yanagida , y. takenaka , a. yoshida , and c. mori , nucl .
instrum .
sec . a , * 353 * , 229 ( 1994 ) .
j. abad , a. morales , r. nez - lagos , and a. f. pacheco , au . fis .
a , * 80 * , 9 ( 1984 ) .
a. griffiths and p. vogel , phys .
c,*46 * , 181 ( 1992 ) .
o. civitarese and j.suhonen , phys .
c , * 58 * , 1535 ( 1998 ) .
h. ejiri , prog . part .
* 64 * , 249 ( 2010 ) .
j. suhonen and o. civitarese , phys .
b , * 668 * , 277 ( 2008 ) .
j. p. schiffer et al.phys . rev .
* 100 * , 112501 ( 2008 ) .
b. p. kay et al.phys .
c , * 79 * , 021301(r ) ( 2009 )
. j. barea , j. kotila , and f. iachello , phys .
c , * 87 * , 014315 ( 2013 ) .
a. faessler , g. l. fogli , e. lisi , v. rodin , a. m. rotunno , and f. imkovic , j. phys .
g : nucl . part .
phys . , * 35 * , 075104 ( 2008 ) . | a calculation of the spectroscopic properties , energy levels and electromagnetic transitions and moments , of the ten nuclei @xmath0te , @xmath0i , @xmath0xe , @xmath1i , @xmath2te within the framework of the interacting boson model ( ibm-2 ) and its extensions ( ibfm-2 and ibffm-2 ) is presented .
the wave functions so obtained are used to calculate single-@xmath3 and @xmath4 matrix elements for @xmath0te @xmath5 @xmath0i @xmath5 @xmath0xe decay .
use of the effective value of the axial vector coupling constant @xmath6 extracted from single-@xmath3 produces results for @xmath4 in agreement with experiment .
* two neutrino double-@xmath3 decay in the interacting boson - fermion model * n. yoshida@xmath7 and f. iachello@xmath8 @xmath7faculty of informatics , kansai university , takatsuki 569 - 1095 , japan + @xmath8 center for theoretical physics , sloane physics laboratory , yale university , + new haven , connecticut 06520 - 8120 , usa |
the proposed space - based gravitational - wave ( gw ) detector , _ lisa _ ( _ laser interferometer space antenna _ ) @xcite , is sensitive to gws in the @xmath4 hz frequency range . within this band , one of the most promising sources
are double white dwarf ( dwd ) binary systems , as it is expected that a large fraction of main - sequence binaries end their lives as close dwds @xcite .
for this reason , the gws emitted by these systems in our galaxy may form a background noise in the low frequency ( @xmath5 hz ) band of _ lisa_. the population of dwds in our galaxy is expected to be dominated by systems that undergo two distinct , long - lived phases of evolution : an `` inspiral '' phase , where both the stars are detached from their roche lobes and the loss of angular momentum in the form of gw emission causes the two stars to slowly spiral in towards each other ; and a `` stable mass transfer '' phase , where the less massive star fills its roche lobe initially and starts transferring mass steadily to its companion .
an example of the stable mass transferring systems are the am cvn type systems , of which 18 @xcite are known through electromagnetic observations .
see @xcite for more details . ] .
apart from dwds , neutron star white dwarf ( nswd ) binary systems are also one of the promising sources of gws for lisa .
various authors @xcite have estimated the gw background from these systems and concluded that the number of nswd systems detectable with lisa is 1 - 2 orders of magnitude less than dwd systems .
similar to dwds , nswd systems also undergo inspiral and stable mass transfer phases .
specifically , several studies ( see for example : * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) have suggested that one of the possible formation scenarios of the so - called ultra compact x - ray binary ( ucxb ) systems , with orbital periods @xmath6 minutes , is that a low mass white - dwarf donor ( @xmath7 ) transferring mass to an accreting neutron star ( ns ) primary in a short orbit . in this scenario ,
a detached nswd binary system initially evolves to a minimum orbital period as angular momentum is lost from the system due to gw radiation . at this minimum orbital period , the companion white dwarf ( wd ) star starts filling its roche lobe and transfers mass to the ns and the system evolves to longer orbital periods . at present
, there are @xmath8 known ucxb systems with measured orbital periods , see table[table1 ] .
some of these systems have accreting millisecond pulsars ( xte j1807 - 294 , xte j1751 - 305 & xte j0929 - 314 : see @xcite , for example ) and several of them are found in globular clusters as the stellar density and close encounters are more common @xcite .
the total population of field ucxbs may be low , @xmath9 @xcite and theoretical studies by @xcite indicates that even at the galactic center , accreting ns systems do not contribute much to the faint x - ray population .
in general , the capabilities of _ lisa _ as a gw detector are usually discussed in the context of the log @xmath10 log @xmath11 domain .
@xcite proposed that an analogy can be drawn between the astronomy community s familiar color - magnitude ( cm ) diagram and _ lisa _ s amplitude - frequency diagram . for lisa sources , an analogous quantity to absolute magnitude
@xmath12 is @xmath13 , where @xmath14 is the distance to the source .
the underlying physical properties of compact binary systems such as dwds and nswds , their evolution , and their relationship to one another in the context of stellar populations can be ascertained only if the observational properties of such systems are displayed in a @xmath15 diagram , rather than in a plot of @xmath16 versus @xmath17 .
@xcite discussed the dwd binary systems in this context of `` absolute '' amplitude - frequency domain assuming that the donors in these systems follow zero temperature mass - radius relation . here
, we will consider the effect of `` warm '' donors as proposed in @xcite and extend the discussion to nswd binary systems and compare the population boundaries between both of them .
we also discuss the limits and applicability of these population boundaries in the log@xmath18 - log@xmath11 space within the context of observed am cvn and ucxb systems .
as mentioned in the introduction , since the donor is a wd star and starts filling its roche lobe to begin mass transfer phase , we will denote the donor wd with subscript _ d _ and the accreting companion as _
a_. this notation will be followed even during the detached inspiral phase of the evolution .
hence , the total mass @xmath19 and the mass ratio @xmath20 . also we assume that the maximum mass of a wd to be chandrasekhar mass , @xmath21 , the minimum mass of a ns is @xmath22 and the maximum mass of ns to be @xmath23 . in the case of a detached dwd or a nswd system inspiralling as a result of loss of angular momentum due to gravitational radiation , we can write , as shown in @xcite , as @xmath24^{1/3 } = 5.38 ~[k^5 f^2]^{1/3}~\mathrm{m } \ , , \label{h_f_relationship}\end{aligned}\ ] ] where the dimensionless mass parameter , @xmath25 @xmath26 is the chirp mass and @xmath27 . notice that , for dwds , the maximum value of @xmath28 occurs at @xmath29 ( @xmath30 ) but for nswd binaries , the maximum value of @xmath31 is @xmath32 and occurs at @xmath33 ( @xmath34 ) .
this is purely due to the upper limit on the mass of the wd , @xmath35 .
similar to a dwd system , where a @xmath36 creates a maximum inspiral trajectory in the log@xmath37-$]log@xmath3 $ ] plane , a @xmath38 makes the limiting inspiral trajectory for nswd binary systems in this plane .
this limiting trajectory , for both dwd @xmath1 nswd is plotted in figs .
[ fig1 ] @xmath1 [ fig2 ] , respectively , as a red line with a slope of @xmath39 ( see eq.([h_f_relationship ] ) ) beyond which a detached inspiralling dwd or nswd system can not be found .
the phase of a detached inspiral evolution terminates when the low mass wd companion comes into contact with it s roche lobe to initiate a phase of mass - transfer .
since the donor is a wd star this contact period can be found by equating the radius of the star with the roche lobe radius . here , for the radius of the wd , we use the mass - radius ( m - r ) relationship of non - zero entropy donor models of @xcite .
these models were initially developed for ultra - compact x - ray binaries ( ucxb ) with wd donors .
later @xcite applied them to model the donors in am cvn systems formed through wd channel . for these models
the donors are assumed to be fully convective and corresponding donor evolution is considered to be adiabatic in nature due to the large mass- transfer rates produced in amcvn systems .
however , @xcite showed that the assumption of adiabatic evolution is applicable during the beginning of mass - transfer phases ( when the mass - transfer rates are high ) and may not work when the system evolves to larger orbital periods with low mass - transfer rates .
instead , they have identified new evolutionary phases and we will also discuss the impact of this most recent study on our population boundaries .
we illustrate these boundaries by assuming that the donor composition is he , but it is straight forward to extend the discussion to carbon ( c ) @xmath1 oxygen ( o ) donors . in this section
, we will discuss in detail the population boundaries of both dwd @xmath1 nswd . in
[ sec3 ] we will discuss the applicability of these boundaries with respect to lisa observations , along with the effects of new phases of evolution detailed in @xcite . finally , we summarize in [ sec4 ] . in fig .
[ fig1 ] , we show the population boundaries for both inspiralling @xmath1 mass - transferring phases of dwd systems . the ( red ) line with slope @xmath40 shows the maximum inspiral trajectory ( @xmath30 ) from a detached dwd system . above this boundary ,
no dwd binary can exist .
the lower ( green cross ) curve shows the locus of the termination ( contact ) points of all detached inspiralling dwd binary systems , which have @xmath41 , assuming that the wd is a he donor and has a temperature of @xmath42 ( which is equivalent to the boundary obtained with @xmath0 m - r relation ; along this isotherm , @xmath43 decreases from top to bottom ) . for comparison
, we have also shown the contact boundary for @xmath41 , generated by eggleton s mass - radius ( m - r ) relationship for @xmath0 white dwarfs , ( blue dashed line ) , as quoted by @xcite and also by @xcite . at higher amplitudes and frequencies ,
both the curves overlap each other .
this is due to the overlap of m - r relations at higher masses as these objects are supported by degeneracy pressure and the @xmath0 m - r relation that we are using is applicable for fully degenerate he donors . at lower amplitude and frequencies ,
the contact boundaries diverge slightly .
this is because there is a turn over in the m - r relation at low masses for @xmath44 as the coulomb interactions dominate thermal contribution at that low temperatures and the m - r relations deviate .
so this ( @xmath41 ) contact curve represents the boundary beyond which no dwds with he donors will be found . the corresponding boundaries for c @xmath1 o would lie just below this contact boundary , except that all three ( he , c @xmath1 o ) boundaries merge at high frequencies and deviate at low frequencies . once the mass transfer phase is initiated , the orbital separation @xmath45 starts increasing and the system evolves to lower amplitudes and frequencies .
the evolutionary trajectories can be then traced out by assuming that the donor wd star is marginally in contact with it s roche lobe and conservative mass transfer ( cmt ) holds true .
for dwds , since we assume that the maximum mass in a dwd system is @xmath35 , we can generate a upper boundary beyond which no mass - transferring dwd system can exist .
this limiting boundary for dwds can be obtained for systems with @xmath46 ( because , then , eventually @xmath47 will exceed @xmath35 ) when their @xmath48 drops below a critical @xmath49 @xmath50 accordingly , the ( green ) dashed line in fig .
[ fig1 ] indicates an isotherm with log@xmath51=7.5 $ ] and @xmath52 .
we note that log@xmath51=7.5 $ ] upper boundary is likely a generous overestimate of the available phase space even if the donors were to evolve adiabatically from contact . assuming that there are no donor wds in a dwd system with log@xmath51 > 7.5 $ ] ,
this curve represents the boundary beyond which no mass transferring dwds can exist .
since the @xcite donor models have two branch nature , with a minimum attainable mass , it is reflected in this boundary as it curves up , with a minimum attainable gw amplitude for a given isotherm
. corresponding boundary for donors with c @xmath1 o composition would lie below this he curve because a higher temperature is required for a c / o donor to fill the roche lobe than for a he donor . for comparison , a similar boundary that arises from @xmath0 m - r relation
is also shown as ( light blue ) dot - dashed curve . as is the case with @xmath41 contact boundary ,
these two curves overlap at high frequencies . from the above discussion
, we note that all mass - transferring dwds with different donor masses , composition and temperatures should be constrained within these two boundaries o composition donors , as discussed above , there is a slight deviation at the lower contact boundary . ] .
compared to t=0 model , as assumed in @xcite , hot donors in mass - transferring dwds occupy larger fraction of space in @xmath53 domain .
this is because hot donors are also more massive , therefore have higher intrinsic gw amplitude and come into contact at lower frequencies because of larger radius .
the inspiral phase of a wd binary evolution is driven by the loss of the angular momentum due to gw radiation and the inspiral evolutionary time scale @xmath54 can be written as : @xmath55 = \frac { 5}{64 \pi^2 } \biggl ( \frac{c}{rh_\mathrm{norm } f^2 } \biggr ) \ , .
\label{tau_chirp}\end{aligned}\ ] ] it is worth noting that the time scale of evolution in both inspiral and cmt phases is of the order of @xmath56 @xcite .
therefore we have drawn `` chirp '' isochrones also in fig.[fig1 ] and each isochrone has a slope of @xmath57 in this log@xmath58 - log@xmath11 space , as can be seen from eq.([tau_chirp ] ) due to the dependence on the product of @xmath59 and @xmath60 .
this implies that a given wd binary system , in either inspiral or cmt phase , spends @xmath54 amount of time corresponding to the gw frequency that the binary is emitting . at lower frequencies ,
the value of @xmath54 is larger compared to that of at higher frequencies .
a consequence of this behavior is that more binary systems accumulate at lower frequencies .
as mentioned in the introduction , the background noise arising due to millions of dwds at the lower frequency band ( @xmath61 hz ) of lisa is due to the fact that the value of @xmath54 for these systems is @xmath62 years , approaching hubble time ( see fig.[fig2 ] ) . .
the bottom ( green ) curve with crosses is drawn using the m - r relation of isentropic donor models from @xcite .
it represents the termination ( contact ) boundary for @xmath41 systems , assuming that the donors are of he composition with @xmath63 . for comparison ,
contact boundary derived using eggleton s @xmath0 m - r relation for fully degenerate he donors is also shown as ( blue ) dashed line .
the top ( green ) dashed curve represents the boundary at which @xmath64 , assuming that donors in dwd systems would have a maximum @xmath65 .
no mass - transferring dwd would be found above this boundary . for comparison
, the ( blue ) dot - dashed curve shows a similar boundary for @xmath0 donor models .
also shown are constant @xmath54 lines along with observed amcvn systems ( blue dots ; see text ) .
explanation for vertical ( brown ) dot - dashed line is given in
, width=600,height=494 ] in fig .
[ fig2 ] , population boundaries for both nswd systems is shown in the log@xmath58 - log@xmath11 space .
the top ( red ) curve represents the maximum inspiral boundary for detached nswd binary systems ( @xmath38 ) beyond which none will be found .
the bottom ( green ) curve with crosses represents the lower contact boundary , drawn using @xcite hot donor models , assuming that the donor is of he composition with @xmath63 .
the mass of the ns on this boundary is set to the minimum mass assumed , @xmath66 . for comparison , a similar boundary using @xmath0 m - r relation for he donors
is also shown as ( blue ) dashed curve . as is the case with dwds
, these two boundaries overlap at higher amplitudes and frequencies because at higher donor masses , the two m - r relations match as they are supported by degeneracy pressure .
furthermore , there is an overlap region between the lower nswd contact boundary ( green dashed curve in fig .
[ fig2 ] ) @xmath1 the @xmath64 dwd contact boundary ( green dashed curve in fig .
[ fig1 ] ) , indicating both types of systems can exist within this region . similarly , an upper contact boundary can also be drawn for nswds , shown as ( green ) dashed curve in fig .
[ fig2 ] , assuming the maximum mass of the ns is @xmath67 and that the maximum donor temperature in a nswd system to be @xmath68 .
the light blue dot - dashed curve is the upper contact boundary for @xmath0 donors . just like dwds , for nswd systems with @xmath69 , the critical mass ratio @xmath70 below which a nswd binary can not exist ,
is @xmath71 so all the mass - transferring nswd binary systems are bounded by the upper @xmath1 lower ( green ) curves and all the detached inspiralling nswds will be bounded between the top ( red ) curve and the bottom ( green ) curve
[ fig2 ] also shows chirp isochrones , indicating the evolutionary timescales .
$ ] -log@xmath3 $ ] plane .
the red line represents the maximum inspiral trajectory from a detached nswd binary system ( @xmath72 ) .
the bottom curve with ( green ) crosses represents the contact boundary , assuming minimum mass of the ns to be @xmath66 with he donors of @xmath63 .
the blue dashed line is drawn assuming a @xmath0 m - r relation for fully degenerate he donors .
these two boundaries overlap at high masses ( high amplitude @xmath1 frequency ) and deviate slightly at lower masses .
the top ( green ) dashed curve shows the upper bound for mass - transferring nswd systems , assuming that the maximum donor temperature in a nswd system to be @xmath68 and maximum mass of ns as @xmath73 . to compare
, an upper boundary drawn using @xmath0 donor model is also shown as ( light blue ) dot - dashed curve . also shown
are currently known ucxb systems ( blue dots ) along with chirp isochrones.,width=600,height=494 ]
the population boundaries for dwd @xmath1 nswd binary systems discussed in previous sections were generated using @xcite non - zero entropy donor models , which assume that the donors are fully convective and undergo adiabatic evolution throughout the mass - loss phase .
recently , @xcite showed that these assumptions may not properly estimate the donor s orbital period evolution and specifically the donor s adiabatic evolution may not hold true for whole mass - transfer phase .
instead they identified three distinct evolutionary phases and here we discuss the implications of their new findings on our population boundaries . during the first phase , which happens during the mass - transfer `` turn - on '' phase ( when the donor comes into contact initially ) , the radius of the donor decreases , the mass - transfer rate @xmath74 increases and the orbital period @xmath75 continues to decrease until the donor radius reaches it s minimum value ( @xmath74 becomes maximum ) and starts expanding again .
@xcite calculated that this turn on phase lasts up to @xmath76 years . in the second phase
, the donor responds to the mass loss adiabatically and starts expanding , which is considered to be the normal am cvn phase . but this phase of adiabatic expansion ends and a third phase of evolution begins at around @xmath77 min , when the mass - transfer rate ( and the donors thermal time ) drop enough for the donor to cool and start contracting to a fully degenerate configuration , stalling the @xmath75 evolution . in the case of dwds ,
the initial turn - on phase will result in the @xmath41 contact boundary ( green curve with crosses in fig .
[ fig1 ] ) to have a width , instead of a sharp boundary as discussed in @xcite .
this is because it is calculated assuming that once the system comes into contact , it would evolve towards lower amplitudes and frequencies , whereas @xmath75 decreases in the turn on phase , to a minimum even after the initial contact .
accordingly , there will be a slight overlap of lower contact boundaries for he , c @xmath1 o donors .
once the system evolves off this initial contact phase , the donor expands adiabatically in response to the mass - transfer and the system follows a typical am cvn evolution , where the gw amplitude @xmath1 frequency keeps decreasing as @xmath75 increases . assuming an adiabatic evolution means that the cooling time of the donor is longer than the mass - transfer time - scale ( @xmath78 ) , it will in turn affects the orbital evolution time - scale .
accordingly , the donor follows a trajectory where it passes through different isotherms of decreasing @xmath48 , after the initial contact . as @xcite illustrate , the lower right curve in fig .
[ fig1 ] shows the contact boundary for @xmath41 systems and similar contact boundaries for lower @xmath48 s would lie to the left of it , but the contact boundaries will shift towards lower amplitudes and frequencies .
this phase of adiabatic evolution comes to an end between @xmath79 min , when the donor starts to cool and contract eventually towards a fully degenerate star .
this will drastically ( almost an order of magnitude , see @xcite fig .
15 ) reduce @xmath74 at these long orbital periods .
hence , after this range of @xmath75 , the systems gw frequency evolution slows down in accordance with the drop in @xmath75 evolution .
if the donor has cooled enough to approximate it as a @xmath0 degenerate model , then it may lie close to one of the @xmath0 contact boundaries , depending upon its @xmath48 .
but these @xmath0 boundaries are bounded within the region constrained by @xmath64 ( light blue dot - dashed ) curve and the @xmath41 contact boundary ( green curve with crosses in fig .
[ fig1 ] ) , because these boundaries are drawn assuming that the donor is a fully degenerate star with @xmath0 and such a system can not exist beyond these curves .
therefore , we have drawn a vertical ( brown ) dot - dashed line in fig .
[ fig1 ] at @xmath80 min ( log@xmath81 ) beyond which we would expect these systems to be within the @xmath0 contact boundaries .
note that this vertical line is not a sharp boundary : some systems may not reach a fully degenerate configuration by this @xmath75 .
rather , a system s @xmath75 evolution slows down at a particular gw frequency once they start cooling towards a degenerate configuration at the above mentioned @xmath75 .
[ fig1 ] also shows the observed am cvn type systems , for which the masses and @xmath75 are taken from @xcite @xmath1 @xcite .
couple of them are clearly outside @xmath0 region ; for some of them , the system s mass function limits the minimum donor mass to a value above that of a roche - filling @xmath0 donor at the same @xmath75 .
it is very difficult to know their exact temperature and/or composition purely from gw observations , as there is a good overlap of systems with these characteristics ) has @xmath82 min and lies to the left of the @xmath80 line and below @xmath83 boundary .
this may seem to imply that this system may be cooling off and trying to reach a @xmath0 degenerate configuration .
but in fig 1 .
of @xcite , they give a temperature range of log @xmath84 and @xcite noted that this system may have a hot donor .
if that is the case , @xcite note that this donor may have evolved with constant entropy that it was born with or may have been heated by either the disk or the accretor .
however , if irradiation is the cause , then @xcite note that it will delay the onset of donor s cooling and increases the temperature of the donor .
so though this system lies below the @xmath83 boundary , it does not necessarily mean that the donor can be approximated as a zero - temperature object . ] . for nswds , similar to dwds , the contact boundaries with he composition drawn assuming ns mass is @xmath66 ( green curve with crosses ) and @xmath67 ( green dashed curve ) will also have a width due to decrease in @xmath75 even after the contact .
furthermore , the lower contact boundary lies _ below _ the upper contact boundary for dwds ( @xmath85 )
. this will make the contact boundaries to `` overflow '' into the dwd region and lisa observations may not be able to distinguish these two types of systems in this region . here
also the donor undergoes adiabatic evolution after @xmath74 reaches a maximum and consequently enters the cooling phase at @xmath86 min .
accordingly , in fig . [ fig2 ] , the brown line shows the orbital period of this transition phase . beyond this line , the donors should start cooling and gw frequency slows down accordingly .
[ fig2 ] also shows the currently observed ucxbs , which we assume to be mass - transferring nswd binary systems .
table [ table1 ] gives the values of the masses of donors and orbital periods that we used .
some of the ucxbs shown in fig .
[ fig2 ] lie _ below _ the lower contact boundary ( green cross curve ) , indicating that probably the donors are not of he composition .
although , this contact boundary has a width , and hence they may have he composition , it is unlikely that we can observe them during this relatively short lived phase .
moreover , these systems are plotted assuming the minimum mass of the donors mass range derived from observations , so this provides additional uncertainty in determining the composition of donors .
it is unlikely that lisa will be able to observe cooling donors in either dwd or nswd systems because of the instrumental and/or dwd background noise . in fig.[fig3 ] , we plot the known ucxbs and am cvns on top of lisa s sensitivity curve ( snr = 1 ) to assess the detectability of these systems . in the case of known ucxbs ,
it is clear that only one system ( 4u 1820 - 30 ) has enough signal strength to be visible to lisa , whereas some of the known am cvns emit gws above the instrumental and dwd background noise .
recalling from [ sec1 ] , in order to transform from log@xmath87 to log@xmath58 space , we need to know the distance @xmath14 to the binary system .
the relation between the unknown binary parameters @xmath14 , @xmath43 and @xmath48 and the observables @xmath88 , @xmath89 and @xmath2 can be written as @xcite @xmath90^{3 } & = & \frac{c^{12 } } { 2^{6}\pi^{2}g^{5}}\frac{h_\mathrm{norm}^{3}}{f^{2 } } , \\
r(1 - 2 g ) & = & \frac{5 c}{24 \pi^{2 } } \frac{\dot f}{h_\mathrm{norm } f^{3 } } \label{r_fdot_relation}\end{aligned}\ ] ] where @xmath91 in the inspiral phase of evolution and hence , it is easy to determine @xmath14 from @xmath88 , @xmath89 and @xmath2 through eq.([r_fdot_relation ] ) . for mass - transferring systems
@xmath92 is a function of @xmath43 and @xmath48 , and they can be related to the observable @xmath89 by the requirement that in the mass transfer phase , @xmath93 .
the determination of @xmath14 and/or the masses of the stars in dwd / nswd binary system depends on the determination of @xmath2 .
if an @xmath2 can not be measured for a system , then there is no way to tell whether that system is a nswd or dwd system based only on lisa observations . if an @xmath2 can be measured , and if it turns out to be negative , then it is possible that particular system is a mass - transferring system ( dwd or nswd ) .
but as shown in figs .
[ fig1 ] @xmath1 [ fig2 ] , it will still not be possible to know the type of the system , at least for low frequency sources ( log@xmath3 \lessapprox -2 $ ] ) .
there is a fairly good overlap in @xmath2 between dwd @xmath1 nswd systems in this region because of the non - zero entropy nature of the donors and also due to the lower limit on the mass of the ns ( @xmath66 ) .
but for high frequency mass - transferring sources ( log@xmath3 \gtrapprox -2 $ ] ) , it _ may _ still be possible to know the type of the system , as the overlap region reduces , but for ns masses higher than @xmath35 , lisa should be able to distinguish both types of systems through the measurement of @xmath2 .
] . the same thing can be said about inspiralling systems because there is a large overlap of dwd and nswd inspirals at low frequencies and even a measurement of positive @xmath2 would unlikely be able distinguish these two types of inspiralling systems .
in the previous sections , we have outlined the construction of population boundaries to illustrate the various evolutionary phases that a dwd or a nswd binary system would undergo in the distance independent `` absolute '' amplitude - frequency domain ( log@xmath37 $ ] - log@xmath3 $ ] ) of lisa . in an update to @xcite , who assumed the donors to be fully degenerate @xmath0 he stars , we consider that the donors in these systems follow the m - r relationship of non - zero entropy donor models of @xcite assuming he composition .
these models assume fully convective and adiabatically evolving donors during the whole episode of the mass - transfer phase . figs .
[ fig1 ] @xmath1 [ fig2 ] show that these `` hot '' donors occupy a larger fraction of the @xmath94 space , than the @xmath0 donors , because hot donors are also more massive , increasing their intrinsic gw amplitude . at high frequencies ,
both the models match each other because at these high masses , the hot donors are supported by degeneracy pressure and the m - r relations match . at low frequencies , the @xmath0 model and @xcite @xmath63 full model ( which is equivalent to @xmath0 )
diverge slightly because the coulomb interactions dominate thermal contribution and m - r relations deviate .
we also discussed the implications of new evolutionary phases found by @xcite on our population boundaries .
the initial `` turn on '' phase , where @xmath74 reaches it s maximum value and @xmath75 decreases even after contact , will result in contact boundaries having a width , instead of an abrupt cut - off .
this will cause some overlap onto the c @xmath1 o isotherms , which lie below he boundary .
soon after the system evolves from this initial contact , the second phase starts where the donor undergoes adiabatic evolution in response to the mass - transfer and the system will follow a typical am cvn evolutionary trajectory with decreasing gw amplitude and frequency .
this will continue until the donor begins to cool and the radius starts contracting .
accordingly , it will reduce @xmath74 , and @xmath75 ( and gw frequency ) evolution slows down and may stall once the donor reaches fully degenerate configuration .
but since lisa s sensitivity in this range is limited by instrumental and dwd background noise , this part of the evolution ( or cooling donors ) will probably be not observable by lisa .
it is unlikely that determination of @xmath2 will shed light on the type ( dwd / nswd ) of the low frequency systems without the help of independent electromagnetic observations .
this is because there is fairly a good overlap of nswd and dwd systems within the resolvable frequency regime ( log@xmath3 > 3 $ ] ) of lisa . and mass - transferring systems
, it may happen that more than one source can reside in a frequency bin , but a measurement of @xmath2 would probably indicate the evolutionary phase . ]
but for high frequency sources , it might be possible to distinguish between them .
combined with the expectation that the relative population of nswd is low compared to dwds and short period systems do not stay longer at those periods , it is likely that lisa will be able to measure @xmath2 for more number of dwds than nswds
. it will be interesting to see how the high frequency regions constrained by these boundaries would be populated through lisa observations and whether indeed we will know the nature of these systems .
we thank an anonymous referee whose suggestions have led to significant improvements in the manuscript .
r. k gratefully acknowledges the support of national science foundation grant no .
phy 06 - 53462 and no .
phy 05 - 55615 , and nasa grant no .
nng05gf71 g , awarded to the pennsylvania state university .
many thanks to joel tohline ( lsu ) for the help and guidance provided in the preparation of this work and manuscript .
cccccc 4u 1820 - 30 & 11 & 0.06 & 1 & 7.6 & 1a + 4u 1543 - 624 & 18 & 0.025 & 2 & 7.0 & 2a + 4u 1850 - 087 & 21 & 0.04 & 3 & 8.2 & 3a + m15 x-2 & 22.6 & 0.02 & 4 & 9.98 & 4a + xte j1807 - 294 & 41 & 0.0053 & 5 & 8.0 & 5a + 4u 1626 - 67 & 42 & 0.04 & 6 & 5 & 6a + xte j1751 - 305 & 42.4 & 0.014 & 7 & 6.7 & 7a + xte j0929 - 314 & 43.6 & 0.008 & 8 & 5.0 & 8a + ngc 6652b & 43.6 & ? & & ? &
+ 4u 1916 - 05 & 50 & ?
& & 8.9 & 10a + 4u 0614 + 091 & 50 & ?
& & 3.0 & 11a + swift j1756.9 - 2508 & 54.7 & 0.0067 & 12 & 7.6 & 12a + + [ table1 ] | in an earlier investigation , we proposed population boundaries for both inspiralling and mass - transferring double white dwarf ( dwd ) systems in the distance independent `` absolute '' amplitude - frequency domain of the proposed space - based gravitational - wave ( gw ) detector , _
lisa_. the degenerate zero temperature mass - radius ( m - r ) relationship of individual white dwarf stars that we assumed , in combination with the constraints imposed by roche geometries , permits us to identify five key population boundaries for dwd systems in various phases of evolution .
here we use the non - zero entropy donor m - r relations of @xcite to modify these boundaries for both dwd and neutron star - white dwarf ( nswd ) binary systems .
we find that the mass - transferring systems occupy a larger fraction of space in `` absolute '' amplitude - frequency domain compared to the simpler @xmath0 donor model .
we also discuss how these boundaries are modified with the new evolutionary phases found by @xcite . in the initial contact phase
, we find that the contact boundaries , which are the result of end of inspiral evolution , would have some width , as opposed to an abrupt cut - off described in our earlier @xmath0 model .
this will cause an overlap between a dwds @xmath1 nswds evolutionary trajectories , making them indistinguishable with only lisa observations within this region . in the cooling phase of the donor , which follows after the adiabatic donor evolution , the radius contracts , mass - transfer rate drops and slows down the orbital period evolution .
depending upon the entropy of the donor , these systems may then lie inside the fully degenerate @xmath0 boundaries , but lisa may be unable to detect these systems as they might be below the sensitivity limit or within the unresolved dwd background noise .
we assess the limits and applicability of our theoretical population boundaries with respect to observations and find that a measurement of @xmath2 by lisa at high frequencies ( log @xmath3 \geq 2 $ ] ) would likely distinguish between dwd / nswd binary . for low frequency sources , gw observations alone would unlikely tell us about the binary components , without the help of electromagnetic observations . |
presently there are a number of efforts to constrain the dense matter equations of state ( esos ) through astronomical observations of compact stars .
the usual approach is to determine the mass and the radius of the stars with the help of various observational features like gravitational redshifts ( z ) from spectral lines , cooling characteristics , khz quasi - periodic oscillations ( qpo ) etc ( lattimer & prakash 2007 , li @xmath3 1999 , zel 2006 , 2008 , zhang @xmath3 2007 ) .
but these methods are not foolproof , @xmath4 the value of z used in zel s ( 2006 ) analysis of exo 0748@xmath5676 can not be reproduced as mentioned by klahn @xmath3 ( 2006 ) . moreover , to constrain esos from qpo observations , one needs to believe in a particular model of qpo which is again a subject of debate .
another alternative method might be the measurement of the moment of inertia from the faster component of the double pulsar system psr j0737 - 3039 ( lattimer & schutz 2005 , bagchi @xmath3 2008 ) . some high mass stars like psr j1903 + 0327 , exo 0748@xmath5676 @xmath6 prefer stiff eos and some other stars like 4u 1728 - 34 ( li @xmath3 1999 ) , exo 1745 - 248 ( zel 2008 ) , prefer soft eos . this fact hints to the possibility of existence of both neutron stars and strange stars .
but even then , i need some constrains as there are a number of esos for neutron stars and also for strange stars . until then , it is interesting to compare the stellar properties for different esos . for
sufficiently fast spinning stars , stellar structures depend upon the spin frequency ( @xmath7 ) . so the study of stellar structures for rotating stars will help in better understanding of the characteristics of fast spinning compact stars like lmxbs and millisecond pulsars .
that is why here i study stellar structures with rotations in section [ sec : stell_struct ] . in section [ sec : rot_param
] , i study different rotational parameters for strange stars and neutron stars .
although stellar structures for rotating neutron stars or strange stars have been already studied by a number of groups ( some of which i discuss in section [ sec : rot_param ] ) , systematic studies of all relevant stellar parameters as well as disk parameters were lacking .
that is why here i report the variation of a number of different stellar parameters as well as disk parameters for different values of star s spin frequency and mass in section [ sec : stell_struct ] and section [ sec : rot_param ] .
also i use one eos for strange stars and another eos for neutron stars whereas in the earlier works people discussed either only neutron star rotations or only strange star rotations , there was no comparison between the properties of rotating neutron stars and rotating strange stars .
in addition i compare my results obtained by using a pseudo - newtonian potential with full general relativistic calculations by other people like haensel & zdunik ( 1989 ) and lattimer & prakash ( 2004 ) and the close matching found implies the correctness of my approach and the validity of the pseudo - newtonian potential . in section [ sec : application ]
i discuss a possible application of the knowledge of the rotational parameters in modeling khz qpos .
i end with a discussion in section [ sec : conclusion ] .
i use two sample esos of the dense matter among the numerous esos available in literature , one for the strange quark matter ( eos a or ssa , bagchi @xmath3 , 2006 ) and the other for the nuclear matter ( eos apr akmal , pandharipande & ravenhall 1998 ) . to find stellar structures with rotations , i use the rns code .
following komatsu , eriguchi & hachisu ( 1989 ) , this code constructs the compact star models by solving stationary , axisymmetric , uniformly rotating perfect fluid solutions of the einstein field equations with tabulated esos ( supplied by the users ) .
the fastest rotating compact star known till date is probably xte j1739 - 285 ( kaaret @xmath3 2007 ) having @xmath8 hz , although the measurement has not been confirmed later .
the second fastest one is j1748 - 2446ad ( hessels @xmath3 2006 ) with @xmath9 hz . in this work ,
i choose the angular frequency ( @xmath10 as 2000 , 3000 , 4000 and 5000 @xmath11 ( which correspond to @xmath7 as @xmath12 318 hz , 477 hz , 637 hz and 796 hz respectively ) .
all of the fast rotating compact stars except xte j1739 - 285 have @xmath7 in that range .
i have also computed non - rotating , spherically symmetric stellar structures by solving tov equations which are sufficient for slow objects like exo 0748@xmath5676 ( @xmath13 hz ) . throughout this work ,
i take the stellar mass ( @xmath14 ) to be always greater than 1.1 @xmath15 as observations usually hint the stellar mass to be greater than that value . in fig
[ fig : em ] i plot the mass against the central density ( @xmath16 ) both for strange stars and neutron stars . for a fixed value of @xmath16 , @xmath14 increases little bit with the increase of @xmath17 . for all of the values of @xmath17 ,
@xmath14 first increases with the increase of @xmath16 ( @xmath18 ) and then after a certain value of @xmath14 ( @xmath19 ) starts to decrease ( @xmath20 ) .
the stars are unstable when @xmath20 .
this instability appears around @xmath21 for strange stars and around @xmath22 for neutron stars ; these values does not change more than @xmath23 with the change of @xmath17 in the chosen range . for any @xmath17 in the chosen range ,
i get @xmath24 at @xmath25 for strange stars and at @xmath26 for neutron stars .
fig [ fig : rm ] shows the mass - radius plots . with the increase of @xmath17 , for both strange stars and neutron stars ,
the maximum mass ( @xmath19 ) increases and for any fixed mass , the radius also increases due to the larger value of the centrifugal force . note that here radius " means the equatorial radius @xmath27 which is always greater than the polar radius @xmath28 . for a fixed @xmath17 ,
the maximum mass for a neutron star is greater than that of a strange star . for fixed values of @xmath17 and @xmath14
, @xmath27 is larger for a neutron star than that of a strange star .
the compactness factor ( @xmath29 ) of strange stars is larger than that of neutron stars and the variation of @xmath14 with @xmath27 follows an approximate @xmath30 law for strange stars in contrast to neutron star s approximate @xmath31 variation .
fig [ fig : mi ] shows the variation of the moment of inertia ( @xmath32 ) with the mass . for any fixed mass ,
the moment of inertia increases with the increase of @xmath17 both for strange stars and neutron stars . for fixed values of @xmath14 and @xmath17 ,
a neutron star possess much higher value of @xmath32 than a strange star because of its larger value of @xmath27 . in fig
[ fig : mob ] , i plot the oblateness parameter @xmath0 the ratio of the polar radius to the equatorial radius ( @xmath33 ) with the mass .
it is clear that for a fixed mass , this ratio decreases with the increase of @xmath17 both for strange stars and neutron stars @xmath0 the star becomes more and more oblate due to the larger values of the centrifugal force .
moreover , for a fixed value of @xmath17 , this ratio decreases with the decrease of the mass as there the centrifugal force becomes increasingly more dominant over the gravitational force .
note that the variation of @xmath33 with @xmath14 is steeper for neutron stars than that for strange stars , but for both of them , the steepness increases with the increase of @xmath17 . in fig
[ fig : ma ] , i plot @xmath34 with @xmath14 where @xmath35 and @xmath36 .
as expected from their expressions , the plot shows that for any fixed mass , @xmath34 increases with the increase of @xmath17 as expected and for a fixed @xmath17 , @xmath34 decreases with the increase of the mass .
for the same value of @xmath14 and @xmath17 , a neutron star has larger value of @xmath34 than that of a strange star because of its larger value of @xmath32 .
@xmath34 is an important parameter of the compact stars as it can be identified as the specific angular momentum of the star and its value determines many other properties of the star . for any other eos ,
the value of @xmath19 and corresponding radius will change depending upon the stiffness of that eos .
but the general trend of the m - r curve will remain the same @xmath0 @xmath37 for strange stars and @xmath38 for neutron stars .
the nature of @xmath39 curve will also remain the same .
with the output of the rns code , @xmath0 @xmath40 and @xmath32 , i calculate some rotational parameters for strange stars and neutron stars .
first i calculate the radius of the marginally stable orbit which is defined as ( bardeen @xmath3 , 1972 ) : @xmath41^{1/2}\}\label{eq : rms}\ ] ] where @xmath42^{1/3}[(1+a / r_g)^{1/3}+(1-a / r_g)^{1/3}]\label{eq : z1}\ ] ] and @xmath43^{1/3}\label{eq : z2}\ ] ] the - " sign in the expression of @xmath44 implies the co - rotating motion and the + " sign implies the counter - rotating motion which i call as @xmath45 and @xmath46 respectively . as the values of @xmath34 are always very small , both @xmath47 and @xmath48 have their values @xmath12 3 .
the keplerian frequency of a particle orbiting around the star at a radial distance @xmath49 can be expressed as @xmath50^{1/2}\label{eq : nuk}\ ] ] where @xmath51 is the force per unit mass . as an example , i take @xmath51 as derived from a pseudo - newtonian potential by mukhopadhyay & misra ( 2003 ) @xmath52\label{eq : fm}\ ] ] in figs [ fig : m_rmsco ] and [ fig : m_rmscounter ] , i plot @xmath45 and @xmath46 respectively with the mass .
for any eos , @xmath45 is always smaller than @xmath46 for any fixed values of @xmath17 and @xmath14 . for a fixed @xmath17 ,
both @xmath45 and @xmath46 increases linearly with the increase of @xmath14 .
but for a fixed @xmath14 , @xmath45 decreases with the increase of @xmath17 and @xmath46 increases with the increase of @xmath17 . for strange stars ,
both @xmath45 and @xmath46 are always greater than @xmath27 . for neutron stars , @xmath45 and @xmath46
are smaller than @xmath27 for low masses and they become greater than @xmath27 for higher masses ; @xmath45 becomes equal to @xmath27 around @xmath53 and @xmath46 becomes equal to @xmath27 around @xmath54 .
all these happen because of the nature of the terms in the expression of @xmath44 ( eqn.[eq : rms ] ) .
as both @xmath55 and @xmath56^{1/2}\}$ ] are positive quantities and the second one is smaller than the first one , their sum ( @xmath46 ) must be greater than their difference ( @xmath45 ) .
as for a fixed value of @xmath17 , both @xmath55 and @xmath56^{1/2}\}$ ] increases with the increase of @xmath14 with the first term having much steeper slope , both @xmath45 and @xmath46 increase with increase of @xmath14 . for a fixed mass , @xmath55 remains almost constant but @xmath56^{1/2}\}$ ] increases with the increase of @xmath17 , so their sum ( @xmath46 ) increases with the increase of @xmath17 and the difference ( @xmath45 ) decreases with the increase of @xmath17 . for strange stars ( eos a ) , @xmath57 for any value of @xmath14 .
so after addition or subtraction of a comparatively small term ( @xmath56^{1/2}\}$ ] ) with it , the expression ( @xmath46 or @xmath45 ) remain always greater than @xmath27 for neutron stars ( eos apr ) , @xmath58 at lower values of @xmath14 where the values for @xmath27 are sufficiently larger , but @xmath59 at larger values of @xmath14 . here
@xmath56^{1/2}\}$ ] is very small in comparison to both @xmath55 and @xmath27 for all values of @xmath14 .
so the addition or subtraction of @xmath56^{1/2}\}$ ] with @xmath55 ( to get @xmath44 for counter - rotating or co - rotating motions respectively ) does not change the overall trend , only the addition ( for counter - rotation ) shifts the transition towards lower values of m whereas the subtraction does the reverse thing . on the other hand , for strange stars , @xmath55 is always much greater than @xmath27 and even after the addition or subtraction of the smaller term @xmath56^{1/2}\}$ ] , it ( @xmath44 ) remains greater than @xmath27 . in fig [
fig : rnuco ] , i plot the variation of @xmath60 as a function of @xmath49 .
no significant difference in @xmath60 ( at any chosen @xmath49 ) between a strange star and a neutron star having the same values of @xmath14 and @xmath17 is observed ( specially at higher values of @xmath49 ) . in comparison to the whole range of @xmath60 ,
the differences between @xmath60 for the co - rotating and the counter - rotating motions ( keeping all of the other parameters fixed ) and the variation of @xmath60 with @xmath17 ( keeping all of the other parameters fixed ) are very small .
@xmath61/ \left[\nu_{k,~counter}(r ) \right ] |$ ] varies around 0.2 to 0.001 for @xmath62 0 to 500 km depending slightly upon the choice of the eos , @xmath14 and @xmath17 ; @xmath63 / \left[\nu_{k,~0.2}(r ) \right]| $ ] varies 0.1 to 0.001 for @xmath62 0 to 500 km ( where the 3rd parameter in the subscript denotes the value of @xmath17 in units of @xmath64 ) depending slightly upon the choice of the eos , @xmath14 and the direction of the motion .
so as an example , the plot in fig [ fig : rnuco ] is only for co - rotating motion with @xmath65 and for only two chosen values of @xmath14 .
but these differences between @xmath66 and @xmath67 become larger for smaller values of @xmath49 where @xmath68 , but at sufficiently higher values of @xmath49 , @xmath69 .
this transition occurs at @xmath70 km for both strange stars and neutron stars with @xmath71 and @xmath72 .
similar trends have been noticed for other values of @xmath14 and @xmath17 .
the value of @xmath49 at which this transition occurs increases slightly with the increase of @xmath14 but does not depend significantly with @xmath17 .
we will now study the variation of @xmath73 @xmath0 @xmath60 at @xmath74 with @xmath14 taking different values of @xmath17 and both co - rotating and counter - rotating motions .
@xmath75 is the keplerian frequency of a particle orbiting the star at the star s equatorial surface and i obtain it by equating the gravitational force with the centrifugal force . in figs [ fig : mnukreqco ] and [ fig :
mnukreqcounter ] i plot @xmath73 with the stellar mass for the co - rotating and the counter - rotating motions . @xmath76
let us concentrate mainly on fig [ fig : mnukreqco ] as @xmath77 can represent the rotational frequency of both the constituent particles of the star and the accreting material at the equatorial surface of the star whereas @xmath78 can represent only the accreting material .
i see that @xmath79 in the chosen range of @xmath14 and @xmath7 .
[ fig : delnu ] shows that the difference @xmath80 decreases at higher @xmath7 and/or lower @xmath14 .
the condition @xmath81 is the mass shedding limit " as for @xmath82 , matter from the star would fly way due to the centrifugal force .
the spin frequency where the mass - shedding " starts ( i.e. @xmath83 ) is the maximum possible rotational frequency of the star and it is known as the keplerian frequency of the star ( @xmath84 ) and the corresponding angular frequency is called as the keplerian angular frequency ( @xmath85 ) of the star . for fixed values of @xmath14 and @xmath7 , @xmath86 .
this implies that the mass shedding limit " will be reached in case of neutron stars at comparatively a lower value for @xmath7 than for strange stars which means that strange stars are more stable than neutron stars against rotations ( we have checked that this fact remains true even if one uses bag model eos for strange stars ) .
the keplerian angular frequency @xmath87 for compact stars has been studied in the literature by different groups like friedman , ipser & parker ( 1986 ) ; haensel & zdunik ( 1989 ) ; lattimer @xmath3 1990 and many other authors .
haensel & zdunik ( 1989 ) proposed an analytic relation between @xmath85 and maximum allowable static mass ( @xmath19 ) and the corresponding radius ( @xmath88 ) @xmath89 with their choice of neutron star esos , lattimer @xmath3 ( 1990 ) found @xmath85 to lie between @xmath90 from eq .
( [ eq : haen_zdun ] ) . using bag model esos for strange stars , prakash , baron & prakash ( 1990 ) found @xmath91 for @xmath92 .
afterwards , lattimer and prakash ( 2004 ) gave a more useful expression of @xmath85 of a star of mass @xmath14 and non - rotating radius @xmath93 .
@xmath94 to test these two simple analytical expressions ( eqns .
[ eq : haen_zdun ] , [ eq : latt_prakash ] ) , i run the task _ kepler _ " in rns code which produces stellar configurations for stars rotating with @xmath85 . using the rns outputs ( @xmath0 , @xmath14 , @xmath93 , @xmath32 , @xmath85 ,
i calculate @xmath77 from eqn .
[ eq : nuk ] . at @xmath85
, i should get @xmath95 ( by the definition of @xmath85 ) . in fig
[ fig : omega_kep ] , i plot the variation of @xmath85 with mass .
the line labeled ( 1 ) is the maximum limit of @xmath85 given by haensel & zdunik ( 1989 ) @xmath0 eqn [ eq : haen_zdun ] .
the curve labeled ( 2 ) is from the analytical expression given by lattimer & prakash ( 2004 ) @xmath0 eqn [ eq : latt_prakash ] .
the curve labeled ( 3 ) is the output of the rns code and the curve labeled ( 4 ) is @xmath96 . for neutron stars ( apr eos ) the curves ( 2 ) , ( 3 ) , ( 4 ) are very close to each other which supports the correctness of both the analytical expression of lattimer & prakash ( 2004 ) ( eqn [ eq : latt_prakash ] ) and the pseudo - newtonian potential of mukhopadhyay & misra ( 2003 ) .
the maximum value of @xmath85 obtained using eqn .
( [ eq : haen_zdun ] ) is @xmath97 . here
the kerr parameter @xmath34 lies around 0.66 which is much greater than the values at lower frequencies ( fig .
[ fig : ma ] ) and the neutron stars are very oblate having @xmath98 . for eos
a , the rns code failed to perform the task _ kepler _ " for @xmath99 and at @xmath100 , i get @xmath101 which is greater than @xmath19 for static configuration .
moreover , at this mass , the value of the kerr parameter @xmath34 is greater than 1 , which is unphysical . at @xmath102 ( @xmath103 ) ,
@xmath34 becomes less than 1 , but very high ( @xmath104 ) .
so no direct comparison of @xmath85 obtained with the rns code with the analytical expressions are possible and in the figure i only plot the analytical expressions . using eqn .
( [ eq : latt_prakash ] ) , i get @xmath105 at @xmath106 and increases with the increase of @xmath16 ( or @xmath14 ) .
this value of @xmath85 is less than the maximum value of @xmath85 obtained using eqn .
( [ eq : haen_zdun ] ) which is @xmath107 . here
the strange stars are very oblate with @xmath108 . to check whether all these facts are intrinsic to strange star properties or depend upon the particular model of strange stars ,
i have even checked with the bag model having model parameters as : @xmath109 where @xmath110 is the bag parameter , @xmath111 , @xmath112 , @xmath113 are masses of s , u and d quarks respectively giving static @xmath114 . here
also i get @xmath115 at lower @xmath16 , @xmath116 when @xmath117 ( @xmath118 ) . using eqn .
( [ eq : latt_prakash ] ) , i get @xmath119 at @xmath120 which increases with increase of @xmath16 ( or @xmath14 ) .
this value of @xmath85 is less than the maximum value of @xmath85 obtained using eqn .
( [ eq : haen_zdun ] ) which is @xmath121 .
the strange stars are very oblate having @xmath122 .
the probable fastest spin frequency of a neutron star ( xte j1739 - 285 ) is @xmath8 hz or @xmath123 , which is less than the value of @xmath85 of both strange stars and neutron stars as derived from eqn .
( [ eq : haen_zdun ] ) or even less than as derived from eqn .
[ eq : latt_prakash ] for a canonical value of the stellar mass as @xmath124 ( see fig . [
fig : omega_kep ] ) .
so there is no problem of this star being either a strange star or a neutron star .
one needs to conclude about its nature by other observational evidences .
several other people performed numerical studies on structures of rapidly rotating compact stars like friedman & ipser ( 1992 ) , weber & glendening ( 1992 ) ; cook , shapiro & teukolsky ( 1994 ) ; eriguchi , hachisu & nomoto ( 1994 ) ; salgado @xmath3 ( 1994a , b ) ; gourgoulhon @xmath3 ( 1999 ) ; bhattacharyya , thampan & bombaci ( 2000 ) ; bombaci , thampan & datta ( 2000 ) , gondek - rosiska @xmath3 ( 2000 ) ; bhattacharyya & ghosh ( 2005 ) ; haensel , zdunik & bejger ( 2008 ) ; haensel @xmath3 ( 2009 ) . in these
, people usually kept themselves confined in studying the properties of stars like @xmath14 , @xmath16 , @xmath27 , @xmath125 etc or at most properties particles rotating at the stellar surface .
but the use of the pseudo - newtonian potential enables us to study the properties of particles orbiting around the star at a much higher distance ( @xmath126 ) and i report quantities like @xmath127 and @xmath128 , @xmath129 , @xmath130 which are useful to study accretion onto rotating neutron stars or strange stars ( see next section ) .
i also report various stellar parameters @xmath14 , @xmath93 , @xmath32 , @xmath33 , @xmath34 , @xmath131
, @xmath132 @xmath133 .
there is no previous work where all this parameters were reported together .
also i use one eos for strange stars and another eos for neutron stars whereas in the earlier works people discussed either only neutron star rotations or only strange star rotations .
i also discuss about khz qpos within the scenario of mukhopadhyay @xmath3 ( 2003 ) in the next section .
khz qpos in the lmxbs are very interesting phenomena . the most popular model to explain khz qpos is the beat frequency model ( strohmayer @xmath3 1996 ) which assumes the upper qpo ( @xmath134 ) as the keplerian orbital frequency @xmath60 of the innermost orbit in the accretion disk around the star , the separation between the two peaks @xmath0 @xmath135 as @xmath7 and the lower qpo ( @xmath136 ) as the beat frequency of @xmath60 and @xmath7 .
this model suggest that for a particular star @xmath137 will remain constant .
however , it has been observed that in many sources , @xmath138 changes with time .
this phenomenon discards the reliability of the beat frequency model for the qpos .
two examples of alternate models can be found in titarchuk and osherovich ( 1999 ) and mukhopadhyay et al .
both of these later two models suggest @xmath139 .
but in the present work , i do nt prefer any particular qpo model over the others .
we take two sample lmxb for which khz qpos have been observed , namely ks 1731@xmath5260 with @xmath140 hz and 4u 1636@xmath553 @xmath141 hz . as none of the above mentioned models for khz qpos
is established beyond doubt , i fit both @xmath134 and @xmath136 to @xmath60 to find the corresponding radial distance @xmath142 ( from eq .
[ eq : nuk ] ) which i call @xmath143 and @xmath144 respectively .
mukhopadhyay @xmath3 ( 2003 ) calculated @xmath144 for these two sources taking mass - radius values for non - rotating strange stars and using @xmath145 ( @xmath146 in their notation ) as an parameter . here
i get @xmath143 and @xmath144 taking exact values of @xmath27 and @xmath145 as generated by the rns code with the specific values of @xmath7 and chosen values of @xmath14 ( as masses of these two stars are not well determined ) .
qpo frequencies for ks 1731@xmath5260 has been taken from wijnands and van der klis ( 1997 ) and those for 4u 1636@xmath553 has been taken from jonker , mendez & van der klis ( 2002 ) .
note that the qpo frequencies of 4u 1636@xmath553 shift with time and hence i have taken only one set of values as an example .
table [ tb : qpo1 ] shows the rotational parameters for ks 1731@xmath5260 and 4u 1636@xmath553 .
table [ tb : qpo2 ] shows the values of the radius of the keplerian orbit ( @xmath147 ) obtained by equating @xmath148 with qpo frequencies .
i choose the stellar mass to be 1.1 @xmath15 , 1.5 @xmath15 and 1.9 @xmath15 . for a given star ( @xmath149 is fixed ) ,
the values of @xmath44 depend significantly on ( a ) whether i consider the co - rotating or the counter - rotating motion , ( b ) on the choice of the eos and ( c ) on the chosen value of @xmath14 .
but the values of @xmath147 depend significantly only on @xmath14 when i keep @xmath7 and @xmath150 fixed . for ks 1731@xmath5260 , @xmath151 km for both strange stars and neutron stars ( independent of the choice of mass ) and for both co - rotating and counter - rotating motions whereas for 4u 1636@xmath553
this value is @xmath152 km .
@xmath153 to @xmath154 for ks 1731@xmath5260 and @xmath155 to @xmath156 km for 4u 1636@xmath553 for co - rotating motions , @xmath157 to @xmath158 km for ks 1731@xmath5260 and @xmath159 to @xmath156 km for 4u 1636@xmath553 for counter - rotating motions ; whereas @xmath160 to @xmath161 km for ks 1731@xmath5260 and @xmath162 to @xmath163 km for 4u 1636@xmath553 for co - rotating motions
@xmath164 to @xmath165 km for ks 1731@xmath5260 and @xmath166 to @xmath167 km for 4u 1636@xmath553 for counter rotating motions .
these values decrease with the increase of @xmath14 and note that at @xmath168 , @xmath169 .
remember , @xmath170 for eos a for the values of @xmath17 of ks 1731@xmath5260 and 4u 1636@xmath553 .
[ cols="^,^,^,^,^,^,^,^,^,^ " , ] [ tb : qpo2 ]
we study the rotational parameters for strange stars and neutron stars using the rns code which consider the generalized axisymmetric metric for rotating stars .
i find that the rotational parameters like @xmath19 , @xmath27 , @xmath32 , @xmath33 , @xmath145 depend on @xmath7 and the choice of the eos .
the value of @xmath44 depends on the stellar mass , the choice of the eos and whether the motion is co - rotating or counter - rotating .
the dependence of @xmath60 on the choice of eos , @xmath17 and the direction of motion is prominent only at low @xmath49 ( @xmath171 ) .
that is why , the values of @xmath147 as obtained by fitting the khz qpo frequencies do not depend much on the choice of the eos and on the direction of motion as here @xmath172 .
the beat frequency model suggests that @xmath143 is the radius of the innermost keplerian orbit of the ( @xmath173 ) of the accretion disk whereas the models by titarchuk and osherovich ( 1999 ) and mukhopadhyay @xmath3 ( 2003 ) suggest that @xmath174 .
the disk parameter @xmath173 can be estimated by x - ray spectral analysis .
as relativistic broadening is more dominant at the innermost edge of the disk , the existence of a relativistically broadened iron k@xmath175 line helps one to determine the value of @xmath173 ( reis , fabian & young 2009 ; di salvo @xmath3 2009 and references therein ) .
it is clear from the table [ tb : qpo2 ] that for a fixed set of @xmath14 and @xmath7 , @xmath173 is different for different esos , but the difference is always @xmath176 , whereas presently @xmath173 is measured only accuracy up to @xmath177 ( reis , fabian & young 2009 ; di salvo @xmath3 2009 ) .
moreover , one should remember that this determination of @xmath173 depends upon the qpo model which is not beyond doubt at the present moment .
indeed there is a strong need of better explanation of khz qpos to understand various features like the shift of the peaks , their correlations ( belloni , mndez and homan 2007 ; yin @xmath3 2007 ) , side - bands ( jonker , mndez and van der klis 2005 ) etc . considering all these facts ,
i conclude that , to constrain dense matter eos by measuring @xmath173 using x - ray spectral analysis , i need much better accuracy and hopefully future advanced technology like astrosat will provide us such ultra - high accuracy . in this work
i have primarily confined myself to a maximum value of @xmath7 as 796 hz .
but then i have also studied the properties of the compact stars if they rotate with the maximum spin frequency @xmath0 the keplerian spin frequency .
bagchi , m. , ray , s. , dey , j. , & dey , m. , 2006 , astron . & astrophys . , 450 , 431 .
akmal , a. , pandharipande , v. r. , ravenhall , d. g. , 1998 , phys .
c 58 , 1804 .
bagchi , m. , dey , j. , konar , s. , bhattacharya , g. , dey , m. , arxiv : astro - ph/0610448 .
bardeen , j. , press , w. h. , teukolsky , s. a. , 1972 , astrophys .
j. 178 , 347 .
belloni , t. , mendez m. , homan j. , 2007 , mon . not .
376 , 1133 .
bhattacharyya , a. , ghosh , s. , 2005 , arxiv : astro - ph/0506202 .
bhattacharyya , s. , thampan , a. v. , bombaci , i. , 2001 , astron . &
astrophys 372 , 925 .
bombaci , i. , thampan , a. v. , datta , b. , 2000 , astrophys .
j. 541 , l71 .
cook , g. b. , shapiro , s. l. , teukolsky , s. a. , 1994 , apj , 424 , 823 . di salvo , t. , d ai , a. , burderi , l. , @xmath3 2009 , arxiv:0904.3318 .
eriguchi , y. , hachisu , i. , nomoto , k. , 1994 , mnras , 266 , 179 .
friedman , j. l. , ipser , j. r , parker , l. , 1986 , astrophys .
j. 304 , 115 .
friedman j. l. , ipser , j. r , 1992 , phil .
lond . a , 340 , 391 .
gondek - rosiska , d. , bulik , t. , zdunik , l. , gourgoulhon , e. , ray , s. , dey , j. , dey , m. , 2000 , astron . & astrophys .
363 , 1005 .
gourgoulhon , e. , haensel , p. , livine , r. , @xmath3 , 1999 , a & a , 349 , 851 .
haensel , p. , zdunik , j. l. , 1989 , nature 340 , 617 .
haensel , p. , zdunik , j. l. , bejger m. , 2008 , new astron . rev . 51 , 785 .
haensel , p. , zdunik , j. l. , bejger , m. , lattimer , j. m. , 2009 , arxiv:0901.1268 hessels , j. w. t. , ransom , s. m. , stairs , i. h. , 2006 , science , 311 , 1901 .
jonker , p. g. , mndez , m. , van der klis , m. , 2002 , mon . not .
soc 336 , 1 .
kaaret , p. , in t zand , j. j. m. , brabdt , s. @xmath3 , 2007 , astrophys .
j. 657 , l97 .
komatsu , h. , eriguchi , y. , hachisu , i. , 1989 , mon . not .
237 , 355 .
klahn , t. , blaschke , d. , typel , s. , @xmath3 , 2006 , phys.rev .
c 74 , 035802 .
lattimer , j. m. , prakash , m. , 2004 , science 304 , 536 .
lattimer , j. m. , prakash , m. , 2007 , physics reports 442 , 109 .
lattimer , j. m. , prakash , m. , masak , d. , yahil , a. , 1990 , astrophys .
j. 355 , 241 .
lattimer , j. m. , schutz , b. f. , 2005 , astrophys .
j. 629 , 979 .
li , x. d. , ray , s. , dey , j. , dey , m. , bombaci , i. , 1999 , astrophys .
j. 527 , l51 .
mukhopadhyay , b. , misra , r. , 2003 , astrophys .
j. 582 , 347 .
mukhopadhyay , b. , ray , s. , dey , j. , dey , m. , 2003 , astrophys .
j. 584 , l83 .
zel , f. , 2006 , nature , 441 , 1115 .
zel , f. , arxiv:0810.1521 .
reis , r. c. , fabian , a. c. , young , a. j. , 2009 , arxiv:0904.2747 .
salgado , m. , bonazzola , s. , gourgoulhon , e. , haensel , p. , 1994a
, a & a , 291 , 155 .
salgado , m. , bonazzola , s. , gourgoulhon , e. , haensel , p. , 1994b
, a & a suppl . , 108 , 455 .
strohmayer , t. e. , zhang , w. , swank , j. h. , @xmath3 1996 , astrophys .
j. 469 , l9 .
titarchuk , l. , and osherovich , v. , 1999 , astrophys .
j. 518 , l95 .
weber , f. , glendenning , n. , 1992 , astrophys .
j. 390 , 541 .
wijnands , r. a. d. and van der klis , m. , 1997 , astrophys .
j. 482 , l65 .
yin , h. x. , zhang , c. m. , zhao , y. h. , @xmath3 2007 , astron . and astrophys . 471 , 381 .
zhang , c. m. , yin , h. x. , kojima , y. , @xmath3 , 2007 , mon . not .
374 , 232 . | i study stellar structures @xmath0 the mass , the radius , the moment of inertia and the oblateness parameter at different spin frequencies for strange stars and neutron stars in a comparative manner .
i also calculate the values of the radii of the marginally stable orbits and keplerian orbital frequencies . by equating khz qpo frequencies to keplerian orbital frequencies ,
i find corresponding orbital radii .
knowledge about these parameters might be useful in further modeling of the observed features from lmxbs with advanced and improved future techniques for observations and data analysis .
= -1.5 true cm @xmath1 tata institute of fundamental research , colaba , mumbai 400005 , india + @xmath2 [email protected] .5 cm keywords : dense matter ; equation of state ; x - rays : binaries ; stars : neutron ; stars : rotation + pacs : 26.60.kp ; 97.10.nf ; 97.10.pg ; 97.60.jd ; 97.80.jp ; 97.10.kc |
a two - qubit maximally entangled state ( mes ) of a two - state system can be written as@xcite , @xmath3 where the coefficients @xmath4 and @xmath5 are normalized to unity @xmath6 it is seen that a mes is defined by a pair of complex numbers @xmath7 . to visualize a mes , @xmath4 and @xmath5
can be parameterized to @xmath8 where @xmath9 is a unit vector , and @xmath10 is an angle between @xmath11 and @xmath0 .
hence a mes can also be written as @xmath12 in the parameter space .
it is easy to check that @xmath13 .
that is , @xmath14 and @xmath15 correspond to the same state except for a global phase factor .
this is just the case of the double - valued representation of the so(3 ) group , which is written as @xmath16 corresponding to a rotation @xmath17 in real space to a two - state particle .
although @xmath18 and @xmath19 are the same rotation , one has @xmath20 .
therefore , there is a one - to - one correspondence between the two - qubit mes and the double - valued representation of so(3 ) .
in fact , any mes can be constructed by a rotation from an initial mes , e.g. , @xmath21 where @xmath22 operates on the first particle .
if a rotation operates on the second particle , one has @xmath23
one could define a so(3 ) sphere with diameter @xmath0 filled by vectors @xmath24 .
due to ( [ p10 ] ) a mes corresponds to a point in the so(3 ) sphere , and an evolution of mes corresponds to a trajectory connecting two points .
the initial state @xmath25 locates at the center of the so(3 ) sphere .
consider an electron in a rotating magnetic field @xmath26 where @xmath27 is the angle between the field and the z - axis , and @xmath28 is the rotating frequency of the field .
the hamiltonian of the electron is given by @xmath29 the two exact solutions of the time - dependent schrdinger equation are given by @xmath30 corresponding to energy eigenvalues @xmath31 and @xmath32 , respectively , where @xmath33 where the values of @xmath34 can be determined by normalization of solutions .
now consider an initial state @xmath35/\sqrt(2 ) \\
\label{evolution1 } & = & |(\alpha , \beta)\rangle = d_1(\omega t,\omega_0 ) |(1,0)\rangle\end{aligned}\ ] ] where new arguments have been assigned to the group element for convenience , and @xmath36 e^{-i\omega_0 t } , \\
\label{b1 } \beta & = & i \sin\frac { \omega t } 2 \frac { a_+ } { a_- } \frac { \hbar ( \omega-2\omega_0)-2b \cos\theta } { 2\hbar \omega_0 } e^{i\omega_0 t}.\end{aligned}\ ] ] it is seen that a rotating magnetic field leads to an evolution of a mes through a continuous trajectory in the so(3 ) sphere .
therefore , a rotating magnetic field is equivalent to a three dimensional rotation in real space to the mes .
it is not surprising that when @xmath37 , the initial state acquires an additional phase of @xmath0 , i.e. , @xmath38 the amazing property is that the above operation can be allocated to two particles of the initial state , i.e. , @xmath39 in general , one does not have such a property for other initial state . if @xmath40 one will have @xmath41 acquiring no additional phase .
hence , one has a choice for the additional phase through selecting the value of @xmath42 .
now we can trace the following evolution @xmath43/\sqrt(2 ) \\ & = & d_1(\omega t,\omega_0 ) d_2(\omega t,\omega_0 ) |(1,0)\rangle\end{aligned}\ ] ] under the choice @xmath44 this evolution makes a closed trajectory in the so(3 ) sphere .
an example is shown in fig.1 , where parameters @xmath27 and @xmath45 are set to meet @xmath46 .
the final time is @xmath47 , that is , both magnetic fields of the two electrons rotate half a cycle . it is seen that this trajectory breaks three times on the surface of the sphere .
it is known that two ends of a diameter of the sphere correspond to the same rotation but the group element , ( [ dka ] ) , changes its sign . in this case , through a whole trajectory , the mes acquires an additional phase of @xmath0 . .
the arrow stands for the beginning state and direction of evolution . ] with proper parameters , one can have closed trajectories with even numbers of breaks , corresponding to a change of @xmath48 in phase .
an example is shown in fig.2 .
hence , one has two classes of trajectories , one of which has an odd number of breaks on the surface of the so(3 ) sphere and the other has an even number of breaks , corresponding to the two classes of the double connectedness of the so(3 ) group .
this is the case that milman and mosseri considered@xcite , whereas their trajectories are hardly possible to be realized , since their magnetic field has to jump through a few discrete points in the parameter space . .
the arrow stands for the beginning state and direction of evolution . ]
it can easily checked that the closed trajectories @xmath49 and @xmath50 and other ones that milman and mosseri considered @xcite belong to the two simplest classes of trajectories which have @xmath11 or @xmath51 breaks , respectively , in the so(3 ) sphere .
therefore , the present work extends their model to include a great number of closed trajectories with even or odd numbers of breaks .
a entangled photon pair emerging from a double refraction crystal@xcite can be in one of four bell states @xmath52 , where @xmath53 and @xmath54 denote a horizontally polarized photon state and a vertically polarized one , respectively .
this two entangled photons separate with each other after emission , and then pass through two negative kerr medium p1 and p3 , and two positive kerr medium p2 and p4 , as seen in fig.3 .
the kerr medium are modulated by electric fields , so that their optical axes are in directions as shown in the lower part in fig.3 .
p1 and p3 point to the same direction , say z - axis , and the directions of p2 and p4 can be adjusted by changing the directions of their electric fields .
p1 and p3 change the relative phase between @xmath53 and @xmath54 , working as the following matrix @xmath55 where @xmath56 is the electric field applied to kerr medium p1 and p3 . since the optical exes of p2 and p4 take an angle of @xmath57 with the z - axis , they work as the following matrix , @xmath58 where @xmath59 is the electric field applied to p3 and p4 .
it is seen that the combination , @xmath60 , is equivalent to a rotating magnetic field . by comparing eqs.([evolution1],[a1],[b1 ] ) and
eqs.([u1],[a],[b ] ) one finds correspondence @xmath61 .
hence , the evolution in eq.([evolution ] ) , as shown in fig.1 and fig.2 , can be exactly traced by varying the electric fields @xmath56 and @xmath59 on the kerr medium . with proper values of electric fields such that @xmath62 one may obtain an additional @xmath0 phase , or zero additional phase if @xmath63 .
the @xmath0 phase can be easily observed by various interference experiments .
for example , according to the scheme described by milman and mosseri considered @xcite , one arm of the entangled photon pair can be transformed into a mach - zender interferometer .
the two wave plates in that scheme are replaced by combinations of kerr medium p1 and p2 , and that of p3 and p4 . in summary , the present paper sets up a representation for the so(3 ) group by maximally entangled two - qubit states .
the evolution of the entangled states showed the double connectedness of the so(3 ) group . in the so(3 ) sphere the evolution path breaks an odd or even number of times .
an odd number of breaks causes an additional @xmath0 phase to the entangled state , but an even number of breaks does not .
the additional @xmath0 phase can be observed by interference experiments of entangled photon pairs .
s. a. werner _
et al . _ ,
* 35 * , 1053(1975 ) .
_ , phys . lett . *
54a * , 425(1975 ) .
p. milman and r. mosseri , phys .
lett . * 90 * , 230403 - 1(2003 ) .
w. t. buttler , _ et al .
_ , phys . rev
. lett . * 81 * , 3283(1998 ) .
a. zeilinger , phys .
world * 11 * , 35(1998 ) .
jian - wei pan _
et al . _ ,
nature * 423 * , 417(2003 ) ; p. g. kwiat , k. mattle , _ et al .
75 * , 4337(1995 ) . | a representation of the so(3 ) group is mapped into a maximally entangled two qubit state according to literatures . to show the evolution of the entangled state , a model
is set up on an maximally entangled electron pair , two electrons of which pass independently through a rotating magnetic field .
it is found that the evolution path of the entangled state in the so(3 ) sphere breaks an odd or even number of times , corresponding to the double connectedness of the so(3 ) group .
an odd number of breaks leads to an additional @xmath0 phase to the entangled state , but an even number of breaks does not .
a scheme to trace the evolution of the entangled state is proposed by means of entangled photon pairs and kerr medium , allowing observation of the additional @xmath0 phase .
it is well known that when the spin of a spin-@xmath1 particle rotates for a whole cycle on the bloch sphere the wave - function of the particle changes a phase of @xmath0 .
this @xmath0 phase has been observed in several experiments@xcite .
this property is commonly attributed to the topological property , i.e. , the double connectedness of the so(3 ) group . the path on the manifold of the so(3 ) group is categorized into two classes under a continuous deformation , one of which leads to a change of @xmath0 in phase to the wave function , the other does not .
however , it was argued by milman and mosseri@xcite that this @xmath0 phase may be shared by the multi - connectedness of both so(3 ) and so(2 ) groups .
they also argued that , in general , the @xmath0 phase is partly geometric and partly dynamic . only in the extreme case that the spin precesses on the @xmath2 plane in the bloch sphere the @xmath0 phase is fully geometric .
especially , the @xmath0 phase still exists when the spin initially points to the same direction of the magnetic field , where there is no rotation at all . in such case
the @xmath0 phase is fully dynamic .
therefore , this @xmath0 phase may not be directly related to the so(3 ) group .
milman and mosseri found a one - to - one correspondence between the representation of the so(3 ) group and the evolution of a maximally entangled state of a two - qubit system(mes)@xcite .
they adopted a discontinuously changing magnetic field , which suddenly jumps from one direction to another .
this is hardly possible to be accomplished in reality . in the present paper
a rotating magnetic field is used to drive the evolution of a mes .
a clearer formalism is presented for the trajectory in so(3 ) .
a mes finds great application to quantum communication and quantum computation techniques , and also to the study of fundamental problems , e.g. , non - locality , of quantum mechanics@xcite .
much attention has been paid to mes s in recent years .
it is interesting that mes can be applied to the representation of the so(3 ) group . |
x - ray emission of late - type stars arises in hot gas contained by magnetic loops above the stellar surface .
these loops are produced by the interaction of convective motion and differential rotation in the stellar envelope ; hence the magnetic activity of a late - type main - sequence star and with it the x - ray emission is found to be higher in rapidly rotating stars ( see , for example , * ? ? ?
as stars age their rotation slows down , and it was a surprise that a _ rosat _ study of the old open cluster m67 discovered a large number of x - ray sources @xcite .
comparison with optical observations provided the explanation : in an old cluster stars may rotate rapidly when tidal forces spin them up towards corotation with the binary revolution @xcite .
x - ray studies of old open clusters help in identifying such , otherwise inconspicuous , active binaries ( abs ) in which tidal forces are or have been effective , and contribute to the understanding of tidal interaction . since the typical timescale for tidal interaction rapidly increases with the ratio of the orbital separation to stellar radius @xcite , only stars in binaries with relatively short periods or relatively large radii experience the effects of tidal locking .
x - ray observations of abs in old clusters thus probe the populations of hard binaries , and the relative number of such systems found in clusters with different properties is an important tool for understanding the effects of the cluster environment on the binary content . thanks to prolonged optical studies , e.g. in the wiyn open cluster study @xcite , our knowledge of the binary population in clusters is reaching completeness levels which enable a detailed comparison with theoretical models of the evolution of stellar clusters and of the binary population in them .
to explain the large number of blue stragglers observed in m67 ( 4 gyr ) , the model by @xcite requires a large initial number of short - period binaries that act as seeds from which blue stragglers are formed .
as the simulated hard - binary fraction and period distribution do not evolve much , the close binaries remain abundant throughout the life time of the cluster .
@xcite point out that comparison with the observed binaries in the old cluster ngc188 ( 6.5 gyr ; @xcite ) finds no evidence of such a high fraction of close binaries .
the number of clusters studied is still small , as is the number of detailed models , and more study is required before a final verdict on these and other discrepancies ( see @xcite for details ) can be made .
by now , observations of globular clusters with the _ chandra x - ray observatory _ have uncovered hundreds of faint ( @xmath4 erg s@xmath1 ) close binaries , only part of which have been classified so far . as these binaries are an important driver of the long - term evolution of a cluster ,
understanding the properties and frequency of the various binary classes is a major goal of investigating ( globular ) cluster x - ray sources .
sometimes the outcome of a dynamical encounter is a binary or multiple system in a very unusual configuration , and this allows us to study source types that are not , or only rarely , found in the field . besides abs , which are the dominant x - ray source class in old open clusters , globular - cluster x - ray sources include cataclysmic variables ( cvs ) , low - mass x - ray binaries in quiescence ( qlmxbs ) , and milli - second pulsars ( msps ) .
the correlation of the numbers of qlmxbs and cvs with the stellar encounter frequency reveals the important role that dynamical encounters play in the creation of such binaries @xcite , but the net balance between formation and destruction is less clear . for a large part
this is due to the lack of a reference point that quantifies their number densities in a dynamically inactive environment .
ideally , the comparison should be made against the galactic field where stellar densities are low , but this is complicated because of the generally limited information on distance and age for stars in the field , and the intrinsic scarcity of sources such as qlmxbs . old open clusters provide an alternative that is worthwhile to explore , but so far only few have been studied in detail in x - rays .
moreover , it should be kept in mind that dynamical encounters can not be completely ignored in open clusters .
observational indications for the occurrence of dynamical interactions have come from individual systems whose properties are difficult to explain by binary evolution as it would take place outside a cluster environment ( e.g. * ? ? ?
in this paper we present the results of the first x - ray observation of ngc6791 , which , at an age of about 8 gyr , is one of the oldest known open clusters in our galaxy .
besides the _ chandra _ data , we obtained optical spectra to classify candidate optical counterparts to the detected x - ray sources .
ngc6791 has been studied extensively in the optical .
the rich body of available literature has proven to be very useful for the classification of our _ chandra _ sources , and for separating cluster members from non - members . for the study of interacting binaries it is a promising target : before this work it was known to harbor two of only three spectroscopically - confirmed cvs found in open clusters @xcite , and many optical - variability studies have revealed dozens of close binaries in the field of the ngc6791 ( see @xcite and references therein ) .
it is therefore interesting to investigate if the x - ray source content of ngc6791 is as rich and varied as was found for the few old open clusters studied previously , viz .
m67 @xcite and ngc188 @xcite .
the cluster lies at a distance of @xmath34.04.3 kpc and is reddened by @xmath5=0.090.16 ; see e.g. @xcite , @xcite , and @xcite for recent determinations of the cluster parameters . throughout the paper
we adopt a distance of 4.1 kpc and a constant reddening @xmath6 , which corresponds to a neutral - hydrogen column density @xmath7 @xmath8 @xcite .
the reddening could vary across the area of the cluster @xcite but the effect is too small to have a significant impact on our results .
ngc6791 stands out in several ways , including its high age and large mass .
remarkably for such an old cluster , the metallicity is higher than solar ( [ fe / h]@xmath9 + 0.4 ) .
the x - ray and optical observations and data reduction are described in sect .
[ sec_obs ] , followed by the data analysis in sect .
[ sec_ana ] .
we detected various types of sources in the cluster which are presented in sect .
[ sec_results ] .
our results are discussed in the context of the populations of interacting binaries in star clusters in sect .
[ sec_disc ] .
we finish with our conclusions in sect .
[ sec_conc ] .
we observed ngc6791 with the advanced ccd imaging spectrometer ( acis ) on _ chandra _ from 2004 july 1 20:51 utc until july 2 10:49 utc for a total exposure time of 48.2 ks ( obsid 4510 ) .
the observation was done in very faint , timed exposure mode with a single - frame exposure time of 3.2 s. in order to achieve optimal sensitivity below @xmath31 kev , we placed the central part of the cluster on the backside - illuminated s3 chip ; the neighboring s1 , s2 , s4 , i2 , and i3 chips were used to cover the outer parts of the cluster .
the cluster center lies at approximately @xmath10=19@xmath1120@xmath1253@xmath13 , @xmath14=+37@xmath154618 ( j2000 ; @xcite ) ; we shifted the center 05 away from the s3 aimpoint in the @xmath16 ( approximately south - east ) direction so that a larger part of the cluster could be imaged on a single chip .
@xcite derive a cluster half - mass radius @xmath17 of @xmath18 .
our observation covers almost the entire area inside @xmath17 ( @xmath19% ) with either the s3 or the s2 chip ; only the southernmost tip falls off the detector area .
[ fig_fov ] shows the outline of the acis chips overlayed on an optical image of the cluster . -band image of ngc6791 from @xcite showing in black the outline of the acis chips ( labelled ) as in our _ chandra _ observation .
the small white , innermost circle indicates the cluster center from @xcite .
the larger white circles mark the area inside the half - mass radius @xmath17 ( 442 ) , and the area inside 8 , which was analyzed in this paper .
north is up , east to the left .
[ fig_fov],width=313 ] our data reduction started with the level-1 event file produced by version 7.6.7.2 of the _ chandra _ x - ray center processing pipeline .
we used the ciao 3.4 package with the caldb 3.3.0 calibration files for further processing following standard procedures , including the very - faint mode background cleaning . in order to improve the precision of the source positions we removed the randomization of event
coordinates that is applied in standard processing . towards the end of the observation , the background level between 0.3 and 7.0 kev increased by a factor of @xmath35 as a result of a background flare . after removing this interval from the observation , the exposure time
is reduced to 42.9 ks .
we obtained low - resolution spectra of candidate optical counterparts to guide the classification of the x - ray sources .
a total of 16 candidate counterparts brighter than @xmath20 were observed with the fast long - slit spectrograph on the 1.5-m tillinghast telescope on mt .
hopkins on 9 nights between 2005 june 7 to september 2 .
we used the 300 lines mm@xmath1 grating , resulting in a wavelength coverage from 3480 to 7400 and a 3- resolution .
exposure time was chosen to achieve a signal - to - noise ratio @xmath21 @xmath22 for @xmath23 , and @xmath24 for fainter sources .
fast spectra were extracted and wavelength - calibrated with a dedicated reduction pipeline .
flux standards were observed on the same nights as the science targets .
candidate optical counterparts fainter than @xmath25 were observed with the fiber - fed multi - object spectrograph hectospec on the 6.5-m multi - mirror telescope .
use of the 270 lines mm@xmath1 grating resulted in spectra that cover 3700 to 9150 with a resolution of 6 .
a total of 16 candidate counterparts were observed on the nights of 2005 may 13 and july 46 .
each setup was repeated 4 to 5 times with individual exposures of 900 s. exposures were combined , and the spectra extracted and wavelength - calibrated , with the hectospec reduction pipeline .
sky background spectra for each setup were created by combining the spectra of fibers positioned at off - source locations .
hectospec observations did not include flux standards .
to correct for the instrumental response we constructed sensitivity curves using the spectra of the subdwarf b ( sdb ) stars b4 , b5 and b3 in ngc6791 @xcite , included in the may 13 , july 4 and july 56 setups , respectively .
we assumed that their intrinsic spectra can be described by blackbodies with temperatures as determined for these sdb stars by @xcite .
the blackbody spectra were then normalized to reproduce the observed @xmath26 magnitudes ( @xcite ) after correction for the cluster reddening . this way we achieved an approximate absolute - flux calibration of the target spectra after applying the resulting sensitivity curves .
flux - calibrated target spectra were assigned spectral types by comparison to standard spectra with similar resolution ( e.g. * ? ? ?
lccccccccc & ( 2 ) & ( 3 ) & ( 4 ) & ( 5 ) & ( 6 ) & ( 7 ) & ( 8) & ( 9 ) & ( 10 ) + cx & cxouj & @xmath10(j2000 ) & @xmath14(j2000 ) & @xmath27 & offset & counts & @xmath28 & @xmath29 & opt + & & ( deg ) & ( deg ) & ( ) & ( ) & & ( erg @xmath8 s@xmath1 ) & ( kev ) & + + + 1 & 192044.9 + 374640 & 290.187175 & + 37.777870 & 0.32 & 1.64 & 213 @xmath30 15 & 318.0 & 1.3 @xmath30 0.1 & + + 2 & 192039.8 + 374354 & 290.165914 & + 37.731752 & 0.39 & 3.54 & 151 @xmath30 13 & 242.9 & 1.00 @xmath30 0.03 & + + 3 & 192035.7 + 374452 & 290.148996 & + 37.747852 & 0.39 & 3.70 & 125 @xmath30 12 & 260.1 & 1.4 @xmath30 0.1 & + + 4 & 192056.3 + 374613 & 290.234628 & + 37.770331 & 0.34 & 0.66 & 116 @xmath30 11 & 170.7 & 1.4 @xmath30 0.1 & + + 5 & 192047.3 + 374318 & 290.197161 & + 37.721768 & 0.43 & 3.20 & 72 @xmath30 9 & 118.0 & 1.4 @xmath30 0.1 & @xmath31 + 6 & 192114.4 + 374530 & 290.310342 & + 37.758603 & 0.63 & 4.32 & 72 @xmath30 9 & 112.0 & 1.4 @xmath30 0.1 & @xmath31 + 7 & 192052.3 + 374550 & 290.218110 & + 37.764148 & 0.35 & 0.47 & 53 @xmath30 8 & 77.9 & 1.2
@xmath30 0.1 & + + 8 & 192038.2 + 374441 & 290.159518 & + 37.744838 & 0.46 & 3.33 & 48 @xmath30 7 & 221.7 & 1.6 @xmath30 0.2 & + + 9 & 192058.4 + 375008 & 290.243732 & + 37.835590 & 0.56 & 3.98 & 37 @xmath30 7 & 59.8 & 1.6 @xmath30 0.3 & + + 10 & 192037.3 + 374612 & 290.155726 & + 37.770073 & 0.43 & 3.09 & 34 @xmath30 6 & 69.3 & 1.4 @xmath30 0.2 & @xmath31 + we restrict the analysis to the area within 8 of the cluster center .
sources further away have relatively large positional errors , which complicates the optical identification .
we focus on the area inside @xmath17 which is close to the largest circular area around the cluster center that is entirely covered by the observation .
we performed source detection in a broad ( 0.37.0 kev ) , soft ( 0.32.0 kev ) and hard ( 2.07.0 kev ) energy band , also used in our _ chandra _ study of m67 @xcite , to facilitate comparison .
the ciao detection routine _ wavdetect _ was run for scales of 1.0 to 11.3 pixels , in steps increasing by a factor @xmath32 , with the larger scales appropriate for large off - axis angles where the point - spread function ( psf ) becomes significantly broader .
we computed exposure maps for the response at 1 kev to account for spatial variations of the sensitivity .
the _ wavdetect _ detection threshold was set to 10@xmath33 , from which we expect two spurious detections per detection scale ( so sixteen spurious detections in total ) in the area that we consider here .
_ wavdetect _ positional errors reflect the statistical uncertainty in centroiding the spatial distribution of the detected events of a given source , but do not include systematic errors that result from data processing and psf asymmetries at large offset angles . to compute positional uncertainties we therefore adopt eq . 5 in @xcite .
this formula relates the 95% confidence radius on the source position @xmath27 to the _ wavdetect _ counts and offset angle from the aimpoint , and is based on extensive simulated detections of artificial sources .
combination of the broad , soft , and hard - band source lists results in a master catalog of 86 distinct sources within 8 of the cluster center , of which 59 lie inside @xmath17 .
table [ tab_src ] summarizes their basic properties .
to investigate the validity of the sources , we also ran _ wavdetect _ with a threshold of 10@xmath34 or an expected number of spurious sources of 1.6 .
the fourteen sources not detected in this run are marked with an asterisk in table [ tab_src ] . in this paper
we adopt a short version of the source names to refer to the sources instead of using their official cxo names ; see columns 1 and 2 in table [ tab_src ] .
the shorthand names are assigned by first sorting the sources within @xmath17 on net counts , and then the sources between @xmath17 and 8 .
we used the _
acis_extract _ package ( version 3.107.2 ; * ? ? ?
* ) to measure net source counts .
events between 0.3 and 7.0 kev were extracted from a region corresponding to 90% of the psf at 1.5 kev ; for a few sources this region was reduced to avoid contamination by a close neighbor .
the background was determined from a source - free annulus centered on the source position .
we convert net count rates to fluxes within sherpa with arf and rmf response files appropriate for the chip location and source - extraction area of each source .
column 7 of table [ tab_src ] lists absorption - corrected ( @xmath35 ) fluxes @xmath28 computed under the assumption that the underlying spectrum is an optically thin plasma ( described by the _ xsmekal _ model ) with @xmath36 kev . for source cx29 that lies near
the aimpoint 1 count s@xmath1 corresponds to @xmath37 erg @xmath8 s@xmath1 .
this temperature is appropriate for the most active abs in the cluster ( see e.g. @xcite ) , while too high for the least active abs and too low for most cvs . for a 1-kev mekal model ,
a power - law spectrum ( _ powlaw1d _ ) with photon index @xmath38 , and a 10-kev thermal - bremsstrahlung model ( _ xsbremss _ ) the conversion factor is 24% smaller , and 30% and 42% larger , respectively . in each case
we account for a column density equal to the cluster value using the _ xsphabs _ model .
for the adopted distance to ngc6791 of 4.1 kpc , a 3-count detection limit corresponds to a limiting unabsorbed x - ray luminosity @xmath39 of @xmath40 erg s@xmath1 @xmath8 ( 0.37 kev ) where the range corresponds to the choice of models specified above . _
_ performs a kolmogorov - smirnov test on the event arrival times to test for variability .
two sources are thus found to be potentially variable , with probabilities that their count rates are constant ( @xmath41 ) smaller than 2.5% : cx19 ( a candidate cv , see sect .
[ sec_cv ] ) for which 15 of a total of 19 counts are detected within 5.5 hours ( @xmath42 ) , and cx37 ( no optical counterpart found ) for which all 9 counts are detected in the first 7 hours of the observation ( @xmath43 ) .
only cx1 has sufficient counts to allow a constraining spectral fit .
a spectrum was extracted from the _
acis_extract _ source region with the ciao tool _
psextract _ , and was grouped to have at least 20 counts bin@xmath1 to warrant use of the @xmath44 statistic ; the background contribution ( @xmath451 count ) can be ignored .
given the classification of cx1 as an active galactic nucleus ( agn ) based on its optical spectrum ( sect .
[ sec_galaxies ] ) , we try to fit the data with an absorbed power - law and find an acceptable result for a photon index @xmath46 and a column density @xmath47 @xmath8 ( 1-@xmath48 upper limit ; @xmath49 , 6 degrees of freedom ) .
the limit on @xmath50 is consistent with the integrated galactic column density in the direction of ngc6791 ( @xmath51 @xmath8 , * ? ? ?
* ) while the value for @xmath52 is typical for an agn ( e.g. * ? ? ?
the remaining sources have too few counts for useful constraints by spectral fitting .
instead we use the method of quantile analysis developed by @xcite , where the median energy ( @xmath29 ) and the 25% and 75% energy quartiles ( @xmath53 and @xmath54 ) of the source photons are used to characterize spectral properties .
the advantage of using energy quantiles as opposed to comparing counts in pre - defined energy bands by means of hardness ratios is that the errors on the diagnostics are less sensitive to the underlying spectral shape we looked for optical counterparts in the deep @xmath55 photometric catalog of ngc6791 compiled by stetson et al .
( 2003 ; s03 hereafter ) , which covers the entire area studied here .
the limiting magnitude is @xmath56 for the central area but the sensitivity is lower for the outer regions . in an attempt to eliminate artifacts from the catalog , we removed entries with photometric - quality indicators that flag them as suspicious , viz .
sources with @xmath57sharpness@xmath57 @xmath58 , and separation index @xmath59 ; see s03 for an explanation of these indicators and a motivation for these criteria .
we first measure and correct for the boresight , i.e. a possible systematic offset between the astrometric reference frames of the _ chandra _ and optical positions .
the optical positions are tied to the international celestial reference system ( icrs ) with an rms accuracy of about 027 ( s03 ) , but the astrometric reference frame of any given _
chandra _ observation as a whole is aligned to the icrs with a typical 06 accuracy ( 90% uncertainty ) .
we measure the boresight correction using matches with optical variables only .
this minimizes the number of chance alignments as the light curves of many variables reveal them as close binaries and therefore as plausible x - ray emitters .
a list of variables was compiled from the studies of @xcite , @xcite , @xcite , @xcite , @xcite , and @xcite ; we only selected variables that we could identify with objects in the s03 catalog so that we could use the s03 positions for the boresighting .
the boresight correction is computed iteratively .
we looked for matches to x - ray sources with more than 10 net counts and @xmath60 .
for each source , the match radius is set to be the quadratic sum of the error on its x - ray position , and the typical 1-@xmath48 error on the optical positions ( 027 ) scaled to a 95% error radius assuming a 2d gaussian error distribution .
the boresight is set to be the weighted ( with @xmath61 ) mean of the x - ray
optical offsets of all matches found .
next , the x - ray positions are corrected for the boresight , and the matching is repeated , now including the statistical error on the mean boresight in the match radius .
this procedure quickly converges to a boresight correction of @xmath62 , @xmath63 ( _ chandra _ minus optical ) based on ten matches with optical variables . matching the boresight - corrected x - ray catalog to the clean s03 catalog results in 51 optical matches ( including 26 variables ) for 47 sources , out of the 86 x - ray sources in total .
the probability to find a match in the s03 catalog by chance depends on the projected star density ( which decreases with distance from the cluster center @xmath64 ) and on the match radius ( which typically increases with @xmath64 because @xmath27 increases ) . to estimate the expected number of spurious matches
, we divide the cluster in a central area defined by @xmath65 , and an outer area of @xmath66 .
the clean s03 catalog gives a projected density of 0.030 optical sources arcsec@xmath67 and 0.0092 sources arcsec@xmath67 for the central and outer area , respectively . if we multiply this by the total area covered by the match circles of the x - ray sources , we find that the expected number of random matches is 5.9 ( center ) , and 3.8 ( outer annulus )
. this amounts to 15% and 29% of the matches found in the respective regions . on the other hand
, a similar calculation shows that all matches with variable stars are likely to be real , with the average number of chance alignments @xmath68 both for the central and outer regions . in order not to overlook any matches
, we repeated the matching using the entire s03 catalog , and inspected the regions around the _ chandra _ sources in the optical fits image from s03 to discard matches with image artifacts or dubious detections .
we thus found nine extra candidate counterparts for eight _ chandra _ sources , which are not included in the clean catalog because the values of the quality flags slightly exceeded the adopted cutoff limits , the object is really extended , or because the object is faint and lies close to a relatively bright star . since selection of these additional counterparts
was not done in a very systematic way but relies on visual inspection , it is not trivial to estimate the number of random coincidences among the new matches .
two new candidate counterparts are matched to a single _ chandra _ source outside @xmath17 , so at least one of the two must be spurious .
the other seven are uniquely matched to seven _
sources inside @xmath17 . based on the higher source density in the entire s03 catalog compared to the clean s03 catalog
, one would expect to find @xmath31.5 extra spurious matches in the central area . keeping in mind that some of these optical `` sources '' are not real
, we estimate that at most one or two of the seven new matches inside @xmath17 are spurious . in total
the s03 optical survey provides 60 astrometric matches for 55 sources .
@xcite used _
hubble space telescope _
( _ hst _ ) multi - epoch imaging with the wide field channel on the advanced camera for surveys ( acs ) to measure proper motions of objects in a central 34 @xmath69 34 region of ngc6791 .
we make use of their results to establish cluster membership of candidate counterparts included in the acs images ( sect .
[ sec_membership ] ) , and refer to @xcite for details of the data and the proper - motion analysis .
we also use these deep images to look for additional faint candidate counterparts . the data set consists of f606w and f814w images taken on 2003 july 16 and 17 ( go-9815 ) , and on 2005 jul 13 ( go-10471 ) .
an astrometric reference frame was created by combining the go-9815 f814w images with the stsci _ multidrizzle _ software , which removes artifacts like cosmic rays and bad pixels , and corrects for the geometric distortion of acs images . in order to align the coordinate system to the icrs , we computed an astrometric solution based on the positions of 273 unsaturated stars from the s03 catalog that could be identified in the stacked image .
fitting for zero point , plate scale , and rotation resulted in a solution where the rms residuals of individual stars are 0035 in right ascension and 0032 in declination ; this is negligible compared to the x - ray positional errors .
the boresighted x - ray positions and error circles were overlayed on this image , and photometry and proper motions for the astrometric matches were extracted from the source catalog created from the entire _ hst _ data set .
we find seven new candidate counterparts for six x - ray sources .
the new matches are all likely extra - galactic , given their proper motion or extended morphology .
from the fact that two of the seven _ hst _
candidate counterparts are matched to _
sources with another candidate counterpart , we estimate that an appreciable fraction of _ hst _ matches could be random , on the order of 2/7 ( @xmath929% ) or even more . as the astrometrically - calibrated image is not readily available , we show the finding charts of these additional identifications in fig .
[ fig_hstid ] .
this brings the final tally of the optical identification to 53 candidate counterparts for 48 of the 57 x - ray sources inside @xmath17 , and 14 candidate counterparts for 12 of the 29 sources between @xmath17 and 8 .
the properties of the candidate counterparts are summarized in tables [ tab_mem ] and [ tab_doubles ] , and their locations in the optical color - magnitude diagrams ( cmds ) are shown in figures [ fig_cmd_mem ] and [ fig_cmd_nmem ] . more information on the light - curve properties of the optical variables ( column 10 of table [ tab_mem ] )
can be found by consulting the original references .
variables with names listed in the format _ nnnnn _ _ were first discovered by @xcite ; table 1 in @xcite gives an overview of the discovery papers for variables with names starting with a v or b , although updated light curves can sometimes be found in more recent papers mentioned in that table .
34 in size , made from the stacked go-9815 f814w image representing 7024 s of exposure time in total . the combined x - ray / optical 95% error circles are shown as circles centered around the boresighted x - ray positions , while tick marks indicate the optical matches . except for the bright candidate counterpart to cx21 on the edge of the error circle
, these matches are found in the _ hst _ data only , and not in the @xcite catalog .
north is up , east to the left .
[ fig_hstid],width=313 ] rrccccccccclll ( 1 ) & ( 2 ) & ( 3 ) & ( 4 ) & ( 5 ) & ( 6 ) & ( 7 ) & ( 8) & ( 9 ) & ( 10 ) & ( 11 ) & ( 12 ) & ( 13 ) & ( 14 ) + cx & oid & @xmath70 & @xmath71 & @xmath26 & @xmath72 & @xmath73 & & @xmath74 & var & @xmath75 & optical spectrum & class & comments + & & ( ) & & & & & & & & ( days ) & & & + + + 3 & 2893 & 0.23 & & 20.64 & @xmath310.82 & 0.63 & 52.3 & @xmath310.06 @xmath30 0.04 & b8 & & & cv & + 4 & 9315 & 0.09 & m & 22.70 & 0.24 & 0.88 & 34.3 & 0.58 @xmath30 0.04 & 06289_9 & & emission lines & cv & + 9 & 10050 & 0.34 & 99 & 16.51 & 1.21 & 1.22 & 12.0 & @xmath312.35 @xmath30 0.08 & & & early - k giant & rg & + 15 & 6371 & 0.16 & 99 & 17.27 & 1.15 & 1.41 & 7.7 & @xmath312.24 @xmath30 0.09 & v9 & 3.187 & early / mid k , filled - in h@xmath10 & ssg & eclipsing binary + 17 & 12652 & 0.26 & 99 & 17.86 & 0.27 & 0.26 & 6.4 & @xmath312.1 @xmath30 0.1 & b7 & & balmer abs .
lines with ems .
cores & cv & + 22 & 12695 & 0.75 & 99 & 17.80 & 0.94 & 1.02 & 5.4 & @xmath312.2 @xmath30 0.1 & v16 & 2.266 & & ab & eclipsing or spotted binary + 23 & 11111 & 0.21 & 98 & 17.13 & 0.99 & 1.07 & 4.6 & @xmath312.5 @xmath30 0.1 & v100 & 12.522 & late g / early k & ab & spotted variable ? + & & & & & & & & & & or 23.947 & & & + 30 & 3626 & 0.31 & 99 & 17.96 & 1.15 & 1.29 & 4.8 & @xmath312.8 @xmath30 0.2 & v17 & 6.366 & late g , filled - in h@xmath10 ? & ssg & + 33 & 10611 & 1.30 & 99 & 19.57 & 1.04 & 1.20 & 3.3 & @xmath311.7 @xmath30 0.2 & v42 & 0.506 & & ab & spotted binary + 39 & 5883 & 0.38 & 99 & 17.20 & 0.89 & 0.94 & 2.6 & @xmath312.8 @xmath30 0.2 & v5 & 0.313 & mid g & ab & contact binary + 41 & 7011 & 0.23 & 99 & 18.27 & 1.06 & 1.12 & 2.3 & @xmath312.4 @xmath30 0.2 & v76 & 4.092 & late g , early k & ssg & + 44 & 12390 & 0.62 & 99 & 17.89 & 0.93 & 1.01 & 2.2 & @xmath312.5 @xmath30 0.2 & v80 & 4.886 & & ab & eclipsing binary + 50 & 8001 & 0.28 & m & 22.72 & 1.31 & 2.48 & 1.7 & @xmath310.7 @xmath30 0.2 & & & late k / early m ; h@xmath10 emission & ab & on binary main sequence + 54 & 11278 & 0.39 & 96,m & 19.91 & 1.07 & 1.35 & 1.4 & @xmath311.9 @xmath30 0.3 & & & k ; filled - in h@xmath10 & ab ? & on binary @xmath73 main seq .
+ 57 & 7397 & 0.20 & 77 & 19.27 & 1.15 & 1.32 & 1.1 & @xmath312.3 @xmath30 0.3 & v41 & 0.482 & & ab & eclipsing or spotted binary + 68 & 15561 & 0.92 & 99,m & 17.65 & & 1.53 & 12.7 & @xmath311.9 @xmath30 0.1 & 01431_10 & 7.640 & & ssg & spotted binary + 77 & 746 & 1.20 & 99,m & 17.96 & 1.35 & 1.39 & 4.5 & @xmath312.2 @xmath30 0.2 & v59 & 13.833 & mid k , weak h@xmath10 emission & ssg & + 79 & 13737 & 0.21 & 90 & 21.06 & 1.45 & 1.93 & 2.7 & @xmath311.2 @xmath30 0.2 & & & & ab & on binary main sequence + 81 & 1185 & 0.79 & 99 & 17.74 & 0.97 & 0.96 & 3.5 & @xmath312.4 @xmath30 0.2 & v7 & 0.488 & early / mid g & ab & eclipsing binary + 86 & 4773 & 1.93 & 99 & 17.55 & 1.04 & 1.03 & 3.5 & @xmath312.5 @xmath30 0.3 & v12 & 1.523 & mid g & ab & eclipsing binary + + + + 18 & 6697 & 0.13 & & 23.99 & @xmath312.4 : & 1.45 & 5.9 & 0.3 @xmath30 0.1 & & & & fb & + 19 & 12116 & 0.40 & & 23.71 & 0.21 & 1.30 & 5.9 & 0.2 @xmath30 0.1 & & & balmer and he ii emission & cv ? & + 25 & 5627 & 0.24 & & 24.09 & 0.4 : & 2.12 & 3.8 & 0.2 @xmath30 0.1 & & & & fb & + 28 & 10837 & 0.56 & & 22.71 & & 0.38 & 3.8 & @xmath310.4 @xmath30 0.2 & & & & fb & + 36 & 6751 & 0.90 & & 20.14 & 0.53 & 0.92 & 3.2 & @xmath311.5 @xmath30 0.2 & & & & fb & + 49 & 5910 & 0.23 & 45 & 20.82 & 1.47 & 1.79 & 1.8 & @xmath311.5 @xmath30 0.2 & & & mid k , h@xmath10 emission & ab ? & on binary main sequence + 56 & 4436 & 0.07 & & 20.83 & 1.29 & 1.69 & 1.1 & @xmath311.7 @xmath30 0.3 & 01724_9 & 1.613 & & ab & spotted variable + 58 & 5913 & 0.35 & & 18.38 & 0.93 & 0.93 & 1.1 & @xmath312.6 @xmath30 0.3 & & & & ab ? & on main sequence + 65 & 14602 & 1.11 & & 23.05 & @xmath310.4 : & 0.54 & 9.0 & 0.1 @xmath30 0.1 & 00670_10 & & & fb & long - term variable + 72 & 13270 & 1.31 & & 22.9 : & & 0.5 : & 5.4 & @xmath310.1 : & & & & fb &
+ + rrccccccccclll ( 1 ) & ( 2 ) & ( 3 ) & ( 4 ) & ( 5 ) & ( 6 ) & ( 7 ) & ( 8) & ( 9 ) & ( 10 ) & ( 11 ) & ( 12 ) & ( 13 ) & ( 14 ) + cx & oid & @xmath70 & @xmath71 & @xmath26 & @xmath72 & @xmath73 & & @xmath74 & var & @xmath75 & optical spectrum & class & comments + & & ( ) & & & & & & & & ( days ) & & & + + + 1 & 5361 & 0.19 & 59 & 20.36 & 0.65 & 0.76 & & @xmath310.09 @xmath30 0.03 & & & agn , @xmath76 & eg & + 2 & 3886 & 0.14 & 0 & 16.22 & 1.09 & 1.35 & & @xmath311.86 @xmath30 0.04 & v33 & 1.173 & mid k , h@xmath10 emission & ab & eclipsing binary + 7 & 7878 & 0.17 & 0 & 15.39 & 1.21 & 1.49 &
& @xmath312.68 @xmath30 0.06 & v19 & & late - g , early - k foreground dwarf & s & irregular variable + 8 & 3472 & 0.17 & & 20.22 & 0.24 & 0.69 & & @xmath310.30 @xmath30 0.07 & & & agn , @xmath77 & eg & + 12 & 12901 & 0.92 & 0 & 16.06 & 1.23 & 1.38 & & @xmath312.65 @xmath30 0.09 & v66 & 50.498 & early - k foreground dwarf & s & long - period variable + 13 & & 0.25 & & & & & & @xmath782.1 & & & & eg & extended + 21 & & 0.08 & nm & 23.96 & & 1.12 & & 0.2 @xmath30 0.1 & & & & eg ?
& proper motion consistent + & & & & & & & & & & & & & with background galaxy + 24 & 6271 & 0.29 & 27 & 16.32 & 0.89 & 1.01 & &
@xmath312.9 @xmath30 0.1 & v1 & 0.268 & & ab & foreground contact binary + 26 & & 0.22 & & 28.5 & & 1.5 & & 1.9 @xmath30 0.2 & & & & eg & extended + 27 & 11376 & 0.04 & 0 & 15.97 & 0.66 & 0.81 & & @xmath313.0 @xmath30 0.1 & v6 & 0.279 & & ab & foreground contact binary + 29 & 7328 & 0.18 & 0 & 13.60 & 0.87 & 1.26 & & @xmath314.0 @xmath30 0.2 & & & mid k & s & + 31 & 11212 & 0.25 & & 22.27 & 0.51 & 0.75 & & @xmath310.6 @xmath30 0.2 & & & agn ,
@xmath79 & eg & + 34 & & 0.33 & & & & & & @xmath781.6 & & & & eg & extended + 35 & 4571 & 0.28 & 0 & 14.62 & 1.38 & 1.35 & & @xmath313.7 @xmath30 0.2 & & & mid - k giant & s & + 38 & 7534 & 0.12 & 11,nm & 21.04 & 1.57 & 2.29 & & @xmath311.2 @xmath30 0.2 & & & & s & + 42 & & 0.15 & & 26.6 & & 0.7 & & 1.0 @xmath30 0.3 & & & & eg & extended + 43 & 3629 & 0.78 & 98 & 22.11 & 1.75 & 2.94 & & @xmath310.6 @xmath30 0.2 & & & & s & too red to be a member + 45 & 4353 & 0.37 & 0 & 18.98 & 1.45 & 1.57 & & @xmath311.7 @xmath30 0.2 & & & @xmath80 & eg & extended + 46 & 10088 & 0.21 & 0 & 11.54 & 1.10 & 1.10 & & @xmath315.1 @xmath30 0.2 & & & mid - g foreground giant & s & + 47 & 10967 & 0.32 & nm & 23.88 & @xmath310.2 & 2.09 & & @xmath310.2 @xmath30 0.2 & & & & eg & extended + 48 & 2426 & 0.66 & 0 & 19.53 & 1.04 & 1.15 & & @xmath311.8 @xmath30 0.2 & v11 & 0.883 & early g & ab & eclipsing binary + 51 & 12037 & 0.41 & 0 & 14.01 & 0.71 & 0.74 & &
@xmath314.2 @xmath30 0.2 & & & late - f / early - g dwarf & s & + 52 & & 0.18 & & 27.68 & & 2.32 & & 1.2 @xmath30 0.3 & & & & eg & extended + 53 & 8705 & 0.20 & nm & 23.7 & 0.3 & 1.3 & & @xmath310.4 @xmath30 0.3 & & & & eg & extended + 55 & & 0.80 & & 18.76 & & & & @xmath312.4@xmath30 0.3 & & & & eg ? & extended + 76 & 14833 & 2.37 & 0 & 16.48 & 1.00 & 1.05 & & @xmath312.9 @xmath30 0.2 & v54 & 8.314 & mid - g , early - k foreground dwarf & ab & spotted variable ? + 78 & 553 & 2.58 & 0 & 16.85 & 1.03 & 1.10 & & @xmath312.7 @xmath30 0.2 & v23 & 0.272 & & ab & foreground contact binary + 80 & 1031 & 2.05 & 0 & 20.36 & 0.85 & 0.94 & & @xmath311.3 @xmath30 0.2 & & & & s & + 82 & 13404 & 0.03 & 0 & 14.47 & 0.87 & 0.89 & & @xmath313.9 @xmath30 0.2 & & & early - k foreground dwarf & s & + 84 & 15434 & 1.49 & 0 & 21.2 : & & 2.9 : & & @xmath311.2 @xmath30 0.3 & 01225_10 & & & s & long - term variable + @xmath81this source has another candidate counterpart , see table [ tab_doubles ] . + @xmath82the counterpart is not included in the s03 catalog .
the listed value is the magnitude in the @xmath83 band @xcite .
+ and @xmath84 is not available , are not shown .
see the electronic edition of the journal for a color version of this figure .
[ fig_cmd_nmem ] ] . for seven x - ray sources
we find two candidate counterparts . in four cases
cx27 , cx57 , cx78 , cx84one of the two is an optical variable , and given the low probability for a chance coincidence we consider the variables to be the true matches .
the variables all lie closer to the x - ray source than the alternative match .
we also tentatively list the closest match to the other three x - ray sources as the likely counterpart in table [ tab_mem ] , but more information is needed to firmly establish which , if any , of the two is the true match .
for completeness we list the properties of the alternative identifications in table [ tab_doubles ] .
cx34 is likely an extra - galactic source , as both _
counterparts are extended . the closest match to cx21
is a faint point source that is unrelated to the cluster ; its _ hst _ proper motion is typical for a background galaxy .
the other match is a proper - motion member on the main sequence ( see also fig .
[ fig_hstid ] ) .
cx82 matches with two bright ( @xmath85 ) stars .
the nearest is a proper - motion non - member and a likely foreground k dwarf .
.alternative optical counterparts[tab_doubles ] [ cols="^,>,^,^,^,^ , < " , ] while almost 80 globular clusters have been studied with _
chandra _ , only a handful have been observed deep enough to uncover a significant fraction of the abs .
here we focus on two globular clusters for which the _ chandra _ sensitivity reaches well below the luminosity cutoff adopted in the previous section .
the nearby core - collapsed cluster ngc6397 was observed down to a limiting luminosity of @xmath3@xmath86 erg s@xmath1 @xcite .
we adopt a cluster mass of @xmath87 @xmath88 @xcite .
@xcite classified most of the 79 _ chandra _ sources inside @xmath89 233 from @xcite using _ hst _ data , and found numerous cvs and abs , and two sub - subgiants ( at least one is in a binary with a msp ) .
about 20% of the sources are brighter than our cutoff value ; these are mostly cvs or neutron - star binaries . while there are plenty of x - ray faint abs ,
there are just a few at most above @xmath2 erg s@xmath1 ; the range listed in table [ tab_oc ] is zero to two , but those two systems have very red colors and are unlikely cluster members .
47tuc is a massive cluster ( @xmath90 @xmath88 ; @xcite ) .
we estimate that within @xmath89 279 , about 28 of the 232 sources bright enough to be included in our selection are background agn , leaving @xmath3200 sources associated with the cluster . as of yet , only part of these have been classified @xcite .
we make a very conservative estimate of the number of cvs and abs . a lower limit
is derived by counting the number of ( candidate ) cvs and abs in table 2 of @xcite , while an upper limit ( given in square brackets in table [ tab_oc ] ) is estimated by adding all unidentified and unclassified sources to either the cvs or abs and subtracting the estimated contribution from agn .
sub - subgiants are more conspicious , so their number is less uncertain . to compute the cluster x - ray luminosity
, we sum the luminosities for the @xmath91 individual sources , and subtract an approximate contribution for the background agn based on the mean luminosity of the unidentified sources . for both clusters we converted the fluxes listed in @xcite and @xcite to the 0.37 kev band using pimms , adopting a 2-kev plasma model .
table [ tab_oc ] readily shows that the total number of abs above @xmath92 erg s@xmath1 , normalized by cluster mass , is much lower in the two globular clusters than in m67 and ngc6791 . for 47tuc
the difference is a factor of 65230 compared to m67 , whereas ngc6397 likely does not have any abs at all above this luminosity .
lowering the x - ray luminosity cutoff value would increase the numbers of abs especially for ngc6397 , but this is not enough to make up for the lower specific x - ray luminosity .
the poor statistics in the open clusters left aside , cvs appear underrepresented in globular clusters as well , by about an order of magnitude .
a likely explanation is that dynamical encounters have a net effect of destroying abs_both _ those with evolved components and wider orbits _ and _ main - sequence systems and cvs . at the same time
, the correlation with encounter rate found by @xcite implies that at least cvs are formed as well ; perhaps this is why the specific frequencies in open and globular clusters are less disparate for cvs than for abs .
ab - like sub - subgiants are underabundant in globular clusters as well , but on the other hand sub - subgiants in ( candidate ) binary msps are not found in open clusters ( note that in table [ tab_oc ] the different types of sub - subgiants are grouped together because some have not been studied well enough to make the distinction ) .
the relative lack of abs in ngc6397 with respect to other globular clusters was already noted by @xcite and @xcite , who ascribed it to its post - collapsed state .
we now show that 47tuc is deficient in abs , too .
m4 , studied by @xcite and not included in table [ tab_oc ] , shows the same effect . even if all fifteen sources or so that are brighter than @xmath2 erg s@xmath1 were abs ( which is not the case as there are also cvs , background and foreground sources , and an msp among them ) , it is not enough to reach the number of abs expected based on scaling by mass from open clusters . @xcite find a lower overall binary fraction in globulars compared to open clusters @xcite .
if this is the result of binary destruction , the x - ray emitting , and thus hardest , binaries are affected as well .
we can not exclude that other factors are important .
there is a trend of decreasing ab frequency with age for all clusters in table [ tab_oc ] except ngc6819 . in @xcite
it was first pointed out that the total _
luminosity per unit mass of most globular clusters is lower than that of m67 . while m67 may be exceptional in some ways ( see sect .
[ disc_oc ] ) , the last column of table [ tab_oc ] shows that the specific x - ray luminosity of globular clusters may be lower than for open clusters in general .
cvs , abs , and sub - subgiants are all responsible for this .
the x - ray emission from qlmxbs and msps in globular clusters , of which there are no confirmed cases in old open clusters , can not make up for the lack of other types of faint sources .
a more extensive investigation , based on a larger sample of globular clusters and old populations , is done by heinke et al .
( in preparation ) to investigate the effects of age , central density , and metallicity on differences in the x - ray emissivity .
we have studied ngc6791 , which is one of the oldest open clusters known , in x - rays for the first time . of the twenty _ chandra _ sources that are identified with proper - motion cluster members , nineteen are binaries .
we find a mix of cvs , abs , and sub - subgiant binaries . with optical follow - up spectroscopy
we confirm the classification of one cv in the cluster , and identify one new candidate cv .
the origin of these and the other two cluster cvs is likely primordial .
a comparison with the x - ray sources in the younger cluster ngc6819 suggests that in order to find numerous ( main - sequence ) abs brighter than @xmath93 erg s@xmath1 in open clusters , one must look at those older than @xmath32.5 gyr .
the mass - specific frequency of main - sequence and sub - giant abs in ngc6791 ( 8 gyr ) is 37 times lower than in m67 ( 4 gyr ) ; it is not clear why this is the case .
comparison with a small number of globular clusters shows that all three source classes that are mainly responsible for the x - ray emission from old open clusters are under - represented in the globulars .
this accounts for the lower total x - ray luminosity per unit mass of globular clusters , and indicates that the net effect of dynamical encounters may be destruction of even the hardest binaries .
the cutoff luminosity for comparing source populations among clusters was set at @xmath94 erg s@xmath1 ( 0.37 kev ) .
a lower cutoff luminosity might be needed to account for effects of variations of the ab x - ray luminosity function with age or metallicity .
while @xcite note that population - ii abs may have lower x - ray luminosities , it remains unclear in this context why ngc6791 , which has a super - solar metallicity , should also have relatively fewer abs above our limiting @xmath95 .
we defer a more thorough investigation to a future study .
deeper data are needed for ngc6791 to match the already deep existing data for m67 , 47tuc , and ngc6397 . the x - ray
derived ab populations may then be compared to those derived from other wavelengths , e.g. optical variability studies . also in this respect , studies of ngc6791 may benefit from being in the field of view of the _ kepler _ satellite .
the authors would like to thank j. hong for providing the routines to compute energy quantiles , c. heinke for comments on the manuscript , and k. cudworth for providing a catalog of proper motions .
this research was supported by _
grant go0 - 11110x .
some of the observations reported here were obtained at the mmt observatory , a joint facility of the smithsonian institution and the university of arizona .
, c. o. , ruiter , a. j. , muno , m. p. , & belczynski , k. 2008 , in american institute of physics conference series , vol .
1010 , a population explosion : the nature & evolution of x - ray binaries in diverse environments , ed .
r. m. bandyopadhyay , s. wachter , d. gelino , & c. r. gelino , 136142 , r. d. 2000 , in astronomical society of the pacific conference series , vol .
198 , stellar clusters and associations : convection , rotation , and dynamos , ed .
r. pallavicini , g. micela , & s. sciortino , 517+ | we present the first x - ray study of ngc6791 , one of the oldest open clusters known ( 8 gyr ) .
our _ chandra _ observation is aimed at uncovering the population of close interacting binaries down to @xmath0 erg s@xmath1 ( 0.37 kev ) .
we detect 86 sources within 8 of the cluster center , including 59 inside the half - mass radius .
we identify twenty sources with proper - motion cluster members , which are a mix of cataclysmic variables ( cvs ) , active binaries ( abs ) , and binaries containing sub - subgiants . with follow - up optical spectroscopy
we confirm the nature of one cv .
we discover one new , x - ray variable candidate cv with balmer and he ii emission lines in its optical spectrum ; this is the first x - ray selected cv confirmed in an open cluster .
the number of cvs per unit mass is consistent with the field , suggesting that the 34 cvs observed in ngc6791 are primordial .
we compare the x - ray properties of ngc6791 with those of a few old open ( ngc6819 , m67 ) and globular clusters ( 47tuc , ngc6397 ) .
it is puzzling that the number of abs brighter than @xmath2 erg s@xmath1 normalized by cluster mass is lower in ngc6791 than in m67 by a factor @xmath337 .
cvs , abs , and sub - subgiants brighter than @xmath2 erg s@xmath1 are under - represented per unit mass in the globular clusters compared to the oldest open clusters , and this accounts for the lower total x - ray luminosity per unit mass of the former .
this indicates that the net effect of dynamical encounters may be the destruction of even some of the hardest ( i.e. x - ray emitting ) binaries . |
ngc 1672 is a @xmath6 barred spiral galaxy in the southern hemisphere that is viewed nearly face on .
it shows dramatic nuclear and extranuclear star - formation activity including starburst regions near the ends of its strong bar ( e.g. , pastoriza 1967 ; osmer , smith & weedman 1974 ; baumgart & peterson 1986 ) .
it may have a seyfert 2 nucleus ( e.g. , vron , vron & zuiderwijk 1981 ; brandt , halpern & iwasawa 1996 , hereafter bhi96 ) , but at present this has not been firmly established .
ngc 1672 is at a distance of 16.3 mpc ( @xmath7 km s@xmath2 mpc@xmath2 ) , and the galactic neutral hydrogen column density in its direction is @xmath8 @xmath5 ( heiles & cleary 1979 ) .
x - ray emission is powerful for probing both the endpoints of stellar evolution and seyfert activity , and ngc 1672 is known to be a prodigious source of x - rays .
bhi96 presented a detailed position sensitive proportional counter ( pspc ) and high resolution imager ( hri ) study of ngc 1672 , and they found several sources of x - ray emission coincident with the galaxy .
the nucleus showed the strongest emission .
this emission was slightly resolved , suggesting at least some starburst x - ray activity , but the data were unable to determine whether the bulk of the emission was from a starburst region or hidden seyfert .
awaki & koyama ( 1993 ) claimed a hard x - ray detection of a @xmath9 erg s@xmath2 seyfert nucleus with the large area counter ( lac ) , but concerns about source confusion were present ( the lac was a non - imaging detector and had a @xmath10 field of view ) .
in addition to the nuclear x - ray emission , bhi96 also found luminous x - ray sources coincident with the starburst regions near the ends of the galactic bar .
it was difficult to determine the nature of these sources due to limited photon statistics .
they could be consistently modeled as superbubbles , collections of x - ray binaries , or ` super - eddington sources ' ( e.g. , fabbiano 1998 ) . to investigate the x - ray properties of ngc 1672 further
, we performed a second hri observation of it in 1997 .
our goals for this observation were ( 1 ) to study the long - term x - ray variability properties of the currently known sources , ( 2 ) to search for new x - ray sources that were not detected before due to their variability , and ( 3 ) to obtain better photon statistics for investigation of any weak point sources or large - scale diffuse x - ray emission . here
we present the results from our second hri observation as well as an improved analysis of the original 1992 hri observation .
we also use data from 1995 to constrain the hard x - ray properties of ngc 1672 .
ngc 1672 was observed with the hri ( e.g. , david et al .
1999 ) during the period 27 january 1997 to 2 march 1997 ( 35.0 ks exposure time ) .
the observation went smoothly , and reduction and analysis of the data were performed with the asterix software system ( allan & vallance 1995 ) .
we created three images for subsequent analysis : ( 1 ) an image from the 1997 hri observation , ( 2 ) an image from the original 1992 hri observation ( 24.5 ks ; see bhi96 ) , and ( 3 ) a ` combined ' image using the data from both the 1992 and 1997 observations ( 59.5 ks of total exposure time ) .
all images were made with @xmath11 per pixel resolution so that the hri point spread function ( psf ) was oversampled by a factor of @xmath12 .
we also used only hri channels 38 , since this channel range provides maximum sensitivity for detecting point sources and low - surface - brightness diffuse emission ( see 3.4 of david et al .
1999 ) .
the three images described in 2.1 were analyzed using the point source searching software pss ( allan 1995 ) .
locates point sources by convolving an hri image with a position - dependent model for the hri psf , and it calculates point source significances using the cash ( 1979 ) statistic .
we used constant backgrounds for each of the three images , and our manual background analyses showed that this is valid in the region around ngc 1672 .
we consider for further study point sources that are within @xmath13 of the center of ngc 1672 ( this corresponds to the main optical extent of the galaxy ) and are found to be significant at @xmath14 in at least one of the three images . given these criteria ,
we expect @xmath15 false sources in total .
also creates a ` significance map ' showing the hri image after it has been convolved with the psf model ; we show the significance map for the 1997 observation in figure 1 . in table
1 we give ( background - subtracted ) source counts , mean count rates , fluxes , and luminosities for all sources .
we have adopted the source naming convention of bhi96 but have also added sources x-7 , x-8 and x-9 .
source counts were calculated using circular regions chosen to fit the morphology of each source . when sources were not detected in one of the images , we calculated @xmath16 upper limits .
we used the ( mukai 1997 ) software package to convert the mean count rates for each source into 0.22.0 kev fluxes and luminosities . for x-1
, we adopted the raymond - smith thermal plasma model of bhi96 to perform this conversion .
for all other sources , we adopted a power - law model with a photon index of @xmath17 and the galactic column density .
the 1997 observation has revealed a new probable x - ray source ( x-8 ) within the optical extent of ngc 1672 .
this source lies near the edge of one of the prominent spiral arms .
it is detected at the @xmath18 level , so there is a small but non - negligible chance that it is due to a statistical fluctuation .
in addition , our analyses of all three images show that the extended source x-3 of bhi96 actually appears to be two separate sources ( labeled here as x-3 and x-7 ; see figures 1 and 2 ) .
this difference between bhi96 and our analysis is understandable because our hri data have been through a revised processing that includes better filtering as well as satellite aspect information from the wide field camera ( m. corcoran 1998 , private communication ) .
these factors improve the spatial resolution .
in addition , bhi96 used the full hri band while we only use channels 38 ( see 2.1 ) .
finally , searching the 59.5 ks combined image reveals another new point source , x-9 ( see figure 2 ) .
this x - ray source is spatially coincident with a bright region along one of the spiral arms that contains three large h ii regions ( numbers 31 , 34 and 39 in table 2 of evans et al .
the alignment of the x - ray and optical emission supports the reality of the x - ray detection . to check the x - ray spatial extent results of bhi96
, we have performed spatial extent analyses for x-1 and x-2 using both the 1992 and 1997 observations .
our results are in general agreement with those of bhi96 : x-2 does not show any clear evidence for spatial extent while x-1 is extended on about the scale of the nuclear ring .
x-1 is more extended than x-2 , and this strongly suggests that its apparent extent is not an artefact of aspect solution errors ( see morse 1994 ) .
we have compared the mean count rates between the 1992 and 1997 observations to search for long - term source variability.% during this time period ( see 2.3.3 of david et al .
] x-3 shows significant mean count rate variability with an @xmath0% increase between 1992 and 1997 . the poisson probability that this variability is merely due to a statistical fluctuation is @xmath19%
we have searched for rapid variability of x-3 within the 1997 data , and we do not find any highly significant variability .
the source x-5 , which is thought to be associated with a foreground star ( see bhi96 ) , also shows significant variability . , and it is thus at a distance of @xmath20 pc . the observed x - ray - to - optical flux ratio is consistent with that expected for a g5 star ( see figure 1 of maccacaro et al .
1988 ) , and the observed x - ray variability is not implausibly large for such a star . ] x-1 and x-2 do not show significant variability .
we place upper limits of @xmath21% and @xmath22% on the variability of x-1 and x-2 , respectively .
we have used the adaptive kernel smoothing algorithm asmooth ( ebeling , white & rangarajan 1999 ) to search for diffuse x - ray emission in the 59.5 ks combined image .
asmooth convolves a gaussian of variable , position - dependent size with an image to allow detection of spatial structure on a wide range of scales , and it provides a quantitative assessment of the significances of observed features .
we find evidence for an @xmath23 area of diffuse x - ray emission between x-1 and x-2 ( see figure 3 ) . we have examined if this emission can be explained as ` spill over ' from x-1 and x-2 , and the observed emission appears to significantly exceed that expected from these two sources ( even after taking into account reasonable aspect solution errors ; see morse 1994 ) .
note that similar emission is not seen between x-1 and x-3 even though these two sources are closer together than x-1 and x-2 ( @xmath24 versus @xmath25 separation ) .
asmooth indicates that the diffuse emission is significant at the @xmath26 level , and our manual calculations using rectangular regions give similar significances .
we estimate @xmath27 counts in total from the diffuse emission , so the count rate is @xmath28 count s@xmath2 .
this emission may be associated with hot , diffuse gas in the interstellar medium of ngc 1672 .
if we adopt a 0.55 kev raymond - smith model with solar abundances and the galactic column density , we estimate a total 0.22.0 kev flux of @xmath29 erg @xmath5 s@xmath2 and a 0.22.0 kev luminosity of @xmath30 erg s@xmath2 ( the flux and luminosity values depend fairly strongly upon the unknown abundances of the gas ) .
of course , we are not able to prove that this emission must be from truly diffuse hot gas . it could also plausibly arise from several weak point sources that are individually below our detection threshold .
ngc 1672 was observed with ( tanaka , inoue & holt 1994 ) on 2324 november 1995 .
both solid - state imaging spectrometer ccd detectors ( sis0 and sis1 ) and both gas imaging spectrometer scintillation proportional counters ( gis2 and gis3 ) were operated .
the sis were operated in 1 ccd mode , and the gis were operated in ph mode .
we have used the revision 2 processed data from goddard space flight center for the analysis below , and we have adopted the standard revision 2 screening criteria ( see pier 1997 ) .
after data screening , the exposure times were 33 ks for sis0/sis1 and 40 ks for gis2/gis3 .
data reduction was performed using xselect ( ingham & arnaud 1998 ) .
the spatial resolution of is a significant limiting factor when studying ngc 1672 , so we will present only the main results of our imaging analysis .
the psfs of the x - ray telescopes have half - power diameters of @xmath31 ( compare with figures 13 and note that x-1 and x-3 are only separated by @xmath24 ) .
however , they also have fairly sharp cores that allow two point sources as close as @xmath32 to be resolved .
in the case of the gis , the final spatial resolution is worsened somewhat further by the detectors spatial responses , and the gis can not effectively study structure on scales smaller than @xmath33 . for sources observed in an approximately on - axis manner , the sis and gis positional uncertainties are @xmath34 and @xmath35 , respectively ( 90% confidence ; gotthelf 1996 and e. gotthelf 1999 , private communication ) .
we have created full ( 0.610 kev for sis and 1.010 kev for gis ) , soft ( 0.62.0 kev for sis and 1.03.0 kev for gis ) , hard ( 210 kev for sis and gis ) , and ultra - hard ( 510 kev for sis and gis ) images for each of the four detectors .
we also created summed sis0+sis1 and gis2+gis3 images in all bands .
these images will be used in the spatial analysis below .
sis images were made with @xmath36 per pixel resolution , and gis images were made with @xmath37 per pixel resolution .
we have used ximage ( giommi , angelini & white 1997 ) and idl to inspect the images created in 3.1 . at low energies our results are consistent with those expected based upon the data .
x-1 is observed to be the dominant x - ray source in ngc 1672 ( see figure 4a ) , and we also detect x-3 and probably x-2 in the sis images . we have compared the and data to examine source variability .
we do not find any strong evidence for variability , although source confusion ( see 3.1 ) significantly limits our variability analyses . in the harder 210 kev and 510 kev images
, however , ngc 1672 has a considerably different appearance ( see figure 4b ) .
the dominant 210 kev emission is observed to be coincident with the hri sources x-3 and x-7 ( in both the sis0+sis1 and gis2+gis3 images ) , and we detect only weak 210 kev emission from x-1 .
note that x-3 was also found to be the hardest source in the pspc band by bhi96 . in the 510 kev gis2+gis3 image ,
the only detectable emission from ngc 1672 is coincident with x-3 and x-7 .
ngc 1672 is not detected in the 510 kev band by the sis detectors , probably due to their smaller effective areas in this energy range ( see tanaka , inoue & holt 1994 ) . due to the limited spatial resolution of ( see 3.1 ) , we have only been able to perform reliable spectral analyses for ngc 1672 as a whole ( see 7.4 of _ the asca data reduction guide _ ) . at energies above @xmath38 35 kev
, x-3 is the dominant source , but at lower energies our spectra contain contributions from x-1 , x-2 and x-3 .
we have extracted spectra from each detector using all of the acceptable exposure time , and we have only used data where ngc 1672 is reliably detected ( 0.65 kev for the sis and 18 kev for the gis ) .
we have grouped our spectra so that there are 15 photons per spectral data point , to allow the use of @xmath39 fitting techniques .
we have first performed preliminary spectral fitting for each of the detectors separately , and although the errors were large we obtained consistent results .
we have therefore jointly fitted the spectra from all four detectors , and we detail these results below . aside from the absolute model normalizations , we tied together the fit parameters across the four detectors .
we use ( arnaud 1996 ) for all spectral fitting below .
we start by fitting the data in the 28 kev band , where x-3 is the dominant source .
we use a basic power - law model with the galactic absorption column .
this model provides an acceptable fit with @xmath40 for 56 degrees of freedom .
we obtain a power - law photon index of @xmath41 ( fit parameter errors are for @xmath42 ) , and there is no evidence for systematic positive residuals near the 6.46.97 kev iron k@xmath43 lines .
the 28 kev flux is @xmath44 erg @xmath5 s@xmath2 , corresponding to a 28 kev luminosity for ngc 1672 of @xmath45 erg s@xmath2 .
we have then applied this same model to the full 0.68 kev band .
we obtain a steeper power law with @xmath46 and @xmath47 for 153 degrees of freedom .
the fit is statistically acceptable , although we observe systematic positive residuals from 0.61.0 kev and systematic negative residuals from 1.11.6 kev .
these are understandable in the context of the spectral fitting presented in 2.3 of bhi96 ; the emission from ngc 1672 below @xmath12 kev appears to be largely from hot , diffuse gas that is associated with x-1 , x-2 and x-3 ( i.e. , it is _ not _ primarily the diffuse emission discussed in 2.3 ) .
if we take the best fitting 28 kev model and extrapolate it downward in energy , we see clear positive residuals that are likely to be associated with the thermal gas emission ( see figure 5 ) .
we note that such emission is commonly seen from galaxies with starburst activity ( e.g. , ptak et al .
1999 ) . we have added a raymond - smith component to our basic power - law model , and this provides a highly significant improvement in overall fit quality .
we obtain a plasma temperature of @xmath48 kev , a photon index of @xmath49 , and @xmath50 for 151 degrees of freedom ( @xmath51 compared to the basic power - law model ; fit parameter errors are for @xmath52 ) . the plasma temperature from our fit
is consistent with that of bhi96 , and the underlying photon index agrees better with that seen in the 28 kev band .
the derived plasma temperature is also in good agreement with the temperatures obtained for other starburst galaxies ( e.g. , figure 3 of ptak et al .
the plasma abundances are poorly constrained , and we have adopted solar abundances in the fit above . the derived plasma temperature , power - law photon index , and x - ray luminosity do not depend sensitively upon the abundance assumptions for abundance enhancements up to a factor of @xmath53 , and we note that storchi - bergmann et al .
( 1996 ) argue for moderately supersolar abundances in the nucleus of ngc 1672 . in reality
, we believe that the thermal emission below @xmath12 kev is likely to arise from gas with a range of temperatures and abundances , and our single raymond - smith component model should not be taken too literally ( see ptak et al .
1999 for further discussion on this matter ) .
comparison of the soft and hard x - ray luminosities of x-3 suggests that this source may suffer from significant intrinsic absorption ( see 4 for further discussion ) , and we have examined if our data constrain an absorption column affecting only the power - law spectral component . unfortunately , our data do not tightly constrain such absorption , and column densities ranging from 0 @xmath5 to @xmath54 @xmath5 are allowed by our data . due to significant - cross - calibration uncertainties ( e.g. , iwasawa ,
fabian & nandra 1999 ) as well as the time variability of x-3 , we have not performed detailed joint spectral fitting of the data from these two missions .
the 28 kev flux we measure from ngc 1672 is @xmath55 times smaller than the 210 kev flux of @xmath56 erg @xmath5 s@xmath2 given by awaki & koyama ( 1993 ) , and our image analyses in 3.2 show that most of the 210 kev flux originates outside the nucleus .
thus , the data do not provide evidence for a highly absorbed nuclear x - ray source as might be expected for a seyfert 2 .
however , the data do not rule out a seyfert 2 nucleus where the torus has a large enough column density to absorb almost all x - rays below 8 kev .
given the emission at other wavelengths ( see 3.1 of bhi96 ) , a nuclear x - ray source with a luminosity up to @xmath57 erg s@xmath2 appears plausible . we can estimate the required torus column density for a plausible scenario using our spectral fits above .
we have added an extra @xmath17 power law to our best 0.68 kev fit and assumed an unabsorbed 28 kev flux of @xmath56 erg @xmath5 s@xmath2 for this power law ( corresponding to a 28 kev luminosity of @xmath58 erg s@xmath2 ) . to represent absorption by the torus
, we added an extra absorption component acting only on this power law . to allow for consistency between data and model
, the fit requires the extra absorption component to have a column density of @xmath4 @xmath5 .
about half of the seyfert 2 population has torus column densities as large as this ( e.g. , risaliti , maiolino & salvati 1999 ) .
many seyfert 2s with large torus column densities show strong iron k@xmath43 emission lines . while we do not see evidence for iron k@xmath43 lines in our data , our constraints are not tight . for a narrow iron k@xmath43
line at 6.4 kev , our 90% confidence upper limit on the equivalent width is 750 ev .
we have presented new hri and data that constrain starburst and seyfert activity in ngc 1672 .
notably , we do not find any evidence for hard x - ray emission associated with the nuclear source x-1 despite the claims for seyfert activity made based upon optical emission - line , infrared emission - line , x - ray , and radio data ( see bhi96 for a review ) .
we also find no evidence for long - term soft x - ray variability of x-1 . if ngc 1672 indeed has a luminous seyfert 2 nucleus , it must have been heavily obscured during the observation by a ` compton - thick ' torus .
our x - ray spectral results stand in contrast to those of awaki & koyama ( 1993 ) , who claimed to detect luminous power - law emission from an obscured seyfert nucleus with @xmath59 @xmath5 .
while it is perhaps possible that the torus column density along our line of sight has increased by a factor of @xmath60 between the and observations , the simplest explanation for the discrepant results is that the data suffered from source confusion .
the other arguments for seyfert activity in ngc 1672 , while entirely plausible , lack robustness .
the reasoning based upon the composite nature of the nuclear optical emission lines ( vron et al .
1981 ) relies upon fairly subtle effects , and we note that storchi - bergmann et al . (
1996 ) have recently found that the nuclear emission - line ratios can be explained via ionization by hot ( @xmath61 k ) stars . in this case
ngc 1672 may be an example of a non - agn liner .
furthermore , moorwood & oliva ( 1988 ) failed to detect the broad brackett @xmath62 line claimed by kawara , nishida & gregory ( 1987 ) , and we are not aware of any confirmation of the radio variability suggested by tovmassian ( 1968 ) . while we must still admit the possibility of a luminous
, highly obscured seyfert nucleus in ngc 1672 , the physical interpretation of its optical emission lines is not yet clear , and the balance of evidence appears to be shifting away from this .
the most interesting extranuclear source in ngc 1672 is x-3 .
we have demonstrated that x-3 shows large - amplitude soft x - ray variability on a timescale of @xmath53 years , suggesting that it is a single ( or at most a few ) object .
x-3 has a peak observed soft x - ray luminosity of @xmath1 erg s@xmath2 , and it probably has an even larger hard x - ray luminosity . if the hard x - ray emission of 3.2 and figure 4b originates from x-3 ( the most probable possibility in our opinion ) , its 0.28 kev x - ray luminosity is @xmath3 erg s@xmath2 .
the fairly large ratio of hard to soft x - ray flux could be explained if x-3 suffers from some absorption within ngc 1672 ; the hard spectral components of starburst galaxies are often absorbed by column densities of @xmath63 @xmath5 ( e.g. , ptak et al .
the variability , luminosity , and extranuclear location of x-3 suggest that it is a powerful x - ray binary , young supernova remnant , or young hypernova remnant .
no supernovae are listed for ngc 1672 in the asiago supernova catalogue ( barbon et al .
if x-3 is an accreting compact object that emits isotropically , the mass required by the the eddington limit is @xmath64 m@xmath65 . to our knowledge
, the only source in our galaxy with a comparable x - ray luminosity is grs 1915 + 105 ( e.g. , greiner , morgan & remillard 1998 ) , and x-3 ranks among the more luminous of the ` super - eddington sources ' seen in nearby galaxies ( e.g. , fabbiano 1998 ; colbert & mushotzky 1999 ; and references therein ) .
the new x - ray sources we find in ngc 1672 ( x-7 , x-8 and x-9 ) are also quite luminous and exceed the eddington limit for a 1.4 m@xmath65 object .
they too are probably x - ray binaries or supernova remnants , and sources with comparable luminosities are seen fairly frequently in nearby spiral galaxies ( e.g. , read , ponman & strickland 1997 , hereafter rps97 ) .
we have compared the number and the luminosity distribution of our extranuclear x - ray sources with those of the nearby galaxies studied by rps97 , and ngc 1672 shows a moderately large although not extreme amount of extranuclear x - ray source activity .
there are almost certainly many more x - ray binaries and supernova remnants lying below our current x - ray sensitivity threshold ( @xmath66 erg s@xmath2 ) .
a moderate - length observation holds particular promise for improving our understanding of ngc 1672 .
separation of the diffuse nuclear x - ray emission from any point - like central source would provide a sensitive probe for low - level seyfert activity .
in addition , would allow an excellent imaging study of the nuclear starburst x - rays , and like the radio emission the x - rays might trace the nuclear ring ( e.g. , lindblad & jrster , in preparation ) .
furthermore , would provide much better spectral constraints upon x-3 and would detect x - ray binaries and supernova remnants down to @xmath67 erg s@xmath2 in this nearby starburst galaxy .
we thank m. corcoran , h. ebeling , e. gotthelf and an anonymous referee for helpful discussions , and we thank h. ebeling for the use of his idl software .
we gratefully acknowledge financial support from nasa grants nag5 - 4826 and nag5 - 6023 ( pjdn ) , and nasa ltsa grant nag5 - 8107 ( wnb ) .
ccccccccc x-1 & 1992 & 04 45 42.2 & @xmath68 14 50.3 & @xmath69 & 22.4 & @xmath70 & 20.9 & 7.4 & 1997 & 04 45 41.9 & @xmath68 14 50.0 & @xmath71 & 24.5 & @xmath72 & 19.6 & 7.0 & comb . &
04 45 42.1 & @xmath68 14 50.3 & @xmath73 & 33.1 & @xmath74 & 19.8 & 7.0 x-2 & 1992 & 04 45 53.0 & @xmath68 14 56.5 & @xmath75 & 10.1 & @xmath76 & 6.6 & 2.6 & 1997 & 04 45 52.5 & @xmath68 14 55.0 & @xmath77 & 11.6 & @xmath78 & 5.2 & 2.1 & comb . &
04 45 52.8 & @xmath68 14 55.4 & @xmath79 & 14.6 & @xmath80 & 5.6 & 2.2 x-3 & 1992 & 04 45 33.9 & @xmath68 14 41.3 & @xmath81 & 7.7 & @xmath82 & 3.6 & 1.4 & 1997 & 04 45 33.6 & @xmath68 14 42.0 & @xmath83 & 9.9 & @xmath84 & 6.2 & 2.5 & comb . &
04 45 33.7 & @xmath68 14 42.1 & @xmath85 & 12.8 & @xmath86 & 5.3 & 2.1 x-5 & 1992 & 04 45 49.6 & @xmath68 12 49.0 & @xmath87 & 5.8 & @xmath88 & 2.7 & & 1997 & & & @xmath89 & @xmath90 & @xmath91 & @xmath92 & & comb . & 04 45 49.6 & @xmath68 12 49.2 & @xmath93 & 6.1 & @xmath94 & 1.4 & x-7 & 1992 & 04 45 31.8 & @xmath68 14 54.8 & @xmath95 & 4.7 & @xmath96 & 2.0 & 0.77 & 1997 & 04 45 31.0 & @xmath68 14 54.0 & @xmath97 & 6.6 & @xmath98 & 2.3 & 0.94 & comb . &
04 45 31.3 & @xmath68 14 53.7 & @xmath99 & 7.4 & @xmath100 & 2.1 & 0.82 x-8 & 1992 & & & @xmath101 & @xmath90 & @xmath102 & @xmath92 & @xmath103 & 1997 & 04 45 57.1 & @xmath68 14 54.0 & @xmath104 & 4.0 & @xmath105 & 0.86 & 0.35 & comb .
& 04 45 57.3 & @xmath68 14 54.9 & @xmath106 & @xmath90 & @xmath107 & @xmath108 & @xmath109 x-9 & 1992 & & & @xmath101 & @xmath90 & @xmath110 & @xmath111 & @xmath112 & 1997 & & & @xmath113 & @xmath90 & @xmath114 & @xmath115 & @xmath116 & comb . &
04 45 29.5 & @xmath68 13 27.3 & @xmath117 & 4.0 & @xmath118 & 0.70 & 0.29 | the nearby barred spiral galaxy ngc 1672 shows dramatic starburst activity and may also host a seyfert 2 nucleus .
we present new x - ray observations that set constraints on starburst and seyfert activity in ngc 1672 .
two hri exposures , taken in 1992 and 1997 , are used to investigate long - term variability of the known x - ray sources and to search for new sources of x - ray emission .
we find large - amplitude ( @xmath0% ) variability from x-3 , one of the off - nuclear sources located near an end of the galactic bar .
x-3 has a peak observed 0.22.0 kev luminosity of @xmath1 erg s@xmath2 , and it is probably a luminous x - ray binary or young supernova remnant .
we do not observe variability of the nuclear source x-1 or the strong off - nuclear source x-2 .
our analyses also reveal two new off - nuclear sources , one of which is associated with a bright region along a spiral arm , and we find evidence for large - scale diffuse x - ray emission throughout part of the disk of ngc 1672 .
furthermore , we use data taken in 1995 to constrain the hard x - ray properties of ngc 1672 . while the nuclear source x-1 is the dominant soft x - ray source in ngc 1672 , we find that the bulk of the 210 kev and 510 kev emission is spatially coincident with the off - nuclear source x-3 , giving it an apparent 0.28 kev luminosity of @xmath3 erg s@xmath2 . a power - law plus raymond - smith model provides an acceptable fit to the full - band spectra .
we do not find any evidence for a luminous but absorbed nuclear x - ray source . if there is a luminous seyfert 2 nucleus in ngc 1672 , it must be obscured by a ` compton - thick ' torus with a column density of @xmath4 @xmath5 . |
there have been many efforts to develop evolutionary population synthesis models ( e.g. bruzual & charlot 1993 ; worthey 1994 ; buzzoni 1995 ; vazdekis et al . 1996 ; maraston 1998 ; lee et al .
2000,2002 ) in order to analyze the integrated spectrophotometric quantities of globular clusters ( gcs ) and galaxies .
ultimately , the primary use of such software packages lies in attempting to derive the ages and metallicities of stellar populations such that quantitative insights into the processses governing galaxy formation can be uncovered .
age and metallicity drive the integrated spectrophotometry of stellar systems , and they also drive the resulting hb morphology of globular clusters . it is interesting therefore to pose the question as to whether this underlying morphology also manifests itself in the integrated spectrophotometry .
there is reason to suspect that age is the _ global _ `` second parameter '' that controls hb morphology ( after the `` first parameter '' , metallicity - e.g. lee , demarque & zinn 1994 ; sarajedini , chaboyer & demarque 1997 ; rey et al .
2001 ; salaris & weiss 2002 ) , although for some gcs a third ( or more ) parameter may be needed to explain their peculiar hb morphologies ( such as the blue tail phenomenon - e.g. recio - blanco et al .
2002 ) . in what follows
, we describe the unique aspects of our recent models which allow for an accurate determination of the role played by blue hb stars in modifying the spectrophotometric indices of ssps .
the reader is directed to lee et al .
( 2000,2002 ) for details .
we first construct a grid of simulated gcs spanning a range of hb morphologies - in order to match the observed morphology of inner halo galactic gcs ( galactocentric radius @xmath1 8 kpc ) at their currently favoured mean age ( @xmath212 gyr ) and age dispersion ( @xmath3_t_=0 gyr ) a reimers ( 1975 ) mass loss parameter @xmath4=0.65 was required . the salaris et al .
( 1993 ) @xmath5-element enhancement correction was applied to our models .
the modified gaussian mass distribution of demarque et al .
( 2000 ) - with @xmath6=0.02 m@xmath7 - was used , shaping the widths of the `` wavy '' features seen in figure 1 .
the temperature - sensitive b - v and h@xmath0 indices shown there illustrate the magnitude of the contribution of blue hb stars ( thick wavy lines - c.f .
maraston & thomas 2000 ) within @xmath84 gyr of system s mean age .
the resulting wavy features stem primarily from our exploration of parameter space 2@xmath94 gyrs _ older _ than the milky way gcs .
the location of these wavy `` peaks '' is determined by the amount of mass loss adopted . ) as a function of metallicity and age ( individual curves ) .
the wavy features for ssps older than the galactic gcs are driven by the inclusion of blue hb stars.,scaledwidth=100.0% ]
perhaps the greatest value that the addition of far - uv photometry to pre - existing optical data provides is the ability to discriminate _ cleanly _ between young ( @xmath101gyr ) , intermediate - age ( 3@xmath95gyr ) , and old ( @xmath1114gyr ) , metal - rich ( @xmath90.8 @xmath1 [ fe / h ] @xmath1 @xmath90.4 ) , clusters - this ability is demonstrated in figure 2 , where it can be seen that a significant far - uv to optical flux ratio is indicative of either very young or very old systems ( for these metallicities ) . in contrast , intermediate - age clusters are relatively faint in the far - uv .
we suggest that the combination of far - uv and optical colours is a most promising chronometer for ssps ( with the admitted caveat that there are peculiar gcs such as ngc 6388 and 6441 with populations of extremely blue hb stars - e.g. rich et al .
_ if _ there are significant age differences between milky way and giant elliptical gc systems , however , in the sense of the latter being @xmath112 gyr older ( as suggested by lee et al .
2000 ) , then we predict that the majority of metal - rich clusters will be far - uv bright compared to their galactic counterparts . in this respect ,
the outcomes of the recent hst far - uv photometry of gc systems in virgo ellipticals ( pids#8643 & 8725 ) are highly anticipated ; future work with hubble s advanced camera for surveys will also be a fruitful avenue of research .
finally , the upcoming galex mission will shed important light on the uv properties of the m31 globular cluster system , perhaps resolving the controversy regarding their supposed young , metal - rich , nature ( e.g. burstein et al .
1984 ; barmby & huchra 2000 ) .
those galactic gcs with minimal foreground extinction ( filled circles : inner halo clusters ; open circles : outer halo clusters ) are contrasted with our models in the ( @xmath12 @xmath9 @xmath13)@xmath14 , ( @xmath13 @xmath9 @xmath15)@xmath14 , ( @xmath16 @xmath9 @xmath17)@xmath14 , and ( @xmath18 @xmath9 @xmath17)@xmath14 vs. [ fe / h ] planes in figure 3 . the data are taken from harris ( 1996 , 1999 june version ) and harris & canterna ( 1977 ) . we believe this to be one of the first attempts at employing integrated broadband colours to derive relative age differences between subpopulations of galactic gcs .
( @xmath12 @xmath9 @xmath13 ) @xmath10 0.2 ] are used to calibrate our models in the ( @xmath12 @xmath9 @xmath13)@xmath14 , ( @xmath13 @xmath9 @xmath15)@xmath14 , ( @xmath16 @xmath9 @xmath17)@xmath14 , and ( @xmath18 @xmath9 @xmath17)@xmath14 vs. [ fe / h ] planes . the dashed and solid lines represent ages of 10 gyr and 12 gyr , respectively.,scaledwidth=100.0% ] it appears that our models reproduce the differences between inner and outer halo clusters in the sense that the inner halo clusters are not only more tightly grouped along the isochrone than the more scattered outer halo counterparts , but also relatively older .
this relative age difference between inner and outer halo milky way gcs is in agreement with that obtained by salaris & weiss ( 2002 ) from their homogeneous age dating of 55 clusters ( as shown in figure 4 ) .
the hb morphology is responsible for driving the non - linearity seen near [ fe / h ] @xmath2 @xmath91.6 , demonstrating that the inclusion of morphological effects are important if one is to attempt subtle relative age - dating between ssps . as most of the linear colour - metallicity transformation relations ( cmrs ) in the literature
are based upon the limited range in colour of galactic gcs , caution must be employed when extrapolating their use to redder extragalactic systems ( or if there are sizable age differences amongst globular clusters within and between galaxies ) .
recently , kissler - patig et al .
( 1998 ) provided a set of cmrs using the globulars of ngc 1399 , but the apparent systematic offset has been questioned by lee et al .
it remains to be seen if the bimodal colour distributions of extragalactic gc systems can be interpreted as simple metallicity or age differences , or whether a more complicated interplay between metallicity and age must be invoked .
lee sincerely thanks the organizers of this beautiful workshop for their hospitality .
support for this work was provided by the postdoctoral fellowship program of the korea science & engineering foundation , the creative research initiatives program of the korean ministry of science and technology , and the australian research council through its linkage international ( lx0346898 ) and large grant ( a00105171 ) programs .
barmby , p. , & huchra , j. p. 2000 , apj , 531 , l29 bruzual , g. , & charlot , s. 1993 , apj , 405 , 538 burstein , d. , faber , s. m. , gaskell , c. m. , & krumm , n. 1984 , apj , 287 , 586 buzzoni , a. 1995 , apjs , 98 , 69 demarque , p. , zinn , r. , lee , y .- w . , &
yi , s. 2000 , aj , 119 , 1398 harris , h. c. , & canterna , r. 1977 , aj , 82 , 798 harris , w. e. 1996 , aj , 112 , 1487 kissler - patig , m. , et al .
1998 , aj , 115 , 105 lee , h .- c . ,
lee , y .- w . , & gibson , b. k. 2002 , aj , in press ( astro - ph/0208047 ) lee , h .- c . ,
yoon , s .- j .
, & lee , y .- w . 2000 ,
aj , 120 , 998 lee , y .- w . , demarque , p. , & zinn , r. 1994 , apj , 423 , 248 maraston , c. 1998 , mnras , 300 , 872 maraston , c. , & thomas , d. 2000 , apj , 541 , 126 recio - blanco , a. , piotto , g. , aparicio , a. , & renzini , a. 2002 , apj , 572 , l71 reimers , d. 1975 , mm .
ser . 8 , 369 rey , s .- c . , et al .
2001 , aj , 122 , 3219 rich , r. m. , et al .
1997 , apj , 484 , l25 salaris , m. , chieffi , a. , & straniero , o. 1993 , apj , 414 , 580 salaris , m. , & weiss , a. 2002 , a&a , 388 , 492 sarajedini , a. , chaboyer , b. , & demarque , p. 1997 , pasp , 109 , 1321 vazdekis , a. , casuso , e. , peletier , r. f. , & beckman , j. e. 1996 , apjs , 106 , 307 worthey , g. 1994 , apjs , 95 , 107 | the effects of blue horizontal - branch ( hb ) stars on the integrated spectrophotometric indices of simple stellar populations ( ssps ) are addressed .
synthetic spectra are drawn from a grid of simulated globular clusters , constructed so as to reproduce the observed range of hb morphologies in the milky way system .
our models indicate that the h@xmath0 line index increases by as much as 0.75 and the integrated b - v colour becomes bluer by upwards of 0.15 mag because of the contribution from blue hb stars .
we discuss the importance of both optical _ and _ far - uv colours as age discriminants for ssps , in particular for high metallicities .
we suggest that relative age differences ( and consequently their differing hb morphologies ) result in non - linear colour - metallicity relations . |
this paper considers a decision problem that involves gussing an invisible state @xmath0 after observing @xmath1 , which is correlated with @xmath0 . in decision theory (
* section 1.5.2 ) , an optimal decision rule , which minimizes the decision error probability , is called the _
bayes decision rule_. let @xmath0 and @xmath1 be random variables that take values in sets @xmath2 and @xmath3 , respectively , where we call @xmath0 a _ state of nature _ or a _
parameter _ and @xmath1 an _
observation_. let @xmath4 be the joint distribution of @xmath5 .
let @xmath6 and @xmath7 be the marginal distributions of @xmath0 and @xmath1 , respectively .
let @xmath8 be the conditional distribution of @xmath0 for a given @xmath1 .
it is well known that an optimal strategy for guessing the state @xmath0 consists of finding @xmath9 , which maximizes the conditional probability @xmath10 depending on a given observation @xmath11 . formally , by taking an @xmath9 that maximizes @xmath10 for each @xmath12 , we can define the function @xmath13 as @xmath14 which is a bayes decision rule . it should be noted that the discussion throughout this paper does not depend on choosing states with the same maximum probability .
when the cardinality @xmath15 of @xmath2 is small , operations ( [ eq : map ] ) and ( [ eq : ml ] ) are tractable by using a brute force search . however , with coding problems , these operations appears to be intractable because @xmath15 grows exponentially as the dimension of @xmath2 grows . in this paper
, we assume a situation where operations ( [ eq : map ] ) and ( [ eq : ml ] ) are intractable . in source coding
, @xmath0 corresponds to a source output and @xmath1 corresponds to a codeword and side information . in channel coding
, @xmath0 corresponds to a codeword and @xmath1 corresponds to a channel output , where the decoding with ( [ eq : map ] ) is called _ maximum a posteriori decoding_. on the other hand , the decoding method that maximizes the conditional probability @xmath16 of a channel is called _ maximum likelihood decoding _ ) as maximizing the likelihood @xmath17 . for this reason ,
we might call @xmath18 a _ maximum - likelihood decision rule_. in a series of papers on the coding problem , we have called @xmath18 a maximum - likelihood decoding based on this idea . ] , which is equivalent to maximum a posteriori decoding when @xmath0 is generated subject to the uniform distribution . in this paper
, we call the decision rule with @xmath18 the _ maximum a posteriori decision rule_. in this paper , we consider a stochastic decision , where the decision is made randomly subject to a probability distribution .
we investigate the relationship between the error probabilities of the stochastic and maximum a posteriori decisions .
then , we introduce the construction of stochastic decoders for source / channel codes .
for a stochastic decision , we use a random number generator to obtain @xmath19 after observing @xmath1 and let @xmath20 be a decision ( guess ) about the state @xmath0 . formally , we generate @xmath20 subject to the conditional distribution @xmath21 on @xmath2 depending on an observation @xmath1 and let an output be a decision of @xmath0 , where @xmath0 and @xmath20 are conditionally independent for a given @xmath1 , that is , @xmath22 forms a markov chain .
the joint distribution @xmath23 of @xmath24 is given as @xmath25 let us call @xmath26 a _ stochastic decision rule_. as a special case , when @xmath26 is given by using a function @xmath27 and is defined as @xmath28 we call @xmath26 or @xmath29 a _ deterministic decision rule_. it should be noted that the maximum a posteriori decision rule is deterministic . throughout this paper , we assume that @xmath2 and @xmath3 are countable sets . it should be noted that the results do not change when @xmath3 is an uncountable set , where the summation should be replaced with the integral . the following lemma guarantees that the right hand side of ( [ eq : map ] ) always exists for every @xmath12 .
this fact , which is implicitly used in this paper , implies that it is enough to assume that @xmath2 is a countable set .
let @xmath30 be a probability distribution on a countable set @xmath2 .
then the maximum of @xmath30 on @xmath2 always exists , that is , there is @xmath31 such that @xmath32 for any @xmath33 .
the lemma is trivial when @xmath2 is a finite set . in the following ,
we assume that @xmath2 is a countable infinite set .
since @xmath34 for all @xmath33 , then @xmath35 always exists , that is , @xmath36 for all @xmath37 , and for any @xmath38 there is a @xmath37 such that @xmath39 .
we prove the lemma by contradiction .
assume that there is no @xmath31 such that @xmath40 .
since @xmath41 and @xmath42 for all @xmath33 , there is @xmath43 such that @xmath44 . from the definition of @xmath35 , there is a @xmath45 such that @xmath46 , where the second inequality comes from the assumption . by repeating this argument
times so that @xmath47 . ]
, we have a sequence @xmath48 such that @xmath49 this implies @xmath50 , which contradicts @xmath41 .
when @xmath2 is a finite dimensional euclidian space , we can make the same discussion by quantizing uniformly from @xmath2 to a countable set , where the decision is interpreted as guessing @xmath0 with a finite precision .
then we can apply the results to parameter estimation problems .
let @xmath51 be a support function defined as @xmath52 then the error probability @xmath53 of a ( stochastic ) decision rule @xmath26 is given as @xmath54 .
\label{eq : error - random}\end{aligned}\ ] ] in the last equality , @xmath55 corresponds to the error probability of the decision rule @xmath26 after the observation @xmath12 , and @xmath53 corresponds to the average of this error probability .
when @xmath26 is defined by using @xmath27 and ( [ eq : deterministic ] ) , the decision error probability @xmath56 of a deterministic decision rule @xmath29 is given as @xmath57 .
\label{eq : error - f}\end{aligned}\ ] ] it should be noted that the right hand side of the first equality can be derived directly from ( [ eq : error - random ] ) and the fact that @xmath58 that is , we have @xmath59 when @xmath29 and @xmath26 satisfy ( [ eq : deterministic ] ) .
in this section , we discuss the optimality of the maximum a posteriori decision .
we introduce the following well - known lemma .
[ prop : bayes ] let @xmath5 be a pair consisting of a state @xmath0 and an observation @xmath1 and @xmath4 be the joint distribution of @xmath5 .
when we make a stochastic decision with a distribution @xmath26 , an optimal decision rule minimizing the decision error probability satisfies @xmath60 for all @xmath61 such that @xmath62 and @xmath63 . in particular , the maximum a posteriori decision rule @xmath18 defined by ( [ eq : map ] ) minimizes the decision error probability , where @xmath26 is defined by @xmath64 and ( [ eq : deterministic ] ) . from the definition of @xmath53 , we have @xmath65 \notag \\ & = 1-\sum_y p_y(y)\sum_x p_{x|y}(x|y)q_{{\widehat{x}}|y}(x|y ) \notag \\ & = 1 -\sum_y p_y(y ) { \left [ { \sum_{x\neq { f_{\mathrm{map}}}(y)}p_{x|y}(x|y)q_{{\widehat{x}}|y}(x|y ) + p_{x|y}({f_{\mathrm{map}}}(y)|y)q_{{\widehat{x}}|y}({f_{\mathrm{map}}}(y)|y ) } \right ] } \notag \\ & = 1 -\sum_y p_y(y ) { \left [ { \sum_{x\neq{f_{\mathrm{map}}}(y ) } p_{x|y}(x|y)q_{{\widehat{x}}|y}(x|y ) + p_{x|y}({f_{\mathrm{map}}}(y)|y){\left[{1-\sum_{x\neq { f_{\mathrm{map}}}(y)}q_{{\widehat{x}}|y}(x|y)}\right ] } } \right ] } \notag
\\ & = 1 -\sum_yp_y(y)p_{x|y}({f_{\mathrm{map}}}(y)|y ) + \sum_yp_y(y ) \sum_{x\neq{f_{\mathrm{map}}}(y)}q_{{\widehat{x}}|y}(x|y){\left[{p_{x|y}({f_{\mathrm{map}}}(y)|y)-p_{x|y}(x|y)}\right]}. \label{eq : optimal}\end{aligned}\ ] ] ( [ eq : optimal ] ) , which is on the top of the next page , where @xmath66 implies that @xmath53 is minimized only when @xmath60 for all @xmath61 such that @xmath62 and @xmath67 ( i.e. , @xmath68 ) .
in this section , we consider the case where @xmath69 for all @xmath70 , that is , we make a stochastic decision with the conditional distribution @xmath8 of a state @xmath0 for a given observation @xmath1 .
it should be noted that @xmath20 is independent of @xmath0 for a given @xmath1 , where the joint distribution @xmath23 of @xmath24 is given as @xmath71 in the following , we call this type of decision a _
stochastic decision with the a posteriori distribution_. it should be noted that it may be unnecessary to know ( or compute ) the distribution @xmath8 to make this type of decision . to make this type of decision ,
it is sufficient that we have a random number generator subject to the distribution @xmath72 with arbitrary input @xmath12 , where the generated random number is independent of @xmath0 for a given @xmath11 .
we have the following theorem . in section [ sec : decoder ] , we apply this theorem to an analysis of stochastic decoders of coding problems .
[ thm : random - fmap ] let @xmath5 be a pair consisting of a state @xmath0 and an observation @xmath1 and @xmath4 be the joint distribution of @xmath5 . when we make a stochastic decision with @xmath8 ,
the decision error probability of this rule is at most twice the decision error probability of the maximum a posteriori decision rule @xmath18 .
that is , we have @xmath73 it should be noted that this theorem can be considerd as a special case of ( * ? ? ?
* corollary 1 of theorem 1 ) ( * ? ? ?
* eq . ( 29 ) ) which assme a general loss function .
although the idea of the proof in ( * ? ? ?
* theorem 1 ) is quite simple by assuming that a loss function is symmetric and satisfies the triangle inequality , we show this lemma in a different way by assuming the loss function @xmath74 .
we have @xmath75 \notag \\ & = \sum_{y}p_y(y ) { \left[{1-\sum_{x}p_{x|y}(x|y)^2}\right ] } \notag \\ & \leq \sum_{y}p_y(y ) { \left[{1-p_{x|y}({f_{\mathrm{map}}}(y)|y)^2}\right ] } \notag \\ & = \sum_{y}p_y(y ) [ 1-p_{x|y}({f_{\mathrm{map}}}(y)|y ) ] [ 1+p_{x|y}({f_{\mathrm{map}}}(y)|y ) ] \notag \\ & \leq 2\sum_{y}p_y(y ) [ 1-p_{x|y}({f_{\mathrm{map}}}(y)|y ) ] \notag \\ & = 2{\mathrm{error}}({f_{\mathrm{map } } } ) , \label{eq : proof - random - deterministic}\end{aligned}\ ] ] where the second inequality comes from the fact that @xmath76 and the fourth equality comes from ( [ eq : error - f ] ) . here
, we introduce the following inequalities , which come from lemma [ prop : bayes ] and theorem [ thm : random - fmap ] . in these inequalities , if either @xmath77 or @xmath78 vanishes as the dimension ( block length ) of @xmath0 goes to infinity , then the other one also vanishes .
[ cor : bound ] @xmath79 here , we introduce another lemma that comes from lemma [ prop : bayes ] and theorem [ thm : random - fmap ] .
[ cor : random ] let @xmath5 be a pair consisting of a state @xmath0 and an observation @xmath1 and @xmath4 be the joint distribution of @xmath5 .
when we make a stochastic decision with @xmath8 , the decision error probability of this rule is at most twice the decision error probability of _ any _ decision rule @xmath26 .
that is , we have @xmath80 for any @xmath26 .
let us consider a situation where @xmath26 is unknown but @xmath53 can be estimated empirically .
then the above corollary implies that the error probability of stochastic decision with the a posteriori distribution is upper bounded by @xmath81 .
for example , when we know empirically that a human being can guess @xmath0 with small error probability explicitly .
] , then the error probability of a stochastic decision with the a posteriori distribution is also small because it is at most twice the error probability of her / his decision rule . inequality ( [ eq : random ] ) is tight in the sense that there is a pair consisting of @xmath8 and @xmath26 such that ( [ eq : random ] ) is satisfied with equality . in fact , by assuming that @xmath82 for all @xmath12 , we have @xmath83\ ] ] and @xmath84 \notag \\ * & \quad + 2\sum_yp_y(y)p_{x|y}(1|y)[1-q_{{\widehat{x}}|y}(1|y ) ] \notag \\ & = 2\sum_yp_y(y)p_{x|y}(0|y)[1-q_{{\widehat{x}}|y}(0|y ) ] \notag \\ * & \quad + 2\sum_yp_y(y)[1-p_{x|y}(0|y)]q_{{\widehat{x}}|y}(0|y ) \notag \\ & = 2\sum_yp_y(y)p_{x|y}(0|y ) \notag \\ * & \quad + 2\sum_yp_y(y)[1 - 2p_{x|y}(0|y)]q_{{\widehat{x}}|y}(0|y ) \notag \\ & = 2\sum_yp_y(y)p_{x|y}(0|y ) -2\sum_yp_y(y)p_{x|y}(0|y)^2 \notag \\ * & = 2\sum_yp_y(y)p_{x|y}(0|y)[1-p_{x|y}(0|y)]\end{aligned}\ ] ] from ( [ eq : error - random ] ) .
let @xmath85 be the variational distance of two probability distributions @xmath30 and @xmath86 on the same set as @xmath87 ( see ( * ? ? ? * eq . ( 11.137 ) ) ) .
we have the following lemma .
[ thm : random - approx ] let @xmath5 be a pair consisting of a state @xmath0 and an observation @xmath1 and @xmath4 be the joint distribution of @xmath5 .
when we make decisions with two stochastic decision rules @xmath88 and @xmath89 for each @xmath12 , we have @xmath90 where @xmath91 and @xmath92 .
the lemma is obtained immediately from ( [ eq : vd - max ] ) by considering the decision error event measured by using the joint probability distributions @xmath93 and @xmath94 .
formally , we have @xmath95 - \sum_{y}p_y(y ) \sum_{x } p_{x|y}(x|y ) [ 1-q'(x|y ) ] } \right| } \notag \\ & = { \left| { \sum_{y}p_y(y ) \sum_{x } p_{x|y}(x|y ) q({\mathcal{x}}\setminus\{x\}|y ) - \sum_{y}p_y(y ) \sum_{x } p_{x|y}(x|y ) q'({\mathcal{x}}\setminus\{x\}|y ) } \right| } \notag \\ & \leq \sum_{y}p_y(y ) \sum_{x } p_{x|y}(x|y ) { \left| { q({\mathcal{x}}\setminus\{x\}|y ) - q'({\mathcal{x}}\setminus\{x\}|y ) } \right| } \notag \\ & \leq \sum_{y}p_y(y ) \sum_{x } p_{x|y}(x|y ) d(q(\cdot|y),q'(\cdot|y ) ) \notag \\ & = \sum_{y}p_y(y ) d(q(\cdot|y),q'(\cdot|y ) ) \notag \\ & = d(q\times p_y , q'\times p_y ) \label{eq : proof - random - approx}\end{aligned}\ ] ] ( [ eq : proof - random - approx ] ) , which appears on the top of the next page , where the second inequality comes from ( [ eq : vd - max ] ) and the last equality comes from ( [ eq : vd ] ) as @xmath96 applying lemma [ thm : random - approx ] by letting @xmath97 and @xmath98 , we have the following theorem from theorem [ thm : random - fmap ] .
let @xmath5 be a pair consisting of a state @xmath0 and an observation @xmath1 and @xmath4 be the joint distribution of @xmath5 .
when we make a stochastic decision with @xmath26 , the decision error probability @xmath53 is bounded as @xmath99
in this section , we assume that a conditional probability @xmath8 is computable . we make a stochastic decision @xmath100 from a random sequence @xmath101 as @xmath102 a typical example of a random sequence @xmath101 is that generated by a markov chain monte carlo method . here
, we assume that @xmath0 and @xmath103 are conditionally independent for a given @xmath1 , that is , the joint distribution @xmath104 of @xmath105 is given as @xmath106 where @xmath107 is a joint probability distribution of @xmath103 for a given @xmath12 .
then , we have a decision error probability @xmath108 as follows ; @xmath109 } \notag \\ & = e_{y{\widehat{x}}^{{\overline{t}}}}{\left [ { 1-p_{x|y}({\widehat{x}}_{t(y)}|y ) } \right ] } \notag \\ & = e_{y{\widehat{x}}^{{\overline{t}}}}{\left [ { 1-\max_{t\in\{1,\ldots,{\overline{t}}\}}p_{x|y}({\widehat{x}}_t|y ) } \right]}. \label{eq : seq - error}\end{aligned}\ ] ] we have the following theorem . [ thm : seq - random ] let @xmath5 be a pair consisting of a state @xmath0 and an observation @xmath1 and @xmath4 be the joint distribution of @xmath5 .
when we make a stochastic decision @xmath110 with a random sequence @xmath103 defined by ( [ eq : ty])([eq : sequence ] ) , the decision error probability @xmath108 defined by ( [ eq : seq - error ] ) satisfies @xmath111 where @xmath112 is a conditional marginal distribution given as @xmath113 from ( [ eq : seq - error ] ) , we have @xmath114 } \notag \\ & \leq e_{y{\widehat{x}}^{{\overline{t}}}}{\left [ { 1-p_{x|y}({\widehat{x}}_t|y ) } \right ] } \notag \\ & = \sum_{y}p_y(y ) \sum_{{\widehat{x}}^{{\overline{t } } } } q_{{\widehat{x}}^{{\overline{t}}}|y}({\widehat{x}}^{{\overline{t}}}|y ) { \left [ { 1-p_{x|y}({\widehat{x}}_t|y ) } \right ] } \notag \\ & = \sum_{y}p_y(y ) \sum_{{\widehat{x}}_t}q_{{\widehat{x}}_t|y}({\widehat{x}}_t|y ) [ 1-p_{x|y}({\widehat{x}}_t|y ) ] \notag \\ & = \sum_{y}p_y(y ) \sum_{{\widehat{x}}_t } p_{x|y}({\widehat{x}}_t|y ) [ 1-q_{{\widehat{x}}_t|y}({\widehat{x}}_t|y ) ] \notag \\ & = { \mathrm{error}}(q_{{\widehat{x}}_t|y})\end{aligned}\ ] ] for any @xmath115 , where the fourth equality comes from the fact that @xmath8 and @xmath116 are probability distributions . from this inequality , we have ( [ eq : seq - random ] ) .
applying lemma [ thm : random - approx ] by letting @xmath117 and @xmath98 , we have the following theorem from theorem [ thm : random - fmap ] .
this theorem implies that if @xmath118 tends towards @xmath119 as @xmath120 ( e.g. ( [ eq : gibbs - approx ] ) in appendix ) then the upper bound of error probability @xmath108 is close to at most twice the error probability @xmath77 of the maximum a posteriori decision .
[ thm : seq - random - approx ] let @xmath5 be a pair consisting of a state @xmath0 and an observation @xmath1 and @xmath4 be the joint distribution of @xmath5 . when we make a stochastic decision @xmath110 with a random sequence @xmath121 defined by ( [ eq : ty])([eq : sequence ] ) , the decision error probability is bounded as @xmath122 in the following , we assume that a random sequence @xmath103 is independent and identically distributed ( i.i.d . ) with a distribution @xmath26 for a given @xmath1 , that is , the conditional probability distribution @xmath123 is given as @xmath124 then we have the following theorem . from this theorem with a trivial inequality @xmath125
, we have the fact that @xmath108 tends towards the error probability @xmath77 of the maximum a posteriori decision , where the difference @xmath126 is exponentially small as the length @xmath127 of a sequence increases .
[ thm : seq - iid ] let @xmath5 be a pair consisting of a state @xmath0 and an observation @xmath1 and @xmath4 be the joint distribution of @xmath5 . when we make a stochastic decision @xmath110 with an i.i.d .
random sequence @xmath103 defined by ( [ eq : ty])([eq : sequence ] ) and ( [ eq : iid ] ) , the decision error probability @xmath108 defined by ( [ eq : seq - error ] ) satisfies @xmath128^{{\overline{t } } } , \label{eq : seq - qiid - fmap}\end{aligned}\ ] ] where @xmath18 is given by ( [ eq : map ] ) .
in particlular , when @xmath129 , we have @xmath130^{{\overline{t } } } \label{eq : seq - iid - fmap } \\ & \leq { \mathrm{error}}({f_{\mathrm{map}}})+{\left[{1-\inf_{y : p_y(y)>0}\max_x p_{x|y}(x|y)}\right]}^{{\overline{t}}}. \label{eq : seq - iid - inffmap}\end{aligned}\ ] ] @xmath131 } \notag \\ & = \sum_{y}p_y(y)\sum_{{\widehat{x}}^{{\overline{t}}}\in{\widehat{{\mathcal{x}}}}^{{\overline{t } } } } { \left [ { 1-\max_{t\in\{1,\ldots,{\overline{t}}\}}p_{x|y}({\widehat{x}}_t|y ) } \right ] } \prod_{t=1}^{{\overline{t}}}q_{{\widehat{x}}|y}({\widehat{x}}_t|y ) + \sum_{y}p_y(y)\sum_{{\widehat{x}}^{{\overline{t}}}\notin{\widehat{{\mathcal{x}}}}^{{\overline{t } } } } { \left [ { 1-\max_{t\in\{1,\ldots,{\overline{t}}\}}p_{x|y}({\widehat{x}}_t|y ) } \right ] } \prod_{t=1}^{{\overline{t}}}q_{{\widehat{x}}|y}({\widehat{x}}_t|y ) \notag \\ & \leq \sum_{y}p_y(y)\sum_{{\widehat{x}}^{{\overline{t}}}\in{\widehat{{\mathcal{x}}}}^{{\overline{t } } } } { \left [ { 1-\max_{{\widehat{x}}}p_{x|y}({\widehat{x}}|y ) } \right ] } \prod_{t=1}^{{\overline{t}}}q_{{\widehat{x}}|y}({\widehat{x}}_t|y ) + \sum_{y}p_y(y)\sum _ { { \widehat{x}}^{{\overline{t}}}:{\widehat{x}}_t\neq{f_{\mathrm{map}}}(y)\ \text{for all}\ t } \prod_{t=1}^{{\overline{t}}}q_{{\widehat{x}}|y}({\widehat{x}}_t|y ) \notag \\ & = \sum_{y}p_y(y ) { \left [ { 1-p_{x|y}({f_{\mathrm{map}}}(y)|y ) } \right ] } + \sum_{y}p_y(y ) \prod_{t=1}^{{\overline{t } } } \sum_{{\widehat{x}}_t\neq{f_{\mathrm{map}}}(y ) } q_{{\widehat{x}}|y}({\widehat{x}}_t|y ) \notag \\ & = { \mathrm{error}}({f_{\mathrm{map } } } ) + \sum_{y}p_y(y ) { \left [ { 1-q_{{\widehat{x}}|y}({f_{\mathrm{map}}}(y)|y ) } \right]}^{{\overline{t } } } \label{eq : proof - iid}\end{aligned}\ ] ] for a given @xmath12 , let @xmath132 be defined as @xmath133 then ( [ eq : seq - qiid - fmap ] ) is shown as ( [ eq : proof - iid ] ) , which appears on the top of the next page , where the inequality comes from the fact that @xmath134 implies @xmath135 and @xmath136 for all @xmath115 . inequality ( [ eq : seq - iid - fmap ] ) is obtained from ( [ eq : proof - iid ] ) by letting @xmath129 . inequalities ( [ eq : seq - iid - inffmap ] ) and ( [ eq : seq - iid - x ] ) are shown by the fact that @xmath137}^{{\overline{t } } } \notag \\ * & \leq \sum_{y}p_y(y ) \sup_{y : p_y(y)>0 } { \left [ { 1-p_{x|y}({f_{\mathrm{map}}}(y)|y ) } \right]}^{{\overline{t } } } \notag \\ & = \sup_{y : p_y(y)>0 } { \left [ { 1-p_{x|y}({f_{\mathrm{map}}}(y)|y ) } \right]}^{{\overline{t } } } \notag \\ & = { \left [ { 1-\inf_{y : p_y(y)>0}\max_x p_{x|y}(x|y ) } \right]}^{{\overline{t}}}.\end{aligned}\ ] ] when @xmath15 is finite , we have @xmath138}^{{\overline{t } } } \notag \\ & \leq { \mathrm{error}}({f_{\mathrm{map } } } ) + { \left [ { 1-\frac1{|{\mathcal{x}}| } } \right]}^{{\overline{t } } } \label{eq : seq - iid - x}\end{aligned}\ ] ] from ( [ eq : seq - iid - inffmap ] ) , where the last inequality comes from the fact that @xmath139 for all @xmath12 implies @xmath140 we have the same bound ( [ eq : seq - iid - x ] ) from ( [ eq : seq - qiid - fmap ] ) when @xmath141 is the uniform distribution on @xmath2 for every @xmath12 .
this implies that the stochastic decision with an i.i.d .
sequence subject to the a posteriori distribution is at least as good as that subject to the uniform distribution .
it should be noted that , when @xmath15 increases exponentially as the dimension of @xmath2 increases , @xmath127 should also increase exponentially to ensure that the second term @xmath142^{{\overline{t}}}$ ] tends towards zero . by letting @xmath143 in ( [ eq : seq - qiid - fmap ] ) , we have the following corollary .
it should be noted that this corollary includes theorem [ thm : random - fmap ] as the case @xmath143 and @xmath129 .
let @xmath5 be a pair consisting of a state @xmath0 and an observation @xmath1 and @xmath4 be the joint distribution of @xmath5 .
when we make a stochastic decision with a distribution @xmath26 the decision error probability @xmath53 defined by ( [ eq : error - random ] ) satisfies @xmath144,\end{aligned}\ ] ] where @xmath18 is given by ( [ eq : map ] ) .
this section introduces applications of the stochastic decision with the a posteriori distribution to some coding problems . for simplicity
, we assume that we can obtain an ideal random number subject to a given distribution . for
a given function @xmath145 on @xmath146 , let @xmath147 .
for a given @xmath148 , let @xmath149 .
this section introduces a stochastic decoder for a fixed - length lossless compression with an arbitrary small decoding error probability .
let @xmath150 be the probability distribution of @xmath151 and @xmath152 be an encoding map , where @xmath153 is the codeword of @xmath154 .
then the joint distribution @xmath155 of a source @xmath156 and a codeword @xmath157 are given as @xmath158 the decoder receives a codeword @xmath159 . by using a stochastic decoder with the distribution @xmath160 we have the bound of error probability from theorem [ thm : random - fmap ] .
this implies that , when we use the encoding map @xmath145 such that the decoding error probability by using the maximum a posteriori decoder vanishes as @xmath161 , the decoding error probability by using the stochastic decoder with @xmath162 also vanishes as @xmath161 .
in addition , for a special case of source coding with no side information at the decoder @xcite , the fundamental limit @xmath163 is achievable with this code , where @xmath163 is the spectrum sup - entropy rate of @xmath164 ( see @xcite ) .
it should be noted that the right hand side of the above equality is the output distribution of the constrained - random - number generator introduced in @xcite . for a given linear code for an additive noise channel , we can use the constrained - random - number generator as the stochastic decoder by letting @xmath164 be a channel noise , @xmath145 be a parity check matrix , and @xmath165 be the syndrome of @xmath164 .
a channel encoder encodes a message to a channel input @xmath166 .
a channel decoder receives a channel output @xmath167 , reproduces a channel noise @xmath168 from the syndrome @xmath169=a{\boldsymbol{x}}$ ] by using the above scheme , and reproduces a message corresponding to @xmath170 , where the decoding is successful when @xmath168 is reproduced correctly from @xmath171 . from the above discussion , the decoding error probability is at most twice the decoding error probability of the maximal a posteriori decoder .
this section introduces a stochastic decoder for the fixed - length lossless compression of a source @xmath164 with an arbitrary small decoding error probability , where the decoder has access to the side information @xmath172 correlated with @xmath164 .
let @xmath173 be the joint distribution of @xmath174 and @xmath152 be an encoding map , where @xmath153 is the codeword of @xmath154 .
then the joint distribution @xmath175 of a source @xmath156 , side information source @xmath176 , and codeword @xmath157 is given as @xmath177 the decoder receives a codeword @xmath159 and side information @xmath167 . by using a stochastic decoder with the distribution @xmath178
we obtain the bound of error probability from theorem [ thm : random - fmap ] .
this implies that , when we use the encoding map @xmath145 such that the error probability of the maximum a posteriori decoder vanishes as @xmath161 , the error probability of the stochastic decoder with @xmath179 also vanishes as @xmath161 .
in addition , the fundamental limit @xmath180 is achievable with this code @xcite , where @xmath180 is the spectrum sup - entropy rate of @xmath164 ( see ( * ? ? ?
* theorems 4 and 5 ) ) .
it should be noted that the right hand side of ( [ eq : side - information - decoder ] ) is the output distribution of the constrained - random - number generator introduced in @xcite .
this section introduces a stochastic decoder for the channel code introduced in @xcite .
let @xmath156 and @xmath176 be random variables corresponding to a channel input and a channel output , respectively .
let @xmath181 be the conditional probability of the channel and @xmath150 be the distribution of the channel input .
we consider correlated sources @xmath174 with the joint distribution @xmath182 .
let @xmath152 be the source code with the decoder side information introduced in the previous section .
let @xmath175 be the joint distribution defined by ( [ eq : joint - side - information ] ) .
the decoder of this source code obtains the reproduction by using the stochastic decoding with the distribution defined by ( [ eq : side - information - decoder ] ) .
let @xmath183 be the error probability of this source code .
we can assume that for all @xmath184 , @xmath185 and all sufficiently large @xmath186 there is a function @xmath145 such that @xmath187 where a maximum a posteriori decoder is not assumed for this code . when constructing a channel code , we prepare a map @xmath188 and a vector @xmath148 .
we use the stochastic encoder with the distribution @xmath189 for a message @xmath190 generated subject to the uniform distribution on @xmath191 .
it should be noted that the right hand side of the above equality is the output distribution of the constrained - random - number generator introduced in @xcite .
the decoder reproduces @xmath154 satisfying @xmath192 by using the stochastic decoder with the distribution given by ( [ eq : side - information - decoder ] ) and reproduces a message @xmath193 by operating @xmath194 on @xmath195 . in the above channel coding , let us assume that @xmath196 , @xmath197 and the balanced - coloring property @xcite of an ensemble @xmath198 , where @xmath199 is the spectrum inf - entropy rate of @xmath164 and @xmath200 .
then , from ( * ? ? ?
* theorem 1 ) and ( [ eq : errora ] ) , we have the fact that for all @xmath185 and all sufficiently large @xmath186 there are @xmath201 and @xmath148 such that the error probability @xmath202 of this channel code is bounded as @xmath203 it should be noted that the channel capacity @xmath204,\ ] ] which is derived in ( * ? ? ?
* lemma 1 ) , is achievable by letting @xmath161 , @xmath205 , @xmath206 , @xmath207 , and @xmath164 be the general source that attains the supremum . here
, we comment on the decoding with a random sequence introduced in section [ sec : seq ] . for decoding
, we can use random sequences generated by markov chains ( random walks ) that converge to the respective stationary distributions ( [ eq : source - decoder ] ) and ( [ eq : side - information - decoder ] ) . in this case
, we can apply theorem [ thm : seq - random - approx ] to guarantee that the decoding error probability is bounded by twice the error probability of the maximum a posteriori decoding as the sequence becomes longer .
we can also use random sequences by repeating stochastic decisions independently with the respective distributions ( [ eq : source - decoder ] ) and ( [ eq : side - information - decoder ] ) , where we can use independent markov chains with independent initial states to generate an i.i.d . random sequence .
in this case , we can use theorem [ thm : seq - iid ] to guarantee that the decoding error probability tends to the error probability of the maximum a posteriori decoding as the sequence becomes longer .
when implementing ( [ eq : ty ] ) for ( [ eq : source - decoder ] ) and ( [ eq : side - information - decoder ] ) , it is sufficient to calculate the value of the numerator on the right hand side of these qualities because the denominator does not depend on @xmath195 .
the numerator value is easy to calculate when the base probability distribution is memoryless .
this paper investigated stochastic decision and stochastic decoding problems .
it is shown that the error probability of a stochastic decision with the a posteriori distribution is at most twice the error probability of a maximum a posteriori decision .
a stochastic decision with the a posteriori distribution may be sub - optimal but acceptable when the error probability of another decision rule ( e.g. the maximum a posteriori decision rule ) is small .
furthermore , by generating an i.i.d .
random sequence subject to the a posteriori distribution and making a decision that maximizes the a posteriori probability over the sequence , the error probability approaches exponentially the error probability of the maximum a posteriori decision as the sequence becomes longer .
when it is difficult to make the maximum a posteriori decision but the error probability of the decision is small , we may use the stochastic decision rule with the a posteriori distribution as an alternative .
in particular , when the error probability of the maximum a posteriori decoding of source / channel coding tends towards zero as the block length goes to infinity , the error probability of the stochastic decoding with the a posteriori distribution also tends towards zero .
the stochastic decoder with the a posteriori distribution can be considered to be the constrained - random - number generator @xcite implemented by using the sum - product algorithm or the markov chain monte carlo method ( see appendix ) .
however , the trade - off between the computational complexity and the precision of these algorithms is unclear .
it remains a challenge for the future .
in this section , we introduce the algorithms for the constrained - random - number generator @xcite , which generates random numbers subject to a distribution @xmath208 for a given matrix @xmath145 and vectors @xmath209 , @xmath210 , where @xmath211 is assumed to be memoryless , that is , there is @xmath212 such that @xmath213 for all @xmath214 and @xmath215 .
we review the algorithms introduced in @xcite and @xcite , which make use of the sum - product algorithm and the markov chain monte carlo method , respectively .
let @xmath217 be a family of subsets of @xmath218 .
for a set of local functions @xmath219\}_{i=1}^l$ ] , the sum - product algorithm @xcite@xcite calculates a real - valued global function @xmath220 on @xmath2 defined as @xmath221 approximately , where the summation @xmath222 is taken over all @xmath154 except for the variable @xmath223 , and the function @xmath224 depends only on the set of variables @xmath225 .
it should be noted that the algorithm calculates the global function exactly when the corresponding factor graph has no loop .
let @xmath226 and @xmath227 be messages calculated as @xmath228 the summation @xmath229 is taken over all @xmath230 , @xmath231 when there is no @xmath232 such that @xmath233 and @xmath234 when @xmath235 .
the sum - product algorithm is performed by iterating the above operations for every message @xmath227 and @xmath226 satisfying @xmath236 and finally calculating the approximation of the global function as @xmath237 where we assign initial values to @xmath226 and @xmath227 when they appear on the right hand side of the above operations and are undefined . in the following ,
we describe an algorithm for a constrained - random - number generator .
let @xmath238 be an @xmath239 ( sparse ) matrix with a maximum row weight @xmath240 , where the set @xmath241 satisfies @xmath242 for all @xmath243 .
then there is a set @xmath244 such that @xmath245 where @xmath246 is a @xmath247-dimensional vector and @xmath248 denotes the inner product of vectors @xmath246 and @xmath249 .
let @xmath250 , where @xmath251 is a null string if @xmath252 .
let @xmath253 .
calculate the conditional probability distribution @xmath255 defined as @xmath256 where @xmath257 is a support function defined by ( [ eq : support ] ) .
it should be noted that the sum - product algorithm can be employed to obtain ( [ eq : sum - product - fk ] ) , where @xmath258 and @xmath259 are local functions , and we substitute the generated sequence @xmath260 for ( [ eq : sum - product - fk ] ) . if @xmath261 is a constant after the substitution of @xmath260 , we can record the constant in preparation for the future .
this section reviews an algorithm that employs the gibbs sampling @xcite , which is a kind of markov chain monte carlo method . in the following algorithm
, it is assumed that @xmath145 is a systematic matrix illustrated as @xmath266 where the left part @xmath267 is an @xmath268 $ ] matrix and the right part of @xmath145 is the @xmath269 identity matrix .
it should be noted that , when a matrix @xmath145 is not systematic , the elementary transformation and the elimination of redundant rows can be used to obtain an equivalent condition represented by a systematic matrix .
let @xmath270 be the @xmath271-th column of @xmath145 and let @xmath272 .
let @xmath273 be the number of iterations .
in the following , we describe an algorithm for a constrained - random - number generator . it should be noted that steps 2 , 5 , and 7 realize the stochastic decision defined by ( [ eq : ty ] ) and ( [ eq : fy ] ) , which may be skipped by outputting @xmath195 instead of @xmath274 in step 9 .
[ lem : gibbs ] for given @xmath293 , let @xmath294 be the probability of @xmath195 at step 7 in the above algorithm , where we can take an arbitrary initial sequence satisfying @xmath192 at step 1 .
then @xmath295 for all @xmath293 .
s. m. aji and r. j. mceliece , `` the generalized distributive law , '' _ ieee trans .
theory _ , vol .
46 , no . 2 , pp . 325343 ,
. j. o. berger , _ statistical decision theory and bayesian analysis _ , springer - verlag new york inc . , 1985 . t. m. cover , `` estimation by the nearest neighbor rule , '' _ ieee trans . inform .
theory _ , vol .
14 , no . 1 ,
pp . 5055 , jan .
t. m. cover and p. e. hart , `` nearest neighbor pattern classification , '' _ ieee trans .
theory _ , vol . 13 , no . 1 ,
pp . 2127 , jan .
t. m. cover and j. a. thomas , _ elements of information theory 2nd . ed .
_ , john wiley & sons , inc . , 2006 .
han and s. verd , `` approximation theory of output statistics , '' _ ieee trans . inform .
theory _ , vol .
it-39 , no . 3 , pp .
752772 , may 1993 . w. k. hastings , `` monte carlo sampling methods using markov chains and their applications , '' _ biometrica _ , vol .
57 , no . 1 ,
pp . 97109 , 1970 . f. r. kschischang , b. j. frey , and h. a. loeliger , `` factor graphs and the sum - product algorithm , '' _ ieee transactions on information theory _
47 , no . 2 , pp .
498519 , feb .
j. muramatsu , `` channel coding and lossy source coding using a generator of constrained random numbers , '' _ ieee trans . inform .
theory _ , vol .
it-60 , no . 5 , pp . 26672686 , may 2014 . j. muramatsu , `` variable - length lossy source code using a constrained - random - number generator , '' _ ieee trans .
theory _ , vol .
it-61 , no . 6 , pp . 35743592 , jun
j. muramatsu and s. miyake , `` construction of a channel code from an arbitrary source code with decoder side information , '' _ proc .
int . symp . on inform .
theory and its applicat .
_ , moterey , usa , 2016 , pp .
extended version is available at arxiv:1601.05879[cs.it ] , 2016 .
y. steinberg and s. verd , `` channel simulation and conding with side information , '' _ ieee trans .
theory _ , vol .
it-40 , no . 3 , pp .
634646 , may 1994 . | this paper investigates the error probability of a stochastic decision and the way in which it differs from the error probability of an optimal decision , i.e. , the maximum a posteriori decision .
it is shown that the error probability of a stochastic decision with the a posteriori distribution is at most twice the error probability of the maximum a posteriori decision .
furthermore , it is shown that , by generating an independent identically distributed random sequence subject to the a posteriori distribution and making a decision that maximizes the a posteriori probability over the sequence , the error probability approaches exponentially the error probability of the maximum a posteriori decision as the sequence length increases .
using these ideas as a basis , we can construct stochastic decoders for source / channel codes .
channel coding , decision theory , error probability , maximum a posteriori decision , source coding , source coding with decoder side information , stochastic decision , stochastic decoding |
when alexander first defined the multivariable alexander polynomial ( mva ) @xcite it was defined only up to sign and powers of the variables .
it was shown by murakami @xcite that once a normalization and an appropriate change of variables is chosen , the polynomial is of finite type . in @xcite a recursive definition of this weight system
was given , whose computation would take exponential time .
we give a closed form formula using determinants , which would be computable in polynomial time .
this explicit formula makes it possible to verify that the relations from @xcite which hold for the single variable alexander polynomial also hold for the multivariable alexander polynomial .
a coloured link ( or chord diagram ) is a link along with a variable associated to each of the components .
we will use the normalization of the multivariable alexander polynomial given in @xcite .
this uses the matrix given by fox free calculus on the wirtinger presentation of the link , with the crossings labeled by the arc starting at that crossing . to compute the multivariable alexander polynomial one first labels all of the arcs , and then labels the crossings by the exiting lower strand
this allows us to create a matrix with rows and colums indexed by the arcs .
the @xmath0 row is given below . 2 [ cols="^,^,^,^",options="header " , ] then once again we can delete the rows and columns corresponding to @xmath1 , @xmath2 , @xmath3 . but note that there is no change in the @xmath4 column .
so the contribution of the @xmath5 column is lost , this is equivalent to deleting the column associated to the marked chord .
this is the matrix that was described in theorem 1 .
as a test of this method , we wish to show that this function obeys certain known relations , such as the 4 t relation and the recursive relations in murakami s paper .
each of these relations involve chord diagrams that differ only at one small area .
this means that the matrix formed in computing the invariant will differ in only one block .
the following lemma shows that the relations need only hold among certain minors of that block .
let b be an @xmath6 block matrix of the following form , where a is an @xmath7 matrix , and @xmath8 .
@xmath9 where @xmath10 refers to the matrix formed from the @xmath11 columns of a and @xmath12 refers to m with the @xmath13 columns removed .
* proof : * expand b along the rows of a using cofactor expansion , and collect like terms .
+ let @xmath14 be matrices as in the previous lemma , with the same @xmath15 but different @xmath16 . to show a relation of the form @xmath17 it is enough to show the relation holds on the appropriate minors the @xmath16 s .
* proof : * follows from previous lemma .
+ when we wish to show relations between chord diagrams differing in a region , we note that the matrices used to compute their weight will be of the above form . using this lemma
let us show the following relation holds . to use the previous lemma we make both sides of the relation have the same number of arcs , so we divide the arcs in the left term in three .
the matrices involved will be as follows . 2 @xmath18 @xmath19 for the first matrix in the example above the minors in question are as follows ; 2 @xmath20 @xmath21 @xmath22 @xmath23 @xmath24 @xmath25 the corresponding minors for the second matrix are , @xmath26 .
you can see that the first are precisely @xmath27 times those of the second , so the relation holds .
if we cross the chords we get the following two relations .
+ + this allows us to derive the following useful relations .
+ the first relation is a direct consequence of the previous . since it does nt matter how these components connect we can omit the connections from our diagrams .
we can draw elements as in the second relation with the understanding that there will always be an even number of such elements in a diagram .
obviously for more complicated relations we are unable to compute by hand all of the minors involved .
the solution to this is to allow mathematica to do all of the computations for us .
for example here is the computation that shows the 4y relation holds . + using the previous relation we can rewrite the relation as .
+ in the following calculation , we have used the substitution @xmath28 , @xmath29 , @xmath30 and @xmath31 , and a common factor of @xmath32 has been factored from all of the matrices . 2 .... mat[1]= { { 1,1,0,0,0,-1,-1,0,0 } , { -x , y,0,0,0,y ,-
x,0,0 } , { -1,0,1,0,1,0,0,0,-1 } , { 0,0,0,-1,0,0,0,1,0 } , { -w,0,-w,0,y,0,0,0,y } } ; mat[2]= { { w,0,0,0,-y,0,w,0,-y } , { -1,1,1,0,0,-1,0,0,0 } , { x ,- y , x,0,0,-y,0,0,0 } , { 0,0,0,-1,0,0,0,1,0 } , { 1,0,0,0,1,0,-1,0,-1 } } ; mat[3]= { { z,0,0,-w,0,0,0,-w , z } , { -1,1,0,0,1,-1,0,0,0 } , { 0,0,-1,0,0,0,1,0,0 } , { 1,0,0,1,0,0,0,-1,-1 } , { x ,- w,0,0,x ,- w,0,0,0 } } ; mat[4]= { { x ,- w,0,0,0,-w,0,0,x } , { 1,1,0,0,0,-1,0,0,-1 } , { 0,0,-1,0,0,0,1,0,0 } , { -1,0,0,1,1,0,0,-1,0 } , { z,0,0,-w , z,0,0,-w,0 } } ; mat[5]= { { z,0,0,-w,0,0,0,-w , z } , { 0,-1,0,0,0,1,0,0,0 } , { -y,0,w,0,-y,0,w,0,0 } , { 1,0,0,1,0,0,0,-1,-1 } , { -1,0,1,0,1,0,-1,0,0 } } ; mat[6]= { { 1,0,1,0,0,0,-1,0,-1 } , { 0,-1,0,0,0,1,0,0,0 } , { -y,0,w,0,0,0,w,0,-y } , { -1,0,0,1,1,0,0,-1,0 } , { z,0,0,-w , z,0,0,-w,0 } } ; mat[7]= { { 1,1,0,0,0,-1,-1,0,0 } , { -x , y,0,0,0,y ,-
x,0,0 } , { z,0,z ,- y,0,0,0,-y,0 } , { -1,0,1,1,0,0,0,-1,0 } , { 0,0,0,0,-1,0,0,0,1 } } ; mat[8]= { { z,0,0,-y,0,0,z ,- y,0 } , { -1,1,1,0,0,-1,0,0,0 } , { x ,- y , x,0,0,-y,0,0,0 } , { 1,0,0,1,0,0,-1,-1,0 } , { 0,0,0,0,-1,0,0,0,1 } } ; .... .... z*(minors[mat[1],5]-minors[mat[2],5])-y*(minors[mat[3],5]- minors[mat[4],5])+x*(minors[mat[5],5]-minors[mat[6],5])- w*(minors[mat[7],5]-minors[mat[8],5 ] ) ; l = simplify[% ] ; take[transpose[l],70 ] { { 0},{0},{0},{0},{0},{0},{0},{0},{0},{0},{0},{0},{0},{0},{0},{0 } , { 0},{0},{0},{0},{0},{0},{0},{0},{0},{0},{0},{0},{0},{0},{0},{0 } , { 0},{0},{0},{0},{0},{0},{0},{0},{0},{0},{0},{0},{0},{0},{0},{0 } , { 0},{0},{0},{0},{0},{0},{0},{0},{0},{0},{0},{0},{0},{0},{0},{0 } , { 0},{0},{0},{0},{0},{0 } } .... from our lemma , since the relation holds for the minors , it holds for the determinants of the original matrices .
so the relation holds for the weight system .
in @xcite additional relations where shown to hold for the weight system of the alexander polynomial , these can be shown to hold for the mva weight system as well . | we derive a formula for the weight system of the multivariable alexander polynomial using determinants , show that it obeys known relations , and satisfies some of the same relations as the single variable polynomial . |
here we briefly discuss how a given function @xmath88 that vanishes at the origin can be fitted by a polynomial of second degree @xmath89 in a given interval @xmath90 .
we require that the deviation of the function from the polynomial is as lower as it can be , or in other words the integral @xmath91 must be extremized . if we set the first derivatives of @xmath92 with respect to @xmath93 equal to zero @xmath94 we obtain a two - by - two system @xmath95 where we have defined @xmath96 for @xmath97 and @xmath98 . for
the given interval and function the above integrals can be computed , and solving the simple algebraic system we find the values of the coefficients @xmath99 given in the text .
the author acknowledges financial support from fpa2008 - 02878 and generalitat valenciana under the grant prometeo/2008/004 .
e. j. copeland , m. sami and s. tsujikawa , int .
j. mod .
d * 15 * ( 2006 ) 1753 [ arxiv : hep - th/0603057 ] . c. armendariz - picon , t. damour and v. f. mukhanov , phys .
b * 458 * ( 1999 ) 209 [ arxiv : hep - th/9904075 ] . c. armendariz - picon , v. f. mukhanov and p. j. steinhardt , phys .
* 85 * ( 2000 ) 4438 [ arxiv : astro - ph/0004134 ] . c. armendariz - picon , v. f. mukhanov and p. j. steinhardt , phys .
d * 63 * ( 2001 ) 103510 [ arxiv : astro - ph/0006373 ] .
a. bandyopadhyay , d. gangopadhyay and a. moulik , arxiv:1102.3554 [ astro-ph.co ] .
d. w. hogg , arxiv : astro - ph/9905116 .
t. padmanabhan and t. r. choudhury , mon . not .
* 344 * ( 2003 ) 823 [ arxiv : astro - ph/0212573 ] .
a. h. guth , phys .
d * 23 * ( 1981 ) 347 .
d. h. lyth and a. riotto , phys .
* 314 * ( 1999 ) 1 [ arxiv : hep - ph/9807278 ] .
e. komatsu _ et al .
_ [ wmap collaboration ] , astrophys .
j. suppl .
* 192 * ( 2011 ) 18 [ arxiv:1001.4538 [ astro-ph.co ] ] .
y. wang and p. mukherjee , astrophys .
j. * 650 * ( 2006 ) 1 [ arxiv : astro - ph/0604051 ] .
d. j. eisenstein _ et al .
_ [ sdss collaboration ] , astrophys . j. * 633 * ( 2005 ) 560 [ arxiv : astro - ph/0501171 ] . | it is assumed that the current cosmic acceleration is driven by a scalar field , the lagrangian of which is a function of the kinetic term only , and that the luminosity distance is a given function of the red - shift . upon comparison with baryon acoustic oscillations ( baos ) and cosmic microwave background ( cmb ) data the parameters of the models are determined , and then the time evolution of the scalar field is determined by the dynamics using the cosmological equations .
we find that the solution is very different than the corresponding solution when the non - relativistic matter is ignored , and that the universe enters the acceleration era at larger red - shift compared to the standard @xmath0 model .
according to our current understanding about the cosmos , we live in a flat universe which expands in an accelerating rate and it is dominated by a dark component .
identifying the origin and nature of dark energy is one of the biggest challenges for modern cosmology .
since we can only speculate about what dark energy could be , many cosmological models have been proposed and studied so far .
the simplest candidate is a cosmological constant with state parameter @xmath1 , but models with a dynamical component with an evolving state parameter @xmath2 also exist in the literature ( for a review on dark energy models see e.g. @xcite ) .
scalar field models with a non - canonical kinetic term have been discussed in k - inflation @xcite , in which inflation is not due to the potential , but rather to the kinetic term in the lagrangian , and in k - essence models .
@xcite , which are designed to address the issue why the cosmic acceleration has recently begun .
very recently , in @xcite the authors studied a k - essence model assuming i ) a closed form parameterization for the luminosity distance ( for distances in cosmology see e.g. @xcite ) , proposed previously in @xcite , and ii ) that the lagrangian of the scalar field does not depend on the field itself .
however , in that work the matter has been ignored , and in the present article we wish to show that if the presence of matter is also taken into account , the time evolution of the scalar field is very different .
we also find that according to this model , the universe enters the acceleration era at a larger red - shift compared to the standard @xmath0 model .
the theoretical framework is basically the same as in @xcite , apart from the inclusion of matter in the cosmological equations , and the use of baos and cmb data to be mentioned later on .
however , we start by giving a summary of all the ingredients we shall be using , so that the present work is self - contained .
we focus here on the case of a flat universe , since this is a robust prediction of inflation @xcite , and it is also supported by current data @xcite .
the framework is based on four - dimensional general relativity coupled to a single scalar field with a general lagrangian @xmath3 , where @xmath4 is the scalar field and @xmath5 is the standard kinetic term .
therefore , our model is described by the action @xmath6 the energy - momentum tensor for the scalar field is given by @xmath7 where @xmath8 denotes differentiation with respect to @xmath9 .
we can recast this energy - momentum tensor into the form of the energy - momentum tensor for a perfect fluid @xmath10 where @xmath11 are the energy density and the pressure of the fluid respectively .
the hydrodynamical quantities @xmath12 are given in terms of @xmath13 as follows @xmath14 therefore , the equations of motion for the system gravity+scalar field are just the first friedmann equation and the equation for energy conservation @xmath15 where @xmath16 is the hubble parameter , @xmath17 is the scale factor and the overdot denotes differentiation with respect to cosmic time . defining the sound speed of the scalar field
@xmath18 the equation for energy conservation takes the form @xmath19 which generalizes the usual klein - gordon equation of a canonical scalar field in a friedmann - robertson - walker background .
if we now include also the non - relativistic matter , the equations of motion for our system are the friedmann equations and the conservation equation @xmath20 where @xmath21 , @xmath22 is newtons constant , @xmath23 is the energy density of matter ( @xmath24 ) , @xmath25 are the energy density and pressure of dark energy respectively , @xmath26 is the total energy density and @xmath27 is the total pressure . assuming no interaction between dark energy and non - relativistic matter , we have two seperate conservation equations @xmath28 where the second equation for dark energy is equivalent to the equation of motion of the scalar field . if we further assume that the lagrangian of the scalar field depends on the kinetic term only , @xmath29 , the equation of motion can be integrated once @xmath30 where @xmath31 is an arbitrary integration constant , which later on will be taken to be @xmath32 , and @xmath33 is the derivative of @xmath34 with respect to the kinetic term .
all in all , there are three independent equations , namely the equation of the scalar field , the conservation equation for dust , and the second friedmann equation . for a given hubble parameter as a function of the red - shift @xmath35 , the time evolution of the scalar field can be found using the set of the basic cosmological equations .
this will be useful later on ( see equations ( 29 ) and ( 30 ) , as well as the definitions ( 27 ) and ( 28 ) ) .
the comparison of a theoretical model against supernovae data relies on the minimization of @xmath36 with @xmath37 the distance modulus , @xmath38 , where @xmath39 is the absolute magnitude and @xmath40 is the apparent magnitude .
the theoretical apparent magnitude @xmath41 as a function of the red - shift @xmath42 is given by @xmath43 where the luminosity distance @xmath44 for a flat universe can be expressed as @xmath45 and @xmath46 is the speed of light , while @xmath35 is the hubble parameter as a function of the redshift .
we also exploit the cmb shift parameter @xmath47 , since it is the least model dependent quantity extracted from the cmb power spectrum , i.e. it does not depend on the present value of the hubble parameter @xmath48 . the reduced distance @xmath47 is written as @xmath49 where @xmath50 is todays value of the matter normalized density , and we use the cmb shift parameter value @xmath51 , as derived in @xcite .
independent geometrical probes are baryon acoustic oscillations measurements . acoustic oscillations in the photon - baryon plasma are imprinted in the matter distribution .
these baos have been detected in the spatial distribution of galaxies by the sdss @xcite at a redshift @xmath52 .
the sdss team reports its bao measurement in terms of the @xmath53 parameter @xmath54 where the function @xmath55 is defined to be @xmath56 and we have used the value @xmath57 .
we shall assume that the scalar field lagrangian , @xmath29 , is chosen so that the solution of the cosmological equations for the hubble parameter @xmath35 leads to the luminosity distance @xmath58 then the cmb shift parameter as well as the bao @xmath53 parameter , for a given matter density @xmath50 , are certain functions of the two parameters @xmath59 , and if we require that @xmath60 and @xmath61 the two parameters can be determined .
taking @xmath62 we find that the model is viable as long as the two free parameters take the values @xmath63 and furthermore the distance modulus @xmath37 can not be distinguished from the corresponding quantity when the matter is ignored , see figure 1 below . and @xmath64 ( dashed line ) , and b ) our model including matter with @xmath62 , @xmath65 and @xmath66 ( solid line ) .
] to find the time evolution of the scalar field , we introduce the dimensionless scalar field and time as follows @xmath67 where @xmath68 is the age of the universe today , and we use the second friedmann equation .
the dimensionless time @xmath69 as a function of the red - shift is given by @xmath70 while the dimensionless field @xmath71 as a function of the red - shift is given by @xmath72 where @xmath73 , and is given in terms of the assumed ( dimensionless ) luminosity distance @xmath74 by @xmath75 with @xmath59 having the values determined before , and we have used that @xmath76 .
finally , eliminating the red - shift we obtain @xmath71 as a function of @xmath69 , and the solution can be seen in figure 2 below . in the same figure
we also show the polynomial of second degree that best fits the solution @xmath77 in the same interval , which is the following @xmath78 we see that the solution exhibits a global minimum at @xmath79 , which corresponds to a red - shift @xmath80 .
this is the value for which the two terms in ( [ eqn ] ) cancel eachother . in our model including matter with @xmath62 , @xmath65 and @xmath66 ( solid line ) , and the polynomial of second degree @xmath81 that best fits the solution ( dashed line ) . ] before ending our discussion , let us compare the deceleration parameter @xmath82 of the standard @xmath0 model with that of our model .
using the definitions , the deceleration parameter as a function of the red - shift is given by @xmath83 and can be seen in figure 3 , where for @xmath0 we have used the values @xmath84 . in our model the universe enters into the acceleration era at @xmath85 , while in the standard cosmological model this takes place at later times .
model with @xmath86 and @xmath87 ( dashed line ) , and b ) our model including matter with @xmath62 , @xmath65 and @xmath66 ( solid line ) . ] in summary , in the present work we have assumed that the current cosmic acceleration is driven by a scalar field , the lagrangian of which is a function of the kinetic term only , and that the luminosity distance is a given function of the red - shift .
upon comparison with baryon acoustic oscillations ( baos ) and cosmic microwave background ( cmb ) data the parameters of the models are determined , and then the time evolution of the scalar field is obtained by integrating the cosmological equations numerically .
we find that the solution is very different than the corresponding solution when the non - relativistic matter is ignored , and that the universe enters the acceleration era at larger red - shift compared to the standard @xmath0 model . |
we capture in an elongated harmonic trap a total of @xmath5 @xmath6 atoms in each of the lowest and third lowest hyperfine states , labeled @xmath7 and @xmath8 .
the particles interact via the van der waals potential , which at the relevant density and energy scales reduces to a contact interaction characterised by the s - wave scattering length @xmath9 .
the scattering length , controlling the interaction strength , is adjusted by setting a homogeneous magnetic field between 673 and 949gauss , covering the regime from @xmath10 to 0.6 , where @xmath11 is the fermi wavevector in the gas , @xmath12 the mass of @xmath6 atoms and @xmath13 the fermi energy in the harmonic trap , with @xmath14 being the geometric mean of its frequencies .
we reach temperatures of 0.15(2)@xmath15 for the strongest and 0.11(2)@xmath15 for the weakest interactions ( si text ) . for strong interaction ,
the temperature is low enough to access the superfluid regime , as sketched in figure [ fig1]a .
starting from the trapped gas , we first imprint a two - dimensional constriction at the center of the cloud using an off - resonant laser beam operating at a wavelength of 532 nm and shaped in a @xmath16-like mode propagating along the @xmath17 axis and hitting the cloud at its centre .
this separates the elongated cloud into two reservoirs smoothly connected by a quasi - two dimensional region , with a maximum vertical trap frequency along @xmath18 of @xmath19khz .
the quantum point contact itself , depicted in fig .
[ fig1]b , is created by imaging a split gate structure on the two - dimensional ( 2d ) region using high - resolution lithography @xcite .
it is characterized by its transverse trapping frequency at the centre , @xmath20 , which is adjustable .
the gaussian envelopes of the beams ensure a smooth connection to the large three - dimensional reservoirs formed at both ends of the cloud , see fig .
an attractive gate potential @xmath21 is realized by a red - detuned laser beam focused on the qpc @xcite which tunes the chemical potential in the qpc and its immediate vicinity , see fig .
the left ( l ) and right ( r ) reservoirs connected to the qpc contain @xmath22 atoms and have chemical potentials @xmath23 , with @xmath24 = @xmath25 , @xmath26 and @xmath27 = @xmath8 , @xmath7 .
because the gate potential has a waist larger than the qpc , it also increases the density at the entrance and exit points ( minima of the effective potential in fig .
[ fig1]c ) , creating a dimple effect that increases the local degeneracy @xcite .
to measure the particle or spin conductances of the qpc , we prepare either an atom number imbalance @xmath28 or a magnetization imbalance @xmath29 , with @xmath30 . these correspond to a chemical potential bias @xmath31 or a spin bias @xmath32 respectively , see fig .
[ fig1]d - g and materials and methods . in the weakly interacting regime
, we do not find deviations from linear response within our experimental uncertainties @xcite . for particle transport in the strongly interacting regime , @xmath33
, we observe the emergence of non - linearities allowing us to identify the superfluid regime ( see @xcite and si text ) .
the interaction - dependent chemical potentials @xmath23 , and the chemical potential at equilibrium @xmath34 are extracted from the known equation of state of the tunable fermi gas @xcite ( see materials and methods ) .
the biases induce a spin current @xmath35 and a particle current @xmath36 defined as @xmath37 where @xmath38 and @xmath39 are the spin and particle conductances , respectively .
the currents are estimated by measuring the number of particles transferred after @xmath40s and @xmath41s of transport time respectively , and compared to the bias to obtain the conductances ( see materials and methods ) . as a function of the gate potential @xmath21 for different interaction strengths @xmath42 in the reservoirs .
each data point represents the mean over 9 measurements and error bars indicate one standard deviation plotted for every third point .
the thin solid lines are quadratic fits used to identify the maxima in @xmath38 .
inset : @xmath38 obtained from a mean - field phenomenological model , reproducing the non - monotonic behaviour of the experimental data .
* b , * two - dimensional color plot of @xmath38 as a function of @xmath42 , with cuts of fig .
2a indicated as grey dotted lines .
the points where @xmath38 is maximum , obtained from a parabolic fit along @xmath21 , are displayed as orange circles for comparison .
the black dashed line represents the superfluid critical line estimated at the entrance and exit regions of the qpc , using the results of @xcite.,width=302 ]
we first investigate the spin conductance @xmath43 as a function of gate potential @xmath21 , at fixed @xmath44khz , but for different interaction strengths . for non - interacting particles , the strengths of the gate determines the number of transport channels that are open @xcite .
the results are presented in fig .
[ fig3]a and b. for the weakest interactions , we observe the onset of spin transport as the first channel opens around @xmath45k , followed by a continuous increase of @xmath38 up to the largest gate potentials .
the second channel is expected to open around @xmath46k . for intermediate interaction strengths
@xmath47 , we observe a broad maximum in @xmath38 as a function of @xmath21 .
the opening of the channel is still indicated by a sharp increase of @xmath38 at an interaction - independent value of the gate potential . with increasing interactions ,
the centre of the broad maximum in @xmath38 shifts to lower @xmath21 , and its height is reduced . for @xmath48
, @xmath38 vanishes over the entire range of gate potentials .
a complete map of @xmath38 as a function of interaction strength and gate potential is shown in fig .
the existence of a maximum in @xmath38 and the negative spin transconductance @xmath49 for strong interactions indicate the appearance of a spin insulating phase . indeed , increasing @xmath21 increases the chemical potential in and around the qpc , reducing the relative temperature @xmath50 .
the decrease of conductance with decreasing temperature is characteristic of an insulating behaviour @xcite .
we use a mean - field approach to capture the phenomenology of the spin transport .
it assumes that excitations are non - interacting , fermionic bogoliubov quasi - particles .
since the cooper pairs are singlets , these excitations carry the spin current and their populations are controlled by the spin bias .
this allows for a generalisation of the landauer approach to spin conductance ( si text ) .
the predictions are shown in the inset of fig [ fig3]a .
the emergence of a maximum as a function of gate potential is reproduced .
it results from the competition between the non - linearly increasing gap at the entrance and exit of the qpc , hindering spin transport , and the opening of conduction channels .
the position of the maximum along the gate potential axis , and the shape of the conductance variations are reproduced , but the predictions for the value of the conductance differ by about a factor of two . this discrepancy could come from interactions between quasi - particles of opposite spin , neglected in the model , in particular inside the contact where the one dimensional geometry enhances scattering @xcite .
the total particle current is expected to be robust against collisions , since they conserve momentum . yet , large interaction effects are observed .
we measured @xmath39 for several interaction strengths as functions of @xmath21 and @xmath20 .
[ fig2]a shows the curves for fixed gate potential @xmath51k as a function of horizontal confinement @xmath20 . for the weakest interaction strength @xmath52
, @xmath39 shows a distinct plateau at @xmath0 in agreement with the landauer picture . for the tightest horizontal confinement , @xmath53khz ,
the qpc is almost pinched off , while when reducing @xmath20 below 8khz , several transverse modes with closely spaced energies get populated . for interaction strengths @xmath54
, a conductance plateau with a reduced length remains visible in fig .
the height of this feature continuously increases above the universal value , and eventually washes out with increasing interaction strength , leaving a visible shoulder as high as @xmath55 for @xmath56 .
a similar observation is made when varying @xmath21 at fixed @xmath57khz , as shown in fig .
there again , plateaux - like features with conductances higher than @xmath0 are observed for interaction strengths @xmath58 .
as interactions are further increased towards the unitary regime ( @xmath59 for fig .
[ fig2]a and c , @xmath60 for fig . [ fig2]b and d )
, no conductance plateaux can be distinguished , and @xmath39 increases continuously from zero to large values .
contrary to variations of @xmath20 , variations of @xmath21 change the density at the entrance and exit of the qpc , which probably causes the disappearance of the plateau already at a lower value of the interaction strength .
the entire crossover from quantized conductance of weakly interacting atoms to its breakdown for strong interactions is mapped out in fig .
[ fig2]c for fixed @xmath21 and varying @xmath20 , and in fig .
[ fig2]d for fixed @xmath20 and varying @xmath21 .
the latter demonstrates most clearly that the conductance plateau , discernible as green area , shrinks gradually when the interaction strength is increased from @xmath61 to @xmath62 . in this regime
the plateau width is well predicted by a mean - field model accounting for intra- and inter - mode attraction in the qpc ( si text ) .
furthermore , we observe little difference between the unitary and the molecular regime in the experimentally accessible region , @xmath63 , where the reservoirs form a condensate of molecules @xcite . in the strongly interacting regime ( deep purple regions in fig .
[ fig2]c and d ) , deviations from a linear response to the bias are observed @xcite in agreement with our previous measurements for a qpc in a unitary superfluid @xcite ( si text ) . indeed , for the temperature imposed by the reservoirs , increasing @xmath21 or the interactions leads to the onset of superfluidity in the minima of the effective potential ( see fig .
[ fig1]c ) , i.e. at the entrance and exit of the qpc .
the local critical temperature at those points thus corresponds to the maximum critical temperature over the entire cloud , and we refer to it as @xmath64 for the remainder . to extract it
, we use the state - of - the - art calculation of @xmath65 @xcite in local density approximation , with @xmath66 being the fermi energy of a homogeneous gas with density @xmath67 .
we estimate @xmath67 at the entrance and exit of the qpc from the trap geometry and the equation of state of the low - temperature , tunable fermi gas ( si text ) .
the resulting critical line is displayed in fig .
[ fig3]b , [ fig2]c and d. it closely tracks the maxima of the spin conductance in fig .
[ fig3]b , as well as the disappearance of the conductance plateaux in fig .
[ fig2 ] .
we now focus on the conductances in the single mode regime , where universal quantization is observed for weak interactions . for this purpose
, we display in fig .
[ fig5 ] the conductances as a function of @xmath68 , measured at the position of the plateau center in the weakly interacting regime .
these are extracted from fig .
[ fig2]c for fixed @xmath69khz and from fig .
[ fig2]d for fixed @xmath70k .
we observe that the resulting conductances now coincide within error bars .
this demonstrates that @xmath68 is a key control parameter of the transition , despite the fact that the two data sets correspond to different geometries in the single mode regime .
the fast increase of particle conductance coincides with a sharp drop in the spin conductance around @xmath71 , demonstrating directly the intimate connection between pairing and superfluidity .
the regime of non - universal quantisation , with a conductance larger than @xmath0 as identified by our measurement method and accuracy , extends from @xmath72 , corresponding to the shoulder observed at @xmath73 , to far above the superfluid transition , up to @xmath74 .
this suggests that in this regime , where @xmath75 at every point in the cloud , current is not carried by fermionic quasiparticles , challenging a description in terms of fermi liquids .
a possible interpretation for the anomalously high conductance , i.e. exceeding @xmath0 in the normal , single mode regime , is the presence of strong superfluid fluctuations in the reservoirs , due to the large critical region around the superfluid transition @xcite .
the critical fluctuations are qualitatively similar to that of the luttinger liquid in one dimension with attractive interactions in the leads , where they are known to yield an enhanced conductance @xcite .
another possibility are preformed pairs above @xmath64 , that could form in particular in the contact region @xcite , leading to a channel with a bosonic character , where large conductances are expected @xcite .
evidence for such non - fermi liquid behaviour in the bec - bcs crossover was found using photoemission spectroscopy @xcite , in contrast to results based on the equation of state @xcite .
our findings , covering the attractively interacting regime , complement the observations made with repulsively interacting electrons in solid state qpcs .
future work could also explore the known conductance anomalies observed in electronic qpcs @xcite .
( closed circles ) and @xmath38 ( open triangles , every second error bar displayed ) for various interaction strengths are presented as a function of the reduced temperature @xmath68 , which varies due to the dependence of @xmath64 on density and scattering length .
blue data points are obtained from the measurements shown in fig . 2 and figs .
3b , d , for @xmath70k and @xmath57khz .
red data points are obtained from the measurements shown in figs .
3a , c , for @xmath76k and @xmath69khz .
@xmath39 tends to the conductance quantum @xmath0 ( horizontal dash - dotted line ) for weak interactions ( @xmath77 ) .
error bars contain statistical and systematic errors ( see materials and methods).,width=302 ]
interacting fermi gases are produced by a two steps evaporative cooling procedure .
we first create a balanced mixture of the lowest two hyperfine states of @xmath6 , and perform evaporative cooling at a magnetic field of @xmath78 g down to temperatures of the order of the fermi temperature .
a landau - zener radio frequency transition then transfers the full population from the second to the third hyperfine state , and the magnetic field is ramped up to @xmath79 g , the centre of a feshbach resonance @xcite , where the s - wave scattering length diverges .
a second step of forced evaporation is then performed using a magnetic field gradient , yielding the low temperature clouds used for the transport measurements .
the magnetic field is then ramped in @xmath80ms to the desired value between 673 and 949gauss in order to vary the interaction strength during transport . in the absence of the point contact and gate beam ,
the atoms reside in a hybrid trap where the confinement along @xmath17 and @xmath18 is ensured by an optical dipole trap , and along @xmath81 by the residual curvature of the magnetic field .
the trap frequencies along the @xmath17 and @xmath18 directions are @xmath82hz and @xmath83hz respectively .
the trap frequency along @xmath81 ranges between @xmath84hz and @xmath85hz , depending on the value of the magnetic field .
to create a symmetric spin bias @xmath86 between the two reservoirs ( see fig .
[ fig1]e ) , we ramp the magnetic field before the landau - zener transfer in 10ms from 302 g down to 52 g , where the lowest and second lowest hyperfine states have different magnetic moments , and apply at the same time a magnetic field gradient along the transport axis .
this induces dipole oscillations with different frequencies and different amplitudes for the two states .
we wait for roughly one period of the faster oscillation before we abruptly switch on an elliptic repulsive gate laser beam separating the reservoirs @xcite .
the magnetic field is then ramped back to a value close to its initial value , from where we transfer all the atoms in state @xmath87 to state @xmath88 using an adiabatic landau - zener radio frequency transfer .
after evaporation at the feshbach resonance , we obtain an opposite atom number imbalance for the two states , @xmath89 .
we ensured that this preparation scheme does not increase the temperature as compared to the one for the particle transport .
for the computation of the conductance we use of adiabatic approximation @xcite , which allows for a separation of longitudinal ( @xmath81 ) and transverse ( @xmath17 , @xmath18 ) coordinates .
we verified numerically that it is a very good approximation for the geometry of our qpc . in the resulting one - dimensional schrdinger equation the transverse energy @xmath90 acts as an additional potential , with @xmath91 describing the spatial variation of the trapping frequencies of the qpc . together with the gate potential @xmath92 , and the harmonic trapping potential @xmath93 along @xmath81 , and a residual repulsive potential @xmath94 arising from residual light in the nodal line of the intensity profile of the @xmath16-like laser mode creating the 2d confinement @xcite , it forms the effective potential @xmath95 , which is drawn in fig .
the involved envelope functions are listed in table [ envelopefunctions ] .
the prefactor in @xmath96 has been calibrated to @xmath97k using a conductance measurement with only the 2d confinement present .
the central maximum in a generic profile @xmath98 is due to the @xmath17 confinement of the qpc , and the two minima to each side of it are a result of the combined potential of @xmath99 and @xmath21 .
we define the entrance and exit of the qpc as the position of these minima .
they represent the positions of highest density and thus of lowest @xmath100 .
.envelope functions determining the effective potential .
[ cols="^,^,^",options="header " , ] [ envelopefunctions ]
to evaluate the interaction dependent chemical potentials @xmath101 and @xmath102 , compressibility @xmath103 and spin bias @xmath104 we use the equation of state of the two component , homogeneous fermi gas @xmath105 @xcite .
integrating it over the trap provides the thermodynamic potential @xmath106 , with @xmath107 the known trapping potential ( including the qpc region ) and @xmath108 the heaviside function . the particle number @xmath109 and the magnetization @xmath110 in a single reservoir are then given by @xmath111 and @xmath112 with @xmath113 and @xmath114 .
the factors of @xmath115 arise because the size of the two identical reservoirs is half of the entire cloud .
given the measured @xmath109 and @xmath110 , one can solve numerically for @xmath101 and @xmath102 , or equivalently , for @xmath34 and @xmath116 .
the compressibility and spin susceptibility of a single reservoir are given by @xmath117 and @xmath118 , respectively .
the additional factors of 1/2 the definitions of @xmath109 and @xmath110 arise because in our definitions of @xmath103 and @xmath119 we require that @xmath120 for @xmath121 . for a symmetric spin bias , we have @xmath122 .
it ranges from @xmath123 to @xmath124 for interaction strengths @xmath125 , corresponding to a mean value of @xmath126 .
@xmath39 is determined within linear response as in @xcite .
we use the relation @xmath127 , yielding @xmath128 we indeed observe an exponential decay of @xmath129 as a function of time ( except for the deep superfluid regime , see si text ) .
the characteristic time @xmath130 is related to @xmath39 through @xmath131 . to determine @xmath39
, we evaluate @xmath103 ( see above ) , and determine @xmath130 by measuring @xmath129 at @xmath132 and after a transport time of @xmath133 . from the solution of eqn .
( [ diffequ ] ) we obtain @xmath134 @xmath38 is extracted slightly differently : a linear response relation similar to eqn .
( [ diffequ ] ) can not be established for @xmath135 because the magnetic susceptibility depends in a non - linear way on @xmath116 .
in particular , @xmath119 starts close to zero for low @xmath116 due to the superfluid gap .
we define @xmath136 as @xmath137 .
we evaluate @xmath104 from the initial magnetization imbalance @xmath138 ( see above ) and determine @xmath35 from @xmath139 where @xmath140 is the time constant of the observed exponential decay of the magnetization imbalance @xmath141 .
@xmath140 is determined by measuring @xmath142 at @xmath132 and after a transport time of @xmath143 , and evaluating @xmath144
error bars in fig .
[ fig3 ] and [ fig2 ] are statistical and indicate one standard deviation .
error bars in fig .
[ fig5 ] represent the uncorrelated combination of one standard deviation statistical and systematic uncertainties .
the systematic uncertainty in the conductance amounts to 11% .
it represents the uncorrelated combination of the uncertainties in the compressibility , which are due to the calibration error in the total particle number , an uncertainty in the overall trapping potential and an uncertainty due to the use of the zero temperature equation of state .
the statistical error in @xmath68 is due to the determination of @xmath145 and amounts to 10% .
the systematic uncertainty in @xmath68 is mainly due to the uncertainty in our estimate of @xmath64 .
it is caused by the overall uncertainty in the effective potential , which is due to the uncertainties in @xmath146 , @xmath20 , @xmath21 , and their spatial dependencies .
none of these uncertainties could explain the departure of conductance from @xmath0 observed in fig.[fig2 ] and [ fig5 ] .
we thank shuta nakajima for experimental assistance and g. haack , j. blatter , t. giamarchi , j. von delft , l.glazman , n. dupuis and w. zwerger for discussions , and m. landini , p.trm and e. demler for their careful reading of the manuscript and for discussions .
we acknowledge financing from nccr qsit , the erc project sqms , the fp7 project siqs , the horizon2020 project quic , swiss nsf under division ii .
jpb is supported by the ambizione program of the swiss nsf .
10 cirac ji , zoller p ( 2012 ) goals and opportunities in quantum simulation .
_ nat phys _ 8:264266 .
t ( 2010 ) fermi - hubbard physics with atoms in an optical lattice .
_ annual review of condensed matter physics _ 1:129152 .
bloch i , dalibard j , nascimbene s ( 2012 ) quantum simulations with ultracold quantum gases . _ nat phys _ 8:267276 .
van wees bj , et al .
( 1988 ) quantized conductance of point contacts in a two - dimensional electron gas .
_ physical review letters _ 60:848850 .
wharam da , et al .
( 1988 ) one - dimensional transport and the quantisation of the ballistic resistance . _ journal of physics c : solid state physics _ 21:l209 .
krinner s , stadler d , husmann d , brantut jp , esslinger t ( 2015 ) observation of quantized conductance in neutral matter .
_ nature _ 517:6467 .
martin t , landauer r ( 1992 ) wave - packet approach to noise in multichannel mesoscopic systems .
b _ 45:17421755 .
batra ip ( 1998 ) origin of conductance quantization .
_ surface science _ 395:4345 .
hassler f , et al . ( 2008 ) wave - packet formalism of full counting statistics .
b _ 78:165330 .
albert m , haack g , flindt c , bttiker m ( 2012 ) electron waiting times in mesoscopic conductors .
_ 108:186806 .
beenakker cwj , van houten h ( 1991 ) josephson current through a superconducting quantum point contact shorter than the coherence length .
_ 66:30563059 .
bauch t , et al .
( 2005 ) correlated quantization of supercurrent and conductance in a superconducting quantum point contact .
b _ 71:174502 . kleinsasser
aw , miller re , mallison wh , arnold gb ( 1994 ) observation of multiple andreev reflections in superconducting tunnel junctions .
_ 72:17381741 .
goffman mf , et al .
( 2000 ) supercurrent in atomic point contacts and andreev states .
_ 85:170173 .
bretheau l , girit co , pothier h , esteve d , urbina c ( 2013 ) exciting andreev pairs in a superconducting atomic contact .
_ nature _ 499:312315 .
thomas kj , et al .
( 1996 ) possible spin polarization in a one - dimensional electron gas . _
_ 77:135138 .
cronenwett s ( 2002 ) low - temperature fate of the 0.7 structure in a point contact : a kondo - like correlated state in an open system " .
; physical review letters _ 88 .
matveev ka ( 2004 ) conductance of a quantum wire in the wigner - crystal regime .
_ 92:106801 .
bauer f , et al .
( 2013 ) microscopic origin of the 0.7-anomaly in quantum point contacts .
_ nature _ 501:7378 .
iqbal mj , et al .
( 2013 ) odd and even kondo effects from emergent localization in quantum point contacts . _
nature _ 501:7983 .
mink mp , et al .
( 2012 ) spin transport in a unitary fermi gas close to the bcs transition .
_ physical review a _ 86:063631 .
sommer a , ku m , roati g , zwierlein mw ( 2011 ) universal spin transport in a strongly interacting fermi gas .
_ nature _ 472:201204 .
koschorreck m , pertot d , vogt e , khl m ( 2013 ) universal spin dynamics in two - dimensional fermi gases .
_ nature physics _ 9:405409 .
bardon ab , et al .
( 2014 ) transverse demagnetization dynamics of a unitary fermi gas . _ science _ 344:722724 .
safi i , schulz hj ( 1995 ) transport in an inhomogeneous interacting one - dimensional system .
b _ 52:r17040r17043 . ponomarenko vv ( 1995 ) renormalization of the one - dimensional conductance in the luttinger - liquid model .
b _ 52:r8666r8667 .
maslov dl , stone m ( 1995 ) landauer conductance of luttinger liquids with leads .
_ physical review b _ 52:r5539r5542 .
zwerger w ( 2011 ) _ the bcs - bec crossover and the unitary fermi gas _ ( springer science & business media ) vol .
836 sagi y , drake te , paudel r , chapurin r , jin ds ( 2015 ) breakdown of the fermi liquid description for strongly interacting fermions . _ physical review letters _ 114:075301 . stamper - kurn dm , et al .
( 1998 ) reversible formation of a bose - einstein condensate .
_ 81:21942197 .
husmann d , et al .
( 2015 ) connecting strongly correlated superfluids by a quantum point contact
. _ arxiv e - print _
p 1508.00578 .
navon n , nascimbne s , chevy f , salomon c ( 2010 ) the equation of state of a low - temperature fermi gas with tunable interactions . _ science _ 328:729732 .
gebhard f ( 2003 ) _ the mott metal - insulator transition : models and methods _ ( springer ) vol . 137 .
olshanii m ( 1998 ) atomic scattering in the presence of an external confinement and a gas of impenetrable bosons .
_ physical review letters _ 81:938941 .
giorgini s , pitaevskii lp , stringari s ( 2008 ) theory of ultracold atomic fermi gases .
_ 80:12151274 .
jendrzejewski f , et al .
( 2014 ) resistive flow in a weakly interacting bose - einstein condensate .
_ 113:045305 .
labouvie r , santra b , heun s , wimberger s , ott h ( 2015 ) negative differential conductivity in an interacting quantum gas .
_ 115:050601 .
haussmann r , rantner w , cerrito s , zwerger w ( 2007 ) thermodynamics of the bcs - bec crossover .
_ physical review a _ 75:023610 .
taylor e ( 2009 ) critical behavior in trapped strongly interacting fermi gases .
a _ 80:023612 .
debelhoir t , dupuis n ( 2015 ) critical region of the superfluid transition in the bcs - bec crossover .
_ arxiv e - print _
p 1507.02818 .
moritz h , stferle t , gnter k , khl m , esslinger t ( 2005 ) confinement induced molecules in a 1d fermi gas .
_ 94:210401 .
papoular d , pitaevskii l , stringari s ( 2015 ) quantized conductance with bosonic atoms .
lee j , et al .
( 2015 ) contact resistance and phase slips in mesoscopic superfluid atom transport .
nascimbne s , et al .
( 2011 ) fermi - liquid behavior of the normal phase of a strongly interacting gas of cold atoms .
_ 106:215303 .
ku mjh , sommer at , cheuk lw , zwierlein mw ( 2012 ) revealing the superfluid lambda transition in the universal thermodynamics of a unitary fermi gas .
_ science _ 335:563567 .
zrn g , et al .
( 2013 ) precise characterization of li6 feshbach resonances using trap - sideband - resolved rf spectroscopy of weakly bound molecules .
_ physical review letters _ 110:135301 .
brantut jp , et al .
( 2013 ) a thermoelectric heat engine with ultracold atoms .
_ science _ 342:713715 .
glazman li , lesovik gb , khmelnitski de , shekhter ri ( 1988 ) reflectionless quantum transport and fundamental ballistic - resistance steps in microscopic constrictions . _ soviet journal of experimental and theoretical physics letters _ 48:238 .
ihn t ( 2010 ) _ semiconductor nanostructures _
( oxford university press ) .
brantut jp , meineke j , stadler d , krinner s , esslinger t ( 2012 ) conduction of ultracold fermions through a mesoscopic channel .
_ science _ 337:10691071 .
for the temperature measurement we ramp the magnetic field back to 689gauss and use the virial theorem valid for a unitary gas @xcite to determine the internal energy from the second moment of the density distribution .
the temperature of the unitary gas is obtained from the internal energy via the known equation of state @xcite . to trace back the temperature in the bcs regime ,
we suppose that the entropy remains constant during the magnetic field sweep , which is based on our observation that the temperature of the unitary gas is independent of the interaction strength set during transport .
we thus set the entropy of the unitary gas , also extracted with the help of the equation of state , equal to the entropy of a gas in the bcs regime , which is given by @xcite @xmath147 temperatures extracted in this way are shown in fig .
[ fig : temperatures]a .
the solid black line is obtained from eqn .
( [ eqn : entropy ] ) using the extracted entropy @xmath148 of the unitary gas .
it corresponds to @xmath149 indicated by the black square in the graph .
since eqn .
( [ eqn : entropy ] ) contains interaction effects only to lowest order , we linearly interpolate the temperature for interaction strengths @xmath150 ( green dashed line in fig . [
fig : temperatures]a ) between the prediction of eqn .
( [ eqn : entropy ] ) for @xmath151 and the measured value at unitarity . here , the upper limit of the validity range of eqn .
( [ eqn : entropy ] ) , @xmath152 , has been chosen such that the equation predicts a value of @xmath153 which is maximally by 20% larger than its value for @xmath154 .
black crosses in fig .
[ fig : temperatures]a mark the interaction strengths that have been sampled in the experiment .
the adiabatic sweep from the unitary regime to the bcs regime reduces @xmath153 by @xmath155 at the weakest value of the interaction strength in the bcs regime , which is @xmath156 . for @xmath157
the temperature has been independently estimated from a degenerate fermi gas fit to the 2d density profile which was acquired by keeping the magnetic field for imaging at the same value as for the transport process .
it yielded @xmath158 in perfect agreement with the value deduced from the entropy of the unitary gas .
the temperatures stated so far correspond to the temperatures at the end of the experimental sequence where the absorption pictures are taken . since in our experimental sequence
some time elapses between the transport process and the imaging , we have to correct those temperatures for heating occurring in this time interval ( typically @xmath159nk ) in order to get the actual temperatures during the transport process .
they are plotted as blue crosses in fig .
[ fig : temperatures]a .
the absolute temperatures are shown in the inset .
since predictions for the critical temperature in the regime between bcs and unitarity are usually made for uniform gases , we convert @xmath153 into @xmath100 and @xmath160 into @xmath161 ( see main text ) . for their computation
we need to know the density at the trap centre ( in the absence of the qpc ) , which we estimate from the known zero temperature equation of state ( see above ) .
the resulting values are plotted as blue crosses in fig.[fig : temperatures]b as a function of @xmath161 .
compared to fig.[fig : temperatures]a these temperatures are shifted towards lower values because @xmath162 increases with density , which increases due to attractive interactions .
we now turn to the local @xmath100 and @xmath163 at entrance and exit of the qpc .
we obtain them by calculating the local density at those points using the known effective potential and the equation of state .
orange crosses show those values for the parameters for which the centre of the conductance plateau is reached in fig .
3b and d , i.e. for @xmath164k and @xmath57khz .
note that in the evaluation of the density , we account for the zero - point energy of the 2d confinement by subtracting @xmath165 from the local chemical potential . while this is correct for the weakly attractive case , this correction
is probably overestimated in the strongly interacting case @xcite .
for the parameters for which the centre of the conductance plateau is reached in the data of fig .
3a and c , i.e. for @xmath166k and @xmath69khz , the effective potential at entrance and exit of the qpc turns out to be close to zero , and hence the corresponding @xmath100 and @xmath163 are represented by the blue crosses the dashed black line in fig.[fig : temperatures]b is the ratio of critical temperature to fermi temperature as predicted by @xcite .
the dash - dotted line is the prediction of bcs theory including the gorkov and melik - barkhudarov corrections @xcite , @xmath167 $ ] .
interactions between the two spin components in the reservoirs lead to a diffusive spin transport in the reservoirs . here
we estimate the effect of this spin diffusion on the measured spin conductance , by modelling the system as consisting of the qpc in series with the reservoirs , to which we assign a spin resistance @xmath168 .
the measured conductance is then @xmath169 to estimate @xmath170 , we use the drude model , stating that the conductivity is @xmath171 , with the density @xmath67 and the scattering time @xmath172 . in the latter expression
, @xmath173 is the scattering cross section , @xmath174 is the fermi velocity and the factor @xmath175 accounts for pauli suppression of scattering at low temperature . considering the reservoir as a box with section @xmath176 and length @xmath25 , the series resistance @xmath168 can be written as @xmath177 . to facilitate comparison with @xmath178 , we introduce a trap frequency @xmath179 corresponding to the level spacing in the qpc , and the corresponding harmonic length @xmath180 .
after simple algebra , we obtain @xmath181 where we have replaced the fermi energy by @xmath182 , which is the sum of the reservoirs chemical potential and the gate potential .
note that even though the expression formally contains @xmath179 , it is actually independent of the qpc trap frequency . taking reasonable parameters for the experiment , @xmath183 m , a scattering length of @xmath184a@xmath185 , a temperature @xmath186 and a gate potential such that @xmath187 , we obtain @xmath188 .
we thus expect the effects of the spin resistance of the reservoirs to be negligible for a qpc in the single mode regime , and we disregard @xmath189 in the interpretation of the data . for the unitary fermi gas , the scaling is expected to be different since @xmath173 is inversely proportional to the fermi wavelength rather than a constant . in this regime , we can estimate the scattering time @xmath190 from dimensional analysis and accounting for the pauli suppression of scattering . taking the same estimates for the chosen geometry yields an even lower estimate for the series resistance .
in the superfluid phase with spin imbalanced populations , the mean - field hamiltonian at each point @xmath81 along the transport direction ( see fig . 2c in the main text ) reads @xcite : @xmath191 with @xmath192 the bogoliubov quasiparticle with momentum @xmath193 and spin @xmath194 , and a dispersion relation @xmath195 . since the paired particles carry no spin , the only contribution to spin transport originates from the excitations @xmath196 on top of the superfluid background .
then , spin transport can be viewed as the scattering of the bogoliubov particles generated in the reservoirs located at @xmath197 ( with an energy just above the gap @xmath198 in the reservoirs ) through a potential barrier representing the space - dependent spin gap @xmath199 .
this approach is valid for the spin current because it is entirely carried by the normal fraction of the gas , in contrast to the particle current which has a genuine superfluid contribution . assuming local thermodynamic equilibrium and applying the wkb approximation , the probability for a bogoliubov quasiparticle to be scattered from one reservoir to the other is given by : @xmath200}\right|^2\,.\ ] ] the spin current is obtained as the sum of the contributions from both branches of the spectrum of bogoliubov quasiparticles : @xmath201 with the transport function @xmath202 , which in the non - interacting limit counts the number of transport channels available for a particle having an energy @xmath203 . in , @xmath204}}$ ] is the distribution of bogoliubov particles with spin @xmath27 in the left ( right ) reservoir .
it immediately shows that the population of excitations carrying spin is driven by the spin potentials @xmath205 .
+ within linear response , and in a configuration such that @xmath206 ( symmetric spin bias ) , one can rewrite the spin current as : @xmath207\,.\end{aligned}\ ] ] inserting in the expression of the transport function gives the following expression for the spin conductance : @xmath208 } } \\ & + \int_{-\mu}^{-\delta_\text{res } } d\varepsilon \frac{\mathcal{t}(\varepsilon , v_g)\vartheta(\varepsilon-(e_n - v_g-\mu))}{4 k_b t \cosh^2{\left[\frac{\varepsilon}{2k_bt}\right]}}\biggr],\end{aligned}\ ] ] which in the non - interacting regime reduces to the usual expression for the conductance of a qpc . the @xmath108-function in eqn .
tends to increase the conductance as @xmath21 is raised , whereas the term @xmath209 is responsible for an exponential suppression of @xmath38 with increasing @xmath21 because the spin gap is proportional to the local fermi energy . for quantitative comparison ,
we take the following expression of the superfluid gap @xcite : @xmath210}\,.\ ] ]
as a function of gate potential @xmath21 and interaction strength in the reservoirs .
points where @xmath211 , corresponding to a pinched - off qpc , are not shown .
the dashed line indicates the expected superfluid transition.,width=336 ] for a mixture of two interacting spin components , the currents for individual spin components are given within linear response by @xmath212 where @xmath35 is the current of spin @xmath27 atoms and @xmath213 is a coefficient describing spin drag @xcite .
in particular , the previous relations yield @xmath214\delta\mu_\uparrow+\frac{\gamma}{g_\downarrow}i_\downarrow\ ] ] the quantity @xmath215 can indeed be interpreted as the fraction of particles of spin @xmath8 dragged by the flow of particles of spin @xmath7 .
+ in the absence of a global spin polarisation we have @xmath216 , and eqn .
( [ eq : spindrag ] ) is diagonalised in the particle - spin basis .
we obtain @xmath217 with @xmath218 we characterise the spin drag by the ratio @xmath219 . in the non - interacting regime
, @xmath213 is zero and so is @xmath220 , while for a maximally correlated flow , @xmath38 is zero and @xmath221 .
note that for magnetic insulators such as produced with repulsive fermions in optical lattices , @xmath220 would be close to @xmath222 .
we use the data of fig .
2b and fig .
3d to extract @xmath220 .
the result is plotted as a function of interaction strength and gate potential in fig .
[ fig4 ] . in the regime of high gate potential and strong interactions ,
@xmath220 is equal to one : the spin channel is entirely closed by pairing and particle transport proceeds only via singlet pairs . for weaker interactions or lower gate potentials , @xmath220 decreases to about @xmath223 but remains significantly larger than zero .
the non - zero value of @xmath220 illustrates the fact that contrary to the particle transport , spin transport turns diffusive and is reduced by interparticle scattering , as collisions transfer momentum between the two spin components . in one - dimensional systems ,
this process is responsible for spin - charge separation .
is obtained from a measurement with a weakly interacting fermi gas at a magnetic field of @xmath224 g .
inset : zoom on the interaction strength studied in this work.,width=336 ] for the data set of fig .
3b and d we extract the plateau width of those conductance curves , for which we observe clear plateaux or a remaining bending of the curve , i.e. for interaction strengths @xmath225 .
we do so by fitting the sum of two sigmoid functions having equal steepness and amplitude .
[ fig : plateauwidth]a shows the fitted plateau widths as a function of @xmath42 and @xmath226 .
the data point to the very left is extracted from a reference measurement using a weakly interacting fermi gas prepared on the other side of the feshbach resonance at a magnetic field of @xmath224 g , where @xmath227 .
the data is in good agreement with a mean - field model of the qpc ( solid red line in fig .
[ fig : plateauwidth ] ) , which includes intra- and inter - mode interactions on a mean field level to determine the occupation of the transverse modes in a self - consistent way .
the conductance is then calculated using the landauer formula and the resulting theory curves are fitted with the same sigmoid fit functions as the data .
attraction between particles of the ground state and the first excited mode leads to an occupation of the latter at an already lower gate potential , and thus to a decreased plateau width .
intra - mode interactions on the other hand increase the steepness of the conductance rise to the plateau values .
this effect however is rather small and not resolved in our measurements .
for a fixed gate potential and horizontal confinement , @xmath228k and @xmath53khz , corresponding to the centre of the conductance plateaux of fig .
3b and d , we measure the decay of the initial particle imbalance as a function of time .
a linear relation between the current and the chemical potential bias implies an exponential decay , like for the discharge of a capacitor in a rc circuit .
deviations from exponential behaviour signal a breakdown of linear response and appear in the deep superfluid regime @xcite .
[ fig : decay ] shows a set of experimental decay curves , together with an exponential least - square fit . for a wide range of interaction strengths , @xmath229 ,
the data is adequately fitted by an exponential . for stronger interactions , systematic deviations from an exponential appear .
we assess the quality of the exponential fit using the reduced @xmath230 obtained from the least square method .
[ fig : decay]b presents the evolution of @xmath230 as a function of interaction strength .
it shows a sharp increase for @xmath231 , which coincides with the transition to the deep purple region in fig . 3c . in this regard , the meaning of @xmath39 as calculated through eqn .
( 8) and ( 9 ) in materials and methods is less straightforward . eqn .
[ 9 ] can be rewritten as : @xmath232 using the thermodynamic relation @xmath127 and the definition of the current in eqn .
( [ eq : linearresp ] ) , eqn .
( [ eqn : estimatedgn ] ) reads @xmath233 the estimated conductance appears as a time average of the particle current - to - bias ratio . assuming that @xmath36 is a concave function of @xmath3 ( as observed in @xcite at unitarity ) , @xmath39 is therefore a lower bound for the linear conductance defined as @xmath234 . | we study particle and spin transport in a single mode quantum point contact using a charge neutral , quantum degenerate fermi gas with tunable , attractive interactions .
this yields the spin and particle conductance of the point contact as a function of chemical potential or confinement .
the measurements cover a regime from weak attraction , where quantized conductance is observed , to the resonantly interacting superfluid .
spin conductance exhibits a broad maximum when varying the chemical potential at moderate interactions , which signals the emergence of cooper pairing .
in contrast , the particle conductance is unexpectedly enhanced even before the gas is expected to turn into a superfluid , continuously rising from the plateau at @xmath0 for weak interactions to plateaux - like features at non - universal values as high as @xmath1 for intermediate interactions . for strong interactions , the particle conductance plateaux disappear and
the spin conductance gets suppressed , confirming the spin - insulating character of a superfluid .
our observations document the breakdown of universal conductance quantization as many - body correlations appear .
the observed anomalous quantization challenges a fermi liquid description of the normal phase , shedding new light on the nature of the strongly attractive fermi gases .
= 1 quantum gas experiments provide a tool to study fundamental concepts in physics , which may be hard to access by other means .
challenges such as the interplay and dynamics of many interacting fermions , are addressed by interrogating a specifically tailored quantum many - body system with controlled parameters , an approach referred to as quantum simulation @xcite .
the outcomes can then be used to benchmark theory , or even as a direct comparison to different experimental realizations of the same concept . during the last years
there has been substantial progress on this path using cold atomic gases to realize important models of condensed matter physics , formulated to describe the bulk properties of materials @xcite . here ,
neutral fermionic atoms are used to model the electrons in a solid . in this article
we use a quantum gas to study the operation of an entire mesoscopic device , a quantum point contact ( qpc ) , in the presence of interactions between the particles .
we observe the transport of a charge neutral quantum degenerate gas of fermionic lithium atoms , which can be prepared in a mixture of two hyperfine states .
these states provide a spin degree of freedom and the attractive interaction between them can be tuned continuously from weak to unitary , a feature unique to cold atomic gases .
the qpc itself is realized by a suitably shaped optical potential , which consists of a short , one - dimensional channel connected to two large reservoirs @xcite .
biasing the reservoirs with different chemical potentials can induce a dc current .
the ratio of the current to the bias is the conductance of the contact , which is independent of the bias in the linear response regime . for spinless
, non - interacting particles at low temperature , the conductance is quantized in units of @xmath0 , the universal conductance quantum for neutral particles @xcite .
an intuitive understanding thereof can be gained by considering the temporal spacing @xmath2 of minimum uncertainty wave packets within one transverse mode , and moving through the channel in response to an applied bias @xmath3 @xcite . at zero temperature , within one energetically available mode
, each wavepacket state is occupied by a single particle in accordance to pauli s principle .
hence the maximum current carried by the mode is @xmath4 with @xmath0 being the upper bound for the contribution of a single mode to conductance .
this bound , set by heisenberg s and pauli s principle , holds for all systems where transport proceeds by fermionic quasiparticles , such as fermi liquids . in a gas with two spin components ,
the conductance of each component is quantized in units of @xmath0 , as long as there are no interactions between them .
the picture becomes more intricate if collisions between the two components play a role .
studies conducted on solid - state systems have shown that the weak correlations in conventional superconductors yields quantized supercurrents @xcite and the emergence of andreev bound states in mesoscopic conductors@xcite .
besides , in semiconductor systems , the coulomb repulsion between charge carriers subtly modifies the conductance quantization @xcite , lowering conductance at low density and leading to strong correlations . to characterize the transport , one may additionally consider the spin current , i.e. the relative current of one component with respect to the other , which is expected to be damped in the presence of interactions @xcite . while measuring and inducing spin currents in clean mesoscopic conductors is a challenge in solid state systems , quantum gases naturally allows for spin resolved observations and manipulations .
for example the damping of spin currents was measured in strongly interacting fermi gases at high temperature , where many - body effects , in particular pairing , are weak @xcite . in contrast , the total particle current is conserved in collisions , thus the quantization of particle conductance in the channel should be robust .
in fact , it was shown , that the applicability of a fermi liquid description in the leads of the contacts guarantees universal conductance quantization @xcite , regardless of the interaction strength .
our study of spin and particle conductance of a quantum point contact with tunable interactions uses an atomic fermi gas in the vicinity of a broad feshbach resonance , realising the bec - bcs crossover regime @xcite .
it features a conventional s - wave paired superfluid for strong attraction and low temperatures , as well as a fermi - liquid phase for weak attraction and high temperatures .
the nature of the state in the intermediate regime remains controversial , as it is governed by a non - trivial interplay of pairing and superfluid fluctuations competing with finite temperature properties of the gas @xcite .
this richness makes our system an ideal test bed to study how transport properties change with interactions . on the one hand ,
a spin - insulating character , not accessible to high temperature measurements @xcite , should emerge as a result of s - wave pairing .
on the other hand , particle transport directly tests the robustness of conductance quantization as many - body correlations emerge . |
we start by describing the _ 321,hexagon avoiding _ permutations in two ways ; first in the context of pattern avoidance , which is the viewpoint we will be exploiting in our enumeration , and then in the context of reduced expressions , which explains the introduction of the term 321,hexagon avoiding ( for brevity , _ 321hex _ ) .
finally , we explain briefly the connection with kazhdan lusztig polynomials in @xcite and how this motivates the work on the present paper . from the first viewpoint
, we consider permutations in @xmath0 as bijections @xmath1 \rightarrow [ n]$ ] , and write them in one line notation as the image of @xmath2 , @xmath3 $ ] .
for @xmath4 we suppress the commas without causing confusion .
let @xmath5 and @xmath6 for some @xmath7 .
we say that @xmath2 _ contains _
@xmath8 if there is a sequence @xmath9 such that the sequences @xmath10 $ ] and @xmath11 $ ] obey the same pairwise relations , i.e. @xmath12 exactly when @xmath13 . in such a case , we write @xmath14 .
if @xmath2 does not contain @xmath8 then we say that @xmath2 _ avoids _ @xmath8 .
we denote by @xmath15 the set of all @xmath8avoiding permutations of length @xmath16 .
for example , the permutation @xmath17 avoids @xmath18 but does not avoid @xmath19 because of its subsequence @xmath20 . for a classification of forbidden subsequences up to length 7 , we direct the reader to @xcite .
the _ 321hex _ permutations are those permutations which simultaneously avoid each of the following five patterns : @xmath21,\,p_1=[46718235],\,p_2=[46781235],\,p_3=[56718234],\,p_4=[56781234].\ ] ] we denote by @xmath22 the set of the four length8 ( _ octagonal _ ) permutations @xmath23 . in order to make sense of the above definition , consider the following equivalent , but perhaps more insightful , reformulation in terms of matrices .
let @xmath6 .
the _ permutation matrix _
@xmath24 is the @xmath25 matrix having a @xmath26 in position @xmath27 for @xmath28 , and @xmath29 elsewhere .
( to keep the resemblance with the `` shape '' of @xmath2 , we coordinatize @xmath24 from the bottom left corner . ) given two permutation matrices @xmath30 and @xmath31 , we say that @xmath30 _ avoids _
@xmath31 if no submatrix of @xmath30 is identical to @xmath31 . @xmath32 a permutation matrix @xmath30 of size @xmath16 is simply a transversal of an @xmath25 matrix
. clearly , @xmath6 contains @xmath5 if and only if @xmath24 contains @xmath33 as a submatrix . under this reformulation ,
[ octagonal ] presents the four octagonal patterns @xmath34 which must be avoided by the 321hex permutations .
the fifth pattern @xmath35 does _ not _ come from a permutation because it is not a transversal .
yet , @xmath35 is a union of the four previous permutation patterns , and it can be easily checked that a permutation matrix @xmath24 avoids all @xmath36 s if and only if no @xmath37 permutation submatrix of @xmath24 can be completely `` covered '' by @xmath35 .
thus , by abuse of notation , we can say that the 321hex permutations are defined as the permutations avoiding both @xmath38 $ ] and @xmath35 .
our enumeration makes use of this last interpretation , exploiting the symmetries in the set @xmath22 of octagonal patterns plus a convenient structural representation of all 321avoiding permutations .
the close relations between the four octagonal permutations are even more clearly revealed from a group theoretical viewpoint when examining their reduced expressions .
below , we briefly review the construction of the _ heaps _ of 321avoiding permutations and their relation to the octagonal patterns @xmath36 ; for more details , see @xcite .
the remainder of the introduction , except for its conclusion , can be skipped by the reader who is interested only in the pattern avoidance interpretation .
the permutation group @xmath0 can be regarded as generated by the set of the _ adjacent transpositions _
@xmath39 , where @xmath40 in _ cyclic _ notation . in this presentation , the generators @xmath41 and @xmath42 commute if @xmath43 ; else @xmath44 .
an _ expression _ is any product of generators .
a _ reduced _ expression @xmath45 for @xmath46 is a shortest
possible expression yielding @xmath2 .
( it is well known that the number of generators in @xmath45 equals the number of inversions in @xmath2 . ) for example , the octagonal pattern @xmath47 $ ] has a reduced expression @xmath48 reduced expressions for the other avoided octagonal patterns are @xmath49 , @xmath50 and @xmath51 .
these expressions can be easily verified by considering the dashed lines in @xmath35 in fig .
[ octagonal ] .
the study of reduced expressions is a major subject of the representation theory of @xmath0 .
after work of viennot @xcite , the @xmath52avoiding permutations @xmath2 can be represented by special ranked posets called _
the elements of @xmath53 are identified with the transpositions @xmath54 in a fixed reduced expression @xmath55 for @xmath2 . by billey jockusch
stanley @xcite , the @xmath52avoiding permutations are those in which no reduced expression contains a substring of the form @xmath56 ; and by tits @xcite , all reduced expressions for a @xmath52avoiding permutation @xmath2 are equivalent up to moves @xmath57 for @xmath43 .
thus , the set of elements in @xmath53 is independent of the choice of @xmath55 .
@xmath58 we now describe the _ rank function _ of @xmath59 , along with a hasse diagram for the poset by embedding its elements in the integer lattice .
one way to define and visualize this embedding is via _ string diagrams_. to form a string diagram of @xmath46 , write the row of numbers @xmath60 $ ] above the row @xmath61 $ ] , thus mimicking the two line notation for a permutation .
connect each number @xmath62 on the bottom line to the corresponding number @xmath62 on the top line , drawing a `` string '' which may change direction , but which at all times is running either due north , northwest or northeast .
strings may cross and recross , but do not run over top of one another , nor do they stray beyond the rectangular bounds formed by the two rows of numbers . for example
, @xmath63 can be realized as the crossing of two strings , as shown in fig .
[ stringdiagram]a . in general ,
two or more adjacent transpositions can be applied simultaneously , provided they commute , i.e. @xmath41 and @xmath42 can occur on the same horizontal level of a string diagram unless @xmath64 .
thus , the crossings of a string diagram can be identified with the @xmath41 s , labelled by the column ( from @xmath26 to @xmath65 ) in which they occur , and can be seen to form a poset in the obvious way .
the linear extensions of this partial order are the expressions for @xmath2 defined above .
if the string diagram has the smallest possible number of crossings , then it is is _ minimal _ and the linear extension is a reduced expression for @xmath2 .
a _ short braid _ is a configuration obtained by applying the non commuting transpositions @xmath41 and @xmath66 in the following order : @xmath67 or @xmath68 ( cf .
[ stringdiagram]b[stringdiagram]c ) .
the string diagram for a braid shows crossings at three of the four points of a small diamond , omitting either the eastern or the western point .
as mentioned earlier , a permutation is 321avoiding exactly when its minimal string diagrams avoid such configurations ; such permutations are therefore also called _ short braid avoiding _ in the literature of coxeter groups .
it is now clear how to embed canonically the poset of string crossings in a minimal string diagram of @xmath2 into the integer lattice .
the resulting hasse diagram is the heap of @xmath2 , @xmath53 .
it is independent of the choice of @xmath69 as long as @xmath2 is 321avoiding .
the heap of the special reduced expression @xmath70 for the octagonal pattern @xmath71 resembles a hexagon : @xmath72 has horizontal and vertical symmetries , with respectively 2,3,4,3,2 lattice points on its five ranks ( see fig . [
heap]a[heap]b ) .
the string diagram of @xmath73 features one extra point on top and one extra point on bottom , corresponding to the crossings of the strings @xmath26 and @xmath74 , and @xmath75 and @xmath76 ( see fig . [
the string diagrams for @xmath77 and @xmath78 have either the top or the bottom extra crossing . in all cases
, @xmath79 contains the hexagonal @xmath80 .
@xmath81 furthermore , it can be shown that the @xmath82avoiding permutations are exactly those 321avoiding permutations whose minimal string diagrams avoid the hexagonal string diagram of @xmath71 .
this justifies the descriptive term _ 321,hexagon avoiding _ and yields the following alternative description ( cf .
@xcite ) : * definition @xmath83 .
* a _ 321hex permutation _ is a permutation whose reduced expressions do not contain a substring of the form @xmath84 for all @xmath85 , and do not contain any of the translates @xmath86 of @xmath80 for all @xmath87 .
we now turn our attention to the kazhdan lusztig polynomials . for definitions and a detailed introduction ,
we direct the reader to @xcite,@xcite . in short , the _ hecke algebra _
@xmath88 of a finite weyl group @xmath89 ( such as @xmath0 ) is an algebra over @xmath90 , with basis @xmath91 , relations for all generators @xmath92 : @xmath93 and well defined inverses due to the presence of @xmath94 .
moreover , the involution of @xmath90 sending @xmath95 extends to an involution @xmath96 of @xmath88 .
the _ kazhdan
lusztig polynomials _ arise in search of a new basis @xmath97 for @xmath88 of @xmath96invariant elements which are linear combinations of `` lower
terms '' @xmath98 for @xmath99 under the bruhat
chevalley order : @xmath99 if every reduced expression for @xmath2 contains a subexpression for @xmath100 .
for any @xmath101 , @xmath88 has a unique @xmath96invariant element @xmath102 , where the degrees of the polynomials @xmath103 $ ] are at most @xmath104 if @xmath105 , and @xmath106 , @xmath107 if @xmath108 .
the kazhdan lusztig polynomials @xmath109 are of fundamental importance in lie theory .
it has been proven that their coefficients are non negative for weyl groups ( see kazhdan
lusztig @xcite , brylinski
kashiwara @xcite ) , but this question remains open for arbitrary coxeter groups .
neither the degrees nor the coefficients of @xmath109 are readily computable ; indeed , only partial results have been derived in certain cases ( see @xcite for detailed references . ) in @xcite , billey and warrington introduce the 321hex permutations @xmath110 , derive simple combinatorial formulas for their kazhdan lusztig and poincar polynomials , and a simple method for determining the singular loci of their schubert varieties @xmath111 .
concretely , let @xmath112 be a reduced expression for @xmath2 , and let @xmath113 denote the _ defect statistic _ of a _ mask _ @xmath114 on @xmath45 whose product is a fixed @xmath115 .
( for definitions of these and other related concepts , see @xcite . ) a permutation @xmath46 is 321hexagon avoiding if and only if one of the following equivalent conditions is satisfied : * the kazhdan lusztig polynomial @xmath116 for @xmath99 .
* the poincar polynomial for the full intersection cohomology group of @xmath111 equals @xmath117 .
* the kazhdan lusztig basis element @xmath118 * the bott
samelson resolution @xmath119 of @xmath111 is small .
* @xmath120 .
[ bw - theorem ] as theorem [ bw - theorem ] shows , the 321hex permutations have properties that make certain algebraic and combinatorial computations easier to perform compared to arbitrary permutations .
another interesting aspect of the 321hex permutations was examined earlier in the introduction : they are one of the few families that are known to be describable both in terms of pattern avoidance and heap avoidance . yet , until now , there was no known recursive , exact or other closed form for the number of 321hex permutations .
for instance , it would be useful to know to how many permutations theorem [ bw - theorem ] applies ; how the number of 321hex permutations changes asymptotically , and how it compares to sizes of other well known sets of permutations , such as @xmath121 , which is enumerated by the catalan numbers .
answers to all these questions are obtained in the present paper , where we find a 7term linear recursive relation and derive from it an explicit exact formula . with this formula at hand ,
answering any enumeration questions about the 321hex permutations becomes a matter of simple observation and calculation .
let @xmath122 denote the set of all 321hex permutations in @xmath0 , and let @xmath123 be the number of such permutations .
the sequence @xmath124 satisfies the following recursive relation for all @xmath125 : @xmath126 [ main - result ] of the six roots of the corresponding characteristic polynomial , four are real : @xmath127 for @xmath128 , and two are complex conjugates : @xmath129 .
this implies the same description of the six coefficients below : @xmath130 for @xmath128 , and @xmath131 .
the number of the 321hexagon avoiding permutations of length @xmath16 equals @xmath132 where the roots and coefficients are rounded off below to 5 digits after the decimal point : @xmath133 for further discussion of this and other results , we refer the reader to sections [ results][extensions ] .
the proof of theorem [ main - result ] follows several steps .
first , we describe the nodes in the generating tree of @xmath122 by using schensted s algorithm for 321avoiding permutations and by introducing 5 parameters for the generating function @xmath134 .
we next observe that this function depends on fewer parameters , yielding therefore relatively few distinct values .
we organize these values in five sequences , @xmath135 and @xmath136 . using the intrinsic symmetries of the set @xmath22 of octagonal patterns
, we deduce recursive relations expressing each sequence in terms of @xmath124 .
the latter turns out to be the number we are looking for , @xmath137 .
finally , putting together all information about the generating function and the five sequences results in the desired formula .
we turn now to the development of a recurrence for the 321hex permutations .
a standard tool in the enumeration of restricted permutations is the _ generating tree _
@xmath138 introduced in @xcite .
begin with an infinite tree whose nodes on level @xmath16 are identified with the permutations in @xmath0 .
the node @xmath2 is a child of @xmath139 $ ] where the omitted value is @xmath140 .
looking at this from the point of view of the parent , we can form all the children of @xmath141 by _ inserting _ the element @xmath142 into each of the @xmath142 _ sites _ of @xmath2 . then , for a given set @xmath143 of forbidden subsequences , prune the tree by deleting all nodes containing any of the forbidden subsequences .
what remains is still connected because if @xmath2 does not contain any forbidden subsequence then clearly @xmath144 does not either . for any node on level @xmath16 of this pruned tree @xmath145
, we call a site in the corresponding permutation _ active _ if inserting @xmath142 at that site yields a node of the tree ; conversely an _ inactive _ site is one where the insertion of @xmath142 creates one or more of the forbidden subsequences in @xmath143 . to give a complete description of a generating tree , we need to associate to each node an appropriate label , and then describe a _
succession rule _ for deriving the labels attached to the set of children of each node .
for instance , we might characterize the original tree @xmath138 generating _ all _ permutations as having a root labelled @xmath146 and a succession rule @xmath147 . in this instance
, the label can be interpreted as revealing directly how many children each node has in the generating tree . in the particular instance of the 321hex permutations
, it will turn out that we need a label containing four integers .
although this is more complicated than the single integer of our motivating example , it is nevertheless a major progress to reduce the amount of information recorded at a node from a full permutation to a label of any bounded size . in particular , it is possible to apply the succession
rule recursively to determine the entire downward structure of any node given only its label , regardless of whether that node is on a level corresponding to permutations on four symbols , or four thousand .
let @xmath2 be any 321hexagon avoiding permutation on @xmath16 symbols .
we divide the elements @xmath148 into two categories : the set of right to left minima ( including @xmath149 ) , and the rest . adapting the terminology of schensted @xcite , we refer to these as the _ first basic subsequence _ and _ second basic subsequence _ of @xmath2 , and denote them by @xmath150 and @xmath151 , respectively . for example , when @xmath152 , @xmath153 $ ] and @xmath154 $ ] .
note that both @xmath150 and @xmath151 decrease from right to left , the former by construction , and the latter because @xmath155 .
now let @xmath156 , @xmath157 and @xmath158 be the three largest ( i.e. rightmost ) elements in @xmath151 ( @xmath159 ) , with each one set to zero if the corresponding element does not exist .
since @xmath30 is the largest ( i.e. rightmost ) element in @xmath151 , it follows that every element to the right of @xmath30 belongs to @xmath150 .
there are @xmath160 elements in this region , which we will call the _
active region_. if @xmath161 because @xmath151 is empty , then we consider the entire permutation to be the active region .
let @xmath162 be the number of elements in the active region which are larger than @xmath163 ; as the elements in the active region decrease from right to left , these @xmath162 elements are @xmath164 .
similarly , @xmath165 be the number of elements in the active region which are larger than @xmath166 , and @xmath167 be the number of such elements larger than @xmath30 ( cf .
[ deletion1 ] . )
now assign the label @xmath168 to @xmath2 . by construction , @xmath169 .
@xmath170elements _ of @xmath2 be those that are counted in @xmath100 , and similarly define the @xmath163 , @xmath166 , and @xmath30elements .
further , let the _
@xmath171elements _ of @xmath2 be the set of @xmath170elements minus the set of @xmath163elements , and similarly define the @xmath172 and @xmath173elements .
for example , the @xmath170elements in @xmath71 are @xmath174 , and @xmath175 .
all children of @xmath2 must avoid the subsequence @xmath38 $ ] .
this restriction by itself renders inactive all sites to the left of @xmath30 , but none of the sites in the active region .
therefore when considering the children of @xmath2 and their labels , we need only consider insertions taking place in the active region to the right of @xmath30 .
thus , if we insert @xmath142 into a site in the active region with @xmath176 elements to its right , it is easy to verify that the resulting permutation will have label : @xmath177 furthermore , the number of @xmath2 s children can be computed as follows . set @xmath178 then the node @xmath2 has @xmath179 children , corresponding to the @xmath179 rightmost insertion sites , where @xmath180 this allows a more compact and complete succession rule for the labels of all @xmath179 children of @xmath2 : @xmath181 we shall not use directly any of the above formulas ( [ labels1])([labels2 ] ) in our calculations , so we leave their verification to the reader , who will find this easier after mastering the material in the rest of the paper .
the importance of the above discussion is that it completely describes the structure of the generating tree @xmath182 , and hence explains in principle why the enumeration in this paper works .
why the resulting final formula for the 321hex permutations is so simple a _ linear _ recursive relation with _ constant _ coefficients is a completely different matter and can be explained only by the structure of the forbidden set @xmath22 of four octagonal patterns , as we shall see later .
let @xmath134 and @xmath183 be the number , resp .
the set , of 321hex permutations in @xmath0 labelled @xmath168 .
according to this definition , @xmath184 is not defined since no permutation is labelled @xmath185 . for convenience ,
denote by @xmath184 and @xmath186 the number and the set of all 321hex permutations in @xmath0 which end in their largest element : @xmath187 .
it is worth noting that @xmath188 for @xmath189 corresponds to permutations @xmath2 in which either @xmath166 is smaller than the final `` tail '' of @xmath2 , or @xmath166 does not exist and @xmath190 .
naively , there are @xmath191 enumerations to be performed for @xmath134 . however , many have the same answers ; indeed , only @xmath192 different values of @xmath193 appear : @xmath194 the operation of deleting all @xmath163elements in a 321hex permutation provides a bijection @xmath195 . hence , for all @xmath196 we have @xmath197 . [ deletion1 ]
proof : let @xmath155 . the @xmath163elements of @xmath2 lie in @xmath150 and are part of the final increasing `` tail '' of @xmath2 ( since they are to the right of @xmath30 ) .
thus , if a @xmath163element @xmath198 were part of an octagonal pattern @xmath199 in @xmath2 , then @xmath198 would lie in @xmath200 .
but then @xmath199 , and hence @xmath2 , would contain at least 3 elements _ larger than _ and _ to the left of _ @xmath198 : this contradicts the definition of a @xmath163element , for which only @xmath166 and @xmath30 are _ larger than _ and _ to the left of _ it .
@xmath201 the above discussion shows that @xmath163elements can not participate in octagonal patterns in @xmath2 , and hence they can be deleted without losing any relevant 321hex information about @xmath2 .
( of course , we have to rescale down appropriately @xmath30 and @xmath166 of @xmath2 to arrive at a permutation of smaller size . )
the resulting map @xmath202 is bijective : to obtain @xmath203 from its image @xmath204 , insert the necessary number of @xmath205elements into @xmath206 and increase appropriately @xmath166 and @xmath30 to fit their definitions in @xmath2 .
this procedure works because we can identify @xmath30 as the largest element in @xmath206 ( after the deletion of all @xmath163elements in @xmath2 ) , and then identify @xmath166 as the second largest element of @xmath206 , which will be necessarily to the left of @xmath30
. then insertion of the appropriate @xmath205elements at the end of @xmath2 requires rescaling @xmath30 and @xmath166 only ( not @xmath163 ) , and hence transforms the 321schensted decomposition of @xmath206 into that of @xmath2 : whether @xmath166 was in @xmath207 or in @xmath208 does not prevent @xmath166 from becoming an element of @xmath150 after applying @xmath209 .
consider a family @xmath210 of permutations in @xmath122 which are described by certain configuration conditions imposed on their basic subsequence decomposition @xmath211 .
for example , @xmath186 is such a family defined by the condition @xmath212 .
let @xmath213 be an element of the permutations in @xmath214 which is identified uniquely in each @xmath215 by the configuration description of @xmath214 .
we say that @xmath213 is _ pattern free _ if its deletion in each @xmath215 ( and appropriate rescaling of @xmath2 ) results in a numerically equivalent family @xmath216 of permutations in @xmath122 ; i.e. @xmath217 with @xmath218 . for @xmath219 ,
denote by @xmath206 the image @xmath220 .
for example , in @xmath186 , the largest element @xmath221 is identified by being in the last position in each @xmath2 , and clearly it is pattern free ( no octagonal pattern @xmath36 has @xmath74 in its last position ) : @xmath222 establishing pattern free elements and identifying the image set @xmath216 is the basis of the enumeration of @xmath122 .
the following technical lemma summarizes the pattern free situations which will be used later in the proof of theorem [ main - result ] .
let @xmath210 be a family in @xmath122 , @xmath223 be the @xmath171elements of @xmath215 , and @xmath224 be the fourth largest element in @xmath151 .
the set @xmath225 imposes the following octagonal conditions on @xmath210 : * all @xmath163elements are pattern free . * if @xmath187 in @xmath210 , or @xmath226 in @xmath210 , then @xmath149 is pattern free . * if @xmath227 and @xmath228 in @xmath210 , then @xmath30 is pattern free . assume now that @xmath229 and @xmath175 in @xmath210 .
* if @xmath230 in @xmath210 ( @xmath224 may not exist in @xmath2 , ) then @xmath231 are pattern free .
* if @xmath232 in @xmath210 , then @xmath233 are pattern free . * if @xmath234 ( forcing @xmath235 ) in @xmath210 , then @xmath30 is pattern free .
@xmath236 proof : ( p1 ) follows from the proof of lemma [ deletion1 ] .
( p2 ) says that if all @xmath215 end in their largest element @xmath16 ( or in their second largest element @xmath65 ) , then this last element is pattern free ( cf .
[ alpha ] ) ; ( p3 ) says that if @xmath30 happens to be the largest element in all @xmath215 and is followed by at most two @xmath170elements , then @xmath30 is pattern free ( cf .
[ beta][gamma ] . ) both ( p2 ) and ( p3 ) follow from the facts : * each octagonal pattern @xmath36 has an empty @xmath237 upper right corner ( see fig . [ octagonal - conditions]a . ) * for ( p2 ) , reinserting @xmath16 or @xmath65 in the last position in @xmath238 does not create @xmath52patterns .
similarly , for ( p3 ) , reinserting @xmath227 into @xmath238 as the first ( rightmost ) element in @xmath151 does not create @xmath52patterns ( since the tail of @xmath239 after @xmath30 is increasing as part of @xmath207 . ) for ( p4 ) , cf .
case 1 in fig .
[ epsilon2 ] .
start by deleting the crosses of the elements in the bottom row and in the rightmost column of each @xmath36 ; this leaves the permutation matrix @xmath240 $ ] , which decomposes into a @xmath241 block @xmath242 in the upper left corner and the fixed @xmath243 block @xmath244 in lower left corner ( corresponding to the original @xmath245 and @xmath246 , see fig .
[ octagonal - conditions]b . ) * the fact that @xmath242 is to the left of and higher than @xmath244 proves that @xmath247 are pattern free : for them , only @xmath205 can fit in such an @xmath242 . in detail , if one of @xmath247 participates in an octagonal pattern @xmath36 , then without loss of generality we may assume that @xmath248 is the rightmost element of @xmath36 .
this forces @xmath205 to participate in @xmath36 too as the only elements larger than @xmath248 .
now @xmath30 , being the largest element of @xmath36 , forces at least three elements after it to participate in @xmath36 ; one is @xmath248 , one could be @xmath249 , hence the third one must be among @xmath250 .
but none of these elements can fit into an @xmath244 as in fig .
[ octagonal - conditions ] because only @xmath251 are larger and to the left of them .
further , reinsertion of @xmath247 into @xmath206 as the largest elements of @xmath150 can not create @xmath52patterns .
thus , @xmath252 are indeed pattern free . * after deleting @xmath247 , the largest element @xmath30 lands in second to last position in @xmath206 , and by ( p3 ) , it is pattern free . for ( p5 ) , cf .
case 2 in fig .
[ epsilon2 ] .
the reasoning here is similar to the case for ( p4 ) . *
first note that there can be no element @xmath253 of @xmath2 between @xmath166 and @xmath30 , or else @xmath254\sim p_1 $ ] . *
if one of @xmath255 participates in a pattern @xmath36 , this forces @xmath205 to participate too , and in order not to run into contradiction with the @xmath256 argument above , we must assume that @xmath249 and @xmath257 are also in @xmath36 , and only one among @xmath255 is in @xmath36 .
then the 1 in @xmath36 will have to be between the two largest elements @xmath166 and @xmath30 , which was ruled earlier .
thus , @xmath255 are pattern free .
* deletion of @xmath255 leaves @xmath258 in third to last position , so by ( p3 ) , @xmath30 is pattern free . but
@xmath166 is the largest element in @xmath259 , with only @xmath249 and @xmath257 after it , so by ( p3 ) again , @xmath166 is pattern free . for ( p6 ) , cf .
case 3 in fig .
[ delta2 ] .
again note that there can be no element @xmath253 of @xmath2 between @xmath166 and @xmath30 , or else @xmath254\sim p_3 $ ] .
if @xmath30 participates in a pattern @xmath36 , then it forces @xmath260 also to participate .
since the `` 1 '' in @xmath36 can not come from between @xmath166 and @xmath30 , we conclude that @xmath166 does not participate in @xmath36 .
but then we can replace @xmath30 by @xmath166 and argue that there is a pattern @xmath36 in @xmath259 , a contradiction .
hence @xmath30 is pattern free .
lemma [ deletion1 ] shows that @xmath134 does not depend on @xmath165 or @xmath167 , but rather on the differences @xmath263 and @xmath264 .
we shall see further that there are only 5 ranges for @xmath264 , that completely determine the values of @xmath265 : each of @xmath266 and @xmath267 corresponds to exactly one of the 5 sequences listed earlier , and defined as follows : @xmath268 where @xmath269 .
all of these definitions are justified by lemma [ deletion1 ] , except for the definition of @xmath136 , which depends only on @xmath16 but not on @xmath270 and whose explanation will be given later . in our search for relations between these sequences
, it will be useful to define each sequence via an alternative description in the terminology of the families @xmath210 discussed in section [ generating ] . by definition ,
@xmath273 is the largest and the rightmost element of any @xmath274 , and hence it is pattern free by ( p2 ) ( see fig .
[ alpha ] . ) deleting @xmath30 results in a family of @xmath52hex permutations of length @xmath16 without any further restrictions , i.e. we have a bijection @xmath275 this justifies the alternative description of @xmath124 , stated in the introduction : @xmath276 @xmath277 by definition , @xmath279 . here
@xmath273 is second from right to left in @xmath280 , and by ( p3 ) it is pattern free ( see fig . [ beta ] . )
thus , deleting @xmath30 imposes only one extra condition : in the new 321hex permutation @xmath281 , the rightmost element in @xmath207 is a @xmath282element or lower , i.e. the largest two elements ( the original @xmath283 and @xmath284 ) belong to @xmath208 .
this justifies the alternative description : @xmath285 @xmath286 in order to find a recursive description of @xmath287 , note that each @xmath288 either ends in @xmath16 or @xmath65 , or both @xmath16 and @xmath65 belong to @xmath151 : @xmath289 by ( p2 ) , if @xmath16 or @xmath65 is the last element in @xmath290 , then it is pattern free , so deleting it results in a bijection : @xmath291 this justifies the use of @xmath292 in ( [ beta - recursive1 ] )
. consequently , we have the recursive relation @xmath293 for @xmath294 . by definition , @xmath296 .
here @xmath273 is third from right to left in @xmath297 , and by ( p3 ) it is pattern free ( see fig .
[ gamma ] . )
deleting @xmath30 imposes the following extra conditions : in the image @xmath281 , the first two elements in @xmath207 are @xmath282elements or lower , i.e. the largest two elements ( the original @xmath283 and @xmath284 ) belong to @xmath151 and there are at least 2 numbers _ after _ @xmath283 .
this justifies the alternative description for @xmath298 : @xmath299 @xmath300 in order to find a recursive description of @xmath301 , note that each @xmath2 counted in @xmath287 ( i.e. @xmath302 ) , falls into one of the following subcases : there is exactly one element _ after _ @xmath16 ( hence in @xmath150 ) , or there are at least two elements _ after _
@xmath16 ( hence in @xmath150 ) : @xmath303 the underbraced set in ( [ gamma - recursive1 ] ) is depicted in the lhs of fig .
[ gamma2 ] . by ( p3 ) , deletion of @xmath283 results in the numerically equivalent set @xmath304 in the rhs of fig .
[ gamma2 ] .
the permutations in @xmath304 can be described as having their largest element @xmath305 . on the other hand
, @xmath306 breaks into two disjoint groups : group @xmath307 consists of the permutations having the largest element @xmath65 in last position , and group @xmath308 is the set @xmath304 : @xmath309 . finally , note that @xmath310 , so that @xmath311 .
this justifies the underbrace notation in ( [ gamma - recursive1 ] ) , and implies the formulas : @xmath312 for @xmath313 .
@xmath314 by definition , @xmath315 for @xmath316 .
let @xmath317 be the @xmath318elements of @xmath319 , where @xmath316 .
there are two cases to consider ( see fig .
[ epsilon2 ] . ) * case 1 . * except for @xmath205 , all other elements of @xmath151 are smaller than @xmath257 . our drawing shows the next element @xmath320 s.t .
@xmath230 ( @xmath224 may not exist in @xmath2 . ) by ( p4 ) , @xmath321 and @xmath30 are pattern free . after deletion , the remaining configuration in @xmath322 is identical to the alternative description of @xmath323 .
@xmath324 * case 2 .
* after @xmath205 , the fourth element @xmath224 of @xmath151 is between @xmath257 and @xmath325 : @xmath232 .
recall from ( p5 ) that there can be no element @xmath253 of @xmath2 between the vertical lines of @xmath166 and @xmath30 ; or else , @xmath326 , and hence @xmath254\sim p_1 $ ] .
further , @xmath327 , @xmath30 and @xmath166 , are pattern free . after deletion ,
the remaining configuration in @xmath322 is identical to the alternative description of @xmath328 .
note that the case @xmath234 is not allowable , or else @xmath329\sim p_2 $ ] or @xmath330 ( when @xmath331 or @xmath332 , respectively . ) incidentally , the above discussion shows that @xmath333 does not depend on @xmath100 as long as @xmath316 , and hence justifies the definition of @xmath136 .
we conclude that @xmath334 for @xmath335 . by definition , @xmath337 .
as above , denote by @xmath260 the @xmath318elements of @xmath338 .
there are three cases to consider ( see fig .
[ delta2 ] . )
@xmath339 * case 1 . * except for @xmath205 , all other elements of @xmath151 are smaller than @xmath257 .
our drawing shows the next element @xmath320 s.t .
@xmath230 ( @xmath224 may not exist in @xmath2 . ) by ( p4 ) , @xmath340 and @xmath30 are pattern free .
the remaining configuration in @xmath341 is identical to the alternative description of @xmath342 .
* case 2 . * after @xmath205 , the fourth element @xmath224 of @xmath151 is between @xmath257 and @xmath325 : @xmath232 . by ( p5 )
, there can be no element @xmath253 of @xmath2 between the vertical lines of @xmath166 and @xmath30 ; or else , @xmath326 , @xmath343 , and hence @xmath254\sim p_1 $ ] .
further , @xmath325 , @xmath30 and @xmath166 are pattern free . after deletion ,
the remaining configuration in @xmath341 is identical to the alternative description of @xmath344 .
* case 3 . * in contrast to
the discussion of @xmath262 , in the case of @xmath336 it is possible to have @xmath234 as long as there is no element @xmath253 of @xmath2 between the vertical lines of @xmath166 and @xmath30 ; otherwise , @xmath254~\sim~p_3 $ ] . by ( p6 ) , the largest element @xmath273 is pattern free . after its deletion ,
the remaining configuration in @xmath306 is identical to the original description of @xmath345 .
we conclude that @xmath346 for @xmath125 .
we summarize the results in this section in the sequences @xmath261 and @xmath262 satisfy the following recursive relations : @xmath347
we are now in a position to combine the recurrence formulas for @xmath348 and @xmath262 into a single recurrence for @xmath271 .
we first use the interpretation of @xmath124 as @xmath349 and expand this in terms of the individual values of @xmath134 .
thus , for a fixed @xmath16 @xmath350 where the sum is taken over @xmath351 .
next , we break the sum into five separate sums depending on the value of @xmath264 ; each such sum corresponds to the definitions of @xmath348 and @xmath262 , respectively .
note that the sum for @xmath271 ( where @xmath352 ) requires two extra special cases for @xmath353 and @xmath354 ; in both cases @xmath355 .
@xmath356 in the next step , we replace the @xmath357 s by the appropriate values of @xmath348 and @xmath262 .
the case of @xmath271 requires some care as we must observe the condition that not all the numerical parameters @xmath358 can be simultaneously equal ; indeed , this only happens in the special case @xmath359 , which was broken out from the main sum earlier .
@xmath360 we replace the indices @xmath162 , @xmath165 and @xmath167 by binomial coefficients : @xmath361 finally , we use the recurrences from sect .
[ recurrences ] to obtain a summation in terms of @xmath271 alone : @xmath362 we simplify this large expression into the following full history , linear recurrence relation with cubic polynomial coefficients , valid for @xmath125 : @xmath363 where @xmath364 and @xmath365 the polynomial @xmath366 arises from three special cases for the polynomial coefficients of @xmath367 and @xmath368 . since @xmath369
, we need 4 successive history eliminations of the form @xmath370 ; combined with the 3 initial terms @xmath371 and @xmath372 , this produces the desired order
six constant coefficient linear recurrence for all @xmath373 : @xmath374 consequently , the number of the 321hexagon avoiding permutations of length @xmath16 is given by the formula : @xmath375 where the roots and coefficients are listed in the introduction . this completes the proof of theorem [ main - result ] .
the sixth degree characteristic polynomial of our recurrence relation ( [ main - recurrence ] ) is irreducible over @xmath376 and has galois group @xmath377 , as calculated by maple .
this means that there are no further algebraic relations among the roots @xmath127 in ( [ formula ] ) , and thus we can not hope for any better closed form results .
however , our numerical approximations of the roots and coefficients can yield exact values for the number of 321hex permutations for any fixed length @xmath16 , exploiting the fact that the value being approximated is known to be an integer . furthermore , since the two roots of modulus less than @xmath26 make such small contributions , they can be dropped and the following formula is exact for all @xmath16 : @xmath378,\ ] ] where the braces denote rounding to the nearest integer .
the study of the octagonal patterns @xmath36 in the present paper was motivated by their apperance in the representation theory of @xmath0 via heap avoidance and thus kazhdan lusztig polynomials and schubert varieties .
we refer to the enumeration of @xmath379 as the `` @xmath37 case '' . from a purely combinatorial viewpoint ,
it is natural to ask what happens in the analogous smaller @xmath380 and @xmath241 cases whose generalized patterns are depicted in fig .
[ extension ] .
@xmath381 to obtain these cases , for each @xmath128 shorten by one the lengths of both @xmath382 and @xmath383 by removing a chosen fixed point appearing in all @xmath36 s .
thus , define @xmath384,[356124],[451623],[456123]\}\,\,\text{and}\\ \mathcal{p}^{4}&=&\{[2143],[3142],[2413],[3412]\}\end{aligned}\ ] ] to be the families of avoided octagonal patterns in the @xmath380 and the @xmath241 cases , respectively .
both of these cases lead again to linear recursive relations with constant coefficients .
the proofs below follow closely the method described in the @xmath37 case , so we leave the details for verification to the reader .
let @xmath385 , and @xmath386 .
then @xmath387 satisfies a 6term linear recursive relation with constant coefficients : @xmath388 consequently , for all @xmath389 , @xmath390 where the roots and coefficients are rounded off below to 5 digits after the decimal point : @xmath391 [ 6-by-6 ] the degree5 characteristic polynomial of the recurrence is irreducible over @xmath376 and has the largest possible galois group , @xmath392 , as calculated by maple .
we can again drop the small roots @xmath393 and @xmath394 , rounding off the remainder to the nearest integer : @xmath395\,\,\text{for all}\,\ , n\geq 1.\ ] ] the first values of @xmath124 are : 1 , 2 , 5 , 14 , 42 , 128 , 389 , 1179 , 3572 , 10825 , 32810 , 99446 . proof(theorem [ 6-by-6 ] ) : we modify the discussion in the proof for the @xmath37 case . define the generating function @xmath396 with one fewer parameter , thus taking into account only the elements @xmath30 and @xmath166 of @xmath397 . as in lemma
[ deletion1 ] , the @xmath166elements are pattern free , and therefore @xmath398 . next ,
define the sequences @xmath399 , @xmath400 , @xmath401 and @xmath402 for @xmath229 .
note that @xmath403 since @xmath124 is the number of @xmath404 in which the largest element @xmath142 is at the end and hence it is pattern free .
further , the relations among the sequences are : @xmath405 now we are ready to express everything in terms of @xmath406 : @xmath407 here the `` 1 '' counts the identity permutation : no @xmath30element in @xmath2 ; in the first sum @xmath166 is smaller than all @xmath170elements ( or @xmath166 does not exist ) ; in the second sum @xmath166 is larger than exactly one @xmath170element ; in the third sum @xmath166 is larger than exactly two @xmath170elements ; and in the forth sum @xmath166 is larger than three or more @xmath170elements and hence @xmath408 .
@xmath409 conveniently , each summand on the rhs can be expressed in terms of @xmath271 , including @xmath410 : @xmath411 let @xmath412 , and @xmath413 .
then @xmath414 in particular , @xmath387 satisfies the 4term linear recursive relation @xmath415 note that in contrast to the @xmath380 and @xmath37 cases , the characteristic polynomial here factors ( completely ) over @xmath376 : @xmath416 .
proof : as expected , we define the generating function @xmath417 with only two parameters , keeping track only of @xmath30 in @xmath418 .
obviously , the @xmath30elements are pattern free , so @xmath419 .
further , letting @xmath420 and @xmath421 for @xmath422 , one quickly discovers that @xmath423 , so @xmath424 for @xmath294 , and @xmath425 for @xmath389 .
express @xmath426 as the sum @xmath427 as before , here the `` 1 '' counts the identity permutation : no @xmath30element in @xmath2 ; in the first sum @xmath30 is larger than exactly one @xmath170element ; and in the third sum @xmath30 is larger than two or more @xmath170elements and hence @xmath428 .
@xmath429 until now , there were relatively few known examples of sets of permutations whose avoidance led to linear , polynomial , or exponential formulas ( see @xcite . ) after the successful enumeration of the @xmath241 , @xmath380 and the 321hex @xmath37 cases , it is tempting to generalize the recursive sequence method of this paper to the corresponding larger sets of @xmath430 patterns . at this point , it is not surprising to conjecture that all these families yield linear recursive relations with constant coefficients .
in fact , when the result of the present paper was publicized , herbert wilf requested that many more examples of such `` linear '' families be found .
these and other related questions will be answered positively in a forthcoming paper by stankova frenkel .
the authors would like to thank sara billey and gregory warrington for sharing their results on kazhdan lusztig polynomials for 321hexagon avoiding permutations and suggesting the problem on enumeration of forbidden subsequences in this paper . | the _ 321,hexagon avoiding _ ( _ 321hex _ ) permutations were introduced and studied by billey and warrington in @xcite as a class of elements of @xmath0 whose kazhdan lusztig and poincar polynomials and the singular loci of whose schubert varieties have certain fairly simple and explicit descriptions . this paper provides a 7term linear recurrence relation leading to an explicit enumeration of the 321hex permutations .
a complete description of the corresponding generating tree is obtained as a by product of enumeration techniques used in the paper , including schensted s 321subsequences decomposition , a 5parameter generating function and the symmetries of the octagonal patterns avoided by the 321hex permutations .
may 2001 |
according to einstein s general theory of relativity , the dynamic properties of a given space - time are fully determined by its total energy content . in the cosmological context , for instance , this amounts to saying that in order to understand the space - time structure of the universe one needs to identify the relevant sources of energy as well as their contributions to the total energy momentum tensor @xmath3 .
matter fields ( e.g. , baryonic matter and radiation ) , are obvious sources of energy . nevertheless , according to current observations , two other components , the so - called dark matter and dark energy ( or _ dark pressure _ ) , whose origin and nature are completely unknown thus far , are governing the late time dynamic properties of our universe . in the standard lore ,
the later actor plays a special role because it is supposed to be driving the current cosmic acceleration .
dark energy has been inferred from a number of independent astronomical observations , which includes distance measurements of type ia supernovae ( sne ia ) @xcite , estimates of the age of the universe @xcite , measurements of the cosmic microwave background ( cmb ) anisotropies @xcite , and clustering estimates @xcite . while the combination of the two latter results suggests the existence of a smooth component of energy that contributes with @xmath4 of the critical density , sne ia observations require such a component to have a negative pressure , which generates repulsive gravity and accelerates the cosmic expansion .
in addition to these recent observational results , the astonishing success of the inflationary cold dark matter ( cdm ) paradigm in explaining high precision measurements of cmb anisotropy , galaxy clustering , the ly@xmath5 forest , gravitational lensing and other astrophysical phenomena can also be thought of as an indirect but important evidence for the existence of a dominant repulsive dark energy component in this short contribution we review some aspects of the so - called dark energy problem .
this paper does not aim to be ( and is far from being ) an exhaustive or complete account of the problem ( for that end we refer the reader to ref .
@xcite ) . here , we focus our attention on three possible mechanisms capable of explaining the current cosmic acceleration , namely , a decaying @xmath2 term , the potential energy density associated with a dynamical scalar field and modifications in gravity due to extra dimensions effects .
as discussed below , all these possibilities ( together with the standard cosmological constant @xmath2 ) seem somewhat to explain the current observational results , which makes the nature of dark energy a completely open question nowadays in cosmology .
there is no doubt that the simplest and most theoretically appealing candidate for dark energy is the cosmological constant ( @xmath2 ) , whose presence modifies the einstein field equations ( efe ) to ( throughout this paper we work in units where the speed of light @xmath6 ) @xmath7 physically , @xmath2 acts as an isotropic and homogeneous source with a constant equation of state ( eos ) @xmath8 , where @xmath9 and @xmath2 is by means of the raychaudhury equation , i.e. , @xmath10 or , equivalently , @xmath11 . clearly , the cosmological constant gives rise to a kind of repulsive force proportional to distance , which in principle could be the mechanism behind the cosmic acceleration evidenced by current sne ia observations . ] . from a more fundamental viewpoint , a physical basis for
the cosmological constant remained unclear until 1967 , when zeldovich @xcite showed that @xmath2 is related to the zero - point vacuum fluctuations of fields and that those must respect lorentz invariance so that @xmath12 . formally , the zero - point energy of a quantum field ( thought of as a collection of an infinite number of harmonic oscillators in momentum space ) must be infinity . if , however , we sum over the zero - point mode energies up to a certain ultraviolet momentum cutoff @xmath13 ( so that the theory under consideration is still valid ) , we find @xmath14 by considering the above expression , we show in table i the expected contribution to the vacuum energy for some _ fundamental _ energy scales in nature .
the ratio of the values appearing in the third column of table i to the current observational estimate of @xmath15 , i.e. , @xmath16 @xmath17 , ranges from 46 - 120 orders of magnitude and is the origin of the famous discrepancy between theoretical and observational estimates for @xmath2 , the so - called cosmological constant problem ( see @xcite for a review on this topic ) . from these arguments one may conclude that although cosmological models with a relic @xmath2 term ( @xmath2cdm ) are extremely successful from the observational viewpoint , the fine - tunning problem involving the vacuum energy density and @xmath2 seem to hamper any definitive conclusion on the cosmological constant as the actual nature of the mechanism behind cosmic acceleration ( see @xcite for a broader discussion on this issue ) .
a phenomenological attempt at alleviating the above problem is allowing @xmath2 to vary - type term that varies in space or time should be thought of as a new _ time - varying field _ and not as a cosmological constant . here , however , we adopt the usual nomenclature of time - varying or dynamical @xmath2 models . ] .
cosmological scenarios with a time - varying or a dynamical @xmath2 term were independently proposed almost twenty years ago in ref .
@xcite ( see also @xcite ) .
in order to build up a @xmath2(t)cdm scenario , we first note that according to the bianchi identities , the einstein equations ( [ efe ] ) implies that @xmath2 is necessarily a constant either if @xmath18 or if @xmath19 is separately conserved , i.e. , @xmath20 .
in other words , this amounts to saying that 1 .
vacuum decay is possible only from a previous existence of some sort of non - vanishing matter and/or radiation ; 2 .
the presence of a time - varying cosmological term results in a coupling between @xmath19 and @xmath2 of the type @xmath21 lcl energy scale & & @xmath15 + + qcd & 0.3 gev & @xmath22 @xmath17 + eletroweak & @xmath23 gev & @xmath24 @xmath17 + gut & @xmath25 gev & @xmath26 @xmath17 + planck & @xmath27 gev & @xmath28 @xmath17 + + to proceed further , one must now specify the function @xmath2(t ) .
however , in the absence of a natural guidance from fundamental physics on a possible time variation of the cosmological constant most of the decay laws discussed in the literature are phenomenological ( based on dimensional , black hole thermodynamics arguments , among others .
see , e.g. , @xcite and references therein ) . in this
regard , a still phenomenological but very interesting step toward a more realistic decay law was recently discussed in ref .
@xcite ( see also @xcite ) , in which the time variation of @xmath2 is deduced from the effect it has on the cdm evolution .
the qualitative argument is the following : since vacuum is decaying into cdm particles , cdm will dilute more slowly compared to its standard evolution , @xmath29 . thus , if the deviation from the standard evolution is characterized by a constant @xmath30 , i.e. , @xmath31 eq .
( [ coupling ] ) yields @xmath32 where @xmath33 is the current cdm energy density , @xmath34 is the cosmological scale factor , @xmath35 is an integration constant , and thermodynamic considerations restrict @xmath30 to be @xmath36 @xcite ( a similar decay law is also obtained from renormalization group running @xmath2 arguments @xcite ) .
note that , differently from previous vacuum decay scenarios , in which the universe is either always accelerating or always decelerating from the onset of matter domination to today , the presence of the residual term @xmath35 in eq .
( [ decayv ] ) makes possible a transition from an early decelerated to a current accelerating phase , as evidenced by sne ia observations @xcite ( see fig .
note also that , in the case of vacuum decay into photons ) and ( [ decayv ] ) become 4 s . ] , the primordial nucleosynthesis arguments discussed in refs .
@xcite ( more specifically the bounds from the primordial mass fraction of helium @xmath37 and the primordial abundance by number of deuterium @xmath38 ) may be no longer valid since even for small values of the parameter @xmath30 the residual term @xmath35 may account not only for the current cosmic acceleration but also for the recent estimates of the total age of the universe . 0.1 in
if vacuum is really transferring energy to the cdm component , the immediate question one may ask is where exactly the vacuum energy is going to or , in other words , where the cdm particles are storing the energy received from the vacuum decay process . in principle , since the energy density of the cdm component is @xmath39 , there are at least two possibilities , namely , the current of cdm particles has a source term @xmath40 ( while the proper mass of cdm particles remains constant ) or the mass of the cdm particles is itself a time - dependent quantity @xmath41 ( while the total number of cdm particles , @xmath42 , remains constant ) .
the former possibility is the traditional approach for the vacuum decay process whereas the latter constitutes a new example of the so - called vamp - type scenarios , in which the interaction of cdm particles with the dark energy field imply directly in an increasing of the mass of cdm particles ( see , e.g. , @xcite and references therein for more about vamp models ) . in fig .
1b we compare the above @xmath43cdm scenario with the standard one ( @xmath2cdm ) , which is formally recovered for @xmath44 .
we use to this end some of the most recent cosmological observations , i.e. , the latest chandra measurements of the x - ray gas mass fraction in 26 galaxy clusters , as provided by allen et al .
@xcite along with the so - called _ gold _ set of 157 sne ia , recently published by riess et al .
@xcite , and the estimate of the cmb shift parameter @xcite . in our analysis
, we also include the current determinations of the baryon density parameter , as given by the wmap team @xcite , i.e. , @xmath45 and the latest measurements of the hubble parameter , @xmath46 , as provided by the hst key project @xcite ( we refer the reader to @xcite for more details on this statistical analysis ) .
the contours stand for confidence regions ( @xmath47 , 95.4@xmath48 and 99.7@xmath48 ) in the plane @xmath49 .
note that , although the limits on the parameter @xmath30 are very restrictive , the analysis clearly shows that the decaying vacuum scenario discussed above constitutes a small but significant deviation from the standard @xmath2cdm dynamics ( as mentioned earlier the standard @xmath2cdm model is very successful from the observational viewpoint so that large deviation from its dynamics are not expected ) .
the best - fit parameters for this analysis are @xmath50 and @xmath51 , with the relative @xmath52 ( @xmath53 is defined as degrees of freedom ) .
note that this value of @xmath54 is similar to the one found for the so - called concordance model " by using sne ia data only , i.e. , @xmath55 @xcite . at 95.4@xmath48 c.l .
we also found @xmath56 and @xmath57 .
we expect that upcoming observational data along with new theoretical developments will be able to confirm or not @xmath43cdm models as realistic dark energy scenarios . 0.1 in
if the cosmological term is null or it is not decaying in the course of the expansion , something else must be causing the universe to speed up . the next simplest approach toward constructing a model for an accelerating universe is to work with the idea that the unknown , unclumped dark energy component is due exclusively to a minimally coupled scalar field @xmath1 ( quintessence field ) which has not yet reached its ground state and whose current dynamics is basically determined by its potential energy @xmath58 .
this idea has received much attention over the past few years and a considerable effort has been made in understanding the role of quintessence fields on the dynamics of the universe @xcite .
examples of quintessence potentials are ordinary exponential functions @xmath59 @xcite , simple power - laws of the type @xmath60 @xcite , combinations of exponential and sine - type functions @xmath61 $ ] @xcite , among others ( see , e.g. , @xcite and references therein ) . an important aspect worth to emphasize
is that all these quintessence scenarios are based on the premise that fundamental physics provides motivation for light scalar fields in nature so that a quintessence field @xmath1 may not only be identified as the dark component dominating the current cosmic evolution but also as a bridge between an underlying theory and the observable structure of the universe .
if , however , it is desirable a more complete connection between the physical mechanism behind dark energy and a fundamental theory of nature , one must bear in mind that an eternally accelerating universe , a rather generic feature of quintessence scenarios , seems not to be in agreement with string / m - theory ( possible candidates of a model for quantum gravity ) predictions , since it is endowed with a cosmological event horizon which prevents the construction of a conventional s - matrix describing particle interactions @xcite .
although the transition from an initially decelerated to a late - time accelerating expansion is becoming observationally established @xcite , the duration of the accelerating phase , depends crucially on the cosmological scenario and , several models , which includes our current standard @xmath2cdm scenario , imply an eternal acceleration or even an accelerating expansion until the onset of a cosmic singularity ( e.g. , the so - called phantom cosmologies @xcite ) .
this dark energy / string theory conflict , therefore , leaves us with the formidable task of either finding alternatives to the conventional s - matrix or constructing a quintessence model of the universe that predicts the possibility of a transient acceleration phenomenon . in this regard , an interesting quintessence scenario whose
accelerating phase is a transient phenomenon has been proposed in ref .
@xcite . in
what follows , we highlight some of its features .
let us first consider a homogeneous , isotropic , spatially flat cosmologies described by the frw flat line element .
the action for the model is given by @xmath62\ ; , \ ] ] where @xmath63 is the ricci scalar and @xmath64 is the planck mass .
the scalar field is assumed to be homogeneous , such that @xmath65 and the lagrangian density @xmath66 includes all matter and radiation fields . by combining the following _ ansatz _ on the scale factor derivative of the energy density @xmath67 with the conservation equation for the quintessence component , i.e. , @xmath68 , the expressions for the scalar field and its potential can be written as @xmath69 and @xmath70\;.\ ] ] in the above expressions @xmath5 and @xmath71 are positive parameters , @xmath72 is the current value of the field @xmath1 , @xmath73 , @xmath74 $ ] , and the generalized function @xmath75 , defined as @xmath76 , reduces to the ordinary logarithmic function in the limit @xmath77 @xcite . the important aspect to be emphasized at this point is that in the limit @xmath78 eqs .
( [ phi1 ] ) and ( [ gpotential ] ) fully reproduce the exponential potential studied by ratra and peebles in ref .
@xcite , while @xmath79 @xmath80 the scenario described above represents a generalized model which admits a wider range of solutions .
the eos for this quintessence component , i.e. , @xmath81 is shown as a function of the redshift parameter ( @xmath82 ) in fig .
( 2a ) for some selected values of the index @xmath5 and @xmath83 .
the eos above ( which must lie in the interval @xmath84 ) is an increasingly function of time , being @xmath85 in the past , @xmath86 today , and becoming more positive in the future ( @xmath87 at @xmath88 and @xmath89 at @xmath90 ) .
such a behavior is typical of the so - called _ thawing _ fields , as discussed in refs .
@xcite . for this kind of fields , ref .
@xcite also provides the following constraint @xmath91 where @xmath92 .
note that , if it is natural to impose such a constraint at the epoch when dark energy starts becoming important , i.e. , at @xmath93 , then the above interval for @xmath94 can be translated into the following bounds on the index @xmath5 @xmath95 clearly , for all the values of @xmath5 ranging in the above interval , the cosmic acceleration is a transient phenomenon since the quintessence field @xmath1 will behave more and more as an attractive matter field . in order to better visualize this transient behavior ,
we show in fig .
( 2b ) the deceleration parameter , @xmath96 , as a function of the redshift for some values of the index @xmath5 and @xmath97 . as can be seen from this figure
, @xmath79 @xmath80 the universe was decelerated in the past , began to accelerate at @xmath98 , is currently accelerated but will eventually decelerate in the future .
note also that , while the acceleration redshift @xmath99 depends very weakly on the value of @xmath5 , the deceleration redshift @xmath100 is strongly dependent , with the latter transition becoming more and more delayed as @xmath78 . as mentioned earlier , a cosmological behavior like the one described above seems to be in agreement with the requirements of string / m - theory ( as discussed in refs .
@xcite ) , in that the current accelerating phase is a transitory phenomenon . as one may also check , the cosmological event horizon , i.e.
, the integral @xmath101 diverges for this transient scalar - field - dominated universe , thereby allowing the construction of a conventional s - matrix describing particle interactions within the string / m - theory frameworks .
a typical example of an eternally accelerating universe , i.e. , the @xmath2cdm model , is also shown in fig .
( 2b ) for the sake of comparison .
so far we have discussed the phenomenon of cosmic acceleration as the result of unknown physical processes involving new fields in high energy physics . however , another possible route to deal with this dark _ pressure _ problem could be a modification in gravity instead of any adjustment to the energy content of the universe .
this idea naturally brings to light another important question at the interface of fundamental physics and cosmology : extra dimensions . as is well known the existence of extra dimensions is required in various theories beyond the standard model of particle physics , especially in theories for unifying gravity and the other fundamental forces , such as string or m theories .
extra dimensions may also provide a possible explanation for the so - called hierarchy problem , i.e. , the huge difference between the electroweak and planck scales [ @xmath102 @xcite . in this regard ,
an interesting scenario was proposed by randall and sundrum who showed that , if our 3-dimensional world is embedded in a 4-dimensional anti - de - sitter bulk , gravitational excitations are confined close to our sub - manifold , giving rise to the familiar @xmath103 law of gravity @xcite . in the cosmological context
, the role of extra spatial dimensions is translated into the so - called brane world ( bw ) cosmologies @xcite .
many attempts to observationally detect or distinguish brane effects from the usual dark energy physics have been recently discussed in the literature . in ref .
@xcite , for instance , sahni and shtanov investigated a class of bw models which admit a wider range of possibilities for the dark pressure than do the usual dark energy scenarios , while maia et al .
@xcite , showed that the dynamics of a dark energy component parametrized by a constant eos ( @xmath104 ) can be fully described by the effect of the extrinsic curvature of a frw universe embedded into a 5-dimensional , constant curvature de - sitter bulk .
0.3 in 0.1 in another interesting bw scenario is the one proposed by dvali et al .
@xcite , widely refered to as dgp model ( see also @xcite for a recent review on the dgp phenomenology ) .
it describes a self - accelerating 5-dimensional bw model with a noncompact , infinite - volume extra dimension whose dynamics of gravity is governed by a competition between a 4-dimensional ricci scalar term , induced on the brane , and an ordinary 5-dimensional einstein - hilbert action , i.e. , @xmath105 where @xmath106 denotes the 5-dimensional reduced planck mass . the friedmann equation which comes from this gravitational action coupled to matter on the brane
take the form @xcite @xmath107^{2 } = h^{2 } + \frac{k}{r(t)^{2}},\ ] ] where @xmath108 is the energy density of the cosmic fluid , @xmath109 is the spatial curvature and @xmath110 is the crossover scale defining the gravitational interaction among particles located on the brane . for scales below the crossover radius @xmath111 ( where the induced 4-dimensional ricci scalar dominates ) ,
the gravitational force experienced by two punctual sources is the usual 4-dimensional @xmath112 force , whereas for distance scales larger than @xmath111 the gravitational force follows the 5-dimensional @xmath113 behavior where @xmath114 is the number of spatial dimensions . ] . from the above equation
it is also possible to define the density parameter associated with the crossover radius @xmath111 , i.e. , @xmath115 note that estimates of @xmath116 means estimates on the length scale @xmath111 .
as shown by several authors @xcite , for values of @xmath117 , the presence of an infinite - volume extra dimension as described above leads to a late - time acceleration of the universe , in agreement with most of the current distance - based cosmological observations , where @xmath118 and @xmath119 . ] . in figure 3a
we show the joint confidence contours at 68.3% , 95.4% , and 99.7% confidence levels in the parametric space @xmath120 arising from the _ gold _ sne ia sample @xcite and the sdss baryon acoustic oscillations @xcite ( see @xcite for more details ) .
the best - fit parameters for this analysis are @xmath121 and @xmath122 .
note that the best - fit value for @xmath116 leads to an estimate of the crossover scale @xmath111 in terms of the hubble radius @xmath123 , i.e. , @xmath124 .
figure 3b illustrates the allowed regions in the @xmath120 plane by using the first - year snls data @xcite in conjunction with the sdss baryon acoustic oscillations .
the best fit for this joint snls plus bao analysis happens at @xmath125 and @xmath126 , which is very closed to the wmap estimates for the clustered matter @xcite . ] . from these and other recent results ( see @xcite for more observational analyses in dgp models ) we note that this class of bw models provide a good description for the current obserbational data , which may be thought of as an indication that the existence of extra dimensions play an important role not only in fundamental physics but also in cosmology .
the results of observational cosmology in the past decade have opened up an unprecedented opportunity to establish a more solid connection between fundamental physics and cosmology .
surely , the most remarkable finding among these results comes from sne ia observations which suggest that the cosmic expansion is undergoing a late time acceleration .
these sne ia results have been checked and confirmed by other cosmological observables , including the angular distribution of the 3-k cmb temperature and the clustering and dynamical estimates on the matter density parameter , which certainly makes a serious case for dark energy .
theoretical efforts to explain the nature of this dark energy component are vast and include , besides the possibilities discussed here , models with mass varying neutrinos @xcite , holographic dark energy @xcite , two - field quintom models @xcite , models driven by dissipative process @xcite , unification dark matter / dark energy @xcite , among many others . here
, we have explored three specific scenarios of dark energy and shown that , although completely different from the physical viewpoint , all of them have attractive features capable of explaining the current cosmic acceleration and other recent astronomical observations .
we emphasize that a possible way to distinguish among some of these dark energy candidates is through the observational limits on the cosmic eos ( @xmath104 ) and its time derivative ( @xmath94 ) , i.e. , if upcoming observations confirrm @xmath127 , then one will need to review the @xmath2 problem , in order to fathom out why the formally infinite quantity @xmath15 is in fact so very small . on the other hand , if either @xmath128 or it is a time - dependent quantity(@xmath129 ) , then one will be able to rule out the cosmological constant as the actual nature of the dark energy and search for more realistic quintessence , bw and other dark energy models .
we expect that future observations along with theoretical developments will be able to decide this matter and shed some light on the origin and nature of the dark energy .
i wish to thank the organizers of the xxvi enfpc for the opportunity to attend such an interesting meeting .
i am also very grateful to j. a. s. lima , n. pires , r. silva , z .- h .
zhu , d. jain , a. dev and f. c. carvalho for valuable discussions .
this work is supported by by cnpq under grants no .
307860/2004 - 3 and 475835/2004 - 2 , and by fundao de amparo pesquisa do estado do rio de janeiro ( faperj ) , no . e-26/171.251/2004 .
j. s. dunlop et al .
, nature * 381 * , 581 ( 1996 ) ; b. chaboyer et al .
, astrophys .
j. * 494 * , 96 ( 1998 ) ; r. cayrel et al .
, nature * 409 * , 691 ( 2001 ) ; g. hasinger et al .
, astrophys .
j. , 573 , l77 ( 2002 ) ; a. friaa , j. s. alcaniz and j. a. s. lima , mon . not .
. soc . * 362 * , 1295 ( 2005 ) .
astro - ph/0504031 .
v. sahni and a. a. starobinsky , int
d9 , 373 ( 2000 ) ; p. j. e. peebles and b. ratra rev .
phys . * 75 * , 559 ( 2003 ) ; t. padmanabhan , phys . rept . * 380 * , 235 ( 2003 ) ; j. a. s. lima , bjp * 34 * , 194 ( 2004 ) ; d. h. weinberg , new astron . rev . * 49 * , 337 ( 2005 ) .
astro - ph/0510196 ; e. j. copeland et al , hep - th/0603057 .
k. freese et al . , nucl . phys . *
b287 * , 797 ( 1987 ) ; w. chen and y - s .
wu , phys .
* d41 * , 695 ( 1990 ) ; j. c. carvalho , j. a. s. lima and i. waga , phys . rev . * d46 * 2404 ( 1992 ) ; i. waga , astrophys . j. * 414 * , 436 ( 1993 ) ; j. a. s. lima and j. m. f. maia , phys .
rev * d49 * , 5597 ( 1994 ) ; j. a. s. lima and m. trodden , phys . rev . *
d53 * , 4280 ( 1996 ) ; a. i. arbab and a. m. m. abdel - rahman , phys.rev . *
d50 * , 7725 ( 1994 ) ; l .f .
bloomfield torres and i. waga , mon . not .
soc . * 279 * , 712 ( 1996 ) ; o. bertolami and p. j. martins , phys . rev . *
d61 * , 064007 ( 2000 ) ; r. g. vishwakarma , grg * 33 * , 1973 ( 2001 ) a. s. al - rawaf , mod .
a14 * , 633 ( 2001 ) ; j. v. cunha , j. a. s. lima and j. s. alcaniz , phys . rev . *
d66 * , 023520 ( 2002 ) .
astro - ph/0202260 ; j. s. alcaniz and j. m. f. maia , phys . rev . * d67 * , 043502 ( 2003 ) .
astro - ph/0212510 ; s. carneiro and j. a. s. lima , ijmp * a20 * 2465 ( 2005 ) ; s. carneiro et al .
phys . rev . * d74 * , 023532 ( 2006 ) .
astro - ph/0605607 .
i. l. shapiro and j. sola , phys .
b475 * , 236 ( 2000 ) ; i. l. shapiro and j. sola , jhep , * 0202 * , 006 ( 2002 ) ; j. sola and h. stefancic , astro - ph/0507110 ; i. l. shapiro , j. sola , c. espana - bonet and p. ruiz - lapuente , phys . lett . * b574 * , 149 ( 2003 ) ; i. l. shapiro , j. sola and h. stefancic , jcap * 0501 * , 012 ( 2005 ) . j. a. casas , j. garcia - bellido and m. quirs , class .
* 9 * , 1371 ( 1992 ) ; g. w. anderson , s. m. carroll , astro - ph/9711288 ; l. amendola , mnras * 342 * , 221 ( 2003 ) ; m. pietroni , phys . rev .
* d67 * , 103523 ( 2003 ) ; g. r. farrar and p. j. e. peebles , , 1 ( 2004 ) .
t. padmanabhan and t. r. choudhury , mon . not .
. soc . * 344 * , 823 ( 2003 ) ; j. a. s. lima , j. v. cunha and j. s. alcaniz , phys . rev . *
d68 * , 023510 ( 2003 ) .
astro - ph/0303388 ; p. t. silva and o. bertolami , , 829 ( 2003 ) ; z .- h .
zhu and m .- k .
fujimoto , , 52 ( 2003 ) ; s. nesseris and l. perivolaropoulos , phys .
* d70 * , 043531 ( 2004 ) ; y. wang and m. tegmark , phys . rev .
lett . * 92 * , 241302 ( 2004 ) ; j. v. cunha , j. a. s. lima and j. s. alcaniz , phys .
rev . * d69 * , 083501 ( 2004).astro - ph/0306319 ; j. s. alcaniz and n. pires , phys . rev .
* d70 * , 047303 ( 2004 ) .
astro - ph/0404146 ; t. r. choudhury and t. padmanabhan , astron . astrophys . * 429 * 807 ( 2005 ) ; h. k. jassal , j. s. bagla and t. padmanabhan , phys .
rev . * d72 * , 103503 ( 2005 ) ; d. rapetti , s. w. allen and j. weller , mon . not .
* 360 * , 546 ( 2005 ) ; r. lazkoz , s. nesseris and l. perivolaropoulos , jcap * 0511 * , 010 ( 2005 ) ; b. feng , x. wang and x. zhang , phys .
b607 * , 35 ( 2005 ) ; a. shafieloo , u. alam , v. sahni and a. a. starobinsky , astro - ph/0505329 ; m. a. dantas et al .
, astro - ph/0607060 .
h. k.jassal , j. s. bagla , t. padmanabhan , astro - ph/0601389 .
r. r. caldwell , r. dave , p. j. steinhardt , phys .
80 * , 1582 ( 1998 ) ; s. m. carroll , phys .
, * 81 * , 3067 ( 1998 ) ; l .- m .
wang , r. r. caldwell , j. p. ostriker , and p. j. steinhardt , astrophys .
j. * 530 * , 17 ( 2000 ) ; j. s. bagla , h. k. jassal and t. padmanabhan , phys .
* d67 * , 063504 ( 2003 ) ; k. r. s. balaji and r. h. brandenberger , phys . rev
. lett . * 94 * , 031301 ( 2005 ) ; a. albrecht and c. skordis , phys .
84 * , 2076 ( 2000 ) ; l. r. w. abramo and f. finelli , phys . lett . * b575 * , 165 ( 2003 ) ; c. rubano et al .
d*69 * , 103510 ( 2004 ) ; v. faraoni and m. n. jensen , class .
* 23 * , 3005 ( 2006 ) .
w. fischler , a. kashani - poor , r. mcnees , and s. paban , jhep * 3 * , 0107 ( 2001 ) ; s. hellerman , n. kaloper and l. susskind , jhep * 3 * , 0106 , ( 2001 ) ; j.m .
cline , jhep 0108 , 35 ( 2001 ) ; e. halyo , jhep 0110 , 025 ( 2001 ) .
r. r. caldwell , phys .
b * 545 * , 23 ( 2002 ) ; v. faraoni , int .
d11 * , 471 ( 2002 ) ; s. m. carroll , m. hoffman and m. trodden , phys .
* d68 * , 023509 ( 2003 ) ; r. r. caldwell , m. kamionkowski and n. n. weinberg , phys .
lett . * 91 * , 071301 ( 2003 ) ; p. f. gonzalez - diaz , phys . rev . * d68 * , 021303 ( 2003 ) ; j. s. alcaniz , phys . rev . * d69 * , 083521 ( 2004 ) .
astro - ph/0312424 ; s. nesseris and l. perivolaropoulos , phy . rev . *
d70 * , 123529 ( 2004 ) ; j. a. s. lima and j. s. alcaniz , phys . lett . *
b600 * , 191 ( 2004 ) .
astro - ph/0402265 ; f. c. carvalho and a. saa , phys . rev . * d70 * , 087302 ( 2004 ) ; j. santos and j. s. alcaniz , phys .
b619 * , 11 ( 2005 ) .
astro - ph/0502031 ; l. r. l. abramo and n. pinto - neto , phys .
rev . * d73 * , 063522 ( 2006 ) ; e. m. barboza jr . and n. a. lemos , gr - qc/0606084 .
r. r. caldwell and e. v. linder , phys .
lett . * 95 * , 141301 ( 2005 ) ; r. j. scherrer , phys . rev . *
d73 * , 043502 ( 2006 ) ; t. chiba , phys . rev . *
d73 * , 063501 ( 2006 ) ; v. barger , e. guarnaccia and d. marfatia , phys . lett . *
b635 * , 61 ( 2006 ) .
t. shiromizu , k. maeda , and m. sasaki , phys .
d 62 , 024012 ( 2000 ) ; r. dick , class . quantum grav .
18 , r1 ( 2001 ) ; c. j. hogan , class .
. grav . * 18 * , 4039 ( 2001 ) ; c. j. hogan , astro - ph/0104105 ; k. ichiki , m. yahiro , t. kajino , m. orito and g. j. mathews , phys .
* d66 * , 043521 ( 2002 ) ; k. freese and m. lewis , phys . lett .
* b540 * , 1 ( 2002 ) ; z. -h . zhu and m. -k .
fujimoto , , * 581 * , 1 ( 2002 ) ; d. langlois , astrophys .
space sci . * 283 * , 469 ( 2003 ) ; z. -h .
zhu and m. -k .
fujimoto , , 52 ( 2003 ) ; z. -h .
zhu and m. -k .
fujimoto , , 12 ( 2004 ) .
h. zhang and z .- h .
zhu , astro - ph/0607531 m. d. maia , e. m. monte and j. m. f. maia , phys .
b585 * , 11 ( 2004 ) .
astro - ph/0208223 ; m. d. maia , e. m. monte , j. m. f. maia and j. s. alcaniz , class .
* 22 * , 1623 ( 2005 ) .
astro - ph/0403072 .
c. deffayet , g. dvali and g. gabadadze , phys . rev .
* d65 * , 044023 ( 2002 ) ; p. p. avelino and c. j. a. p. martins , apj * 565 * , 661 ( 2002 ) ; j. s. alcaniz , phys .
d * 65 * , 123514 ( 2002 ) .
astro - ph/0202492 ; d. jain , a. dev and j. s. alcaniz , phys .
rev . * d66 * , 083511 ( 2002 ) .
astro - ph/0206224 ; j. s. alcaniz , d. jain and a. dev , phys .
* d66 * , 067301 ( 2002 ) .
astro - ph/0206448 ; a. lue , phys . rev .
* d67 * , 064004 ( 2003 ) ; a. lue and g. d. starkman , phys . rev . *
d67 * , 064002 ( 2003 ) ; a. lue , r. scoccimarro and g. d. starkman , phys . rev . *
d69 * , 124015 ; e. v. linder , phys .
90 * , 091301 ( 2003 ) ; z. -h .
zhu and j. s. alcaniz , astrophys .
j. * 620 * , 7 ( 2005 ) .
astro - ph/0404201 ; j. s. alcaniz and z. -h .
zhu , phys . rev . *
d71 * , 083513 ( 2005 ) .
astro - ph/0411604 ; n. pires , z .- h . zhu and j. s. alcaniz , phys . rev . * d73 * , 123530 ( 2006 ) .
astro - ph/0606689 .
p. gu , x. wang and x. zhang , phys . rev .
* d68 * , 087301 ( 2003 ) ; r. fardon , a. e. nelson , n. weiner , jcap * 0410 * , 005 ( 2004 ) ; d. b. kaplan , a. e. nelson and n. weiner , phys .
* 93 * , 091801 ( 2004 ) .
m. li , phys . lett . *
b603 * , 1 ( 2004 ) ; s. d. h. hsu , phys . lett . *
b594 * , 13 ( 2004 ) ; b. wang , y .- g . gong and e. abdalla , phys . lett . *
b624 * , 141 ( 2005 ) ; b. wang , c .- y .
lin and e. abdalla , phys . lett . *
b637 * , 357 ( 2006 ) ; z .- y . huang , b. wang , e. abdalla and r .- k .
su , jcap * 0605 * , 013 ( 2006 ) ; j. p. b. almeida and j. g. pereira , phys . lett .
* b636 * , 75 ( 2006 ) ; b. wang , j. zang , c .- y .
lin , e. abdalla and s. micheletti , astro - ph/0607126 ; w. zimdahl and d. pavon , astro - ph/0606555 .
a. kamenshchik , u. moschella and v. pasquier , phys .
b511 * , 265 ( 2001 ) ; n. bilic , g. b. tupper and r. d. violler , phys . lett . *
b535 * 17 ( 2002 ) ; m. c. bento , o bertolami and a. a. sen , phys . rev . * d66 * , 043507 ( 2002 ) ; a. dev , j. s. alcaniz and d. jain , phys .
rev * d67 * , 023515 ( 2003 ) .
astro - ph/0209379 ; j. s. alcaniz , d. jain and a. dev , phys . rev . * d67 * , 043514 ( 2003 ) .
astro - ph/0210476 ; m. makler , s. q. de oliveira and i. waga , phys .
b555 * , 1 ( 2003 ) ; d. carturan and f. finelli , phys . rev . *
d68 * , 103501(2003 ) ; r. r. r. reis , i. waga , m. o. calvao and s. e. joras , phys . rev . * d68 * , 061302 ( 2003 ) ; z .- h .
zhu , astron .
astrophys . * 423 * , 421 ( 2004 ) ; o. bertolami , a. a. sen , s. sen and p. t. silva , mnras * 353 * , 329 ( 2004 ) ; a. dev , d. jain and j. s. alcaniz , astron . astrophys . * 417 * , 847 ( 2004 ) .
astro - ph/0311056 ; m.c .
bento , o. bertolami , m.j .
rebouas and p.t .
silva , phys .
* d73 * , 043504 ( 2006 ) ; h. zhang and z .- h .
zhu , phys . rev . *
d73 * , 043518 ( 2006 ) ; j. a. s. lima , j. v. cunha and j. s. alcaniz , astro - ph/0608469 .
i. prigogine , j. geheniau , e. gunzig and p. nardone , gen .
* 21 * , 767 ( 1989 ) ; m. o. calvao , j. a. s. lima and i. waga , phys . lett . *
a162 * , 223 ( 1992 ) ; j. a. s. lima and a. s. m. germano , phys . lett . *
a170 * , 373 ( 1992 ) ; j. triginer , w. zimdahl and d. pavon , class .
. grav . * 13 * , 403 ( 1996 ) ; j. a. s. lima , j. s. alcaniz , astron . astrophys . * 348 * , 1 ( 1999 ) .
astro - ph/9902337 ; j. s. alcaniz , j. a. s. lima , astron . astrophys . * 349 * , 729 ( 1999).astro - ph/9906410 ; j. a. s. lima , a. s. m. germano , l. r. w. abramo , phys . rev . * d53 * , 4287 ( 1996 ) ; l. p. chimento , a. s. jakubi , d. pavon and w. zimdahl , phys . rev . * d67 * , 083513 ( 2003 ) . y. qiang and t - j .
zhang , astro - ph/0503123 . | the high - quality cosmological data , which became available in the last decade , have thrusted upon us a rather preposterous composition for the universe which poses one of the greatest challenges theoretical physics has ever faced : the so - called dark energy . by focusing our attention on specific examples of dark energy scenarios , we discuss three different candidates for this dark component , namely , a decaying vacuum energy or time - varying cosmological constant [ @xmath0 , a rolling homogeneous quintessence field ( @xmath1 ) , and modifications in gravity due to extra spatial dimensions . as discussed , all these candidates [ along with the vacuum energy or cosmological constant ( @xmath2 ) ] seem somewhat to be able to explain the current observational results , which hampers any definitive conclusion on the actual nature of the dark energy . |
the study of galaxies in rich clusters at earlier epochs has long been seen as one of the best routes to understanding the role of environment in the evolution of galaxy populations , as well as more general issues of galaxy formation .
for this reason a number of groups have studied the photometric and spectroscopic properties of galaxies in distant clusters ( e.g. butcher & oemler 1978 , 1984 ; couch & sharples 1987 ; dressler & gunn 1992 ; aragn - salamanca et al .
1993 ; barger et al .
one of the most interesting conclusions of these studies has been the realisation that the luminous elliptical population which dominates local cluster cores ( dressler 1980 ) has been in place for a considerable time , at least prior to @xmath5 ( smail et al . 1997a ; ellis et al .
1997 ) and possibly earlier than @xmath6 ( faber et al . 1997 ; lubin et al .
all indications are that this population has undergone only passive evolution in the recent past ( van dokkum & franx 1996 ; ellis et al .
1997 ; barger et al .
in contrast to the relative stability of the elliptical population , the other major component of the local cluster population , s0 galaxies , are claimed to be relatively rare in distant clusters ( dressler et al .
the recent formation or transformation of this population and the nature of the process responsible for it are therefore of considerable interest for understanding the extent of environmental influences on galaxy morphology .
the most dramatic evidence for evolutionary change in the cluster population , however , is seen in the blue galaxy populations of these regions ( butcher & oemler 1978 , 1984 ; couch & sharples 1987 ; dressler & gunn 1992 ) . butcher & oemler s extensive study of a heterogenous sample of distant clusters indicated a substantial increase in the blue populations of rich clusters at @xmath7 .
recent hst work on these systems has shown that these objects are mainly disk galaxies , some of which are interacting ( couch et al .
1994 , 1997a , 1997b ; dressler et al . 1994 ; smail et al .
the absence of this population from similar environments in local clusters may connect with the large present - day population of s0s in these regions .
mechanisms proposed for removing or transforming these star - forming galaxies from the centres of clusters include ram - pressure stripping from the icm ( gunn & gott 1972 ) , tidal effects due to the cluster potential ( byrd & valtonen 1990 ) and interactions with other cluster galaxies ( moore et al .
1996 ) . to date , however , few well - defined samples of distant clusters have been available to test the different correlations predicted by these mechanisms ( c.f . the @xmath8@xmath9 relation shown by edge & stewart ( 1991 ) or the mass@xmath9 relation of smail et al .
such samples are also necessary to link the observations at different redshifts , by providing an understanding of the parallel evolution of the structures which the galaxies inhabit . here
we describe a statistical analysis of the galaxy populations in a large , well - defined sample of 10 distant clusters at @xmath100.3 , selected on the basis of their x - ray emission .
our sample focuses on the highest @xmath8 systems , from this we hope to understand the variation in the properties of massive clusters across a relatively narrow mass range , and the effects this has had on the galaxy populations in such regions . in particular , we will concentrate on the variation in the star - formation histories of the luminous elliptical populations between clusters , as determined from their uv optical colours .
this is a technique previously only applied to small samples of 23 clusters ( bower , lucey & ellis 1992 , ble ; ellis et al .
1997 ) , here we can compare a homogeneous sample of 10 clusters imaged in identical conditions to robustly determine the cluster to cluster scatter .
in particular , we illustrate how our combination of @xmath0 imaging is particularly useful for searching for the signature of recent star - formation in more quiescent systems .
we show the advantages of working in the restframe mid
uv for detecting even small traces of past activity in certain populations ( dorman et al .
we also study the blue populations of these clusters , using the large positive k correction in @xmath11 for these galaxies to identify them relative to the passive cluster galaxies .
we focus on the varying proportion of this blue , star - forming population in the clusters and the dependence of this component on the global properties of the clusters .
our final aim is to understand the relationships between the various galactic components of the clusters within the framework of a simple model for their growth and evolution .
this study is particularly timely given recent theoretical work in this area ( bower 1991 ; moore et al . 1996 ; baugh et al . 1996 , 1997 ) .
the integration of spectral evolution models with simple dynamical models for structure formation , allow us for the first time to begin to compare theory and observations in this field ( e.g. baugh et al .
detailed n - body simulations are also now well placed to follow the dynamical evolution of galaxies within cluster potentials ( e.g. moore et al .
1996 , 1997 ) .
it is these types of quantitative comparisons which will benefit from the well - defined cluster sample we use .
this paper is thus part of an on - going program to map out the properties of the galaxy populations in the most massive collapsed structures in the universe as a function of epoch out to @xmath12 .
the data described here deals with the statistical analysis of the galaxy populations in luminous x - ray clusters in the range @xmath13 .
a similar study using very wide field ccd imaging and spectroscopy of a sample of luminous x - ray clusters at @xmath14 is underway ( ohely et al . , in prep ) , this will provide continuous coverage of the evolution of the most massive clusters and their galaxy populations from @xmath150.3 .
previous work linking the galaxy population of distant clusters with their global characteristics has concentrated on the blue , star - forming or `` butcher - oemler '' galaxies in distant clusters ( lea & henry 1988 ) and has found little evidence relating their characteristics to the x - ray properties of the clusters ( although see wang & ulmer 1997 ) . nevertheless , there is still much discussion of the role of a cluster s x - ray - emitting gas in effecting the properties of its galaxy populations .
we note , however , that the necessity for such processes to power the butcher - oemler effect is diminishing .
it is no longer apparent that the star - formation rates seen in the distant bo galaxies are in any way extreme given the characteristics of the surrounding field population ( couch et al .
while the only class of objects thought to be uniquely associated with the bo effect in the distant clusters , the post - starburst galaxy ( psg ) or `` e+a '' , have recently been discovered in the local field ( zabludoff et al .
1996 ) . hence , although our study focuses on a sample of luminous x - ray clusters , our emphasis is to use the x - ray properties of the clusters to provide a well - defined sample of massive clusters and also to give some insight into their dynamical states , rather than to simply test possible interactions between the galaxy populations and the cluster x - ray gas .
we briefly describe the structure of this paper .
section 2 details the observations and their reduction and analysis . in section 3 we catalogue the galaxy populations in our clusters before describing our analysis of these catalogues to study the properties of the cluster populations , both the evolved red spheroidal systems and the star forming galaxies associated with the clusters .
finally , in section 4 we discuss our results and give our main conclusions .
the cluster sample we elected to use for this study is ideally suited for investigating the properties of the galaxy populations in the most massive collapsed structures at @xmath16 .
the sample contains a total of 24 of the most luminous x - ray clusters in the redshift range [email protected] selected from the northern section of the rosat all sky survey , as given in an early compilation of the brightest cluster sample of ebeling et al .
these clusters all have x - ray luminosities in excess of @xmath180.12.4@xmath19 erg s@xmath20 , kms / sec / mpc and @xmath21 unless otherwise stated . ] indicating that their typical masses are : @xmath22 and they thus represent some of the most massive collapsed structures known .
this is also supported by the richnesses of these systems as demonstrated by our analysis below .
the x - ray luminosities derived from rosat hri imaging are presented in table 1 and are discussed in more detail in edge et al .
( 1997a , in prep ) .
the sample has also been subjected to extensive x - ray observations , providing deep , high resolution x - ray images of all the clusters and x - ray spectroscopy of all but one .
these x - ray images show that the majority of the clusters have relatively simple morphologies , the most obvious exception being the bimodal cluster a1758 ( fig .
1 ) which exhibits a highly elongated x - ray morphology .
most importantly for the simple comparison of their galaxy populations , the whole sample spans only a modest range in redshift , minimizing differential k corrections .
we have obtained deep and wide - field , multi - colour imaging of a subsample of 10 of these clusters ( those whose right ascensions lie in the range 1322@xmath23 ) for the analysis here .
one particular feature of this project has been our use of a blue sensitive ccd , allowing us to acquire deep @xmath11-band ( restframe @xmath24 in the mid uv ) images of the clusters .
the @xmath11 filter used for this study is a copy of bessell s @xmath11 ( bessell 1990 ) kindly made available to us by dr .
j. schombert .
our standard johnson @xmath25 and cousins @xmath26 exposures provide complimentary information on the cluster galaxies at restframe wavelengths around @xmath27 and @xmath28 , very close to @xmath11 and @xmath29 respectively .
the catalogue of candidate gravitationally lensed features from our survey is given in edge et al .
( 1997b ) , while the galaxy populations in some of the individual clusters are briefly discussed in smail et al .
( 1995b ) and allen et al .
x - ray imaging of the clusters in this sample are presented in edge et al .
( 1997a ) .
the data discussed here comprise @xmath0 imaging of all 10 clusters .
these data were acquired using the cosmic imaging spectrograph and a thinned 2048@xmath30 tek detector on the 5-m hale telescope at palomar during the nights of june 9 - 12 1994 and july 5 - 7 1994 .
this detector provides 0.284 arcsec sampling over a large , @xmath31 arcmin , field with very good response into the near - ultraviolet .
the bulk of the data analysed here was taken in good conditions during the june run .
the median seeing in the @xmath26-band during these nights was 1.10 arcsec and the nights appeared stable and photometric .
a log giving the field identifications , positions , exposure times and seeing is presented in table 1 .
the reddening values come from the estimated hi column densities in these fields , using the conversion @xmath32 ) .
the exposures typically comprise a 3000 s integration in @xmath11 , along with shorter 500 s integrations in @xmath25 and @xmath26 .
these exposures were split into @xmath334 sub - exposures to allow for cosmic - ray rejection during processing .
observations of landolt ( 1992 ) standard fields were interspersed between the science exposures .
standard iraf reduction procedures were used to process both the science frames and standards .
firstly the images had their bias levels calculated from the overscan region and this subtracted from the exposure , which was then trimmed .
flatfields were constructed from either dome ( @xmath11 ) or twilight observations ( @xmath25 and @xmath26 ) to roughly remove sensitivity gradients across the detector .
having initially flattened the frames , we then cleaned them of the brighter galaxies and stars and used a median algorithm to stack all the cleaned frames in a given filter for each night .
this created an illumination correction frame which was smoothed with a large box filter , normalised and divided through the science frames to remove any remaining mismatch between the `` true '' sky illumination and that removed by the flatfields .
these frames were then aligned and coadded with a rejection algorithm to eliminate the cosmic rays .
examples of the final reduced images are given in fig .
1 . total exposure times and seeing from the final stacked images are listed in table 1 .
analysis of the standard star exposures through the nights confirms our impression that all the nights used for this project were photometric .
photometric calibration of our science exposures indicates average zero point errors contribute to the scatter in the colours at the level of @xmath34 , @xmath35 .
analysis of the colour terms for our filter+detector combinations shows that these are negligible compared to the reddening corrections ( table 1 ) and so we have applied only the latter correction ( @xmath36 and @xmath37 ) . combining the uncertainty in the reddenings ( @xmath38 ) with the zero point errors
, we estimate that our colours are accurate to @xmath39 and @xmath40 on average .
we have tested the accuracy of our photometry ( and the reddening corrections ) using the distribution of reddening - corrected colours for objects identified as stars on our frames . comparing the distribution in
colours for stars identified in each cluster to that given in landolt ( 1992 ) using a two dimensional k - s test we find only two instances of significant ( @xmath41% ) offsets , these are in the @xmath25 photometry of a2390 ( @xmath42 ) and a2261 ( @xmath43 ) . applying these corrections and then combining the stellar colour distributions for 314 stars in the ten clusters we determine limits on the maximum offsets in our mean colours of @xmath44 and @xmath45 .
the errors on these offsets are consistent with those claimed for the individual passbands .
sample which had blue colours , but were very red in .
these objects have @xmath46 , @xmath47 , @xmath48 and @xmath49 , @xmath50 , @xmath51 .
their colours indicate that they are most likely to be @xmath52 qsos , in which case we would estimate a space density of such objects of @xmath53 degree@xmath54 .
the two objects lie in a1758 ( @xmath55 , @xmath56 ) and a2261 ( @xmath57 , @xmath58 ) , at radii of 330 and 240 arcsec from the central galaxies .
the object in a1758 is also detected as an x - ray source in the pspc / hri images of this cluster . ] having aligned and calibrated our @xmath0 images we now create object catalogues from the @xmath26-band images of each cluster .
for this we use the sextractor package ( bertin & arnouts 1996 ) to detect and analyse galaxies on the @xmath26 frames .
we adopt a detection criteria of 10 contiguous pixels each 1@xmath59 ( of the pixel - to - pixel noise ) above the local background , after convolution with a @xmath60 pixel top - hat filter .
these catalogues are then visually inspected and cleaned of spurious objects ( e.g. diffraction spikes and noise objects in the halos of a few very bright stars ) .
the coordinates from the cleaned object catalogues so created are then used to measure colours ( and ) for all the objects detected . before measuring colours , however , we first match the seeing in the @xmath0 frames from the profiles of stars in the images .
we have chosen to use a fixed angular size photometry aperture , 3.0 arcsec in diameter ( 6.87.9 @xmath61 kpc ) , rather than a fixed metric aperture to simplify the field corrections for our catalogues .
we discuss the uncertainties which arise from this later in the paper .
these apertures are sufficiently large that we should detect the bulk of the light from both bulge and disk components of luminous galaxies in these clusters .
the final frames typically cover an area of @xmath62 sq .
arcmin to an 80% completeness limit of @xmath6323.0 ( table 1 ) , where these values are estimated from comparing the observed galaxy counts with those expected from deeper field observations ( smail et al .
1995a ) . to determine galaxy luminosities we use the best magnitude provided by sextractor , which is based on a `` kron - type '' magnitude for brighter galaxies , with a fixed minimum aperture size for fainter objects . in our analysis we restrict ourselves to an @xmath64 sample to ensure @xmath65% completeness in all our fields .
analysis of the other passbands shows that our exposures provide photometry to 20% accuracy to median limits of @xmath66 and @xmath67 .
the final catalogues contain a total of 11,211 objects brighter than @xmath68 over an area of 898 sq .
arcmin , of which 1138 have profiles consistent with being stars , giving an average surface density of @xmath69 galaxies arcmin@xmath54 , where the scatter is field - to - field . to determine the field correction for our cluster frames we used the deep @xmath11 exposure analysed by hogg et al .
this 28.0 ksec exposure covers an area of 81.0 sq .
arcmin in a high - latitude blank field ( table 1 ) and was taken with the same instrument as used for our observations , although using a different @xmath11 filter . to supplement this observation we acquired 500s exposures of the field in @xmath25 and @xmath26 during the night of january 31 1997 ( table 1 ) .
these frames were reduced in the same manner as the cluster observations .
calibration of these data was provided from observations of landolt ( 1992 ) standard stars , observed before and after the science exposures , giving zero point errors of @xmath70 and @xmath71 .
we confirm the accuracy of the colours determined from these fields using the locus of stars on the colour - colour plane as was done for the cluster observations .
the objects in this field were catalogued in a similar manner to the cluster observations , starting with an @xmath26-selected sample and measuring colours in seeing - matched apertures .
the precision of these colours is similar to that achieved in our cluster observations due to equivalent exposure times for the @xmath25 and @xmath26 observations .
the final catalogue contains 377 galaxies brighter than @xmath68 , or a surface density of @xmath72 galaxies arcmin@xmath54 , where the variation is estimated from independent sub - regions within the field .
as we will show the clusters studied here are sufficiently rich , given their redshifts , that field contamination only becomes an important correction at the faintest magnitudes discussed .
to @xmath73 the typical cluster frame contains 315 galaxies , of which 90 ( 28% ) are expected to be field contamination , predominantly blue galaxies , this fraction increases to 40% at @xmath74 , the faintest limit used in this work . in our analysis
we will ignore the effects on the estimated field correction of gravitational amplification of the background field population .
the majority of our analysis focuses on galaxies with red optical colours and for these we expect the effects of lensing to actually lower the background field density ( c.f .
broadhurst et al .
1997 ) , leading to a slight over - correction for field contamination .
we start by describing the general colour distribution of galaxies within our clusters , before identifying particular groups of objects and discussing their properties in more detail in the following sections .
we show in fig .
2 the distribution of galaxies on the colour - colour ( c - c ) plane in each of the 10 clusters , after correcting for field contamination .
we also show the composite colour distribution for all the clusters combined .
all of these plots use only those objects classified as galaxies and with magnitudes brighter than @xmath75 .
no differential k corrections , to compensate for the different cluster redshifts , have been applied to any of these plots . in fig .
3 we present the combined cluster sample as three independent magnitude slices , along with the whole sample . on these plots we have marked the locus of colours expected for the spectral energy distributions ( sed ) representative of the spectral types of galaxies with e , sab , sbc , scd and sdm morphologies in the local universe ( effectively a series of increasing ratio of current to average star - formation rate ) , as they would be observed at @xmath76 .
it can be seen that the clusters contain galaxies spanning the whole range of star - formation rates observed in the local field . at bright magnitudes , however , the distribution is dominated by a population of galaxies whose and colours similar to unevolved elliptical galaxies at these redshifts . as we look at fainter luminosities within the clusters we find a gradual increase in the number of blue galaxies , with colours ranging across those expected from galaxies with star - formation rates similar to local sbc to sdm galaxies .
a similarly wide range in galaxy colours is also seen for the brightest cluster galaxies ( fig .
3 ) , these are typically d or cd galaxies , and although a number of them show the red colours expected for such massive spheroidal systems , three show substantially bluer colours indicating significant on - going star - formation ( allen 1995 ) .
we now identify three groups of cluster galaxies on the colour - colour plane to @xmath74 and
use these to highlight the various components of the cluster populations . at this depth
the median photometric errors are @xmath77 and @xmath78 , with 80% of the objects having errors less than @xmath79 and @xmath80 , adequate to define independent regions of the colour - colour plane .
in particular the accuracy of our photometry allows us to robustly separate samples of galaxies using this colour , which measures the strength of their 4000 breaks .
the three groups we identify are : 1 ) a large population of galaxies with colours similar to passive elliptical galaxies at the cluster redshift ; 2 ) the cluster population with colours typical of star - forming galaxies at @xmath81 ; 3 ) a number of galaxies whose colours indicate a strong 4000 break , representative of an evolved stellar population , but which have bluer colours than group # 1 showing that they are , or recently have been , forming stars . for ease of use we have called these three groups : `` red '' , `` blue '' and `` uv+ '' . to estimate the relative proportions of these groups within the clusters we define rough limits of their boundaries on the , colour - colour plane as : [ 0.1:0.9,2.9:3.6 ] for red , [ @xmath821.1:0.9,1.9:2.9 ] for blue and [ @xmath821.1:0.1,2.9:3.6 ] for uv+ .
note that the exact definitions of the samples used in our analysis below vary according to what questions we are attempting to answer , we have used the regions above simply to illustrate the distribution of objects in the various groups .
the typical proportions of the three cluster populations are then : [ 0.95,0.05,0.00 ] to @xmath83 and [ 0.66,0.29,0.05 ] to @xmath84 .
we give the proportions in the individual clusters in table 2 .
the average radial distributions of the three galaxy populations in the clusters within 500 @xmath61 kpc of the cluster centres are shown in fig .
4 . outside of @xmath85
kpc the profile of the red galaxies drops as @xmath86 , close to the value expected for an isothermal distribution .
the uv+ population follows a radial distribution close to that of the red galaxies , although with a large scatter , @xmath87 . while the blue galaxy distribution is considerably flatter with @xmath88 .
finally , we can identify one further sample of galaxies on the c - c plane : those galaxies with colours substantially redder than the cluster elliptical sequence in ( which provides a good measure of the strength of the 4000 break at the cluster redshifts ) .
this last group consists of galaxies which are probably background to the cluster , as the elliptical sequence represents galaxies with the strongest 4000 breaks at the cluster redshift , and as such we do not discuss them further here .
the number of objects lying substantially redward of the cluster sequence in is consistent with the number expected from our field distributions . the most obvious feature of the plots in figs .
2 and 3 is the strong clump of red galaxies in all the clusters .
the clump is populated with the luminous elliptical galaxies which dominate the bright end of the cluster population .
replotting the catalogues as colour - magnitude ( c - m ) diagrams illustrates another property of this population a very narrow colour - luminosity relation in both @xmath26 and @xmath26 ( fig . 5 ) .
in both panels the elliptical sequence slants down , indicating that the spheroidal galaxies have bluer colours at fainter luminosities , a result of lower mean metallicities in the low luminosity galaxies ( kodama & arimoto 1997 ) . in the fainter samples in fig 3 , however , this population appears to disperse , a feature which may be associated with the apparent `` break '' at @xmath89 in the red elliptical sequence on the @xmath26 plane .
we return to discuss this feature later .
the first question we wish to investigate is the variation in the typical colours of the brighter spheroidal galaxies across the clusters . as we discussed in the introduction ,
the dispersion in the restframe uv optical colours of galaxies is a sensitive test of their recent star - formation , and the variation of the mean colours for the spheroidal population can give powerful constraints on their formation epoch ( ble ; ellis et al .
1997 ) as well as the formation of the larger structures they inhabit . to study this scatter we fit to the colour - luminosity relation in our clusters and then correct these fits to a fiducial cluster at @xmath76 ( the mean of our sample ) to compare the various values .
we fit a linear relation ( @xmath90 or the equivalent for ) to the c - m relation using the same technique applied by ble and ellis et al .
( 1997 ) : huber s robust estimator , after removal of probable field contamination .
galaxies are rejected from the fit interactively until the fit is stable to the inclusion of remaining points , the expected number of field galaxies is used as a guide to indicate the numbers of galaxies to be removed ( typically @xmath91 galaxies , or @xmath92% of the sample ) .
we have confirmed that these fits are not sensitive to modest changes in the definition of the red sub - sample region on the c - c plane .
we then fit the resulting sample across a fixed range in absolute luminosity corresponding to @xmath93 $ ] ( @xmath94 ) .
for simplicity we parameterise the variation in apparent magnitude across the redshift range of our sample , for a @xmath95 galaxy with a non - evolved elliptical sed the apparent magnitude is : @xmath96 .
the fits are evaluated at an absolute magnitude corresponding to @xmath97 at @xmath76 , roughly @xmath98 .
we then apply differential k corrections to the intercept values to convert them to a cluster observed in or at @xmath76 , the corrections assume a non - evolved elliptical sed and are parameterised as : k@xmath99 and k@xmath100 to 2% accuracy .
these corrections amount to @xmath101 in and @xmath102 in , with uncertainties of @xmath103 at a fixed luminosity .
the different k corrections associated with the variation in galaxy colour along the elliptical sequence will introduce a scatter into the corrected colours . from the slope of the c - m relations we expect this to be at the level of @xmath104 and @xmath105 , where the slope is steeper in both colours at higher redshifts .
this is substantially larger than the colour variations introduced between clusters due to our adoption of a fixed aperture size for our photometry , @xmath106 , and so we ignore the effects of the variation in metric aperture size across the different clusters .
we obtain the fits listed in table 3a and 3b ( for and respectively ) . from these
we determine average slopes for the c - m sequence of @xmath107 and @xmath108 , where the scatter within the sample is comparable to the typical errors on the fit .
we have therefore simply adopted these average slopes for all the clusters .
next to simplify the comparison of the clusters we fix the relevant slopes at these average values and refit to our data , across two absolute magnitude ranges : @xmath93 $ ] ( @xmath109 ) and a more restrictive range @xmath110 $ ] ( @xmath111 ) .
the results of these fits ( evaluated at @xmath97 ) are listed in table 4 and we illustrate the values for the complete sample in fig . 6 .
these values show a remarkably small dispersion , @xmath112 and @xmath113 , while for the smaller , bright subsamples we obtain @xmath114 and @xmath115 .
we find no statistically significant correlation between the mean colours in the two passbands .
the typical colours of the bright elliptical cluster population derived above are similar to those expected from a local l@xmath116 elliptical galaxy placed at @xmath76 : @xmath117 and @xmath118 .
the observed colours are offset by @xmath119 and @xmath120 compared to these .
these offsets are in a similar sense to the limits we placed on the relative offset of the colours of stars in our fields compared to the distribution in landolt ( 1992 ) . taking the uncertainty in our absolute colour system into consideration we can only place weak limits on the extent of colour evolution of cluster ellipticals out to @xmath76 .
looking at the average c - m slope we derived from our clusters , we can crudely compare this with that measured in @xmath121 using the slope observed in @xmath122 for spheroidal galaxies in coma by ble .
they find @xmath123 , the expected slope in @xmath124 is @xmath125 ( kodama & arimoto 1997 ) , indicating a predicted slope of @xmath126 , only slightly steeper than the observed value in our distant clusters of @xmath127 .
unfortunately , there is little information on the uv properties of cluster ellipticals shortward of the atmospheric cut - off and so we can not compare the observed slope in with a locally determined value . turning to the scatter in the mean colours of the elliptical sequences between the clusters .
the random errors in the calibration of the clusters are approximately @xmath128 and @xmath129 , removing these contributions from the observed scatter would indicate intrinsic @xmath130 dispersions of @xmath131 and @xmath132 .
we therefore conclude that the red , spheroidal populations brighter than @xmath133 are remarkably homogeneous across all 10 clusters in our sample at a lookback time of @xmath134 gyrs .
we also note that the small spread in colours for the brighter ellipticals in the clusters would allow them to be used as a crude , but economical , redshift indicator provided the colour evolution of the population can be calibrated for a number of clusters spanning the redshift range of interest .
the observed scatter in the mean colour of the bright cluster ellipticals , including measurement errors , is @xmath135 , corresponding to an accuracy in the estimated redshift of @xmath136 ( c.f .
belloni et al . 1995 ) .
a complete analysis of the expected scatter in the mean colour of the elliptical sequences within rich clusters will require a combined spectral modelling and press - schechter approach ( e.g. baugh et al .
nevertheless , we can use a simple , if somewhat contrived , model to understand what general constraints our observations provide on the discreteness of the structures which coalesce to form rich clusters . in the model , clusters form from sub - units ( groups ) of galaxies each containing a fixed number of elliptical galaxies , @xmath137 .
the elliptical galaxies within a given groups have synchronised formation at some epoch ( this could be thought of as relating to the collapse redshift of the group ) which we distribute uniformly in time between the big bang and a cut - off time , @xmath138 , when elliptical formation ceases in all structures .
we define @xmath138 relative to @xmath76 and hence @xmath139 means that elliptical formation continues right up to @xmath76 , while @xmath140 gyrs ( @xmath141 ) corresponds to a time 5 gyrs prior to @xmath76 or roughly @xmath142 . in the press - schechter models
the redshift cut - off and the distributions in relative formation times for ellipticals within the groups would probably both be functions of group mass , as well as cosmology , but the current model is sufficient to illustrate the technique .
we now determine what joint constraints our observed cluster - to - cluster scatter in , restframe @xmath121 , provides on the numbers of structures from which the clusters were built , and the most recent epoch of elliptical formation in these structures .
we use the spectral evolution model of fioc & rocca - volmerange ( 1997 ) for elliptical galaxies to determine the evolution in the rest - frame @xmath121 colours of the galaxies .
starting from the observed number of ellipticals in the cluster , we randomly select groups of @xmath137 galaxies and assign them a formation epoch uniformly in the available time interval , from this we determine their @xmath121 colour at @xmath76 . we repeat this until we have fully populated the cluster and then calculate the mean colour of the elliptical population in the cluster before moving onto the next cluster , until all have their full complements of elliptical galaxies . with values for the mean colours of the elliptical population in each of the 10 model clusters
we next estimate the scatter in the colours _ between
_ the clusters .
this proceedure is repeated 10,000 times for each combination of @xmath137 and @xmath138 to estimate the likelihood that the scatter amongst the 10 model clusters would be less than or equal to that observed for those parameters .
looking at fig 7 it is apparent that as we form the cluster from larger and larger sub - units ( higher @xmath137 ) then the scatter between the clusters should increase .
remembering that all the galaxies within a given sub - unit form together , but that the sub - units themselves can form at any point within the available time period , we can see that this constraint arises from the increasing shot - noise in the average colour of the cluster population when it is formed from a smaller and smaller number of independent sub - units . at one extreme
where each cluster forms as a single group of ellipticals , @xmath143 , which themselves all formed together at a given time , then the scatter between the clusters simply reflects the scatter between the formation times of the original groups and these are therefore constrained to all have formed in a relatively short period of time ( to minimise the scatter between their present day colours ) .
such a model might represent the result of a top - down formation of clusters from large zeldovich pancakes with the bright ellipticals being formed in the first collapse .
looking at fig .
7 we could constrain the elliptical formation to occur at least @xmath1447 gyrs ( @xmath141 ) before @xmath76 or @xmath145 . at the other extreme , where the clusters form from accreting individual elliptical galaxies , @xmath146,whose formation epochs are uncorrelated
, we see that given the large number of ellipticals within each cluster the mean colours of the clusters have a very small scatter and hence the constraint on the last period of star - formation in the elliptical population is much weaker ( equivalent to @xmath147 ) .
taking an intermediate case , if the clusters typically form from sub - clumps and groups containing @xmath92 ellipticals , and the formation of the galaxies within each group is synchronised , then these structures must have formed their galaxies at least 5 gyrs ( @xmath141 ) before @xmath76 ( 95% c.l . ) , equivalent to @xmath148 .
requiring that the bulk of the ellipticals form much earlier than this ( e.g. @xmath52 , ellis et al .
1997 ) would mean that we could only rule out their coherent , stochastic formation in large structures ( @xmath149 ) .
the colours of the bright cluster ellipticals indicate that this population is remarkably homogeneous between clusters .
we can further test this homogeneity by investigating the scatter in the amount of baryonic material locked up as stars in the cluster ellipticals , as compared to the total mass of the cluster .
the simplest method to achieve this is to compare the luminosity in the elliptical sequence to the cluster x - ray temperatures for the 7 clusters with published temperatures ( mushotzky & scharf 1997 ) .
we determine the total spheroidal luminosity in our cluster by simply integrating the light in the cluster sequences defined above down to @xmath133 across the whole field ( an effective radius of @xmath150 mpc in a typical cluster ) .
the values obtained , along with a bootstrap estimate of their uncertainties , are given in table 5 .
we find a reasonable linear correlation ( fig .
8) between the the integrated red galaxy luminosities and the x - ray temperatures : @xmath151 , where the fit has been constrained to pass through [ 0,0 ] .
the dispersion around this line is small , only 17% , ( compared to @xmath152% when @xmath8 is used in place of @xmath153 ) indicating that the integrated luminosity of the red galaxy population is a good tracer of the total mass of the cluster .
in fact the `` true '' scatter may be less when proper account is taken of the roles of mergers and cooling flows in changing the observed x - ray temperatures .
parameterising the ratio of the cluster mass to the luminosity of the elliptical population as : @xmath154 , and assuming an isothermal gas distribution with @xmath155 we find @xmath156 for @xmath157 in solar units and @xmath153 in kev .
thus the observed best fit slope corresponds to @xmath158 in the restframe @xmath159-band . including the observed passive evolution of the stellar populations ( see 3.2 ) this corresponds to an equivalent @xmath160 value of @xmath161 .
this compares well to the value for the coma cluster of @xmath162 within @xmath163 mpc ( transformed from the @xmath25-band value of fusco - femiano & hughes 1994 ) . while the proportion of galaxies falling in the uv+ region of the fig .
3 is small at bright magnitudes , constituting only @xmath164% of the cluster population brighter than @xmath165 , it substantially increases as we reach fainter into the cluster population ( table 2 ) .
the colour selection used to define this sample is sufficiently wide that it will include all cluster galaxies with strong 4000 breaks ( e s0sa ) , which also show blue colours .
the accuracy of our colours means that we expect little contamination of the sample from later spectral types ( with bluer 4000 colours ) .
we now investigate the relation between this population and the apparent break in the c - m relation for the cluster ellipticals in @xmath26 . to do this we first look at the luminosity distribution of galaxies lying along the elliptical c - m sequences defined above .
we select a wedge - shaped region parallel to the c - m sequence with a width varying from @xmath1660.33 ( @xmath1670.43 ) from @xmath16822.0 , the increasing width of the region at fainter limits compensates for the increased photometric errors .
these regions contain 90% of the morphologically classified e and s0 galaxies brighter than @xmath169 in the hst / wfpc-2 image of a2390 ( see 4 )
. using these areas we determine counts from the combined cluster sample and correct these using the same regions of the field galaxy colour - magnitude plane . the magnitudes and colours for the cluster galaxies
are corrected to those of a fiducial cluster at @xmath76 using the k corrections suitable for a non - evolved elliptical sed ( consistent with the typical colours observed for this population ) .
we plot the luminosity distributions determined from both colour - magnitude planes ( fig .
9 ) . to parameterise the differences between these two lfs we have fitted composite gaussian+schechter functions ( e.g. wilson et al . 1997 ) to the distributions and these are over - plotted in fig . 9 .
the only free parameter in the fit is the faint end slope of the schechter function .
the parameters of the gaussian function ( @xmath170 and @xmath171 ) have been fixed to agree with the luminosity function of elliptical galaxies in local clusters ( binggeli , sandage & tammann 1985 , bst ; thompson & gregory 1993 ; biviano et al .
1995 ) , after allowance is made for the expected luminosity evolution of this population .
we derive with that given by colless ( 1989 ) for the distributions in 14 rich clusters , where he estimated @xmath172 for @xmath173 , or @xmath174 assuming a mean color of @xmath175 .
we therefore fit a single schecter function to our distributions having fix @xmath173 and estimate @xmath176 from our distribution and @xmath177 from .
these correspond to @xmath178 , or @xmath179 mag brighter than the @xmath180 value .
this is in good agreement with the expectation from luminosity evolution of a passively evolving stellar population formed at high redshift ( @xmath181 , 3.3 ; barger et al .
the characteristic magnitude of the schechter function is fixed at @xmath169 , @xmath182 .
the main feature of interest in fig .
9 is the difference in the relative numbers of galaxies lying on the red cluster sequence beyond @xmath183 , there being considerably fewer in the sequence .
this is quantified by the different values of the faint end slope , @xmath184 , determined for the two samples , @xmath185 and @xmath186 .
thus we confirm the visual impression from the @xmath26 c - m diagram that there is a decline in the number of red sequence members at magnitudes fainter than @xmath89 ( @xmath187 ) . splitting the selected sample on the basis of radius from the cluster center we see at most a very marginal preference for the decrement to be greater in the cluster centre ( @xmath188 kpc ) .
this is as expected given the similarity of the radial profiles of the uv+ and red galaxies shown in fig .
4 . to search for the cause of the dip we select those galaxies lying along the elliptical sequence in the -@xmath26 c - m plane and study the distribution of their colours .
we show in fig .
10 the distributions for two independent magnitude slices , @xmath18920.5 and @xmath16922.0 , which bracket the position of the dip .
the median photometric errors for galaxies within these two samples are @xmath190 and @xmath191 .
we show the observed distributions in colours in fig 10 , the galaxies lying at @xmath192 are those which correspond to the original definition of `` uv+ '' .
the two samples have intrinsic dispersions of @xmath193 and @xmath194 after removing the contributions from the photometric errors on the objects and the cluster - to - cluster variations in mean colours .
restricting the initial selection to cover a similar colour range for the samples brighter and fainter than @xmath169 does not reduce the colour range seen for the @xmath195 sample . further splitting the @xmath16922.0 sample on the basis of their colours
, we confirm that both halves show similar median colours as well as dispersions : @xmath196 and @xmath197 for the redder and bluer halves respectively , indicating that there is no strong colour - colour correlation within this sub - sample .
we therefore conclude that the red galaxies fainter than @xmath198 have a substantially broader range of mid - uv colours than those galaxies brighter than this limit . to identify morphologically the faint , red cluster population we turn to archival hubble space telescope wfpc-2 imaging .
unfortunately , such images are only available for a small region in one of our clusters , a2390 ks ) of an area of a1758 have been taken but these are useless for our purposes . ] .
the a2390 data comprises 8.4 ks integration in f555w ( @xmath159 ) and 10.5 ks in f814w ( @xmath26 ) , they are discussed further in kneib et al .
( 1997 ) . of the 78 objects with @xmath199 which fall on the c - m sequence used for fig .
10 , only nine lie within the wfpc-2 field , of these we would expect @xmath200 to be field contamination .
the morphologies of these galaxies were visually determined using the scheme of smail et al .
( 1997a ) .
the distribution is 3 e s , 3 s0 s , an s0/a , an sc and a gravitationally lensed arc ( the giant arc in a2390 ) .
the colours of the arc are at the blue end of the distributions , especially in , while the sc galaxy is the reddest of the nine galaxies in , but the second bluest in ( after the arc ) , indicating that it is likely to be background . ignoring these field objects ,
the remaining seven galaxies split roughly equally into s0 and e. with only three galaxies in each class ( we remove the s0/a ) , our conclusions are limited , however , we note that the s0 galaxies tend to be bluer in both ( @xmath201 versus @xmath202 , where the errors on the means are bootstrap estimates ) and ( @xmath203 versus @xmath204 ) , as well as spanning a wider range in colours than the ellipticals , @xmath205 as opposed to @xmath206 . to determine the significance of this difference in the dispersions , we simulate a gaussian distribution with a dispersion equal to that of the combined e+s0 sample and ask how often we expect the ratio of the relative dispersions of random subsets of three galaxies to exceed that observed ( @xmath207 ) .
this occurs in only 2% of cases , which would imply that there maybe a real difference between the colour ranges of e and s0 galaxies .
we conclude that the majority of the cluster galaxies with @xmath183 and red colours are spheroidal systems ( e or s0 ) and that the bluer half of these contains 70% s0 galaxies .
the relatively strong 4000 breaks in the uv+ galaxies means that in the absence of dynamical effects , we would expect them to undergo only modest fading in their restframe optical luminosities between @xmath208 and today , not much above that expected for the elliptical population , @xmath209 .
we therefore would predict that this group would contribute @xmath210% of the population with strong 4000 breaks at typical luminosities of around @xmath211 in massive local clusters .
we compare the expected characteristics of this population with those observed for different local cluster galaxies in 4 .
after the red spheroidal sequence , the next most prominent feature in figs .
3 and 5 is the population of faint , blue galaxies . while some proportion of these objects in fig .
5 will be field galaxies , the distributions shown in fig .
3 have been corrected for this contamination and hence the population of blue galaxies seen there is associated with the cluster , although they do not show the strong central concentration of the red population ( fig . 4 ) .
first we determine the blue fractions ( @xmath212 ) in our clusters , following the prescription of butcher & oemler ( 1984 , bo ) .
they estimate the radius in the cluster which contains 30% of the population brighter than @xmath133 . using this radius
they then calculate the fraction of the @xmath213 population within that aperture which have restframe @xmath214 colours bluer than 0.2 mag below the elliptical sequence . defining the cluster centre from our x - ray images , and correcting for the field counts , we determine the values for @xmath215 and @xmath216 , the number of cluster galaxies within this radius , given in table 5 .
table 5 also gives the concentration index , @xmath217 , as defined by bo .
the only remaining step is to determine the equivalent colour boundary for our passbands .
we have the colour of the elliptical sequence in each cluster ( 3.1 ) , we need to transform an offset of @xmath218 in the restframe into a difference in the observed colour at @xmath208 . for this purpose
we simply fit the variation in restframe @xmath214 colour with observed colour for the non - evolved seds used earlier . from these fits
we can derive the offset which provides the equivalent colour difference in the restframe @xmath214 .
a colour of @xmath219 is between an sab and sbc sed ( 0.67 sab + 0.33 sbc ) and equivalent to an offset of @xmath220 from the observed elliptical c - m sequence . using this limit and the observed elliptical sequences we determine the blue fractions given in table 5 .
we have one cluster in common with the original bo study : a1758 . for this
we derive @xmath221 , @xmath222 arcmin and @xmath223 , in good agreement with the values published by bo : @xmath224 , @xmath225 arcmin and @xmath226 .
the median @xmath212 for the concentrated clusters in our sample , those with @xmath227 , is @xmath228 .
this is slightly lower than the value of @xmath229 for a similar sample of @xmath100.3 clusters in butcher & oemler s original work , perhaps indicating a tendency for these massive clusters to contain a smaller proportion of active galaxies than typically seen in a broad range of collapsed structures at their epochs .
it should be a goal of any future studies to investigate such possibilities by contrasting the properties of galaxy populations in massive and less massive structures at a single epoch .
while the individual cluster cores contain so few luminous , blue galaxies ( typically 23 per cluster ) that the individual values of @xmath212 are not well determined , the presence of such a small number of star - forming galaxies in the core regions of these clusters does mean that they do not show any butcher - oemler effect .
the scatter in our @xmath230 sample is considerable , @xmath231 , and somewhat intriguing given that the clusters are all concentrated and in addition span both a small range in mass and a restricted redshift range .
looking at the concentrated clusters with the two highest @xmath212 values ( we ignore the two @xmath232 clusters ) we can see no obvious distinction between them and the remainder of the sample . similarly focusing on the most bimodal cluster , a1758
, we also see no indication of atypical blue fractions in this system .
evidence has been presented for recent mergers in both a2219 ( smail et al .
1995b ) and a2390 ( pierre et al .
1996 ) from detailed lensing analysis , although the blue populations in neither cluster shows any strong evidence for a recent influx of star - forming galaxies .
we conclude that the wide range in @xmath212 seen in our sample probably results from small accretion events which , while adding a few star - forming galaxies to the cluster , do not substantially alter its morphology , mass or x - ray luminosity .
this is not to say that the accretion of a more massive structure would not effect the cluster , just that such events appear not to substantially alter the mix of galaxies in the cores of massive clusters at @xmath160.3 . in the light of this
we address the recent claim by wang & ulmer ( 1997 , wu ) of a correlation between the blue fraction , @xmath212 , and the ellipticity of the x - ray emission on large scales in a sample of clusters at @xmath2330.6 .
we show the original data from wang & ulmer ( 1997 ) in figure 11 , where we also give similar observations of 5 clusters in our sample ( those for which pspc images are available ) .
we have remeasured the ellipticities of the x - ray emission in all the clusters , both those in wu and the new clusters added here , inside a radius of 750 @xmath61 kpc , to confirm that our new measurements are on the same scale as the earlier values .
we find a mean offset between the wu values and our measurements of : @xmath234 , with a typical deviation of @xmath235 between the estimates .
we conclude that our ellipticity measurements are in good agreement with those presented in wu and hence we can add our clusters to their sample .
the addition of our high luminosity clusters to the already heterogenous sample from wu obviously does not improve the correlation seen in their data ( although neither does it completely destroy it ) .
moreover , we note that the presence of two points at high@xmath236 and high@xmath212 appears to be responsible for much of the significance of their original correlation
. one of these is the cluster a2125 , which has an x - ray luminosity of only a tenth of the next lowest luminosity cluster in the sample , and as such is clearly a very different object to the remaining clusters in the figure . removing this single object to provide
a more homogenous sample also removes any statistically significant correlation between @xmath212 and @xmath236 .
we therefore caution that the @xmath212@xmath236 correlation reported by wang & ulmer ( 1997 ) may be simply the result of a small and diverse sample .
the lack of any correlation between @xmath212 and @xmath236 for luminous x - ray clusters at @xmath237 should not be surprising given the generally low level of butcher - oemler activity in these clusters and their wide range of morphologies .
while our sample of clusters does lack a large population of luminous blue galaxies , looking at the values in table 2 we can see that they do harbour a substantial population of star - forming galaxies , albeit at considerably fainter magnitudes than those used to trace the butcher - oemler effect . as we show below the characteristic @xmath159-band luminosity of the blue populations in these clusters
is considerably fainter than the cluster ellipticals , in contrast to the situation in more distant clusters at @xmath12 where the blue , spiral population have a characteristic luminosity brighter than that of the ellipticals ( smail et al .
this may be hinting at a redshift
luminosity relation for the active populations in the clusters , with the characteristic luminosity of the star - forming class of galaxies increasing with redshift . a similar proposal
has recently been made for the evolution of the star - forming population in the distant field ( lilly et al .
lilly et al .
( 1995 ) claim that the characteristic luminosity of the bluer field galaxies may brighten by up to a magnitude out to @xmath12 . combining this evolution with the increased accretion rate onto massive clusters at moderate redshifts it may be possible to explain the redshift evolution of the butcher - oemler effect using a simple infall model ( bower 1991 ) .
we now look in more detail at the colour distribution within this faint blue cluster population to better understand its subsequent evolution .
we address this using the two dimensional distributions from fig . 3 and project these onto the axis defined by the non - evolved seds seen in that figure .
this allows us to crudely classify the various galaxy populations , in terms of their equivalent present - day spectral type .
we see a broad distribution of spectral classes , the frequency and breadth of this distribution is at odds with that seen at similar luminosities in comparably massive local clusters .
fitting to the luminosity distribution of the population with colours of an sab or bluer in , we find @xmath238 and a steeply rising faint end slope , @xmath239 .
thus the characteristic luminosity of this population is @xmath240 , about 2.5 mag fainter than the elliptical population .
this is in contrast to more distant clusters where the spiral population has a characteristic luminosity similar to that of the cluster ellipticals ( smail et al .
furthermore , the bulk of this population have colours similar to present - day sbc scd galaxies , with comparatively few as blue as would be expected for vigorously star - forming sdm galaxies .
although it should also be kept in mind that the absence of objects as blue as very late - type spirals from our clusters may be partly explained by the limiting magnitude of our sample ( @xmath241 ) . with the advent of mid uv surveys of local clusters ( brosch et al .
1997 ) , it will become possible to trace the evolution of the uv luminosity function ( and hence the mean star - formation rate ) in cluster environments from @xmath160 and compare this to the strong increase in star - formation seen in the distant field , amounting to a factor of @xmath242 increase in the luminosity density at 2800 out to @xmath243 ( lilly et al .
such a comparison would determine whether the mean star - formation rates in clusters evolves in a similar manner to that in the surrounding field , or whether the cluster environment produces an additional long term decline in the star - formation of cluster galaxies at recent epochs .
if the cluster environment is producing a widespread decline in the star - formation of member galaxies , we can ask how the clusters we see at @xmath81 might appear today , in particular where the various galaxy populations we have identified would appear in the luminosity distribution of a local rich cluster . to do this
we need to model the evolution in their star - formation rates with time .
this can only be done approximately , of course , given our lack of knowledge of their detailed dynamical history and previous star - formation .
nevertheless , at this stage it sufficies to determine what the simplest model for the star - formation rate of the cluster population might predict at the present day .
we start with the expectation that the star - formation in all cluster galaxies will decline towards the present , and hence for simplicity we have chosen to truncate the star - formation in all the cluster galaxies at @xmath76 and evolve them forward to the present day , allowing the galaxy s stellar populations to fade and redden by the amount expected from the models .
one attractive model for stopping star - formation in cluster galaxies is to deplete their gas reservoirs by removing their halos through interaction with the intracluster medium ( larson , tinsley & caldwell 1980 ) . while other more drastic alternatives are possible , and may be necessary to form some of the more extreme spectral features observed in some cluster galaxies ( couch et al .
1994 ) , this basic mechanism is sufficient to produce the drop in star - formation required in our simple model .
we will next assume that a galaxy whose colours are similar to a particular spectral type at @xmath76 would , if left undisturbed , retain that spectral type until the present day .
this allows us to use the evolutionary models of fioc & rocca - volmerange ( 1997 ) which fit the local spectral types ( e sdm ) to describe the star - formation histories of the various galaxy populations in our clusters .
we can then take these models and truncate the star - formation at @xmath76 before evolving them to @xmath160 . using fioc & rocca - volmerange s spectral synthesis code ( pegase , fioc & rocca - volmerange 1997 ) we
can then determine the amount of fading and reddening the various spectral types would undergo , in their restframe @xmath29 and @xmath121 respectively .
the evolution of the stellar populations after the star - formation ( if any ) is truncated produces different degrees of subsequent reddening between @xmath244 for an elliptical to @xmath245 for an sdm , the equivalent fading in the restframe @xmath29 is @xmath246 for an elliptical and @xmath247 for an sdm . and @xmath248 , giving a look - back time to @xmath76 of 4 gyrs . ]
the predicted fading and reddening from our model should be viewed as the maximum allowed for a continously star - forming galaxy , both are lower in the event that either the star - formation was halted prior to @xmath76 and we are observing the already fading remnant , or the star - formation continues to lower redshift before being truncated
. however , the predicted fading and reddening can be exceeded if the star - formation is burst - like , as appears to be necessary for modelling some of the spectral properties of distant cluster galaxies ( e.g. couch et al .
more substantial fading of this population would also result if dynamical processes ( stripping , harassment , etc ) removed stellar material from the galaxies .
moreover , we would expect different fading rates across this population in those scenarios where the efficiency of removing material is a function of bulge strength ( i.e. hubble type ) , as appears to be the case for the harassment mechanism ( moore et al .
1996 , 1997 ) .
indeed , the harassment model predicts substantial stripping from later - types ( sdm / irr ) and low surface brightness galaxies , possibly amounting to @xmath24995% of their stellar material , while leaving early - type sab s relatively untouched ( apart from the truncation of the star - formation in their disks ) .
thus it is conceivable that the strong - bulge , early - type spirals ( sab ) would suffer only the fading expected from the stellar population model , whereas the later - type spirals would fade considerably more than our predictions , possibly up to 24 mags ( moore et al .
1997 ) , due to the stripping of stars from these systems .
better predictions from the various theoretical models are required before we can make more detailed comparisons . to begin with
we ignore possible dynamical effects , and simply apply the evolution expected from our stellar models to the original distribution of galaxies in the cluster , as a function of their spectral type ( figure 12 ) .
we show the results of this in figure 13 , where we give both the original distribution at @xmath76 and that which would be observed for the same cluster at @xmath160 . as can be seen in figure 13 , the truncation of all the star - formation in the clusters at @xmath76 results in a large population of faint , mid - type spiral galaxies moving towards the elliptical sequence . concentrating on these fading spirals
we would predict a characteristic magnitude of @xmath250 at the present day , and the population would retain the steep faint end slope characteristic of the original blue cluster population ( @xmath239 ) . adding the dynamically - driven
, type - dependent fading discussed above would tend to produce two families of remnents .
the stronger bulged , and typically brighter , early - type spirals would end up around @xmath251 , perhaps identifying them as precursors to the uv+ group . while the bluer and later - type spirals would fade to @xmath252 , providing a natural source of the large population of dwarf elliptical ( de ) galaxies seen at these magnitudes in local clusters ( e.g. bst ; gregory & thompson 1993 ) .
for completeness we repeat the fit to the luminosity distribution of galaxies lying along the red sequence predicted for the local clusters ( using our simple fading model ) .
we find that the evolution of the blue population leads to a general steepening of the faint end of the distribution .
repeating the fit of a composite gaussian+schechter function to the luminosity distribution along the red sequence , we find that the faint end slope is expected to steepen by @xmath253 from @xmath254 to closer to @xmath255 as a result of the truncation of star - formation at @xmath76 . here
we have evolved the parameters describing the brighter elliptical component to reflect the expected evolution to @xmath160 .
the resulting faint - end slope lies close to that seen in the coma cluster today , @xmath256 ( biviano et al . 1995 ) .
while our simple model has provided a possible outline of where the remnents of the galaxy populations identified in the distant clusters might be found today , we reiterate that much more detailed models , including information on both the dynamics and the previous evolutionary history of the galaxy populations , are needed to conclusively tackle this problem ( c.f .
baugh et al .
we have analysed the galaxy populations in the 10 luminous x - ray clusters in our sample on an individual cluster - by - cluster basis .
the interest of this sample comes from its selection from the most massive clusters at @xmath16 , allowing a simple comparison to the properties of the richest local clusters .
furthermore , using the uniformity of our sample and dataset we have been able to combine the galaxy populations across all the clusters to provide a more detailed view of their typical populations
. the galaxy mix in our clusters at bright magnitudes is characterised by a population of luminous red galaxies , with only a small fraction of blue , star - forming galaxies ( @xmath257% ) and an even smaller fraction of uv+ objects ( @xmath164% ) .
we find that the typical colours of the luminous ( @xmath258 ) red galaxies are highly homogenous across all the clusters in our sample .
the intrinsic scatter in the typical colours of this population between clusters is @xmath2% in restframe @xmath121 . in our bluer colour , restframe @xmath259@xmath260
, the scatter may be larger although our photometric precision is not as high , leading to a limit on the cluster - to - cluster variation of @xmath261% in their mean colours .
these observations thus extend the studies of ble and ellis et al .
( 1997 ) across a larger sample of clusters and into the mid uv .
we conclude that the brightest cluster galaxies comprise a homogeneous population longward of 2900 in their restframe .
we illustrate the type of constraints on cluster formation which this observation can provide using a very simple model .
the model allows us to convert the observed cluster - to - cluster scatter in the mean colours of cluster ellipticals into a joint constraint on the lower limit of the formation epoch for this population and the maximum size of the structures in which they can coherently form . while the model we use is not physically well motivated ,
it does provide some insight into the information on cluster formation which these observations can provide .
a rigorous theoretical analysis of the scatter expected in the colours of luminous ellipticals both within clusters ( e.g. ellis et al .
1997 ) and between clusters is urgently required .
a comparison between the integrated luminosity of the red cluster galaxies with the x - ray temperatures of 7 of the clusters in our sample shows a good correlation between in the two observables with a scatter of only @xmath262% .
we conclude that the total luminosity of the elliptical population in the clusters is a reasonable tracer of the cluster mass .
again this is an interesting observation to tackle with theoretical models of cluster formation and growth , which can be robustly applied to our sample due to the well - defined cluster selection originally used .
the picture of wide spread uniformity abruptly changes as we probe fainter in the clusters .
beyond @xmath263 we start to detect an increasing population of blue cluster galaxies .
these objects fall into two groups , galaxies blue in both and ( about @xmath152% of the population ) and those which while having blue colours are red in : the uv+ galaxies , these are @xmath257% of the total cluster population and comprise @xmath210% of the faint cluster population with red colours . the latter class of galaxies defines an extension of the narrow red sequence of the luminous ellipticals in the c - m plane .
the same is not true in where they exhibit a substantial spread in colours , in contrast to the brighter ellipticals .
a galaxy s flux around 2900 , as probed by our colours , traces residual star - formation ( dorman et al .
1995 ) , being @xmath210% more sensitive to low levels of star - formation than our colour for standard imfs .
thus the wide range in colours for this population indicates a spread in current star - formation , albeit at a low enough level that it does nt substantially perturb their red colours .
the lack of a strong signature in the 4000 colours of these galaxies means they are not included in the standard butcher - oemler definition for the blue fraction of the cluster population , although neither are they completely passive .
the fates of these two blue populations in present - day clusters are of considerable interest .
we propose that , in the absence of dynamical processes or a reactivation of their star - formation , the uv+ population will undergo only modest fading in their optical luminosities from that observed at @xmath81 . as the clusters we study are amongst
the most massive formed at their epoch , we assume that their galaxy populations should evolve to be similar to that seen in the richest local clusters . therefore , these galaxies should contribute a large proportion of the quiescent , red @xmath264 population in the most massive local clusters ( @xmath22 ) .
of the @xmath183 galaxies in the distant clusters with red colours , roughly 50% lie in a blue tail in .
the high proportion of such objects indicates that a large fraction of the quiescent @xmath265 population in local clusters must have passed through this stage . the morphological mix in the coma cluster at these magnitudes is dominated by s0 galaxies , 65% of the population , with ellipticals making up 30% and sa s most of the remaining 5% ( dressler 1980 ) .
thus while some of the uv+ galaxies may continue to form stars at a low level and hence appear as sa galaxies in the local clusters , the majority must be precursors to the s0 or e populations .
moreover , the typical magnitude of these galaxies corresponds to a luminosity close to the characteristic value for the s0 populations in local clusters .
bst claim a mean luminosity of @xmath266 for the s0 galaxies in virgo ( assuming @xmath267 ) , while thompson & gregory ( 1993 ) find @xmath268 for the s0 population of coma . in the light of this and
the observations of a rapid decline in the numbers of s0 galaxies in @xmath269 clusters ( dressler et al .
1997 ) , we suggest that the faint galaxies observed in our @xmath160.3 clusters with red colours , but blue may be the precursors of today s s0 cluster population , whose characteristics they would closely match . by
morphologically classifying a small number of the @xmath183 red and uv+ population using wfpc-2 imaging of a2390 we demonstrate that the faint , red cluster population does consist of spheroidal galaxies and we also show marginal evidence that the s0 component of this population may exhibit a wider range , and typically bluer , mid
uv colours than the faint ellipticals .
the substantial deficit of s0s in clusters at @xmath12 ( dressler et al .
1997 ) , along with our suggested identification of the uv+ population with the precursor s of the s0 population of local rich clusters , would indicate that @xmath15 s0s should show some signs of their recent transformation .
the spectral signatures of recent star - formation have been seen in samples of spheroidal galaxies from the coma cluster ( caldwell et al .
1993 ) , although these have not been linked specifically to the s0 population .
moreover , in a comparison of the dispersions in the @xmath122 colours of luminous ellipticals and s0s in coma , ble concluded that both populations showed essential no intrinsic scatter .
clearly issues of luminosity and morphology must be addressed before we can fully understand the evolution of the galaxy populations of clusters .
central to this understanding will be further work on classifying samples of faint galaxies in @xmath2700.3 clusters , these will enable us to robustly track the morphological evolution of this population ( couch et al .
1997a ) .
the fate of the blue , star - forming galaxies in these distant clusters is more speculative .
these are typically low luminosity systems ( @xmath271 ) , exhibiting relatively modest star - formation .
we illustrate one possible evolutionary pathway , which the population would follow if their star - formation was terminated at @xmath76 , this produces a population of faint , red galaxies in local clusters with a steep luminosity distribution .
these characteristics are similar to those of the dwarf elliptical population which dominates local clusters at faint magnitudes ( bst ) .
recent spectroscopic observations of similar objects in a cluster at @xmath272 has shown them to have very low masses ( koo et al .
1997 ) and so we caution that these low luminosity galaxies may be be quite fragile and hence the cluster environment could have a significant impact on their luminosity and morphological evolution through dynamical processes ( moore et al .
if substantial amounts of baryonic material are removed from late - type galaxies during their accretion onto the clusters we would expect this material to be deposited into the cluster potential as either intracluster gas or intracluster light .
tracing the evolution of these components within rich clusters at @xmath150.5 may provide the most direct test of the stripping and harassment mechanisms ( moore et al .
1997 ) . in conclusion
, there is a growing body of evidence which indicates that the passive red galaxy populations in local clusters are produce by a diverse range of processes .
the most luminous elliptical galaxies appear to have formed at very early epochs ( @xmath52 , ellis et al .
1997 ) , while the bulk of the s0 population have come into being much more recently ( @xmath273 , dressler et al .
1997 ) as may the lower luminosity dwarf ellipticals ( koo et al .
1997 ) . by combining the galaxy samples from our uniform survey of clusters we have identified a class of galaxies in our distant clusters which would have characteristics similar to those of the s0 population at low redshift , but which still show residual traces of star - formation , consistent with their being previously more active .
determining the morphologies of a large sample of these objects in intermediate redshift clusters using hst will provide a strong test of their relationship to both the butcher - oemler populations of distant clusters and the local s0s ( c.f .
larson , tinsley & caldwell 1980 ) .
we would like to thank steve allen , harald ebeling , andy fabian and hans bhringer for their work in defining the x - ray sample on which our study is based .
we also thank jim mccarthy and jim westphal for their work in providing an efficient uv imaging capability on the 5-m , essential for our observations , as well as jim schombert for kindly loaning us his @xmath11 filter , david hogg for providing his deep @xmath11 field exposure , alan dressler for donating 5-m time to undertake the @xmath274 imaging of the same field and david buote for the analysis of the shapes of the cluster x - ray emission used in fig .
we thank the referee , phil james , for his thorough reading of the paper and constructive comments which improved the text , particularly in regard to the discussion of the models .
finally , we thank rebecca bernstein , richard bower , julianne dalcanton and roger davies for extensive discussions on the nature and evolution of cluster galaxies .
irs acknowledges support from a pparc advanced fellowship and ace from a royal society fellowship .
rdb acknowledges nsf grants ast 92 - 23370 and ast 95 - 29170 .
the research published here used the starlink computer resources at the university of durham .
lrrrrr 6@xmath213 & @xmath275 & @xmath276 & @xmath277 & @xmath278 & @xmath279 a1682 & 133.0 & 2.0 & 1.7 & 0.01 & 0.01 a1704 & 69.6 & @xmath824.4 & 1.8 & @xmath820.07 & 0.03 a1758 & 160.1 & 20.1 & 1.9 & 0.11 & 0.01 a1763 & 143.3 & @xmath823.7 & @xmath821.2 & @xmath820.03 & @xmath820.01 a1835 & 148.7 & 11.7 & 5.0 & 0.07 & 0.03 zw7160 & 69.3 & 12.3 & 0.9 & 0.15 & 0.01 a2146 & 66.0 & 0.0 & @xmath821.3 & 0.00 & @xmath820.02 a2219 & 180.0 & 8.0 & @xmath820.3 & 0.04 & @xmath820.00 a2261 & 131.9 & 12.9 & @xmath821.3 & 0.09 & @xmath820.01 a2390 & 105.0 & @xmath825.0 & @xmath821.3 & @xmath820.05 & @xmath820.01 all & 1206.9 & 53.9 & 5.9 & 0.04 & 0.01 6@xmath280 a1682 & 258.9 & 135.7 & 47.3 & 0.31 & 0.11 a1704 & 131.0 & 45.6 & 7.8 & 0.25 & 0.04 a1758 & 225.0 & 127.1 & 8.2 & 0.35 & 0.02 a1763 & 247.4 & 59.1 & 14.5 & 0.18 & 0.05 a1835 & 224.4 & 128.5 & 45.7 & 0.32 & 0.11 zw7160 & 108.6 & 104.6 & 5.4 & 0.48 & 0.02 a2146 & 94.9 & 35.7 & 1.3 & 0.27 & 0.01 a2219 & 307.8 & 83.7 & 13.3 & 0.21 & 0.03 a2261 & 184.7 & 79.3 & 0.2 & 0.30 & 0.00 a2390 & 118.8 & 27.6 & 0.3 & 0.19 & 0.00 all & 1901.5 & 826.9 & 144.0 & 0.29 & 0.05 lcccccc & @xmath281 & @xmath282 & @xmath283 & @xmath284 & @xmath285 & @xmath286 a1682 & 0.226 & 0.473 & 0.017 & @xmath820.021 & 0.016 & 177 a1704 & 0.220 & 0.399 & 0.020 & @xmath820.026 & 0.018 & 107 a1758 & 0.280 & 0.588 & 0.017 & @xmath820.070 & 0.017 & 206 a1763 & 0.228 & 0.500 & 0.015 & @xmath820.055 & 0.014 & 209 a1835 & 0.253 & 0.444 & 0.018 & @xmath820.032 & 0.018 & 200 zw7160 & 0.256 & 0.370 & 0.024 & @xmath820.031 & 0.026 & 126 a2146 & 0.234 & 0.493 & 0.024 & @xmath820.044 & 0.025 & 100 a2219 & 0.228 & 0.501 & 0.013 & @xmath820.018 & 0.012 & 230 a2261 & 0.225 & 0.456 & 0.016 & @xmath820.044 & 0.015 & 197 a2390 & 0.233 & 0.638 & 0.019 & @xmath820.075 & 0.018 & 198 lcccccr & @xmath281 & @xmath287 & @xmath288 & @xmath289 & @xmath290 & @xmath286 a1682 & 0.226 & 3.225 & 0.021 & @xmath820.062 & 0.019 & 173 a1704 & 0.220 & 3.048 & 0.020 & @xmath820.073 & 0.018 & 100 a1758 & 0.280 & 3.088 & 0.019 & @xmath820.077 & 0.021 & 216 a1763 & 0.228 & 3.085 & 0.018 & @xmath820.100 & 0.018 & 202 a1835 & 0.253 & 3.098 & 0.016 & @xmath820.046 & 0.016 & 195 zw7160 & 0.256 & 3.073 & 0.031 & @xmath820.044 & 0.031 & 109 a2146 & 0.234 & 3.105 & 0.020 & @xmath820.092 & 0.020 & 89 a2219 & 0.228 & 3.138 & 0.015 & @xmath820.049 & 0.014 & 221 a2261 & 0.225 & 3.034 & 0.013 & @xmath820.067 & 0.013 & 163 a2390 & 0.233 & 3.121 & 0.020 & @xmath820.066 & 0.019 & 194 lcccccc 7@xmath291 & @xmath282 & @xmath283 & @xmath292 & @xmath287 & @xmath288 & @xmath293 a1682 & 0.472 & 0.022 & 37 & 3.219 & 0.019 & 31 a1704 & 0.389 & 0.028 & 27 & 3.085 & 0.022 & 21 a1758 & 0.606 & 0.016 & 50 & 3.133 & 0.014 & 43 a1763 & 0.490 & 0.022 & 42 & 3.135 & 0.019 & 36 a1835 & 0.440 & 0.018 & 44 & 3.114 & 0.014 & 36 zw7160 & 0.379 & 0.032 & 34 & 3.140 & 0.032 & 17 a2146 & 0.485 & 0.023 & 27 & 3.098 & 0.017 & 26 a2219 & 0.486 & 0.015 & 51 & 3.170 & 0.017 & 42 a2261 & 0.458 & 0.017 & 52 & 3.057 & 0.014 & 45 a2390 & 0.660 & 0.021 & 47 & 3.150 & 0.017 & 38 7@xmath213 a1682 & 0.487 & 0.013 & 177 & 3.210 & 0.009 & 173 a1704 & 0.423 & 0.016 & 107 & 3.086 & 0.009 & 100 a1758 & 0.557 & 0.013 & 206 & 3.125 & 0.009 & 216 a1763 & 0.481 & 0.010 & 209 & 3.115 & 0.007 & 202 a1835 & 0.444 & 0.013 & 200 & 3.133 & 0.007 & 195 zw7160 & 0.373 & 0.016 & 126 & 3.125 & 0.012 & 109 a2146 & 0.475 & 0.016 & 100 & 3.112 & 0.012 & 89 a2219 & 0.506 & 0.011 & 230 & 3.171 & 0.006 & 221 a2261 & 0.446 & 0.014 & 197 & 3.068 & 0.008 & 163 a2390 & 0.614 & 0.013 & 198 & 3.149 & 0.010 & 194 lccccr & c & @xmath215 & @xmath216 & @xmath294 & @xmath212 & & arcmin & & @xmath295 a1682 & 0.32 & 1.75 & 45 & @xmath296 & @xmath297 a1704 & 0.63 & 1.18 & 21 & @xmath298 & @xmath299 a1758@xmath300 & 0.45 & 2.10 & 65 & @xmath301 & @xmath302 a1763 & 0.44 & 1.75 & 53 & @xmath303 & @xmath304 a1835 & 0.42 & 1.37 & 52 & @xmath305 & @xmath306 zw7160 & 0.37 & 2.22 & 36 & @xmath307 & @xmath308 a2146@xmath30 & 0.45 & 0.80 & 14 & @xmath309 & @xmath310 a2219 & 0.40 & 1.56 & 60 & @xmath311 & @xmath312 a2261 & 0.32 & 2.13 & 66 & @xmath313 & @xmath314 a2390 & 0.38 & 1.37 & 49 & @xmath315 & @xmath316 cl 1 ) & c , @xmath215 , @xmath216 and @xmath212 calculated using s.e . component as centre .
2 ) & c , @xmath215 , @xmath216 and @xmath212 calculated using the mean of the position of the two bright galaxies as centre . | we present a deep , multi - colour ( @xmath0 ) ccd survey using the palomar 5-m telescope of a sample of high x - ray luminosity , distant clusters selected from the rosat all - sky survey .
the 10 clusters lie in redshift range @xmath10.28 , an era where evolutionary effects have been reported in the properties of cluster galaxy populations .
our clusters thus provide a well - defined sample of the most massive systems at these redshifts to quantify the extent and variability of these evolutionary effects .
the relatively low redshifts of these clusters also means that simple connections can be made between the galaxy populations of these clusters and their immediate descendents , local rich clusters .
moreover , by concentrating on a narrow redshift range , we can take advantage of the homogeneity of our cluster sample to combine the galaxy catalogues from all the clusters to analyse statistically the bulk properties of their populations .
we present an analysis of the cluster galaxy populations using our multi - colour data to probe the distribution , luminosities and star - formation histories of galaxies in these regions .
our aim is to chart the characteristics of the galaxy populations of massive intermediate redshift clusters and to combine these into a wider scheme for galaxy evolution in high density environments .
the core regions of clusters in our sample contain only a small proportion of star - forming galaxies , and they therefore do not exhibit a classical `` butcher - oemler '' effect . focusing on the redder cluster galaxies we find that their integrated luminosity is well correlated with the cluster s x - ray temperatures , and hence with cluster mass .
furthermore , the typical restframe uv optical colours of the luminous elliptical sequences in the clusters exhibit a remarkably small cluster - to - cluster scatter , @xmath2% , indicating that these galaxies are highly homogeneous between cluster environments .
however , at fainter magnitudes we observe a marked increase in the range of mid uv colours of galaxies possessing strong 4000 breaks , as determined from photometry in @xmath3 kpc diameter apertures . in the light of the apparent decline in the population of s0 galaxies seen in distant , @xmath4 , clusters ( dressler et al .
1997 ) , and in view of the luminosities and colours of this population , we propose that they may be the progenitors of the dominant s0 population of local rich clusters , caught in the final stage before they become completely quiescent .
further studies of this population will provide a necessary link to connect the evolution observed in cluster populations at high redshift with the nature of the final remnants locally .
observations in the restframe uv will be important in these studies owing to the relative ease of detecting the signature of previous activity in this spectral region .
[ firstpage ] cosmology : observations clusters galaxy evolution galaxies : photometry galaxies : luminosity function x - ray astronomy . |
one of the main tools used in the study of dust grain properties is extinction curves , particularly ultraviolet ( uv ) extinction curves .
the uv is where dust extinction is strongest and shows the large variations from region to region in the milky way @xcite , large magellanic cloud ( lmc ) @xcite , and small magellanic cloud ( smc ) @xcite .
the work of @xcite found that most of the variation in milky way extinction curves could be described by an empirically relationship based on the single parameter @xmath3 .
this was a major step forward in our understanding of dust properties and was possible only due to the existence in the literature of near - infrared photometry for a subset of stars in the @xcite sample .
this allowed @xcite to determine @xmath0 values and transform the @xcite extinction curves to an absolute scale ( i.e. , normalized to @xmath4 instead of @xmath5 ) .
these @xmath4 normalized curves had variations which were correlated with @xmath0 allowing @xcite to empirically derive the @xmath0 dependent ccm relationship .
since the @xmath0 value is a rough measure of average dust grain size , this gave a physical basis for the variations in extinction curves .
it is worth noting that significant , small deviations from the ccm relationship are seen for individual sightlines @xcite .
the one caveat on the result of @xcite is that the strength of uv dust extinction limits the measurement of uv dust extinction curves to low to moderate reddening sightlines .
this results in a significant bias in measured uv extinction curves in the milky way to regions surrounding the sun . in the one of the largest , detailed study of uv dust extinction curves in the milky way to date , @xcite presented curves for 78 sightlines and the average distance probed was 1.3 kpc .
the ccm relationship was derived from a subset of the @xcite sample and , as such , this relationship might only be valid for dust in our region of the milky way .
@xcite tested the validity of the ccm relationship for a larger region of the milky way by measuring uv extinction curves along very low - density sightlines .
the 26 extinction curves in their sample had an average distance of 5.2 kpc .
they found that 19 of the 26 extinction curves in their sample were qualitatively consistent with the ccm relationship .
the remaining 7 curves had shapes which are not described by the ccm relationship and were qualitatively similar to those seen in the part of the large magellanic cloud ( lmc ) @xcite associated with the lmc2 supershell ( near the 30 dor star formation region ) .
these 7 curves were all clustered in the same region in the sky and this sightline through the galaxy displays evidence for shocked dust @xcite .
the ccm relationship seems to be a good description of milky way extinction curves with a few exceptions .
this raises the question : does the ccm relationship describe the dust outside of the milky way ?
in other words , are the extinction curves in other galaxies quantitatively similar to those in the milky way ? only in the magellanic clouds can this question be answered as these are the only two galaxies with reliably measured uv extinction curves . a full answer this question requires @xmath0 values for all the magellanic cloud extinction curves and this has only been possible with the recent release of the 2mass observations of the magellanic clouds .
this is the motivation for this paper .
even without all the needed @xmath0 measurements , previous work has gone a long way in answering the above question .
it was quickly realized with the first few measured extinction curves in the lmc and smc the that clouds had curves which were similar to milky way curves as well as curves which were quite different .
for example , the sightlines towards the lmc star sk -69 108 @xcite and the smc star azv 456 @xcite display milky way - like extinction curves . on the other hand , sightlines near 30-dor in the lmc @xcite and in the star forming bar of the smc @xcite show definite non - milky way - like extinction curves especially in their 2175 bump and far - uv rise strengths . in order to move from a qualitative to a quantitative comparison of magellanic cloud and milky way extinction curves , @xmath0 values are needed for each extinction curve .
this allows for the normalization of the extinction curves by @xmath4 instead of the usual @xmath5 which the @xcite work proved was vitally important in understanding the true differences between extinction curves . while , the studies of @xcite and @xcite concentrated on deriving all the uv extinction curves possible with _ international ultraviolet explorer _ ( iue ) archival data in the smc and lmc , respectively , they also presented @xmath0 values for a subset of the smc and lmc curves . ideally , the @xmath0 values for each extinction curve should be derived from near - infrared photometry of both the reddened and comparison stars which make up each curve .
this ensures that the measured extinction curve and @xmath0 value correspond to the same dust column and both are similarly corrected for foreground milky way dust .
due to the paucity of near - infrared photometry for their reddened and ( especially ) comparison stars , a majority of the @xmath0 values presented in @xcite and @xcite were based on assumed near - infrared intrinsic colors .
in addition , the reddened stars colors were not corrected for milky way foreground dust .
this interjected a significant error in the @xmath0 values .
for example , the foreground extinction can be up to 25% of the total extinction for the lmc extinction curves @xcite .
this results in significant differences between @xmath0 values reported in these two studies and previous work @xcite .
the release of the 2mass data for the magellanic clouds makes it possible to correctly and accurately compute @xmath0 values for all the known extinction curves in the magellanic clouds @xcite using the 2mass near - infrared photometry of the reddened and comparison stars .
we combine archival international ultraviolet explorer ( iue ) ultraviolet spectra , optical photometry , and the 2mass @xcite and denis @xcite near - infrared photometry with new hubble space telescope ( hst ) space telescope imaging spectrograph ( stis ) ultraviolet spectroscopy and optical photometry to derive ultraviolet through near - infrared extinction curves for 24 sightlines in the magellanic clouds in [ sec_data ] . in the same section
, we also measure the @xmath0 and column for all 24 extinction curves . in
[ sec_discuss ] , we quantitatively compare milky way and magellanic cloud extinction curves as well as discuss various average magellanic cloud extinction curves and the existence or lack thereof of the 2175 bump in the smc bar .
we present data on all the sightlines in the magellanic clouds which have uv extinction curves .
the sightlines include 23 based on iue data which have been published previously @xcite and one based on stis data which are published for the first time in this paper . for each sightline
, we have gathered uv spectra and optical and near - infrared photometry for both the reddened and comparison stars . from these data ,
we have constructed uv to near - ir extinction curves using the standard pair method and measured @xmath0 and @xmath2 values for each sightline .
the optical and near - infrared photometry for the reddened and comparison stars is given in table [ tab_star_data ] .
ubvri photometric data for all of the smc stars as well as two of the lmc stars ( sk -67 36 and sk -68 26 ) , were obtained on observing runs during 1998 august and september , 1999 january , and 2001 august .
the data were acquired at the 1.5-m telescope of the cerro tololo inter - american observatory ( ctio ) .
a c31034a gaas photomultiplier , ubvri filter set # 3 , and the standard photoelectric data acquisition system were used .
extinction and transformation relations , including non - linear transformation relations were applied to the instrumental data .
the final magnitudes and color indices are on the photometric system defined by @xcite .
the ubv optical photometry for the remainder of the lmc stars was taken from @xcite .
the jhk near - infrared photometry for most of the smc and lmc stars was taken from the results of the 2mass project @xcite .
the jhk photometry for the smc stars azv 456 and azv 462 was taken from @xcite . for two of the three lmc stars without 2mass photometry , we used the denis @xcite jk photometry converted to the 2mass system @xcite .
for the one lmc star without 2mass or denis photometry due to a nearby saturating star , we used jhk images of sk -68 26 and its surrounding region which were taken using the cerro tololo infrared imager ( cirim ) at the ctio 1.5 m .
these images were taken on 15 dec 1999 in non - photometric weather and each band was observed 4 times offsetting between positions for a total exposure time of 4 seconds .
the field - of - view of the coadded images was approximately @xmath6 . in this region , 5 of the bright stars
have 2mass observations and we used these stars to calibrate the jhk fluxes for sk -68 26 using differential photometry .
the uncertainty in these measurements was calculated as the standard deviation of the mean of the the 5 measurements of the jhk magnitudes of sk -68 26 , one measurement per 2mass star .
the uv spectra for all but two of the stars in this paper were taken from archival iue observations .
the specific iue observations we used are given by @xcite for the smc stars and @xcite for the lmc stars .
the individual iue observations were downloaded from the mast archive at space telescope science institute ( stsci ) and were coadded to produce a single spectrum from 1150 to 3225 with a resolution of approximately 400 .
the calibration of these spectra was improved using the results of @xcite who found that the signal - to - noise of iue low - dispersion data could be significantly improved over that provided in the iue archive .
the uv spectra for the final two stars ( azv 23 and 404 ) were taken with the stis instrument on hst as part of our go program # 8198 .
the spectra were taken using the 52x0.5@xmath7 slit with the g140l and g230l gratings .
the individual observations were coadded to produce spectra extending from 1140 to 3140 with a resolution of approximately 1000 .
these two spectra are presented in fig .
[ fig_stis ] . the excellent match between the reddened and comparison spectral types can be easily seen in this figure .
in addition , the superior nature of stis ultraviolet spectra as compared to iue spectra can be seen by comparing this figure with fig . 1 of @xcite which displays iue spectra for similar spectral type stars .
the extinction curves for sightlines in the magellanic clouds were derived using the standard pair method .
the reddened / comparison star pairs used are listed in table [ tab_ext_data ] .
we rederived the 23 extinction curves presented in @xcite and @xcite to take into account the new calibration of iue low - dispersion data @xcite and the new optical and near - infrared photometry .
the effects of the new calibration of the iue spectra on the uv extinction curves was small , mainly reducing the noise in the curves .
the one new extinction curve in the smc for the azv 23 sightline is presented in fig . [ fig_stis ] at it s full spectral resolution .
we calculated all 24 extinction curves and their associated uncertainties as outlined in @xcite .
in addition , we removed the effects of the ly@xmath8 absorption using our measurements of the column ( see [ sec_hi ] ) .
these extinction curves give the difference in extinction between the two sightlines to the reddened and comparison stars .
the milky way foreground component is effectively removed as long as the reddening due to this component towards the reddened and comparison star pairs are similar @xcite .
this results in an extinction curve that only measures dust in the magellanic clouds .
all 24 extinction curves are plotted in figs . [ fig_smc_curves]-[fig_lmc_gen_curves ] .
we fit each curve with the fm parameterization of the shape of the uv extinction curve @xcite .
the fm parameterization is @xmath9 where @xmath10 , @xmath11 and @xmath12 for @xmath13 and @xmath14 for @xmath15 .
we determined the fm parameters for the extinction curves by numerically minimizing the @xmath16 in a manner similar to that used by @xcite .
first , @xmath17 and @xmath18 were fixed and the values of @xmath19 , @xmath20 , @xmath21 , and @xmath22 were determined by minimizing the @xmath16 .
next , @xmath19 , @xmath20 , @xmath21 , @xmath22 , and @xmath18 were fixed and the value of @xmath17 was determined by minimizing the @xmath16 . finally , @xmath19 , @xmath20 , @xmath21 , @xmath22 , and @xmath17 were fixed and the value of @xmath18 was determined by minimizing the @xmath16 .
these two steps were repeated until the @xmath16 no longer changed by a significant amount .
this three step method gives a smaller @xmath16 than doing a single @xmath16 minimization while simultaneously fitting all 6 parameters .
it is worth noting that there are probably only 5 independent parameters in the fm equation . while all evidence points to @xmath19 and
@xmath20 being correlated in the milky way and the magellanic clouds @xcite , we have not assumed this correlation .
for the 4 smc bar extinction curves , we set @xmath23 and @xmath24 as the weak to nonexistent 2175 bumps in these curves precludes fitting all three bump parameters .
this allows for a more realistic measurement of the bump strength or an upper limit for these weak bump sightlines .
the uncertainties in the fm parameters for low @xmath5 sightlines are dominated by the uncertainty in @xmath5 and the random uncertainties of each wavelength point .
the uncertainties in the extinction curves were calculated using eq . 2 of @xcite .
from this equation , it can be seen that an uncertainty in @xmath5 affects the extinction curve in a correlated way .
uncertainties in @xmath5 do not add random noise to each wavelength point , but shift the entire extinction curve up and down .
the random uncertainties at each wavelength point contribute a smaller uncertainty , except in the case of @xmath17 and @xmath18 .
the contribution of uncertainties in @xmath5 ( @xmath25 $ ] ) to the fm parameter uncertainties were calculated by fitting the two additional curves for each measured extinction curve which describe the effects of the @xmath5 uncertainty .
these curves are @xmath26 and @xmath27 .
the @xmath5 uncertainties in each fm parameter were then one half the difference between the parameters for these two curves .
the fm parameter uncertainties due to the random uncertainties ( @xmath28 ) were calculated using a monte carlo approach .
determining the random component of the fm parameter uncertainties is not usually done , but we found that the random component dominates the @xmath17 and @xmath18 uncertainties .
this method consisted of generating 100,000 possible fm fits similar to the best fm fit and determining which ones fit equally well within @xmath29 using the f - test where the @xmath30 is only that due to random flux uncertainties .
the resulting @xmath31 uncertainties were determined by dividing by three .
the reported fm uncertainties ( see table [ tab_fm_param ] ) were determined by summing in quadrature the @xmath5 associated ( @xmath25 $ ] ) and random ( @xmath28 ) uncertainties .
the @xmath5 uncertainties dominate the @xmath19 , @xmath20 , @xmath21 , and @xmath22 uncertainties , while the random uncertainties dominate the @xmath17 and @xmath18 uncertainties .
the value of @xmath3 for each sightline was determined using a @xmath16 minimization method .
this method relies on the invariance of the rijhk portion of the extinction curve @xcite .
the @xmath0 value for each extinction curve is the value which minimizes the @xmath16 between the measured rijhk extinction and the @xcite curve .
the equation giving @xmath32 was solved and the analytic solution for @xmath0 was found . from this equation , the uncertainty for @xmath0 was derived .
see the appendix for details of this derivation .
the @xmath0 values and uncertainties are tabulated in table [ tab_ext_data ] .
our method of using the rijhk extinction curve itself to determine @xmath0 values is different from what is usually done .
usually the observed vrijhk or just vk photometry of the reddened star is compared to assumed intrinsic colors @xcite . in the magellanic clouds ,
the reddened stars are usually first corrected for the average milky way foreground reddening @xcite .
instead , we have used the measured vrijkh photometry for the matched comparison star and avoided having to assume intrinsic colors of the reddened star .
this naturally removes the foreground milky way reddening and ensures that the measured @xmath0 value corresponds to the same dust column as the extinction curve . the ccm relationship for these @xmath0 values for each measured extinction curve
is given in figs .
[ fig_smc_curves]-[fig_lmc_gen_curves ] .
the effects of the uncertainty in @xmath0 on the ccm relationship is shown in these plots using dashed curves .
we determined the column for each extinction curve using a variant of the method outlined in @xcite .
this method relies on fitting the wings of the ly@xmath8 absorption profile .
as we are only interested in a measurement of the difference in the column between the reddened and comparison stars , we made this measurement in the ratio spectrum for each reddened and comparison star pair .
the uncertainty in this measurement was estimated by eye , varying the value until it was noticeable incorrect .
two examples of this are shown in fig .
[ fig_hi_measure ] .
basically , the strength of a model ly@xmath8 profile is adjusted until the division of the ratio spectrum by the model spectrum yields a straight line .
the central wavelength of the ly@xmath8 line for the lmc and smc is corrected for the heliocentric velocity of the lmc and smc ( 280 and 130 km / s , respectively ) .
the central region of the ly@xmath8 line is ignored as it is contaminated by geocoronal emission .
this measurement is straightforward for the stis data . for iue data ,
this measurement is much more uncertain as the contamination by geocoronal emission is much larger and the spectra blueward of ly@xmath8 are very noisy .
the red wing of ly@xmath8 was mainly used in the measurement of columns for the iue data .
the measured values and uncertainties are tabulated in table [ tab_ext_data ] .
now that we have produced full ultraviolet to near - infrared extinction curves ( in units of @xmath33 ) for all the known magellanic cloud reddened / comparison star pairs , we can quantitatively compare them to milky way extinction curves .
the @xmath0 dependent ccm relationship and the ccm sample extinction curves provides a nice , compact form for representing the properties of milky way dust found in the local interstellar medium .
thus , we can compare milky way and magellanic clouds dust by seeing if the ccm relationship is applicable to any of the individual extinction curves or describes aspects of the sample behavior of these curves . to test
if the ccm relationship accurately describes any of the individual extinction curves , we compare the full measured uv to nir extinction curves to the appropriate ccm curve for the measured @xmath0 .
this is done in figs .
[ fig_smc_curves]-[fig_lmc_gen_curves ] and includes curves showing the uncertainty in the measured curve as well as the uncertainty in the ccm curve due to uncertainty in the measured @xmath0 value .
it should be remembered that the ccm relationship gives the average behavior at a particular @xmath0 and that individual curves can have small deviations from this average behavior and still follow the ccm relationship .
it is clear from fig .
[ fig_smc_curves ] that the four curves located in the star - forming bar of the smc ( azv 18 , 23 , 214 , & 398 ) do not follow the ccm relationship . for the one curve located outside of the smc bar ( azv 456 ) ,
the ccm curve does a better job following the measured curve , but there still are significant deviations at the 2175 bump and in the far - uv ( @xmath34 ) .
thus , there are no measured extinction curves in the smc which follow the ccm relationship . in the lmc ,
there is one curve in the lmc2 sample ( sk -69 280 ) and three curves in the lmc - average sample ( sk -66 19 , sk -68 23 , & sk -69 108 ) which follow the ccm relationship within their uncertainties ( figs . [ fig_lmc_lmc2_curves]-[fig_lmc_gen_curves ] ) . thus , there is evidence that the ccm relationship is at work in the lmc , but only in a limited sense as it describes only 4 out of the 19 measured extinction curves . the most direct way to quantify how applicable the ccm relationship is in the magellanic clouds is to plot @xmath35 versus @xmath36 .
this was how the ccm relationship was originally presented by @xcite .
[ fig_al_rv ] shows such plots for @xmath37 1200 , 1500 , 2200 , & 2800 and u and j bands .
since the ccm relationship represents the average extinction behavior as a function of @xmath0 , saying extinction measurements do not follow ccm is to say that they are beyond the scatter of the extinction curves which were used to derive the ccm relationship . in fig .
[ fig_al_rv ] , the small open circles without error bars give data for the extinction curves used to derive ccm and can be used to determine the scatter which is consistent with the ccm relationship .
these plots show that at far - uv wavelengths ( @xmath37 1200 & 1500 ) the ccm relationship forms a lower bound to the values of @xmath33 at a particular value of @xmath0 . at longer wavelengths ( @xmath38 ) , the magellanic clouds measurements are indistinguishable from the ccm relationship within their uncertainties . a final way to probe how well the ccm relationship works in the magellanic clouds is to examine behavior of the fm parameters as a function of @xmath0 .
the fm parameters describe the uv extinction curve with only 6 parameters allowing for more sensitive tests to be performed .
the fm parameters @xmath19 , @xmath20 , @xmath21 , and @xmath22 have a factor of @xmath0 embedded in them , we have plotted the equivalent @xmath0-independent coefficients versus @xmath35 in fig .
[ fig_fm_rv ] .
these @xmath0-independent coefficients can be derived by examining eq .
[ fm_eq ] converted from @xmath39 to @xmath40 units .
the equation expressed this way is @xmath41 the behavior of the @xmath42 , @xmath43 , @xmath44 , and @xmath45 coefficients do not follow the ccm relationship directly , but the ccm relationship does form a bound on the values of these coefficients as a function of @xmath0 .
for example , all the values of @xmath43 lie on or above the line defining the ccm relationship . in the plots of the other three coefficients ,
the ccm relationship forms an upper bound on their behaviors . in the case of @xmath17 and @xmath18
, ccm does not predict much of a dependence on @xmath0 and we do not see one for the magellanic clouds either .
the values of @xmath17 and @xmath18 for the magellanic clouds have a much larger scatter than seen in the extinction curves defining ccm . for @xmath17
this scatter is consistent with the uncertainties in @xmath17 . for @xmath18 ,
the scatter is larger than can be accounted for by measurement uncertainties .
thus , it is either real or the result of measurement uncertainties we have not accounted for in our error analysis . by examining the behavior of the magellanic cloud extinction curves with @xmath0 three different ways and comparing that behavior to that predicted by the @xmath0 dependent ccm relationship
, we find evidence that the general behavior of milky way extinction curves is seen in the magellanic clouds .
not only are 4 lmc extinction curves indistinguishable from milky way extinction curves , but the general behavior of milky way extinction curves forms a bound on the general behavior of magellanic cloud extinction curves .
while a small number of magellanic cloud extinction curves do seem to be well described by the ccm relationship , the majority do not .
yet there is strong evidence that the ccm relationship serves as a bound on the behavior of all the magellanic cloud extinction curves .
this fact raises the question : is there some more general relationship dependent on @xmath0 and at least one other parameter which describes the average behavior of the milky way and magellanic cloud extinction curves ?
an indication that such a relationship might exist was presented by @xcite for low density sightlines in the milky way which display some of the same deviations from ccm seen in the magellanic clouds .
they found that the strength of the 2175 bump ( @xmath46 ) and the steepness of the ultraviolet extinction ( @xmath20 ) were anti - correlated along the low density sightlines as well as for the average curves in the magellanic clouds @xcite .
we present a similar plot in fig .
[ fig_bump_c2 ] for the magellanic cloud extinction curves with the equivalent @xmath0 independent measures of the 2175 bump strength and ultraviolet extinction steepness .
this plot gives evidence for an anti - correlation between the strength of the 2175 bump and the steepness of the ultraviolet extinction .
one possible second parameter could be the measured gas - to - dust ratio ( @xmath47 ) . to test this
we have plotted in fig .
[ fig_hiav ] the gas - to - dust ratios versus the 2175 bump strength and ultraviolet steepness values for the magellanic cloud extinction curves .
the scatter in these two plots is quite large as are the uncertainties on the individual points .
there might be real correlations in both plots , but higher quality data are needed .
there would be reason to expect a correlation between the gas - to - dust ratio and the behavior of extinction curves .
the gas - to - dust ratio is known to correlate with metallicity on a galaxy wide basis @xcite . on a local scale
, it could be a measure of the formation and destruction history of dust grains .
for example , the dust self - shielding will decrease with increasing gas - to - dust ratio making it easier for dust grains to be destroyed by the ambient radiation field . if the dust grains responsible for the 2175 bump are easier to destroy than those responsible for the underlying ultraviolet extinction , a behavior like that seen in fig .
[ fig_hiav ] would be expected .
the extremely weak or absent 2175 bump in the four smc bar extinction curves makes these curves unique . in all other measured extinction curves ,
the 2175 bump is quite prominent .
the obvious question is whether the weak 2175 bump in the smc bar is due to milky way contamination or is intrinsic to the smc .
while only one of the four curves has a bump which is detected at greater than @xmath29 , the other three are all @xmath48 detections .
the construction of the curves using smc comparison stars should remove all the milky way foreground , but small differences between the reddened and comparison stars foreground could result in a weak 2175 bump . a foreground contamination like
this would result in a positive bump as often as a negative bump .
interestingly , one out of the four smc bump curves ( azv 214 ) has a @xmath48 detection of a negative bump and a negative far - uv curvature ( @xmath22 ) .
the foreground contamination ( in percent of total @xmath5 with @xmath49 uncertainties ) needed to produce spurious bump detections at the observed levels would have to be ( @xmath50)% , ( @xmath51)% , ( @xmath52)% , and ( @xmath53)% for azv 18 , 23 , 214 , and 398 , respectively .
these foreground contaminations are not unlikely given the low @xmath5 values for these sightlines , especially given the large uncertainties .
while it is difficult to definitively decide , the current evidence suggests that the carriers of the 2175 bump could be completely absent from the dust in the smc bar sightlines . since the magellanic cloud extinction curves display significantly different shapes than that seen in the milky way , average extinction curves of the lmc - average , lmc2-supershell , and smc - bar samples are of interest .
these average curves capture the large scale variations in dust properties at higher signal - to - noise than the individual curves .
while past work has produced such average curves @xcite , we can create more accurate average curves as the result of our determinations of @xmath0 values .
we have averaged the individual curves in each sample in @xmath33 units instead of @xmath54 units as was done in previous work .
@xmath54 units are a relative measure of dust properties whereas @xmath33 units are an absolute measure of the dust properties and , as a result , more accurately represent the effects of dust . for the lmc2 supershell sample
, we have not used the results for sk -69 256 in producing the lmc2 supershell average .
it s extremely low @xmath0 value of @xmath55 is likely the result of poor jhk photometry , contamination by nearby hot dust , or contamination by a red companion .
the sample average curves are given in figure [ fig_ave_curves ] and tabulated in table [ tab_ave_curves ] .
the averages were calculated weighting by the uncertainties at each wavelength for each individual curve and in 0.25 @xmath56 wide bins .
the average values of @xmath0 and n()/a(v ) were also calculated weighting by the individual uncertainties and are tabulated in table [ tab_ext_data ] .
the average curves were fit with the fm parameterization and the best fit 6 fm parameters are given in table [ tab_fm_param ] .
the fm fits were done for wavelengths @xmath57 @xmath58 as this is the bluest wavelength in common between the iue and stis data .
this fitting limit only affected the smc bar average . for the smc bar ,
we find that @xmath59 and n()/a(v ) @xmath60 which is consistent with the results of @xcite and @xcite
. however , we do not find that the gas - to - dust ratio for azv 456 ( sk 143 ) is similar to that of the milky way .
for the milky way , n()/a(v ) @xmath61 @xcite and we find n()/a(v ) @xmath62 for azv 456 whereas @xcite found n()/a(v ) @xmath63 assuming @xmath64 ( the average @xmath0 for the entire smc ) .
the differences between our measurement and @xcite can be traced to our lower @xmath5 and a lower @xmath0 value .
these differences illustrate the difficulties of measuring such quantities for low reddening sightlines . for the lmc2 supershell sample ,
we find that @xmath65 and n()/a(v ) @xmath66 .
our gas - to - dust ratio is roughly consistent with results of @xcite and @xcite who found n()/a(v ) @xmath67 and @xmath68 , respectively .
our value of @xmath0 is lower than that of @xcite who found @xmath69 but consistent with @xcite . for the lmc average sample , we find that @xmath70 and n()/a(v ) @xmath71
. these values are significantly different from those for the lmc2 supershell sample .
the different gas - to - dust ratios is in conflict with the conclusions of @xcite who state that there are no gas - to - dust measurable differences between the non-30 dor and 30 dor samples .
but , examination of their fig .
6 shows that the non-30 dor points fall consistently below 30 dor gas - to - dust ratio implying that we are not in conflict with earlier work .
the lower gas - to - dust ratio is especially interesting as it supports a systematic trend of more extreme extinction curves ( smaller 2175 bump and steeper far - uv rise ) with rising gas - to - dust ratio ( see [ sec_sccm ] ) .
we have presented a quantitative , exhaustive comparison of all the known magellanic cloud extinction curves with a representative sample of milky way extinction curves . like previous studies @xcite
we find that both the lmc and smc have examples of extinction curves qualitatively similar to those found in the milky way . unlike previous studies , we are able to take this comparison one step further and make quantitative comparisons as we determined @xmath0 values for all magellanic cloud extinction curves .
this allows for the comparison of magellanic cloud and milky way extinction curves to be done using measurements based on absolute dust properties ( @xmath33 ) instead of relative dust properties ( @xmath54 ) .
the importance of this difference is well illustrated by the work of @xcite where the comparison of milky way curves in @xmath33 units allowed for the derivation of the @xmath0 dependent ccm relationship .
we conclude that 4 extinction curves in the lmc are _ indistinguishable _ from milky way extinction curves and milky way extinction curves form a bound on the behavior of magellanic cloud extinction curves .
the majority of the magellanic cloud extinction curves are significantly different than milky way extinction curves and this is likely a result of magellanic cloud extinction curves probing quite different environments than the ccm sample of milky way extinction curves .
the ccm sample is based on fairly quiescent sightlines and most of our extinction curve measurements in the magellanic clouds are biased towards quite active regions as they are based on ob supergiants .
the systematic behavior of the curves studied in this paper hint at the existence of a multiparameter relationship ( possibly dependent on @xmath0 and @xmath47 ) describing both quiescent ( ccm - like ) and active extinction curves .
the different biases between the majority of the measured milky way and magellanic cloud extinction curves point to a weakness in our understanding of dust properties .
the magellanic cloud extinction curves are biased towards active star forming regions ( ob supergiant sightlines ) and low dust columns .
the milky way sightlines are less biased towards active star forming regions as they are mainly measured along sightlines towards ob main sequence stars .
they are limited in their variety of dust properties probed as the measurements are generally limited to our local region of the milky way .
evidence that the type of dust seen in the magellanic clouds does exist in the milky way is seen along milky way low density sightlines @xcite .
higher dust column sightlines and sightlines towards main sequence stars need to be measured in the magellanic clouds to allow us to truly explore the possible range of dust properties .
the results of this work and those of @xcite imply that the common usage of discussing milky way , lmc , and smc dust as separate is inaccurate .
in reality , a continuum of dust properties exists between that seen in the quiescent and active regions ( eg .
, see fig . [ fig_bump_c2 ] ) .
the usual average curves can then be arranged into a rough version of this sequence and it would be milky way ( ccm ) lmc average
smc wing ( azv 456 )
lmc2 supershell smc bar .
ordering all the individual known extinction curves would produce an even greater mixing of milky way , lmc , and smc curves .
this work describes a qualitative view of the continuum of dust extinction curves .
quantifying the variation between quiescent and active region extinction curves would give valuable clues to the identities of dust grain properties .
support for proposal # 8198 was provided by nasa through a grant from the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas 5 - 26555 .
we thank the referee for comments which motivated us to improve the presentation of this paper .
this work was partially supported through jpl contract # 960785 .
aul acknowledges support from nsf grant ast0097895 .
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 .
this research made use of the simbad database and vizeir catalogue access tool , cds , strasbourg , france .
the derivation of the equations to determine value of @xmath0 and its uncertainty are given in this appendix . to the knowledge of the authors
, these equations have not appeared in the literature before and are included in this paper as they may be of use to others .
the equations below are equivalent to determining @xmath0 from individual colors and averaging the results weighted by the appropriate uncertainties .
the equation giving the @xmath16 between the measured extinction curve and the @xcite curve is @xmath72 ^ 2\ ] ] where @xmath73 r_v\end{aligned}\ ] ] and @xmath74 is given in table 3 of @xcite .
differentiating @xmath16 with respect to @xmath0 and setting the result to zero gives @xmath75 } { \sum_i \left [ y(\lambda_i)/\sigma(\lambda_i)^2 \right]}.\ ] ] using eq .
6.19 of @xcite gives the uncertainty in @xmath0 as @xmath76 } { \sum_i \left\ { \left [ a(\lambda_i)/a(v ) - 1 \right]^2 / \sigma(\lambda_i)^2 \right\ } } .\ ] ] aiello , s. , barsella , b. , chlewicki , g. , greenberg , j. m. , patriarchi , p. , & perinotto , m. 1988 , , 73 , 195 azzopardi , m. , & vigneau , j. 1982 , , 50 , 291 bevington , p. r. & robinson , d. k. 1992 , data reduction and error analysis for the physical sciences ( new york : mcgraw - hill , inc . ) bohlin , r. 1975 , , 200 , 402 bohlin , r. c. , savage , b. d. , & drake , j. f. 1978 , , 224 , 132 bouchet , p. , lequeux , j. , maurice , e. , prevot , l. , & prevot - burnichon , m. l. 1985 , , 149 , 330 calzetti , d. , kinney , a. l. , & storchi - bergmann , t. 1994 , , 429 , 582 cardelli , j. a. , clayton , g. c. , & mathis , j. s. 1989 , , 345 , 245 ( ccm ) clayton , g. c. , gordon , k. d. , & wolff , m. j. 2000 , , 129 , 147 clayton , g. c. & martin , p. g. 1985 , , 288 , 558 cioni , m - r .
2000 , , 144 , 235 cutri , r. m. et al . 2000 ,
explanatory supplement to the 2mass second incremental data release diplas , a. & savage , b. d. 1994 , , 427 , 274 epchtein , n. et al . 1999 , , 349 , 236 fitzpatrick , e. l. 1985 , , 299 , 219 fitzpatrick , e. l. 1986 , , 92 , 1068 .
fitzpatrick , e. l. & massa , d. 1988 , , 328 , 734 fitzpatrick , e. l. & massa , d. 1990 , , 72 , 163 ( fm ) garmany , c. d. , conti , p. s. & massey , p. 1987
, , 93 , 1070 gordon , k. d. , calzetti , d. , & witt , a. n. 1997 , , 487 , 625 gordon , k. d. & clayton , g. c. 1998 , , 500 , 816 issa , m. r. , maclaren , i. , & wolfendale , a. w. 1990 , , 236 , 237 koornneef , j. 1982 , , 107 , 247 landolt , a. u. 1992 , , 104 , 340 lennon , d. j. 1997 , , 317 , 871 lequeux , j. , maurice , e. , prvot - burnichon , m .-
l . , prvot , l. , rocca - volmerange , b. 1982 , , 113 , l15 martin , p. g. & whittet , d. c. b. 1990 , , 357 , 113 massa , d. & fitzpatrick , e. l. 2000 , , 126 , 517 massey , p. , lang , c. c. , degioia - eastwood , k. , & garmany , c. d. 1995 , , 438 , 188 mathis , j. s. & cardelli , j. a. 1992 , , 398 , 610 misselt , k. a. , clayton , g. c. , & gordon , k. d. 1999 , , 515 , 128 morgan , d. h. & nandy , k. 1982 , , 199 , 979 nandy , k. & morgan , d. h. 1978 , , 276 , 478 neubig , m. m. s. , & bruhweiler , f. c. 1997 , , 114 , 1957 prvot , m. l. , lequeux , j. , maurice , e. , prvot , l. , rocca - volmerange , b. 1984 , , 132 , 389 rieke , g. h. , & lebofsky , m. j. 1985 , , 288 , 618 skrutskie , m. f. et al .
1997 , assl vol . 210 : the impact of large scale near - ir sky surveys , 25 witt , a. n. , bohlin , r. c. , & stecher , t. p. 1984
, , 279 , 698 zaritsky , d. , harris , j. , & thompson , i. 1997 , , 114 , 1002 zaritsky , d. , harris , j. , thompson , i. b. , grebel , e. k. , & massey , p. 2002
, , 123 , 855 .
lrrrrrrrr + azv 18 & @xmath77 & @xmath78 & @xmath79 & @xmath80 & @xmath81 & @xmath82 & @xmath83 & @xmath84 + azv 23 & @xmath85 & @xmath86 & @xmath87 & @xmath88 & @xmath89 & @xmath90 & @xmath91 & @xmath92 + azv 70 & @xmath93 & @xmath94 & @xmath95 & @xmath96 & @xmath97 & @xmath98 & @xmath99 & @xmath100 + azv 214 & @xmath101 & @xmath102 & @xmath103 & @xmath104 & @xmath105 & @xmath106 & @xmath107 & @xmath108 + azv 289 & @xmath109 & @xmath110 & @xmath111 & @xmath112 & @xmath113 & @xmath114 & @xmath115 & @xmath116 + azv 380 & @xmath117 & @xmath118 & @xmath119 & @xmath120 & @xmath121 & @xmath122 & @xmath123 & @xmath124 + azv 398 & @xmath125 & @xmath126 & @xmath127 & @xmath128 & @xmath129 & @xmath130 & @xmath131 & @xmath132 + azv 404 & @xmath133 & @xmath134 & @xmath135 & @xmath136 & @xmath137 & @xmath138 & @xmath139 & @xmath140 + azv 456 & @xmath141 & @xmath142 & @xmath143 & @xmath144 & @xmath145 & @xmath146 & @xmath147 & @xmath148 + azv 462 & @xmath149 & @xmath150 & @xmath151 & @xmath152 & @xmath153 & @xmath154 & @xmath155 & @xmath156 + + sk -65 15 & @xmath157 & @xmath158 & @xmath159 & & & @xmath160 & @xmath161 & @xmath162 + sk -65 63 & @xmath163 & @xmath164 & @xmath165 & & & @xmath166 & @xmath167 & @xmath168 + sk -66 19 & @xmath169 & @xmath170 & @xmath171 & & & @xmath172 & @xmath173 & @xmath174 + sk -66 35 & @xmath175 & @xmath176 & @xmath177 & & & @xmath178 & @xmath179 & @xmath180 + sk -66 88 & @xmath181 & @xmath182 & @xmath183 & & & @xmath184 & @xmath185 & @xmath186 + sk -66 106 & @xmath187 & @xmath188 & @xmath189 & & & @xmath190 & @xmath191 & @xmath192 + sk -66 118 & @xmath193 & @xmath194 & @xmath195 & & & @xmath196 & @xmath197 & @xmath198 + sk -66 169 & @xmath199 & @xmath200 & @xmath201 & & & @xmath202 & @xmath203 & @xmath204 + sk -67 2 & @xmath205 & @xmath206 & @xmath207 & & & @xmath208 & @xmath209 & @xmath210 + sk -67 5 & @xmath211 & @xmath212 & @xmath213 & & & @xmath214 & @xmath215 & @xmath216 + sk -67 36 & @xmath217 & @xmath218 & @xmath219 & @xmath220 & @xmath221 & @xmath222 & @xmath223 & @xmath224 + sk -67 78 & @xmath205 & @xmath225 & @xmath226 & & & @xmath227 & & @xmath228 + sk -67 100 & @xmath229 & @xmath230 & @xmath195 & & & @xmath231 & @xmath232 & @xmath233 + sk -67 168 & @xmath234 & @xmath235 & @xmath201 & & & @xmath236 & @xmath237 & @xmath238 + sk -67 228 & @xmath239 & @xmath194 & @xmath240 & & & @xmath241 & @xmath242 & @xmath243 + sk -67 256 & @xmath244 & @xmath188 & @xmath245 & & & @xmath246 & @xmath247 & @xmath248 + sk -68 23 & @xmath249 & @xmath250 & @xmath251 & & & @xmath252 & @xmath253 & @xmath254 + sk -68 26 & @xmath255 & @xmath256 & @xmath257 & @xmath258 & @xmath259 & @xmath260 & @xmath261 & @xmath262 + sk -68 40 & @xmath263 & @xmath176 & @xmath264 & & & @xmath265 & @xmath266 & @xmath267 + sk -68 41 & @xmath268 & @xmath269 & @xmath270 & & & @xmath271 & @xmath272 & @xmath273 + sk -68 129 & @xmath274 & @xmath275 & @xmath276 & & & @xmath277 & @xmath278 & @xmath279 + sk -68 140 & @xmath280 & @xmath281 & @xmath282 & & & @xmath283 & @xmath284 & @xmath285 + sk -68 155 & @xmath280 & @xmath275 & @xmath240 & & & @xmath286 & @xmath287 & @xmath288 + sk -69 108 & @xmath289 & @xmath290 & @xmath291 & & & @xmath292 & & @xmath293 + sk -69 206 & @xmath294 & @xmath295 & @xmath296 & & & @xmath297 & @xmath298 & @xmath299 + sk -69 210 & @xmath300 & @xmath301 & @xmath302 & & & @xmath303 & @xmath304 & @xmath305 + sk -69 213 & @xmath306 & @xmath307 & @xmath308 & & & @xmath309 & @xmath310 & @xmath311 + sk -69 228 & @xmath312 & @xmath313 & @xmath296 & & & @xmath314 & @xmath315 & @xmath316 + sk -69 256 & @xmath317 & @xmath275 & @xmath282 & & & @xmath318 & @xmath319 & @xmath320 + sk -69 265 & @xmath321 & @xmath170 & @xmath322 & & & @xmath323 & @xmath324 & @xmath325 + sk -69 270 & @xmath326 & @xmath295 & @xmath327 & & & @xmath328 & @xmath329 & @xmath330 + sk -69 279 & @xmath169 & @xmath313 & @xmath276 & & & @xmath331 & @xmath332 & @xmath333 + sk -69 280 & @xmath334 & @xmath335 & @xmath336 & & & @xmath337 & @xmath338 & @xmath339 + sk -70 116 & @xmath340 & @xmath341 & @xmath342 & & & @xmath343 & @xmath344 & @xmath345 + sk -70 120 & @xmath346 & @xmath347 & @xmath177 & & & @xmath348 & @xmath349 & @xmath350 + llcccr + azv 18 & azv 462 & @xmath351 & @xmath352 & @xmath353 & @xmath354 + azv 23 & azv 404 & @xmath355 & @xmath356 & @xmath353 & @xmath357 + azv 214 & azv 380 & @xmath358 & @xmath359 & @xmath360 & @xmath361 + azv 398 & azv 289 & @xmath362 & @xmath363 & @xmath364 & @xmath365 + & & @xmath366 & & @xmath367 + + azv 456 & azv 70 & @xmath368 & @xmath369 & @xmath370 & @xmath371 + + sk -68 140 & sk -68 41 & @xmath372 & @xmath373 & @xmath374 & @xmath375 + sk -68 155 & sk -67 168 & @xmath372 & @xmath376 & @xmath377 & @xmath378 + sk -69 228 & sk -65 15 & @xmath379 & @xmath380 & @xmath381 & @xmath382 + sk -69 256 & sk -68 41 & @xmath383 & @xmath384 & @xmath360 & @xmath385 + sk -69 265 & sk -68 40 & @xmath386 & @xmath387 & @xmath388 & @xmath389 + sk -69 270 & sk -67 228 & @xmath386 & @xmath390 & @xmath391 & @xmath392 + sk -69 279 & sk -65 63 & @xmath393 & @xmath394 & @xmath374 & @xmath395 + sk -69 280 & sk -67 100 & @xmath396 & @xmath397 & @xmath398 & @xmath399 + sk -70 116 & sk -67 256 & @xmath386 & @xmath400 & @xmath381 & @xmath401 + & & @xmath402 & & @xmath403 + + sk -66 19 & sk -66 169 & @xmath404 & @xmath405 & @xmath406 & @xmath407 + sk -66 88 & sk -66 106 & @xmath408 & @xmath409 & @xmath410 & @xmath411 + sk -67 2 & sk -66 35 & @xmath379 & @xmath412 & @xmath413 & @xmath414 + sk -68 23 & sk -67 36 & @xmath415 & @xmath416 & @xmath417 & @xmath418 + sk -68 26 & sk -66 35 & @xmath419 & @xmath420 & @xmath381 & @xmath421 + sk -68 129 & sk -68 41 & @xmath383 & @xmath422 & @xmath374 & @xmath423 + sk -69 108 & sk -67 78 & @xmath424 & @xmath425 & @xmath426 & @xmath427 + sk -69 206 & sk -67 5 & @xmath428 & @xmath429 & @xmath364 & @xmath430 + sk -69 210 & sk -66 118 & @xmath431 & @xmath432 & @xmath433 & @xmath434 + sk -69 213 & sk -70 120 & @xmath435 & @xmath436 & @xmath417 & @xmath437 + & & @xmath438 & & @xmath439 + lrrrrrr + azv 18 & @xmath440 & @xmath441 & @xmath442 & @xmath443 & @xmath444 & @xmath445 + azv 23 & @xmath446 & @xmath447 & @xmath448 & @xmath449 & @xmath444 & @xmath445 + azv 214 & @xmath450 & @xmath451 & @xmath452 & @xmath453 & @xmath444 & @xmath445 + azv 398 & @xmath454 & @xmath455 & @xmath456 & @xmath457 & @xmath444 & @xmath445 + average & @xmath458 & @xmath459 & @xmath460 & @xmath461 & @xmath444 & @xmath445 + + azv 456 & @xmath462 & @xmath463 & @xmath464 & @xmath465 & @xmath466 & @xmath467 + + sk -68 140 & @xmath468 & @xmath469 & @xmath470 & @xmath471 & @xmath472 & @xmath473 + sk -68 155 & @xmath474 & @xmath475 & @xmath476 & @xmath477 & @xmath478 & @xmath479 + sk -69 228 & @xmath480 & @xmath481 & @xmath482 & @xmath483 & @xmath484 & @xmath485 + sk -69 256 & @xmath486 & @xmath487 & @xmath488 & @xmath489 & @xmath490 & @xmath491 + sk -69 265 & @xmath492 & @xmath493 & @xmath494 & @xmath495 & @xmath496 & @xmath497 + sk -69 270 & @xmath498 & @xmath499 & @xmath500 & @xmath501 & @xmath502 & @xmath503 + sk -69 279 & @xmath504 & @xmath505 & @xmath506 & @xmath507 & @xmath508 & @xmath509 + sk -69 280 & @xmath510 & @xmath511 & @xmath512 & @xmath513 & @xmath514 & @xmath515 + sk -70 116 & @xmath516 & @xmath517 & @xmath518 & @xmath519 & @xmath520 & @xmath521 + average & @xmath522 & @xmath523 & @xmath524 & @xmath525 & @xmath526 & @xmath527 + + sk -66 19 & @xmath528 & @xmath529 & @xmath530 & @xmath531 & @xmath532 & @xmath533 + sk -66 88 & @xmath534 & @xmath535 & @xmath536 & @xmath537 & @xmath538 & @xmath539 + sk -67 2 & @xmath540 & @xmath541 & @xmath542 & @xmath543 & @xmath544 & @xmath545 + sk -68 23 & @xmath546 & @xmath547 & @xmath548 & @xmath549 & @xmath550 & @xmath551 + sk -68 26 & @xmath552 & @xmath553 & @xmath554 & @xmath555 & @xmath556 & @xmath557 + sk -68 129 & @xmath558 & @xmath559 & @xmath560 & @xmath561 & @xmath562 & @xmath563 + sk -69 108 & @xmath564 & @xmath565 & @xmath566 & @xmath567 & @xmath568 & @xmath569 + sk -69 206 & @xmath570 & @xmath571 & @xmath572 & @xmath573 & @xmath574 & @xmath575 + sk -69 210 & @xmath576 & @xmath577 & @xmath578 & @xmath579 & @xmath580 & @xmath581 + sk -69 213 & @xmath582 & @xmath583 & @xmath584 & @xmath585 & @xmath586 & @xmath587 + average & @xmath588 & @xmath589 & @xmath590 & @xmath591 & @xmath592 & @xmath593 + ccccc 2.198 & 0.455 & @xmath594 & @xmath595 & @xmath596 + 1.650 & 0.606 & @xmath597 & @xmath598 & @xmath419 + 1.250 & 0.800 & @xmath599 & @xmath600 & @xmath601 + 0.810 & 1.235 & @xmath602 & & + 0.650 & 1.538 & @xmath603 & & + 0.550 & 1.818 & @xmath604 & @xmath605 & @xmath605 + 0.440 & 2.273 & @xmath606 & @xmath607 & @xmath608 + 0.370 & 2.703 & @xmath609 & @xmath610 & @xmath611 + 0.296 & 3.375 & @xmath612 & @xmath613 & @xmath614 + 0.276 & 3.625 & @xmath615 & @xmath616 & @xmath617 + 0.258 & 3.875 & @xmath618 & @xmath619 & @xmath620 + 0.242 & 4.125 & @xmath621 & @xmath622 & @xmath623 + 0.229 & 4.375 & @xmath624 & @xmath625 & @xmath626 + 0.216 & 4.625 & @xmath627 & @xmath628 & @xmath629 + 0.205 & 4.875 & @xmath630 & @xmath631 & @xmath632 + 0.195 & 5.125 & @xmath633 & @xmath634 & @xmath635 + 0.186 & 5.375 & @xmath636 & @xmath637 & @xmath638 + 0.178 & 5.625 & @xmath639 & @xmath640 & @xmath641 + 0.170 & 5.875 & @xmath642 & @xmath643 & @xmath644 + 0.163 & 6.125 & @xmath645 & @xmath646 & @xmath647 + 0.157 & 6.375 & @xmath648 & @xmath649 & @xmath650 + 0.151 & 6.625 & @xmath651 & @xmath652 & @xmath653 + 0.145 & 6.875 & @xmath654 & @xmath655 & @xmath656 + 0.140 & 7.125 & @xmath657 & @xmath658 & @xmath659 + 0.136 & 7.375 & @xmath660 & @xmath661 & @xmath662 + 0.131 & 7.625 & @xmath663 & @xmath664 & @xmath665 + 0.127 & 7.875 & @xmath666 & @xmath667 & @xmath668 + 0.123 & 8.125 & @xmath669 & @xmath670 & @xmath671 + 0.119 & 8.375 & @xmath672 & @xmath673 & @xmath674 + 0.116 & 8.625 & @xmath675 & & + | we present an exhaustive , quantitative comparison of all of the known extinction curves in the small and large magellanic clouds ( lmc and smc ) with our understanding of the general behavior of milky way extinction curves .
the @xmath0 dependent ccm relationship and the sample of extinction curves used to derive this relationship is used to describe the general behavior of milky way extinction curves .
the ultraviolet portion of the smc and lmc extinction curves are derived from archival iue data , except for one new smc extinction curve which was measured using hst / stis observations .
the optical extinction curves are derived from new ( for the smc ) and literature ubvri photometry ( for the lmc ) .
the near - infrared extinction curves are calculated mainly from 2mass photometry supplemented with denis and new jhk photometry .
for each extinction curve , we give @xmath1 and @xmath2 values which probe the same dust column as the extinction curve .
we compare the properties of the smc and lmc extinction curves with the ccm relationship three different ways : each curve by itself , the behavior of extinction at different wavelengths with @xmath0 , and behavior of the extinction curve fm fit parameters with @xmath0 . as has been found previously , we find that a small number of lmc extinction curves are consistent with the ccm relationship , but majority of the lmc and all of the smc curves do not follow the ccm relationship . for the first time , we find that the ccm relationship seems to form a bound on the properties of all of the lmc and smc extinction curves .
this result strengthens the picture of dust extinction curves exhibit a continuum of properties between those found in the milky way and the smc bar .
tentative evidence based on the behavior of the extinction curves with dust - to - gas ratio suggests that the continuum of dust extinction curves is possibly caused by the environmental stresses of nearby star formation activity . |
in two and three dimensions , spin wave excitations are well studied in pure isotropic and anisotropic insulating antiferromagnets @xcite .
magnetic excitations are significantly modified by magnetic dilution introduced by site substitution of the magnetic ions with diamagnetic ones . for isotropic systems near the magnetic percolation threshold concentration , @xmath5 , both well - resolved
, local spin excitations as well as crossover from spin wave excitations to fracton excitations on a fractal - like lattice @xcite , have been characterized near @xmath5 in two and three dimensions .
local spin excitations have been observed in dilute two - dimensional anisotropic systems near @xmath5 @xcite . in three dimensions ,
magnetic excitations have been studied for the magnetically dilute anisotropic systems @xmath6 @xcite and @xmath7 @xcite .
the parent compounds @xmath8 and @xmath9 exhibit comparable exchange energies and corresponding spin wave dispersions , but the @xmath9 system has an order of magnitude larger anisotropy and a correspondingly larger spin wave gap .
the excitations in the @xmath6 system were interpreted in terms of spin wave to fracton crossover as the scattering wavevector @xmath10 increases .
the behavior of the @xmath7 , for @xmath11 , has been interpreted @xcite as showing both spin wave and local spin excitations for small @xmath10 and local spin excitations for large @xmath10 . in this study
, we examine magnetic excitations with high resolution neutron scattering experiments and computer simulations in @xmath7 as @xmath12 approaches @xmath3 .
the well - characterized random - exchange antiferromagnet @xmath7 , with its simple structure and interactions , is an ideal anisotropic three - dimensional ( @xmath13 ) system in which to study magnetic excitations through inelastic neutron scattering measurements and theoretical modeling and simulation .
magnetic excitations in the @xmath13 anisotropic antiferromagnet @xmath9 have been very well characterized.@xcite the structure of @xmath9 and diamagnetic @xmath14 are similar.@xcite the antiferromagnetic spins in @xmath9 form a tetragonal lattice with two interpenetrating sublattices .
the dominant antiferromagnetic inter - sublattice @xmath15 exchange interaction is between the body - center and body - corner magnetic ions .
the intra - sublattice ferromagnetic @xmath16 and frustrating antiferromagnetic @xmath17 exchange interactions are much smaller ( cf .
iii for further details ) .
best fit values from inelastic neutron scattering measurements are shown in table i. @xmath9 and diamagnetic @xmath14 mix well during crystal growth to form @xmath7 , which is a dilute , anisotropic , three - dimensional antiferromagnet .
the occupation of sites by @xmath18 ions with @xmath19 or diamagnetic @xmath20 ions appears close to random , though slight clustering can not be ruled out .
it appears that @xmath15 does not vary significantly with dilution.@xcite there is limited information about the effect of dilution on the anisotropy , but it also does not appear to vary by a large amount.@xcite magnetic ordering in @xmath7 has been experimentally studied previously @xcite at magnetic concentrations @xmath12 equal to or near @xmath21 , the magnetic percolation threshold for the body - centered tetragonal magnetic structure with an interaction between the body - centered and corner ions ( @xmath15 in @xmath7 ) .
the @xmath22 random - exchange transition should be expected for @xmath23 if there is only the dominant exchange interaction @xmath15 .
however , the small @xmath16 and @xmath17 interactions in @xmath7 could become influential near @xmath5 .
the prior experiments in zero - field have demonstrated that for concentrations @xmath24 @xcite there is , at best , very weak long - range antiferromagnetic order at low temperatures .
the system exhibits spin - glass - like behavior , dominated by slow dynamics near the percolation threshold , possibly a result of the frustrating @xmath17 interaction .
early inelastic neutron scattering measurements in @xmath7 were compared to a simple treatment with the excitation energies assigned @xcite as @xmath25 , where @xmath26 represents the number of neighbors in @xmath9 , @xmath27 is the possible number of neighbors of a given spin in the magnetically dilute system , and @xmath28 is the spin wave energy as a function of the scattering wave vector @xmath10 in @xmath7 .
the intensities are assigned by the combinatorial probabilities of finding @xmath29 neighbors of a given spin .
while giving a fairly accurate description of the overall spread in energy , this description fails in the detailed structure of the excitation spectrum when higher energy resolution measurements resolve individual peaks .
similar results were found in far - infrared absorption experiments , high magnetic field pulsed laser absorption , and inelastic neutron scattering experiments for @xmath30.@xcite it was observed in the pulsed laser absorption measurements that the peaks become more easily resolvable for @xmath31 and that for this case the simplistic modeling described above proves wholly inadequate;@xcite the spacings of the resolved peaks do not correspond to the simple model .
the excitations were found to be largely localized , having little dispersion . here , we present a high resolution neutron - scattering study of @xmath32 close to its percolation threshold . in agreement with the aforementioned work ,
the excitations show little or no dispersion .
we show that a model of localized excitations accounts for our spectra even quantitatively .
the layout of the remaining part of this paper is as follows . in sec .
[ sect : experiments ] , we report our new experimental results .
the obtained spectra are rationalized through a simple model in sec .
[ sect : model ] . finally , we present our conclusions in sec . [ sect : discussion ] .
the results at @xmath31 motivated experiments closer to the percolation threshold and we have conducted inelastic neutron scattering studies for @xmath1 and @xmath4 .
neutron scattering measurements were carried out using the high energy - resolution triple - axis spectrometer c1 - 1 installed at the jrr-3 m reactor of jaea in tokai operating with a horizontally focusing analyzer with a final neutron energy of @xmath33 mev .
the energy resolution at the elastic position is 0.09 mev ( full width at half maximum ) but it increases to 0.46 mev as the energy transfer increases to 8 mev .
single - crystal samples were mounted in a closed - cycle refrigerator with the c - axis perpendicular to the scattering plane .
we examined a @xmath34 single crystal with a mass of @xmath35 g and a @xmath36 single crystal with a mass of @xmath37 g .
the magnetic concentrations of the optical - quality crystals were determined using density measurements .
the resulting scattering spectra of the experiments , shown in figs .
[ excite ] and [ excite_t ] , indicate localized excitations since the excitation energies are largely independent of the scattering wave vector @xmath10 . the simple approximation described above yields peaks similar to the resolved experimental peaks in the spectra , but the energies
are not well predicted , as discussed below .
it was clear that a more realistic calculation was needed to describe the peaks and to elucidate what governs the details of the @xmath38 vs @xmath10 spectra .
we discuss simulations below that capture essential characteristics of the experimental results .
experimental intensities vs energy for @xmath1 and @xmath2 for the zone center , @xmath39 , for @xmath40 , and for the zone boundary , @xmath41 .
the data for @xmath1 were taken at @xmath42k and those at @xmath4 were taken at @xmath43k .
the curves are guides to the eye constructed from gaussian peaks .
the gaussian peak locations are the same for each value of @xmath10 .
the ratio of intensities for @xmath1 and @xmath4 is arbitrary.,width=340 ] experimental intensities vs energy for @xmath1 at @xmath42 , @xmath44 , and @xmath45k to show the temperature dependence of the neutron scattering spectra.,width=340 ] we have remarked in the previous section that , to a large extent , the scattering spectra are independent of the scattering wave vector , which suggests that the underlying excitations are spatially localized and therefore can be described by a local model . local spin hamiltonians for pure @xmath46 have been extensively discussed,@xcite and include both an on - site interaction characterized by an anisotropy parameter @xmath47 , and heisenberg exchange interactions characterized by three coupling constants @xmath16 , @xmath15 and @xmath17 .
the strongest exchange is the antiferromagnetic @xmath15 that couples nearest neighbors belonging to different sublattices of the body - centered tetragonal magnetic lattice of @xmath48 atoms , while @xmath16 and @xmath17 couple atoms belonging to the same sublattice , i.e. , bonds parallel to the edges of the unit cell ( fig . [ unitcells ] ) . ( a ) unit cell with seven occupied sites of which the central site is active and the six dimmed sites are held frozen in an antiferromagnetic configuration .
( b ) the same unit cell with all the occupied sites being active ( for clarity , only two of them are marked with flipping arrows ) .
the corresponding frozen environment is the next shell of neighbors ( not shown in the figure).,width=340 ] following the same pattern , we model the disordered sample with the following hamiltonian : @xmath49 ^ 2 + j_2 \sum_{\langle i , j\rangle } \epsilon_i \epsilon_j { \boldsymbol s}_i\cdot{\boldsymbol s}_j \nonumber\\ & & { } + j_1\sum_{\langle\langle i , j\rangle\rangle } \epsilon_i \epsilon_j { \boldsymbol s}_i\cdot{\boldsymbol s}_j + j_3\sum_{\langle\langle\langle i , j\rangle\rangle\rangle}\epsilon_i \epsilon_j { \boldsymbol s}_i\cdot{\boldsymbol s}_j , \label{eq : hamil}\end{aligned}\ ] ] where the @xmath50 are @xmath19 spin operators that represent the @xmath48 atoms , and where the substitutional disorder is represented by statistically independent random variables @xmath51 which take the value one with probability @xmath12 ( the fraction of @xmath48 atoms ) and the value zero with probability @xmath52 ( the fraction of nonmagnetic @xmath53 atoms symbolized as empty sites in fig .
[ unitcells ] ) .
.best fit parameters , taken from ref . , for the spin hamiltonian of pure @xmath46.[tab : param ] [ cols="^,^,^,^",options="header " , ] table [ tab : param ] shows the parameters obtained by hutchings _ et al . _
@xcite by fitting inelastic neutron scattering data of @xmath46 to the spin hamiltonian , and it is often assumed that the same values can be used for the analysis of the dilute antiferromagnet @xcite ( we will elaborate on this point later ) . the fact that @xmath54 allows us to get an estimate of the spectrum at low temperatures by ignoring in the hamiltonian the terms proportional to @xmath16 and @xmath17 and making the approximation @xmath55 in the terms proportional to @xmath15 . for
the sake of definiteness let us consider a site @xmath56 on the a sublattice , with @xmath57 @xmath48 neighbors ( the probability distribution function for @xmath58 is binomial ) .
since at very low temperatures the magnetic state of the sample is essentially the nel state ( @xmath59 if the site @xmath56 belongs to the a sublattice and @xmath60 if @xmath56 is in the b sublattice ) , the local magnetic field felt by the spin @xmath56 due to the @xmath57 surrounding atoms is @xmath61 .
hence , the contribution of spin @xmath56 to the energy is @xmath62 ^ 2 - 2n_2j_2s_i^{(z)}$ ] .
an incoming neutron typically causes a spin flip @xmath63 ; the energy of such a transition is @xmath64 and therefore to a first approximation the spectrum consists of nine evenly spaced zero - width peaks with a binomial distribution of intensities .
the anisotropy parameter @xmath47 determines the average position of the peaks , while the spacing between peaks is proportional to @xmath15 .
actually , a more accurate description can be obtained by averaging the immediate generalization of eq .
@xmath65 over the respective number of neighboring sites @xmath66 , @xmath67 , @xmath68 . in fig .
[ formula ] we show the result of this calculation with the parameters of the pure sample @xcite and ( zero - width ) intensities proportional to the products of the respective combinatorial weights .
this figure shows how the coupling constants @xmath16 and @xmath17 contribute to the effective spread of the peaks .
in fact , comparison with the fit to the experimental data suggests that all the pure sample parameters need to be modified to describe the highly diluted sample and , in particular , that the anisotropy parameter @xmath47 is too large . as we will see in the forthcoming discussion , the parameters that provide the best fit to the experimental data depend on the approximation scheme used to study the model eq . .
however , the semiclassical picture given by eq . and illustrated in fig . [ formula ] remains qualitatively correct in the full simulation .
semiclassical approximation to the spectrum provided by eq . averaged over the respective number of neighboring sites with the pure sample parameters .
@xcite the solid line is a fit to the experimental data of fig .
[ excite ] corresponding to @xmath4 and @xmath41 , and suggests that the pure sample parameters need to be modified to describe the highly diluted sample.,width=340 ] our approach to simulate the experimental spectra is a two - step procedure . in the first step we use the full hamiltonian but maintain the approximation @xmath55 for the exchange terms , so that we can generate typical local environments by a classical monte carlo simulation . more concretely
, we generate equilibrium spin configurations at @xmath69 using a heat bath combined with a cluster method in lattices with @xmath70 sites with an @xmath48 density @xmath4 .
after we equilibrate ten such lattices ( samples ) we pick at random on each sample 1000 nonempty sites .
we call this site , together with its @xmath71 nearest neighbors , as illustrated in fig .
[ unitcells](b ) , a dynamic shell , and the next shell of atoms [ not shown in fig .
[ unitcells](b ) ] , the local environment .
figure [ unitcells](a ) illustrates a simpler version of this idea , in which there is only one spin in the dynamic shell ( the central spin marked by a flipping arrow ) and there are six frozen atoms in an antiferromagnetic state that constitute the local environment . in the second step of our procedure we use the full hamiltonian with full quantum spin operators for each atom on the dynamic shell , while the nondynamic spins that constitute the local environment are kept fixed and act in effect as boundary conditions for the dynamic shell .
since the third component of the total spin for each dynamic shell @xmath72 commutes with the hamiltonian , we find the ground state ( g.s . ) within the subspace corresponding to the nel state @xmath73 , which amounts to diagonalizing a square matrix with up to @xmath74 states , and the excited states by diagonalizing the hamiltonian restricted to the subspaces @xmath75 allowed by the selection rules .
the transition energies are simply the energy differences @xmath76 the exact calculation of the intensities in this experimental setting involves the matrix elements of a rather complicated interaction hamiltonian.@xcite we settle for an estimate of the relative intensities of these transitions and consider the simplest possible interaction operator ( in fact , one of the terms appearing in the full expression ) , which is proportional to @xmath77 , where @xmath56 denotes the central atom of the dynamic shell .
the contribution of each transition to the total intensity is proportional to @xmath78 we have found that the main effect of the simplified transition matrix element is to suppress the contributions of high - energy transitions .
note also that the combinatorial factor is already included in our sampling of the simulation results . finally , we add up the contributions of all the possible transitions in our 1000 samples ,
calculate the convolution of this result with the measured instrumental resolution function ( a gaussian with moderate energy - dependent width ) , and normalize the result to match the maximum count of the experimental curves ( it would seem better to match the integrated intensity in the experimental range , but as we will see , the resulting widths are too narrow ) . experimental data for @xmath41 , numerical fitting and simulated spectrum at @xmath69 with the parameters of the pure sample @xcite for : ( a ) a one - site calculation corresponding to the dynamic shell of fig .
[ unitcells](a ) ; ( b ) calculation corresponding to the dynamic shell of fig .
[ unitcells](b).,title="fig:",width=302 ] experimental data for @xmath41 , numerical fitting and simulated spectrum at @xmath69 with the parameters of the pure sample @xcite for : ( a ) a one - site calculation corresponding to the dynamic shell of fig . [ unitcells](a ) ; ( b ) calculation corresponding to the dynamic shell of fig .
[ unitcells](b).,title="fig:",width=302 ] experimental points , numerical fitting and simulated spectrum with fitted values of @xmath47 and @xmath15 corresponding to the dynamic shell of fig .
[ unitcells](b ) for : ( a ) @xmath69 ; ( b ) @xmath79 ( i.e. , purely antiferromagnetic environment).,title="fig:",width=302 ] experimental points , numerical fitting and simulated spectrum with fitted values of @xmath47 and @xmath15 corresponding to the dynamic shell of fig .
[ unitcells](b ) for : ( a ) @xmath69 ; ( b ) @xmath79 ( i.e. , purely antiferromagnetic environment).,title="fig:",width=302 ] figure [ stdparam ] shows the results of these procedures for the two dynamic shells illustrated in fig .
[ unitcells ] in the energy range between @xmath80 and @xmath81 .
the figures show the experimental points , a numerical fitting to these points , and our simulation results .
the one - site calculation reproduces quite well the main features of the experimental spectrum , including the average position of the peaks ( controlled by the anisotropy parameter @xmath47 ) , although the separation between peaks ( controlled by the coupling constant @xmath15 ) is too large and , as we anticipated , the widths are too narrow . although the simulation results corresponding to the dynamic shell of fig .
[ unitcells](b ) feature wider widths , the average position is clearly shifted to high energies , which suggests that , not only the value of @xmath15 , but also the value of the anisotropy parameter @xmath47 may be too large in this context . a possible explanation
might be related to the method by which the pure sample parameters @xcite are obtained , whereby a semiclassical approximation is used to determine the spectrum parametrically as a function of @xmath47 , @xmath16 , @xmath15 and @xmath17 , and later these parameters are fitted to match the experimental results .
in essence , this procedure involves a calculation to first order in @xmath82 , which should give better results for @xmath83 i.e. , for the one - site approximation .
this possibility has already been noticed .
for example , in ref . certain empirical relations between the parameters of the pure of the diluted sample are proposed . following these ideas , in fig .
[ bestparam](a ) we show the result of an optimization of the parameters @xmath15 and @xmath47 to match the experimental results , which yielded @xmath84 and @xmath85 .
unfortunately , the correlation between these parameters and the uncertainties prevents a more accurate determination of these values or the simultaneous optimization of the less significant @xmath16 and @xmath17 that we have kept fixed .
although the widths of the peaks are still too narrow , the intensities are quite well accounted for , and even a last peak at @xmath86 seems to be reproduced . finally , as an estimate of the thermal effects in our simulations , in fig .
[ bestparam](b ) we show a similar calculation at @xmath79 , i.e. , with a purely antiferromagnetic state ( no thermal disorder ) of the environment .
note the distinctly narrower widths and poorer intensity relations between the peaks , particularly at high energies .
there are a variety of possible reasons to explain the larger experimental widths : our scattering operator is oversimplified , as it does not take into account the relative orientation of the lattice and the wave vector of the incoming neutron , which we also assume perfectly well defined ( i.e. , we neglect the spread of the neutron beam ) ; the dynamic shells and their environments have been obtained from a classical ( rather than quantum ) monte carlo , which surely overestimates the spin ordering at low temperatures ; and the dynamic shells are limited to spins and their immediate neighbors .
in summary , we have presented high - resolution spectra from neutron scattering experiments , conducted over @xmath87 close to its percolation threshold .
we model these spectra in terms of a site diluted heisenberg model , containing both ferromagnetic and antiferromagnetic exchange interactions . in spite of its simplicity and with only a moderate adjustment of the parameters , the proposed model accounts quite well for the position and intensity relations of the peaks in the spectra .
this success is probably due to the validity of our main hypothesis , namely the local nature of the spin excitations in these systems which lie close to the percolation threshold for the @xmath48 lattice . whereas local spin excitations dominate the energy spectrum for @xmath12 near @xmath5 in @xmath0 , we can not rule out a very small contribution from fracton excitations in a similar energy range .
fracton excitations as well as local spin excitations coexist in isotropic systems and both may exist in the small anisotropy @xmath88 system as @xmath12 approaches @xmath5 . in that case , modeling the local spin excitations could aid in separating the two types of excitations , allowing the characterization of local spin excitations as well as the persistence of fracton excitations under conditions of weak anisotropy .
we acknowledge partial financial support from micinn , spain , ( grant nos .
fis2009 - 12648-c03 and fis2011 - 22566 ) , and from ucm - banco santander ( gr32/10-a/910383 , gr58/08 - 910556 ) .
thanks the _ del amo _ foundation and the hospitality of the physics department of u. california - santa cruz , where part of this work was performed .
we thank the members of the neutron scattering laboratory , institute for solid state physics , the university of tokyo for supporting our experiments . | the low temperature excitations in the anisotropic antiferromagnetic @xmath0 for @xmath1 and @xmath2 , at and just above the magnetic percolation threshold concentration @xmath3 , were measured using inelastic neutron scattering .
the excitations were simulated for @xmath4 using a localized , classical excitation model , which accounts well for the energies and relative intensities of the excitations observed in the scattering experiments . |
the deformation of the nucleon and/or @xmath7 is reflected in the non - zero values of the ratios @xmath8_{e2}}{[\bar{\gamma}]_{m1 } } \\ r_{sm}&= & \frac{[\bar{\gamma}]_{c2}}{[\bar{\gamma}]_{m1}},\end{aligned}\ ] ] where @xmath9_{\alpha}$ ] denotes the @xmath2 transition with multipolarity @xmath10 ( magnetic m1 ) , @xmath11 ( electric e2 ) , and @xmath12 ( coulomb c2 ) .
as explained , for example , in ref.@xcite , these two ratios at the resonance energy @xmath13 mev can be calculated from the imaginary ( @xmath14 ) parts of the multipole amplitudes of pion electroproduction reactions @xmath15}{im[m^{3/2}_{1^+ } ] } \\ r_{sm } & = & \frac{im[s^{3/2}_{1^+}]}{im[m^{3/2}_{1^+}]}\end{aligned}\ ] ] here the standard notations ( @xmath16 with @xmath17 @xmath18 , @xmath19 , @xmath20)@xcite of the multipole amplitudes are used in the right hand sides of the above equations .
the @xmath5 and @xmath6 generated from the considered three models are compared in fig.1 along with the empirical values obtained from performing the amplitude analyses of the data from mit - bates@xcite , mainz@xcite , and jefferson laboratory ( jlab)@xcite . the comparison shown in fig.1 was already given in the publication@xcite by the jlab clas collaboration .
we see from fig.1 that the @xmath6 ( right panel ) ratios predicted by the sl model ( solid curve ) and dmt model ( dashed curve ) are strikingly different in the @xmath0 ( gev / c)@xmath1 region
. however their corresponding predictions of the ratio @xmath5 ( left panel ) are very close in the same low @xmath21 region .
as we will see in figs.3 - 10 , all three models give almost the same @xmath22 amplitude at the resonance position @xmath23 mev .
thus their differences seen in fig.1 can be understood from fig.2 where the @xmath21-dependence of the predicted ratios @xmath24/im[e_{1^+}]$ ] are compared .
we see that the sl model prediction at @xmath25 is very close to the long wavelength limit @xmath26 ( @xmath27 ) which is consistent with the results given in the papers by amaldi , fubini , and furlan@xcite and also by capstick and karl@xcite . on the other hand , the @xmath24/im[e_{1^+}]$ ] of dmt(maid2003 ) model approaches to 2 ( 3 ) at @xmath25 .
thus their predictions of @xmath6 are very different , as seen in the right panel of fig.1 .
as explained in section iv of ref.@xcite , the long wave length limit is used in the sl model to define the bare @xmath12 transition strength as @xmath28g_e(0)$ ] . with @xmath29
determined from fitting the pion photoproduction data , @xmath30 is fixed in the sl model @xmath31 making use of the pion electroproduction data .
in fig.1 we see that the @xmath6 data point at @xmath32 ( gev / c)@xmath1 from mit - bates@xcite and mainz@xcite disagrees with the prediction from the sl model , while it agrees with the results generated from the dmt and maid models .
this difference between the sl model and the dmt model marks the large descrepancies between these two dynamical models in describing the pion cloud effects . as illustrated in fig.10 of ref.@xcite , the pion cloud effect can enhence the @xmath2 @xmath12 transition ( @xmath33 ) ) by a factor of about 2 at @xmath34 ( gev / c)@xmath1 and has a very prounced @xmath21-dependence .
thus the experimental verifications of the predictions in the entire low @xmath21 region given in fig.1 will lead to a detailed understanding of the pion cloud effects .
to give more detailed information for analyzing the new data , we compare in figs .
3 - 10 the multipole amplitudes generated from the sl model ( solid curves ) , dmt model ( dashed curves ) , and maid2003 ( dotted curves ) .
we see that they are in excellent agreement only in the imaginary part of the m1 multipole ( @xmath35 ) . for other resonant multipole amplitudes ( @xmath36 and @xmath37 ) ,
the sl model differs significantly from the other two models .
these differences can lead to rather different predictions on various interference cross sections @xmath38 ,
@xmath39 , and @xmath40
. for non - resonant multipole amplitudes in figs.3 - 10 , the differences between the considered three models are very large in some cases .
however it is not easy to identify experimental observables which are most effective in testing these weaker amplitudes . finally , we list in tables 1 - 4 the multipoles amplitudes generated from the sl model at @xmath41 0 .
, 0.05 , 0.1 , 0.20 , 0.5 .
the corresponding values from dmt and maid can be obtained from the web site listed in ref.@xcite .
more results from the sl model can be obtained from the authors .
this work was supported in part by the u.s .
department of energy , office of nuclear physics , under contract no .
w-31 - 109-eng-38 and japan society for the promotion of science , grant - in - aid for scientific research ( c ) 15540275 .
( left panel ) and @xmath6 ( right panel ) predicted by the sl model[1,2 ] ( solid curves ) , dmt model[3,4 ] ( dashed curves ) , and maid2003[5 ] ( dotted curves ) are compared with the empirical values from mit - bates[7 ] , mainz[8 ] , and jefferson laboratory[9,10].,title="fig:",width=264 ] ( left panel ) and @xmath6 ( right panel ) predicted by the sl model[1,2 ] ( solid curves ) , dmt model[3,4 ] ( dashed curves ) , and maid2003[5 ] ( dotted curves ) are compared with the empirical values from mit - bates[7 ] , mainz[8 ] , and jefferson laboratory[9,10].,title="fig:",width=264 ] | pion electroproduction amplitudes predicted by three theoretical models in the @xmath0 ( gev / c)@xmath1 region are compared .
the objective is to facilitate the analyses of the data from new experiments on investigating the pion cloud effects on the @xmath2 transition form factors . in recent years
, the data of pion electroproduction reactions have been analyzed by using various models for extracting the @xmath3 transition form factors .
the large pion cloud effects predicted by the dynamical models@xcite have motivated new measurements in the low mometum transfer region @xmath4 ( gev / c)@xmath1 . to facilitate the analyses of these new data , we here compare the predictions from three models : the sato - lee(sl )
model@xcite , the dubna - mainz - taiwan(dmt ) model@xcite , and the mainz unitary isobar model ( maid)@xcite . the details of each of the considered models can be found in their original papers and therefore will not repeated here .
we only point out some noticeable differences between them , which are relavent to the understanding of the pion cloud effects . |
let @xmath1 be an odd prime and @xmath6 an integer . consider the @xmath7-sphere @xmath8 @xmath9-times ) .
given integers @xmath10 relatively prime to @xmath1 , the map @xmath11 , where @xmath12 , defines a free action of @xmath13 on @xmath14 .
the orbit spaces of @xmath15 and the subgroup @xmath16 are the lens spaces @xmath17 and @xmath3 , respectively .
thus , we have a free action of @xmath18 on @xmath19 with the orbit space @xmath17 . by a mod @xmath1 cohomology lens space ,
we mean a space @xmath2 whose ech cohomology @xmath20 is isomorphic to that of a lens space @xmath21 .
we will write @xmath22 to indicate this fact .
if @xmath23 acts on a mod @xmath1 cohomology lens space @xmath2 , then the fixed point set of @xmath15 on @xmath2 has been investigated by su @xcite . in this paper
, we determine the cohomology ring ( mod @xmath1 ) of the orbit space @xmath4 , when @xmath15 acts freely on @xmath2 .
the following theorem is established .
_ let @xmath23 act freely on a finite - dimensional cw - complex @xmath24 .
if @xmath25 then @xmath26 is one of the following graded commutative algebras@xmath27 _ 1 .
@xmath28/i,$ ] where @xmath29 is the homogeneous ideal @xmath30 & \hskip -1.25pc \quad\,-c_{hh'}y^{(h+h')/2 } , u_hu_{h'}-b'_{hh'}zxu_{h+h'-2p-1}-c'_{hh'}zy^{(h+h'-2p)/2}\rangle \,,\end{aligned}\ ] ] @xmath31 @xmath32 @xmath33 @xmath34 @xmath35 @xmath36 and @xmath37 when @xmath38 .
@xmath39 is the mod-@xmath1 bockstein homomorphism associated with the sequence @xmath40 .
@xmath41/\langle x^2,z^m),$ ] where @xmath42 and @xmath43 .
in this section , we recall some known facts which will be used in the proof of our theorem . given a @xmath15-space @xmath2 , there is an associated fibration @xmath44 , and a map @xmath45 , where @xmath46 and @xmath47 is the universal @xmath15-bundle .
when @xmath15 acts freely on @xmath2 , @xmath48 is homotopy equivalance , so the cohomology rings @xmath49 and @xmath50 ( with coefficients in a field ) are isomorphic . to compute @xmath49
, we exploit the leray serre spectral sequence of the map @xmath51 .
the @xmath52-term of this spectral sequence is given by @xmath53 ( where @xmath54 is a locally constant sheaf with stalk @xmath55 and group @xmath56 and it converges to @xmath49 , as an algebra .
the cup product in @xmath57 is induced from that in @xmath58 via the isomorphism @xmath59 .
when @xmath60 operates trivially on @xmath61 , the system of local coefficients is simple ( constant ) so that @xmath62 in this case , the restriction of the product structure in the spectral sequence to the subalgebras @xmath63 and @xmath64 gives the cup products on @xmath65 and @xmath61 respectively .
the edge homomorphisms @xmath66 & h^q(x_g)\to e_\infty^{0,q}=e_{q+1}^{0,q}\subset \cdots\subset e_2^{0,q}=h^q(x)\end{aligned}\ ] ] are the homomorphisms @xmath67 respectively .
the above results about spectral sequences can be found in @xcite .
we also recall that @xmath68/ ( s^2 ) = \lambda ( s ) \otimes \zz_p[t],\ ] ] where @xmath69 , @xmath70 and @xmath71 .
the following fact will be used without mentioning it explicitly .
@xmath72 _ suppose that @xmath73 acts on a finite - dimensional cw - complex space @xmath2 with the fixed point set @xmath74 . if @xmath75 for @xmath76 then the inclusion map @xmath77 induces an isomorphism @xmath78 for @xmath79 @xmath80see theorem @xmath81 in chapter vii of @xcite ) .
the example of free action of @xmath23 on the lens space @xmath21 described in the introduction is a test case for the general theorem .
all cohomology groups in the proof should be considered to have coefficients in @xmath82 since @xmath83 acts trivially on @xmath61 , the fibration @xmath44 has a simple system of local coefficients on @xmath84 .
so the spectral sequence has @xmath85 let @xmath86 and @xmath87 be generators of the cohomology ring @xmath61 . as there are no fixed points of @xmath15 on @xmath2 , the spectral sequence does not collapse at the @xmath52-term .
consequently , we have either @xmath88 and @xmath89 or @xmath90 and @xmath91 . if @xmath92 and @xmath93 , then we have @xmath94 and @xmath95 for @xmath96 . so @xmath97= m t\otimes ab^{m-1}$ ] , which is true iff @xmath98 .
suppose that @xmath99 .
then @xmath100 is an isomorphism if @xmath101 is even and @xmath102 , and @xmath103 if @xmath101 is odd or @xmath104 .
so @xmath105 for all @xmath106 if @xmath107 or @xmath108 , where @xmath109 ; @xmath110 if @xmath101 is odd and @xmath111 , and @xmath112 , otherwise .
it is easily seen that @xmath113 , for example , if @xmath114 and @xmath115 ( a\in\zz_p)$ ] , then , for @xmath116\in e_3^{2,0}$ ] , we have @xmath117\rightarrow a = 0 $ ] .
a similar argument shows that the differentials @xmath118 are all trivial . if @xmath119 is also trivial , then @xmath120 is trivial for every @xmath106 and @xmath101 , because every element of @xmath121 can be written as the product of an element of @xmath122 by @xmath123 and @xmath124 is obviously trivial . if @xmath125 , then @xmath126 , where the top and bottom lines survive .
this contradicts our hypothesis ; so @xmath127 . if @xmath128 = [ st^p\otimes 1]$ ] , then it can be easily verified that @xmath129 = q[st^p\otimes b^{(q-1)p}]\quad \text{and}\\[.2pc ] & d_{2p+1}[1\otimes ab^{(q+1)p-1 } ] = q[st^p\otimes ab^{qp-1}]\end{aligned}\ ] ] for @xmath130 . consequently , @xmath131 = n ( st^p\otimes ab^{np-1}),\ ] ] which is not true for @xmath132 . on the other hand , if @xmath133 is trivial , then @xmath134 is also trivial for every @xmath106 and @xmath101 , as above .
now @xmath135 for every @xmath136 , so several lines of the spectral sequence survive to infinity .
this contradicts our hypothesis .
therefore , @xmath137 must be non trivial .
assume that @xmath138 = [ t^p\otimes 1]$ ] .
then @xmath139 is an isomorphism for @xmath107 , @xmath140 , and is trivial homomorphism for other values of @xmath101 .
accordingly , we have @xmath141 , and hence @xmath142 the elements @xmath143 and @xmath144 , for @xmath145 are permanent cocyles ; so they determine elements @xmath146 and @xmath147 , respectively . obviously , @xmath148 , @xmath149 and @xmath150 .
let @xmath151 and @xmath152 .
then @xmath153 , @xmath154 , and , by the naturality of the bockstein homomorphism @xmath155 , we have @xmath156 and @xmath157 but @xmath158 .
it follows that the total complex tot @xmath159 is the graded commutative algebra @xmath160/\langle x^2,y^p , z^n , w_hw_{h'},w_hy\rangle,\end{aligned}\ ] ] where @xmath161 , @xmath162 .
choose @xmath163 representing @xmath164 for @xmath165 .
then @xmath166 , @xmath167 and @xmath168 .
it follows that @xmath169/i,\ ] ] where @xmath29 is the ideal generated by the homogenous elements @xmath170 and @xmath171 .
here @xmath172 , @xmath173 , @xmath174 , @xmath175 and , when @xmath176 , @xmath177 .
if @xmath178 , then we must have @xmath179 , @xmath180 .
it can be easily observed that @xmath181 is a trivial homomorphism for @xmath101 even and an isomorphism for @xmath101 odd .
it is now easy to see that @xmath182 for every @xmath183 .
so @xmath184 for @xmath185 and @xmath186 .
therefore , we have @xmath187 if @xmath188 is determined by @xmath189 , then @xmath190 .
the multiplication by @xmath191 @xmath192 is an isomorphism for @xmath101 even .
the element @xmath193 is a permanent cocycle and determines an element @xmath194 .
we have @xmath195 and @xmath196 . therefore , the total complex tot @xmath159 is the graded commutative algebra @xmath197/\langle x^2,z^m\rangle\,.\ ] ] notice that @xmath198 is @xmath199 for @xmath200 even and @xmath201 for @xmath200 odd .
hence , @xmath202/\langle x^2 , z^m\rangle,\ ] ] where @xmath42 and @xmath43 .
this completes the proof .
we realize here the second case of our theorem . recall that @xmath23 acts freely on @xmath21 with the orbit space @xmath204 . we claim that @xmath205/\langle x^2,z^m\rangle,\ ] ] where @xmath206 .
it is known that @xmath207 is a cw - complex with 1-cell of each dimension @xmath208 and the cellular chain complex of @xmath207 is @xmath209 where each @xmath210 .
accordingly , the co - chain complex of @xmath207 with coefficients in @xmath18 is @xmath211 where each coboundary operator is the trivial homomorphism .
therefore @xmath212 to determine the cup product in @xmath213 , we first observe that the inclusion @xmath214 induces isomorphism @xmath215 for @xmath216 so that we can identify them . for @xmath217 ,
this follows from the cohomology exact sequence of the pair @xmath218 . the exact cohomology sequence of the pairs @xmath219 and
@xmath220 show that @xmath221 and @xmath222 ; the latter because the homomorphism @xmath223 is surjective .
now , we choose generators @xmath224 and @xmath225 .
then obviously @xmath226 and @xmath227 .
we can assume , by induction , that @xmath228 and @xmath229 generate @xmath230 and @xmath231 , respectively , for @xmath232 .
then , there is an element @xmath233 such that @xmath234 generates @xmath235 ( see corollary 3.39 of @xcite ) .
we must have @xmath236 , otherwise the order of @xmath237 would be less than @xmath1 .
thus @xmath238 generates @xmath235 , and this is true only if @xmath239 generates @xmath240 .
hence our claim .
it is clear from the proof of the theorem that if @xmath178 , then only the second possibility of the theorem holds .
furthermore , if @xmath24 and @xmath241 , then there exists a simply connected space @xmath242 with a free action of @xmath243 such that @xmath244 and @xmath245 ( theorems 3.11 and 2.6 of @xcite ) .
if @xmath23 acts freely on @xmath2 , then the liftings of transformations ( on @xmath2 ) induced by the elements of @xmath15 form a group @xmath246 of order @xmath247 which acts freely on @xmath242 and hence @xmath246 must be cyclic .
it is clear that @xmath246 contains the group @xmath248 of deck transformations of the covering @xmath249 , and @xmath250 .
so @xmath251 .
since @xmath244 , the mod @xmath1 cohomology algebra of @xmath252 is a truncation of @xmath253 .
thus , in this case also , only the second possibility of the theorem holds irrespective of the condition whether or not @xmath254 .
we recall that a paracompact hausdorff space @xmath2 is called finitistic if every open covering of @xmath2 has a finite dimensional open refinement ( see p. 133 of @xcite ) . our theorem and its proof go through for finitistic spaces .
we would like to thank the referee for his / her valuable suggestions which have improved exposition of the paper . | let @xmath0 , @xmath1 an odd prime , act freely on a finite - dimensional cw - complex @xmath2 with mod @xmath1 cohomology isomorphic to that of a lens space @xmath3 . in this paper ,
we determine the mod @xmath1 cohomology ring of the orbit space @xmath4 , when @xmath5 . # 1\{#1 } # 1#1 |
the borsuk - ulam is the following theorem in classical topology : if @xmath0 is continuous , then there is some point @xmath1 such that @xmath2 , where @xmath4 is the antipodal point of @xmath5 .
the typical layman s statement of the borsuk - ulam theorem is that there is always a pair of opposite points on the surface of the earth having the same temperature and barometric pressure .
since the theorem first appeared ( proved by borsuk ) in the 1930s , many equivalent formulations , applications , alternate proofs , generalizations , and related ideas have been developed , and active work continues today .
see the nice book @xcite by matouek for a survey .
@xmath6 this paper aims to derive a similar theorem in the context of digital images which are defined by a discrete set of pixels . to get a feel for our aims ,
consider the grayscale digital image pictured in figure [ haystackfig ] , which measures @xmath7 pixels .
the brightness of each pixel varies across the image , so we can consider the brightness as a function @xmath8 , where @xmath9 is the set of pixels in the image .
the rectangular boundary @xmath10 of the image is topologically analagous to @xmath11 , and our brightness function naturally restricts to a function @xmath12 .
then the classical borsuk - ulam theorem suggests that , subject to some continuity condition on @xmath13 , there should be a pair of opposite points on the boundary having the same brightness . for digital images it turns out that we can not in general expect opposite points with exactly the same brightness .
we will revisit the image in figure [ haystackfig ] in section [ exlsection ] after we derive our results .
our approach uses the topological theory of digital images which has been developing since the 1980s .
several different models of `` digital topology '' have been proposed by various authors .
we will use the graph - like theory of digital images and continuous functions based on adjacency relations . see boxer @xcite for a basic reference .
we will review the basic ingredients here .
a _ digital image _ is a pair @xmath14 , where @xmath15 is a set and @xmath16 is a symmetric antireflexive relation on @xmath15 , called the _
adjacency relation_. we sometimes refer elements of @xmath15 as _ pixels _ , and the relation @xmath16 tells us which pixels are adjacent to one another .
when @xmath17 and @xmath18 and @xmath19 are related by @xmath16 , we say @xmath18 and @xmath19 are _
@xmath16-adjacent _ , and we write @xmath20 . if the adjacency relation @xmath16 is understood , we simply write @xmath21 . if @xmath21 or @xmath22 , we write @xmath23 . typically it is assumed that @xmath24 for some @xmath25 , and the adjacency relation @xmath16 is a restriction of some globally defined adjacency relation on @xmath3 .
the most obvious applications occur in the case where @xmath15 is a finite subset of @xmath26 or @xmath27 .
the topological theory of digital images has , to a large part , been characterized by taking ideas from classical topology and `` discretizing '' them .
typically @xmath28 is replaced by @xmath29 , and so on .
topology in @xmath30 typically makes use of the standard topology on @xmath30 .
viewing @xmath29 as a digital image , it makes sense to use the following adjacency relation : @xmath31 are called _ @xmath32-adjacent _ if and only if @xmath33 .
this adjacency relation corresponds to connectivity in the standard topology of @xmath28 ( the notation `` @xmath32 '' is explained below ) .
unfortunately there is no canonical `` standard adjacency '' to use in @xmath3 which corresponds naturally to the standard topology of @xmath30 . in the case of @xmath26 , for example , at least two different adjacency relations seem reasonable : we can view @xmath26 as a rectangular lattice connected by the coordinate grid , so that each point is adjacent to 4 neighbors ; or we can additionally allow diagonal adjacencies so that each point is adjacent to 8 neighbors .
this is formalized in the following definition from @xcite ( though these adjacencies had been studied for many years earlier ) : let @xmath34 be positive integers with @xmath35 .
then define an adjacency relation @xmath36 on @xmath3 as follows : two points @xmath37 are @xmath36-adjacent if their coordinates differ by at most 1 in at most @xmath38 positions , and are equal in all other positions .
thus @xmath39 is the two - dimensional integer lattice described above in which each point has 4 neighbors , and @xmath40 is the variation in which each point has 8 neighbors .
the usual adjacency described above on @xmath29 is simply @xmath32 . in @xmath27
the three relations @xmath41 differ with respect to which kinds of diagonal adjacencies are allowed : @xmath32 allows no diagonal adjacencies , @xmath42 allows `` face diagonals '' across unit squares , while @xmath43 additionally allows `` solid diagonals '' . in this paper
we focus on these @xmath36 relations .
almost all digital images in this paper will be subsets of @xmath3 for some @xmath25 , with adjacency given by @xmath36 for some @xmath35 .
our results in section [ dim1section ] will hold for more general adjacency relations , and even digital images @xmath15 with no particular embedding in any @xmath3 . without restricting to subsets of @xmath3 ,
the theory of digitial images resembles abstract graph theory ( see @xcite , which never assumes that the images are embedded in @xmath3 ) .
continuity of functions is a basic ingredient in a digital topological theory .
the following definition is equivalent to one posed by rosenfeld in @xcite : let @xmath14 and @xmath44 be digital images , and let @xmath45 be a function .
we say @xmath13 is _ @xmath46-continuous _ ( or simply _ continuous _ , if the adjacencies are understood ) if @xmath47 implies @xmath48 .
we will also make use of the notion of connectedness .
for @xmath49 and @xmath50 , let @xmath51_\z$ ] denote the set @xmath52 .
this set is called the _ digital interval _ from @xmath53 to @xmath54 .
given two points @xmath55 , a _
@xmath16-path from @xmath18 to @xmath19 _ is a @xmath56-continuous function @xmath57_\z \to x$ ] with @xmath58 and @xmath59 .
when the adjacency relation is understood , a @xmath16-path is called simply a _
path_. a digital image @xmath14 is _ @xmath16-connected _ ( or simply _ connected _ ) when any two points of @xmath15 can be joined by a @xmath16-path .
a subset @xmath60 is a @xmath16-connected component of @xmath15 when it includes all points which can be connected by a @xmath16-path to its elements . in many papers
the concepts defined above are always prefixed by the word `` digital '' ( e.g. `` digital path '' , `` digitally continuous '' , etc ) .
this is usually unnecessary for us , but we will sometimes use the word for emphasis . the rest of this paper is organized as follows : in section [ lipsection ] we develop some necessary terminology and machinery for working with non - continuous functions .
this material is used in section [ dim1section ] to obtain a borsuk - ulam theorem for @xmath29-valued functions on digital images .
our results in section [ dim1section ] generalize results by boxer in ( * ? ? ?
* section 5 ) . in section [ exlsection ]
we demonstrate a specific application to the grainstack image . in section [ higherdimsection ]
we give some partial results and state some open questions for higher dimensional images .
we would like to thank lawrence boxer for several comments and helpful references .
the proof of our borsuk - ulam theorem will require us to consider some functions which are not continuous .
all functions we will consider do , however , have the following weaker property .
let @xmath14 and @xmath44 be digital images , and let @xmath45 be a function .
we say @xmath13 has _
[ @xmath46]-lipschitz constant _
@xmath62 if @xmath47 implies that there is a @xmath63-path of length @xmath64 from @xmath65 to @xmath66 .
the terminology is motivated by the classical definiton of the lipschitz constant , which would require that @xmath67 . in our case to avoid assuming a particular metric on @xmath15 we only refer to @xmath16-adjacent points in which case it makes sense to interpret @xmath68 , and the distance from @xmath65 to @xmath66 is interpreted in terms of the length of a @xmath63-path connecting them .
this notion corresponds to a lipschitz condition given by rosenfeld ( * ? ? ?
* section 5 ) , which assumes that @xmath15 and @xmath69 are metric spaces .
rosenfeld does not explore the idea very much , and we will require a bit more development . by definition
a function is digitally continuous if and only if it has lipschitz constant 1 .
thus , unlike in classical analysis , the existence of a lipschitz constant does not imply continuity .
in fact in typical cases , any function ( continuous or not ) has a lipschitz constant : let @xmath70 be any function ( not necessarily continuous ) such that @xmath71 is a finite set contained in a single connected component of @xmath69 . in this case
any two points of @xmath71 are joined by a path in @xmath69 .
since @xmath71 is finite , there must be some maximum length required to realize all of these paths .
let @xmath64 be this maximum , and then @xmath13 has lipshitz constant @xmath64 . for integer - valued functions @xmath72 , it is easy to see that @xmath13 has lipschitz constant @xmath64 if and only if @xmath73 implies @xmath74 .
lipschitz constants for such functions behave predictably with respect to addition and scalar multiplication : [ lipaddition ] let @xmath75 have lipschitz constants @xmath64 and @xmath76 respectively , and let @xmath77 . then : * @xmath78 has lipschitz constant @xmath79 . *
@xmath80 has lipschitz constant @xmath81 .
the proofs are routine .
since continuous functions have lipschitz constant 1 , we have : [ ctsaddition ] if @xmath82 are continuous , then @xmath83 and @xmath84 each have lipschitz constant 2
. lipschitz constants also behave predictably with respect to compositions . again
the proof is routine .
if @xmath85 has lipschitz constant @xmath76 and @xmath86 has lipschitz constant @xmath64 , then @xmath87 has lipschitz constant @xmath88 . since a continuous function
has lipschitz constant 1 , the above gives : [ lipcomposition ] if @xmath85 is continuous and @xmath86 has lipschitz constant @xmath64 , then @xmath87 has lipschitz constant @xmath64 .
@xmath46-lipschitz constants can be equivalently formulated in terms of continuity with respect to a different adjacency . for a digital image @xmath44 ,
let @xmath89 be the adjacency relation defined by : @xmath18 is @xmath89-adjacent to @xmath19 if and only if there is a @xmath63-path of length @xmath38 from @xmath18 to @xmath19 .
this exponentiation operation has the following simple properties : * @xmath90 * @xmath91 most useful for us will be the following relation between powers of @xmath63 and lipschitz constants .
again we omit the routine proof .
[ powerlip ] let @xmath14 and @xmath44 be digital images , and let @xmath45 be any function .
then @xmath13 has @xmath46-lipschitz constant @xmath64 if and only if @xmath13 is @xmath92-continuous .
in dimension 1 , the borsuk - ulam theorem says that any continuous map @xmath93 has a point @xmath5 with @xmath2 .
the classical proof takes the map @xmath94 , and applies the intermediate value theorem to show that @xmath95 must have the value 0 for some @xmath5 . our adaptation of
this proof will require a digital analog of the intermediate value theorem which was proved by rosenfeld in ( * ? ? ? * theorem 3.1 ) .
after completing our work we learned of the material by boxer in @xcite , which gives a special case of our result in this section .
boxer uses essentially the same ideas , but focuses only on continuous functions ( rather than functions with lipschitz constant @xmath96 ) , and focuses only on the case where the domain is a `` digital simple closed curve '' .
we are grateful to l. boxer for bringing this material to our attention . the difficulty in adapting the classical proof of the borsuk - ulam theorem to digital images is that , even when @xmath13 is continuous
, the function @xmath97 may not be continuous , and thus rosenfeld s intermediate value theorem does not apply . by corollary [ ctsaddition
] , however , @xmath97 will have lipschitz constant 2 .
thus we will need to generalize rosenfeld s theorem for functions with lipschitz constant greater than 1 .
the theorem below is a direct generalization of the digital intermediate value theorem : when @xmath13 is continuous we have @xmath98 and the conclusion reads @xmath99 .
[ ivt ] let @xmath15 be connected and @xmath100 have lipschitz constant @xmath101 , and let @xmath102 and @xmath77 with @xmath103 .
then there is a point @xmath104 with @xmath105 .
if @xmath106 or @xmath107 then the conclusion follows trivially .
thus we will assume that @xmath108 .
let @xmath109_\z \to x$ ] be a path from @xmath5 to @xmath110 . since @xmath111 and @xmath112 , some values of @xmath113 are greater than or equal to @xmath53 and some are less than @xmath53 .
thus is some @xmath114 such that @xmath115 .
let @xmath116 , and we have : @xmath117 by corollary [ lipcomposition ] , @xmath118_\z \to \z$ ] has lipschitz constant @xmath64 , which gives the last inequality above .
we have shown that @xmath119 as desired .
recall that the standard dimension 1 borsuk - ulam theorem concerns functions @xmath120 . for our digital analogue we will replace @xmath28 by @xmath121 , and
it seems most natural to use a simple cycle of points to replace @xmath11 .
let @xmath122 be a digital image given by @xmath123 with the adjacency relation @xmath124 given by @xmath125 if and only if @xmath126 .
for a point @xmath127 , define the `` antipodal point '' @xmath128 by @xmath129 , where @xmath130 .
the natural statement of the borsuk - ulam theorem in our context would then be the following : if @xmath131 is continuous , then there is a point @xmath132 with @xmath2 .
this is not true , as the following counterexample demonstrates : ( a similar example appears in ( * ? ? ?
* figure 7 ) ) let @xmath133 be the map given by : @xmath134 then @xmath13 is @xmath135-continuous , and @xmath136 have the following values : @xmath137 and we can see that @xmath138 for all @xmath139 .
the classical borsuk - ulam theorem fails to hold for the above example because the function @xmath97 fails to be continuous , and thus the intermediate value theorem does not hold .
we will need to weaken the conclusion in our borsuk - ulam theorem in light of the slightly weaker conclusion to our intermediate value theorem from theorem [ ivt ] .
it turns out , as we will see below , that the domain space need not be a cycle @xmath122 , but needs only be a digital image with some continuous `` antipodality '' function .
a function @xmath140 is an _ involution _ when it is continuous and @xmath141 for all @xmath142 .
if @xmath143 has no fixed points , it is a _
free involution_. we say that a digital image @xmath14 is _ @xmath143-involutive _ if it admits a @xmath16-continuous free involution @xmath140 . by analogy with @xmath144 ,
we will write @xmath145 when no confusion will arise , and simply call the image _ involutive_.
we are now ready to prove our digital borsuk - ulam for @xmath29-valued functions .
we follow exactly the classical proof .
( see ( * ? ? ?
* theorem 5.1 ) for a slightly weaker form of this result . ) [ budim1 ] let @xmath14 be a connected involutive digital image , and let @xmath72 have lipschitz constant @xmath146 .
then there is a point @xmath142 such that @xmath147 .
since @xmath148 has lipschitz constant @xmath64 , it is easy to see that the function @xmath149 also has lipschitz constant @xmath64
. let @xmath150 be defined by @xmath151 . by lemma [ lipaddition ] , @xmath95 has lipschitz constant @xmath152 . to obtain a contradiction ,
assume that @xmath153 for all @xmath5 , that is , @xmath154 for all @xmath5 .
since the function @xmath155 is an involution , we have @xmath156 for all @xmath5 . since @xmath95 never takes the value 0 , this means that @xmath157 is positive for some @xmath5 and negative for others .
thus there is some point @xmath158 such that @xmath159 and thus @xmath160 .
thus we have @xmath161 , and by theorem [ ivt ] there is some @xmath162 with @xmath163 .
this contradicts our assumption that @xmath154 for all @xmath5 .
equivalently we can state the theorem in terms of powers of the @xmath32 adjacency on @xmath29 . by theorem [ powerlip ]
we have : [ budim1cor ] let @xmath14 be a connected involutive digital image , and let @xmath164 be @xmath165-continuous for @xmath166 .
then there is a point @xmath142 such that @xmath167 . in the case where @xmath13 is continuous
we can replace @xmath64 above by 1 , and we obtain : let @xmath14 be a connected involutive digital image , and let @xmath72 be continuous .
then there is a point @xmath142 with @xmath168 .
we conclude the section with one more restatement of theorem [ budim1 ] .
since we make no assumptions on @xmath15 as embedded in @xmath3 with any particular adjacency , we can state the result in the context of pure graph theory . in this context
, the digital image @xmath14 is viewed as a simple graph ( undirected , no looped or multiple edges ) with vertex set @xmath15 and edges corresponding to @xmath16-adjacencies .
the existence of a continuous free involution , in graph theoretic terms , is equivalent to @xmath15 having an automorphism which is a free involution .
let @xmath15 be a simple graph with an automorphism @xmath143 which is a free involution , and let @xmath169 be any function on the vertex set , and assume that @xmath170 whenever @xmath5 and @xmath110 are connected by an edge .
then there is some point @xmath142 with @xmath171 .
a typical bitmapped digital image on a computer ( for example in jpg or png format ) is stored as a rectangular array of color values . for simplicity
we will consider grayscale images , which can be stored as a two - dimensional array of integers , where the integer represents the brightness of each pixel .
consider figure [ haystackfig ] , which is a low - resolution grayscale image of monet s _ grainstack ( sunset)_. at this resolution and color - depth , this image is represented as a @xmath7 array of integers in the range @xmath172_\z$ ] , where 0 indicates a pure black pixel , and 255 indicates pure white .
we adopt the standard computer - graphics convention that pixel @xmath173 is in the top - left corner of the displayed image . in our abstract model ,
the grainstack image is the 2-box @xmath174_\z\times [ 0,117]_\z$ ] , equipped with a `` brightness '' function @xmath175_\z$ ] .
let @xmath15 be the `` boundary '' of @xmath9 , the perimeter cycle of 536 points .
the natural free involution on @xmath15 is the function @xmath176 which exchanges `` opposite '' points . for an image of this size
, it is easy to check continuity of @xmath13 by comparing the brightness of adjacent pixels .
the function @xmath13 will be @xmath177-continuous if all @xmath36-adjacent pixels have brightnesses differing by at most 1 .
few typical grayscale images will be @xmath177 continuous , since this precludes any sharply - defined features in the image .
the @xmath177-lipschitz constant ( typically not equal to 1 ) is a measure of how sharply the brightness can change between any adjacent pixels .
this can easily be computed in our example image . in our specific example , the restriction of @xmath13 to @xmath15 has @xmath178-lipschitz constant 23 .
the greatest change in brightness between adjacent pixels on the boundary occurs on the right edge of the image where the haystack meets the sky : between pixels ( 149,29 ) and ( 149,30 ) the brightness changes by 23 .
theorem [ budim1 ] applies , and states that there is a point @xmath142 with @xmath179 .
that is , there is some pair of opposite points whose brightness differ by less than 46 .
this is also easy to verify for an image of this size , and in fact the antipodal points with the closest brightness values are @xmath180 and @xmath181 , which have brightness differing by 13 .
these are marked on figure [ haystackmarked ] .
@xmath182{hs150blurannotated}\ ] ] informally speaking , in the case of typical grayscale images , our result states that the the largest possible difference in brightness between antipodal boundary points is two times the largest change in brightness between adjacent boundary points .
a higher dimensional borsuk - ulam theorem would consider maps from certain involutive digital images into @xmath3 with @xmath183 .
two factors will complicate matters in this more general setting . for @xmath183
, we have several competing options for the adjacency to use on the codomain space @xmath3 .
it is also not obvious what the assumptions on the digital image @xmath15 should be , in order that it resemble the sphere @xmath144 .
several authors have considered `` sphere - like '' digital images .
we adapt a formulation from @xcite , where boxer defines an @xmath25-sphere as @xmath184 given by @xmath185^{n+1}_\z - \{\mathbf 0\}$ ] , where @xmath186 is the origin in @xmath187 .
we will consider slightly more general digital images : our results hold for any image which is the boundary of an @xmath25-box : a set @xmath188 is called a _
[ digital ] @xmath25-box _ if it has the form : @xmath189_\z \times \dots \times [ a_n , b_n]_\z,\ ] ] where @xmath190 with @xmath191 for each @xmath139 .
if @xmath192 for each @xmath139 , we say @xmath9 is a _ unit _
@xmath25-box . for an @xmath25-box @xmath9 ,
let @xmath10 denote the _ boundary _ of @xmath9 , defined by : @xmath193 the @xmath25-box @xmath9 carries a canonical involution @xmath194 given by : @xmath195 this is an `` antipodal '' reflection of @xmath9 , which restricts to a free involution on @xmath10 . as usual , we will denote @xmath145 ( note that this will not equal the usual additive inverse of @xmath196 ) . our main approach will take a digitally continuous function @xmath202 and find the desired point @xmath5 by replacing @xmath13 by a continuous map on @xmath203 and applying the classical borsuk - ulam theorem . for any subset @xmath204 ,
let @xmath205 be its convex hull . in particular if @xmath206 is an @xmath25-box , we have @xmath207 \times \dots \times [ a_n , b_n],\ ] ] and @xmath208 is homeomorphic to the closed @xmath25-ball in @xmath30 .
let @xmath211 be any triangulation of @xmath30 whose vertex set is @xmath3 .
any function @xmath212 induces a piecewise linear real - valued map @xmath213 by extending @xmath13 linearly along the simplices of @xmath211 .
this @xmath214 we call the _ linearization of @xmath13 on @xmath211_. such linearizations respect simplices in the following sense : let @xmath218 .
the set @xmath219 is @xmath214 applied to the convex hull of @xmath220 , and since @xmath214 is linear on the interior of @xmath221 , this is the convex hull of @xmath222 . since @xmath214 and @xmath13 agree on the integer lattice and @xmath223 for each @xmath139 , we have that @xmath219 is the convex hull of @xmath224 , which equals @xmath225 .
let @xmath211 be any antipodally symmetric triangulation of @xmath209 whose 1-skeleton is a subset of the edges given by @xmath227 .
( `` antipodally symmetric '' means that if @xmath216 is a simplex of @xmath211 then @xmath230 is a simplex of @xmath211 , where the negation denotes the antipodality involution of @xmath10 . )
this proof is made possible by the rigid structure of simplices in @xmath32 .
any @xmath32-simplex consists simply of a pair of points , and its convex hull is a line segment .
two @xmath32-simplexes whose convex hulls intersect must share a vertex . by lemma [ simplexlem ]
there is a @xmath227-simplex @xmath216 with @xmath229 , and since both @xmath240 and @xmath241 are @xmath32-simplices with intersecting convex hulls , @xmath242 contains a vertex @xmath243 .
the above proof makes use of very specific properties of @xmath32 adjacency , and seems difficult to generalize to other adjacencies in a straightforward way . in the case
@xmath248 the above corresponds to the @xmath98 case of corollary [ budim1cor ] . as we saw in section [ exlsection ] , typical digital images will not be @xmath32 continuous so it is important to obtain a result for @xmath249 adjacency in place of @xmath32 .
in fact we can generalize to higher values of @xmath64 .
we require a special construction for @xmath249-simplices in @xmath3 .
let @xmath216 be a @xmath249-simplex in @xmath3 , and let @xmath250 be the point whose coordinate in each position @xmath139 is the minimum of the @xmath139-th coordinates of the points of @xmath216 .
then let @xmath251 be the set : @xmath252 where @xmath253 is the sum of the coordinates of @xmath5 .
the points of @xmath254 form a @xmath249-simplex .
we have @xmath255 , and so @xmath256 . also , given any two @xmath249-simplices @xmath216 and @xmath257 , if @xmath258 then @xmath259 is nonempty and contains at least one point of @xmath3 .
again we begin with lemma [ simplexlem ] to obtain a @xmath227-simplex @xmath216 with @xmath229 .
since @xmath240 and @xmath241 are @xmath249-simplices , the set @xmath262 is nonempty and contains some integer point @xmath263 .
let @xmath264 .
then @xmath265 , and so @xmath266 since @xmath110 and @xmath148 are both in @xmath267 which is a @xmath249-simplex .
similarly we have @xmath247 and so @xmath268 . concatenating the @xmath32-paths from @xmath148 to @xmath110 and from @xmath110 to @xmath149
gives @xmath261 as desired .
since @xmath216 and @xmath257 are @xmath227-simplices , each is a subset of a unit @xmath25-box .
let @xmath9 and @xmath272 be unit @xmath25-boxes containing @xmath216 and @xmath257 , respectively .
then we have @xmath273 and @xmath274 , and so @xmath275 , and thus @xmath276 .
since @xmath9 and @xmath272 are unit boxes , their intersection @xmath277 is a `` face '' of a box , which by construction must contain a vertex from each of @xmath216 and @xmath257 .
thus there is some vertex @xmath244 with @xmath278 . since @xmath279 ,
every vertex of @xmath272 is @xmath227-adjacent to any vertex of @xmath257 .
thus @xmath5 is @xmath227-adjacent to every vertex of @xmath257 as desired .
again by lemma [ simplexlem ] there is a @xmath227-simplex @xmath216 with @xmath282 .
then since @xmath13 is @xmath280-continuous , @xmath240 and @xmath241 are @xmath269-simplices of @xmath283 , and so by lemma [ cnlemma ] there is a vertex @xmath284 which is @xmath269-adjacent to all vertices of @xmath241 .
let @xmath244 be some point with @xmath245 , and then we have @xmath281 . the proofs used for theorems [ bucnc1 ] and [ bucncn-1 ]
generalize directly to any adjacency relation satisfying the following condition : we say an adjacency relation @xmath63 on @xmath3 is _ regular _ when , if @xmath216 and @xmath257 are @xmath63-simplices and @xmath258 , then there is some vertex @xmath244 which is @xmath63-adjacent to all points of @xmath257 .
lemma [ cnlemma ] states that @xmath227 is regular as an adjacency on @xmath3 , and the arguments used in the proof of theorem [ bucnc1 ] show that @xmath32 is regular .
subject to the conjecture above we can expect a nice borsuk - ulam result for @xmath286-continuous functions on @xmath25-boxes with @xmath287 .
it is natural to ask if this type of result can hold for @xmath288-continuous functions with both @xmath289 and @xmath287 .
the adjacency conclusion from theorems [ bucnc1 ] and [ bucncn-1 ] will not hold in this more general case .
the following example shows a @xmath26-valued function on the boundary of a @xmath290-box which is @xmath291-continuous but has no point @xmath5 with @xmath239 .
@xmath292 \filldraw[fill = white , xshift=-2cm , yshift=-2 cm ] ( 45:1.2 ) \foreach \x in { 135,225,315,45 } { -- ( \x:1.2 ) } ; \filldraw[fill = white , xshift=-2cm , yshift=0 cm ] ( 45:1.2 ) \foreach \x in { 135,225,315,45 } { -- ( \x:1.2 ) } ; \filldraw[fill = white , xshift=-2cm , yshift=2 cm ] ( 45:1.2 ) \foreach \x in { 135,225,315,45 } { -- ( \x:1.2 ) } ; \filldraw[fill = white , xshift=0cm , yshift=-2 cm ] ( 45:1.2 ) \foreach \x in { 135,225,315,45 } { -- ( \x:1.2 ) } ; \filldraw[fill = white , xshift=0cm , yshift=0 cm ] ( 45:1.2 ) \foreach \x in { 135,225,315,45 } { -- ( \x:1.2 ) } ; \filldraw[fill = white , xshift=0cm , yshift=2 cm ] ( 45:1.2 ) \foreach \x in { 135,225,315,45 } { -- ( \x:1.2 ) } ; \filldraw[fill = white , xshift=2cm , yshift=-2 cm ] ( 45:1.2 ) \foreach \x in { 135,225,315,45 } { -- ( \x:1.2 ) } ; \filldraw[fill = white , xshift=2cm , yshift=0 cm ] ( 45:1.2 ) \foreach \x in { 135,225,315,45 } { -- ( \x:1.2 ) } ; \filldraw[fill = white , xshift=2cm , yshift=2 cm ] ( 45:1.2 ) \foreach \x in { 135,225,315,45 } { -- ( \x:1.2 ) } ; \node ( ) at ( -2cm,-2 cm ) { $ ( 0,1)$ } ; \node ( ) at ( -2cm,0 cm ) { $ ( 0,0)$ } ; \node ( ) at ( -2cm,2 cm ) { $ ( 0,0)$ } ; \node ( ) at ( 0cm,-2 cm ) { $ ( 0,0)$ } ; \node ( ) at ( 0cm,0 cm ) { $ ( 0,0)$ } ; \node ( ) at ( 0cm,2 cm ) { $ ( 0,0)$ } ; \node ( ) at ( 2cm,-2 cm ) { $ ( 0,0)$ } ; \node ( ) at ( 2cm,0 cm ) { $ ( 0,0)$ } ; \node ( ) at ( 2cm,2 cm ) { $ ( 0,0)$ } ; \end{tikzpicture } \qquad \begin{tikzpicture}[scale=.5 ] \filldraw[fill = white , xshift=-2cm , yshift=-2 cm ] ( 45:1.2 ) \foreach \x in { 135,225,315,45 } { -- ( \x:1.2 ) } ; \filldraw[fill = white , xshift=-2cm , yshift=2 cm ] ( 45:1.2 ) \foreach \x in { 135,225,315,45 } { -- ( \x:1.2 ) } ; \filldraw[fill = white , xshift=0cm , yshift=-2 cm ] ( 45:1.2 ) \foreach \x in { 135,225,315,45 } { -- ( \x:1.2 ) } ; \filldraw[fill = white , xshift=0cm , yshift=2 cm ] ( 45:1.2 ) \foreach \x in { 135,225,315,45 } { -- ( \x:1.2 ) } ; \filldraw[fill = white , xshift=2cm , yshift=-2 cm ] ( 45:1.2 ) \foreach \x in { 135,225,315,45 } { -- ( \x:1.2 ) } ; \filldraw[fill = white , xshift=2cm , yshift=2 cm ] ( 45:1.2 ) \foreach \x in { 135,225,315,45 } { -- ( \x:1.2 ) } ; \filldraw[fill = white , xshift=-2cm , yshift=0 cm ] ( 45:1.2 ) \foreach \x in { 135,225,315,45 } { -- ( \x:1.2 ) } ; \filldraw[fill = white , xshift=2cm , yshift=0 cm ] ( 45:1.2 ) \foreach \x in { 135,225,315,45 } { -- ( \x:1.2 ) } ; \node ( ) at ( -2cm,-2 cm ) { $ ( 1,1)$ } ; \node ( ) at ( -2cm,2 cm ) { $ ( 0,1)$ } ; \node ( ) at ( 0cm,-2 cm ) { $ ( 1,0)$ } ; \node ( ) at ( 0cm,2 cm ) { $ ( 0,1)$ } ; \node ( ) at ( 2cm,-2 cm ) { $ ( 1,0)$ } ; \node ( ) at ( 2cm,2 cm ) { $ ( 0,0)$ } ; \node ( ) at ( -2cm,0 cm ) { $ ( 0,1)$ } ; \node ( ) at ( 2cm,0 cm ) { $ ( 1,0)$ } ; \end{tikzpicture } \qquad \begin{tikzpicture}[scale=.5 ] \filldraw[fill = white , xshift=-2cm , yshift=-2 cm ] ( 45:1.2 ) \foreach \x in { 135,225,315,45 } { -- ( \x:1.2 ) } ; \filldraw[fill = white , xshift=-2cm , yshift=0 cm ] ( 45:1.2 ) \foreach \x in { 135,225,315,45 } { -- ( \x:1.2 ) } ; \filldraw[fill = white , xshift=-2cm , yshift=2 cm ] ( 45:1.2 ) \foreach \x in { 135,225,315,45 } { -- ( \x:1.2 ) } ; \filldraw[fill = white , xshift=0cm , yshift=-2 cm ] ( 45:1.2 ) \foreach \x in { 135,225,315,45 } { -- ( \x:1.2 ) } ; \filldraw[fill = white , xshift=0cm , yshift=0 cm ] ( 45:1.2 ) \foreach \x in { 135,225,315,45 } { -- ( \x:1.2 ) } ; \filldraw[fill = white , xshift=0cm , yshift=2 cm ] ( 45:1.2 ) \foreach \x in { 135,225,315,45 } { -- ( \x:1.2 ) } ; \filldraw[fill = white , xshift=2cm , yshift=-2 cm ] ( 45:1.2 ) \foreach \x in { 135,225,315,45 } { -- ( \x:1.2 ) } ; \filldraw[fill = white , xshift=2cm , yshift=0 cm ] ( 45:1.2 ) \foreach \x in { 135,225,315,45 } { -- ( \x:1.2 ) } ; \filldraw[fill = white , xshift=2cm , yshift=2 cm ] ( 45:1.2 ) \foreach \x in { 135,225,315,45 } { -- ( \x:1.2 ) } ; \node ( ) at ( -2cm,-2 cm ) { $ ( 1,1)$ } ; \node ( ) at ( -2cm,0 cm ) { $ ( 1,1)$ } ; \node ( ) at ( -2cm,2 cm ) { $ ( 1,1)$ } ; \node ( ) at ( 0cm,-2 cm ) { $ ( 1,1)$ } ; \node ( ) at ( 0cm,0 cm ) { $ ( 1,1)$ } ; \node ( ) at ( 0cm,2 cm ) { $ ( 1,1)$ } ; \node ( ) at ( 2cm,-2 cm ) { $ ( 1,1)$ } ; \node ( ) at ( 2cm,0 cm ) { $ ( 1,1)$ } ; \node ( ) at ( 2cm,2 cm ) { $ ( 1,0)$ } ; \end{tikzpicture}\ ] ] let @xmath9 be the @xmath290-box given by @xmath293_\z^3 $ ] .
then @xmath294 . let @xmath295 be given as follows : @xmath296 ; @xmath297 for all other @xmath298 ; @xmath299 ; @xmath300 ; @xmath301 if @xmath302 and @xmath303 if @xmath304 ; @xmath305 , and @xmath306 for all other @xmath298 .
see figure [ exlfig ] .
it is easy to check that @xmath13 is @xmath307-continuous , but there is no point @xmath5 with @xmath168 .
note that @xmath13 is not @xmath308- or @xmath309-continuous , since @xmath310 and @xmath311 are @xmath42- and @xmath43-adjacent , but @xmath312 and @xmath313 are not @xmath32-adjacent .
as we have seen in section [ exlsection ] , for many important examples the function @xmath13 will not be continuous , but rather have some lipschitz constant @xmath64 .
equivalently , @xmath13 will be @xmath314-continuous for some @xmath96 .
this case presents new obstacles , and there is considerable opportunity for further work . in particular we pose the following question : given an @xmath25-box @xmath9 and a function @xmath200 which is @xmath314-continuous for positive integers @xmath64 and @xmath287 and @xmath315 , what constant @xmath316 can be chosen which guarantees a point @xmath238 with @xmath317 ? we have answered the above question only in some restricted cases . by theorem [ budim1 ] , in the case @xmath248 , @xmath318 , and @xmath319 we can take @xmath320 . by theorem [ bucnc1 m ] , for any @xmath25 in the case @xmath321 , @xmath318 we can take @xmath322 .
we believe these values of @xmath272 ( especially the latter ) may be able to be improved ( reduced ) , which is also an interesting question . | the borsuk - ulam theorem states that a continuous function @xmath0 has a point @xmath1 with @xmath2 .
we give an analogue of this theorem for digital images , which are modeled as discrete spaces of adjacent pixels equipped with @xmath3-valued functions .
in particular , for a concrete two - dimensional rectangular digital image whose pixels all have an assigned `` brightness '' function , we prove that there must exist a pair of opposite boundary points whose brightnesses are approximately equal .
this theorem applies generally to any integer - valued function on an abstract simple graph .
we also discuss generalizations to digital images of dimension 3 and higher .
we give some partial results for higher dimensional images , and show a counter example which demonstrates that the full results obtained in lower dimensions can not hold generally . |
single - dot spectroscopy makes it possible to probe dot - to - dot changes in the excitonic properties of self - assembled ingaas / gaas quantum dots .
@xcite both single - particle and many - particle aspects of these properties depend non - trivially on the quantum dots size and shape,@xcite reflecting not only simple quantum - confinement physics , but also electronic structure effects such as interband , intervalley , spin - orbit and strain - induced state coupling .
the description of these effects require an atomistic multi - band approach .
@xcite here we adopt a method that is based on screened pseudopotentials and the configuration - interaction approach , and address the changes with height of recombination ( emission ) energies , polarization and radiative lifetimes of various neutral and charged excitons : the neutral exciton ( @xmath0 ) , negatively- ( @xmath1 ) and positively - charged ( @xmath2 ) excitons , and the biexciton ( @xmath3 ) in lens - shaped , self - assembled alloyed in@xmath4ga@xmath5as / gaas quantum dots .
our predictions compare reasonably well with available spectroscopic data .
we also compare our findings in in@xmath4ga@xmath5as / gaas dots with those in pure , non - alloyed inas / gaas dots .
we describe the basic electronic structure of the excitonic manifolds ( fig .
[ fig_1 ] ) before describing recombination processes .
all excitonic states are based on mixing and excitations of the single - particle states @xmath23 , @xmath24 , @xmath25 , @xmath26 and @xmath27 , @xmath28 , @xmath29 , @xmath26 , for holes and electrons , respectively .
these states are solutions to the single - particle schrdinger equation @xmath30 where both the external ( pseudo ) potential @xmath31 due to the ion - ion and ion - electron interaction and the screening response to such external potential @xmath32 are expressed as a superposition of screened atomic pseudopotentials @xmath33.\ ] ] here , @xmath34 is a non - local spin - orbit interaction;@xcite @xmath35 is a screened pseudopotential for atom of type @xmath36 that depends on strain , and it has been fitted to _ bulk _ properties of gaas and inas , including bulk band structures , experimental deformation potentials and effective masses , as well as local - density - approximation ( lda)-determined band offsets.@xcite the single - particle shrdinger equation [ eq . ( [ sp - effective_eq ] ) ]
includes not only quantum - confinement effects ( as in simple , one - band particle - in - a - box models ) , but also multi - band coupling ( light hole , heavy hole , conduction ) ; inter - valley ( @xmath37-@xmath38-@xmath39 ) coupling ; and spin - orbit coupling .
strain effects are present through the relaxation , via a valence force field,@xcite of the atomic positions @xmath40 within the simulation supercell ( quantum dot+gaas - matrix ) ; and directly `` felt '' by the potential @xmath41 $ ] .
states @xmath23 , @xmath24 , @xmath25,@xmath26 and @xmath27 , @xmath28 , @xmath29,@xmath26 form a basis for the excitonic states .
we indicate the dominant single - particle configuration of each excitonic level by @xmath42 , where @xmath43 , @xmath44 , @xmath45 , and @xmath46 ( = 0 , 1 , 2 ) indicate , correspondingly , the occupation of the electron levels @xmath47 and @xmath48 , and hole levels @xmath49 and @xmath50 .
figure [ fig_1 ] illustrates the excitonic manifolds using a @xmath51 - tall lens - shaped in@xmath52ga@xmath5as / gaas quantum dot ( base diameter @xmath53 ) , as obtained from configuration - interaction calculations based on the pseudopotential single - particle description .
_ monoexciton .
_ the monoexciton @xmath0 has a ground - state @xmath54 created by occupying @xmath55 and @xmath56 , denoted @xmath57 [ fig [ fig_1](a ) ] .
however , configuration - interaction@xcite ( ci ) also mixes into @xmath54 other states such as @xmath58 .
this is done by expanding the monoexciton states @xmath59 ( @xmath60 ) in a basis of slater determinants ( configurations ) @xmath61 constructed in the subspace of @xmath62 and @xmath63 electron and hole confined single - particle levels , respectively : @xmath64 are the ci coefficients and @xmath65 is a composite index that labels each slater determinant . since @xmath56 and @xmath55 are each two - fold degenerate due to spin , @xmath57 has four - fold degeneracy at the single - particle level of eqs .
( [ sp - effective_eq ] ) and ( [ pseudo ] ) .
we next allow coulomb electron - electron and hole - hole interactions .
direct coulomb is given by @xmath66 and exchange by @xmath67 , with the coulomb scattering matrix elements given by @xmath68^{*}\left[\psi^{(\mu^{\prime})}_j({\bf r}^{\prime})\right]^{*}\left[\psi^{(\mu^{\prime})}_k({\bf r}^{\prime})\right]\left[\psi^{(\mu)}_l({\bf r})\right]}{\epsilon({\bf r},{\bf r}^{\prime})|{\bf r}-{\bf r}^{\prime}|}.\ ] ] here @xmath69 ; and @xmath70 is a microscopic , phenomenological dielectric constant.@xcite we also allow electron - hole direct coulomb interaction @xmath71 and electron - hole exchange @xmath72^{*}\left[\psi^{(e)}_j({\bf r}^{\prime})\right]^{*}\left[\psi^{(e)}_k({\bf r})\right]\left[\psi^{(h)}_l({\bf r}^{\prime})\right]}{\epsilon({\bf r},{\bf r}^{\prime})|{\bf r}-{\bf r}^{\prime}|}.\ ] ] inclusion of these coulomb interactions splits @xmath57 into four distinct levels:@xcite the lowest two are spin - forbidden ( `` dark '' ) in the absence of spin - orbit coupling , and the highest two are allowed ( `` bright '' ) .
the bright - dark splitting of the ground - state levels shown in fig .
[ fig_1](a ) is @xmath73 .
the magnitude of this splitting increases up to @xmath74 for a @xmath75 - tall dot . at @xmath76
only the dark states are populated , thus , the transition to the ground state @xmath77 is long - lived .
figure [ fig_1](a ) ( shaded area ) shows transitions @xmath78 and @xmath79 from the bright states to @xmath77 .
the low - lying excited states of the monoexciton correspond to excitations of the hole , i.e. @xmath80 , @xmath81 and @xmath82 , due to the much smaller spacing of hole single - particle energy levels compared to that of the electrons .
thus , the spacing of the excited states fingerprint the hole energy level structure .
figure [ fig_1](a ) shows the first twelve excited states that are derived from @xmath80 , @xmath81 and @xmath82 .
each of these states are four - fold degenerate and their fine structure is shown schematically .
we find that three out of the four levels that arise from @xmath81 are optically allowed .
in particular , one of the levels , indicated as @xmath83 , emits light that is polarized along @xmath84 $ ] , while the remaining two emit in - plane polarized light .
these excited levels become optically forbidden as the dot becomes flatter .
_ biexciton .
_ in contrast to the four levels comprising the monoexciton , the biexciton @xmath3 has a singly - degenerate , _ closed - shell _ ground state ( @xmath85 ) that is bright .
thus , even at @xmath76 both emissions @xmath86 and @xmath87 [ fig . [ fig_1](b ) ; shaded area ] of the biexciton are fast ( @xmath19 ) .
the biexciton has a non - trivial ladder of excited states that is determined primarily by the relative magnitude of the direct coulomb hole - hole interaction and the single - particle energy splittings of the hole states .
the ladder begins with states derived from @xmath88 at about @xmath89 above the ground state and follows with @xmath90 and @xmath91 at about @xmath92 and @xmath93 , respectively . due to the two - fold kramers degeneracy of the hole levels
, @xmath88 is four - fold degenerate at the single - particle level , these states split into four distinct states due to hole - hole and electron - hole exchange [ fig [ fig_1](b ) ] .
similarly , the four states in @xmath90 split in two groups of two . in this case , remarkably , the splitting is about twice as big as that in @xmath88 and nearly five times bigger than in @xmath91 ( @xmath94 ) .
similarly to the monoexciton case , the lowest split - off pair in @xmath90 is optically active and emits light polarized along @xmath84 $ ] [ @xmath83 , fig . [ fig_1](b ) ] .
further , these states become darker as the dot becomes flatter , being forbidden at @xmath95 .
the subsequent excited state in the ladder derives from the singly - degenerate closed - shell state @xmath96 and is closely - spaced with @xmath91 .
note that albeit the splitting between these states is small , the state derived from the closed - shell configuration @xmath96 is relatively `` inert '' in that it mixes weakly with other states . in particular
the weight of this configuration in the ci expansion is @xmath97 .
below , we discuss the energy , polarization and lifetime of @xmath86 and @xmath87 .
_ trions . _
the negatively- ( @xmath6 ) and positively - charged ( @xmath7 ) trions have ground states that are bright , two - fold degenerate and arise from occupying @xmath98 and @xmath99 , respectively .
due to configuration - interaction mixing , the ground states of @xmath6 and @xmath7 mix with @xmath100 and @xmath101 , respectively . as in the case of @xmath3 , even at @xmath76 the recombination @xmath102 of @xmath6 [ fig .
[ fig_1](c ) ] into state @xmath103 and that of @xmath7 [ fig .
[ fig_1](d ) ] into state @xmath104 is fast .
for both trions , while there are only two dipole - allowed transitions in the absence of spin - orbit coupling , we predict four allowed transitions .
the first few excited states of @xmath6 correspond to occupying states derived primarily from @xmath100 , @xmath105 and @xmath106 [ fig .
[ fig_1](c ) ] .
( naturally , these states are mixed in with other states of @xmath6 . )
these excited states are two - fold degenerate and lie , correspondingly , about @xmath107 , @xmath108 and @xmath109 above the ground state , as shown in fig .
[ fig_1](c ) .
note the similarity of the low - lying excited states energy spacing with that in @xmath0 .
in addition , note that one of the two states derived from @xmath105 is optically active and emits light polarized along @xmath84 $ ] [ @xmath83 in fig .
[ fig_1](c ) ] .
again , these states become optically forbidden for a @xmath75 - tall dot .
the excited states of @xmath7 are more complex and present a high density of states [ fig .
[ fig_1](d ) ] ; a consequence of the closely spaced ( @xmath110 ) single - particle hole levels : in contrast to @xmath6 , all low - lying excited states show fine - structure splitting due to electron - hole exchange .
the first excited states at about @xmath111 corresponds to occupying @xmath101 , due to the two - fold degeneracy of each of these single - particle states , there are eight excited states that split in two groups of four , due to hole - hole exchange . occupying @xmath112 leads to the next eight excited states around @xmath113 , which are also split in two quartets . in this case
, there is an optically active ( polarization @xmath114 $ ] ) pair of excited states that belong to the lower - energy quartet [ @xmath83 , fig .
[ fig_1](d ) ] . as in the other excitons ,
these states become dark as the dot height decreases .
the next ten excited states about @xmath115 are a mixture of eight configurations derived from @xmath116 and two from @xmath117 . because @xmath56 is _ half - filled _ , @xmath117 is very `` reactive '' in the sense that it heavily mixes via configuration interaction .
note that this is significantly different from the case of configuration @xmath85 in the biexciton [ fig .
[ fig_1](b ) ] .
higher excited states correspond to occupying @xmath118 and @xmath119 .
below , we discuss the energy , as well as polarization and lifetime of @xmath102 for @xmath6 and @xmath7 .
we define the recombination energy upon recombination of an electron - hole pair in @xmath120 as the difference between the total energies of the initial state @xmath121 and the final state @xmath122 .
namely , @xmath123 where @xmath124 and @xmath125 label an initial and final state , respectively ; @xmath126 and @xmath127 are , respectively , the single - particle energy of the electron and hole in the final state ; and @xmath128 is the multi - particle , configuration - interaction energy of state @xmath129 . [ in eq .
( [ omega_if.eq ] ) , the energy of the ground state of the system in the absence of excited electron - hole pairs is taken to be zero . ]
optical experiments like photoluminescence probe electron - hole recombination transitions that are allowed ( bright).@xcite figure [ fig_1](a)-(d ) indicates with arrows the lowest bright recombination transitions of exciton @xmath120 , for a @xmath51 - tall lens - shaped in@xmath130ga@xmath131as / gaas dot with base diameter @xmath53 .
the final states upon electron - hole recombination are presented in the shaded area of fig .
[ fig_1 ] .
figure [ fig_2 ] shows the recombination energies @xmath132 , @xmath133 and @xmath134 calculated at the many - particle , configuration - interaction level as a function of dot height .
these recombination energies are shown relative to the lowest recombination energy @xmath135 of the monoexciton ( the latter energy is shown in the top axis of fig .
[ fig_2 ] ) .
thus , the results correspond to _ spectroscopic shifts _ that are currently measured by several groups .
prominent features : ( a ) recombination energies @xmath132 and @xmath133 , and @xmath134 ( see text and fig . [ fig_1 ] ) in in@xmath130ga@xmath131as / gaas quantum dots ( base @xmath53 ) with different heights .
the energy is shown as a spectroscopic shift relative to @xmath135 [ see fig .
[ fig_1](a ) ] .
( b ) screened direct coulomb interaction @xmath136 versus dot height . at @xmath137 a nearly `` symmetric '' regime [ @xmath138 is attained.,width=321 ] \(i ) the recombination energy @xmath139 blue - shifts as height increases ; in contrast , @xmath140 red - shifts .
these trends have been explained by bester and zunger in ref . by adopting the hartree - fock approximation .
\(ii ) for the flattest dot ( @xmath141 ) the ordering @xmath142 of the emission energies and the relative magnitude of the spectroscopic shifts @xmath143 agree with photoluminescence ( pl ) data .
@xcite these relationships are explained at the hartree - fock level of approximation , as in ref . , in which one derives @xmath144-j^{\,(eh)}_{00},\\ \omega_{00}(x^+)&=&[{\cal e}^{\,(e)}_0-{\cal e}^{\,(h)}_0]+j^{\,(hh)}_{00}-2\,j^{\,(eh)}_{00},\\ \omega_{00}(x^-)&=&[{\cal e}^{\,(e)}_0-{\cal e}^{\,(h)}_0]+j^{\,(ee)}_{00}-2\,j^{\,(eh)}_{00}.\end{aligned}\ ] ] then , the relationship @xmath145 that holds at @xmath95 [ fig . [ fig_2](b ) ] reveals the ordering of the emission energies and the magnitude of the spectroscopic shifts . regarding the magnitude of the spectroscopic shifts , @xmath146 and
@xmath147 agree reasonably well with ( a ) @xmath148 and @xmath149 , respectively , that ware and coworkers have recently observed in an inas / gaas dot ( size unspecified ) with monoexciton emission at @xmath150 ( ref . ) ; also with ( b ) @xmath151 measured by smith _
_ in an ( in , ga)as / gaas dots ( size unspecified ) with emission energy at @xmath152 ( ref . )
; and with ( c ) the value of @xmath153 for @xmath154 observed by finley and coworkers in an ( in , ga)as / gaas dots [ @xmath155 , @xmath156 ] with emission at @xmath157 ( ref . ) .
\(iii ) as height increases , @xmath134 red - shifts at small heights , reaches a maximum shift of nearly @xmath158 at @xmath159 , and then it moderately blue - shifts for taller dots .
in addition , at @xmath160 the emission energy of @xmath3 coincides with that of @xmath2 . as in ( ii ) , these results are explained at the hartree - fock level , which predicts @xmath161 here , we have neglected the small ( @xmath162-@xmath163 ) splitting of the monoexciton bright states .
then , by analyzing the height dependence of the direct coulomb interactions @xmath164 , @xmath165 and @xmath166 [ fig .
[ fig_2](b ) ] we find that eq .
( [ wxx0 ] ) predicts the observed height dependence of @xmath134 , although the actual magnitude of the emission is not quantitatively predicted due to correlations.@xcite in addition , eq .
( [ wx+ ] ) reveals the coincidence of the emission of @xmath3 and that of @xmath7 as a result of balancing the magnitudes of @xmath164 and @xmath166 at @xmath137 .
the latter balance arises from the similar degree of localization of @xmath167 and @xmath168 within the dot.@xcite \(iv ) @xmath169 at @xmath160 agrees with the value of @xmath170 for @xmath3 measured by finley and coworkers in pl experiments in an ( in , ga)as / gaas dot [ @xmath155 and @xmath156 ] with exciton ground - state emission at @xmath171 ( ref . ) .
in addition , this result of @xmath172 for the biexciton shift agrees remarkably well with the value of @xmath158 measured by rodt _
et al _ in inas / gaas dots ( @xmath173 , height unspecified ) with monoexciton emission energies ranging between @xmath174 and @xmath175 ( ref . ) ; and also with the measured shift of @xmath176 observed by urbaszek and coworkers in ( in , ga)as / gaas dots ( size unspecified ) with monoexciton emission energy of @xmath177 ( ref . ) .
the value of about @xmath178 for @xmath179 that we find at @xmath180 and @xmath181 agrees well with the value of @xmath182 observed by bayer and coworkers in ( in , ga)as / gaas dots [ @xmath183 ] with exciton emission at @xmath184 ( ref . ) .
our results also agree satisfactorily with the value of @xmath185 measured by findeis and coworkers in dots ( size unspecified ) with monoexciton emission energy at @xmath186 ( ref . ) .
the binding energies of excitons @xmath120 are defined as @xmath187-e^{(0)}(x^0 ) \\ \delta(x^-)&=&\big[{\cal e}^{\,(e)}_0+e^{(0)}(x^0)\big]-e^{(0)}(x^- ) \\ \delta(x^+)&=&\big[-{\cal e}^{\,(h)}_0+e^{(0)}(x^0)\big]-e^{(0)}(x^+ ) \\
\delta(xx^0)&=&2e^{(0)}(x^0)-e^{(0)}(xx^0 ) .
\end{array } } \ ] ] exciton @xmath120 is said to be _ bound _ when the binding energy @xmath188 is positive .
conversely , @xmath189 implies the exciton is unbound .
note that the binding energy is defined with respect to the ground - state energy of _ dissociated _ excitonic complexes .
for instance , the binding energy of @xmath0 is defined with respect to the energy of a non - interacting electron - hole pair . in turn , the binding energy of the biexciton @xmath3 is defined with respect to the total energy of two non - interacting monoexcitons [ eq .
( [ binding.eqs_ci ] ) ] . binding energy @xmath188 [ eq .
( [ binding.eqs_ci ] ) ] of @xmath0 , @xmath6 and @xmath7 , and @xmath3 as a function of dot height .
the ground - state energy @xmath190 of @xmath0 is also shown ( top axis ) . for @xmath6 , @xmath7 , and @xmath3 ,
height drives a bound - to - unbound crossover.,width=321 ] figure [ fig_3.mp ] shows @xmath191 , @xmath192 , @xmath193 and @xmath194 as well as @xmath190 as a function of dot height .
@xmath191 decreases with increasing height , and it is well approximated by @xmath195 as correlation effects are relatively small ( @xmath196 ) .
the height dependence of the binding energy of @xmath6 , @xmath7 , and @xmath3 follows the height dependence of the spectroscopic shifts shown in fig . [
fig_2](a ) .
this is so because the bright - dark splitting for the monoexciton @xmath0 is small @xmath197-@xmath198 . as expected from the spectroscopic shifts results [ fig .
[ fig_2](a ) ] , the height dependence of the binding energy is qualitatively different for each exciton @xmath120 .
( i ) @xmath192 is bound ( @xmath199 ) for the flattest dot ( @xmath141 ) and it decreases almost linearly , becoming unbound for dots taller than @xmath200 .
( ii ) in contrast , @xmath201 is unbound for the flattest dot and increases up to @xmath202 , becoming bound slightly below @xmath203 .
( iii ) @xmath194 does not depend monotonically on the gap , reaching a maximum around @xmath160 .
in addition , @xmath3 is bound [ @xmath204 for all heights above @xmath205 while unbound for the flattest dot ( @xmath141 ) .
the latter is due to an interplay between hartree - fock and correlation contributions to binding , @xmath206 and @xmath207 , respectively , that results in correlation being insufficient to bind @xmath3 .
namely , @xmath208\}+\delta(xx^0 ) \nonumber \\ & = & -1.6\;{\rm mev}+1.4\;{\rm mev } \nonumber \\ & = & -0.2\;{\rm mev } , \nonumber\end{aligned}\ ] ] it should be noted that rodt and coworkers@xcite observed ( in photoluminescence ) a bound - unbound crossover for @xmath3 as the monoexciton emission energy of ingaas / gaas quantum dots decreased .
those authors also calculated the binding energy of @xmath3 , using an 8-band @xmath209 model , and suggested that the reduction of correlation effects was responsible for unbinding @xmath3 as the gap increased.@xcite
when an electron - hole pair in exciton @xmath120 recombines optically , the transition is characterized by both the transition energy @xmath210 [ eq .
( [ omega_if.eq ] ) ] and the transition dipole matrix element @xmath211 here , @xmath212 is the electron momentum and @xmath213 is the polarization vector of the electromagnetic field.@xcite the dipole matrix elements @xmath214 [ eq . ( [ dipole.element ] ) ] depend on the polarization vector @xmath213 ; so , it is natural to quantify what is the degree of polarization anisotropy between different polarizations @xmath215 and @xmath216 .
therefore , we introduce the recombination ( emission or pholuminescence ) intensity spectrum of exciton @xmath120 for polarization @xmath213 , @xmath217.\ ] ] here , @xmath218/k_b\,t\}\ ] ] is the occupation ( boltzmann ) probability of the initial state @xmath121 at temperature @xmath219 ; @xmath220 is a normalization constant such that @xmath221 and @xmath222 is the boltzmann constant .
then , as _ in - plane _
polarizations @xmath223 $ ] and @xmath224 $ ] have been probed extensively , we introduce the in - plane polarization anisotropy parameter @xmath225 , @xmath226)}(\omega , t;\chi^q)-i^{([1\bar 10])}(\omega , t;\chi^q)}{i^{([110])}(\omega , t;\chi^q)+i^{([1\bar 10])}(\omega , t;\chi^q)}.\ ] ] thus , @xmath227 indicates an optical transition that is _ fully _ polarized along the @xmath8 $ ] direction , while @xmath228 indicates one _ fully _ polarized along @xmath229 $ ] .
in - plane polarization anistropy @xmath225 [ eq .
( [ lambda_eq ] ) ] for the lowest optical transitions of ( a ) @xmath0 , ( b ) @xmath3 , ( c ) @xmath6 , and ( d ) @xmath7 .
( see fig .
[ fig_1 ] . )
the transitions of @xmath0 and @xmath3 are polarized along @xmath8 $ ] and @xmath229 $ ] , and the degree of this polarization strongly depends on height . for @xmath6 and @xmath7 ,
the transitions are not polarized neither along @xmath8 $ ] nor @xmath229$].,width=321 ] figure [ fig_4 ] shows the in - plane polarization anisotropy @xmath225 for the lowest optical transitions of ( a ) @xmath0 , ( b ) @xmath3 , ( c ) @xmath6 , and ( d ) @xmath7 as a function of dot height .
two features are prominent : \(i ) the bright transitions @xmath135 and @xmath230 are polarized and the polarization anisotropy depends on height ; see fig . [
fig_4](a ) .
similarly , the lowest transitions of the biexciton @xmath134 and @xmath231 are polarized and the degree of polarization also depends on height .
these transitions correspond to the _ decay _ into the two bright states of the monoexciton [ fig . [ fig_1](d ) ] .
clearly , the biexciton transitions inherit the polarization of the monoexciton transitions . for both @xmath0 and @xmath3
, we see that @xmath225 switches sign as a function of height . in particular , at @xmath137 the transitions of @xmath0 present no in - plane polarization anisotropy
[ see arrow in figs . [ fig_4](a ) ] because the bright doublet formed by @xmath232 and @xmath233 is degenerate , which according to our calculation results in @xmath234)}(\omega_{20})\simeq i^{([1\bar 10])}(\omega_{20})$ ] .
\(ii ) the lowest optical transitions of @xmath6 and @xmath7 have degenerate transition ( recombination ) energies [ fig . [ fig_1](b ) and ( c ) ] and , according to our calculations , this results in @xmath234)}(\omega_{00})\simeq i^{([1\bar 10])}(\omega_{00})$ ] and , thus , negligible in - plane polarization anisotropy regardless of height .
in addition to ( i ) and ( ii ) , we find that ( iii ) the in - plane polarization anisotropy of the lowest transitions of @xmath0 and @xmath3 depends dramatically on the dot s alloy randomness ( disorder realization ) ; as shown in figure [ fig_5.new ] for seven alloy realizations in a @xmath235 - tall dot .
for a given transition in both @xmath0 and @xmath3 , @xmath225 changes sign depending on alloy randomness .
further , while some alloy realizations like @xmath236 and @xmath237 result in transitions nearly fully polarized ( @xmath238 ) , others such as @xmath14 , @xmath239 and @xmath240 present small anisotropy ( @xmath241 ) .
in - plane polarization @xmath225 for the lowest optical transitions of ( a ) @xmath0 and ( b ) @xmath3 for different realizations of alloy randomness in a @xmath235 - tall in@xmath130ga@xmath242as / gaas dot ( base @xmath53 ) . for both @xmath0 and @xmath3
, @xmath225 depends dramatically on alloy realization , changing not only its sign but its magnitude too.,width=321 ]
the _ characteristic _ radiative lifetime @xmath243 of a transition @xmath244 follows from both the magnitude of the dipole matrix element of the transition @xmath245 and the recombination energy @xmath246.@xcite namely , @xmath247 here , @xmath248 and @xmath249 are the charge and mass of the electron , respectively , and @xmath250 is the velocity of light in vacuum . in addition , the refractive index @xmath251 of the dot material accounts for the material s effects on the photon emission .
the linear dependence of @xmath252 on refractive index is applicable only when considering dot and matrix materials with similar dielectric constants , as it is the case in ingaas / gaas dots . in a more general case , more complicate dependences have been proposed.@xcite note that the characteristic radiative lifetime [ eq .
( [ intrinsic.lifetime ] ) ] does not depend on temperature nor on the occupation probability of the initial state @xmath121 , as it is a characteristic property of the transition @xmath244 . on the other hand ,
the actual radiative lifetime @xmath253 of exciton @xmath120 depends both on the probability @xmath254 of having the initial states @xmath121 of exciton @xmath120 occupied and the number of final states @xmath122 of exciton @xmath255 available for recombination , as well as on the characteristic radiative lifetimes @xmath243 .
we calculate @xmath253 from @xmath256 here , @xmath257 and @xmath258 where @xmath219 is the temperature of the system . in general
, the calculation of @xmath254 involves solving a system of rate equations.@xcite however , in the case that the intra - level relaxation of exciton @xmath120 is much faster than the radiative lifetimes , @xmath254 is given by the boltzmann weight @xmath259 [ eq .
( [ bolzmann.n ] ) ] .
_ monoexciton .
_ at @xmath20 , and assuming fast non - radiative relaxation to the dark ground - state of @xmath0 , the radiative lifetime of the monoexciton @xmath11 equals the characteristic radiative lifetime @xmath260 and it is long ( @xmath12 ) . at finite temperatures , all four levels [ fig . [ fig_1](a ) ] of the @xmath0 ground state are thermally populated and emit light with their own characteristic lifetime [ eq .
( [ intrinsic.lifetime ] ) ] . for temperatures such that the lowest four monoexciton states are occupied ,
while the occupation of excited states is negligible , we calculate the radiative lifetime from eq .
( [ lifetime ] ) and find @xmath261\simeq 2\,\tau_{20}(x^0).\ ] ] here , we have used a result of our calculations that predict @xmath262 regardless of dot height . in addition , we have neglected the long - lived recombinations @xmath102 and @xmath263 .
figure [ fig_5](a ) shows the characteristic @xmath264 versus dot height .
we find that this lifetime depends weakly on height .
a calculation of @xmath11 at @xmath265 predicts a moderate decrease in the monoexciton radiative lifetime as height increases [ fig . [ fig_5](b ) ] . it should be noted that @xmath11 is actually bigger than the approximate value of @xmath266 [ eq .
( [ tau.ht ] ) ] due to the actual occupation probability @xmath267 of the initial states . by measuring time - resolved photoluminescence in inas / gaas dots (
@xmath268-@xmath269 , height unspecified ) with monoexciton emission ranging from @xmath270-@xmath271 , karachinsky and coworkers have recently observed the opposite trend,@xcite i. e. @xmath11 increases with dot size from @xmath272 in a dot with emission energy of @xmath273 to @xmath274 in one with emission energy of @xmath275 .
the values of @xmath11 we predict for tall dots ( @xmath200 and @xmath276 ) agree satisfactorily with the value of @xmath272 measured by buckle _
et al _ for an inas / gaas dot ( @xmath277 , @xmath278 ) with gap @xmath279 ( ref . )
at a temperature of @xmath280 .
further , our predictions also agree well with the value of @xmath281 extracted from time - resolved photoluminescence experiments at @xmath282 performed in inas / gaas dots ( @xmath283 , @xmath141 ) by bardot and coworkers.@xcite ( color online ) ( a ) characteristic radiative lifetime @xmath243 [ eq .
( [ intrinsic.lifetime ] ) ] of the lowest optical transition of @xmath0 , @xmath6 , @xmath7 , and @xmath3 as a function of dot height . for @xmath0 and @xmath3 ,
the characteristic lifetimes are nearly the same and depend weakly on height .
a stronger dependence on height is found for @xmath6 and @xmath7 .
( b ) @xmath265 radiative lifetime @xmath253 [ eq .
( [ lifetime ] ) ] of the lowest optical transition of @xmath0 , @xmath6 , @xmath7 , and @xmath3 as a function of dot height .
the lifetimes depend weakly on height for all these excitons . for @xmath0 and @xmath3 ,
the lifetimes are nearly the same .
a similar trend is found for @xmath6 and @xmath7 . at all heights ,
the radiative lifetimes of @xmath6 and @xmath7 are about twice as big as that of @xmath3.,width=321 ] _ biexciton .
_ in contrast to the monoexciton , the biexciton ground state is singly - degenerate and bright [ fig . [ fig_1](b ) ] .
thus , at @xmath284 , the radiative lifetime is given from eq .
( [ lifetime ] ) as @xmath285\simeq \frac{1}{2}\,\tau_{03}(xx^0).\ ] ] similarly to the @xmath0 case , in eq .
( [ tau.xx0 ] ) we have used the relationship @xmath286 that our calculations predict , and we have neglected the long - lived dark recombination channels .
figure [ fig_5](a ) shows @xmath287 as a function of height .
we find a weak dependence with height , as in the monoexciton case .
remarkably , we find that @xmath288 regardless of height .
the latter leads to the following relationship between the radiative lifetime of the monoexciton and biexciton , @xmath289 figure [ fig_5](b ) shows that at @xmath265 the calculated @xmath18 depends weakly on height , changing by about @xmath290 in the entire range of heights studied .
@xmath18 is in excellent agreement with the value of @xmath291 measured by ulrich _
et al _ in an ( in , ga)as / gaas quantum dot ( @xmath292-@xmath293 , @xmath294-@xmath205 ) with an exciton gap of @xmath295 ( ref . ) .
_ trions . _ both @xmath6 and @xmath7 have a two - fold degenerate ground state that is bright [ figs . [
fig_1](c ) and ( d ) ] .
in each of these excitons , there are four lowest transitions ; namely , @xmath132 [ fig . [ fig_1](c ) ] , @xmath296 , @xmath297 , and @xmath298 ; and @xmath133 [ fig .
[ fig_1](d ) ] , @xmath299 , @xmath300 , and @xmath301 , for @xmath6 and @xmath7 , respectively .
in turn , each of these transitions have a corresponding characteristic radiative lifetime . for the latter ,
our calculations predict @xmath302 and @xmath303 as well as @xmath304 and @xmath305 .
thus , at @xmath20 , the radiative lifetimes of @xmath6 and @xmath7 are given by @xmath306+\frac{1}{2}\left[\frac{1}{\tau_{10}(x^-)}+\frac{1}{\tau_{11}(x^-)}\right]\right\}^{-1}=\frac{\tau_{00}(x^-)\tau_{01}(x^-)}{\tau_{00}(x^-)+\tau_{01}(x^-)},\nonumber\\ \\
\tau(x^+)=\left\{\frac{1}{2}\left[\frac{1}{\tau_{00}(x^+)}+\frac{1}{\tau_{01}(x^+)}\right]+\frac{1}{2}\left[\frac{1}{\tau_{10}(x^+)}+\frac{1}{\tau_{11}(x^+)}\right]\right\}^{-1}=\frac{\tau_{00}(x^+)\tau_{01}(x^+)}{\tau_{00}(x^+)+\tau_{01}(x^+)}.\nonumber\\\end{aligned}\ ] ] figure [ fig_5](a ) shows the height dependence of @xmath307 and @xmath308 , and @xmath309 and @xmath310 .
these _ characteristic _ radiative lifetimes depend strongly and non - monotonically on height .
this non - monotonic dependence translates into a rather simple and monotonic @xmath16 and @xmath17 , as shown in fig . [
fig_5](b ) .
for flat dots @xmath311 , whereas for taller dots these lifetimes become slightly different .
our predicted @xmath16 are in satisfactory agreement with the value of @xmath312 recently observed by smith and co - workers in ( in , ga)as / gaas dots ( size unspecified ) with exciton ground - state emission at @xmath313 ( ref . ) .
we find that the radiative lifetimes of the charged trions satisfy the relationship @xmath314 .
for completeness , we briefly compare the binding and recombination energies , polarization anisotropy , and radiative lifetimes in lens - shaped pure , non - alloyed inas / gaas dots with in@xmath4ga@xmath5as / gaas dots .
note that williamson , wang , and zunger have already compared results for @xmath0 for several alloy profiles.@xcite in ref .
we predicted that hole localization takes place at the dot - gaas matrix interface as the height of these dots increases above @xmath315 .
thus , we discuss here two flat dots ( @xmath141 and @xmath315 ) with base @xmath316 .
\(i ) recombination energies are smaller in inas / gaas dots than in in@xmath52ga@xmath131as / gaas dots with the same geometry . for instance , @xmath317 and @xmath318 for @xmath141 and @xmath315 , respectively .
\(ii ) the spectroscopic shifts show the same trends with height in flat inas / gaas as those in flat in@xmath130ga@xmath131as / gaas dots .
so do the binding energies @xmath192 and @xmath193 , and @xmath194 . however , there are two important differences bewteen the pure , non - alloyed inas / gaas dots and their in@xmath130ga@xmath131as / gaas counterparts : ( a ) for the @xmath205 - tall inas / gaas dot , while the binding energies still satisfy @xmath319 , we find that @xmath7 and @xmath3 are _ bound _ , with @xmath320 and @xmath321 , respectively .
this is so because at this height the inas / gaas dot is in the nearly `` symmetric '' regime : @xmath322 , so the hartree - fock component of the binding energy [ see eq .
( [ correlation ] ) for @xmath3 case ] is much smaller than in the in@xmath130ga@xmath5as / gaas dot with same height .
thus , correlation becomes capable of binding @xmath7 and @xmath3 in the pure , non - alloyed dot .
( b ) for the @xmath315 - tall inas / gaas dot , we find _ ordering reversal _ , i.e. @xmath323 . in in@xmath4ga@xmath5as / gaas dots ,
this ordering is attained at @xmath160 ( fig .
[ fig_3.mp ] ) .
\(iii ) in contrast to the findings in in@xmath130ga@xmath131as / gaas dots , transitions @xmath135 and @xmath230 are fully polarized along @xmath8 $ ] ( @xmath227 ) and @xmath229 $ ] ( @xmath228 ) , respectively , regardless of height . consequently , @xmath134 and @xmath231 are fully polarized along @xmath229 $ ] and @xmath8 $ ] , respectively .
these polarizations are expected from a dot with @xmath324 symmetry , like a lens - shaped pure inas / gaas dot .
\(iv ) radiative lifetimes @xmath11 , @xmath17 and @xmath16 , and @xmath18 are similar to those in in@xmath130ga@xmath325as / gaas dots with the same geometry .
for instance , @xmath326 and @xmath327 for @xmath203 ; and @xmath328 and @xmath329 for @xmath141 .
we have addressed the height dependence of recombination energies , polarization and radiative lifetimes of the lowest optical transitions of the neutral exciton ( @xmath0 ) , negatively- ( @xmath1 ) and positively - charged ( @xmath2 ) trions , and the biexciton ( @xmath3 ) in lens - shaped , self assembled in@xmath4ga@xmath5as / gaas quantum dots . we have predicted the following .
\(i ) the recombination energy of the lowest transition of @xmath6 , @xmath7 and @xmath3 , correspondingly , @xmath132 , @xmath133 and @xmath134 shows qualitatively different behavior for each excitonic complex .
namely , @xmath132 blue - shifts as height increases , whereas that of @xmath133 red - shifts .
on the other hand , as height increases , @xmath134 shows a red - shift at small heights , reaches a maximum shift , and then blue - shifts for taller dots .
this behavior is explained by the height dependence and relative magnitude of @xmath164 , @xmath165 and @xmath166 .
\(ii ) the binding energies @xmath192 , @xmath193 and @xmath194 follow the height dependence of the emission spectroscopic shifts .
changes in the dot height drives a bound - to - unbound crossover for each of these complexes .
\(iii ) the in - plane polarization anisotropy @xmath225 of the lowest transitions of @xmath0 ( @xmath78 ) and @xmath3 ( @xmath86 ) strongly depends on dot height as well as on alloy randomness ( disorder realization ) .
in contrast , the lowest transitions of @xmath6 and @xmath7 present negligible @xmath225 regardless of height .
\(iv ) the ground state of @xmath0 encompasses four states that split off in a low - energy pair that is dark and a high - energy pair that is bright , with a bright - dark splitting that increases as height decreases .
thus , at @xmath330 the radiative lifetime @xmath11 of @xmath0 is long . on the other hand , at @xmath13
both dark and bright states are populated ; so , @xmath11 becomes fast , moderately decreases as height increases , and its magnitude ranges from @xmath14-@xmath331 .
in contrast , @xmath16 , @xmath17 and @xmath18 are fast even at @xmath332 , as a consequence of these excitons having ground states that are bright .
these radiative lifetimes depend weakly on height .
further , @xmath333 , while @xmath334 .
s. rodt , r. heitz , r. l. sellin , a. schliwa , k. ptschke , and d. bimberg , physica e * 21 * , 1065 ( 2004 ) ; s. rodt , r. heitz , a. schliwa , r. l. sellin , f. guffarth , and d. bimberg , phys .
b * 68 * , 035331 ( 2003 ) . l. landin , m. s. miller , m .- e .
pistol , c. e. pryor , and l. samuelson , science * 280 * , 262 ( 1998 ) ; l. landin , m .- e .
pistol , c. pryor , m. persson , l. samuelson , and m. miller , phys .
b * 60 * , 16640 ( 1999 ) . , d. bimberg , m. grundmann and n. n. ledenstov ( wiley , chichester , 1999 ) ; _ nanostructures theory and modeling _ , c. delerue and m. lannoo ( springer - verlag , berlin , 2004 ) ; t. takagahara , phase transitions * 68 * , 281 ( 1999 ) ; _ optical properties of semiconductor quantum dots _ , u. woggon ( springer - verlag , berlin , 1997 ) s. lee , o. l. lazarenkova , p. von allmen , f. oyafuso , and g. klimeck , phys
b * 70 * , 125307 ( 2004 ) ; s. lee , f. oyafuso , p. von allmen , and g. klimeck , phys . rev .
b * 69 * , 045316 ( 2004 ) ; r. santoprete , b. koiller , r. b. capaz , p. kratzer , q. k. k. liu , and m. scheffler , phys . rev .
b * 68 * , 235311 ( 2003 ) .
m. bayer , g. ortner , o. stern , a. kuther , a. a. gorbunov , a. forchel , p. hawrylak , s. fafard , k. hinzer , t. l. reinecke , s. n. walck , j. p. reithmaier , f. klopf , and f. schfer , phys .
b * 65 * , 195315 ( 2002 ) . | we calculate the height dependence of recombination energies , polarization and radiative lifetimes of the optical transitions of various excitonic complexes : neutral excitons ( @xmath0 ) , negatively- ( @xmath1 ) and positively - charged ( @xmath2 ) trions , and biexcitons ( @xmath3 ) in lens - shaped , self assembled in@xmath4ga@xmath5as / gaas quantum dots . by using an atomistic pseudopotential method combined with the configuration - interaction method , we predict the following .
( i ) the recombination energy of the lowest transition of @xmath6 blue - shifts as height increases , whereas that of @xmath7 red - shifts .
remarkably , the recombination of @xmath3 shows a red - shift at small heights , reaches a maximum shift , and then blue - shifts for taller dots .
this feature results from the height dependence and relative magnitude of the inter - electronic direct coulomb interaction .
( ii ) changes in dot height lead to a bound - to - unbound crossover for @xmath6 , @xmath7 and @xmath3 . (
iii ) when considering the @xmath8 $ ] and @xmath9 $ ] directions , the lowest transitions of @xmath0 and @xmath3 manifest @xmath8 $ ] vs @xmath9 $ ] in - plane polarization anisotropy that switches sign as a function of height as well as alloy randomness .
@xmath6 and @xmath7 show transitions with negligible polarization anisotropy regardless of height .
( iv ) the ground state of @xmath0 is split in a low - energy pair that is forbidden ( dark ) and a high - energy pair that is allowed ; thus , at @xmath10 the radiative lifetime @xmath11 is long ( @xmath12 ) due to the dark exciton . on the other hand , at @xmath13 ,
@xmath11 decreases moderately as height increases and its magnitude ranges from @xmath14-@xmath15 .
the ground state of @xmath6 and @xmath7 , and that of @xmath3 is allowed ( bright ) ; so , @xmath16 , @xmath17 and @xmath18 are fast ( @xmath19 ) even at @xmath20 .
these radiative lifetimes depend weakly on height .
in addition , @xmath21 , while @xmath22 .
we compare our predictions with available spectroscopic data . |
we summarize essential concepts of the one - dimensional ( 1d ) spin-1/2 heisenberg model of a magnetic ring ( see e.g. @xcite or @xcite ) .
we denote by @xmath33 the set @xmath34 of the integers @xmath35 and by @xmath36 the set of all functions @xmath37 .
we assign to every function @xmath38 the sequence @xmath39 of its values over the set @xmath33 .
then the hilbert space of a ring model with @xmath1 nodes and a spin - alphabet of @xmath40 letters ( see @xcite ) is the set of all formal complex linear combinations @xmath41 of the @xmath42 in which the set @xmath43 of all @xmath42 is considered a set of linearly independent elements .
we equip @xmath0 with the scalar product @xmath44 obviously , @xmath0 is an hilbert space of dimension @xmath45 , in which @xmath46 is an orthonormal basis .
the original 1d spin-@xmath47 heisenberg ring arises for @xmath48 , where @xmath49 represents an _ up spin _ @xmath50 and @xmath51 a _ down spin _
@xmath52 at site @xmath53 .
according to @xcite , the hamiltonians @xmath54 ( @xmath55 ) of a 1d spin-@xmath47 heisenberg ferromagnet ( antiferromagnet ) of @xmath1 sites with periodic boundary conditions @xmath56 , @xmath57 , are defined by @xmath58\,,\hspace*{0.5cm}h_a : = - h_f\,,\hspace*{0.5cm}j = \mathrm{const . }
> 0\ ] ] where @xmath59 , @xmath60 and @xmath61 are spin flip operators .
symmetry classes classes in @xmath0 can be defined in analogy to the definition of symmetry classes of tensors ( see @xcite ) .
first we define symmetry operators .
let @xmath0 be the hilbert space of a ring model of @xmath1 sites and @xmath40 letters , @xmath2 be the symmetric group of the permutations of @xmath35 and @xmath3 $ ] be the complex group ring of @xmath2 .
* following @xcite we define the action of a permutation @xmath62 on a basis vector @xmath63 by @xmath64 . *
every group ring element @xmath65 $ ] acts as so - called symmetry operator on the vectors @xmath66 by @xmath67 a symmetry operator is a linear mapping @xmath68 which possesses the following properties : * it holds @xmath69 for all @xmath70 , where @xmath71 is defined by @xmath72 . *
it holds @xmath73 for all @xmath74 $ ] and @xmath75 , where @xmath76 denotes the multiplication @xmath77 of group ring elements .
* it holds @xmath10 for all @xmath4 $ ] and @xmath11 , where @xmath5 denotes the complex conjugate of @xmath6 and the element @xmath7 $ ] of @xmath8 is defined by @xmath9 .
now we define symmetry classes in @xmath0 .
let @xmath16 $ ] be a right ideal of @xmath3 $ ] .
then @xmath78 is called the symmetry class of @xmath0 defined by @xmath17 . as for tensors one can prove ( see e.g. ( * ? ? ?
4 ) or @xcite ) . let @xmath79 $ ] be a generating idempotent of a right ideal @xmath16 $ ] , i.e. @xmath80 $ ] .
then a @xmath81 is in @xmath82 iff @xmath83 .
if @xmath84 are generating idempotents of a right ideal @xmath16 $ ] , then @xmath85 iff @xmath86 .
we denote by @xmath87 the set @xmath88\;|\;au = 0\ ; \forall\,u\in\mathcal{h}\}$ ] . if @xmath89 then @xmath90 , because then the idempotent @xmath91 satisfies @xmath92 for all @xmath93 .
@xmath87 is a two - sided ideal of @xmath3 $ ]
. every two - sided ideal of @xmath3 $ ] has one and only one generating idempotent which is central
. let @xmath94 be the generating idempotent of @xmath87 .
then @xmath95 is also a central idempotent which is orthogonal to @xmath94 , i.e. @xmath96 , and which generates a two - sided ideal @xmath97 $ ] that fulfills @xmath3 = \mathcal{j}\oplus \mathcal{j}_0 $ ] .
@xmath98 contains all those symmetry operators @xmath6 for which @xmath99 .
these are the symmetry operators in which we are interestet .
( compare @xcite . ) a special type of symmetries are symmetries of elements @xmath81 with respect to commutations of the nodes of the ring .
let @xmath100 be a subgroup of @xmath2 and @xmath101 be a homomorphism of @xmath102 onto a finite subgroup in the group @xmath103 of complex units .
we say that @xmath81 possesses the commutation symmetry @xmath104 if @xmath105 for all @xmath106 .
[ prop2.14]let @xmath104 be a commutation symmetry .
* then the group ring element @xmath107 is an idempotent of @xmath108\subseteq\mathbb{c}[\mathcal{s}_n]$ ] . *
a @xmath81 has the symmetry @xmath104 iff @xmath109 .
consequently , a commutation symmetry @xmath110 defines the symmetry class @xmath82 of the right ideal @xmath111 $ ] .
in the present paper we are interested in symmetry operators @xmath4 $ ] which are self - adjoint with respect to the scalar product @xmath112 of @xmath0 .
such symmetry operators can lead to _ observables _ of the ring model .
first we find [ prop2.1 ] two symmetry operators @xmath113 $ ] satisfy @xmath114 for all @xmath11 iff @xmath115 with @xmath116 . the relations @xmath117 and @xmath114 for all @xmath11 lead to @xmath118 for all @xmath11 . from this
we obtain @xmath119 for all @xmath120 .
@xmath121 [ lem2.2 ] 1 . if @xmath122 , then @xmath123 , too .
2 . if @xmath124 , then @xmath125 , too . we show ( ii ) . if @xmath126 for all @xmath81 , then @xmath127 for all @xmath128 , i.e. @xmath129 .
further , the condition @xmath126 for all @xmath81 leads to @xmath130 for all @xmath131 .
thus we obtain @xmath132 for all @xmath120 , i.e. @xmath133 . finally , a combination of @xmath134 yields @xmath135 .
@xmath121 + now it follows from proposition [ prop2.1 ] and lemma [ lem2.2 ] a symmetry operator @xmath122 is self - adjoint iff @xmath136 . according to proposition [ prop2.1 ] @xmath6 is self - adjoint iff @xmath137 with @xmath116 .
but since @xmath138 we obtain also @xmath139 such that @xmath140 .
we say that a symmetry operator @xmath4 $ ] has property ( s ) if @xmath6 satisfies the condition + every @xmath4 $ ] with property ( s ) is self - adjoint .
however , an arbitrary symmetry operator @xmath141 $ ] with @xmath142 , @xmath143 has only to fulfil @xmath144 to be a self - adjoint operator .
we investigate in our paper operators @xmath4 $ ] which have property ( s ) .
now we search for classes of symmetry operators which possess property ( s ) .
obviously , young symmetrizers @xmath145 $ ] do not have property ( s ) in general .
for instance , one can easily check that the young symmetrizer @xmath146 + [ 2,1,3 ] - [ 3,1,2 ] - [ 3,2,1]\in\mathbb{c}[\mathcal{s}_3]$ ] of the standard tableau @xmath147 does not fulfil ( s ) .
the following theorem yields a big manifold of symmetry operators with property ( s ) .
let @xmath148 $ ] be a subgroup and @xmath13 be an irreducible character of @xmath14 . then denotes the cardinality of @xmath14 .
] @xmath149 is an idempotent which satisfies ( s ) , too .
the assertion follows from ( * ? ? ?
ii.1.47 ) , @xmath150 and @xmath151 .
if @xmath152 is a commutation symmetry , then @xmath153 is a 1-dimensional character of @xmath14 which is irreducible .
consequently , the idempotent @xmath154 of every commutation symmetry has property ( s ) .
the assertion follows from ( * ? ? ?
ii.1.48 ) . @xmath121 + already the commutation symmetries lead to a big manifold of symmetry operators with property ( s ) .
let us consider the @xmath155 .
a list of all commutation symmetries belonging to subgroups of @xmath2 with @xmath156 is given in ( * ? ? ?
* appendix a.1 ) .
the @xmath155 possesses 55 conjugacy classes of subgroups @xmath157 ( solvable and non - solvable ) .
the number of solvable subgroups @xmath157 of @xmath155 is equal to 1428 ( see ( * ? ? ? *
iii.2.2 ) ) . on all theses
subgroups one can define between 1 and 8 commutation symmetries .
this yields a big number of operators with property ( s ) .
however , if we use irreducible character for the construction of operators with property ( s ) , the number of such operators becomes still bigger .
[ thm2.8 ] let @xmath158 $ ] be a minimal right ideal with @xmath159 .
( @xmath17 defines a symmetry class ) .
then the set of ( primitive ) generating idempotents of @xmath17 contains one and only one idempotent @xmath160 with property ( s ) .
@xmath160 can be formed from an arbitrary generating idempotent @xmath18 of @xmath17 by @xmath161 a corresponding statement holds true for minimal left ideals @xmath162 $ ] .
for left ideals one has to replace ( [ eq6 ] ) by @xmath163 .
a statement similar to theorem [ thm2.8 ] was proven by h. weyl in @xcite .
we give here a new proof .
since @xmath17 is minimal and @xmath159 we have @xmath164 .
let @xmath18 be a ( primitive ) generating idempotent of @xmath17 . obviously , @xmath165 is a group ring element with property ( s ) .
( note that @xmath166 . )
we show that @xmath167 is essentially idempotent .
since @xmath168 there exists an @xmath81 such that @xmath169 . using the scalar product we obtain @xmath170 and @xmath171 .
consequently , we have @xmath172 and @xmath173 .
the relation @xmath174 can be proved when we start a similar consideration from @xmath175 , where @xmath120 is an element with @xmath176 .
now we consider @xmath177 .
this leads to @xmath178 .
if @xmath18 is a primitive idempotent then @xmath179 is proportional to @xmath18 for all @xmath180 $ ] . since @xmath178
we obtain @xmath181 with @xmath182 and @xmath183 .
thus @xmath167 ist essentially idempotent .
next we show that @xmath184 . from @xmath181
it follows @xmath185 . if we multiply this relation from the left by @xmath18 we obtain @xmath186 and @xmath187 .
so @xmath188 is an idempotent which has property ( s ) because @xmath184 .
finally @xmath189 = ( e\cdot\overline{e}^{\ast})\cdot\mathbb{c}[\mathcal{s}_n]$ ] is a right ideal with @xmath190 .
since @xmath191 and @xmath17 is minimal we obtain @xmath192 , i.e. @xmath160 generates @xmath17 .
now we show the uniqueness of @xmath160 .
assume @xmath17 possesses two generating idempotents @xmath193 , @xmath194 with property ( s ) .
we denote by @xmath82 the symmetry class defined be @xmath17 and by @xmath195 the orthogonal complement of @xmath82 with respect to @xmath112 .
every idempotent @xmath196 satisfies @xmath197 for all @xmath198 .
since every @xmath199 has a structure @xmath200 with @xmath201 we obtain @xmath202 for all @xmath203 and all @xmath201 .
this leads to @xmath204 for all @xmath203 .
we see that both idempotents @xmath193 and @xmath194 have the same effect on @xmath82 and @xmath195 . consequently , we have @xmath205 for all @xmath201 , from which follows @xmath206 .
but since @xmath164 , @xmath193 and @xmath194 lie in @xmath98 and @xmath207 .
in this section we present an algorithm for the decomposition of an idempotent @xmath18 with property ( s ) into pairwise orthogonal , primitive idempotents @xmath196 which have property ( s ) , too .
[ thm3.1 ] every idempotent @xmath168 with @xmath19 has a decomposition @xmath208 into primitive idempotents @xmath209 which fulfill @xmath210 the idempotent @xmath18 generates a right ideal @xmath80\subseteq\mathcal{j}$ ] .
the decomposition ( [ eq7 ] ) can be constructed by the following @xmath211 + [ alg3.2 ] 1 .
form a set @xmath212 of primitive idempotents @xmath213 $ ] such that @xmath3 = \bigoplus_{y\in\mathcal{y } } y\cdot\mathbb{c}[\mathcal{s}_n]$ ] .
2 . determine a @xmath214 such that @xmath215 .
the group ring element @xmath216 generates a minimal right ideal @xmath217 $ ] with @xmath218 .
3 . determine a primitive generating idempotent @xmath219 of @xmath220 from @xmath216 .
4 . calculate the unique generating idempotent @xmath221 of @xmath220 .
5 . calculate the rest @xmath222 .
determine further idempotents @xmath223 by iteration of the steps ( ii ) , , ( v ) starting with the rests @xmath224 instead of @xmath18 .
stop this iteration when a rest @xmath225 was reached .
the set @xmath212 could be constracted from the young symmetrizers of all standard tableaux of @xmath3 $ ] . in step ( ii )
the set @xmath212 is always the original set @xmath212 .
we do not delete from @xmath212 the elements @xmath226 found in previous steps ( ii ) .
step ( iii ) can be carried out by a procedure given in ( * ? ? ? *
prop.i.2.1 ) , ( * ? ? ?
* prop.1 ) or ( * ? ? ?
* prop.4.1 ) now we show that algorithm [ alg3.2 ] produces the decomposition ( [ eq7 ] ) . obviously , the rest @xmath227 has property ( s ) .
further it follows @xmath228 from @xmath218 .
this leads to @xmath229 .
now we can carry out the calculation @xmath230 which yield @xmath231 ( for all @xmath232 ) and @xmath233 since @xmath234 . from @xmath233
we obtain @xmath235 , i.e. @xmath236 is an idempotent with property ( s ) .
thus the start of a second iteration of ( ii ) , , ( v ) with @xmath236 instead of @xmath18 is correct .
the result of this iteration are idempotents @xmath237 which fulfil @xmath238 , @xmath239 .
consider the symmetry classes @xmath240 , @xmath241 , @xmath242 of the right ideals @xmath243 $ ] , @xmath244 $ ] , @xmath245 $ ] .
we obtain @xmath246 from ( [ eq9 ] ) and @xmath247 from @xmath239 .
consequently , it holds @xmath248 , too .
now a calculation similar to ( [ eq9 ] ) with @xmath194 instead of @xmath236 yields @xmath249 . since these considerations can be carried out in all iteration steps of algorithm [ alg3.2 ] theorem [ thm3.1 ]
is correct .
computer calculations in big group rings @xmath3 $ ] have high costs in calculation time and memory because of the high number of permutations in @xmath2 .
one has to use a _ discrete fourier transform _ for @xmath2 to reduse these costs .
( see @xcite for discrete fourier transforms of groups . )
a discrete fourier transform for @xmath2 is an isomorphism @xmath250 \;\rightarrow\ ; \bigotimes_{\lambda\vdash n } { \mathbb{c}}^{d_{\lambda}\times d_{\lambda } } & , & d:\;a = \sum_{p\in\mathcal{s}_n } a_p\,p \;\mapsto\ ; d(a ) = \left ( \begin{array}{cccc } a_{{\lambda}_1 } & 0 & 0 & 0 \\ 0 & a_{{\lambda}_2 } & 0 & 0 \\ \vdots & \vdots & \ddots & \vdots \\ 0 & 0 & 0 & a_{{\lambda}_l } \\ \end{array } \right)\hspace*{0.5cm}\ ; \label{eq10}\end{aligned}\ ] ] according to wedderburn s theorem which maps the group ring @xmath3 $ ] onto an outer direct product @xmath251 of full matrix rings @xmath252 .
we denote by @xmath253 the natural projection @xmath254\rightarrow { \mathbb{c}}^{d_{\lambda}\times d_{\lambda}}$ ] , @xmath255 , belonging to the partition @xmath30 of @xmath1 .
the dimension @xmath256 of the @xmath257-block matrices @xmath258 can be computed from @xmath259 by means of the _ hook length formula _ ( see e.g. @xcite or @xcite ) .
the group ring @xmath3 $ ] decomposes into _ minimal two - sided ideals _
@xmath29 , @xmath260 = \bigoplus_{\lambda\vdash n } \mathcal{z}_{\lambda } = \bigoplus_{\lambda\vdash n } \mathbb{c}[\mathcal{s}_n]\cdot z_{\lambda}\ , .
\label{eq11}\ ] ] every @xmath29 is generated by a unique idempotent @xmath31 which is _ centrally primitive_. @xmath31 can be calculated by means of ( [ eq5 ] ) from the irreducible character @xmath261 belonging to @xmath30 .
consequently every @xmath31 has property ( s ) .
the structure of block matrices in ( [ eq10 ] ) reflects the decomposition ( [ eq11 ] ) since @xmath262 is an isomorphism .
elements @xmath263 $ ] which lie in minimal left , right or two - sided ideals of @xmath3 $ ] , have only one non - vanishing block matrix in ( [ eq10 ] ) .
in particular this holds true for all primitive idempotents .
the product @xmath264 of such an element @xmath6 by another element @xmath265 is in @xmath3 $ ] the product of two long sums @xmath8 , @xmath266 of length @xmath267 , however in @xmath251 only the product of two @xmath257-matrices @xmath268 which are much smaller .
this leads to a considerable reduction of computation costs .
the steps ( i)-(iii ) of algorithm [ alg3.2 ] were taken from another ideal decomposition algorithm which we described in ( * ? ? ? * chap.i.2 ) or @xcite .
the algorithm from @xcite runs both in @xmath3 $ ] and in @xmath251 without a necessity of a fourier transformation between these two spaces during the run of the algorithm .
the version in @xmath251 is much more efficient than the version in @xmath3 $ ] .
because of the relationship between algorithm [ alg3.2 ] and the algorithm from @xcite one can transfer algorithm [ alg3.2 ] to @xmath251 by means of the remarks in @xcite , ( * ? ? ?
* sec.5 ) , @xcite .
only the computation of @xmath269 for an element @xmath4 $ ] can not be transfered to @xmath251 in a simple way .
the calculation of the matrix @xmath270 from a matrix @xmath271 can not be carried out in @xmath251 without fourier transformations between @xmath251 and @xmath3 $ ] . in the next sections we present algorithms for the efficient computation of @xmath272 from @xmath26 . for @xmath2 three discrete fourier transforms
are known : 1 .
young s natural representation of @xmath2 ( see @xcite ) , 2 .
young s semi - normal representation of @xmath2 ( see @xcite ) , 3 .
young s orthogonal representation of @xmath2 ( see @xcite ) .
the dft ( b ) is the basis of the fast fourier transform for @xmath2 by m. clausen and u. baum @xcite .
we use ( a ) as discrete fourier transform in our mathematica package perms @xcite .
the three dft map permutations @xmath62 to following types of matrices : [ cols="^,^,^,^",options="header " , ]
in this section we present an algorithm which determines the sets @xmath27 for the algorithm [ alg5.5 ] ( see figure 1 ) .
we can not simply consider the set @xmath273 und determine @xmath274 linearly independent vectors in this set , because this set contains @xmath267 vectors from @xmath3 $ ] of length @xmath267 .
this data set is very large .
the algorithm calculates @xmath275 for permutations @xmath62 and checks whether these matrices are linearly independent . the algorithm starts with the identity permutation @xmath276 and @xmath277 . to characterize the 1-dimensional linear space spanned by @xmath278 , the algorithm uses a vector @xmath279 of variables @xmath280 .
the algorithm forms the scalar product @xmath281 $ ] of the vectors @xmath282 and @xmath283 $ ] .
@xmath282 is a linear function .
the algorithm determines one of the variables @xmath280 from the equation @xmath284 and obtains a relation @xmath285rest of the @xmath286 . by means of this relation the variable @xmath287
is eliminated in @xmath288 .
we obtain a vector @xmath288 which fulfil @xmath289 exactly for the elements @xmath290 of the 1-dimensional span of @xmath291 $ ] .
now the algorithm investigates the next permutation @xmath292 . to select @xmath292 ,
the algorithm uses a procedure * next - permutation * which determines for a given permutation @xmath293 the successor @xmath62 according to the lexicographical order of permutations
. then the algorithm checks the condition @xmath294
\label{eq21}\ ] ] if ( [ eq21 ] ) is satisfied , then @xmath292 is canceled . if , however , ( [ eq21 ] ) is not fulfilled then a new linear equation @xmath284 arises which the algorithm uses to eliminate a further variable in @xmath288 .
@xmath288 caracterizes now the span of two linearly independent vectors @xmath291 $ ] , @xmath295 $ ] . the new found permutation @xmath292 ist added to the set @xmath27 .
this procedure is iterated until the cardinality of @xmath27 has reached the value @xmath296 .
the algorithm needs not much memory because only one matrix @xmath297 is stored in the computer memory in every step of the algorithm .
a halving of the calculation time can be obtained by the following theorem .
let a set @xmath27 be known for a partition @xmath30 .
then @xmath298 is a basis of @xmath299 , where @xmath300 denotes that partition whose young frame is transposed to the young frame of @xmath259 .
[ thm6.1 ] the generating idempotents @xmath31 , @xmath301 of @xmath29 , @xmath299 fulfil @xmath302 where @xmath261 , @xmath303 denote the irreducible characters of @xmath304 belonging to @xmath259 , @xmath305 .
it holds true @xmath306 , because @xmath307 , @xmath308 and @xmath309 .
further , it follows from ( * ? ? ?
246 , p. 65
, statement 4.11 ) , that @xmath310 because of these facts we can write @xmath311 where @xmath312 now we assume that @xmath313 is a basis of @xmath29 , but @xmath298 is a set of linearly dependent vectors from @xmath299 . consequently , we can form a non - trivial vanishing linear combination of the vectors from @xmath314 , @xmath315 now the element @xmath316 has a decomposition @xmath317 , where @xmath318 and @xmath319 are sums which run only through the even or odd permutations of @xmath27 , respectively . now we obtain @xmath320 it follows from ( [ eq27 ] ) that @xmath321 since @xmath322 and @xmath323 contain only even or odd permutations , respectively .
now we consider the non - trivial linear combination @xmath324 of the permutations from @xmath27 .
the product of @xmath325 and @xmath31 yields @xmath326 so , @xmath327 is a non - trivial vanishing linear combination of the vectors from @xmath328 , i.e. @xmath328 would not be a basis of @xmath29 in contradiction to the assumptions of theorem [ thm6.1 ]
we finish with an example . we consider the young symmetrizer @xmath329 $ ] of the young tableau @xmath330 , which belongs to the partition @xmath331 .
@xmath332 has the length 48 .
it is proportional to a primitive idempotent @xmath333 and generates a 4-dimensional minimal right ideal of @xmath334 $ ] .
now our algorithms yield @xmath335 @xmath253 was calculated by means of young s natural representation of @xmath336 .
@xmath337 represents the primitive idempotent @xmath160 with property ( s ) which generates the same right ideal as @xmath332 ( see theorem [ thm2.8 ] ) .
@xmath160 has the length 120 .
one can easily check the properties @xmath338 , @xmath339 , @xmath340 , @xmath341 , @xmath342 .
the relation @xmath343 can be verified by means of our mathematica package perms @xcite .
we implemented all algorithms presented in this paper in perms .
mathematica notebooks of the calculation for this paper can be downloaded from @xcite .
10 b. fiedler .
symmetry classes connected with the magnetic heisenberg ring . in tadeusz lulek , andrzej wal , and barbara lulek ,
editors , _ symmetry and structural properties of condensed matter ( sspcm 2007 ) _ , volume 104 of _ journal of physics : conference series _ , bristol , philadelphia , 2008 . iop publishing .
paper : 012035 .
online at jpconf.iop.org .
p. jakubczyk , t. lulek , d. jakubczyk , and b. lulek .
construction of kostka matrix at the level of bases . in tadeusz lulek ,
andrzej wal , and barbara lulek , editors , _ symmetry and structural properties of condensed matter ( sspcm 2007 ) _ , volume 30 of _ journal of physics : conference series _ , bristol , philadelphia , 2008 .
iop publishing .
paper : 012039 .
online at jpconf.iop.org .
bernd fiedler .
ideal decompositions and computation of tensor normal forms . in _
sminaire lotharingien de combinatoire _ , 2001 .
electronically published : ` http://www.mat.univie.ac.at/~slc ` .
b45 g , 16 pp .
archive : ` http://arxiv.org/abs/math.co/0211156 ` .
p. jakubczyk , t. lulek , d. jakubczyk , and b. lulek .
the duality of weyl and linear extension of kostka matrices . in tadeusz lulek ,
andrzej wal , and barbara lulek , editors , _ symmetry and structural properties of condensed matter ( sspcm 2005 ) _ , volume 30 of _ journal of physics : conference series _ , pages 203208 , bristol , philadelphia , 2006 .
iop publishing .
online at jpconf.iop.org .
d. jakubczyk , t. lulek , p. jakubczyk , and b. lulek .
geometry and rigged strings in bethe ansatz . in tadeusz lulek , andrzej wal , and barbara lulek ,
editors , _ symmetry and structural properties of condensed matter ( sspcm 2005 ) _ , volume 30 of _ journal of physics : conference series _ , pages 188196 , bristol , philadelphia , 2006 .
iop publishing .
online at jpconf.iop.org .
bernd fiedler .
an algorithm for the decomposition of ideals of the group ring of a symmetric group . in adalbert
kerber , editor , _ actes @xmath344 sminaire lotharingien de combinatoire , thurnau , 1997 _ , publ .
i.r.m.a . strasbourg .
institut de recherche mathmatique avance , universit louis pasteur et c.n.r.s .
( ura 01 ) , 1998 .
electronically published : ` http://www.mat.univie.ac.at/~slc ` .
b39e , 26 pp .
gordon d. james and adalbert kerber . ,
volume 16 of _ encyclopedia of mathematics and its applications_. addison - wesley publishing company , reading , mass . ,
london , amsterdam , don mills , ont . ,
sidney , tokyo , 1981 . | in @xcite we defined symmetry classes , commutation symmetries and symmetry operators in the hilbert space @xmath0 of the 1d spin-1/2 heisenberg magnetic ring with @xmath1 sites and investigated them by means of tools from the representation theory of symmetric groups @xmath2 such as decompositions of ideals of the group ring @xmath3 $ ] , idempotents of @xmath3 $ ] , discrete fourier transforms of @xmath2 , littlewood - richardson products . in the present paper we consider symmetry operators @xmath4
$ ] on @xmath0 which fulfil + ( s ) + where @xmath5 denotes the complex conjugate of @xmath6 and the element @xmath7 $ ] of @xmath8 is defined by @xmath9 .
such symmetry operators are _ self - adjoint _ because of the relation @xmath10 which holds true for all @xmath4 $ ] and @xmath11 .
they yield consequently _ observables _ of the quantum mechanical heisenberg model .
we prove the following results : * ( i ) * let @xmath12 be an arbitrary subgroup and @xmath13 be an irreducible character of @xmath14 then @xmath15 is an idempotent of @xmath3 $ ] which has property ( s ) .
this leads to a big manifold of observables .
in particular every commutation symmetry belonging to @xmath14 yields a 1-dimensional character of @xmath14 from which we can build one of the idempotents @xmath13 described here . *
( ii ) * let @xmath16 $ ] be a minimal right ideal of @xmath3 $ ] .
then the set of all generating ( primitive ) idempotents of @xmath17 contains one and only one idempotent @xmath18 which satisfies @xmath19 .
this is a result by h. weyl @xcite for which we give a new proof . *
( iii ) * every idempotent @xmath18 with property ( s ) can be decomposed into primitive idempotents @xmath20 which have also property ( s ) and satisfy @xmath21 if @xmath22 .
we give a computer algorithm for the calculation of such decompositions . in big
group rings @xmath3 $ ] computer calculations are only possible by the use of a diskrete fourier transform @xmath23\rightarrow\bigoplus_{\lambda } \mathbb{c}^{d_{\lambda}\times d_{\lambda}}$ ] .
we present two algorithms which allow the calculation of the matrix @xmath24 from a matrix @xmath25 of an @xmath4 $ ] without to determine the inverse fourier transform of @xmath26 .
these algorithms use as precomputed data permutation sets @xmath27 which define bases @xmath28 of the minimal two - sided ideals @xmath29 belonging to the partitions @xmath30 .
( @xmath31 is the central - primitive generating idempotent of @xmath29 ) .
the sets @xmath27 can be used for any discrete fourier transform @xmath32 . in our investigations
we use computer calculations by means of the mathematica packages perms and hring . |
nuclear spin polarization is a necessary condition for many types of physics experiments including studies of nuclear structure , nuclear reactions , fundamental interactions , and condensed matter physics .
many of these experiments require the polarization of radioactive ion beams ( ribs ) .
for example , ground state nuclear magnetic dipole ( @xmath1 ) and electric quadrupole moments ( @xmath2 ) are critical in the study of nuclear structure , especially as one moves away from the valley of stability .
currently , the best reach for such measurements has been demonstrated by ribs polarized in fragmentation reactions . quenched magnetic moment values are observed for the extremely neutron - deficient nuclei @xmath3c and @xmath4cu .
the small @xmath1 value of the proton drip line nucleus @xmath3c ( @xmath5=1.3 mev ) leads to a large deviation of the deduced isoscaler expectation value @xmath6 for the @xmath3li-@xmath3c mirror pair when compared to observed trends for @xmath7 nuclei @xcite .
shell model calculations were shown to better reproduce @xmath6 when isospin mixing was included @xcite .
intruder mixing with the ground state of @xmath3c caused by proton shell quenching may also account for the deviant @xmath1 value @xcite .
@xmath4cu represents the heaviest @xmath8 nuclide with a known @xmath1 value .
the magnetic moment is smaller than predicted in shell model calculations that assume @xmath9ni is a good doubly - magic nucleus @xcite .
the quadrupole moments of the b isotopes also provide evidence for changes in nuclear structure .
@xmath10 was found to be twice as large as expected from shell model calculations @xcite .
the larger @xmath2 was considered evidence of a proton halo covering a neutron core . near the neutron drip line
, the experimental @xmath2 moments for odd - mass @xmath11b , @xmath12b , and @xmath13b were found to be nearly constant , independent of neutron number @xcite .
the calculated @xmath2 obtained with standard values of effective charges overpredict the experimental @xmath2 for @xmath12b and @xmath13b isotopes .
the result reveals that both proton and neutron quadrupole polarization charges may be reduced near the neutron drip line @xcite . from these examples
, it is apparent that moment measurements play an important role in nuclear structure studies . as stated earlier , in order to perform any moment measurement , the nuclei must be spin polarized .
there is a strong dependence of polarization on the momentum of the fragment nucleus , therefore , it is crucial to know the magnitude of polarization prior to the experiment . while fragmentation reactions are key in producing exotic nuclei , these nuclei come with low rates .
a figure of merit for measurements involving polarization is @xmath14 , where @xmath15 represents polarization and @xmath16 is yield .
the optimization of polarization with yield is important .
improvements in yield will come with the development of new rib facilities , but while yields are small , the ability to accurately predict the expected polarization is required for experimental success .
broader applications for the polarized beams will be achieved with new rib facilities , but at the same time , better predictive power is needed . in general ,
the spin orientation of an ensemble of quantum states is usually specified by a statistical tensor @xcite .
the tensor can be denoted by @xmath17 , where @xmath18 is the spin of the state and @xmath19 , the rank of the statistical tensor , takes the values @xmath20 if @xmath18 is an integer , and @xmath21 if @xmath18 is half - integer .
@xmath22 takes integer values between @xmath23 and @xmath24 .
the statistical tensor is related to the distribution of magnetic substates with respect to a chosen coordinate frame .
it is usually possible to specify the magnetic substate distribution by the populations @xmath25 of the @xmath26 @xmath27-substates , where it is assumed that the @xmath25 are normalized so that @xmath28 . for an @xmath27-state distribution with axial symmetry , only the @xmath29 components of @xmath17 are non - zero .
spin polarization is defined in terms of the @xmath30 statistical tensor , denoted @xmath31 , relative to its value for maximum polarization .
specifically , @xmath32 and @xmath33 so the spin polarization is @xmath34 thus , spin polarization is a measure of the orientation of the total angular momentum relative to a fixed axis .
spin polarization of projectile - like residues from intermediate - energy heavy - ion reactions was first observed at riken @xcite .
fragments detected at small angles with respect to the normal beam axis were shown to have a polarized spin in the low - intensity wings of the momentum distribution .
a qualitative description of the polarization mechanism was found in a model that considers conservation of linear and angular momentum , and assumes peripheral interactions between the fast projectile and target .
[ fig : excelsimremoval ] presents a schematic of the expected polarization and yield for the nucleon removal process for fragmentation of a projectile on a heavy target .
a systematic study of the spin polarization following few - nucleon removal from light projectiles as a function of energy and target was completed by okuno _
_ @xcite .
this study demonstrated that the relation between the outgoing fragment momentum and sign of spin polarization depended on the mean deflection angle , @xmath35 .
near - side reactions occur for high-@xmath36 targets , where the coulomb deflection will dominate the internuclear potential between projectile and target , giving the polarization dependence shown in fig .
[ fig : excelsimremoval ] .
the nucleon - nucleon potential energy governs removal reactions on low-@xmath36 targets .
far - side reactions prevail in this case , and the sign of the observed polarization is reversed . target .
the yield and polarization curves are given relative to the incident projectile momentum .
the removal schematic is given in the projectile - like rest frame.,width=264 ] the spin polarization has a near - zero value at the peak of the fragment yield curve for both near- and far - side dominated reactions , since @xmath37@xmath35@xmath37 is large .
this behavior can be qualitively understood from the projectile rest - frame diagram in fig .
[ fig : excelsimremoval ] .
the removed nucleons have momentum * k*. the @xmath38-component of the induced angular momentum of the projectile - like species is @xmath39 , where @xmath40 are the localized cartesian coordinates of the removed nucleon(s ) , and @xmath41 are the momentum components of the removed nucleons in the reaction plane . if the nucleon removal occurs uniformly in the overlap region , @xmath42 , then @xmath43 .
zero polarization will therefore result when the fragment momentum equals the projectile momentum , since @xmath44 in the projectile rest frame under these conditions .
if nucleon removal is not uniform in the overlap region , @xmath45 and the term @xmath46 can contribute to @xmath47 .
such a contribution will only be observed experimentally when @xmath37@xmath35@xmath37 is small .
the final scattering angle of the fragment is @xmath48@xmath35@xmath49 , where @xmath50 is the change in angle caused by the transverse momentum component of the removed nucleons , @xmath51 . here
@xmath52 is the total momentum of the projectile - like fragment . in reactions where @xmath37@xmath35@xmath53
, it is the transverse momentum component of the removed nucleon(s ) that `` kicks '' the fragments to small angles , and the resulting polarization is negative since @xmath54 to give positive @xmath50 and @xmath55 for non - uniform nucleon removal as illustrated in fig . [ fig : excelsimremoval ] .
a monte carlo code was developed @xcite based on the ideas discussed above to simulate the spin polarization generated in nucleon removal reactions at intermediate energies .
the general behavior of the spin polarization as a function of projectile - like momentum was achieved , although a scaling factor of 0.25 was needed to reproduce the magnitude of polarization observed experimentally .
spin polarization via nucleon pickup reactions at intermediate energies was first demonstrated at nscl @xcite .
positive spin polarization was determined for @xmath56k species collected at small angles in the reaction of @xmath57ar projectiles on a @xmath3be target at 150 mev@xmath58nucleon .
[ fig : excelsimpickup ] illustrates the spin polarization and yield from nucleon pickup reactions .
the key to understanding the observed spin polarization in the pickup process is that the picked - up nucleon must have an average momentum equal to the fermi momentum oriented parallel to the beam direction .
et al . _
@xcite reached this conclusion from the observed shifts in the centroids of the momentum distributions for one and two - nucleon pickup products .
the average projectile - like momentum @xmath59 was found to satisfy the relation @xmath60 , with @xmath61 the average momentum of the incident particle and @xmath62 the average momentum of the picked up nucleon equal to the fermi momentum . in the rest frame of the projectile - like species ,
the momentum of the picked - up nucleon will be antiparallel to the incoming projectile momentum .
the @xmath38-component of orbital angular momentum induced by the nucleon pickup process is @xmath63 , assuming a peripheral interaction where the nucleon is picked up to a localized position on the projectile given by @xmath64 in fig .
[ fig : excelsimpickup ] . the spin polarization will be zero when the momentum of the picked - up nucleon matches the momentum of the incoming projectile ( @xmath65 ) .
the zero crossing occurs at the projectile - like momentum @xmath66{{p}_{p}}$ ] , where @xmath67 and @xmath68 are the mass number and momentum of the projectile , respectively .
a linear increase in @xmath47 is expected with a decrease in the momentum of the outgoing pickup product .
et al . _
@xcite found that proton pickup reactions follow the trend shown in fig .
[ fig : excelsimpickup ] , except for the low momentum side of the momentum distribution . at low momentum values of the pickup products , the matching conditions for pickup
will no longer be satisfied , and the spin polarization is observed to rapidly approach zero .
turz _ et al .
_ showed that neutron pickup reactions at intermediate energies behave in a similar manner @xcite .
they extended the monte carlo simulation of ref .
@xcite to include nucleon pickup and the momentum considerations discussed by groh _
et al . _
@xcite . qualitative agreement of the observed spin polarization as a function of the projectile - like product was realized , as was the case with nucleon removal reactions . but
again , a scaling factor of 0.25 was needed to reproduce the magnitude of the observed spin polarization .
scaling factors of the same magnitude required for both nucleon removal and nucleon pickup suggest that the same quantitative correction factors should apply to both . the kinematical model proposed by asahi _ et al .
_ has been successfully employed to qualitatively explain spin polarization at intermediate energies for both nucleon removal and pickup reactions .
this paper introduces additional considerations to the kinematical model aimed to improve the quantitative agreement of the monte carlo simulation with the experimentally - observed spin polarization in nucleon removal and pickup reactions .
the simulation code developed by okuno _
et al . _
@xcite was used to compute the momenta of the removed nucleons in monte carlo fashion .
the kinematical equations from ref .
@xcite were then applied to calculate the spin polarization @xmath69 as a function of fragment momentum .
the nucleon removal positions @xmath40 on the projectile surface are determined by the relations @xmath70 and @xmath71 , where @xmath72 as shown in fig .
[ fig : excelsimremoval ] . improvements made to the original simulation code , which advance toward better quantitative agreement with spin polarization in both nucleon removal and nucleon pickup reactions , are discussed below .
one shortcoming of the original simulation code was that the projectile - target interaction was one - sided .
we have adopted an absolute coordinate system to maintain the correct relationship between the sign of the polarization , the emission angle of the fragment , and the momentum of the fragment .
interactions between projectile and target are permitted in the reaction plane , on both sides " of the target , as shown in fig .
[ fig : sidedness ] .
the coordinate system is defined as follows .
positive @xmath73 is the beam direction , positive @xmath74 is defined to the right of the target relative to the beam direction , and positive @xmath38 is perpendicular to the scattering plane forming a right - handed coordinate system .
positive angles are defined to the left of the @xmath73-axis , or toward negative @xmath74 . for left - sided interactions , a near - side(far - side ) collision will scatter to the left(right ) , or to positive(negative ) angles . for the interactions on the right side of the target
, a near(far)-side interaction must scatter to the defined negative(positive ) angles , thus the signs of the mean deflection angles are changed for these events .
one of the fundamental phenomena in fragmentation reactions related to polarization is that no fragment spin polarization is observed when the mean fragment angle , @xmath75 , is zero degrees @xcite . shown in fig .
[ fig : zerodeg ] is the calculated polarization for 109.6 mev@xmath58nucleon @xmath12n fragmented in a target of @xmath76au to make @xmath11b at a mean fragment angle of 0@xmath77 with an angular acceptance of 0.25@xmath77 .
the calculated polarization is equivalent to zero within statistical error .
allowance for right- or left - sided collisions correctly accounts for the observed absence of spin polarization . on average ,
half of the events detected at 0@xmath77 are from a right - sided interaction , while the other half are from left - sided interactions .
the resulting polarization for each type of interaction is equal in magnitude , but opposite in sign , giving an average of zero polarization .
nucleon @xmath12n in a target of @xmath76au to make @xmath11b at a fragment angle of 0@xmath77 .
@xmath78 is the momentum of the incident projectile and @xmath52 is the momentum of the outgoing fragment .
the grey band represents the range of simulation results within a @xmath79 distribution.,width=264 ] previously , an average value for the location @xmath40 of the removed nucleons in the projectile was calculated based on the rotation angle @xmath80 .
the same removal location was then used for every event in the simulation .
the removal location is now calculated based on the volume of intersection of a cylinder and sphere following the prescription of gosset _ et al . _
the number of nucleons removed is directly proportional to the overlap volume .
a removal position was calculated in monte carlo fashion for each nucleon in the overlap region and then averaged to give the position @xmath81 of the group of removed nucleons .
the rotation angle @xmath80 is accommodated in this approach by offseting the @xmath16 position by a fixed value , yet ensuring that the offset @xmath16 value remains in the overlap region .
the mean deflection angle @xmath35 is calculated by way of numerical integration .
required input parameters are the @xmath36 and @xmath82 of both projectile and target , the energy of the projectile , the distance of closest approach , @xmath83 , and the real part of the optical model potential for the nucleus - nucleus interaction , @xmath84 .
the relations of gossett _ et al . _
@xcite are used to determine @xmath83 . in the original incarnation of the simulation code ,
all resulting fragments scatter to a single angle defined by @xmath35 , based on the number of nucleons removed plus the angular impulse given to the fragment in the transverse direction .
however , for a given number of nucleons removed , fragments will scatter to a range of angles whose mean is the mean deflection angle . to better simulate experimental conditions ,
a distribution of scattering angles whose average is @xmath35 has been considered , including rutherford , fermi , and step function distributions .
the rutherford distribution describes classical elastic scattering and is strongly peaked at 0@xmath77 .
this distribution was problematic because fragmentation reactions are not classical elastic scattering , and the rutherford distribution is incompatible with a mean angle larger than about 1@xmath77 .
the step function distribution ranged in angle from zero to twice the mean deflection angle , with the functional form shown below : @xmath85 even though the mean is @xmath35 , the angular distribution is non - physical .
in particular , the maximum number of fragments did not occur at 0@xmath77 as observed in experiment .
the fermi distribution ( i.e. a woods - saxon shape ) has the form @xmath86 where @xmath87 , @xmath88 is the scattering angle and the value of @xmath74 is varied to make the mean of the distribution equal to the mean deflection angle .
the correct mean for the distribution could only be achieved with a negative value for @xmath74 , which is unphysical .
a compromise was made on a straight - line distribution with negative slope .
this distribution has the advantage of peaking at 0@xmath77 , and by changing the slope , the desired mean deflection angle can be imposed on the distribution . a calculated angular distribution with @xmath35=3@xmath77 is shown in fig .
[ fig : defangle ] .
= 3@xmath77.,width=264 ] in the original simulation code , projectile - target interactions that occurred only in the @xmath89 plane were considered .
however , fragments should interact with the target on the top , the bottom , and all the angles in between , not just on the right or left side of the target in the horizontal x - y plane .
such non - equatorial interactions would decrease the @xmath38-component of angular momentum in the equatorial ( reaction ) plane , as shown in fig .
[ fig : outofplane ] .
the primed frame represents non - equatorial scattering and can be represented as a rotation of the coordinate system about the y - axis through the angle @xmath90 .
plane ( unprimed axes ) .
labels defined in the text.,width=264 ] the resulting @xmath38-component of angular momentum for the non - equatorial scattering is @xmath91 , where @xmath92 is the component of angular momentum projected on the @xmath74 axis and @xmath90 is the rotation angle . in theory
, @xmath90 can range from @xmath93 to @xmath94 , since the projectile interaction can occur anywhere in the @xmath95 plane .
integrating all possible contributions from the out - of - plane acceptance , and including a @xmath96 normalization factor from the interval of integration gives @xmath97 experimental devices have a limited angular acceptance , and thus the range on @xmath90 will be smaller than @xmath93 to @xmath94 .
[ fig : outofplane2 ] shows a schematic of the beam view of an angular acceptance of 2.0@[email protected]@xmath77 horizontally and @xmath982@xmath77 vertically .
@xmath980.5@xmath77 and the vertical acceptance is @xmath982@xmath77.,width=264 ] the out - of - plane acceptance of @xmath982@xmath77 limits the range on @xmath90 to -53@xmath77 to + 53@xmath77 for this case .
equation [ eq : lz ] with these limits of integration and a re - normalization of 106@xmath77 rather than 180@xmath77 gives @xmath99 .
as the acceptance window moves further from the beam axis , the range on the angle @xmath90 will decrease .
this causes the value of @xmath100 to approach that of @xmath101 .
various example acceptance windows , @xmath90 angles ( i.e. ranges on @xmath90 ) and the resulting corrections for @xmath101 in terms of @xmath100 are listed in table [ tab : outofplane ] .
a vertical acceptance of @xmath982@xmath77 is assumed in all cases , common in most fragment separators .
.polarization correction factors due to non - equatorial scattering . [ cols="^,^,^,^",options="header " , ] comparison of the spin polarization measurements for intermediate - energy reactions that include two - nucleon removal in ref .
@xcite with the simulation results that include considerations for angular distributions , out - of - plane scattering , and @xmath0-ray deorientation , are presented in fig .
[ fig : removaldatasim ] .
no scaling factor was used to adjust the simulated spin polarization .
relative to the primary beam momentum @xmath102 for the removal reactions @xmath76au(@xmath103n,@xmath104b ) at 39.4 mev@xmath58nucleon , @xmath76au(@xmath12n,@xmath11b ) at 68 mev@xmath58nucleon , @xmath76au(@xmath12n,@xmath11b ) at 109.6 mev@xmath58nucleon , @xmath105nb(@xmath12n,@xmath11b ) at 67.3 mev@xmath58nucleon , and @xmath106al(@xmath12n,@xmath11b ) at 68 mev@xmath58nucleon .
the red squares are the experimental data points and the blue dashed lines are the previous simulation results , both from ref .
the black band represents the range of the present simulation results within a @xmath79 distribution .
momentum is given relative to the peak of the yield distribution.,width=491 ] the predicted polarization trend as a function of outgoing fragment momentum for a three - nucleon removal case is shown in figure [ fig:55ni ] .
ni beam at 140 mev@xmath58nucleon on a a@xmath107 @xmath3be target and b@xmath107 @xmath108w target .
both targets will be used during the experiment , to provide better comparison to the simulation.,width=566 ] the polarized @xmath109ni nucleus from a @xmath110ni primary beam is proposed to determine the ground state magnetic moment measurement of @xmath109ni at nscl @xcite .
the expected rate of @xmath109ni is 500 pps , thus optimized polarization at peak yield is critical to realize a successful measurement .
the proposed measurement will complete the @xmath1 measurements of nuclei one nucleon removed from @xmath9ni .
the two @xmath111 nuclei , @xmath109co and @xmath4ni , both have known magnetic moment values that compare favorably with results of shell model calculations . on the other side of @xmath112 ,
the @xmath8 nucleus @xmath4cu has a @xmath1 value that deviates significantly from shell model expectations , suggesting a weakened core in @xmath9ni .
the measurement of @xmath1(@xmath109ni ) is key to understanding how the @xmath9ni core may be polarized when coupled to a neutron hole .
the simulation results for the one nucleon pickup processes are shown in fig .
[ fig:37kband ] and [ fig:34alband ] .
angular distributions , out - of - plane scattering , and @xmath0-ray deorientation were implemented in the nucleon pickup process as well . in the momentum distribution calculations discussed previously , souliotis _ et al . _
@xcite used the `` typical '' fermi momentum of 230 mev@xmath58c . in order for the simulation to encompass a range of targets ,
the fermi momentum was taken not as 230 mev@xmath58c , but was calculated based on data taken from moniz _ et al . _
the fermi momentum ranges from 170 mev@xmath58c for the lightest targets to 260 mev@xmath58c for heavier targets .
the results of the simulation for proton pickup are shown in fig .
[ fig:37kband ] , calculated for the one proton pickup reaction , @xmath3be(@xmath57ar,@xmath56k)x , first observed by groh _
et al . _
the momentum matching conditions @xcite for simple surface - to - surface pickup are best met for the two data points on the high momentum side of the yield distribution , where the simulation agrees with the data . on the low momentum side of the peak of the yield curve , the picked - up nucleon has a momentum less than the fermi momentum , and the momentum matching conditions for direct pickup are poorly satisfied .
more complex transfer mechanisms are therefore required to describe the polarization on the low momentum side @xcite .
relative to the primary beam momentum @xmath102 for the one - proton pickup reaction @xmath3be(@xmath57ar,@xmath56k ) ( 150 mev@xmath58nucleon ) .
the red squares are the experimental data points from ref . @xcite and the grey band represents the range of simulation results within a @xmath79 distribution.,width=302 ]
the simulation code was also used to model the neutron pickup data obtained in ref .
@xcite , as shown in fig .
[ fig:34alband ] .
again , no scaling factor was required to achieve a better quantitative agreement with experiment .
relative to the momentum at the peak of the yield distribution @xmath102 for the one - neutron pickup reaction @xmath3be(@xmath57s,@xmath113al ) ( 77.5 mev@xmath58nucleon ) .
the red squares are the experimental data points and the blue dashed line is the previous simulation result , both from ref .
the grey band represents the range of the present simulation results within a @xmath79 distribution.,width=302 ]
a statistical monte carlo code for nuclear spin polarization has been modified to include corrections to the kinematical model that aim to improve quantitative agreement with experiment .
the positions of removed nucleons are assigned in a monte carlo fashion , and projectiles are allowed to interact on either side of the target .
the real part of the optical potential , @xmath84 , is obtained with a folding model calculation and then used in the calculation of the mean deflection angle , @xmath35 .
fragments are permitted to scatter to a distribution of deflection angles rather than a single mean angle , and the out of reaction plane acceptance has been taken into account . the process of de - orientation due to @xmath0-ray relaxation is also included .
the angular distribution implementation reduces the calculated polarization magnitude by about 10% and the out - of - plane acceptance reduces the polarization magnitude by about 40% .
the calculated polarization magnitude is corrected by about 50% due to @xmath0-ray de - orientation .
these corrections account for the 0.25 scaling factor needed by both ref .
@xcite and ref .
@xcite in their simulations .
it should be noted that the corrections are approximations .
a primitive function is used to calculate the distribution of deflections angles , and no attempt was made to rigorously reproduce the scattering angles observed in fragmentation reactions .
the out - of - plane scattering correction is implemented as a constant multiplicative factor based on the expected acceptances of typical fragment separators .
the deorientation correction due to @xmath0-ray relaxation is calculated in a statistical manner considering average excitation energy and entry spin . even within these limitations ,
the magnitude with which the corrections are shown to improve the results indicate that these are important contributions to consider in the simulation of polarization in intermediate - energy reactions .
we acknowledge the assistance of k. asahi and w.d .
schmidt - ott for providing the initial monte carlo simulation codes for spin polarization .
we thank d.j .
morrissey and p.g .
hansen for helpful discussions on spin polarization .
the work was supported in part by the national science foundations grants phy-06 - 06007 and phy-99 - 83810 .
k. matsuta , m. fukuda , m. tanigaki , t. minamisono , y. nojiri , m. mihara , t. onishi , t. yamaguchi , a. harada , m. sasaki , t. miyake , s. fukuda , k. yoshida , a. ozawa , t. kobayashi , i. tanihata , j.r .
alonso , g.f .
krebs , and t.j.m .
symons , nucl .
phys . a**588 * * , 153c ( 1995 ) .
h. ogawa , k. asahi , k. sakai , t. suzuki , h. izumi , h. miyoshi , m. nagakura , k. yogo , a. goto , t. suga , t. honda , h. ueno , y.x .
watanabe , k. yoneda , a. yoshimi , n. fukuda , y. kobayashi , a. yoshida , t. kubo , m. ishihara , n. imai , n. aoi , w.d .
schmidt - ott , g. neyens , and s. teughels , phys .
c * 67 * 064308 ( 2003 ) .
k. asahi , m. ishihara , n. inabe , t. ichihara , t. kubo , m. adachi , h. takanashi , m. kouguchi , m. fukuda , d. mikolas , d.j .
morrissey , d. beaumel , t. shimoda , h. miyatake , and n. takahashi , phys .
* b251 * , 488 ( 1990 ) . h. okuno , k. asahi , h. sato , h. ueno , j. kura , m. adachi , t. nakamura , t. kubo , n. inabe , a. yoshida , t. ichihara , y. kobayashi , y. ohkubo , m. iwamoto , f. ambe , t. shimoda , h. miyatake , n. takahashi , j. nakamura , d. beaumel , d.j .
morrissey , w .- d .
schmidt - ott , and m. ishihara , phys .
b335 * , 29 ( 1994 ) .
k. turz , p. himpe , d.l .
balabanski , g. blier , d. borremans , j.m .
daugas , g. georgiev , f. de oliveira santos , s. mallion , i. matea , g. neyens , yu .
e. penionzhkevich , ch .
stodel , n. vermeulen , and d. yordanov , phys . rev .
c * 73 * , 044313 ( 2006 ) .
k. yoneda , h. sakurai , t. gomi , t. motobayashi , n. aoi , n. fukuda , u. futakami , z. gacsi , y. higurashi , n. imai , n. iwasa , h. iwasaki , t. kubo , m. kunibu , m. kurokawa , z. liu , t. minemura , a. saito , m. serata , s. shimoura , s. takeuchi , y. x. watanabe , k. yamada , y. yanagisawa , k. yogo , a. yoshida , and m. ishihara , phys . lett . *
b499 * , 233 ( 2001 ) .
r. grzywacz , r. braud , c. borcea , a. emsallem , m. glogowski , h. grawe , d. guillemaud - mueller , m. hjorth - jensen , m. houry , m. lewitowicz , a.c .
mueller , a. nowak , a. plochocki , m. pftzner , k. rykaczewski , m.g .
saint - laurent , j.e .
sauvestre , m. schaefer , o. sorlin , j. szerypo , w. trinder , s. viteritti , and j. winfield , phys .
lett . * 81 * , 766 ( 1998 ) . | intermediate - energy heavy - ion collisions can produce a spin polarization of the projectile - like species .
spin polarization has been observed for both nucleon removal and nucleon pickup processes .
qualitative agreement with measured spin polarization as a function of the momentum of the projectile - like fragment is found in a kinematical model that considers conservation of linear and angular momentum and assumes peripheral interactions between the fast projectile and target .
better quantitative agreement was reached by including more realistic angular distributions , de - orientation caused by @xmath0-ray emission , and by correcting for the out - of - plane acceptance .
the newly introduced corrections were found to apply to both nucleon removal and nucleon pickup processes . |
9 bauer e d , frederick n a , ho p - c , zapf v s , and maple m b 2002 _ phys . rev . _ b * 65 * 100506(r ) kotegawa h , yogi m , imamura y , kawasaki y , zheng g - q , kitaoka y , ohsaki s , sugawara h , aoki y and sato h 2003 _ phys .
lett . _ * 90 * 027001 aoki y , namiki t , ohsaki s , saha s r , sugawara h and sato h 2002 _ j. phys .
jpn . _ * 71 * 2098 vollmer v , faisst a , pfleiderer c , lhneysen h v , bauer e d , ho p - c and maple m b 2002 ( _ preprint _ cond - mat/02072225 ) izawa k , nakajima y , goryo j , matsuda y , osaki s , sugawara h , sato h , thalmeier p and maki k 2002 ( _ preprint _ cond - mat/0209553 ) izawa k , yamaguchi h , matsuda y , shishido h , settai r and nuki y 2001 _ phys .
_ * 87 * 057002 takegahara k , harima h and yanase a 2001 _ j. phys .
jpn . _ * 70 * 1190 harima h and takegahara k ( _ unpublished _ ) goto t ( _ private communications _ ) ozaki m and machida k 1989 _ phys .
_ b * 39 * 4145 sugawara h , osaki s , saha s r , aoki y , sato h , inada y , shishido h , settai r , nuki y , harima h and oikawa k 2002 _ phys .
rev . _ * 66 * 220504(r ) nakajima s 1973 _ prog .
phys . _ * 50 * 1101 miyake k , schmitt - rink s and varma c m 1986 _ phys . rev .
_ b * 34 * 6554 y. aoki _
et al _ 2002 to be published in _
c ( _ proceedings of lt23 _ ) cox d l 1987 _ phys .
_ * 59 * 1240 ; 1993 _ physica _ b * 186 - 188 * 312 leggett a j 1975 _ rev
phys . _ * 47 * 331 5e .
miyake k , matsuura t , jichu h and nagaoka y 1984 _ prog .
phys . _ * 72 * 1063 anderson p w and brinkman b f 1973 _ phys .
lett . _ * 30 * 1108 kuroda y 1975 _ prog .
phys . _ * 53 * 349 miyake k and kohno h ( _ unpublished _ ) ishikawa m 1977 _ prog .
phys . _ * 57 * 1836 ; ishikawa m , miyake k and usui t 1980 _ prog .
phys . _ * 63 * 1083 miyake k and usui t 1980 _ prog .
phys . _ * 63 * 711 kita t 1996 _ j. phys .
jpn . _ * 65 * 1355 ; _ ibid _ 664 ; 1998 _ j. phys .
jpn . _ * 67 * 216 see , for example , vollhardt d and wlfle p. 1990
_ the superfluid phases of helium 3 _ ( taylor & francis : london ) 7.3 and 9.3 ishida k ( _ private communications _ ) kusunose h , miyake k , shimizu y and sakai o 1996 _ phys . rev .
lett . _ * 76 * 271 yotsuhashi s , miyake k and kusunose h 2002 _ physica _ b * 312 - 313 * 100 tou h , kitaoka y , ishida k , asayama k , kimura n , nuki y , yamamoto e , haga y and maezawa k 1998 _ phys . rev .
* 80 * 3129 ; tou h ( _ private communications _ ) | it is shown that unconventional nature of superconducting state of pros@xmath0sb@xmath1 , a pr - based heavy electron compound with the filled - skutterudite structure , can be explained in a unified way by taking into account the structure of the crystalline - electric - field ( cef ) level , the shape of the fermi surface determined by the band structure calculation , and a picture of the quasiparticles in f@xmath2-configuration with magnetically singlet cef ground state .
possible types of pairing are narrowed down by consulting recent experimental results . in particular , the chiral p "-
wave states such as @xmath3 is favoured under the magnetic field due to the orbital zeeman effect , while the p "- wave states with two - fold symmetery such as @xmath4 can be stabilized by a feedback effect without the magnetic field .
it is also discussed that the double superconducting transition without the magnetic field is possible due to the spin - orbit coupling of the triplet " cooper pairs in the chiral state .
recently , the superconductivity has been found in heavy electron compound pros@xmath0sb@xmath1 with the crystal structure of the filled skutterudite @xcite . since the specific heat jump @xmath5 at the superconducting transition temperature @xmath6=1.8k is quite enhanced as @xmath7 mj / k@xmath8mole
, heavy quasiparticles are responsible for the cooper pair formation .
quite recently , a measurement of the longitudinal relaxation rate , @xmath9 , of nqr at sb site has been performed and very unusual temperature ( @xmath10 ) dependence was revealed both for @xmath11 and @xmath12 @xcite , while the normal state properties had been known to be also quite unconventional @xcite .
unconventional behaviors of @xmath9 are summarized as follows : 1 ) pseudo gap behavior is seen in @xmath13 at @xmath14 , in which the resistivity @xmath15 also shows a pseudo - gap behavior @xcite .
2 ) there is no trace of the coherence ( hebel - slichter ) peak around @xmath16 at all .
3 ) @xmath9 appears to exhibit an exponential @xmath10-dependence below @xmath17 , giving the superconducting gap @xmath18 in the low temperature limit as @xmath19 , although a possibility is not ruled out that the crossover to the @xmath20-dependence begins to be observed at around @xmath21 the lowest temperature covered by experiments . very recently , the anomaly of specific heat near @xmath6 has been observed , which suggests a double transition at @xmath22 and @xmath23 ( @xmath24 ) @xcite it also turned out very recently on the basis of measurements of the angular dependence of the thermal conductivity @xmath25 under the magnetic field @xmath26 @xcite that there exsist at least two different superconducting phases in the @xmath10-@xmath26 phase diagram . in the low - field phase , the 2-fold component of @xmath27 along the @xmath28-direction is observed as a function of the angle of the direction of @xmath26 around the @xmath28-axis , while the 4-fold one is observed in the high - field phase .
the phase boundary approaches the lower critical temperature @xmath23 as @xmath29 .
this is in marked contrast with the case of a heavy electron superconductor cecoin@xmath30 where the data can be interpreted by a simple d "- wave model @xcite .
these behaviors suggest that a novel type of heavy electron superconductivity is realized in pros@xmath0sb@xmath1 . the purpose of this letter is to present a scenario explaining such anomalous behaviors in a unified way on the basis of the crystalline - electric - field ( cef ) level , inferred from the experiment @xcite and theoretical study @xcite , a topology of the fermi surface ( fs ) offered from the band structure calculations @xcite , and a picture of the quasiparticles of @xmath31-based heavy electrons with a non - kramers ( non - magnetic ) doublet cef ground state . the cef level scheme proposed by bauer _
et al . _
@xcite in the point group @xmath32 is as follows @xcite : the lowest level is the non - kramers doublet @xmath33 ( @xmath34 in the representation of the point group @xmath35 ) @xmath36 and the first excited state is one of the triplet states @xmath37 or @xmath38 ( @xmath39 or @xmath40 for @xmath35 ) , wavefunction of which has the form @xmath41 where the coefficients @xmath42 s and @xmath43 s are not universal but depend on the details of the cef parameters @xcite . the lowest excitation energy of cef levels has been estimated as @xmath44k @xcite .
other excited cef levels have excitation energies higher than 100k so that their effects are negligible in the low temperature region of the order of @xmath6 .
however , the possibility of @xmath45-@xmath39 cef level scheme , with @xmath44k , can not be ruled out only from the analysis of the static susceptibility and the specific heat @xcite the ultrasonic measurements performed quite recently strongly suggest that the @xmath46 is the cef gorund state @xcite around @xmath4710k , these lowest excited cef levels give considerable contribution not only to the thermodynamic quantities , such as the specific heat @xmath48 and the magnetic susceptibility @xmath49 , but also to the nqr relaxation rate @xmath9 , since the spin - flip " process can occur among the states forming @xmath50 , e.g. , between @xmath51 and @xmath52 , giving the nqr relaxation .
it is noted that each cef level broadens due to the hybridization with the conduction electrons so that the energy conservation law is satisfied in the nqr or nmr relaxation process .
indeed , if we assume that the energy level of excited cef level is broadened such that its spectral weight is approximated by the lorentzian with the width @xmath53 , and that the processes across the cef ground states , ( [ gamma23 + ] ) and ( [ gamma23- ] ) , and @xmath39 s are neglected , the imaginary part of the spin - flip susceptibility @xmath54 is given simply , in the limit @xmath55 , as @xmath56 where const .
is given by a combinatin of the coefficients @xmath57 s and @xmath58 s in ( [ gamma40 ] ) , and the clebsch - gordon coefficeints .
therefore , the nqr / nmr relaxation rate @xmath59 due to the excited cef level is given as @xmath60 the width of cef level arises from the hybridization between f- and conduction electrons and is of the order of the width of the renormalized quasiparticle band .
in the present case , @xmath44k is comparable to the bandwidth of heavy electrons , so that @xmath53 is also expected to be highly renormalized by the correlation effect .
namely , if the temperature is decreased well below @xmath61=7k , the relaxation processes are gradually killed , leading to the pseudo - gap behavior ( [ rate ] ) at such temperature region .
we have , however , the usual relaxation process due to the quasiparticles of the fermi liquid in addition .
therefore , @xmath10-dependence of @xmath62 at @xmath11 will be rather complicated one , since both contributions , from @xmath63 cef level ( @xmath64 ) and the quasiparticles ( @xmath65 ) , to @xmath9 are decreasing in such @xmath10-region with different @xmath10-dependence in general .
in particular , one has to be careful in drawing the structure of the superconducting gap from the @xmath10-dependence of @xmath9 at @xmath66 . in this letter
, we discuss the nature of the gap structure specific to the present system .
first of all , it may be reasonable to assume that the strong on - site repulsion , the possible origin of the heavy electron state , can not be avoided in a manifold of the conventional s - wave pairing state .
this is also consistent with the absence of the coherence peak in @xmath9 @xcite although the possibility of the strong coupling effect is not completely ruled out .
another crucial aspect of pros@xmath0sb@xmath1 uncovered by the band structure calculation is that the fs of the heavy electron band is missing in the directions of [ 1,0,0 ] and [ 1,1,1 ] and their equivalents as shown in fig .
[ fig:1 ] @xcite .
there exists a small fs surrounding the @xmath67-point whose mass is not heavy @xcite and has been detected by the de haas - van alphen experiment @xcite . due to this porous structure of fs
, even the anisotropic pairing state can have finite gap over the fs .
however , even with such fs features being taken into account , the anisotropy of the gap due to such features of fs does not seem to fully explain the exponential - like @xmath10-dependence of @xmath9 .
+ + _ odd - parity pairing due to quadrupolar fluctuations _
we adopt here the odd - parity pairing to explain the unconventional nature of the superconducting state mentioned above .
there are at least three circumstantial evidences for favouring the odd - parity pairing .
first , the pairing interaction should be mainly mediated by the mode which gives rise to mass enhancement of quasiparticles , the quadrupolar fluctuations in the present case .
quadrupolar susceptibility @xmath68 is expected to be enhanced at large wave vector because the main fs has the nesting tendency as shown in fig .
[ fig:2 ] which can induce the attraction in both the d"- and p "- wave channel since the spin factor @xmath69 does not exist in contrast to the case of spin - fluctuation mechanism @xcite .
second , a scenario for the bouble transition is more easily constructed in the odd - parity pairing with degeneracy due to the time - reversal symmetry than the even - parity pairings .
third , the so - called maki parameter @xmath70 under the magnetic field exhibits no paramagnetic limitation @xcite .
as mentioned above , the pairing is also expected to be induced by exchanging the quadrupolar fluctuations of the non - kramers doublet @xmath33 .
the quadrupolar coupling between essentially localized 4f@xmath2-states and the quasiparticles containing considerable weight of the conduction electrons , arises through the hybridisation with the local symmetry of @xmath71 and @xmath72 in the cubic representaion of cef state of @xmath73-manifold , as discussed by cox in the context of the quadrupolar kondo effect @xcite .
the fact that the heavy quasiparticles contains the considerable weight of conduction electrons is a salient feature of f@xmath2-based heavy electron state , which is in marked contrast with the f@xmath74- or f@xmath75-based ones where the quasiparticles are dominated by f - electrons .
it is also consistent with the result of band structure calculation which shows that the f - component of the heaviest band at fs is only several % @xcite .
the propagator of the quadrupolar fluctuations @xmath76 may be given as @xmath77 where @xmath78 and @xmath79 denote the propagator for local fluctuations of quadrupolar moment , and the polarisation function of quasiparticles , respectively , and @xmath80 is the coupling constant among them .
as shown in fig .
[ fig:2 ] , the fs has a nesting tendency and @xmath79 is expected to have peaks at @xmath81 , and @xmath82 , and their equivalent positions .
it is noted that the fs is rather flat in the @xmath28-direction near the nesting position as can be seen in fig .
[ fig:1 ] .
then , the pairing interaction in the static approximation , @xmath83 , can be parameterised as @xmath84 + \gamma_{2}\cos{q_{x}\over 2}\cos{q_{y}\over 2 } \nonumber \\ & & \quad + ( { \hbox{cyclic permutations of $ q_{x}$ , $ q_{y}$ , and $ q_{z}$ } } ) , \label{gamma1}\end{aligned}\ ] ] where @xmath85 s are positive constants and @xmath86 is rather smaller than than @xmath45 .
the term of @xmath86 represents the effect that tendency of the nesting at @xmath82 is less than that at @xmath81 as seen in fig .
[ fig:2 ] . by putting @xmath87 , @xmath88
is represented near the peak as follows : @xmath89 + \cdots \label{gamma2}\ ] ] this gives the attractive interactions in the following channels : @xmath90 among these states , @xmath91 and its equivalents , @xmath92 and @xmath93 , will be the most favourable ones because they have maximum amplitude on the fs . indeed , other states have more nodes on the fs as seen in fig . [ fig:2 ] .
the simplest odd - parity states with equal - spin - pairing " ( esp ) allowed in cubic symmetry are given as follows : @xmath94{\rm i}(\sigma_{y}\sigma_{x } ) , \label{triplet1 } \\
\hat{\delta}_{k}&=&\delta[p_{x}(k)+{\rm i}p_{y}(k ) ] { \rm i}(\sigma_{y}\sigma_{x } ) , \label{triplet2}\end{aligned}\ ] ] here , @xmath95 is the @xmath96-th component of the pauli matrix , @xmath97 , and @xmath98 s are bases of irreducible representations with
p "-
symmetry : @xmath99 , etc .. these gaps vanish along the direction of [ 1,1,1 ] or [ 1,0,0 ] , and its equivalent direction .
however , since there exists no fs in those directions , @xmath9 exhibits an exponential @xmath10-dependence in the lowest temperature region in spite of the anisotropic gap .
it is noted that the fermi surface of light electrons , detected by the de haas - van alphen experiment @xcite , is closed surrounding the @xmath67-point and these gaps have nodes at points on this fs .
however , since @xmath9 is proportional to the square of the density of states at the fermi level , the effect of such light electrons should hardly be seen by the @xmath10-dependence of @xmath9 .
other possible states in the odd - parity manifold are @xmath100 such states are less favourable compared to the chiral states ( [ triplet1 ] ) and ( [ triplet2 ] ) in the so - called weak - coupling case where the feedback effect is not taken into account . this can be seen from the structure of the gl free energy .
for instance , in the case of odd - parity class of esp with p "- symmetry , it is given as follows @xcite : @xmath101^{2}+ 2[{\rm re}(\delta_{x}\delta_{y}^{*})]^{2}+ 2[{\rm re}(\delta_{x}\delta_{y}^{*})]^{2}\biggr\ } , \label{gl1}\end{aligned}\ ] ] where @xmath18 s are the coefficient of each ( normalized ) irreducible representations of the gap , @xmath102 is the strength of pairing interaction of
p "- symmetry , and @xmath103^{2}{\tanh(\xi_{k}/2t)\over 2\xi_{k } } , \label{gl2 } \\
\chi_{\rm diag}&\equiv & \sum_{k}[p_{x}(k)]^{4}\left(-{{\rm d}\over { \rm d}\xi_{k}^{2 } } { \tanh(\xi_{k}/2t)\over 2\xi_{k}}\right)>0 , \label{gl3 } \\
\chi_{\rm off}&\equiv & \sum_{k}[p_{x}(k)p_{y}(k)]^{2}\left(-{{\rm d}\over { \rm d}\xi_{k}^{2 } } { \tanh(\xi_{k}/2t)\over 2\xi_{k}}\right)>0 .
\label{gl4 } k\end{aligned}\ ] ] since @xmath104 due to the schwarz inequality , the gap ( [ triplet3 ] ) can not minimize @xmath105 in general . in case @xmath106 ,
the gap ( [ triplet1 ] ) minimizes @xmath105 , while the gap ( [ triplet2 ] ) minimizes @xmath105 in case @xmath107 .
the low - field phase , in the @xmath10-@xmath26 phase diagram @xcite , having 2-fold symmetry is consistent with the gap ( [ triplet3 ] ) , which can not be realized in the weak - coupling theory .
this is also the case in the singlet manifold .
the so - called bw - like state is known to be the most stable state in the weak - coupling approximation , and there is no reason in principle to rule out its possibility from the first .
however , bw - like state looks inconsistent with the thermal conductivity measurement under the magnetic field @xcite and other thermodynamic measurements .
+ + _ feedback effect _ in order that the gap with 2-fold symmetry such as ( [ triplet3 ] ) to be realised , we need a feedback effect . among them , the following mechanism may be promising .
the polarisation function @xmath79 appearing ( [ chiquadrupolar ] ) in the superconducting state is given as @xmath108 where @xmath109 .
if the nesting were perfect at @xmath81 , and @xmath82 , and their equivalent positions , the following relations would hold , @xmath110 , @xmath111 , and @xmath112 , for @xmath113 near the fs . then
, the expression , ( [ pi ] ) , would be reduced to @xmath114 then , the polarisation mediating the pairing interaction depends on the type of pairing itself .
indeed , the pairing ( [ triplet3 ] ) is expected to give larger ( [ pi2 ] ) than the pairing ( [ triplet2 ] ) , because the gap function of ( [ triplet3 ] ) @xmath115 vanishes on the planes , @xmath116 , and @xmath117 , which pass through near the fs , while the gap function ( [ triplet2 ] ) @xmath118 \label{pi4}\ ] ] vanishes only on the lines , @xmath119 , @xmath120 , @xmath121 , etc .
, which are located away from the fs .
although the explicit band structure calculation is hard in practice for the moment , the tendency mentioned above is expected to remain valid .
therefore , the state ( [ triplet3 ] ) may be stabilized against ( [ triplet2 ] ) by the feedback effect . in the bw - like state , @xmath122 , ( [ pi2 ] ) ,
is suppressed more severely than in the state ( [ triplet2 ] ) , and is destabilized against ( [ triplet3 ] ) .
this kind of feedback effect is an analogue of that due to the ferromagnetic spin - fluctuation mechanism discussed in superfluid @xmath123he @xcite , in which the spin - fluctuation spectrum depends on the gap structure of the triplet states .
+ + _ double transition due to spin - orbit coupling _
the spin - orbit interaction @xmath124 due to the mutual coulomb interaction between electrons and relative motion is given by @xmath125\right ] , \label{spinorbit1}\ ] ] where @xmath126 is the bohr magneton , @xmath127 electron mass , @xmath128 the band mass , and @xmath129 is defined as @xmath130 , @xmath131 being the effective magnetic moment @xmath132 .
the appearance of the factor @xmath133 in ( [ spinorbit1 ] ) can follows from the ward - pitaevskii - identity @xcite . by the procedure similar to that described in ref .
@xcite for the dipole interaction , the interaction ( [ spinorbit1 ] ) leads to the spin - orbit free energy @xmath134 for cooper pairs which is spin triplet and chiral , such as ( [ triplet2 ] ) , with the pair angular momentum @xmath135 as follows@xcite : @xmath136 where @xmath137 where @xmath138 is the strength of the dipole coupling in the esp "- superconducting state , and we have used @xmath139 for quasiparticles consisting of @xmath72 f@xmath74-cef state , and @xmath140 for @xmath71 . the free energy due to the dipole - dipole interaction is given as @xcite @xmath141 therefore the spin - orbit interaction , in the non - unitary state with @xmath142 , dominates the dipole - dipole interaction , in the unitary state with @xmath143 , because @xmath144 far exceeds @xmath138 in @xmath72-band considering ( [ spinorbit3 ] ) and @xmath145 .
following the calculation in the spherical model @xcite , @xmath138 is given by @xmath146^{2 } , \label{spinorbit4}\ ] ] where @xmath147 is the density of states ( dos ) of the quasiparticles , and @xmath148 is the condensation free energy @xmath149 where @xmath25 is the average of square of the magnitude of normalized gap function over the fs .
the second factor in ( [ spinorbit4 ] ) is estimated as @xmath150^{2 } \simeq \left({\mu_{\rm eff}\over\mu_{\rm b}}\right)^{2 } \times1.4\times 10^{-3 } , \label{spinorbit5}\ ] ] where we have assumed that the renormalized fermi energy is @xmath151 k , the number density of quasiparticles @xmath152 , @xmath153 being the distance between two nearest pr ions . therefore , the spin - orbit coupling @xmath144 , ( [ spinorbit3 ] ) , is estimated as @xmath154 the free - energy difference between ( [ triplet3 ] ) and ( [ triplet2 ] ) , its non - unitary version with @xmath155 , is of the order of 10% of @xmath148 in general , and @xmath145 according to the band structure calculation @xcite .
therefore , if the stable state is ( [ triplet3 ] ) due to the feedback effect as discused above , there occurs a double transition with splitting of the transition temperature being ( @xmath156 , because the state ( [ triplet2 ] ) of non - unitary version is stabilized , due to the spin - orbit interaction ( [ spinorbit2 ] ) , against ( [ triplet3 ] ) which is real state and has no spin - orbit coupling such as ( [ spinorbit2 ] ) .
this width of splitting of double transition is consistent with the experimental observations @xcite . especially , the results by aoki _ et al _
@xcite , suggesting the double transition remains rather robustly under the magnetic field , can be explained by the present mechanism .
the chiral state ( [ triplet1 ] ) has also intrinsic magnetic moment along ( 1,1,1 ) direction , so that it can also give rise to the double transition as above .
however , this state gives the angualr dependence of the thermal conductivity opposite to the observation in the high - field phase @xcite although the 4-fold behaviour is expected .
+ + _ two phases in t - h phase diagram _
finally , the multiphase diagram in @xmath10-@xmath26 plane determined by the thermal conductivity measurements under the magnetic field @xcite may be understood as follows : a crucial point is that the triplet state ( [ triplet2 ] ) has the intrinsic magnetic moment @xmath157 associated with the intrinsic angular momentum @xmath158 as @xmath159 , where @xmath160 is the effective mass of heavy quasiparticles , and @xmath158 is given as @xmath161 where @xmath162 is the order of the superfluid electron density @xmath163 @xcite , while the lively disputes were performed concerning the size of @xmath162 , whether @xmath164 or @xmath165 with @xmath166 1 or 2 , about a quarter of century ago in the context of superfluid @xmath75he @xcite .
very recently , the reality of this intrinsic magnetic moment has caused a renewed interest in the magnetic property of the chiral superconducting state of sr@xmath167ruo@xmath0 @xcite . at low enough temperature @xmath168 ,
the intrinsic magnetic moment is @xmath169 .
therefore , the state ( [ triplet2 ] ) is stabilized under the magnetic field @xmath26 over the state ( [ triplet3 ] ) , which has no intrinsic magnetic moment .
the transition between the two states occurs when @xmath170 where the left - hand side represents the difference of the condensation energy between the two superconducting phases at @xmath168 , and the right - hand side the energy gain in the chiral state ( [ triplet2 ] ) with the intrinsic angular mementum in the magnetic field @xmath26 .
we have included in ( [ boundary ] ) the so - called the demagnetisation factor @xmath171 which depends on the sample shape . in ( [ boundary ] ) , the energy due to the magnetic field arising from the intrinsic magnetisation itself is neglected because it is much smaller than the external field @xmath172 in question .
the magnetic field giving the phase boundary beween low- and high - field phases , determined by the thermal conductivity , roughly agrees with the present estimation is in the same order as given by ( [ boundary ] ) , because @xmath173 , @xmath174 , @xmath175 , and @xmath176 .
a crucial prediction of the present scenario is that the phase boundary in the @xmath10-@xmath26 plane is dependent on the sample shape through the demagnetisation factor @xmath171 .
the angular dependence of the thermal conductivity @xmath27 reported in ref .
@xcite may also be explained by the present scenario .
if the state ( [ triplet3 ] ) , @xmath177 , is realised in the low - field phase due to the boundary effect , which works to align the extention of pair wavefunction , @xmath27 takes maximum ( minimum ) , when the magnetic field @xmath178 is @xmath179 ( @xmath180 ) , in consistent with ref .
@xcite .
this can be understood by applying the argumet similar to ref .
@xcite .
if a type of the state ( [ triplet2 ] ) is realized in the high - field phase , the free energy takes minimum when the quantisation axis of intrinsic angular momentum is parallel to @xmath178 for which @xmath27 is minimum @xcite .
therefore , when the direction of @xmath178 is rotated by angle @xmath181 from the @xmath182-axis in the plane perpendicular to @xmath28-axis , @xmath27 increases and reaches maximum at @xmath183 above which the stable quantisation axis changes from @xmath182- to @xmath184-axis , and then @xmath27 decreases up to @xmath185 where the stable configuration is reached again .
namely , the 2-fold dependence of @xmath186 taking maximum at @xmath187 can be possible in the low - field phase , and the 4-fold one taking minimum at @xmath187 and @xmath185 in the high - field phase @xcite .
+ in conclusion , it is remarked that the magnetic susceptibility , both longitudinal and transverse , can be enhanced by electron correlations even if the mass enhancement arises from the degeneracy due to the non - kramers doublet , i.e. , electric quadrupolar moment , provided that there exists a perturbation which breaks the particle - hole symmetry , such as the repulsion among conduction electrons , as shown by the numerical renormalization group calculations for the impurity model @xcite .
this is in marked contrast with the case of heavy electrons based on f@xmath2-configuration with the singlet cef ground state @xcite , where the static susceptibility along the easy axis due to quasiparticles is not enhanced while the nmr / nqr relaxation rates given by the dynamical transverse susceptibility is enhanced in proportion to a square of the mass - enhancement factor as observed in upt@xmath188 @xcite .
+ we have benefited from informative conversations with y. aoki , t. goto , k. izawa , y. kitaoka , h. kotegawa , y. matsuda , h. sato , h. sugawara .
this work was supported by the grant - in - aid for coe research program ( no .
10ce2004 ) by the ministry of education , culture , sports , science and technology .
sb@xmath1 relevant to the heavy electrons given by band structure calculation @xcite .
fs is missing in the direction of [ 1,1,1 ] or [ 1,0,0 ] , and its equivalent ones . ]
cut by the @xmath189-@xmath190 plane @xmath191 . the nesting tendency in the heavy electron band at @xmath192 and @xmath193 as shown by arrows , and their equivalent positions , remains in the direction rather rubustly for finite @xmath194 . ] |
to define quasilocal energy in general relativity , one can begin with a suitable action functional for the time history @xmath3 of a spatially bounded system @xmath1 . here
`` suitable '' means that in the associated variational principle the induced metric on the time history @xmath4 of the system boundary @xmath5 is fixed . in particular
, this means that the lapse of proper time between the boundaries of the initial and final states of the system @xmath1 must be fixed as boundary data .
the quasilocal energy ( qle ) is then defined as minus the rate of change of the classical action ( or hamilton - jacobi principal function ) corresponding to a unit increase in proper time.@xcite so defined , the qle is a functional on the gravitational phase space of @xmath1 , and is the value of the gravitational hamiltonian corresponding to unit lapse function and zero shift vector on the system boundary @xmath6 .
although other definitions of quasilocal energy have been proposed ( see , for example , the references listed in @xcite ) , the qle considered here has the key property , which we consider crucial , that it plays the role of internal energy in the thermodynamical description of coupled gravitational and matter fields.@xcite in this paper we define the energy of a perfectly isolated system at a given retarded time as the suitable limit of the quasilocal energy @xmath0 for the partial system enclosed within a finite topologically spherical boundary . for our choice of asymptotic reference frame
the energy that we compute equals what is usually called the bondi - sachs mass.@xcite as we shall see , our asymptotic reference frame defines precisely that infinitesimal generator of the bondi - metzner - sachs ( bms ) group corresponding to a pure time translation.@xcite we also show that in the same null limit the lapse - arbitrary , shift - zero hamiltonian boundary value defines a physically meaningful element in the space dual to supertranslations
. this dual space element , it turns out , coincides with the supermomentum " discussed by geroch.@xcite our results are then specialized to an expression for the full bondi - sachs four - momentum in terms of hamiltonian values .
it is already known that when @xmath6 is the two - sphere at spacelike infinity , the quasilocal and arnowitt - deser - misner@xcite notions of energy - momentum agree.@xcite our results therefore indicate that the quasilocal formalism provides a unified hamiltonian framework for describing the standard notions of gravitational energy - momentum in asymptopia . before turning to the technical details ,
let us first present a short overview of our approach . consider a spacetime @xmath3 which is asymptotically flat at future null infinity @xmath7 and a system @xmath8 of bondi coordinates thereon.@xcite the retarded time @xmath9 labels a one - parameter family of outgoing null hypersurfaces @xmath10 .
the coordinate @xmath11 is a luminosity parameter ( areal radius ) along the outgoing null - geodesic generators of the hypersurfaces @xmath10 .
the bondi coordinate system also defines a two - parameter family of topologically spherical two - surfaces @xmath12 .
it suits our purposes to consider only a single null hypersurface of the family @xmath10 , say @xmath13 , the one determined by setting @xmath9 equal to an _ arbitrary _ constant @xmath14 .
the collection @xmath15 of two - surfaces foliates @xmath13 , and in the @xmath16 limit these two - surfaces converge on an infinite - radius _ round _ sphere @xmath17 . to streamline the presentation
, we refer to our generic null hypersurface simply as @xmath18 ; and we use the plain letter @xmath6 to denote both the @xmath18-foliating collection @xmath15 and a single generic two - surface of this collection .
now , should we desire a more general @xmath19-foliating collection of two - surfaces , we could , of course , introduce a new radial coordinate @xmath20 . for a fixed retarded time
@xmath21 the new two - surfaces would then arise as level surfaces of constant @xmath20 .
however , we shall not consider such a new radial coordinate , because the new two - surfaces would not necessarily converge towards a round sphere in the asymptotic limit . at any rate
, we could handle such an additional kinematical freedom , were it present , by assuming that along each outgoing null ray @xmath20 approached @xmath11 at a sufficiently fast rate in the asymptotic limit .
our first goal is to compute the qle within a two - surface @xmath6 in the limit as @xmath6 approaches a spherical cut of @xmath7 along the null surface @xmath18 , and to show that this result coincides with the bondi - sachs mass : @xmath22 here @xmath23 is the quasilocal energy surface density with @xmath24 ( in geometrical units ) and @xmath25 is the determinant of the induced metric on @xmath6 .
recall that @xmath26 denotes the mean curvature of @xmath6 as embedded in some _ spacelike _ spanning three - surface @xmath1 .
since both @xmath6 and @xmath1 are embedded in the _ physical _ spacetime @xmath3 , we sometimes use the notation @xmath27 .
also recall that @xmath28 denotes the mean curvature of a surface which is isometric to @xmath6 but which is embedded in a three - dimensional _ reference _ space different than @xmath1 . here
we choose the reference space to be flat euclidean space @xmath29 , i. e. we assign a flat three - slice of minkowski spacetime the zero value of energy.@xcite although a definition of the zero - energy reference in terms of flat space is neither always essential nor appropriate@xcite , it is the correct choice for the analysis of this paper . in order to define @xmath26 , we must select such a three - surface spanning @xmath6 for each @xmath11 value .
( for a single @xmath6 many different spanning three - surfaces will determine the same @xmath26 .
in fact , @xmath26 is determined solely by @xmath6 and a timelike unit vector field @xmath30 on @xmath6 , which can be considered as the unit normal of a slice @xmath1 .
thus , the continuation of @xmath1 away from @xmath6 is not needed ; moreover , such a continuation of @xmath1 might not be defined throughout the interior of @xmath3 . therefore , though we speak of choosing a @xmath1 three - surface to span @xmath6 for each @xmath11 value , we are really fixing only a timelike unit normal vector field at @xmath6 . ) for generality , we leave the choice of spanning three - surface @xmath1 essentially arbitrary at each @xmath11 value , but we do enforce a definite choice asymptotically .
heuristically , as @xmath31 the @xmath1 three - surface spanning @xmath6 approaches an asymptotic three - surface @xmath32 which spans a round infinite - radius spherical cut of @xmath7 ( see the figure ) .
our construction is , as expected , sensitive to the choice of asymptotic three - surface @xmath32 .
said another way , the qle depends on the fleet of observers at @xmath6 whose four - velocities are orthogonal to the spanning three - surface at @xmath6 .
therefore , one expects _ a priori _ the expression on the right - hand side of ( [ limit ] ) to depend on the choice of asymptotic fleet associated with the two - sphere at @xmath7 .
the asymptotic fleet we choose corresponds to a pure bms time translation : each member of the asymptotic fleet rides along @xmath33 . note that
, although @xmath33 is everywhere timelike in @xmath3 ( at least in the relevant exterior regions ) , the _ extension _ of @xmath33 to @xmath7 in a conformal completion @xmath34 of the physical spacetime @xmath3 is in fact a null vector which lies in @xmath7 .
( while we occasionally find it clarifying to make reference to the concept of a conformal completion , we do not explicitly use conformal completions in this paper . )
therefore , heuristically , one should envision @xmath32 as a spacelike slice which becomes null asymptotically ( see the figure ) .
this paper is organized as follows . in a preliminary section
we write down the familiar bondi - sachs form@xcite of the spacetime metric as well as asymptotic expansions for the associated metric coefficients .
we also introduce on @xmath35 two future - pointing null vector fields @xmath36 and @xmath37 ( do not confuse @xmath36 with the mean curvature @xmath26 ) .
both vector fields point everywhere normal to our collection @xmath12 of two - surfaces , and @xmath38 .
next , we construct on @xmath19 a timelike vector field @xmath39 ( equality restricted to @xmath19 ) , which in our analysis will _ define _ for each @xmath6 along @xmath18 a spacelike spanning three - surface @xmath1 . in @xmath40ii
we use the @xmath1 three - surfaces determined by @xmath41 to define an unreferenced energy surface density @xmath27 for each @xmath6 slice of @xmath18 and then examine the asymptotic limit of @xmath42 . in @xmath40iii we consider the asymptotic expression for the flat - space reference density @xmath43 , but give the derivation of this expression in the appendix .
next , we assemble the results of the previous two sections and prove the main claim ( [ limit ] ) . in @xmath40iv
we examine the `` smeared energy surface density , '' which is the hamiltonian value corresponding to an arbitrary supertranslation .
we then specialize our result for the smeared energy surface density to express the full bondi - sachs four - momentum in terms of hamiltonian values .
in @xmath40v we examine the smeared energy surface density _ via _ the spin - coefficient formalism , and show that it equals the `` supermomentum '' of geroch@xcite as written by dray and streubel.@xcite the appendix is devoted to a detailed analysis of the reference term .
in terms of a bondi coordinate system the metric of our asymptotically flat spacetime @xmath3 takes the standard form@xcite @xmath44 where @xmath45 are @xmath6 indices running over @xmath46 .
we assume the following expansions for the various metric coefficients above : remainder terms , our expansions in the radial coordinate @xmath11 coincide with those given by sachs ; however , we do not assume that the @xmath47 remainder terms are necessarily expandable in powers of inverse @xmath11 , an assumption which would be tantamount to what sachs calls the `` outgoing radiation condition.''@xcite recently , chruciel _ et al .
_ have shown that `` polyhomogeneous '' expansions in terms of @xmath48 also provide a consistent framework for solving the characteristic initial value problem of the bondi - sachs type.@xcite they argue that the so - called outgoing radiation condition is overly restrictive . ]
@xmath49r + \delta_{\sigma_{ab}}{\ , } .
\eqnum{\ref{sachs}d } \addtocounter{equation}{1}\end{aligned}\ ] ] here @xmath50 and @xmath51 are respectively the real and imaginary parts of the asymptotic shear @xmath52 , @xmath53 is the all - important _ mass aspect _ , @xmath54 is the metric of a unit - radius round sphere , and commas denote partial differentiation . in the appendix
we examine the form of the two metric @xmath55 in more detail .
remainder terms , denoted by the @xmath47 symbol , always fall off faster ( or have slower growth , as the case may be ) than the terms which precede them .
for instance , @xmath56 denotes a term which falls off _ faster _ than @xmath57 .
introduce the future - directed null covector field @xmath58 , where the scalar function @xmath59 is a point - dependent boost parameter .
the null covector @xmath60 is orthogonal to the spheres @xmath12 , and the function @xmath61 gives us complete freedom in choosing the extent of @xmath60 at each point of any @xmath6 two - surface .
we shall find it necessary later to assume that @xmath61 falls off _ faster _ than @xmath62 on every outgoing ray .
also define another future - directed null vector field @xmath37 which is orthogonal to the @xmath12 and normalized so that @xmath38 . as one - forms these null normals
are @xmath63 while as vector fields they are @xmath64 now define @xmath65 and @xmath66 along @xmath19 as the timelike and spacelike unit normals of the @xmath6 two - surfaces . for each slice @xmath6 of the null hypersurface @xmath19 , the normals @xmath30 and
@xmath67 determine a spanning spacelike three - surface @xmath1 . as mentioned previously
, the three - surface @xmath1 is not unique and , moreover , need not be defined throughout @xmath3 .
indeed , there is no guarantee that @xmath30 as defined is even surface - forming .
( that is , in general @xmath68 does not satisfy the frbinius condition @xmath69}=0 $ ] . ) nevertheless , our construction provides us with what we need : a unit timelike vector @xmath41 orthogonal to @xmath6 .
we can therefore obtain an unreferenced energy surface density @xmath42 which is the same for any slice or partial slice @xmath1 that contains @xmath6 and has timelike unit normal which agrees with @xmath30 at @xmath6 .
our construction implies @xmath70 on each ray as @xmath31 .
now , the standard realization of the bms - group lie algebra ( as vector fields on future null infinity ) identifies the _ extension _ of @xmath33 to @xmath7 ( in a conformal completion @xmath34 of @xmath3 ) with a pure time translation.@xcite therefore , asymptotically , our fiducial surface @xmath32 determines precisely the pure time - translation generator of the bms group .
we do _ not _ claim that @xmath41 generates an `` infinitesimal asymptotic symmetry transformation '' in the sense of sachs@xcite , _
i. e. _ that the various coefficients associated with the transformed metric @xmath71 satisfy the fall - off conditions ( [ sachs ] ) ; however , this is unimportant for our construction .
we now turn to the task of calculating an expression for the unreferenced quasilocal energy surface density @xmath27 .
our starting point is the definition @xmath72 , where the two - metric @xmath73 serves as the projection operator into @xmath6 .
we find it convenient to write does not depend on how @xmath74 is extended off @xmath6 . ]
@xmath75 , where in the standard notation of the spin - coefficient formalism@xcite @xmath76 and @xmath77 are , respectively , the expansions associated with the inward null normal and outward null normal to @xmath6 .
these are given by the formulae @xmath78 as a technical tool , it proves convenient to introduce fiducial vector fields @xmath79 and @xmath80 determined from ( [ vectors ] ) by setting @xmath81 on @xmath18 . from the middle expressions above , it is obvious that @xmath82 and @xmath83 , where the easier - to - calculate expressions @xmath84 and @xmath85 are built exactly as in ( [ spin ] ) but with the fiducial null normals @xmath79 and @xmath80 .
therefore , we may assume that @xmath81 while calculating the spin coefficients in ( [ spin ] ) and then simply multiply the @xmath81 results by the appropriate factor to get the correct general expressions .
let us sketch the calculation .
first , from ( [ bassae]a ) with @xmath81 note that @xmath86 , because @xmath87 is a gradient .
next , using both expressions ( [ bassae ] ) with @xmath81 , one can work the second term inside the parenthesis of ( [ spin]a ) into the form @xmath88 .
finally , one writes the covariant - divergence terms as ordinary divergences ; for example , @xmath89 , where the square root of ( minus ) the determinant of the spacetime metric is @xmath90 . following these steps and multiplying by the appropriate boost factors at the end of the calculation , one finds @xmath91 here the over - dot denotes partial differentiation by @xmath33 , the prime denotes partial differentiation by @xmath92 , and @xmath93 denotes the @xmath6 covariant derivative . since @xmath11 is an areal radius , we may take @xmath94 [ see the form of the @xmath6 metric given in ( [ bmetric ] ) ] .
therefore , we obtain the compact expressions @xmath95 adding twice ( [ mu2]a ) to ( [ mu2]b ) , we arrive at our desired expression @xmath96 which has the asymptotic form @xmath97 note that we have chosen not to expand the @xmath98 pure divergence term @xmath99 .
our assumption about the fall - off of @xmath61 ensures that a term @xmath100 which appears in the asymptotic expression for @xmath26 can be swept into @xmath101 .
write the total quasilocal energy as @xmath102 , with the total _ unreferenced _
quasilocal energy @xmath103 taken as @xmath104 plugging the expansion ( [ kexpansion ] ) into the above expression , using @xmath105 for our choice of coordinates , and integrating term - by - term , one finds @xmath106 here the bondi - sachs mass associated with the @xmath21 cut of @xmath7 is the two - surface average of the mass aspect evaluated at @xmath21,@xcite @xmath107 we use the notation @xmath108 to denote proper integration over the unit sphere ( which is identified with a spherical cut of @xmath7 ) . in passing from ( [ unreferencede1 ] ) to ( [ unreferencede2 ] ) ,
we have made an appeal to stokes theorem to show that the `` dangerous '' @xmath109 term that arises from proper integration over the pure - divergence term @xmath110 in ( [ kexpansion ] ) does indeed vanish .
hence , this term does not contribute to the bondi - sachs mass and does not spoil the result ( [ unreferencede2 ] ) .
the reference point contribution to the energy is @xmath111 where the asymptotic expression for @xmath28 must be determined from the specific asymptotic form ( [ bmetric ] ) of the sachs two - metric .
we present this calculation in the appendix .
the result is @xmath112 note the absence of an @xmath109 term in @xmath113 .
the result ( 3.5 ) has just the right form , in that it removes the part of @xmath103 which becomes singular as @xmath31 but does not itself contribute to the mass .
therefore the total quasilocal energy for large @xmath11 is @xmath114 this is the energy of the gravitational and matter fields associated with the spacelike three - surface @xmath1 which spans a @xmath6 slice of @xmath18 and which tends toward @xmath32 .
our main claim ( [ limit ] ) follows immediately from ( 3.6 ) .
consider the expression @xmath115 for the on - shell value of the gravitational hamiltonian appropriate for a spatially bounded three - manifold @xmath1 , subject to the choice of a vanishing shift vector at the boundary @xmath116 : @xmath117 we refer to @xmath118 as the smeared energy surface density .
addition of this boundary term to the smeared hamiltonian constraint ensures that as a whole the sum is functionally differentiable.@xcite in this section we consider the @xmath16 limit of @xmath118 along the null hypersurface @xmath18 in exactly the same fashion as we considered the limit ( 1.1 ) of the quasilocal energy previously . before evaluating @xmath119 ,
let us discuss its physical significance .
consider a particular spherical cut @xmath17 of @xmath7 .
a general bms supertranslation pushes @xmath17 forward in retarded time @xmath9 in a general angle - dependent fashion . as is well - known
, the infinitesimal generator corresponding to such a supertranslation has the form @xmath120 , where @xmath121 is any twice differentiable function of the angular coordinates.@xcite as we have seen , @xmath33 is heuristically the hypersurface normal @xmath41 at @xmath17 of an asymptotic spanning three - surface @xmath32 . in other words , each member of the fleet of observers at @xmath17 rides along @xmath33 .
therefore , again heuristically , the on - shell value of the hamiltonian generator of a general bms supertranslation is @xmath122 this symbolic expression coincides with the @xmath16 limit of the smeared energy surface density ( 4.1 ) , where we set @xmath123 ( suitable fall - off behavior for @xmath124 is assumed ) .
thus , @xmath119 defines a physically meaningful element in the dual space of general supertranslations . in this respect
it is like the supermomentum " of geroch.@xcite in the next section we show explicitly that , in fact , @xmath125 is precisely geroch s supermomentum . "
note , however , that it might be better to call such an expression the `` superenergy , '' as it arises entirely from the `` energy sector '' of the hamiltonian s boundary term ( that is , the sector with vanishing shift vector ) but also incorporates the `` many - fingered '' nature of time ( that is , an arbitrary lapse function ) .
let us now evaluate the @xmath16 limit of the smeared energy surface density @xmath118 . as we have stated , there is no @xmath109 contribution to @xmath113 .
the absence of this contribution stems from the fact that the two - sphere average of the coefficient @xmath126 of the @xmath98 piece of the reference term @xmath28 vanishes .
as spelled out in the appendix , this fact follows directly from an equation governing the required isometric embedding of @xmath6 into euclidean three - space .
moreover , as seen in section 2 , the coefficient @xmath127 of the @xmath98 piece of the physical @xmath26 is not solely twice the mass aspect but also contains a unit - sphere divergence term .
now , in the present case @xmath23 is smeared against a function @xmath124 , so one might worry that the limit is spoiled in some way by the presence of the smearing function . however , as we now show , _ for solutions of the field equations _ , the _ unintegrated _ expression @xmath128 is precisely the mass aspect of the system in the @xmath31 limit .
this striking result rests on an exact cancellation between @xmath126 and the aforementioned unit - sphere divergence part of @xmath129 . with the machinery set up in the previous sections and the appendix
[ see in particular equations ( [ kexpansion ] ) and ( a2 ) ] , we find the following limit : @xmath130 \ , .
\label{limk - k0}\ ] ] here we set @xmath24 ( in geometrical units ) and use @xmath131 to denote the coefficient of the @xmath132 piece of the @xmath6 ricci scalar . also , the coefficients @xmath133 of the leading @xmath98 pieces of @xmath134 are listed in ( 1.2c ) .
inspection of ( [ sachs]c , d ) shows that @xmath135 is expressed in terms of the same functions , @xmath136 and @xmath137 , that appear in the @xmath57 piece of @xmath138 .
furthermore , a short calculation with the @xmath6 metric shows that for these solutions @xmath139 may be expressed in terms of @xmath133 as follows : @xmath140 therefore , the term in ( [ limk - k0 ] ) which is enclosed by square brackets vanishes , and we obtain @xmath141 for the desired limit .
this result shows that the @xmath16 limit of the smeared energy surface density equals the smeared mass aspect .
coupled with the findings of the next section , it follows that geroch s supermomentum " is just the smeared mass aspect .
this simple result does not appear to be widely known . finally , recall that the bondi - sachs four - momentum components @xmath142 correspond asymptotically to a pure translation . in terms of the smeared energy surface density ,
one obtains a pure translation for a judicious choice of lapse function on @xmath17 ; namely , @xmath143 , where the @xmath144 are constants and @xcite @xmath145 therefore , we write @xmath146 for the appropriate limiting value of @xmath124 , and thereby obtain the bondi - sachs four - momentum as a hamiltonian value .
in this section we show that the null limit of the smeared hamiltonian boundary value , eq .
( 4.1 ) , is the supermomentum " of geroch.@xcite to be precise , we show that in the null limit @xmath115 equals geroch s supermomentum " as written by dray and streubel.@xcite the spin - coefficient formalism is required for this analysis.v we deal exclusively with smooth expansions in inverse powers of an _ affine _ radius , as we know of no work examining the standard spin coefficient approach to null infinity within a more general framework such as the polyhomogeneous one .
the expansions we borrow from @xcite are valid for einstein - maxwell theory . ]
apart from a few minor notational changes we adopt the conventions of dougan.@xcite geometrically , the scenario is nearly the same as the one described in the previous sections . however , we now work with a slightly different type of bondi coordinates .
namely , @xmath147 , where @xmath148 is an _ affine _ parameter along the null - geodesic generators of @xmath19 and @xmath149 is the stereographic coordinate .
dougan picks and @xmath150 respectively correspond to @xmath151 and @xmath152 in @xcite , where @xmath153 is dougan s retarded time .
the minus sign difference between our definition for @xmath60 and dougan s definition for @xmath154 stems from a difference in metric - signature conventions [ ours is @xmath155 .
the convention for metric signature does not affect the spin coefficients ( [ spinexpansions ] ) . ] @xmath156 as the first leg of a null tetrad , which is the same normal as given in ( [ bassae]a ) if @xmath81 . for convenience ,
in this section we ignore the kinematical freedom associated with the @xmath61 parameter , setting it to zero throughout .
as before , the vector field @xmath157 ( equality restricted to @xmath19 ) defines a three - surface @xmath1 spanning each @xmath6 slice of @xmath19 .
it follows that @xmath158 as @xmath159 , and hence our asymptotic slice @xmath32 again defines a pure bms time translation .
let us first collect the essential background results from @xcite which we will need .
first , the required spin coefficients have the following asymptotic expansions : as both the stem letter for the @xmath6 two - metric and as the spin coefficient known as the shear .
we have used @xmath25 twice in order to stick with the conventions of our references as much as possible . in all but equation ( [ spinexpansions ] ) , where @xmath25 has the spin - coefficient meaning
, it carries a `` @xmath160 '' superscript denoting the asymptotic piece . ] @xmath161r^{-2 } + o(r^{-3 } ) \nonumber \\
\sigma & = & \sigma^{0 } r^{-2 } + o(r^{-4 } ) \label{spinexpansions } { \ , } , \end{aligned}\ ] ] where @xmath162 is dougan s @xmath163 , the term @xmath164 is a certain asymptotic component of the weyl tensor , and @xmath165 is the asymptotic piece of the shear .
like before , an over - dot denotes differentiation by @xmath33 . as fully described in @xcite
, @xmath166 is the standard differential operator from the compacted spin - coefficient formalism , here defined on the _ unit _ sphere .
the expansion for the corresponding full operator in spacetime is @xmath167 + o(r^{-3 } ) { \ , } , \ ] ] where @xmath168 sw@xmath169 , @xmath170 being the spacetime scalar on which @xmath171 acts and sw denoting _
spin weight_. the commutator of @xmath171 and @xmath172 is @xmath173 now consider the following ansatz for the @xmath6 intrinsic ricci scalar : @xmath174 if we insert this expansion into ( [ commutator ] ) and expand both sides of the equation [ assuming @xmath175 with sw@xmath176 , then to lowest order , namely @xmath177 , we get a trivial equality .
however , equality at the next order demands that @xmath178 this will prove to be a very important result for our purposes .
finally , dougan gives the following expansion for the @xmath6 volume element : @xmath179 ( here @xmath25 is the determinant of the @xmath6 metric and @xmath180 is the asymptotic piece of the shear . )
we now consider the spin - coefficient expression for the smeared energy surface density introduced in @xmath40iv . again with @xmath75 , in the present notation we find @xmath181r^{-2 } + o(r^{-3 } ) { \ , } .\ ] ] moreover , by an argument identical to the one found in the last paragraph of the appendix ( although here with the affine radius @xmath148 rather than the areal radius @xmath11 ) , we know that the result ( [ result ] ) determines @xmath182 as the appropriate asymptotic expansion for the reference term
. therefore , ( @xmath183 times ) the full quasilocal energy surface density is @xmath184r^{-2 } + o(r^{-3 } ) { \ , } .\ ] ] at this point we consider again a smearing function @xmath124 , with appropriate fall - off behavior and limit @xmath185 . using the results amassed up to now , one computes that the limit of the smeared energy surface density is @xmath186\right|_{w = w _ { * } } { \ , } .
\label{supere}\ ] ] the right - hand side of this equation is the `` supermomentum '' of geroch as written by dray and streubel [ see equation ( a1.12 ) of @xcite and set their @xmath187 for a bondi frame as we have here ] ; and the `` supermomentum '' is known to be the charge integral " associated with the ashtekar - streubel flux@xcite of gravitational radiation at @xmath7 ( in the restricted case when the flux is associated with a supertranslation).@xcite dray and streubel have discussed the importance of the particular factors of @xmath188 which multiply the last two terms within the square brackets on the right - hand side of equation ( [ supere ] )
. it is evident from our approach that the origin of these @xmath188 factors stems from the flat - space reference of the quasilocal energy ( flat - space being the correct reference in the present context ) .
when @xmath189 determines a pure bms translation , the last two terms in the integrand integrate to zero .
for instance , setting @xmath190 , one finds that the strict energy @xmath191\right|_{w = w _ { * } } { \,}\ ] ] is the standard spin - coefficient expression for the bondi - sachs mass @xmath192.@xcite
we thank h. balasin , p. t. chruciel , t. dray , and n. murchadha for helpful discussions and correspondence .
we acknowledge support from national science foundation grant 94 - 13207 .
s. r. lau has been chiefly supported by the `` fonds zur frderung der wissenschaftlichen forschung '' in austria ( fwf project p 10.221-phy and lise meitner fellowship m-00182-phy ) .
in this appendix we prove that the subtraction - term contribution @xmath193 to the quasilocal energy obeys @xmath194 for large @xmath11 .
moreover , we derive the result @xmath195 relating the coefficient @xmath196 of the @xmath98 piece of the reference term @xmath28 and the coefficient @xmath131 of the @xmath132 piece of the @xmath6 ricci scalar @xmath197 .
the @xmath193 term is constructed as follows .
consider a generic @xmath6 slice of @xmath18 .
assume that @xmath6 may be embedded isometrically in euclidean three - space @xmath29 and that the embedding is suitably unique ( we address these issues below ) .
let @xmath198 represent the extrinsic curvature of @xmath6 as isometrically embedded in @xmath29 .
the flat - space reference density is @xmath199 , and in terms of this density @xmath200 since it is the intrinsic geometry of @xmath6 which determines the reference term @xmath28 , let us first collect a few results concerning this geometry
. in the bondi coordinate system the two - metric of @xmath6 takes the following form:@xcite @xmath201 in terms of the functions @xmath136 and @xmath137 , @xmath202 and @xmath203 have the expansions [ see ( [ sachs]d ) ] @xmath204 as mentioned , we do not necessarily enforce sachs `` outgoing radiation condition '' which would , in fact , imply that in the expansions for @xmath202 and @xmath203 the terms that appear after the leading @xmath57 terms are @xmath132.@xcite from the form of the line element ( [ bmetric ] ) one easily verifies ( [ sachs]d ) and that @xmath105 . a bit more work establishes that the scalar curvature @xmath197 of @xmath6 has the asymptotic form @xmath205 where @xmath206 .
we have given the explicit form ( [ calr = dw ] ) of the coefficient @xmath207 corresponding to the asymptotic solutions considered here , and we have found this coefficient to be a pure divergence on the unit sphere . that this term integrates to zero can also be shown _ via _ the argument to follow , which does not assume that the einstein equations hold .
the two - sphere average of @xmath139 vanishes , _
i. e. _ @xmath208 . to prove the lemma ,
start with the gauss - bonnet theorem @xcite @xmath209 in the second line we have simply expanded @xmath210 and used @xmath105 . on the right side of the equation , integration over the first term inside the brackets gives @xmath211 .
therefore , we arrive at @xmath212 since @xmath213 falls off _ faster _ than @xmath132 , the @xmath214 limit of this last equation proves the lemma .
now let us discern the @xmath11 dependence of the reference term @xmath28 .
we must first address the issue of whether or not @xmath6 may be isometrically embedded in @xmath29 in a suitably unique way .
it is known that a riemannian manifold possessing two - sphere topology and everywhere positive scalar curvature can be globally immersed in @xmath29 .
an immersion differs from an embedding by allowing for self - intersection of the surface ( seemingly allowable provided that @xmath28 remains well defined ) . the cohn - vossen theorem states that any compact two - surface contained in @xmath29 whose curvature is everywhere positive is unwarpable .
unwarpable means that the surface is uniquely determined by its two - metric , up to translations or rotations in @xmath29 ( which of course do not affect @xmath215).@xcite our surface @xmath6 has two - sphere topology , and for @xmath11 sufficiently large the scalar curvature @xmath210 is everywhere positive .
it thus follows that for suitably large @xmath11 , the reference term @xmath28 is well defined .
the key equation for determining the form of @xmath28 is the standard gauss - codazzi relation:@xcite @xmath216 this equation ( essentially a two - dimensional version of the hamiltonian constraint ) is an integrability criterion , obeyed by our embedding , which relates certain components of the _ vanishing _ riemann tensor of @xmath29 to the intrinsic @xmath6 curvature scalar and the desired reference extrinsic curvature tensor @xmath198 .
next , since @xmath6 approaches a perfectly round sphere as @xmath31 , we take the following expansions for the various pieces of @xmath217 as an ansatz : @xmath218 to avoid clutter , for the various remainder terms we have used in ( [ expand ] ) simply a subscript @xmath219 in place of what should be @xmath220 , _
etc_. plugging the expansions ( [ curvexpand ] ) and ( [ expand ] ) into ( [ key ] ) , one finds that to lowest order , namely @xmath98 , the equation is identically satisfied . at the next order , namely @xmath132 , the equation ( [ key ] ) yields the result ( a2 ) .
therefore , we have found the following asymptotic expansion for @xmath28 : @xmath221 the first lemma we proved above has an important consequence .
it ensures that the `` dangerous '' @xmath109 term in the integral ( [ integral ] ) vanishes ( regardless of whether or not ( [ calr = dw ] ) holds , which requires that the einstein equations are satisfied asymptotically ) . hence we get the result ( [ mainresult ] ) .
see j. w. york , phys .
rev . * d33 * , 2092 ( 1986 ) ; b. f. whiting and j. w. york , phys .
* 61 * , 1336 ( 1988 ) ; j. d. brown , e. a. martinez , and j. w. york , phys .
lett . * 66 * , 2281 ( 1991 ) ; j. d. brown and j. w. york , jr . ,
phys . rev . * d47 * , 1420 ( 1993 ) ; to be published in _ the black hole 25 years after _ , edited by c. teitelboim and j. zanelli ( plenum press ) ( gr - qc/9405024 ) ; and references therein .
see , for example , m. walker , in _ gravitational radiation _ ; nato advanced study institute , edited by nathalie deruelle and tsvi piran , ( north - holland , amsterdam , 1983 ) .
proceedings of the les houches school , two-21 june , 1982 . in this figure one dimension of the two - surfaces @xmath15 is suppressed .
the shaded , partially cut - away , conical surface depicts the null hypersurface @xmath222 determined by a constant value @xmath21 of retarded time .
heuristically , in the limit @xmath31 the @xmath1 slice spanning @xmath223 becomes the asymptotic slice @xmath32 which spans a round spherical cut of @xmath7 , and one should envision the spacelike slice @xmath32 as becoming null asymptotically .
although @xmath33 is timelike everywhere in the _ physical _ spacetime @xmath3 ( at least in relevant exterior regions ) , the _ extension _ to @xmath7 of @xmath33 ( in a conformal completion @xmath34 of @xmath3 ) is a null vector which lies in @xmath7 . again
heuristically , on @xmath17 the @xmath32 hypersurface normal @xmath224 is @xmath33 . | we define the energy of a perfectly isolated system at a given retarded time as the suitable null limit of the quasilocal energy @xmath0 .
the result coincides with the bondi - sachs mass .
our @xmath0 is the lapse - unity shift - zero boundary value of the gravitational hamiltonian appropriate for the partial system @xmath1 contained within a finite topologically spherical boundary @xmath2 .
moreover , we show that with an arbitrary lapse and zero shift the same null limit of the hamiltonian defines a physically meaningful element in the space dual to supertranslations .
this result is specialized to yield an expression for the full bondi - sachs four - momentum in terms of hamiltonian values . |
with the discovery of superconductivity in lafeaso@xmath9f@xmath10,@xcite a new family of iron - based high - temperature superconductors was found .
its simplest member is fese@xmath9 , which consists of a stack of fese layers.@xcite its superconducting transition temperature @xmath11 k increases drastically with external pressure , reaching @xmath12 k.@xcite interestingly , a similar high @xmath13k is attained in the iron - selenide family @xmath14fe@xmath1se@xmath2 by intercalating alkali atoms ( @xmath15 = k , rb , cs ) between the fese layers.@xcite however , @xmath16 is found to decrease with pressure and is fully suppressed at 9 gpa for k@xmath0fe@xmath1se@xmath2 ( ref . ) and at 8 gpa for cs@xmath0fe@xmath1se@xmath2.@xcite the critical temperature is almost insensitive to pressure below 1 gpa , suggesting that @xmath17 is almost independent of small variations of the lattice constants .
this provides an opportunity to study the temperature dependence of physical quantities without being affected by changes of the lattice constants due to thermal expansion .
early @xmath18sr experiments on cs@xmath0fe@xmath1se@xmath2 indicate that superconductivity and antiferromagnetism coexist on microscopic length scales.@xcite the nel temperature @xmath19 k ( ref . ) of k@xmath20fe@xmath1se@xmath2 is substantially higher than @xmath21 k. several experiments point towards nanoscale phase separation between superconducting , vacancy - disordered domains and vacancy ordered antiferromagnetic ( afm ) domains.@xcite in contrast to the slightly hole doped fese@xmath9 , the intercalation of alkali ions @xmath15 into the fese structure introduces a large amount of electrons into the system.@xcite other experiments@xcite suggested that this highly electron doped system contains no hole - like sheets at the fermi surface , and thus electron scattering between hole and electron - like bands is impossible.@xcite moreover , fe vacancies in the crystalline structure order at @xmath22 k.@xcite in k@xmath0fe@xmath1se@xmath2 , the presence of vacancies appears detrimental to superconductivity.@xcite this intriguing microscopic coexistence of vacancy ordering , antiferromagnetism , and superconductivity in @xmath14fe@xmath1se@xmath2 points to an unconventional thermodynamic behavior of the superconducting state .
+ recently , the lower critical field @xmath23 was investigated in tetragonal k@xmath0fe@xmath1se@xmath2 for magnetic fields @xmath24 along different crystallographic directions , _
i.e. _ parallel to the @xmath25-plane and parallel to the @xmath26-axis,@xcite revealing a surprisingly low and isotropic @xmath27 mt . invoking the phenomenological relation between @xmath23 and the effective magnetic penetration depth @xmath4 for an isotropic superconductor @xmath28,@xcite a value of @xmath29 @xmath18 m is obtained ( assuming an approximate ginzburg - landau parameter @xmath30 ) .
this low field estimate of @xmath4 deviates remarkably from the small in - plane magnetic penetration depth @xmath31 @xmath18 m derived from nmr experiments at 8.3 t on a similar sample.@xcite moreover , the numerous observations of an anisotropic vortex lattice in high magnetic fields@xcite contrast with the isotropic behavior of @xmath23 in low magnetic fields . in order to illuminate this intriguing field dependence of the superconducting properties
, we performed a detailed magnetic study of rb@xmath0fe@xmath1se@xmath2 .
a rb@xmath0fe@xmath1se@xmath2 single crystal with composition rb@xmath32fe@xmath33se@xmath2 as refined by x - ray fluorescence was grown from a melt by the bridgman method , using a presynthesised ceramic precursor of fese@xmath34 and metallic rubidium . for the precursor synthesis high purity ( at least 99.99% alfa ) powders of iron and selenium
were mixed in the molar proportion 1 fe : 0.98 se and pressed into a rod .
this nominal stoichiometry , chosen on the basis of our previous studies@xcite of the fe - se chemical phase diagram , provides an iron selenide of pure tetragonal phase .
the mixture was prereacted in a sealed silica ampoule at 700c for 15 hours and then grounded in an inert atmosphere , pressed again into a rod , sealed in an evacuated double wall quartz ampoule and resintered at 700c . after 48 hours
the furnace was cooled down to 400c and kept at this temperature for 36 hours more . for the single crystal growth a piece of the fe - se rod
was sealed in an evacuated silica bridgman ampoule with an appropriate amount of pure alkali metal placed in an additional thin silica tube ; 5% excess of rb was added to compensate its loss during synthesis .
the bridgman ampoule was sealed in another protecting evacuated quartz tube .
the ampoule was heated at 1030c for 2 hours for homogenization , followed by cooling down the melt to 750c at a rate of 6c / h . finally , the furnace was cooled down to room temperature at a rate of 200c / h . after synthesis
the ampoules were transferred to a he - glove box and opened there to protect the crystal from oxidation in the air .
the superconducting properties of the plate - like crystal of dimensions @xmath35 mm@xmath36 ( thickness of 0.2 mm along the @xmath26-axis ) were characterized with a _ quantum design _ mpms xl squid magnetometer .
the temperature dependence of @xmath37/@xmath24 , where @xmath37 is the magnetic moment is shown in fig .
[ m_th ] for various magnetic fields @xmath24 applied after zero - field cooling . the onset transition temperature for this sample
is estimated to be @xmath38 k. for both studied orientations ( parallel to the @xmath26-axis shown in fig .
[ m_th]a , and parallel to the @xmath25-plane shown in fig .
[ m_th]b ) , the magnetic properties are very similar . for @xmath39 mt a sharp diamagnetic transition is observed for both field orientations . in higher fields , the diamagnetism is rapidly suppressed , indicating that @xmath40 for rb@xmath0fe@xmath1se@xmath2 is very low ( @xmath41 0.3 mt , including demagnetizing field correction.@xcite ) + a small piece of of rectangular shape and approximate dimensions @xmath42m@xmath36 was cleaved off the above rb@xmath0fe@xmath1se@xmath2 single crystal .
magnetic torque investigations were carried out using a home - made magnetic torque sensor.@xcite for a uniaxial superconductor like tetragonal rb@xmath0fe@xmath1se@xmath2 , the angular dependence of the magnetic torque @xmath43 in the london approximation ( @xmath44 ) can be written according to@xcite @xmath45 here @xmath46 is the angle between @xmath24 and the crystallographic @xmath26-axis , @xmath47^{1/2}$ ] is the angular scaling function , @xmath48 is the @xmath26-axis upper critical field , and @xmath49 is a dimensionless parameter of the order of unity . without loss of generality @xmath49
is restricted to 1 within this work.@xcite the anisotropy parameter @xmath50/@xmath51 is the ratio of the out - of - plane and in - plane magnetic penetration depths , whereas @xmath52 denotes the effective magnetic penetration depth .
it is possible that our sample presents phase separation between superconducting non - magnetic regions and antiferromagnetic , non - superconducting regions , with a domain size smaller than the penetration depth.@xcite in that case , applying the kogan model yields a @xmath4 which may be renormalized to a larger value than the superconducting parameter @xmath53 .
the field penetrates more easily in this phase separated material , and the kogan model yields an `` averaged '' effective bulk penetration depth . the good agreement between eq .
( [ kogan ] ) and the data ( fig . [ raw]b ) shows that the kogan model is still useful , albeit with a broader interpretation of its parameters .
+ the magnetic torque experiments at various @xmath54 and @xmath24 were performed by turning @xmath24 around the sample in a plane containing the @xmath26-axis ( see inset of fig .
[ raw]a ) , and measuring the resulting torque . as an example ,
[ raw]a shows the angular dependence of the torque signal measured at 15 k in 1.4 t. note that the torque signal is affected by an angular irreversibility between the clockwise ( @xmath55 ) and counter - clockwise ( @xmath56 ) branches .
such irreversible angular dependent torque signals are usually observed in hard superconductors due to vortex pinning.@xcite due to the tetragonal structure of rb@xmath0fe@xmath1se@xmath2 , twinning in the crystal manifests itself in the results only through pinning on the twin boundaries , and all in - plane parameters are not separated along @xmath57- and @xmath58- axes in the analysis . in this work ,
the so - called vortex - shaking technique@xcite was successfully applied to reduce irreversibility , allowing a more reliable determination of the superconducting parameters .
this was done by applying a small ac field orthogonal to @xmath24 in order to enhance the vortex relaxation towards thermodynamic equilibrium . as seen in fig .
[ raw]a the vortex shaking clearly reduces the irreversible component , especially for @xmath24 close to the @xmath25-plane . in fig .
[ raw]b the average torque @xmath59 is presented .
the torque signal consists of a superconducting component described by eq .
( [ kogan ] ) and a magnetic background component @xmath60 .
the variables @xmath61 and @xmath62 denote the magnetic susceptibilities along the crystallographic axes . here
@xmath63 ( as also mentioned in ref . ) and @xmath64 is large , consistent with a bulk antiferromagnetic phase having the magnetic moments aligned along the @xmath26-axis.@xcite the subtraction of this antiferromagnetic background can be either performed by adding a sinusoidal component in the fitting routine of eq .
( [ kogan ] ) , or by directly removing the symmetric sinusoidal component of the data as discussed in ref . .
all results presented in this work are independent of the background treatment . + the parameters @xmath65 , @xmath5 , and @xmath4 can be extracted simultaneously by analyzing the magnetic torque data with eq .
( [ kogan ] ) .
however , in order to reduce the amount of free fit parameters , @xmath48 was fixed according to a werthamer - helfand - hohenberg ( whh ) temperature dependence.@xcite the slope @xmath66 at @xmath16 was fixed to @xmath67 t / k , in concordance with resistivity results of a similar sample.@xcite small variations in @xmath66 do not affect the results of the analysis , since @xmath68 contributes only logarithmically in eq .
( [ kogan ] ) and has no weight in the determination of @xmath5 and @xmath4 in low magnetic fields.@xcite magnetic torque curves with the antiferromagnetic background subtracted are presented in fig .
[ torqueth ] for various @xmath54 ( fig .
[ torqueth]a ) and @xmath24 ( fig . [ torqueth]b ) .
the insets show the normalized magnetic torque @xmath69 $ ] close to the @xmath25-plane .
a change of the shape of @xmath70 qualitatively reflects a change of @xmath5 .
note that the slope of @xmath71 vs. @xmath46 changes strongly with @xmath24 , but not with @xmath54 .
this demonstrates that @xmath5 is field dependent , but not temperature dependent .
as the field direction is approaching the @xmath25-plane , the screening currents start to flow not only in plane , but also out of plane , which makes the torque depend strongly on @xmath72 _ and _ @xmath73 in this angular region . the temperature and field effect on the anisotropy @xmath5
are consequently already visible on the data taken around the @xmath25-plane , independently of the strict validity of eq .
( [ kogan ] ) in a phase separated material .
+ the temperature dependence of @xmath4 presented in fig .
[ gamma_t]a can be fitted with the empirical power law @xmath74 $ ] with @xmath3 k , @xmath75 and @xmath76 @xmath18 m .
this rather large value of the effective penetration depth may be due to phase separation , as mentioned above .
the value of @xmath77 is quite large compared to what is expected from the two - fluid model the fitting expression comes from .
this might be related to phase separation , as the links between superconducting domains could depend on temperature , changing the temperature dependence of the penetration depth .
figure [ gamma_t]b shows @xmath78 for various @xmath24 .
it is almost temperature independent between @xmath16 and @xmath16/2 for all fields studied .
the slight drop observed at higher temperatures in 1.4 t may be due to the proximity of the transition , which decreases the superconducting signal .
the increase of @xmath79 at 1.4 t at low temperatures is most likely due to pinning effects.@xcite most importantly , a remarkable monotonous field dependence of @xmath5 is observed ( fig .
[ gamma_t]d ) , with the strongest dependence in the lowest fields . in low fields , @xmath80 .
an extrapolation of the measured values of @xmath5 towards zero field yields @xmath81 . in fig .
[ gamma_t]c the field dependence of @xmath82 is shown for various @xmath54 .
the effective penetration depth @xmath82 tends to a constant value for @xmath83 1 t for all studied @xmath54 .
the extrapolated @xmath84 @xmath18 m is surprisingly large compared to that of other iron - based superconductors . in the related iron selenide fese@xmath85te@xmath85 a much lower value of @xmath86 @xmath18 m was reported.@xcite however , for rb@xmath0fe@xmath1se@xmath2 such a large @xmath87 is consistent with the very small @xmath88 mt observed in this work and in k@xmath0fe@xmath1se@xmath2 by others.@xcite the small @xmath89 at very low fields is also in agreement with an isotropic @xmath23 .
the increase of @xmath90 with @xmath24 and the field independent @xmath52 imply that @xmath91 decreases with increasing @xmath24 .
this is consistent with the high - field nmr result.@xcite the small high - field value of @xmath51 reported in ref .
, in combination with the saturating @xmath82 and the field dependence of @xmath92 observed here , suggests that @xmath5 continues to increase at higher fields .
note that the anisotropy of @xmath93 ( ref . ) observed in very high fields by upper critical field measurements can not be directly compared with the magnetic penetration depth anisotropy @xmath5 investigated here , because in general @xmath94.@xcite + a field dependent @xmath5 might be associated with a complex band structure , since in the case of multiple superconducting gaps originating from different bands the superconducting screening currents , related to @xmath51 and @xmath95 , may give rise to an unusual behavior of @xmath5 .
a similar behavior was observed in mgb@xmath2 , where the two - gap excitation spectrum yields a strongly field dependent anisotropy.@xcite in mgb@xmath2 , as the field increases , the gap from the 3d @xmath96-band is closed , so the main part of the superconducting fluid density comes from the 2d @xmath97-band .
the 2d character of the remaining band implies a larger anisotropy .
+ band structure calculations@xcite in @xmath14fe@xmath1se@xmath2 yield multiple bands , with four 2d sheets on the sides of the brillouin zone and one cylindrical sheet at the center .
depending on doping , this cylinder can be split into two 3d cones in k@xmath20fe@xmath2se@xmath2,@xcite which are completely detached ( 3d ) in the calculation of ref . ,
although when the authors use experimental lattice parameters these cones are replaced by a cylinder ( 2d ) . according to ref . , the center band is a cylinder . in csfe@xmath2se@xmath2 and cs@xmath20fe@xmath2se@xmath2
, this band consists in two almost detached cones.@xcite in tl@xmath98rb@xmath99fe@xmath100se@xmath2,@xcite two gaps of different amplitudes were observed on the side and center bands , but the candidate for a 3d inner band was to small for a gap to be observed .
it is therefore not yet clear whether for rb@xmath0fe@xmath1se@xmath2 with the doping used in this work a 3d band is present .
clearly , more material specific work is needed in order to clarify the interplay between multiband superconductivity and the anisotropy of rb@xmath0fe@xmath1se@xmath2 . however , if a 2d-3d band scenario is the origin of the field dependence of the anisotropy of rb@xmath0fe@xmath1se@xmath2 , it must be temperature dependent as well,@xcite which is not the case according to fig .
[ gamma_t]d .
+ the temperature independent @xmath5 suggests that the origin of its field dependence is not related to the superconducting gap energy , which is strongly temperature dependent in the examined temperature range .
however , there is one energy scale which is almost temperature independent in the superconducting state : the nel temperature @xmath19 k.@xcite a superconductor coexisting with an antiferromagnetic phase is expected to behave in a peculiar way , although the temperature range studied here is far too low to excite any change in this strongly coupled antiferromagnet .
however , magnetic field modifications can lead to changes in the domain structure , and therefore changes of coupling between superconducting areas .
this could result in variations of the `` averaged '' effective anisotropy . as was shown for various iron - based superconductors , the lattice parameters , in particular the pnictogen height in the iron - pnictides ,
are directly related to superconductivity.@xcite importantly , such scaling also works for the iron - selenide layer.@xcite it is also possible that magnetostrictive effects , which are expected to increase with magnetic field , may influence the lattice parameters and by that the anisotropy of the system .
in such a case the strong curvature of @xmath68(@xmath54 ) in the vicinity of @xmath16 often observed in pnictides / chalcogenides may be explained by the change of @xmath16 with @xmath24 caused by the change of pnictogen height .
in summary , we present an investigation of the magnetic properties of rb@xmath0fe@xmath1se@xmath2 revealing a strong field dependence of @xmath5 ranging from @xmath101 to @xmath102 . this behavior stands out among other iron - based superconductors , consistent with the singular coexistence of magnetic ordering and superconductivity . in accordance with lower critical field measurements ,
our data suggest that in very low fields @xmath89 . at 1.4 t
the effective magnetic penetration depth is @xmath103 m .
the vortex phase in rb@xmath0fe@xmath1se@xmath2 is best described by an isotropic three dimensional state in low fields which becomes strongly anisotropic with increasing field . in this respect
the novel iron - selenide rb@xmath0fe@xmath1se@xmath2 could be a potential candidate for magnetic - field tuned applications of superconductivity .
this work was supported by the swiss national science foundation , in part by the nccr program manep and the sciex - nmsch ( project code 10.048 ) , and by national science centre ( poland ) based on decision no .
dec-2011/01/b / st3/02374 .
42ifxundefined [ 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 * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) `` , '' ( , ) chap . , pp . @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) | the anisotropic superconducting properties of a rb@xmath0fe@xmath1se@xmath2 single crystal with @xmath3 k were investigated by means of squid and torque magnetometry , probing the effective magnetic penetration depth @xmath4 and the penetration depth anisotropy @xmath5 .
interestingly , @xmath5 is found to be temperature independent in the superconducting state , but strongly field dependent : @xmath6 and @xmath7 .
this unusual anisotropic behavior , together with a large zero - temperature @xmath8 m , is possibly related to a superconducting state heavily biased by the coexisting antiferromagnetic phase . |
the experiments at hera measure deep - inelastic electron - proton scattering in the previously unexplored small @xmath1 regime , @xmath4 . as usual , @xmath1 is the bjorken variable @xmath5 and @xmath6 , where @xmath7 and @xmath8 are the four momenta of the incoming proton and the virtual photon probe respectively .
the deep - inelastic observables reflect the small @xmath1 behaviour of the gluon , which is by far the dominant parton in this kinematic region .
in particular the small @xmath1 behaviour of the structure function @xmath9 is driven by the @xmath10 transition . at such values of @xmath1 soft gluon emission and
the associated virtual gluon corrections give rise to powers of @xmath11 , which have to be resummed . to leading @xmath12 order , this is carried out by the bfkl ( or lipatov ) equation@xcite which , in a physical gauge , effectively amounts to the summation of gluon ladder diagrams .
two characteristic features of the result of this procedure are , first , the @xmath13 growth of the gluon distribution @xmath14 as @xmath1 decreases , with @xmath15 , and , second , the relaxation of the strong - ordering of the transverse momenta , @xmath16 , of the gluons along the ladder ( which is characteristic of the `` large '' @xmath1 , large @xmath17 `` altarelli - parisi '' regime )
. the recent measurements of @xmath9 at hera@xcite show a rise with decreasing @xmath1 which is entirely consistent with the @xmath13 bfkl perturbative qcd prediction@xcite .
however the observed small @xmath1 behaviour of @xmath18 can also be mimicked by conventional dynamics based on the altarelli - parisi equation@xcite .
because of the inclusive nature of @xmath18 , it is not a sensitive discriminator between bfkl and conventional dynamics . for this purpose
it is necessary to look into properties of the final state . here
we investigate whether or not the emitted transverse energy , @xmath0 , which accompanies a deep - inelastic event ( see fig .
1 ) , can be used to identify the bfkl dynamics . due to the relaxation of the strong - ordering of the gluon @xmath16 s along the chain we expect to find more transverse energy in the central region ( between the current jet and the proton remnants ) than would result from conventional evolution .
indeed a monte carlo study , in the context of heavy quark production , of the @xmath0 distribution as a function of rapidity hints at an increase of the @xmath0 flow if qcd all - loop " dynamics @xcite ( which incorporates bfkl effects ) is employed , see fig . 9 of ref
. @xcite . now the @xmath0 emitted by the gluon chain is given by the integral over the inclusive distribution for the emission of one gluon weighted by its @xmath0 .
thus , for a given element of @xmath19 , the distribution of transverse energy as a function of @xmath20 is given by @xmath21 where @xmath22 denotes the cross section in the @xmath19 interval , and where @xmath23 and @xmath24 are the longitudinal momentum fraction and the transverse momentum carried by the gluon jet in the photon - proton centre - of - mass frame .
the variable @xmath20 is closely related to rapidity , see sec .
[ sec : twob ] . in terms of the proton structure functions , the differential cross section in eq .
( [ aa ] ) is @xmath25 \label{ab}\ ] ] where @xmath26 with @xmath27 being the centre - of - mass energy of the colliding @xmath28 system . with these definitions
, we can calculate the energy flow , @xmath29 of eq .
( [ aa ] ) , by using the diagrams of fig . 2 to evaluate the differential structure functions .
we have @xcite @xmath30 with @xmath31 , where @xmath32 and @xmath33 describe , respectively the soft gluon resummation above and below the emitted gluon of fig .
that is @xmath32 and @xmath33 satisfy bfkl equations , provided of course that @xmath34 and @xmath23 are sufficiently small . the bfkl equation for @xmath33 may be written in the differential form @xmath35 \equiv k \otimes f , \label{ad}\ ] ] with a similar form @xcite for @xmath32 . as usual
, @xmath33 is the unintegrated gluon distribution in the proton , that is with the final integration over @xmath36 not performed .
the expression in brackets in ( [ ac ] ) arises from the ( square of the ) bfkl vertex for real gluon emission , where @xmath37 and @xmath24 are the gluon transverse momenta in the photon - proton frame .
we may use eqs ( [ aa])([ac ] ) , together with the solutions of the bfkl equations for @xmath33 and @xmath32 , to determine the @xmath0 flow , @xmath38 , which accompanies deep - inelastic scattering , provided @xmath39 .
valuable insight can be obtained by using the analytic solutions of the bfkl equation for fixed @xmath40 .
therefore , before presenting the full numerical calculation , we perform in sec .
[ sec : two ] an analytic study to establish general characteristics of the @xmath0 distribution .
first , we make the simplifying assumption of strong - ordering at the vertex for real gluon emission ; in fact in addition to the normal `` strong - ordering , @xmath41 , in the expression in brackets in eq .
( [ ac ] ) , we also need to include the contribution from the anomalous '' ordering @xmath42
. then we present an exact analytic study by using the bfkl form of the vertex , which incorporates all orderings automatically . in sec .
[ sec : four ] we present the predictions of the @xmath0 distribution obtained when the bfkl equation is solved numerically . at each stage
we compare with the @xmath0 flow pattern predicted by coventional dynamics based on the altarelli - parisi equation .
section [ sec : five ] contains our conclusions .
at first sight the prediction for the shape of the @xmath0 distribution as a function of @xmath20 looks simple .
the leading small @xmath43 behaviour of the solutions of the bfkl equation for fixed @xmath40 have the form @xmath44 where @xmath45 , and similarly for @xmath32 .
thus for @xmath34 and @xmath23 sufficiently small , we see that the relevant combination which occurs in eq .
( [ ac ] ) has the behaviour @xmath46 which appears to give an @xmath0 distribution which is independent of @xmath23 , and also of @xmath1 .
thus well away from both the current jet and the proton remnants , we anticipate a plateau distribution of @xmath0 as a function of @xmath47 , which remains unchanged as @xmath1 varies .
however the situation is more subtle .
we have overlooked the @xmath1 and @xmath23 dependences which arise from the integrations over the transverse momenta . to see this we need to look more closely at the analytic solution of the bfkl equation .
recall that the solution may be obtained by taking the mellin transform @xmath48 of @xmath33 : @xmath49 @xmath50 with @xmath51 .
if we substitute eq .
( [ bd ] ) into ( [ ad ] ) we obtain @xmath52 where @xmath53 , the transform of the bfkl kernel , is @xmath54 \label{bf}\ ] ] with @xmath55 . on substituting the solution of eq .
( [ be ] ) into eq .
( [ bc ] ) we have @xmath56 as @xmath57 varies along the contour of integration , the maximum value of @xmath58 is found to occur at @xmath59 .
the region @xmath60 therefore dominates the small @xmath43 behaviour of @xmath33 . to obtain the leading behaviour we expand the various terms in eq .
( [ bg ] ) about the point @xmath59 and find @xmath61^{\frac{1}{2 } } } \exp\left ( \frac{-\log ^2(k^2/\bar{k}^2 ) } { 2\lambda^{\prime\prime}\log ( 1/z ) } \right ) , \label{bh}\ ] ] with an analogous relation for the @xmath32 @xmath62^{\frac{1}{2 } } } \exp \left ( \frac{-\log ^2(k^2/aq^2)}{2\lambda^{\prime\prime}\log ( 1/z ) } \right ) \label{bi}\ ] ] where @xmath63 and @xmath64 are given in terms of the boundary conditions .
the quantities @xmath65 and @xmath66 arise from the expansion of @xmath58 of eq .
( [ bf ] ) about the point @xmath59 : @xmath67 with @xmath65 given by eq .
( [ ba ] ) and @xmath68 , where the riemann zeta function @xmath69 .
note that the scale of the @xmath32 is provided by @xmath17 and also that , with our definition @xcite , the @xmath32 have the dimensions of @xmath17 . in eq .
( [ bh ] ) we have used a slightly different notation for the function @xmath70 at the boundary ( which for simplicity we take to be @xmath71 ) .
we have introduced a dimensionless boundary function @xmath72 defined by @xmath73 similarly for the @xmath32 we have introduced dimensionless boundary functions @xmath74 defined by @xmath75 to estimate the characteristics of the @xmath0 flow we may omit the longitudinal structure function and assume @xmath76 .
it is straightforward to include the small correction arising from @xmath77 .
thus ( [ aa ] ) becomes @xmath78 where the integrand is specified by ( [ ac ] ) with @xmath33 and @xmath79 given by eq .
( [ bh ] ) and eq .
( [ bi ] ) respectively , for sufficiently small @xmath23 and @xmath34 .
thus to determine the @xmath0 flow we must perform the integrations @xmath80 over the transverse momenta , subject of course to the conservation of transverse momentum as embodied in the delta function in ( [ ac ] ) .
we can obtain valuable insight by making the assumption that the transverse momenta are strongly - ordered at the vertex for real gluon emission . at first sight
, we would expect that this meant just keeping the contribution with @xmath81 in fig . 2 .
however for the @xmath0 weighted distribution we shall find that there is an equally big contribution coming from the region with anomalous " ordering @xmath82 .
we evaluate these two contributions in turn . for the normal " strong - ordered case with @xmath41 we have @xmath83 ;
then eq .
( [ ac ] ) simplifies , since the @xmath84 integration over @xmath33 just gives the integrated gluon distribution @xmath85 .
in fact in this approximation we recover the formula for deep - inelastic + energetic jet events @xmath86 see , for example , eq . ( [ bd ] ) of ref . @xcite . using the explicit solutions ( [ bh ] ) and ( [ bi ] ) , and inserting ( [ bn ] ) into ( [ bm ] ) , we obtain @xmath87^{\frac{1}{2 } } } \label{bo}\ ] ] where @xmath88 the factor 2 in eq .
( [ bo ] ) arises when we translate formula ( [ bh ] ) for @xmath89 into one for @xmath90 .
this factor can be inferred by using mellin transform techniques and noting that the main contribution occurs when @xmath91 .
the other contribution , which we denote @xmath92 , coming from the anomalous `` ordered region @xmath82 , can be readily shown to be equal to @xmath93 . to see this , we note that for this case @xmath94 and so we have @xmath95 where the factor @xmath96 arises from @xmath97 if we insert eqs ( [ bh ] ) and ( [ bi ] ) then we find @xmath92 is also given by eq.([bo ] ) .
this so - called anomalous contribution @xcite comes from ordering of transverse momenta that are opposite to that in the ordinary altarelli - parisi evolution of the structure function .
we have found that in the small @xmath1 limit and for fixed ( albeit large ) @xmath17 , that it gives a contribution to the @xmath0 flow equal to that arising from the normal strong - ordering . why does this anomalous ''
contribution not occur , for instance , in the conventional double leading logarithm ( dll ) limit of , say , @xmath9 ? in the total cross section ( or in an inclusive structure function ) the dominant or leading anomalous `` contribution is cancelled by the virtual corrections which lead to gluon reggeisation @xcite ( or , equivalently , to the non - sudakov form factor '' of ref.@xcite ) .
the cancellation ensures that , at small @xmath1 , we do not encounter the double logarithmic terms of the form @xmath98 .
moreover the virtual corrections guarantee that the conventional dll terms @xmath99 only come from the usual strongly - ordered region of transverse momenta , i.e. @xmath100 .
it may also be interesting to observe that , unlike the unweighted distributions , the @xmath0 flow is an infrared finite quantity since the factor of @xmath101 in the integrand of ( [ aa ] ) removes the unintegrable singularity which would otherwise occur at @xmath102 , see ( [ ac ] ) .
after this aside , we now return to complete the analytic calculation of the @xmath0 flow
. the integral @xmath103 of eq .
( [ bp ] ) can readily be evaluated if we change the variable from @xmath104 to @xmath105 @xmath106 .
then eq .
( [ bp ] ) becomes @xmath107 where @xmath108 . if we evaluate this gaussian integral , and insert the result into eq .
( [ bo ] ) and into the identical expression for @xmath109 , we finally obtain the following analytic prediction for the @xmath0 flow @xmath110 ^ 2 } { \log ( 1/x ) } \right ) , \label{br}\ ] ] provided @xmath34 and @xmath23 are sufficiently small , where @xmath111 . to obtain eq .
( [ br ] ) we have used the bfkl prediction for @xmath18 in the leading @xmath112 approximation @xcite @xmath113 where @xmath114 is the quark box ( and crossed box ) contribution shown in fig .
2(b ) , and where the integration over the gluon longitudinal momentum has been performed .
if we insert expression ( [ bh ] ) for @xmath33 and we take the small @xmath1 limit ( where the gaussian factor may be neglected ) , then the integral becomes proportional to the mellin transform @xmath115 at @xmath59 .
using eq .
( [ bl ] ) we obtain @xmath116^{\frac{1}{2 } } } .
\label{bt}\ ] ] this result for @xmath18 has been inserted into eq .
( [ bo ] ) . equation ( [ br ] ) for the @xmath0 flow applies in the central region away from the current jet and the proton remnants , that is when @xmath117 and @xmath118 .
contrary to our initial expectation , ( [ bb ] ) , the expression ( [ br ] ) depends on @xmath1 , and also on @xmath23 .
the main characteristics of the @xmath0 flow can be readily identified from eq .
( [ br ] ) .
we see that the @xmath0 distribution has a broad gaussian shape in @xmath47 ( or in rapidity , see sec .
[ sec : twob ] below ) with * a peak which moves away from @xmath119 as @xmath17 increases , * a width which grows as @xmath120 as @xmath1 decreases , * an @xmath2 dependence such that it grows with decreasing @xmath1 , * a @xmath121 dependence such that it grows with increasing @xmath17 .
although the shape of the distribution is specified , the normalisation is controlled by @xmath122 , the parameter which gives the centre of the input ( non - perturbative ) @xmath123 gaussian distribution of @xmath33 , see eq .
( [ bh ] ) and ref .
@xcite .
we have calculated the @xmath0 flow as a function of @xmath47 , a variable which is closely related to the rapidity , @xmath124 .
the result , @xmath125 , is the gaussian form of eq .
( [ br ] ) . for practical purposes
the rapidity distribution @xmath126 is more useful , so here we show how to evaluate the integral @xmath103 of eq .
( [ bq ] ) in terms of @xmath124 , rather than @xmath47 . in this way we will find that the @xmath0 flow as a function of rapidity also has a gaussian form .
we define the rapidity @xmath124 in the photon - proton c.m .
frame in the standard way @xmath127 which leads to the following relation between @xmath124 and @xmath47 @xmath128 in terms of @xmath124 , the integral @xmath103 of eq .
( [ bq ] ) assumes the ( non - gaussian ) form @xmath129 } - \frac{(t - t)^2}{\lambda^{\prime\prime}[2y - t+t + \log ( 1/x ) ] } \right ) \label{bw}\ ] ] where the limits @xmath130 now , to obtain an analytic form we approximate the @xmath131 factors in the denominators in eq .
( [ bw ] ) by @xmath132 which denotes the value of @xmath131 for which the integrand in eq .
( [ bq ] ) is a maximum in the leading @xmath133 approximation .
it is straightforward to evaluate the integral approximated in this way .
we find @xmath134 ^ 2}{\log ( 1/x ) } \right ) .
\label{by}\ ] ] thus we obtain a gaussian @xmath0 distribution as a function of rapidity @xmath124 , identical in structure to that in terms of @xmath47 , except for the two non - leading terms in the exponential .
equation ( [ br ] ) , though valuable to see the origin of the @xmath1 and @xmath23 dependence of @xmath0 , is based on assumptions which are not necessary even for an analytic treatment . in this subsection
we show that it is not necessary to assume ( i ) strong - ordering at the gluon emission vertex and ( ii ) the approximate gaussian diffusion patterns of eq .
( [ bh ] ) and eq .
( [ bi ] ) . the procedure is to evaluate ( [ aa])([ac ] ) using mellin transform techniques .
we have @xmath135 @xmath136 the function @xmath137 , introduced as above , is now replaced by the inverse mellin transform @xmath138 with @xmath139 in the range @xmath140 . proceeding in this way
we see that the @xmath141 and @xmath142 integrations , over @xmath33 and @xmath143 respectively , now factorize and , moreover , are simply of the form of the mellin transforms @xmath48 and @xmath144 , defined as in eq .
( [ bd ] ) .
thus eq .
( [ bz ] ) becomes @xmath145 the mellin transform @xmath146 satisfies the transformed bfkl equation ( [ be ] ) which has the solution @xmath147 see eq .
( [ bk ] ) .
similarly the solution of the equation for @xmath144 is @xmath148 recall eq .
( [ bl ] ) .
the explicit form of the function @xmath149 can be simply deduced from , for example , appendix b of ref .
@xcite . inserting solutions
( [ bzc ] ) and eq .
( [ bzd ] ) into eq .
( [ bzb ] ) , we obtain @xmath150 we see that in the limit where @xmath117 and @xmath151 , and in particular in the central region away from the current jet and the proton remnants where @xmath119 ( or to be more precise where @xmath152 , that the integral is dominated by contributions from the region with @xmath153 .
we therefore expand the relevant quantities about @xmath154 and evaluate the resulting gaussian integral .
we find the master analytic formula @xmath155 which may be compared with result ( [ br ] ) obtained by the simplified analytic treatment . here
we see we have recovered the same general structure as before , that is a gaussian @xmath0 distribution as a function of @xmath47 , with the characteristics ( i ) to ( iv ) listed in sec .
[ sec : twoa ] .
the quantities are now more precisely determined . first the parameter @xmath157 , which controls the growth of the distribution with decreasing @xmath1 , is given by @xmath158 which would reduce to the previous result , @xmath111 , if the quadratic form eq .
( [ bj ] ) of @xmath58 were valid to @xmath159 .
the parameters @xmath160 and @xmath161 denote the derivatives of @xmath53 at @xmath159 @xmath162 and finally the normalization @xmath163 where @xmath164 as before we can translate the gaussian @xmath0 distribution in @xmath47 to one in terms of rapidity @xmath124 .
again the distribution turns out to have the form of eq .
( [ bzf ] ) , @xmath165 but we see that there are additional , non - leading , terms in the exponential .
we compare the bfkl forms of the @xmath0 distribution , eq.([br ] ) and eq .
( [ bzf ] ) , with that obtained from conventional dynamics based on altarelli - parisi evolution in the small @xmath1 limit . to make a strict comparison we take fixed @xmath40 .
thus rather than inserting into eq .
( [ bn ] ) the bfkl solutions ( [ bh ] ) and ( [ bi ] ) , we replace them by the double leading logarithm ( dll ) forms @xmath166^{\frac{1}{2 } } \right ) \label{ca}\ ] ] @xmath167^{\frac{1}{2 } } \right ) .
\label{cb}\ ] ] we substitute these forms into eq .
( [ bn ] ) and eq .
( [ bm ] ) , and change the @xmath104 integration variable to @xmath168 .
we then obtain @xmath169^{\frac{1}{2 } } + 2 \left [ \bar{\alpha}_s t \log \frac{1}{x_j } \right]^{\frac{1}{2 } } \right ) \label{cc}\ ] ] where @xmath170 .
it is informative to evaluate the @xmath0 distribution for @xmath171 , that is for @xmath172 a point midway " between the current jet and the proton remnants .
this choice facilitates an analytic study and allows us to gain insight into the @xmath1 and @xmath17 dependence of the @xmath0 flow .
when @xmath173 we can evaluate eq .
( [ cc ] ) using saddle point methods .
we find @xmath174 \right ) \label{ce}\ ] ] where @xmath175 is a known function of the ratio @xmath176 . to be precise @xmath177 ^{-2}\right\ } ^{\frac{1}{2 } } \right ) .
\label{cf}\ ] ] the @xmath178 in the square brackets in eq .
( [ ce ] ) arises from substituting @xmath179^{\frac{1}{2 } } \right ) \label{cg}\ ] ] into eq .
( [ cc ] ) , rather than using eq .
( [ bt ] ) . after some algebra , it is possible to show that the approximate estimate of the integral in eq .
( [ cc ] ) predicts a slow increase of @xmath0 with decreasing @xmath23 .
we inspect the properties of the @xmath0 flow , eq .
( [ ce ] ) , in two different limits .
first in the double leading logarithm limit , where @xmath180 but @xmath181 and @xmath182 , we see that @xmath183 , and so @xmath184 thus in the dll limit we find that the @xmath0 distribution grows essentially as @xmath185 with increasing @xmath17 . in the second limit , @xmath186 at finite @xmath17 , we see that @xmath187 .
thus we find @xmath188 moreover it is possible to show from eq .
( [ ce ] ) that with decreasing @xmath1 the distribution decreases towards its final limit , which is indicated by eq .
( [ ci ] ) ( modulo slowly varying logarithmic factors ) .
this should be contrasted with the bfkl behaviour of eq .
( [ bzf ] ) which increases as @xmath189 with decreasing @xmath1 .
above we have obtained analytic forms for the leading behaviour of the @xmath0 distribution in the @xmath190 , @xmath191 limit , for fixed @xmath40 . these expose characteristic features of the expected @xmath0 flow in deep - inelastic scattering in the central region between , but well away from , the current jet and proton remnants . to obtain a more realistic estimate we solve the appropriate bfkl equations numerically , and use running @xmath40 . from eqs ( [ ac ] ) and
( [ bm ] ) we see that the @xmath0 flow , for a given @xmath192 , can be expressed as @xmath193 in this paper we focus on the central region with @xmath194 and @xmath195 . thus to calculate the @xmath0 flow we have to solve three different bfkl equations , one for @xmath33 which resums the lower gluon ladders of fig .
2(b ) , one for @xmath79 which resums the upper gluon ladders , and finally one for @xmath9 . to be precise
we determine @xmath89 for @xmath196 by solving the bfkl equation , as described in ref.@xcite , starting from a gluon distribution at @xmath197 obtained from the mrs set of partons@xcite .
the quantity @xmath198 may be identified with the structure function of a gluon of ( approximate ) virtuality @xmath199 , integrated over its longitudinal momentum .
the function @xmath200 is determined for @xmath201 by solving the bfkl equation , as described in ref.@xcite , starting from the boundary condition @xmath202 at @xmath203 , where @xmath114 is the contribution resulting just from the quark box ( and crossed box ) in fig
finally @xmath9 is determined by the convolution of the gluon ladder with the quark box ( and crossed box ) , which symbolically may be written @xmath204 , as described in @xcite .
the infrared cut - off on the transverse momenta integrations is taken to be @xmath205 gev@xmath206 throughout , a value which makes the calculated values of @xmath18 consistent with the recent observations at hera @xcite .
the predictions for the @xmath0 flow are less sensitive to the choice of the cut - off than those of @xmath18 .
the ultraviolet cut - off on the integrations over the transverse momenta is taken to be @xmath207 gev@xmath206 ; the results for @xmath208 and @xmath18 are not sensitive to reasonable variations about this value .
we shall see below , however , that the @xmath0-weighted integration of eq .
( [ da ] ) is sensitive to the ultraviolet cut - off as @xmath23 decreases towards the current jet .
numerical results for the @xmath0 flow are presented in fig . 3 for @xmath209 gev@xmath206 and values of @xmath210 .
we show results for such low values of @xmath1 so as to establish a sufficiently large central region in order to compare the general features of the distribution with the expectations of the leading order analytic treatment of sec.[sec : two ] from fig .
3 we indeed see a broad gaussian shape in @xmath47 , with a peak below @xmath171 and a width which grows with decreasing @xmath1
. moreover the full numerical treatment shows an _ increase _ of the height of the distribution with decreasing @xmath1 , but not as rapid as anticipated by the leading order analytic result .
we also show in fig .
3 the @xmath0 distributions obtained using altarelli - parisi ( ap ) evolution for two values of @xmath1 .
these confirm the expectation that they _
decrease _ with decreasing @xmath1 .
we come back to the bfkl comparison with ap when discussing fig . 5 , but first in fig .
4 we explore the sensitivity to the choice of the ultraviolet cut - off in eq.([da ] ) .
having established the general features of the @xmath0 flow , we focus on the distribution for @xmath211 , a value which is accessible at hera . in fig .
4 the sequence of dashed curves show the effect of choosing the ultraviolet cut - off in eq.([da ] ) to be @xmath212 gev@xmath206 , and the continuous curve is the result obtained if the cut - off is chosen to be @xmath213 as implied by energy - momentum conservation @xcite , where @xmath214 . at relatively low values of @xmath17 , such as @xmath209 gev@xmath206 , which are relevant to experiments at hera , energy conservation has a significant effect on the @xmath0 flow arising from bfkl gluon emission . fig . 5 compares the @xmath0 flow for @xmath209 gev@xmath206 with that for @xmath215 gev@xmath206 with , in each case , @xmath216 .
also shown are the expectations from altarelli - parisi evolution .
as anticipated we see a significant difference between the @xmath0 flow arising from bfkl and ap gluon emission , particularly at lower @xmath17 values .
we may use eq .
( [ bu ] ) to translate the @xmath0 flow as a function @xmath23 into a distribution in terms of the rapidity @xmath124 in the virtual - photon - proton c.m . frame . for a reasonably flat @xmath0 distribution a good estimate of @xmath124
is obtained if we insert the average value of @xmath104 into eq .
( [ bu ] ) . for @xmath209
gev@xmath206 we see that @xmath217 gev per unit of rapidity and so we take @xmath218 gev@xmath206 . the result for @xmath124
is the scale shown above fig .
if the @xmath0 distribution observed by the h1 collaboration @xcite is compared with monte carlo expectations , then reasonable agreement is found for @xmath219 , but for @xmath220 more @xmath0 appears to be present in the region between the current jet and the proton remnants than is implied by the favoured monte carlo calculations ( which at present do not allow for bfkl emissions ) .
the deficiency is most evident in the rapidity region @xmath221 , which extends beyond the central region shown in fig .
however it is encouraging that a simple extrapolation of the bfkl curves of fig . 5 into this region indicates a sizeable enhancement of the predicted @xmath0 distribution over that expected from conventional ap dynamics .
our main objective is to find characteristics which are unique to bfkl dynamics at small @xmath1 .
one striking feature of bfkl dynamics is the very strong @xmath13 , with @xmath222 , increase of the deep - inelastic structure function @xmath9 as @xmath1 decreases in the hera regime @xcite .
however this increase can be mimicked by conventional altarelli - parisi evolution from a very low starting scale @xmath223 gev@xmath206 @xcite .
global or inclusive quantities like structure functions do not provide conclusive tests of bfkl dynamics and it has been advocated that the bfkl effect can , in principle , be more definitively identified by studies of the final state in deep - inelastic scattering .
one possibility is mueller s proposal @xcite of measuring the @xmath34 dependence of deep - inelastic events containing a forward - going identified jet of longitudinal momentum @xmath224 .
another , which we have studied here , is to measure the @xmath0 flow accompanying deep - inelastic events at small @xmath1 .
a distinctive feature of bfkl dynamics is the diffusion pattern of the transverse momentum flow , in contrast to the strong - ordering of @xmath16 towards @xmath17 which is characteristic of conventional altarelli - parisi dynamics .
indeed it is the absence of strong @xmath16-ordering which leads to a strongly enhanced @xmath0 distribution at small @xmath1 .
we analysed this effect both analytically and numerically , focussing attention on the central region between the current jet and the proton remnants . in sec .
[ sec : three ] we performed a detailed analytic analysis of the @xmath0 flow .
we found that bfkl dynamics with fixed @xmath40 gives a broad gaussian @xmath0 distribution as a function of rapidity ( or rather @xmath47 ) which grows as @xmath2 with decreasing @xmath1 where @xmath2252log2 .
this should be contrasted with the much smaller @xmath0 flow obtained assuming strong @xmath16-ordering , which gives an @xmath0 distribution that decreases with decreasing @xmath1 , for fixed @xmath17 . in section 4
we confirmed these qualitative features using numerical solutions of the appropriate bfkl equations .
as expected we found that the @xmath0 flow which accompanies deep - inelastic events is much larger for bfkl than for conventional ap dynamics .
we also showed , at the low @xmath17 values relevant to hera , that energy conservation should be used to limit the energy - weighted integrations over transverse momentum .
taking , for example , deep - inelastic events with @xmath211 and @xmath209 gev@xmath206 we found that the @xmath0 flow in the central rapidity region @xmath226 between the current jet and the proton remnants is about 2 gev per unit of rapidity .
this result is very encouraging for the identification of bfkl dynamics .
however for a quantitative comparison with experiment we will need to include contributions to the @xmath0 flow coming from the hadronization of the produced partons , not only from the bfkl ladder , but also from the partons radiated from the current jet . also there is a contribution radiated from the colour string which connects the current jet to the proton remnants .
an analysis of this problem is under study and its results will be reported elsewhere .
we thank yuri dokshitzer and genya levin for useful discussions .
jk and kgb thank grey college and the department of physics at the university of durham for warm hospitality .
two of us ( jk , adm ) thank the european community for fellowships and two of us ( kgb , pjs ) thank the polish kbn - british council collaborative research programme for partial support . this work has also been supported in part by polish kbn grant no .
2 0198 91 01 , by the uk science and engineering research council and eu under contract no .
chrx - ct92 - 0004 .
kuraev , l.n .
lipatov and v.s .
fadin , sov .
jetp * 45 * 199 ( 1977 ) ; ya.ya .
balitsky and l.n .
lipatov , sov . j. nucl
* 28 * 822 ( 1978 ) h1 collaboration : i. abt et al .
, nucl.phys . *
b407 * 515 ( 1993 ) zeus collaboration : m. derrick et al .
, phys . lett .
* b316 * 412 ( 1993 ) a.j .
askew , j. kwieciski , a.d .
martin and p.j .
sutton , phys .
rev * d47 * 3775 ( 1993 ) a.j .
askew , k.golec-biernat , j.kwieciski , a.d .
martin and p.j .
sutton , krakow inst .
report no . 1653/ph ,
november 1993 , phys .
( in press ) m. ciafaloni , nucl . phys .
b*296 * 49 ( 1988 ) s. catani , f. fiorani and g. marchesini , phys .
b*234 * ( 1990 ) 339 ; nucl .
b*336 * 18 ( 1990 ) g. marchesini , proc . of workshop qcd at 200 tev `` , erice , june 1990 , ed .
l. cifarelli and yu.l .
dokshitzer ( plenum , new york ) p. 183
webber , proc . of the workshop physics at hera '' , eds.w .
buchmller and g. ingelman , desy 1992 , vol . 1
, p. 285 .
gribov , e.m . levin and m.g .
ryskin , phys . rep . * 100 * 1 ( 1983 ) k. charchula and e.m .
levin , proc . of workshop physics at hera " eds .
w. buchmller and g. ingelman , vol . 1 , p 223 ( 1992 ) e. laenen and e.m
. levin , j. phys .
g*19 * , 1582 ( 1993 ) j. kwieciski , a.d .
martin and p.j .
sutton , phys . rev . *
d44 * 2640 ( 1991 ) l.v .
gribov , yu.l .
dokshitzer , v.a .
khoze and s.i .
troyan , sov .
jetp * 67 * 1303 ( 1988 ) , phys .
b*202 * 276 ( 1988 ) a.d .
martin , w.j . stirling and r.g .
roberts , phys .
b306 * 145 ( 1993 ) j. kwieciski , a.d .
martin and p.j .
sutton , phys . rev . *
d46 * 921 ( 1992 ) j.r .
forshaw , p.n .
harriman and p.j .
sutton , nucl .
( in press ) ; j. phys .
g*19 * 1616 ( 1993 ) h1 collaboration , a. de roeck , to be published in proc . of int .
europhysics conf . on hephysics , marseille , 1993 , desy preprint 94005 ; j. feltesse , to be published in proc . of int
. conf . on deep - inelastic scattering , eilat , 1994 .
askew , j. kwieciski , a.d .
martin and p.j .
sutton , mod .
a*8 * 3813 ( 1993 ) ; phys .
d*49 * may 1st issue ( 1994 ) m. glck , e. reya and a. vogt , z. phys .
c*53 * 127 ( 1992 ) ; phys .
b*306 * 391 ( 1993 ) a.h .
mueller , nucl .
b ( proc .
suppl . ) * 18*c 125 ( 1990 ) ; j. phys .
g*17 * 1443 ( 1991 ) | the distribution of transverse energy , @xmath0 , which accompanies deep - inelastic electron - proton scattering at small @xmath1 , is predicted in the central region away from the current jet and proton remnants .
we use bfkl dynamics , which arises from the summation of multiple gluon emissions at small @xmath1 , to derive an analytic expression for the @xmath0 flow .
one interesting feature is an @xmath2 increase of the @xmath0 distribution with decreasing @xmath1 , where @xmath3 .
we perform a numerical study to examine the possibility of using characteristics of the @xmath0 distribution as a means of identifying bfkl dynamics at hera . |
the basic view that we have of the solar wind is that of a magnetofluid pervaded by fluctuations over a wide range of scales which are strongly modified by the effects of the dynamics during the expansion into the interplanetary medium .
these effects are more relevant within the inner heliosphere and on the ecliptic where the stream stream dynamics more strongly reprocesses the original plasma and the large velocity shears add new fluctuations to the original spectrum ( coleman , 1968 , roberts et al . , 1991 ) .
this scenario has reconciled the `` wave '' point of view proposed by belcher and davis ( 1971 ) , i.e. solar origin of the fluctuations , and the `` turbulence '' point of view , i.e. local generation due to velocity shears , proposed by coleman(1968 ) .
the first consequence of this scenario is that large fluctuations of solar origin containing energy interact non linearly with other fluctuations of local origin giving rise to an energy exchange between different scales , which can be interpreted as the usual energy cascade towards smaller scales in fully developed turbulence . as a matter of fact
, spacecraft observations have shown that the spectral slope of the power spectrum of these fluctuations changes with the radial distance from the sun ( bavassano et . , 1982 ,
denskat and neubauer , 1983 ) .
this behaviour was recognized ( tu et al . , 1984 ) as a clear experimental evidence that cascade processes due to
non linear interaction between opposite propagating alfvn waves were active in the solar wind with .
one of the consequences of this radial evolution was the observed radial decrease of the correlation of velocity and magnetic field fluctuations ( generally known as cross helicity , or alfvnicity)(roberts et al .
these observations finally answered to the question of whether the observed fluctuations were remnants of coronal processes or were dynamically created during the expansion .
however , successive theoretical models ( see review by tu and marsch , 1995 ) which tried to obtain the radial spectral evolution of the solar wind fluctuations had to deal with peculiarities of the observations that they could not reproduce within the framework of solely non linear interacting waves . the lack of a strict self similarity of the fluctuations and the consequent non applicability of strict scale invariance ( marsch and liu , 1993 ) , the strong anisotropy shown by velocity and field fluctuations ( bavassano et al . , 1982 ,
tu et al . , 1989 , roberts , 1992 ) , the different radial evolution of the minimum variance direction for magnetic field and velocity ( klein et al . , 1993 ) , the lack of equipartition between magnetic and velocity fluctuations ( matthaeus and goldstein , 1982 , bruno et al . , 1985 )
all contributed to suggest the idea that fluctuations could possibly be due to a mixture of propagating waves and static structures convected by the wind .
some kind of filamentary structure , similar to flux tubes , was firstly proposed by mccracken and ness ( 1966 ) and the observed spectral radial evolution of the large scale fluctuations has been attributed to the interaction of outward propagating alfvn waves with these structures ( tu and marsch , 1993 , bruno and bavassano , 1991 ; bavassano and bruno , 1992 ) .
incompressible magnetic structures were found by tu and marsch ( 1991)and magnetic fluctuations with a large correlation length parallel to the ambient magnetic field , suggested the idea of a quasi two dimensional , incompressible turbulence for which @xmath0 ( matthaeus et al . , 1990 ) .
thus , solar wind fluctuations are not isotropic and scale invariant , two of the fundamental hypotheses at the basis of k41 kolmogorov s theory ( 1941 ) .
this theory is based on an important statistical relation , which characterizes turbulent flows , between velocity increments @xmath1 , measured along the flow direction @xmath2 , and the energy transfer rate @xmath3 at the scale separation @xmath4 , that is @xmath5 or , more in general , @xmath6 .
if @xmath3 is constant , the previous relation simply reads @xmath7 and fluctuations are said to be _ self similar _ , and our signal is a simple fractal . however , as remarked by landau ( kolmogorov , 1962 , obukhov , 1962 ) , if @xmath3 statistically depends on scale due to the mechanism that transfers energy from larger to smaller eddies , @xmath3 will be replaced by @xmath8 and a new scaling has to be evaluated @xmath9 . expressing @xmath10 via a scaling relation with @xmath11
, we obtain @xmath12 and , consequently , @xmath13 where @xmath14 is generally a nonlinear function of @xmath15 .
this means that the global scale invariance required in the k41 theory would release towards a local scale invariance where different fractal sets characterized by different scaling exponents can be found .
one of the consequences of this lack of a universal scale invariance , directly observed in experimental tests , is that the shape of the probability density functions ( pdfs ) of the velocity increments at a given scale is not the same for each scale but roughly evolves from a gaussian shape , near the integral scale , to a distribution whose tails are much flatter than those of a gaussian , resembling a stretched exponential near the dissipation scale .
this means that the largest events , contained in the tails of the distribution , do not follow the gaussian statistics but show a much larger probability .
this phenomenon is also called intermittency and , in practice , fluctuations of a generic time series affected by intermittency , alternate intervals of very high activity to intervals of quiescence .
because of this lack of gaussianity , the study of the fluctuations based on conventional spectral analysis is strongly limited , and the second order moment of the distribution is not longer the limiting order . an alternative way for characterizing the fluctuations is to investigate directly the differences of a fluctuating field over all the possible spatial scales and look at moments of orders higher than 2 , adopting the so - called multifractal approach ( parisi and frisch , 1985 ) . a convenient statistical tool to perform this study is the so called @xmath16 order structure function ( sf ) defined as @xmath17 and @xmath18 is expected to scale as @xmath19 .
sfs are then computed for various orders as a function of all the possible scales and each order provides a value of the scaling exponent @xmath20 .
if observations show a non linear departure from the simple @xmath21 ( or @xmath22 for the mhd case ( carbone , 1993 ) ) this is an indication that intermittency is present .
this method was introduced for the first time in space plasma studies by burlaga ( 1991 ) who studied the exponents @xmath20 of structure functions based on voyager s observations of solar wind speed at 8.5 au .
this author found that , similarly to what is found in ordinary laboratory turbulent fluids , the exponent @xmath20 was not equal to @xmath23 , as expected in the k41 theory .
this exponent was found to scale non - linearly with the order @xmath15 and to be consistent with a variety of newer theories of intermittent turbulence , including kolmogorov
obukhov ( 1962 ) .
the first results obtained by burlaga ( 1991 ) and carbone et al .
( 1995 ) not only revealed the intermittent character of interplanetary magnetic field and velocity fluctuations but also showed an unexpected similarity to those obtained for laboratory turbulence ( anselmet et al.,1984 ) .
these results showed consistency between observations on scales of 1 au and laboratory observations on scales of meters , suggesting a sort of universality of this phenomenon , which was independent on scale .
while previous results referred to observations in the outer heliosphere , marsch and liu ( 1993 ) firstly investigated solar wind scaling properties in the inner heliosphere .
for the first time they provided some insights on the different intermittent character of slow and fast wind , on the radial evolution of intermittency and on the different scaling characterizing the three components of velocity .
they also concluded that the alfvnic turbulence observed in fast streams starts from the sun as self similar but then , during the expansion , decorrelates becoming more multifractal .
this evolution was not seen in the slow wind supporting the idea that turbulence in fast wind is mainly made of alfvn waves and convected structures ( tu and marsch , 1993 ) as already inferred by looking at the radial evolution of the level of cross helicity in the solar wind ( bruno and bavassano , 1991 ) . as we will see in the following ,
although the tools used in our analysis differ from those used by marsch and liu ( 1993 ) our results fully confirm their results but also add some more inferences on the radial evolution of solar wind intermittency . successively , several other papers tried to understand the phenomenon of intermittency in the solar wind looking for the best model which could fit the observations or could establish whether the observed scaling was closer to that shown by an ordinary fluid or rather by a magnetofluid as predicted by kolmogorov ( 1941 ) and kraichnan ( 1965 ) , respectively .
ruzmainkin et al . , ( 1995 ) studying fast wind data observed by ulysses developed a model of alfvnic turbulence in which they reduced the spectral index of magnetic field fluctuations by an amount depending on the intermittency exponent .
they found a close agreement with the expected kraichnan scaling for a magnetofluid ( @xmath24 ) and concluded that their results were consistent with a turbulence based on random phased alfvn waves ( kraichnan , 1965 ) .
tu et al . , ( 1996 ) re elaborated the tu ( 1988 ) model of developing turbulence including intermittency derived from the p - model of meneveau and sreenivasan ( 1987 ) .
they obtained a new expression for the scaling exponent that took into account that , for turbulence not fully developed , the spectral index is not defined yet .
carbone et al . , ( 1995 ) , for the first time adopted the extended self similarity ( ess ) concept ( benzi et al . , 1993 ) to interplanetary data collected by voyager and helios , and looked for differences in the scaling properties between interplanetary magnetofluid and ordinary fluid turbulence obtained in laboratory .
ess is a powerful method to easily recover the scaling exponent of the fluctuations exploiting the interdependency of the structure functions of various orders .
these authors concluded that , differences exist between scaling exponents in ordinary ( unmagnetized ) fluid flows and hydromagnetic flows .
horbury and balogh ( 1997 ) performed a comprehensive structure function analysis of ulysses data and concluded that interplanetary magnetic field fluctuations are more kolmogorov like rather than kraichnan like .
veltri and mangeney ( 1999 ) , adopting a method based on the discrete wavelet decomposition of the signal identified for the first time intermittent events .
successively , using conditioned structure - functions , they excluded any contribution from intermittent samples and were able to recover the scaling properties of the mhd fluctuations .
in particular , the radial component of the velocity displayed the characteristic kolmogorov slope @xmath21 while the other components displayed the kraichnan slope @xmath22 . all previous works dealt with the scaling exponents @xmath20 of the structure functions @xmath18 , aiming to show that they follow an anomalous scaling with respect to that expected from k41 theory for turbulent fluids .
this anomalous scaling is strictly related to the way probability distribution functions ( pdfs ) of the increments change with scale .
it is interesting to notice that if we consider fluctuations that follow a given scaling , say @xmath25 and introduce a change of scale , say @xmath26 ( @xmath27 ) , we end up with the following transformation @xmath28 .
the importance of this relation is that the statistical properties of the left and right hand side members are the same ( frisch , 1995 ) , i.e. @xmath29 .
this means that if @xmath30 is unique , the pdfs of the standardized variables @xmath31 reduces to a unique pdf highlighting the self similar ( fractal ) nature of the fluctuations . in other words ,
if all the pdfs of standardized fluctuations @xmath32 collapse to a unique pdf , fluctuations are not intermittent .
intermittency implies multifractality and , as a consequence , an entire range of values for @xmath30 .
castaing et al .
( 1990 ) developed a model based on the idea of a log normal energy cascade and showed that the non - gaussian behavior of the probability distribution functions ( pdf s ) at small scales can be represented by a convolution of gaussians whose variances are distributed according to a log - normal distribution whose width is represented , for each scale @xmath11 , by the parameter @xmath33 .
this model has been adopted , for the first time in the solar wind context , by sorriso et al . , ( 1999 ) to fit the departure from a gaussian distribution of the pdfs of solar wind speed and magnetic field fluctuations at small scales . as a matter of fact , marsch and tu ( 1994 )
had already shown that the pdfs closely resemble a gaussian distribution at large scales but , at smaller scales , their tails become more and more stretched as result of the fact that large events have a probability to happen larger than for a gaussian distribution .
their results showed that values of @xmath33 relative to magnetic field were higher than those relative to velocity throughout the inertial range , confirming that pdf s of magnetic field fluctuations are less gaussian than those relative to wind speed fluctuations ( marsch and tu , 1994 ) .
the same authors determined also the codimension of the most intermittent magnetic and velocity structures , suggesting that within slow wind intermittency is mainly due to compressive phenomena .
moreover , the use of techniques recently adopted in the context of solar wind turbulence ( veltri and mangeney , 1999 , bruno et al .
2001 ) based on wavelet decomposition allowed to identify those events causing intermittency .
those events were identified as either compressive phenomena like shocks or planar sheets like tangential discontinuities separating contiguous regions characterized by different total pressure and bulk velocity , possibly associated to adjacent flux tubes . lately , padhye et al . , ( 2001 ) used the castaing approach to describe directly the pdfs of the fluctuations of the overall interplanetary magnetic field components .
these authors concluded that all the components followed a rather gaussian statistics but they were not able to relate their results to those obtained by marsch and tu ( 1994 ) and sorriso et al .
( 1999 ) who compared pdfs for different time scales . as a matter of fact ,
padhye and co - workers referred to fluctuations respective to the mean field and not increments as it was done in the previous mentioned studies and in the present study . in this paper , we base our analysis on the concept of intermittency as given by frisch ( 1995 ) , following which a random function @xmath34 is said to be intermittent if the flatness @xmath35 grows without bound as we filter out the lowest frequency components of our signal and consider only smaller and smaller scales .
thus , we will define a given time series to be intermittent if @xmath36 continually grows at smaller scales and , we will define the same time series to be more intermittent if @xmath36 grows faster .
moreover , if @xmath36 remains constant within a certain range of scales , it will indicate that those scales are not intermittent but simply self similar and , a value of @xmath37 ( 3 is the value expected for a gaussian ) would simply indicate that those scales do not have a gaussian statistics .
this is a simpler way than that used by sorriso et al .
( 1999 ) to look at the behaviour of the flatness to infer the intermittency character of the fluctuations but , what we gain in simplicity we loose in effectiveness to quantify the degree of intermittency and , we will only be able to evaluate whether a given sample is more or less intermittent than another one . in the following sections we will analyze and discuss the radial evolution of intermittency in the inner heliosphere and on the ecliptic plane evaluating the behavior of @xmath36 as previously illustrated .
the present analysis was performed using plasma and magnetic field data recorded by helios 2 during its first solar mission in 1976 when the s / c repeatedly observed the same corotating stream at three different heliocentric distances on the ecliptic plane , during three consecutive solar rotations . in order to compare intermittency between high and low speed plasma , low speed regions ahead of each corotating high speed stream ,
were also studied .
the three streams , named `` 1 '' , `` 2 '' and `` 3 '' , respectively , can be identified in figure 1 where the wind speed profile and the spacecraft heliocentric distance are shown for the whole helios 2 primary mission to the sun .
the exact location of the selected intervals , lasting 2 days each , is shown by the rectangles drawn on the data profile .
beginning and end of each time interval are shown in table 1 where , we also show the average heliocentric distance , the average wind speed , the angle @xmath38 between magnetic and velocity vectors , the angle @xmath39 between magnetic field vector and radial direction and the angle @xmath40 between velocity vector and radial direction . while the velocity vector is always closely aligned with the radial direction , magnetic field vector generally follows the expected archimedean spiral configuration although this agreement is larger during fast wind than during slow wind time intervals .
the data set is made of 81 sec averages of magnetic and plasma observations recorded in solar
ecliptic reference system se where , the @xmath41 axis is oriented towards the sun , the @xmath42 axis lies on the ecliptic and it is oriented opposite to the s / c direction of motion and , the @xmath43 axis completes the right handed reference system .
these fast wind streams are notorious for being dominated by alfvn waves and have been widely studied since they offer a unique opportunity to observe the radial evolution of mhd turbulence within the inner heliosphere ( for a rather complete review of existing literature related to this topic see tu and marsch , 1995 ) .
the aim of the present study is to investigate the behavior of magnetic field and wind velocity intermittency as a function of heliocentric distance and type of wind ( i.e. fast and slow ) .
although intermittency refers to the statistical behavior of the fluctuations in the spatial domain , it can be estimated from measurements made in the temporal domain simply adopting the taylor s frozen in hypothesis .
this assumption , which is fully acceptable within the usual conditions of strongly supersonic and super alfvnic solar wind , allows to treat , with good approximation , each fluctuation as an eddy and spatial @xmath11 and temporal @xmath44 coordinates can be mutually exchanged via the relation @xmath45 where @xmath46 is the solar wind bulk speed . in order to study intermittency we computed the following estimator of the flatness factor @xmath36
@xmath47where @xmath48 is the scale of interest and @xmath49 is the sf of order @xmath15 of the generic function @xmath50 .
this definition slightly differs from that given by frisch ( 1995 ) since we compute the factor @xmath36 for each single scale while frisch calculates @xmath36 using a high pass filter whose cutoff frequency is repeatedly shifted towards higher and higher frequencies each time . however , in both cases a given function is considered intermittent if the factor @xmath36 increases when considering smaller and smaller scales or , equivalently , higher and higher frequencies .
a vector field , like velocity and magnetic field , encompasses two distinct contributions , a compressive one due to intensity fluctuations that can be expressed as @xmath51 and a directional one due to changes in the vector orientation @xmath52 obviously , relation [ directional ] takes into account also compressive contributions and the expression @xmath53 is always true . in the following
we will study the flatness factor @xmath36 obtained from sfs computed for both compressive @xmath54 and directional @xmath55 fluctuations . as regards this last quantity , we like to stress that we verified that magnetic sector changes do not appreciably influence its value and that in this study only interval ( 72:0074:00 ) contains a magnetic sector change . comparing the radial dependence of these two quantities for fast and slow wind and for magnetic field and velocity will turn out to be useful to better interpret the radial evolution of intermittency as observed in the solar wind mhd turbulence .
our analysis will be based on the following definitions : 1 ) a given time series is defined intermittent if the factor @xmath36 monotonically increases moving from larger to smaller scales , 2 ) the same time series is defined more intermittent than another one if @xmath36 begins to increase at larger scales since , following castaing et al .
( 1990 ) , this implies a larger inertial range and , consequently , a larger number of steps along the cascade with intermittency increasing at each step , 3)if @xmath36 starts to increase at the same scale for two different time series , we will consider more intermittent the one for which @xmath36 grows more rapidly .
moreover , we like to remind that a gaussian statistics would show values of @xmath36 close to 3 for all scales , indicating the self similar character of our fluctuations . however , if @xmath36 fluctuates around a value somewhat different from 3 our fluctuations are still self similar although not gaussian . anyhow , in both cases these fluctuations are not considered intermittent .
thus , our definition of intermittency will be limited to a qualitative definition rather than quantitative since the aim of the present work is only to compare the radial evolution of intermittency for different solar wind parameters and within different solar wind conditions .
values of @xmath36 for both scalar and vector differences for magnetic field as a function of temporal scale @xmath56 expressed in seconds are shown in figure 2 .
the factor @xmath36 has been computed for slow ( left column ) and fast ( right column ) wind and for three distinct radial distances as indicated by the different symbols used in the plots .
in addition , errors relative to each value of @xmath36 are also shown .
it is readily seen that magnetic field fluctuations in both slow and fast wind are intermittent since @xmath36 increases at small scales .
values of @xmath36 for compressive fluctuations within slow wind ( panel a ) start to increase well beyond @xmath57 sec reaching values larger than 20 at the smallest scale .
moreover , there is no evidence for any radial dependence since all the curves overlap to each other within the error bars . on the contrary ,
compressive fluctuations for fast wind ( panel b ) show a clear radial dependence . as a matter of fact ,
the three curves intersect each other at large scales down to @xmath58 sec but clearly separate at smaller scales indicating that intermittency increases with the radial distance from the sun .
moreover , since within slow wind @xmath36 starts to grow at larger scales and reaches higher values at small scales , we can say that magnetic compressive fluctuations in slow wind , at least at 0.3 and 0.7 au , are more intermittent than those within fast wind .
slow wind directional fluctuations ( panel c ) are also rather intermittent since @xmath36 starts to increase around @xmath59 sec , at frequencies slightly higher than for slow wind compressive fluctuations
. however , these fluctuations are less intermittent than compressive fluctuations in the same type of wind since @xmath36 increases more slowly at small scales .
moreover , there is no radial dependence .
panel d shows the behavior of @xmath36 for directional fluctuations in fast wind .
also in this case as for panel b , there is a clear radial dependence of intermittency on the radial distance .
the flatness factor @xmath36 remains approximately constant and rather similar for the three distances at large scales down to @xmath60 sec and then increases more rapidly for larger heliocentric distances .
thus , our sample at 0.9 au is more intermittent than that at 0.7 au , which , in turn , is more intermittent than that at 0.3 au . considering that the scales at which @xmath36 starts to increase is only around @xmath61 sec and that the values reached at small scales are lower , these fluctuations are less intermittent than the corresponding ones within slow wind and the compressive ones within the same fast wind .
moreover , the fact that in panel d @xmath36 starts to increase at much smaller scales than in slow wind ( panel c ) , is strongly indicative that the inertial range in this case is much less extended as we already know from the existing literature ( see review by tu and marsch , 1995 ) .
results relative to velocity fluctuations are shown in figure 3 in the same format adopted in the previous figure .
values of @xmath36 for slow speed compressive fluctuations start to increase around @xmath57 sec ( panel a )
. however , the three curves intersect each other various times along the whole range of scales showing that there is no clear radial dependence although , the smallest scale would suggest some radial evolution which , in addition , would be opposite to what is observed in fast wind .
however , the large associated error bars do not allow us to draw any realistic conclusion .
taking into account that the the scale at which @xmath36 starts to increase is of the same order of that relative to magnetic field compressive fluctuations in slow wind but @xmath36 reaches much lower values at small scales , we conclude that velocity compressive fluctuations are less intermittent than magnetic compressive fluctuations in slow wind .
panel b , relative to velocity compressive fluctuations in fast wind , shows a clear radial dependence of @xmath36 .
the three curves start to increase around @xmath61 sec and separate at smaller scales .
then , intermittency increases from 0.3 to 0.9 au since @xmath36 increases more rapidly for larger heliocentric distances .
this result is similar to what we observed for magnetic compressive fluctuations in fast wind although the overall intermittency in this case is much reduced taking into account that @xmath36 starts to increase at much smaller scales and reaches smaller values .
panel c shows results relative to velocity directional fluctuations in slow wind .
these curves , although less stable than the corresponding curves relative to magnetic field ( figure 2c ) , show a very similar behavior and no hint for a possible radial dependence .
also panel d , where we report values of @xmath36 for velocity directional fluctuations in fast wind , shows results very similar to those shown for magnetic field in figure 2d to the extent that these two sets of curves overlap to each other , within the error .
this last result , as it will be discussed later on in this paper , clearly derives from the strong contribution due to alfvnic fluctuations populating the fast corotating streams that we selected ( bruno et al . , 1985 ) . differently from magnetic field fluctuations ,
velocity directional fluctuations seem to be only slightly less intermittent than compressive fluctuations .
although , other authors ( marsch and liu , 1993 , marsch and tu , 1994 ) already addressed the study of the radial evolution of intermittency for magnetic field and velocity components , we like to provide a complete picture of this radial dependence adding a study performed in the mean field coordinate system ( mf , hereafter ) which , for magnetic field , is more appropriate than the usual rtn or se coordinate systems . as a matter of fact , the large scale interplanetary magnetic field configuration breaks the spatial symmetry and introduces a preferential direction along the mean field . as a consequence ,
a natural reference system is the one for which one of the components , that we call @xmath62 , is along the mean field @xmath63 outwardly oriented , and the other two are perpendicular to this direction . in our case
we chose one of the two perpendicular components @xmath64 to be perpendicular to the plane identified by @xmath63 and the mean solar wind velocity @xmath65 , so that @xmath66 and , the remaining direction @xmath67 descends from the vector product @xmath68 , where the symbol @xmath69 indicates a unitary vector . in the top panel of figure 4
, we show values of @xmath36 for the three magnetic components in se reference system ( i.e. @xmath70 , @xmath71 and @xmath72 ) at the three different heliocentric distances previously chosen and , in the bottom panel , we show the components in the mf reference system ( i.e. @xmath62 , @xmath73 and @xmath64)for the same heliocentric distances . in the top
panel @xmath36 increases for all the components as the radial distance increases .
while at 0.9 au the three components show the same behavior , at 0.3 and 0.7 au the curve relative to @xmath70 runs above the other two curves .
since the distance between these curves slightly increases at small scales we might conclude that the radial component is slightly more intermittent than the other two components .
unfortunately , this conclusion is not corroborated by the size of the errors associated to each point , which are quite large .
the bottom panel shows results relative to the mf reference system . at 0.3 and 0.7 au
there is not much difference with the situation discussed in the previous panel because the orientation of the two reference systems is not very different either , given that the magnetic field is almost radially oriented ( see table 1 ) . on the contrary , moving to 0.9 au and comparing these results with those obtained in the other reference system , we clearly observe a decrease of @xmath36 for the two perpendicular components and an increase for the parallel component .
since the three curves lie on the same level at large scales and end up with remarkable different values at small scales , we conclude that the component parallel to the local magnetic field is more intermittent than the perpendicular components
. moreover , the two perpendicular components are less intermittent in the mf reference system than in se .
this means that in the mf reference system we enhance on one hand the stochastic character of the fluctuations perpendicular to the local field direction and , on the other hand , the coherent character of the fluctuations along the local field direction . in figure 5 we show , for the slow wind , the same elements discussed in the previous figure .
the much steeper behavior of these curves suggests that , generally , slow wind is more intermittent than fast wind .
moreover , especially at 0.3 au , there is a tendency for both @xmath70 , in the top panel , and @xmath62 , in the bottom panel , to be steeper than the other components at small scales , suggesting a higher intermittency but , this tendency is not confirmed at larger heliocentric distances . as discussed in the following , the reason for this appreciable different behavior of @xmath70 and @xmath62 might be due to the fact that so close to the sun the contribution due to alfvnic fluctuations , mainly acting on the perpendicular components , is not negligible even within slow wind ( bruno et al . , 1991 ) . as a matter of fact
, the stochastic nature of the fluctuations due to alfvn waves tends to make more gaussian the pdfs of magnetic and velocity fluctuations perpendicular to the mean field direction . in conclusion
, an overall view reveals that the behavior of @xmath36 within slow wind is not very sensitive to this change of reference system . for sake of completeness
we have rotated into the mf reference system also velocity fluctuations although this reference system is not the most appropriate for this parameter given that the wind expands radially .
the two panels of figure 6 show , for the three heliocentric distances , the behavior of @xmath36 for fast wind velocity fluctuations in the se reference system and in the mf reference system , respectively .
we like to remark that in the se reference system @xmath74 resembles very closely the behavior of the fast wind speed shown in figure 3 since the average wind velocity vector is always close to the radial direction
. moreover , the two perpendicular components , @xmath75 and @xmath76 in se and @xmath77 and @xmath78 in mf , at 0.3 and at 0.7 au closely recall the behavior of the corresponding magnetic components within fast wind .
however , the presence of a large plateau in the central part of @xmath74 and @xmath79 makes it more difficult to estimate the degree of intermittency of these components with respect to the perpendicular ones .
at 0.9 au , due to a weaker stationarity in the data , the situation looks even more complex and does not allow to estimate which component is the most intermittent one .
@xmath36 remarkably increases with distance at small scales for all the components in both reference systems .
however , as expected , the rotation into the mf reference system does not have a large influence at 0.3 au but it causes a general increase of @xmath36 at 0.9 au .
the enhancement is such that the two perpendicular components have the same behavior , and differences with the parallel component become appreciably smaller .
this is due to the fact that in this reference system the fluctuations of the components are not longer independent from each other as it would be in se reference system . finally ,
in figure 7 , we show results relative to the slow wind in the same format of the previous figure . here , the very confused behavior of the curves and the large associated errors , especially at 0.7 and 0.9 au , suggest a rather weak stationarity of the data and make it difficult to compare the behavior of different components .
a general comment that we can easily make is that these curves are much steeper and start to increase at much larger scales than in fast wind . as a consequence ,
velocity components in slow wind are generally more intermittent than in fast wind .
in addition , the rotation from se to mf does not influence much our results , as expected .
however , it is such that the behavior of the three velocity components at 0.3 au looks more similar to that of the corresponding magnetic components ( figure 5 , lower panel ) .
this suggests that alfvn waves , although less relevant than in fast wind , might play a role even in this sample of slow wind .
we studied the radial dependence of solar wind intermittency looking at magnetic field and velocity fluctuations between 0.3 and 1 au .
in particular , we analyzed compressive and directional fluctuations for both fast and slow wind .
our analysis exploits the property that probability distribution functions of a fluctuating field affected by intermittency become more and more peaked at smaller and smaller scales .
since the peakedness of a distribution is measured by its flatness factor we studied the behavior of this parameter at different scales to estimate the degree of intermittency of our time series .
our general results can be summarized in the following points : \1 ) magnetic field fluctuations are more intermittent than velocity fluctuations ; \2 ) compressive fluctuations are more intermittent than directional fluctuations ; \3 ) slow wind intermittency does not show radial dependence ; \4 ) fast wind intermittency , for both magnetic field and velocity , clearly increases with distance .
\5 ) magnetic and velocity fluctuations have a rather gaussian behavior at large scales , as expected , regardless of type of wind or heliocentric distance .
point 4 is particularly interesting because we found that both compressive and directional fluctuations become more intermittent with distance . as a matter of fact
, if we think of relations [ compressive ] and [ directional ] we easily realize that while intermittency of directional fluctuations can be fully uncompressive , it is not possible to avoid that intermittency of compressive fluctuations contaminates directional fluctuations . in the latter case
, the limiting condition would be the same intermittency level for both kind of fluctuations .
thus , intermittency of directional fluctuations contains also contributions due to compressive fluctuations .
this distinction plays an important role in discussing our results since the intermittency character of directional fluctuations reflects the contribution of both compressive phenomena and uncompressive fluctuations like alfvn waves . now , there are at least two questions that we should address : 1 ) why directional fluctuations are always less intermittent than compressive fluctuations ? and , 2 ) why only fast wind shows radial evolution ?
we can explain our observations simply assuming that the two major ingredients of interplanetary mhd fluctuations are compressive fluctuations due to a sort of underlying , coherent structure convected by the wind and stochastic alfvnic fluctuations propagating in the wind .
the coherent nature of the first ingredient would contribute to increase intermittency while the stochastic character of the second one would contribute to decrease it .
if this is the case , coherent structures convected by the wind would contribute to the intermittency of compressive fluctuations and , at the same time , would also produce intermittency in directional fluctuations .
however , since directional fluctuations are greatly influenced by alfvnic stochastic fluctuations , their intermittency will be more or less reduced depending on the amplitude of the alfvn waves with respect to the amplitude of compressive fluctuations .
thus , compressive fluctuations would always be more intermittent than directional fluctuations . before addressing the second question , we like to recall that several papers have already shown ( see review by tu and marsch , 1995 ) that slow wind alfvnicity does not evolve with increasing the radial distance from the sun . as a matter of fact , power spectra exhibit a spectral index close to that of kolmogorov and a rather good equipartition between _ inward _ and _ outward _ modes ( tu et al . , 1989 ) .
thus , once the inertial range is established , the alfvnicity of the fluctuations freezes to a state that , successively , is convected by the wind into the interplanetary space without major changes ( bavassano et al . , 2001 ) . on the other hand , within fast wind
, turbulence is dominated by outward propagating alfvn waves , which strongly evolve in the inner heliosphere becoming weaker and weaker during the wind expansion , to the extent that at 1 au , on the ecliptic plane , their amplitude is much reduced and of the order of that of inward propagating alfvn waves . at that point
, the resulting alfvnicity resembles the one already observed in the slow wind close to the sun ( tu and marsch , 1990 ) . keeping this in mind , taking into account that convected structures experience a much slower radial evolution because they do not interact with each other non linearly as alfvn waves do , considering that alfvn waves are mainly found in fast rather than in slow wind , it comes natural to expect that intermittency would radially evolve within fast rather than slow wind . obviously
, this would explain why directional fluctuations become more intermittent only within fast wind but would not explain why also compressive fluctuations become more intermittent within fast wind .
in reality , if we consider that compressive events cause intermittency ( veltri and mangeney , 1991 , bruno et al .
, 2001 ) , we might ascribe this different behavior to the fact that fast wind becomes more and more compressive with radial distance while the compressive level of slow wind remains approximately the same , as shown by marsch and tu , ( 1990 ) .
our analysis performed on the components can also help to understand , although partially , the topology of these convected structures . in se reference system , fluctuations along the radial component are more intermittent than those perpendicular to it as already found by marsch and liu ( 1993 ) , although this feature , especially for the magnetic field , tends to vanish around 1 au .
the reason is that perpendicular components are more influenced by alfvnic fluctuations and as a consequence their fluctuations are more stochastic and less intermittent .
this effect largely reduces during the radial excursion mainly because the se reference system is not the most appropriate one for studying magnetic field fluctuations . as a matter of fact
, the presence of the large scale spiral magnetic field breaks the spatial symmetry introducing a preferential direction parallel to the mean field .
consequently , we showed , that if we rotate our magnetic data into the mean field reference system , especially at 0.9 au , the intermittency of the perpendicular components decreases and that of the parallel component increases .
moreover , the two perpendicular components show a remarkable similar behavior as expected if they experience alfvnic fluctuations . on the other hand , results obtained on velocity fluctuations suggest that a reference system with an axis parallel to the radial direction looks more appropriate to perform a similar study showing that the radial component seems to be the most intermittent component .
one further observation is that generally most of the curves relative to velocity fluctuations came out to be less stable than those relative to magnetic field fluctuations and affected by larger errors due to a weaker stationarity with respect to magnetic field data . finally , our results can not establish whether magnetic and velocity structures causing intermittency are convected directly from the source regions of the solar wind or they are locally generated by stream stream dynamic interaction or , as an alternative view would suggest ( primavera et al . , 2002 )
, they are locally created by parametric decay instability of large amplitude alfvn waves .
probably all these origins coexist at the same time and confirm that in any case the the radial dependence of the intermittency of interplanetary fluctuations is strongly related to the turbulent radial evolution of their spectrum . 1.5truecm we thank f. mariani and n. f. ness , pi s of the magnetic experiment and , h. rosenbauer and r. schwenn , pi s of the plasma experiment onboard helios 1 and 2 , for allowing us to use their data .
we also thank both referees for their valuable comments and suggestions .
anselmet , f. , gagne , y. , hopfinger , e. j. , and antonia , r.a . , 1984 , high order velocity structure functionsin turbulent shear flows , j. fluid mech . , 140 , 6389 bavassano , b. , m. dobrowolny , f. mariani , and n.f . ness , 1982 , radial evolution of power spectra of interplanetary alfvnic turbulence , j. geophys
86 , 36173622 bavassano , b. , m. dobrowolny , g. fanfoni , f. mariani , and n.f .
ness , 1982 , statistical properties of mhd fluctuations associated with high speed streams from helios2 observations , solar phys .
, 78 , 373384 b. bavassano , and r. bruno , 1992 , on the role of interplanetary sources in the evolution of low - frequency alfvnic turbulence in the solar wind , j. geophys .
res . , 97 , 1912919137 bavassano , b. , pietropaolo , e. , bruno r. , 2001 , radial evolution of outward and inward alfvnic fluctuations in the solar wind : a comparison between equatorial and polar observations by ulysses , j. geophys . res . , 106 , 1065910668 belcher , j. w. , and l. davis , jr . , 1971 , large amplitude alfvn waves in the interplanetary medium , 2 , j. geophys
76 , 35343563 .
benzi , r. , ciliberto , s. , tripiccione , r. , baudet , c , massaioli , f. , succi , s. , 1993 , extended sel similarity in turbulent flows , phys .
e 48 , r29 bruno , r. , bavassano , b. and villante , u. , 1985 , evidence for long period alfvn waves in the inner solar system , j. geophys .
90 , 43734377 .
bruno , r. and bavassano , b. , 1991 , origin of low cross
helicity regions in the inner solar wind , j. geophys .
96 , 78417851 .
bruno , r. , v. carbone , p. veltri , e. pietropaolo and b. bavassano , 2001 , identifying intermittent events in the solar wind , planetary space sci . , 49 , 12011210 .
burlaga , l. , 1991 , intermittent turbulence in the solar wind , j. geophys .
96 , 58475851 .
carbone , v. , 1993 , cascade model for intermittency in fully developed magnetohydrodynamic turbulence , phys .
71 , 1546 - 1548 .
carbone , v .
, veltri , p. , bruno , r. , 1995 , experimental evidence for differences in the extended self
silimarity scaling laws between fluid and magnetohydrodynamic turbulent flows , phys .
75 , 31103120 .
castaing , b. , gagne , y. , and hopfinger , 1990 , velocity probability density functions of high reynolds number turbulence , physica d 46 , 177200 .
coleman , p. j. , 1968 , turbulence , viscosity and dissipation in the solar wind plasma , astrophys .
j. , 153 , 371388 .
denskat , k. u. , and neubauer , f. m. , 1983 , observations of hydromagnetic turbulence in the solar wind , solar wind v , edited by m. neugebauer , nasa conf .
publ . , cp2280 , 8191 .
frisch , u. , 1995 , turbulence : the legacy of a. n. kolmogorov , cambridge university press horbury , t. a. , balogh , a. , forsyth , r. j. , smith , e. , 1996 , magnetic field signatures of unevolved turbulence in solar polar flows , j. geophys .
101 , 405413 .
l. klein , r. bruno , b. bavassano and h. rosenbauer , 1993 , anisotropy and minimum variance of magnetohydrodynamic fluctuations in the inner heliosphere , j. geophys .
res . , 98 , 1746117466 kolmogorov , a. n. , 1941 , the local structure of turbulence in incompressible viscous fluid for very large reynolds numbers , c. r. akad .
sssr 30 , 301 .
kolmogorov , a. n. , 1962 , a refinement of previous hypotheses concerning the local structure of turbulence in a viscous incompressible fluid at high reynolds number , _ j. fluid mech .
_ , 13 , 8285 kraichnan , r. h. , 1965 , inertial range spectrum of hydromagnetic turbulence , phys .
fluids 8 , 13851387 .
marsch , e. and tu , c .- y , 1990 , on the radial evolution of mhd turbulence in the inner heliosphere , j. geophys .
95 , 82118229 .
marsch , e. , and liu , s. , 1993 , structure functions and intermittency of velocity fluctuations in the inner solar wind , ann .
geophysicae 11 , 227238 .
marsch , e. and tu , c .- y , 1994 ,
non gaussian probability distributions of solar wind fluctuations , ann .
geophysicae 12 , 11271138 .
matthaeus , w. h. , goldstein , m. l. , 1982 , measurements of the rugged invariants of magnetohydrodynamic turbulence in the solar wind , j. geophys .
res . , 87 , 60116028 matthaeus , w. h. , goldstein , m. l. , roberts , d. a. , 1990 , evidence for the presence of quasi two dimensional nearly incompressible fluctuations in the solar wind , j. geophys . res . 95 , 2067320683 .
mccracken , k.g .
, and n.f .
ness , 1966 , the collimation of cosmic rays by the interplanetary magnetic field , j. geophys .
71 , 33153332 .
meneveau , c. , and sreenivasan , k. r. , 1987 , simple multifractal cascade model for fully developed turbulence , phys .
59 , 14241427 .
obukhov , 1962 , some specific features of atmospheric turbulence ,
_ j. fluid mech .
_ , 13 , 7781 padhye , n. s. , smith , c. w. and matthaeus , w. h.,2001 , distribution of magnetic field components in the solar wind plasma , j. geophys .
9106 , 18635 - 18650 .
parisi , g. and , u. frisch , 1985 , on the singularity structure of fully developed turbulence , in _ turbulence and predictability in geophysical fluid dynamics , proceed .
school of physics e . fermi , 1983 , varenna , italy _ , 8487 , eds .
m. ghil , r. benzi and g. parisi , north holland , amsterdam primavera , l. , f. malara , and p. veltri , 2002 , parametric instability in the solar wind : numerical study of the nonlinear evolution , paper presented at solar wind 10 conference , 1721 june 2002 , pisa , italy roberts , d.a . ,
m. goldstein , l. klein , and w. matthaeus , 1987 , origin and evolution of fluctuations in the solar wind : helios observations and helios - voyager comparisons , j. geophys .
, 92 , 1202312035 .
roberts , d. a. , m. l. , goldstein , w. h. matthaeus and s. gosh , 1991 , mhd simulation of the radial evolution and stream structure of solar wind turbulence , phys .
lett . , 67 , 37413765 . roberts , d. a.,1992 , observation and simulation of the radial evolution and stream structure of solar wind turbulence , in e. marsch and r. schwenn ( eds ) , solar wind seven , cospar colloquia series vol .
3 , pergamon press , oxford , 533538 ruzmaikin , a. , feynman , j. , goldstein , b. , and balogh , a. , 1995 , intermittent turbulence in solar wind from the south polar hole , j. geophys .
100 , 33953404 .
sorriso valvo , l. , carbone , v. , veltri , p. , consolini , g. , bruno , r. , 1999 , intermittency in the solar wind turbulence through probability distribution functions of fluctuations , geophys .
26 , 18011804 .
tu , c. -y . ,
pu , z .- y , and wei , f .- s , 1984 ,
the power spectrum of interplanetary alfvnic fluctuations : derivation of its governing equation and its solution , j. geophys .
89 , 96959702 .
tu , c. -y .
, 1988 , the damping of interplanetary alfvnic fluctuations and the heating of the solar wind , j. geophys .
93 , 720 .
tu , c .- y and marsch , e. , 1990 , evidence for a `` background '' spectrum of the solar wind turbulence in the inner heliosphere , j. geophys .
95 , 43374341 .
tu , c .- y and marsch , e. , and thime , k. m. , 1989 basic properties of solar wind mhd turbulence analysed by means of elssser variables , j. geophys .
95 , 1173911759 .
tu , c .- y and marsch , e. , 1993 , a model of solar wind fluctuations with two components : alfvn waves and convective structures , j. geophys .
98 , 12571276 .
tu , c .- y and marsch , e. , 1995 , mhd structures , waves and turbulence in the solar wind : observations and theories , space sci .
, 73 , 1210 tu , c .-
y , marsch , e. , rosenbauer , h. , 1996 , an extended structure function model and its aplication to the analysis of solar wind intermittency properties , ann .
14 , 270285 .
veltri , p. , and mangeney , a. , 1999 , in solar wind ix , edited by s. habbal , aip conf .
, 543546 .
.from left to right : time interval in dd : hh , heliocentric distance in au , average wind velocity in km / s , angle between field and velocity vectors , angle between field vector and radial distance and , finally , angle between velocity vector and radial distance . [ cols="^,^,^,^,^,^",options="header " , ]
| we analyzed intermittency in the solar wind , as observed on the ecliptic plane , looking at magnetic field and velocity fluctuations between 0.3 and 1 au , for both fast and slow wind and for compressive and directional fluctuations .
our analysis focused on the property that probability distribution functions of a fluctuating field affected by intermittency become more and more peaked at smaller and smaller scales .
since the peakedness of a distribution is measured by its flatness factor we studied the behavior of this parameter for different scales to estimate the degree of intermittency of our time series .
we confirmed that both magnetic field and velocity fluctuations are rather intermittent and that compressive magnetic fluctuations are generally more intermittent than the corresponding velocity fluctuations .
in addition , we observed that compressive fluctuations are always more intermittent than directional fluctuations and that while slow wind intermittency does not depend on the radial distance from the sun , fast wind intermittency of both magnetic field and velocity fluctuations clearly increases with the heliocentric distance .
we propose that the observed radial dependence can be understood if we imagine interplanetary fluctuations made of two main components : one represented by coherent , non propagating structures convected by the wind and , the other one made of propagating , stochastic fluctuations , namely alfvn waves . while the first component tends to increase the intermittency level because of its coherent nature , the second one tends to decrease it because of its stochastic nature . as the wind expands , the alfvnic contribution is depleted because of turbulent evolution and , consequently , the underlying coherent structures convected by the wind , strengthen further on by stream stream dynamical interaction , assume a more important role increasing intermittency , as observed . obviously , slow wind does nt show a similar behavior because alfvnic fluctuations have a less dominant role than within fast wind and the alfvnicity of the wind has already been frozen by the time we observe it at 0.3 au .
finally , our analysis suggests that the most intermittent magnetic fluctuations are distributed along the local interplanetary magnetic field spiral direction while , those relative to wind velocity seem to be located along the radial direction . |
for an @xmath4 matrix @xmath5 over the finite field @xmath6 the * weight * of @xmath5 , denoted @xmath7 , is the number of non - zero entries . in the hamming metric on matrices
it is the distance from @xmath5 to @xmath8 .
there is some relationship between the rank and the weight of a matrix .
for example , if @xmath9 , then @xmath10 , and if @xmath11 , then @xmath12 . on the other hand , there are matrices of rank one and maximal weight @xmath13 , such as a matrix with every entry a one . in this article
we determine the average weight of rank @xmath2 matrices in terms of @xmath2 , @xmath0 , @xmath1 , and @xmath3 . without fixing the rank ,
the average weight of @xmath4 matrices is @xmath14 and the weight has a binomial distribution .
however , the full probability distribution of the weight for matrices of rank @xmath2 is yet to be determined .
the tools are those of elementary combinatorics and linear algebra .
nothing special is used from the theory of finite fields other than the understanding that the fundamental ideas of linear algebra work over all fields and not just the real or complex numbers .
we need a modest amount of background material .
we use three basic formulas .
the number of ordered @xmath2-tuples of linearly independent vectors in @xmath15is @xmath16 the first vector is any non - zero vector and each succeeding vector must avoid the span of the previous vectors .
the number of @xmath2-dimensional subspaces of @xmath15 is given by the @xmath3-binomial coefficient @xmath17 the numerator is the number of bases of all @xmath2-dimensional subspaces , while the denominator is the number of bases of any given subspace .
the number of @xmath4 matrices of rank @xmath2 is @xmath18 for a fixed @xmath2-dimensional subspace @xmath19 , the number of matrices with @xmath20 as the column space is equal to the number of @xmath21 matrices of rank @xmath2 .
such a matrix is given by the @xmath2 linearly independent row vectors of length @xmath1 .
the number of those is given by formula 1 .
the number of @xmath2-dimensional subspaces of @xmath22 is @xmath23 and the product is the number of rank @xmath2 matrices given in the first line . by the same reasoning , the second line counts the number of @xmath24 matrices of rank @xmath2 , which is the same .
a special case of formula 3 is worth noting .
the number of invertible @xmath25 matrices is @xmath26
the average weight of a rank @xmath2 matrix is the sum of the average weights of the entries , and the average weight of the @xmath27 entry is the probability that the entry is not zero : @xmath28 the probability that @xmath29 for a rank @xmath2 matrix @xmath5 is the same for all @xmath30 and @xmath31 .
the probability that the @xmath27 entry is not zero is the quotient whose numerator is the number of matrices @xmath5 of rank @xmath2 with @xmath32 , and whose denominator is the number of matrices of rank @xmath2 .
consider the map on the @xmath33 matrices that switches rows 1 and @xmath30 and switches columns 1 and @xmath31 .
this map preserves rank and gives a bijection between the subset of matrices of rank @xmath2 with a non - zero in the 1,1 location and the subset of matrices of rank @xmath2 with a non - zero in the @xmath34 location .
thus , @xmath35 .
call this common value * the average weight per entry*. now we focus on the upper left corner of the matrices of rank @xmath2 .
our analysis depends on the reduced row echelon form .
we recall the definition @xcite .
a rectangular matrix is in _ row echelon form _ if it has the following three properties : 1 .
all non - zero rows are above any rows of all zeros .
each leading entry of a row is in a column to the right of the leading entry of the row above it .
all entries in a column below a leading entry are zero . if a matrix in echelon form satisfies the following additional conditions , then it is in _ reduced row echelon form _ :
the leading entry in each nonzero row is 1 2 .
each leading 1 is the only non - zero entry in its column .
the @xmath21 matrices in reduced row echelon form correspond bijectively with the @xmath2-dimensional subspaces of @xmath15 .
the rows of the matrix give a basis of the corresponding subspace .
when an @xmath36 matrix @xmath5 is reduced to reduced row echelon form by row operations , the result is an @xmath4 matrix whose first @xmath2 rows form a basis of the row space of @xmath5 .
let @xmath37 be the @xmath21 matrix consisting of the @xmath2 non - zero rows of the reduced form
. then @xmath5 and all matrices with the same row space can be constructed from @xmath37 by multiplying @xmath37 on the left by an @xmath38 matrix @xmath39 of rank @xmath2 .
the matrix @xmath39 is unique .
this gives a factorization of @xmath5 as @xmath40 . in terms of the associated linear maps ,
@xmath5 is a linear map from @xmath15 to @xmath22 , which factors into a surjective map onto @xmath41 followed by an injective map from @xmath41 to @xmath22 .
recall that knowing the row space of a matrix is equivalent to knowing the kernel of the associated linear map .
thus , when the reduced matrix @xmath37 is held fixed and @xmath39 is varied , the product @xmath42 gives all maps with the same row space ( i.e. same kernel ) .
for @xmath4 matrices of rank @xmath2 , the average weight per entry is @xmath43 we consider the random selection of a rank @xmath2 matrix such that each such matrix is equally probable . with the factorization @xmath40 ,
this can be done by selecting @xmath39 uniformly from all @xmath44 matrices of rank @xmath2 and by selecting @xmath37 independently from among all reduced row echelon matrices , which is the same as selecting the row space uniformly from all @xmath2-dimensional subspaces of @xmath15 .
the upper left corner of @xmath5 is @xmath45 .
but because @xmath37 is in reduced row echelon form , the first column of @xmath37 is either all zeros or has a leading 1 followed by zeros .
thus , @xmath46 . in order for @xmath47 to be non - zero , both @xmath48 and @xmath49 must be non - zero .
since the selection of @xmath37 is independent of the selection of @xmath39 , @xmath50 the columns of @xmath39 are @xmath2 linearly independent vectors of length @xmath0 and so the first column is not the zero vector .
that means there are @xmath51 possible first column vectors .
there are @xmath52 choices for @xmath53 and @xmath54 choices for the remaining entries of the first column .
therefore , @xmath55
now @xmath56 , and @xmath57 when the row space of @xmath37 contains nothing in the direction of the vector @xmath58 , which is to say that the row space is contained in the @xmath59-dimensional space @xmath60 . therefore , @xmath61 @xmath62 using formula 2 one easily obtains @xmath63 putting these results together we have @xmath64 with this result we have a clear picture of the effect of the parameters @xmath2 , @xmath0 , and @xmath1 on the average weight .
the factor @xmath65 is the average weight per entry without the rank condition , in which case the matrix size does not matter .
note that with @xmath0 and @xmath1 fixed , it is more likely for an entry to be non - zero for matrices of larger rank , an intuitively plausible result because both weight and rank are some measure of distance from the zero matrix . also , one can see that as @xmath0 , @xmath1 , and @xmath2 simultaneously go to infinity , the probability approaches @xmath66 , which is again the unconditioned probability . for invertible matrices of size @xmath1 ( i.e. @xmath67 )
the probability of a non - zero entry is @xmath68
for the matrices of rank one a more complete analysis of the weight distribution is possible . in this case
@xmath39 is a non - zero column vector of length @xmath0 and @xmath37 is a non - zero row vector of length @xmath1 whose leading non - zero entry is 1 .
the rank one matrix @xmath40 is given by @xmath69 , and so the weight of @xmath5 is the product of the weights of @xmath39 and @xmath37 .
the weight of @xmath39 has a binomial distribution conditioned on the weight being positive ( the entries of @xmath39 can not all be zero ) @xmath70 likewise for @xmath37 the weight distribution is given by @xmath71 ( to select a random @xmath37 , choose a random non - zero vector of length @xmath1 and then scale it to make the leading non - zero entry 1 .
the scaling does not change the weight . )
the weight on rank one matrices is the product of these two binomial random variables , each conditioned to be positive .
@xmath72 not all weights between 1 and @xmath13 occur for rank 1 matrices since the weight is a product with one factor no greater than @xmath0 and the other factor no greater than @xmath1 .
plots of actual probability distributions show spikes and gaps .
plots of cumulative distributions are smoother and lead us to expect a limiting normal distribution .
see figures 1 and 2 .
the weight random variable for rank one matrices of size @xmath36 is the product of independent binomial random variables conditioned on being positive .
define @xmath73 , where @xmath74 , @xmath75 , and @xmath76 and @xmath77 are independent bernoulli random variables with probability @xmath78 of being 0 .
then @xmath20 is the sum of @xmath0 independent identically distributed random variables @xmath79 .
conditioning @xmath20 on @xmath80 is the weight of rank one matrices . by the central limit
theorem the distribution of @xmath20 converges , as @xmath81 , to a normal distribution after suitable scaling .
now conditioning on @xmath20 being positive does not change this result because the probability that @xmath80 is @xmath82 , which goes to 1 as @xmath81 .
@xmath86 expanding and using the independence of the random variables @xmath87 and the fact that @xmath88 and @xmath89 , we get @xmath90 the variance of the weight is @xmath91 furthermore @xmath92 combining these expressions we get @xmath93 we get a good approximation to the variance of the weight when @xmath0 and @xmath1 using the unconditioned @xmath20 , essentially the factors in the denominator by 1 .
thus , @xmath94 which can be simplified to give @xmath95 | define the weight of a matrix to be the number of non - zero entries .
one would like to count @xmath0 by @xmath1 matrices over a finite field by their weight and rank .
this is equivalent to determining the probability distribution of the weight while conditioning on the rank .
the complete answer to this question is far from finished . as a step in that direction
this paper finds a closed form for the average weight of an @xmath0 by @xmath1 matrix of rank @xmath2 over the finite field with @xmath3 elements . the formula is a simple algebraic expression in @xmath0 , @xmath1 , @xmath2 , and @xmath3 . for rank one
matrices a complete description of the weight distribution is given and a central limit theorem is proved .
random matrices , finite fields , weight , rank .
15a52 ( primary ) , 15a03 , 15a33 , 60c05 ( secondary ) . |
wind energy is a major source of renewable energy .
however , conventional wind turbines are restricted by physical and economic limits .
airborne wind energy has the potential to overcome some of the limitations , using tethered flying devices to reach altitudes of 400 to 600 m where the wind is stronger and steadier @xcite .
the fact that airborne wind energy systems do not require towers reduces the costs per installation significantly .
the focus of this paper is the modelling of airborne wind energy systems that use the traction power of a tethered inflatable wing in a pumping cycle , as described in @xcite and @xcite .
the main components of such a single - tether kite power system ( kps ) are the wing , the kite control unit ( kcu ) suspended below the wing by means of a bridle system , the tether and the drum - generator module , which is part of the ground station .
it is the objective to develop a system model that is real - time capable and of sufficient accuracy for the development and verification of flight path and ground station controllers . a dynamic model of a two - line kite
is derived in @xcite .
variations of the angle of attack are not taken into account and the simplicity of the model allows for an analytical derivation of a state space representation based on four dynamic states . further expanding on this model
, @xcite proposed a kite power system model with three degrees of freedom ( dof ) , in which the kite is represented as a point mass at the end of the straight tether of variable length .
assuming a rigid wing with constant aerodynamic properties , the steering forces are derived as functions of the roll angle .
a discretisation of the tether as a multibody system has been proposed by @xcite , using a lagrangian approach to derive the equations of motion in generalised coordinates .
the advantage of this approach is the direct incorporation of constrains which results in a compact problem formulation .
this model used rigid tether segments , connected by spherical joints , which is not sufficient for modelling the tether force and implementing the force control loop .
in addition it is adding and removing point masses during the simulation to simulate reel - out and reel - in of the tether . according to our experience
this causes artificial discontinuities in the model which makes it difficult to implement the force control loop . for the kite it also used a point mass model .
a model that uses a discretised tether with point masses connected by springs was published in @xcite .
the aerodynamics of the kite were modelled using the vortex lattice method , which means that it is using an advanced kite model . on the other hand
it was not mentioned if the dynamics of the winch were modelled at all and no details were published on the question how reeling in and out was modelled .
other authors presented detailed generator and winch models @xcite , but no or only a very simple model for the kite and the tether .
coupling fluid and structural dynamic solvers for wind turbine applications has been studied by @xcite , while fluid - structure interactions methods have been applied to kite aeroelastic behaviour by @xcite .
these kind of models might be useful for the design of improved kites , but they are very computational intensive and currently at least one order of magnitude slower than real - time @xcite
. this paper presents a model where the dynamics of all major system components - the tether , the kite and the generator - are taken into account , with a focus on a novel discretised tether model which allows smooth reel - in and reel - out .
it is soft real - time capable and thus suitable for the training of kite pilots and winch operators , but can also be used for software in the loop testing of kps control systems , the development of estimation algorithms and for the optimization of flight trajectories .
an improved one - point kite model is presented , that allows to change the angle of attack during simulation time and uses look - up tables to calculate the lift and drag as function of the angle of attack .
it also takes the increased drag when flying around corners into account .
in addition it uses a correction term to match the influence of gravity .
this model can already be sufficient for optimizing flight trajectories . for controller development a four - point kite model
is devised , the most simple point mass model that has rotational inertial in all axis .
this avoids discontinuities in the kite orientation which make the one - point kite model uncontrollable in curtain flight manoeuvres .
in addition it is very close to a fully physical model : many model parameters like the height and width of the kite and the height of the bridle can just be measured and do not need to be identified . only the steering sensitivity parameters need to be identified because they depend on the flexibility of the kite which is not explicitly modelled .
this article will first explain the atmospheric model , then the tether model and the two kite models and finally the winch model .
furthermore , the control system is briefly explained .
subsequently a systematic approach for the model calibration is presented , with the goal to match the conditions of a real flight as good as possible . in the results section major parameters like force , speed , power and flight trajectory as obtained from the point mass model and the four point model
are compared with data , measured using the hydra kite of delft university of technology .
finally conclusions are drawn about the performance and accuracy of the described models and which improvements are still needed .
one of the requirements when building the model was , that it has to be ( soft- ) real - time capable .
on the other hand , the programming effort should be limited and it should be easy to adapt the model to different kite power systems .
it was found that high - level modelling tools like simulink or modellica were not capable to simulate a discretised tether that is reeling in or out in real - time .
therefore general purpose programming language was used that makes low - level optimizations of the modelling code possible .
we are modelling the kite and the tether as a particle system , using discrete point masses which are connected by spring - damper elements .
this has the advantage of a coherent model structure for which efficient mathematical methods for solving the stiff equation system exist @xcite
. for describing the positions of the particles a ground fixed reference frame is used , where the x - axis is pointing east , the y - axis north and the z - axis upwards .
the origin is placed at the ground station .
the state vector of the system was constructed using the states of the tether particles , the states of the kite particles ( only needed for the the four point kite model , because otherwise the last tether particle also represents the kite ) and the scalar states of the winch ( generator ) . because no accurate , real - time measurements of the wind speed at the height of the kite were available , an atmospheric model , describing the wind profile ,
was also needed . to determine the wind speed @xmath4 at the height of the kite and at the height of each tether segment ,
the power law @xcite and the log law @xcite are used .
input parameters are the ground wind speed @xmath5 and the current height @xmath6 of the kite or tether segment .
the ground wind speed used in this paper was measured at @xmath7 .
the power law establishes the relationship between @xmath4 and @xmath5 as @xmath8 with the exponent @xmath9 as fitting parameter .
the logarithmic law , which according to @xcite is more accurate than the least - square power law , can be written in the following form @xmath10 where @xmath11 is the reference height and @xmath12 is the roughness length . for this paper
not only the ground wind speed @xmath5 is measured , but once per flight additionally the wind speed at two more heights , @xmath13 and @xmath14 .
then , a wind profile is fitted to these three wind speeds . to make a fit with three ( speed , height ) pairs possible , eqns .
( [ eq : loglaw ] ) and ( [ eq : powerlaw ] ) are combined in the following way @xmath15 the fit is done by varying the surface roughness @xmath12 and @xmath16 until @xmath17 according to eq .
( [ eq : windprofile ] ) matches the measured wind speed at all three heights .
the exponent @xmath9 is chosen according to @xmath18 which results in @xmath19
. an average sea - level density of @xmath20 kg / m@xmath21 is assumed , and the height dependency is calculated according to @xmath22 where @xmath6 is the height and @xmath23 km .
an example for a fitted wind profile is shown in fig .
[ fig : wind_profile ] , using the parameters from table [ tab : simulation_parameters ] .
the tether is modelled as a fixed number of lumped masses , connected with n spring damper elements as shown in fig . [
fig : fourpointkite ] . to simulate reel - in and
reel - out the initial length of the tether segments @xmath24 is varied according to @xmath25 where @xmath26 is the tether length at the beginning of the i - th time step , @xmath27 the reel - out velocity , @xmath28 the simulation time and @xmath29 the simulation time at the beginning of the i - th time step .
this length is then used to calculate the spring and damping constants @xmath30 @xmath31 where @xmath32 is the initial length of the tether segments at the beginning of the simulation .
the differential equations of the particle system are formulated as an implicit problem @xmath33 the state vector @xmath34 of the particle system is defined as @xmath35 where @xmath36 and @xmath37 comprise the positions and velocities of the particles , respectively . for solving the problem only the residual @xmath38 needs to be programmed .
the vector @xmath39 consists of two partitions , the residual of the position vectors and its derivatives , and the residual of the velocity vectors and its derivatives , @xmath40 the first partition can be calculated from newton s law , @xmath41 . to calculate the second partition the particle forces must be known . on each particle
are acting the forces of the spring damper element above and below the particle .
in addition , half of the aerodynamic drag forces of the tether segments above and below of each particle have to be taken into account . with @xmath42 denoting the tensile force of segment @xmath0 and @xmath43 denoting the aerodynamic drag force of this segment as calculated in eq .
( [ eq : drag_force ] ) , the forces acting on the i - th particle can be calculated according to @xmath44 the first and the last particle have to be treated differently : for @xmath45 the lower spring force has to be replaced with the tether force as experienced by the ground station , and for the last tether particle the aerodynamic force of the kite has to be taken into account .
the spring forces are calculated according to hooke s law @xmath46 with @xmath24 , @xmath47 and @xmath48 calculated according to eqns .
( [ eq : inital_length ] ) , ( [ eq : spring_constant ] ) , ( [ eq : damping_constant ] ) and with @xmath49 we use linear springs with a different stiffness for the extension and compression regimes .
the stiffness for compression is much lower to model the behaviour of flexible bridle and tether lines , yet provide some structural stability .
the aerodynamic drag of any tether segment is calculated in the following way : first the wind speed at the height of the i - th tether segment @xmath50 is calculated using eq .
( [ eq : windprofile ] ) .
then , the average segment velocity is calculated as @xmath51 which leads to the apparent air velocity @xmath52 the drag of a cylinder is mainly caused by the component of @xmath53 that is perpendicular to the tether segment @xmath54 calculated as @xmath55 using this the drag force on the tether segment is resulting in @xmath56 where @xmath57 is the tether drag coefficient and @xmath58 the tether diameter . the point mass model proposed in @xcite represents the kite as a discrete mass moving under the action of an aerodynamic force vector .
it is also denoted as one point " or 1p " model .
steering is incorporated by an aerodynamic side force which depends linearly on the steering input .
this model does not account for rotational inertia , assuming that the wing is always aligned with the local relative flow experienced during flight .
expanding on the original work , the model presented in the following allows for tether deformation and a variable angle of attack .
the kite reference frame ( @xmath59 ) is defined on the basis of the local tether geometry and relative flow conditions . as illustrated in fig .
[ fig : kitereferenceframe ] the @xmath6-axis is aligned with the last tether segment .
( 1,0.70571463 ) ( 0,0 ) ) of the point mass kite model .
the physical wing is included here for the purpose of illustrating the concept of angle of attack and the assumed alignment with the relative flow.,title="fig : " ] ( 0.52309783,0.13566052)(0,0)[lb ] ( 0.52397632,0.21066957)(0,0)[lb ] ( 0.40889866,0.14865481)(0,0)[lb ] ( 0.47125528,0.02732639)(0,0)[lb ] ( 0.37025204,0.39141717)(0,0)[lb ] ( 0.29652592,0.25200616)(0,0)[lb ] ( 0.24819902,0.25390188)(0,0)[lb ] ( 0.34901797,0.16299502)(0,0)[lb ] ( 0.24657469,0.14116669)(0,0)[lb ] ( 0.1793676,0.44498375 ) ( 0.46805146,0.47298171)(0,0)[lb ] ( 0.56999222,0.31736892)(0,0)[lb ] ( 0.45157671,0.30673491)(0,0)[lb ] ( 0.79163634,0.02851584)(0,0)[lb ] ( 0.71812262,0.64055205 ) the @xmath60- and @xmath61-axes are constructed such that the apparent air velocity vector @xmath62 is in the @xmath63-plane .
this is based on the assumption that the wing is always aligned with the apparent wind velocity and that the sideslip velocity vanishes correspondingly .
the vector base is calculated as @xmath64 the unit vector @xmath65 is also called _ heading _ , because it describes the orientation of the wing .
the external force @xmath66 acting on the point mass representation of the kite comprises contributions of aerodynamic lift @xmath67 and drag @xmath68 , the aerodynamic side force @xmath69 and the gravitational force @xmath70 @xmath71 it should be emphasised that the drag force increases as the kite is steered due to kite deformation .
also , the steering force is based on the side area of the kite rather than the top area of the kite . the factored term @xmath72 represents a parametrized description of a kite s geometry .
the constant @xmath73 describes the steering sensitivity of the kite and has to be determined experimentally .
the influence of the steering on the drag is described by @xmath74 .
the empirical value of @xmath75 is used .
the variable @xmath76 is a correction term for the influence of gravity on the turn rate of the kite .
it is calculated as follows @xmath77 equation ( [ eq : correction ] ) is derived from the turn rate law as presented in eq .
( [ eq : psi ] ) .
the correction factor @xmath78 must be chosen such that the identified parameter @xmath79 of the one - point model matches the measurements . without this correction
the influence of gravity in this model was more than a factor of two higher .
we make the following assumptions : * the kite - tether angle depends linearly on the depower settings @xmath80 ; * the kite - depower angle has the value @xmath81 for @xmath82 ; * the maximal depower value of @xmath83 corresponds to a kite - tether angle of @xmath84 then , the angle of attack can be calculated with the following formula @xmath85 where @xmath81 is the angle between the kite and the cable when the kite is fully powered as shown in fig .
[ fig : alpha_0 ] and @xmath86 is the additional angle resulting from reeling out the depower line @xmath87 where @xmath88 is the value of the depower control input that is needed for the fully powered kite ( maximal l / d ) and @xmath89 and @xmath90 the values for @xmath80 and @xmath86 respectively that are needed for the fully depowered kite .
figure [ fig : lift_and_drag ] shows the lift coefficient @xmath91 and the drag coefficient @xmath92 as functions of the angle of attack @xmath9 .
the curves are established using the models of lift and drag coefficients of stalled and unstalled airfoils from @xcite , yet experience based modifications were made to better fit the coefficients of the non - ordinary wing section of a leading edge inflatable tube kite .
the point mass kite model can be sufficient to simulate and optimize the flight path of a power kite , because it is controllable during the power cycle and the simulated tether forces are close to the measured values . in addition the point mass model can be used to calculate the initial orientation of more complex models .
however , it is not a good choice for the development and optimization of flight - path control algorithms , because the reaction of the kite to steering inputs is problematic : a point mass kite has no rotational inertia , therefore its yaw angle is _ jumping _ when the sign of @xmath93 is changing .
this is non - physical behaviour . in these situations
controllability is lost .
therefore , we will now investigate a four - point kite model ( 4p model ) in order to obtain a more realistic and robust model .
the most simple particle - system based kite model that has rotational inertia in all axis is a four point kite model , which we will use from now on .
the points of the this model are defined in fig .
[ fig : fourpointkite ] .
( 1,0.92916667 ) ( 0,0 ) , @xmath94 , @xmath95 and @xmath96 .
points @xmath97 to @xmath98 discretize the tether.,title="fig : " ] ( 0.83646118,0.90893707)(0,0)[lb ] ( 0.9484904,0.739199)(0,0)[lb ] ( 0.68812368,0.74897258)(0,0)[lb ] ( 0.81051717,0.69804108)(0,0)[lb ] ( 0.83700134,0.50594693)(0,0)[lb ] ( 0.79741557,0.3770418)(0,0)[lb ] ( 0.73868073,0.26231895)(0,0)[lb ] ( 0.65997101,0.17086036)(0,0)[lb ] ( 0.5535286,0.10655021)(0,0)[lb ] ( 0.42991496,0.06415127)(0,0)[lb ] ( 0.17465035,-0.00237206)(0,0)[lb ] ( 0.19122147,0.27742785)(0,0)[lb ] ( 0.39659175,-0.00056181)(0,0)[lb ] ( 0.33631983,0.11220507)(0,0)[lb ] ( 0.84278941,0.77431226)(0,0)[lb ] the kite mass @xmath99 is distributed to points @xmath100 to @xmath96 according to eqns .
( [ eq:0 ] ) to ( [ eq:4 ] ) while the mass of the kite control unit @xmath101 plus half of the mass of the last tether segment are used as the mass of @xmath98 @xmath102 where @xmath103 is the nose mass fraction of the wing , @xmath104 the number of tether segments , @xmath105 the current tether length and @xmath106 the linear mass density of the tether .
the simulation of typical flight manoeuvres at low apparent air velocities has shown that a value of @xmath107 reproduces well the dive - down behaviour of the leading edge inflatable ( lei ) tube kites employed in the current study .
the virtual centre position of the kite , @xmath108 is defined as @xmath109 the origin of the kite reference frame is at @xmath94 .
the unit vectors @xmath110 and @xmath111 are defined as @xmath112 to parametrize the shape of the kite only three values need to be defined : the height of the kite @xmath113 ( distance between @xmath114 and @xmath94 ) , the height of the bridle @xmath115 ( distance between @xmath114 and @xmath98 ) and the width of the kite @xmath116 ( the distance between @xmath95 and @xmath96 ) . to calculate the initial positions of the kite particles , the point mass kite model from sect .
[ subsec : point - mass_model ] is used .
the initial unit vectors of the kite reference frame ( @xmath117 and @xmath118 ) are calculated using the kite position , the orientation of the last tether segment and the apparent air velocity ( eqns .
( [ eq : z_vector ] ) , ( [ eq : y_vector ] ) and ( [ eq : x_vector ] ) ) .
when these vectors are known , the positions of the kite particles at zero force can be defined by the following equations @xmath119 where @xmath120 is the relative nose distance , a kite dependant factor in the order of @xmath121 . in combination with the nose mass fraction
@xmath103 the factor @xmath120 can be used to tune the rotational inertia and the centre of gravity . the distance from @xmath95 to @xmath96 is calculated using the tip - to - tip distance of the kite @xmath116 multiplied with the relative kite width @xmath122 which is a factor in the order of @xmath123 and reflects the fact that the aerodynamic steering forces do not act on the tips of the kite , but a little bit further inwards . during the simulation , the aerodynamic forces
are applied to surfaces that are attached to the kite particles .
this causes them to change their positions , and from the current positions the unit vectors of the kite reference frame can then be calculated using eqns .
( [ eq : x_4point ] ) , ( [ eq : y_4point ] ) and ( [ eq : z_4point ] ) .
the aerodynamic model assumes surfaces attached to the top particle @xmath94 and to the side particles @xmath95 and @xmath96 .
the sole purpose of particle @xmath100 is to achieve rotational inertia and to realistically place the centre of gravity , therefore no aerodynamic force is attached to this particle .
the lift forces are determined based on the part of the apparent velocity that is perpendicular to the leading edge as suggested in @xcite . for the surface attached to the top particle , this is the apparent velocity in the @xmath63-plane @xmath124 . for the surfaces attached to the side particles ,
the apparent velocity in the @xmath125-plane @xmath126 is needed .
these can be calculated as follows @xmath127 for the top surface of the kite the angle of attack can be calculated as follows @xmath128 the angle @xmath86 is the change of the angle between the kite and the last tether segment due to the change of the depower settings .
the value of @xmath86 is between zero when fully powered and - for the leading edge inflatable tube kites used at delft university of technology - about when fully depowered .
if the reel - out length of the depower tape , the height of the bridle , the height of the kite and the power - to - steering - line distance are known , @xmath86 can be calculated geometrically ; in many cases the linear approximation given by eq .
( [ eq : alpha_d ] ) is sufficient .
for the side surfaces of the kite the angles of attack can be calculated as follows @xmath129 where @xmath130 is the change of the angle of attack caused by the steering line difference .
for @xmath131 a value of @xmath132 is assumed . with @xmath133 a steering offset - which is in practice unavoidable and caused by asymmetries in the steering system
- it can be calculated as follows @xmath134 the value of @xmath135 ( in the order of must be chosen such that the steering sensitivity of the kite model matches the steering sensitivity of the kite to be simulated .
the factor @xmath136 describes the influence of the depower angle @xmath86 on the steering sensitivity : depending on the geometry of the bridle it has a value in the range of @xmath137 .
a value of @xmath138 means that the fully depowered kite needs @xmath139 times the steering input as a fully powered kite to achieve the same turn rate ( under the condition that the apparent wind speed is the same ) .
steering is accomplished by changing the angle of attack for the side surfaces differentially .
the aerodynamic forces that act on @xmath94 , @xmath95 and @xmath96 can be calculated according to eqns .
( [ eq : lp2 ] ) to ( [ eq : dp4 ] ) , where @xmath140 is the relative side area of the kite and @xmath141 the air density .
@xmath142 the coefficient @xmath143 is required to achieve the same lift - to - drag ratio for the straight flying four point kite as for the one point kite .
it can be calculated from @xmath144 where @xmath145 was needed to compensate the higher drag coefficients of the side areas , compared to the top area , caused by @xmath131 .
we view the winch as the assembly of an asynchronous generator , a gearbox and a drum around which the tether is wound .
the generator is used as motor during the reel - in phase and the sign of the generator s torque determines the direction of the energy flow .
we modelled the winch by combining the differential equations for the inertial system and an expression for the torque - speed characteristics of the generator .
the differential equations for the winch are again defined as an implicit problem @xmath146 the vector @xmath147 is the extended state vector of the implicit problem and consists of tether length @xmath26 and the tether velocity @xmath27 @xmath148 in order to solve this problem the residual @xmath149 is to be calculated , with @xmath150 defined as @xmath151 here , @xmath152 is the acceleration of the tether at the ground station . under the assumption of an inelastic interconnection of the generator and
drum through the gearbox , the acceleration can be calculated as @xmath153 where @xmath154 is winch inertia as seen from the generator , @xmath155 the drum radius , @xmath104 the gearbox ratio , @xmath156 the generator torque , @xmath157 torque exerted by the drum on the generator and @xmath158 the friction torque .
the torque exerted by the drum depends on the tether force that is exerted on the drum , which equals the norm of the force on the first tether particle @xmath159 we modelled the friction as the combination of a viscous friction component with friction coefficient @xmath160 and static friction @xmath161
@xmath162 to determine the torque - speed profile of the asynchronous generator , we used the equivalent circuit representation as in @xcite . under the assumption of negligible stator resistance ,
@xmath163 can be expressed as a function of @xmath27 and the synchronous generator speed @xmath164 as @xmath165 we assumed that the generator voltage @xmath166 is increasing linearly with the set speed , up to the nominal voltage @xmath167 at the nominal synchronous speed @xmath168 of the generator @xmath169 as derived in @xcite , the parameters @xmath9 and @xmath170 can be expressed as @xmath171 where @xmath172 is the rotor resistance and @xmath173 is the generator s self inductance .
these generator parameters could either be measured or estimated based on known torque data .
.properties of the ground station of delft university of technology [ cols= " < , > , < " , ] ( the value @xmath174 is the average mechanical power over the whole cycle , and @xmath175 is the cycle efficiency , the quotient of the mean mechanical power and the average mechanical reel - out power ) .
a two dimensional projection of flight trajectory , height of the kite vs. the ground distance , is a suitable means for visualisation and comparison of different flights . in fig .
[ fig : flight_path ] the measured and the simulated flight path of one cycle is shown .
the maximum height differs by less then 5% .
the minimum height differs by about 45 m. one reason for this are the inaccuracies of the global navigation satellite system ( gnss ) based height measurement .
the computed dynamic response of the kite to steering inputs compares well to the response measured during test flights . in all situations the estimated turn rate of the wing was within @xmath176 of the full range of the experimentally measured values while the standard deviation was only @xmath177 of the full range .
similar results , but limited to the steering of a kite on a tether of a fixed length were presented in @xcite and @xcite . by modifying empirically the parameters @xmath73 and @xmath78 , the proposed point mass model can be adapted to match all parameters of the turn rate law . compared to the proposed four point model it runs faster but is less accurate and can become dynamically unstable at low tether forces . in @xcite
it was assumed that the turn - rate law derived in @xcite and @xcite would only be valid for ram - air kites .
we found , that it is valid for leading edge inflatable tube kites , too .
the parameters of the mechanistic four point model can be derived from the physical properties of any soft kite and any asynchronous generator .
only small changes are required for other kites and generators .
it is well suited for controller development and can be used not only for the pumping cycle operation of the kite , but also for the simulation of launching , landing and airborne parking . for a full model validation of a specific system , two enhancements of the test design
are needed : first , accurate wind measurements at the height of the kite .
second , an accurate measurement of the maximum and minimum l / d of the kite and of the depower offset @xmath88 .
the presented models have shown to be easily adaptable and well suited for flight path optimization and the development of kps estimators and kps controllers .
while the corrected one - point model with an adapted flight path controller can be sufficient for flight - path optimization , the four point model is better suited for controller validation in a broader range of flight conditions . even though the accuracy of the predicted power output is not yet sufficiently validated , the one point model , using the correction according to eq .
[ eq : correction ] is predicting the influence of gravity on the turn rate much better than uncorrected point mass models and the four point model has a much more realistic dynamic response to the steering input then simpler models while still being real - time capable .
the source code is published under the gnu lgpl license in the context of the freekitesim @xcite project .
the authors want to thank filip saad and rachel leudhold for their participation in the development and documentation of a sport kite simulator , which in turn inspired the development of the kps simulator as described in this paper .
in addition they want to thank william anderson and axelle vir for proof reading .
r. van der vlugt , j. peschel , r. schmehl , design and experimental characterization of a pumping kite power system , in : u. ahrens , m. diehl , r. schmehl ( eds . ) , airborne wind energy , springer , berlin heidelberg , 2013 , ch .
http://dx.doi.org/10.1007/978-3-642-39965-7_23 [ ] .
u. fechner , r. schmehl , model - based efficiency analysis of wind power conversion by a pumping kite power system , in : u. ahrens , m. diehl , r. schmehl ( eds . ) , airborne wind energy , springer , berlin heidelberg , 2013 , ch .
http://dx.doi.org/10.1007/978-3-642-39965-7_14 [ ] . m. ahmed , a. hably , s. bacha , power maximization of a closed - orbit kite generator system , in : proceedings of the 50th ieee conference on decision and control and european control conference ( ieee cdc - ecc 2011 ) , ieee , orlando , fl , usa , 2011 , pp .
http://dx.doi.org/10.1109/cdc.2011.6160914 [ ] .
p. williams , b. lansdorp , w. j. ockels , modeling and control of a kite on a variable length flexible inelastic tether , in : proceedings of the aiaa modeling and simulation technologies conference and exhibit , hilton head , sc , usa , 2007 . http://dx.doi.org/10.2514/6.2007-6705 [ ] .
f. gohl , r. h. luchsinger , simulation based wing design for kite power , in : u. ahrens , m. diehl , r. schmehl ( eds . ) , airborne wind energy , springer , berlin heidelberg , 2013 , ch .
j. coleman , h. ahmad , e. pican , d. toal , non - reversing generators in a novel design for pumping mode airborne wind energy farm , in : u. ahrens , m. diehl , r. schmehl ( eds . ) , airborne wind energy , springer , berlin heidelberg , 2013 , ch .
34 , pp . 587597 . http://dx.doi.org/10.1007/978-3-642-39965-7_34 [ ] .
a. vir , j. xiang , f. milthaler , p. e. farrell , m. d. piggott , j .-
latham , d. pavlidis , c. c. pain , modelling of fluid - solid interactions using an adaptive mesh fluid model coupled with a combined finite - discrete element model , ocean dynamics 62 ( 10 - 12 ) ( 2012 ) 14871501 . http://dx.doi.org/10.1007/s10236-012-0575-z [ ] .
a. bosch , r. schmehl , p. tiso , d. rixen , dynamic nonlinear aeroelastic model of a kite for power generation , aiaa journal of guidance , control and dynamics 37 ( 5 ) ( 2014 ) 14261436 . http://dx.doi.org/10.2514/1.g000545 [ ] .
b. eberhardt , o. etzmu , m. hauth , implicit - explicit schemes for fast animation with particle systems , in : n. magnenat - thalmann , d. thalmann , b. arnaldi ( eds . ) , computer animation and simulation 2000 , eurographics , springer , vienna , 2000 , pp .
http://dx.doi.org/10.1007/978-3-7091-6344-3_11 [ ] .
u. fechner , r. schmehl , design of a distributed kite power control system , in : proceedings of the ieee international conference on control applications , dubrovnik , croatia , 2012 , pp .
http://dx.doi.org/10.1109/cca.2012.6402695 [ ] .
e. hairer , g. wanner , solving ordinary differential equations ii , stiff and differential - algebraic problems , 2nd edition , springer , berlin heidelberg , 1996 . http://dx.doi.org/10.1007/978-3-642-05221-7 [ ] .
m. erhard , h. strauch , theory and experimental validation of a simple comprehensible model of tethered kite dynamics used for controller design , in : u. ahrens , m. diehl , r. schmehl ( eds . ) , airborne wind energy , springer , berlin heidelberg , 2013 , ch .
http://dx.doi.org/10.1007/978-3-642-39965-7_8 [ ] .
l. fagiano , a. u. zgraggen , m. morari , m. khammash , automatic crosswind flight of tethered wings for airborne wind energy : modeling , control design and experimental results , ieee transactions on control systems technology 22 ( 4 ) ( 2014 ) 122 . http://dx.doi.org/10.1109/tcst.2013.2279592 [ ] .
c. jehle , r. schmehl , tracking control on the unit sphere applied to traction kites for power generation , aiaa journal guidance , control and dynamics 37 ( 4 ) ( 2014 ) 12111222 . | converting the traction power of kites into electricity can be a low cost solution for wind energy .
reliable control of both trajectory and tether reeling is crucial .
the present study proposes a modelling framework describing the dynamic behaviour of the interconnected system components , suitable for design and optimization of the control systems .
the wing , bridle , airborne control unit and tether are represented as a particle system using spring - damper elements to describe their mechanical properties .
two kite models are proposed : a point mass model and a four point model .
reeling of the tether is modelled by varying the lengths of constituent tether elements .
dynamic behaviour of the ground station is included .
the framework is validated by combining it with the automatic control system used for the operation of a kite power system demonstrator .
the simulation results show that the point mass model can be adjusted to match the measured behaviour during a pumping cycle .
the four point model can better predict the influence of gravity and inertia on the steering response and remains stable also at low tether forces .
compared to simple one point models , the proposed framework is more accurate and robust while allowing real - time simulations of the complete system .
kite power , airborne wind energy , kite power system model , kite model , tether model , kite control damping coefficient of tether segment [ ns / m ] unit damping coefficient [ ns ] steering coefficient ( one point kite model ) [ - ] tether diameter [ m ] relative depower input of kite control unit ( 0 , 1 ) [ - ] relative steering input of kite control unit ( -1 , 1 ) [ - ] spring constant of tether segment [ n / m ] unit spring
constant [ n ] steering - induced drag coefficient [ - ] tether length at beginning of time step @xmath0 [ m ] mass of kite control unit [ kg ] mass of kite [ kg ] number of tether segments [ - ] initial length of tether segment [ m ] relative depower setting of kite control unit ( 0 , 1 ) [ - ] relative steering setting of kite control unit ( -1 , 1 ) [ - ] tether reel - out speed [ m / s ] horizontal wind velocity at 6 m height [ m / s ] height of kite or tether segment [ m ] vector of accelerations of tether particles [ m / s@xmath1 drag force vector of tether segment @xmath0 vectors of the gravity and steering forces of kite [ n ] lift and drag force vectors of kite [ n ] vector of positions of tether particles [ m ] position vectors of the front and top kite particles [ m ] position vectors of the right and left kite particles [ m ] vector of the residual of the implicit problem/ model vector from the tether particle i to the particle i+1 [ m ] velocity of tether particle @xmath2 relative to particle @xmath0 [ m / s ] vector of apparent air velocity [ m / s ] vector of wind velocity at the height of kite [ m / s ] unit vector of the x , y and z - axis of the kite - reference frame state vector of the implicit problem/ model angle of attack and elevation angle [ rad ] air density [ kg@xmath3 |
fast and ultra - sensitive displacement and force detection is of great interest for a broad range of fundamental applications .
it is an essential requirement for the proposed detection of quantum superposition states@xcite , quantum squeezed states@xcite , and single - spin magnetic resonance in mrfm@xcite which has also been proposed for the read - out stage of a solid - state nmr quantum computer.@xcite the common requirement of these examples is the need to perform the measurement using a micro - to - nanoscale - mechanical detector at , or even below , the
standard quantum limit " ( sql ) @xmath4 , and in the radio - frequency range .
considering , for instance , a micron - sized cantilever with a resonant frequency of @xmath2 and quality factor @xmath3 , the detection of quantum squeezed states requires a displacement sensitivity of about @xmath0.@xcite as another example , in detecting single - spin magnetic resonance , the interaction force between a nano - magnet with a targeted electron spin is typically on the order of about @xmath5.@xcite if a micron - sized cantilever on which the nano - magnet is mounted is directly coupled at the spin precession frequency ( @xmath2 or higher ) , the magnetic force will cause the cantilever to oscillate at an amplitude of about @xmath6 for a single spin . following the development of the rf - set,@xcite it has been proposed to use an rf - set - based displacement detector for measuring sub - angstrom motion of micron - scale and smaller mechanical oscillators.@xcite in the rf - set , @xmath7 noise which limits the applications of the conventional dc - set
is considerably improved .
a shot noise limit of @xmath8 may be achievable by using an rf - set displacement detector,@xcite provided the coupling ( gate voltage ) between the set and micro - mechanical oscillator is strong enough .
however , the back - action of the set onto the mechanical oscillator must also be taken into account:@xcite the fluctuating force acting on the oscillator due to the fluctuating island charge produces another contribution to the displacement noise .
the total displacement noise will be the sum of contributions from two sources - the shot noise in the tunneling current and the force noise due to the fluctuating island charge . in this paper , we give a calculation of the displacement noise of the rf - set displacement detector . by using the intrinsic noise analysis method of korotkov,@xcite the spectral density of the shot noise @xmath9 in the tunneling current and the spectral density of the force noise @xmath10
are determined , and the total spectral density of the displacement noise @xmath11 is then deduced .
we also present two simple expressions from which good estimates of the displacement noise due to the shot noise and the fluctuating island charge can be obtained .
results for the ultimate intrinsic noise limit of the detector for different parameter values are presented and discussed .
the displacement and force sensitivities are estimated for an optimized rf - set displacement detector and we find that they meet the requirement for detecting quantum squeezing and single - spin magnetic resonance .
the scheme of the rf - set displacement detector is shown in fig .
1 . the basic principle of the device involves locating one of the gate capacitor plates of the set on the cantilever so that , for fixed gate voltage bias , a mechanical displacement is converted into a polarization charge fluctuation .
the stray capacitance @xmath12 of the leads contacting the set and an inductor @xmath13 form a tank circuit with resonant frequency @xmath14 , loaded by the set .
a monochromatic carrier wave is sent down the cable . at the resonant frequency
the circuit impedance is small and the reflected power provides a measure of the set s differential resistance @xmath15.@xcite when the gate capacitor is biased , mechanical motion of the cantilever is converted into differential resistance changes , hence modulating the reflected signal power . to simplify the algebra and make the final results concise
, we shall in fact consider a symmetric dc - set in the calculations with junction resistances @xmath16 and junction capacitances @xmath17 .
the differences between the rf- and dc - set noise formulae are inessential , with appropriate time - averages required in the former@xcite resulting in only small quantitative differences in the noise values .
referring to fig . 1 , when the drain - source voltage @xmath18 is small compared to the voltage @xmath19 , where @xmath20 is the total capacitance of the set , and also the thermal energy @xmath21 , the peaks of the tunneling current @xmath22 will be well separated in gate voltage @xmath23 . to derive an expression for @xmath22 in the tunneling region , we consider only the two most probable island electron numbers @xmath24 and @xmath25 , resulting in the following approximation:@xcite @xmath26\sigma(n)+e[b_{1}(n+1)-t_{1}(n+1)]\sigma(n+1 ) .
\label{current1}\ ] ] the probabilities @xmath27 and @xmath28 are also given there and the tunneling rates @xmath29 from the bottom ( top ) across the @xmath30th junction of the rf - set take the usual form [ see expression ( [ apprates ] ) in the appendix ] .
in the previous calculation of the sensitivity of the rf - set based detector , the shot noise formula @xmath31 was used.@xcite this expression , which gives the maximum current noise , is only approximately valid when @xmath18 is close to the tunneling threshold .
taking as our tool the intrinsic noise analysis of the set by korotkov , we obtain the following more accurate expression for the spectral density of the current noise @xmath32 [ see appendix a for an outline of the deviation ] : @xmath33 } { a(n)+b(n+1)}\frac{a^{2}(n ) g(n+1 ) -b^{2}(n+1 ) f(n)}{[a(n)+b(n+1)]^{2}+\omega^{2 } } , \label{currentnoise}\ ] ] where @xmath34 the spectral density of the displacement noise due to the shot noise in the tunneling current is : @xmath35 the dependence on the displacement @xmath36 enters through the gate capacitance which is approximately @xmath37 for @xmath38 where @xmath39 is the cantilever electrode - island electrode gap .
the fluctuation of the island charge will produce a fluctuating force on the cantilever , therefore leading to another contribution to the displacement noise .
the ultimate force noise is only determined by the stochastic character of the tunneling process . by using the same method of intrinsic noise analysis,@xcite
the following expressions are obtained for the spectral density of the force noise @xmath40 [ see appendix b for an outline of the deviation ] : @xmath41^{2}+\omega^{2 } } , \label{sf}\end{aligned}\ ] ] where @xmath42 and @xmath43 we model the cantilever as a simple harmonic oscillator , so that the spectral density of the displacement noise due to the fluctuating force is : @xmath44 where @xmath45 is the motional mass of the cantilever , @xmath46 is its resonant frequency , and @xmath47 is the quality factor . for a cantilever with geometry @xmath48 ( length@xmath49width@xmath49thickness ) and made from material of density @xmath50 and young s modulus @xmath51 , the resonant frequency of the lowest - order flexural mode is@xcite @xmath52 the motional mass of the cantilever
is given by @xmath53 and the effective spring constant @xmath54 at the cantilever s tip is @xmath55 .
the total displacement noise will be given as the sum of the above two contributions ( [ sxi ] ) and ( [ sxf ] ) : @xmath56 in addition , the following expressions are given here only for simple estimates of the displacement noise at the resonant frequency of the cantilever [ see appendix c for an outline of their derivation ] : @xmath57^{2 } , \label{approxsxi}\ ] ] @xmath58^{2 } , \label{approxsxf}\ ] ] where @xmath59 is a constant .
in this paper , we consider crystalline si cantilevers with mass density @xmath60 and young s modulus @xmath61 . to make the picture clear
, we compare the displacement noise for different cantilevers .
table i lists the resonant frequencies @xmath62 and masses @xmath63 ( the effective mass of a cantilever is @xmath64 ) of cantilevers with the same width and thickness but different lengths . the resonant frequencies @xmath65 are obtained from expression ( [ frequency ] ) .
we mainly focus our attention on the detector using the cantilever with geometry @xmath66 in @xmath67 , quality factor @xmath68 , resonant frequency @xmath69 , and effective spring constant @xmath70 .
the standard quantum limit of the fundamental flexural mode of this cantilever is @xmath71 , so that the minimum detectable force at the resonant frequency is @xmath72 .
2 shows the dependence of the displacement noise @xmath73 on the gate voltage @xmath23 at the resonant frequency @xmath69 of cantilever .
the result assumes a symmetric set with junction capacitance @xmath74 , junction resistance @xmath75 , gate capacitance @xmath76 , gate capacitor gap @xmath77 , source - drain voltage @xmath78 , and temperature @xmath79 , which are all routinely achievable .
the solid line is obtained by using expressions ( [ current1 ] ) , ( [ currentnoise ] ) , and ( [ sxi ] ) and the dashed line using ( [ approxsxi ] ) .
the minimum displacement noise occurs near the edges of the tunneling current @xmath22 , and decreases approximately in inverse proportion to increasing gate voltage .
the displacement noise diverges in the center of tunneling current peaks and in the coulomb blockade region where the tunneling current is suppressed .
the result here is similar to fig . 2 in ref .
the dependence of the displacement noise @xmath80 on the gate voltage @xmath23 at the resonant frequency @xmath69 is illustrated in fig .
the result assumes the same parameters as above .
the solid line is obtained by using expressions ( [ sf ] ) and ( [ sxf ] ) , and the dashed line using ( [ approxsxf ] ) .
the maximum displacement noise increases linearly with the increasing gate voltage .
4 shows the total displacement noise @xmath81 as a function of gate voltage .
the tunneling current is shown as the dashed lines . considering the opposite dependences of @xmath82 and @xmath83 on gate voltage
, we will expect an optimum gate voltage value for the lowest displacement noise and best sensitivity .
but there are two problems involved in the determination .
the first problem is that near the edges the tunneling current @xmath84 may be too small to be measured accurately .
the second is that because @xmath82 and @xmath83 change in different ways with increasing gate voltage @xmath23 , it is rather involved finding the most optimum value of @xmath23 at which the total noise is minimum ; the positions of the minima in the displacement noise doublets will change relatively to the tunneling current peaks with increasing gate voltage .
an alternative , simpler strategy is to investigate the displacement noise at the gate voltage for which the tunneling current is fixed and measurable , say between @xmath85 and @xmath86 of its maximum value @xmath87 . fig .
5 shows the total displacement noise @xmath81 for our rf - set displacement detector using different cantilever parameter values listed in table i. the above method is utilized with the displacement noise determined where the tunneling current is half its maximum value .
for the detector using a cantilever with resonant frequency @xmath69 , the minimum displacement noise is @xmath88 , corresponding to an absolute displacement sensitivity of @xmath89 and force sensitivity @xmath90 .
note that the minimum displacement noise increases and shifts to lower gate voltage for increasing cantilever length . as is evident from ( [ approxsxf ] ) , this is a direct consequence of the effective spring constant dependence : obviously a stiffer cantilever will be displaced less by the same applied force .
6 shows the detector displacement noise for different junction capacitance values .
both @xmath82 and @xmath83 decrease with decreasing @xmath91 . because small displacements are detected by monitoring changes in tunneling current , taking a smaller junction capacitance serves another advantage for the experiment : with @xmath92 kept fixed
, a smaller junction capacitance corresponds to a larger source - drain voltage , so that the maximum tunneling current @xmath87 increases according to expression ( [ currentmax ] ) .
7 shows the displacement noise when the tunneling current is biased at various values between @xmath93 and @xmath94 of @xmath87 .
we see that performing detection at a smaller percentage of the maximum tunneling current becomes another effective way to further decrease the displacement noise .
the lowest displacement noise we obtain is @xmath95 , assuming @xmath74 , @xmath76 , and @xmath96 , biased at @xmath93 of its maximum value .
the optimum gate voltage is @xmath97 in this case . in conclusion ,
approaching the sql will become possible for a carefully engineered rf - set displacement detector .
a sensitivity at , or even beyond the sql may be achievable by using the same parameters as in the above analysis but with the smaller junction capacitance @xmath98 .
such a sensitivity would enable us to observe quantum squeezed states in micro - mechanical systems .
the proposed displacement detector also meets the crucial requirement for advances in mrfm .
the radio - frequency micro - mechanical cantilever used in our model allows direct coupling at the spin precession frequency , therefore allowing fast read - out , while the predicted force sensitivity would allow the measurement of single - spin resonance .
the authors thank keith schwab and rob schoelkopf for pointing out the need to include the backaction on the cantilever due to the fluctuating island charge and martin wybourne for helpful discussions .
this work was supported in part by the national security agency ( nsa ) and advanced research and development activity ( arda ) under army research office ( aro ) contract number daag190110696 , and by an award from research corporation .
the tunneling current shot noise is given by expression ( 30 ) in ref . .
in our notation , it is : @xmath99 b_{m',m}\cr & & \times\left\{\left[t_{2}(m-1)-b_{1}(m-1)\right]\sigma(m-1)+ \left[t_{1}(m+1)-b_{2}(m+1)\right]\sigma(m+1)\right\ } , \label{appsi}\end{aligned}\ ] ] where @xmath100 \label{u}\ ] ] and @xmath101 is the real part of the inverse of the tridiagonal matrix @xmath102 , where @xmath103 is the unit matrix and @xmath104 with @xmath105 the tunneling rates are @xmath106/ \left[1-e^{-\delta e_{i}^{-}(n)/k_{b}t}\right],\cr t_{i}(n)&=&\left[\delta e_{i}^{+}(n)/e^{2}r_{j}\right]/\left[1-e^{-\delta e_{i}^{+}(n)/k_{b}t}\right ] , \label{apprates}\end{aligned}\ ] ] where @xmath107{e}/{c_{\sigma}},\cr \delta e_{2}^{\pm}(n)&=&\left[-{e}/{2}\ \mp\left(en+(c_{j}+c_{g } ) v_{\rm ds } -c_{g } v_{g}\right)\right]{e}/ { c_{\sigma}}. \label{energies}\end{aligned}\ ] ] a convenient form for the tunneling rates is @xmath108},\cr b_{2}(m)&=&\frac{1}{r_{j}c_{\sigma}}\frac{m - n-1-\delta n+\tilde{v}_{\rm ds}}{1-\exp\left[-(m - n-1-\delta n+\tilde{v}_{\rm ds})/\tilde{t}\right]},\cr t_{1}(m)&=&\frac{1}{r_{j}c_{\sigma}}\frac{m - n-1-\delta n-\tilde{v}_{\rm ds}}{1-\exp\left[-(m - n-1-\delta n-\tilde{v}_{\rm ds})/\tilde{t}\right]},\cr t_{2}(m)&=&\frac{1}{r_{j}c_{\sigma}}\frac{n - m+\delta n-\tilde{v}_{\rm ds}}{1-\exp\left[-(n - m+\delta n-\tilde{v}_{\rm ds})/\tilde{t}\right ] } , \label{appconvrates}\end{aligned}\ ] ] where @xmath109 and @xmath24 , @xmath25 are the most probable island electron numbers , determined by the condition @xmath110 which is equivalent to @xmath111 . under the condition @xmath112 ( equivalently @xmath113 ) , we have that @xmath114 , @xmath115 , @xmath116 , @xmath117 for small drain - source voltage , and furthermore @xmath118 for all @xmath119 and @xmath120 for all @xmath121 . because the probabilities @xmath122 for all @xmath123 , @xmath25 , only the terms @xmath124 and @xmath125 need be considered in the summation over @xmath63 in eqs .
( [ appsi ] ) and ( [ u ] ) . from the form of the matrix ( [ gamma ] ) and
the above approximate conditions on the probabilities and tunneling rates , we then also have that only @xmath126 and @xmath127 need be considered in the sum over @xmath128 . the matrix elements
@xmath129 , @xmath130 , @xmath131 , and @xmath132 contain only the tunneling rates @xmath133 , @xmath134 , @xmath135 , and @xmath136 and it is valid to just consider @xmath137 matrices for @xmath138 and @xmath139 : @xmath140 the probabilities @xmath27 and @xmath28 are given in ref . as @xmath141,\cr \sigma(n+1)&=&a(n)/[a(n)+b(n+1 ) ] .
\label{rhon}\end{aligned}\ ] ] substituting ( [ matrix ] ) and ( [ rhon ] ) into ( [ appsi ] ) and ( [ u ] ) and neglecting all small terms , we get @xmath33 } { a(n)+b(n+1)}\frac{a^{2}(n ) g(n+1 ) -b^{2}(n+1 ) f(n)}{[a(n)+b(n+1)]^{2}+\omega^{2}}. \label{appcurrentnoise}\ ] ]
the electrostatic interaction between the fixed island and cantilever electrodes of the gate capacitor is @xmath142^{2 } , \label{appforce}\ ] ] where @xmath143 is the background charge and @xmath144 define @xmath145 where the second equality follows from ( [ dimdef ] ) . neglecting @xmath143 in ( [ appforce ] ) ,
the force noise is given as @xmath146\right\}. \label{appsf}\ ] ] similarly to expression ( 27 ) in ref .
, we have @xmath147 using expressions ( [ matrix ] ) , ( [ rhon ] ) , and ( [ correlations ] ) to obtain @xmath148 , @xmath149 , and @xmath150 $ ] , the final result for the force noise is @xmath151^{2}+\omega^{2}}. \label{appsf2}\end{aligned}\ ] ] note that both @xmath32 and @xmath40 have a lorenzian spectrum .
a simple , approximate formula for the tunneling current is@xcite @xmath152 , \label{appcurrent}\ ] ] where @xmath153 is the value of the gate voltage at which the tunneling current is maximal : @xmath154 subsituting ( [ appcurrent ] ) and the shot noise limit formula @xmath155 into ( [ sxi ] ) , we obtain @xmath156^{2}\frac{\cosh^{4 } \left[e c_{g}(v_{g,{\rm res}}-v_{g})/2.5k_{b}tc_{\sigma}\right ] } { \sinh^{2}\left[e c_{g}(v_{g,{\rm res}}-v_{g})/2.5k_{b}tc_{\sigma}\right]}. \label{appsxi}\ ] ] using expression ( [ current1 ] ) for the tunneling current with @xmath157 and expressions ( [ appconvrates ] ) for the tunneling rates with @xmath158 gives for the maximum tunneling current : @xmath159 to make a simpler estimate of the displacement noise , we might replace the dimensionless , parameter - independent function @xmath160/\sinh^{2}[x]$ ] by a constant @xmath161 , say its minimum value which is 4 .
however , this function poorly approximates the correct doublet shape and it is found numerically that the @xmath59 provides a good fit .
thus we obtain the final result for the displacement noise due to the shot noise in the tunneling current : @xmath57^{2}. \label{appapproxsxi}\ ] ] a simple approximation to the maximum force noise is obtained by setting @xmath162 in ( [ appsf2 ] ) and neglecting small terms : @xmath163^{2}. \label{approxsf1}\ ] ] substituting in the definition ( [ a ] ) for @xmath164 and using @xmath165 at maximum tunneling current , we obtain : @xmath166^{2 } , \label{appapproxsxf}\ ] ] at the resonant frequency @xmath46 of the cantilever , which is assumed to be much smaller than the tunneling rates in magnitude .
s. bose , k. jacobs , and p. l. knight , phys .
a * 59 * , 3204 ( 1999 ) .
m. p. blencowe and m. n. wybourne , physica b * 280 * , 555 ( 2000 ) ; xxxiv rencontres de moriond : quantum physics at mesoscopic scale , edited by d. c. glattli , m. sanquer , and j. trn thanh vn , ( edp sciences , france , 2000 ) , pp .
. j. a. sidles , j. l. garbini , k. j. bruland , d. rugar , o. zger , s. hoen , and c. s. yannoni , rev .
* 67 * , 249 ( 1995 ) .
g. p. berman , g. d. doolen , p. c. hammel , and v. l. tsifrinovich , phys .
b * 61 * , 14694 ( 2000 ) .
r. j. schoelkopf , p. wahlgren , a. a. kozhevnikov , p. delsing , and d. e. prober , science * 280 * , 1238 ( 1998 ) .
t. fujisawa and y. hirayama , appl .
lett . * 77 * , 543 ( 2000 ) .
a. aassime , g. johansson , g. wendin , r. j. schoelkopf , and p. delsing , phys .
lett . * 86 * , 3376 ( 2001 ) .
m. p. blencowe and m. n. wybourne , appl .
* 77 * , 3845 ( 2000 ) .
k. c. schwab ( unpublished ) .
a. n. korotkov , phys .
b * 49 * , 10381 ( 1994 ) . a. n. korotkov and m. a. paalanen , appl . phys .
* 74 * , 4052 ( 1999 ) . c. w. j. beenakker , phys .
b * 44 * , 1646 ( 1991 ) . | we investigate the intrinsic noise of a micro - mechanical displacement detector based on the radio - frequency single - electron transistor ( rf - set ) . using the noise analysis of a set by korotkov as our starting point
, we determine the spectral density of the displacement noise due to the tunneling current shot noise .
the resulting mechanical displacement noise decreases in inverse proportion to the increasing gate voltage .
in contrast , the displacement noise due to the fluctuating set island charge increases approximately linearly with increasing gate voltage . taking into account both of these noise sources results in an optimum gate voltage value for the lowest displacement noise and hence best sensitivity .
we show that a displacement sensitivity of about @xmath0 and a force sensitivity of about @xmath1 are predicted for a micron - sized cantilever with a realizable resonant frequency @xmath2 and quality factor @xmath3 .
such sensitivities would allow the detection of quantum squeezing in the mechanical motion of the micro - mechanical cantilever and the detection of single - spin magnetic resonance in magnetic resonance force microscopy ( mrfm ) .
= 9truein = .5 truein |
a generic prediction of the standard cosmological paradigm is that small structures form first while big ones are assembled later by hierarchical merging .
because the power spectrum is not truncated at any scales relevant to galaxy formation as it evolves , early massive galaxies are comparatively much rarer than galaxies of the same mass that assembled later .
these later massive galaxies have assembled their stellar bodies in ways and over time scales that are rather different from the early ones and thus must generally have different properties .
thus , the discovery of old and massive galaxies at high redshift that have rather different structural properties than those of most early type galaxies ( etgs ) in the second half of the hubble time is interesting because of the possibility it offers to directly explore additional mechanisms of formation of massive galaxies in general , and of quenching of the star formation in particular .
passive galaxies with large stellar mass , e.g. m @xmath3 10@xmath4 m@xmath5 , have been identified at redshift as high as @xmath6 , @xmath7% of the current age of the universe .
a striking characteristic of these young members of the population of `` etgs '' is that they are often of very small size , up to @xmath8 times smaller than galaxies of comparable mass in the local universe , and hence have very high stellar density , up to two orders of magnitude higher than local counterparts ( see , for example , * ? ? ?
* ; * ? ? ?
? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in fact , at @xmath9 the population of etgs is dominated by the compact ones , with more than @xmath10% of them having size smaller than the lower @xmath11 of local etgs of the same mass @xcite . in the local universe ,
these compact etgs seem to be exceedingly rare , although there is still ongoing debate on whether this apparent paucity is real or is due to bias in local surveys , such as the sdss @xcite given this apparent spectacular evolution , it is no surprise that a lot of effort went into exploring the possible evolutionary mechanisms that have driven it . for example , it has been suggested that individual compact ellipticals might form extended stellar halos either by in-situ star formation or by dry merging and accretion @xcite .
in particular it has also been proposed that interactions and repeated minor merging events , even if they do not increase the stellar mass by a large amount , can energize the innermost stellar orbits and `` puff up '' the compact galaxies @xcite .
concurrently , the size evolution of the population of etgs as a whole can also be driven by the addition of new members coming from the late quenching of massive , large galaxies @xcite .
in fact , from the analysis of the evolution of etgs of different stellar densities as a function of redshift , @xcite conclude that the addition of new , larger , etgs is required to explain the overall increase in their numbers from @xmath12 to the present @xcite . in any case , it is important to realize that an accurate census of compact galaxies in the local universe is still missing , since the sdss samples are very likely biased against such systems @xcite and also because the descendants of the compact galaxies might not be easily recognized in the local universe if they became the core of systems that developed extended stellar halos ( e.g. @xcite ) .
regardless of the problem of their subsequent evolution , however , it seems clear that compact etg were very abundant at high redshift , and in fact largely dominate the population of passive galaxies at redshift @xmath132.8 ( cassata et al .
2013 ) , at least at large mass ( @xmath14 m@xmath5 ) .
this raises the question of how such massive systems could form in such a relatively short time .
there are indications that whatever process is responsible for quenching star formation in massive galaxies , it is largely controlled by the star formation rate , with more actively star forming galaxies being more likely to quench ( e.g. * ? ? ?
* ; * ? ? ?
* ) , and that more compact systems are more likely to quench more effectively than those with more diffuse mass distribution @xcite . but even before they quench , the problem of how massive galaxies with such high stellar density could form , and why they dominate the population of massive passive galaxies at high redshift , is interesting because it seems to imply a specific formation mechanism different from other galaxies . is the physics that shuts off star - formation in high - redshift galaxies producing _ only _ compact remnants as a by - product ? or , does it preferentially affect those galaxies with high stellar densities ?
one popular mechanism to both shut off the star formation in a galaxy and also produce spheroidal morphologies are major mergers @xcite .
evidence of this mechanism is seen in the local universe @xcite , and evidence of merging has been observed out to high redshift @xcite . during a merging event , however , a substantial fraction of the pre existing stars of the merging galaxies are dispersed to larger radii , and its difficult to produce compact remnants unless the progenitors themselves are also very compact , and in any case the fraction of stars scattered to large radii is not negligible @xcite .
gas rich mergers may produce remnants with a very compact core through in - situ star - formation thanks to the highly dissipative nature of the gas @xcite , but the pre - existing stars are still dispersed to large radii , and the gas fractions must be high ( e.g. @xmath1570% ) in order to produce a large fraction of the stellar mass in a compact remnant @xcite . while it is possible that current data have not yet probed the low surface brightness regions surrounding compact etgs to the sensitivity required to to rule out evidence of major merger activity , tidal debris , or dispersed stars , there is some general evidence that these galaxies are truly compact in size , with no diffuse or extended structure surrounding them @xcite .
there is still much debate on the role of major merging in the buildup of the stellar mass of massive galaxies in general , regardless if compact or not , @xcite and in particular for morphologically selected spheroidal galaxies @xcite .
theoretical work have shown that violent disk instability ( vdi ; * ? ? ?
* ) , driven by intense accretion of cold gas from the cosmic web @xcite , can lead to the formation of compact massive galaxies .
the gas - rich disks are toomre unstable , with large - scale transient perturbations and massive bound clumps .
the mutual interactions between these perturbations exert torques that drive angular - momentum out and mass in , in the form of clump migration and gas inflow in the inter - clump disc medium @xcite .
as long as the inflow rate is more rapid than the star - formation rate in the disk , the inflow to the center is gas rich , and the product is compact ( dekel & burkert , in preparation ) .
the induced central starburst can eventually lead to quenching , by gas consumption into stars @xcite , by outflows driven by stellar or agn feedback @xcite , or by morphological quenching @xcite .
the star formation quenching may also be related to the shutting off of cold gas supply .
theory , confirmed by simulations , predict that after @xmath16 , for dark matter halos with masses of @xmath17 and above , the incoming gas is heated by a virial shock that can be supported because of the long cooling times @xcite .
the extent to which these processes affect the formation of compact etgs at @xmath18 remains unconstrained .
thus , progress is likely to come from the identification and empirical studies of their progenitors before and while they quench , namely while they are still in the star forming phase and when they shut it down .
some have proposed potential progenitors based on matching the stellar mass and the volume density of the etgs with those of star forming galaxies together with assumptions of possible evolutionary paths @xcite .
@xcite in particular have identified progenitors of the most massive compact etgs ( m@xmath19 m@xmath5 ) among the most massive ( m@xmath20 m@xmath5 ) galaxies at z@xmath213 by projecting their observed stellar masses and sfrs assuming various sfhs . but
whether or not the morphological properties , star formation rate and stellar mass of the more general population of putative progenitors were consistent with the compactness of their descendants among all passive galaxies at z@xmath222 and their specific star formation rate have not been considered in detail , something we set to do here . in this paper
, we try to answer the following question . since we do not know of any physical mechanism capable to transform a non compact stellar system into a compact one , do star forming galaxies exist at suitably high redshifts ( i.e. such that there is time for quenching ) that are _ plausible _ progenitors of the compact and massive etg , namely that are compact themselves and with stellar mass and star formation rate such to plausibly explain their descendants at @xmath23 ? to answer this question , we identify star - forming galaxies at @xmath6 that , if they quench , can reproduce both the mass and the stellar density , and hence size , of the observed compact galaxies at @xmath23 . in other words ,
we try to see if we can identify the progenitors based on the _ evolutionary consistency _ assuming only that 1 ) the star
forming galaxies quench early enough to be passive at @xmath9 and 2 ) that no ( unknown ) physical mechanism transform non - compact stellar systems into compact ones . with a sample of plausible progenitors
, we then can compare their properties to those of other star forming galaxies that are not plausible progenitors and see if there are differences that might offer some insight into the formation of the etgs .
we present the sample , and its selection , in sections 2 and 3 . in section 4
, we study the properties of these plausible progenitors , and compare them with the rest of the normal star - forming galaxy population at @xmath24 , and we investigate the distribution of stellar populations of different ages in the galaxies . in section 5
we discuss the implications of our results for the evolution of compact etgs , in the context of the evolutionary drivers and quenching mechanisms affecting these galaxies . throughout this paper
we assume a cosmology with @xmath25 , @xmath26 , and @xmath27 km s@xmath28 mpc@xmath28 .
in this paper we use data from the great observatories origins deep survey ( goods ; * ? ? ? * ) , and 4-epoch depth observations with _
hst_/wfc3 from the cosmic assembly near - infrared deep extragalactic legacy survey ( candels ) .
this covers 113 square arc - minutes of the goods - south field , which includes the candels deep region @xcite and the early release science ( ers ) @xcite .
the h - band images in the 4-epoch deep and ers regions reach 1-@xmath29 fluctuations of 26.6 and 26.3 ab arcsec@xmath30 , respectively . in total we make use of panchromatic photometry in goods - south , including u band data from the visible multiobject spectrograph ( vimos ) on the very large telescope ( vlt ; * ? ? ?
* ) , hst / acs b , v , i , z - band , hst / wfc3 j , h - band , vlt / isaac k@xmath31 photometry @xcite , and _
spitzer_/irac 3.6 , 4.5 , and 5.7 @xmath32 m imaging ( m. dickinson et al . , in preparation ) , _ spitzer_/mips 24@xmath32 m imaging ( m. dickinson et al .
, in preparation ) , and goods-_herschel_/pacs 100@xmath32 m imaging @xcite .
we measure photometry ( in irac channels and blue - ward ) for galaxies in the 4-epoch candels data using the object template - fitting method ( tfit ; * ? ? ?
* ) software package , which allows us to construct spectral energy distributions ( seds ) with mixed - resolution datasets .
all details about the construction of the multi - wavelength photometry constructed using tfit is discussed in @xcite .
the colors and ages of the high - redshift etgs are such that they should be star - forming at @xmath33 @xcite .
therefore we select star - forming galaxies at redshift @xmath33 from the acs z - band imaging using the lyman - break color selection @xcite , including z - band detections with z @xmath3 26.5 ( ab magnitudes ) .
the acs z - band is @xmath22 90% complete down to 26.5 for galaxies with size less than 0.3 arcseconds in half - light radius @xcite .
band dropout selection criteria are @xmath34 where @xmath35 refers to the logical and operator .
we additionally require signal to noise of at least 3 in the @xmath36 and @xmath37 bands to ensure robust color measurements .
we calibrate our u
band dropout selection using redshifted , continuously star - forming stellar population synthesis models of @xcite with varying values of dust extinction , following @xcite .
the lbg selection , with these varying stellar population models , are shown in figure [ lbgsel ] , and includes 943 objects .
we additionally remove from the lbg sample those galaxies without unique wfc3 h - band detections ( 36% of the original sample ) , and those with photometric redshifts less than 2 ( an additional 14% ) , to ensure only the most robust sample of z@xmath213 galaxies are used in the following analysis .
our final sample of lbgs includes 517 galaxies , 180 are h @xmath38 25 , and all of which have z@xmath39 .
as we shall see later , a sample of candidate progenitors of compact massive early type galaxies will be selected among the most compact lbg which , given the sensitivity of the goods images , is @xmath40 - 90% complete down to z@xmath39 .
the redshift distribution for h @xmath38 25 lbgs is shown in figure [ zdist ] . [ !
t ] we measure photometric redshifts and stellar masses for our sample of lbgs by fitting stellar population synthesis models to their observed seds .
photometric redshifts are derived using pegase 2.0 @xcite , where we integrate the probability distribution function of redshift to derive the photometric redshift .
20@xmath41 of our lbgs have spectroscopically confirmed redshifts ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* stern et al . in prep ) . using these photometric redshifts ( or spectroscopic where available )
, stellar masses are derived through fitting the stellar population synthesis models of @xcite with a salpeter initial mass function , as described in @xcite .
the models also use the calzetti dust extinction law @xcite and the madau ( 1995 ) cosmic opacity , and a number of star formation histories including exponentially decreasing ( @xmath42-models with varying time scale @xmath42 ) , constant , and two - component ( delay ) models comprised of linearly increasing and exponentially decreasing components ( e.g. * ? ? ?
while we found that there generally is good quantitative agreement between the stellar mass derived using these three star formation histories , we ended up adopting the exponentially declining or constant star
formation history that minimizes the @xmath43 .
t ] 25 lbgs . ] we measure sfrs of our lbgs using the observed slope of the rest frame ultraviolet ( uv ) sed , i.e. we do not use the sfr derived from the sed fitting procedure .
we make use of the correlation between the dust obscuration and the slope of the rest
frame uv sed of starburst galaxies @xcite to derive the dust - corrected uv luminosity , and subsequently the dust - corrected uv sfr using the conversion factor by @xcite .
morphological measurements in the h - band ( restframe - optical at @xmath2 ) of our lbgs are performed using the galfit package @xcite .
nearby objects are masked using segmentation maps produced by sextractor in the same configuration as that used for the initial source detection .
we fit light profiles using a sersic model .
we use a psf constructed from an average of unsaturated stars . to ensure robust morphological measurements , we remove galaxies from our sample
if galfit indicated the morphological measurement was questionable , namely such that we could not confidently rule out that they were stars , or the signal to noise in the h - band photometry was less than 15 .
previous investigations of galfit measurements with low signal to noise data indicate that a signal to noise of at least 10 is required to produce unbiased measurements @xcite . these criteria remove @xmath44 of our lbg sample , 2 lbgs on the basis of galfit error , and 8 lbgs on the basis of low signal - to - noise in the h - band .
our study uses a measure of circularized half - light radius , defined as @xmath45 where @xmath46 is the length of the semi - major axis in arc - seconds , and @xmath47 is the axis ratio of the galaxy .
all measurements of circularized half - light radius are converted to kpc using the photometric redshift of the object ( or spectroscopic redshift if available ) .
the uncertainty on the physical size of each galaxy is taken to be 20@xmath41 of the measurement , based on the simulations by @xcite , as the error returned by galfit does not include systematic errors .
our goal in this section is to see if among star forming galaxies at redshift @xmath2 , e.g. u band dropout lbgs , there are some that are plausible progenitors of the compact etgs at @xmath23 .
the redshift range between @xmath6 and 1.6 encompasses about 2 gyr , so essentially _ every _ lbg at @xmath6 that quenches its star formation soon and quickly enough after the epoch of the observation will satisfy our definition of `` passive galaxy '' . but
while quenching is an obvious necessary condition to be classified as passive , it is not sufficient , as not every quenched star - forming galaxy will have 1 ) the stellar mass , 2 ) the specific star formation rate , and 3 ) the stellar density of the compact , massive and passive galaxies that we are considering here , i.e. those selected using the criteria by @xcite thus , to identify the likely progenitors of this specific group of galaxies , let s first exactly define what we mean by _ massive , passive , compact galaxies_. we use the same criteria adopted by cassata et al .
( 2011 , 2013 ) to define their sample .
specifically , galaxies were selected to be in the redshift range @xmath48 by means of photometric redshift ( or spectroscopic ones when available ) , and for having stellar mass @xmath49 @xmath50 , and specific star
formation rate @xmath51 gyr@xmath28 . in the case of the etg sample , both stellar mass and @xmath52 have been estimated from sed fitting to stellar population synthesis models , assuming the star formation history as an exponentially declining one , which is appropriate for the case of galaxies that are completing the cessation of their star formation activity .
we note that our requirement of passivity is a relatively stringent one , with the @xmath52 limit being about 1/10 of the current value of the milky way .
recent selections of passive galaxies in the literature use ssfr thresholds which are an order of magnitude or even higher than our criteria ( e.g. * ? ? ?
this is required , in our opinion , to clearly distinguish a passive galaxy , like present day ellipticals , from relatively low level but continuous star formation , such as that of massive disks , since the star
formation history of these two types of galaxies are very different .
compact galaxies are defined in terms of their size and stellar mass , hence average stellar mass surface density within the effective ( half - light ) radius @xmath46 , namely @xmath53 . to be classified as compact , passive galaxies
of given mass must have stellar mass surface density larger than the value that exceeds the @xmath54 scatter of the stellar mass - size relationship for local ( @xmath55 ) etgs of the same mass ( thus , by definition @xmath56% of local etg are compact ) . ultra
compact galaxies are those that are 0.4 dex smaller than the @xmath54 value . in terms of projected density ,
the local mass size relationship is roughly parallel to the line of constant @xmath57 , so that the above definitions translate into the following conditions : @xmath58 m@xmath5 kpc@xmath30 for compact galaxies and @xmath59 m@xmath5 kpc@xmath30 for ultra compact ones ( see cassata et al .
we adopt these more general classifications in the remainder of this study .
the sample by @xcite includes a total of 107 etgs with stellar mass @xmath14 m@xmath5 in the range @xmath60 , with average redshift @xmath61 , of which 76 are compact according to the above definition , 42 of which are also ultra compact , and the remaining 31 are normal etgs , i.e. within the @xmath54 scatter in the mass size relationship observed at @xmath55 . at high redshift
ultra compact galaxies appear to dominate the population of massive passive galaxies . of the 21 galaxies of this sample that have @xmath62 , 4 are normal , 3 are compact and 14 are ultra compact , suggesting that compact and ultra compact galaxies dominate the population of passive galaxies at high redshift and thus were the first to become passive .
as the mass - size distributions of galaxies in the top panel of figure [ sizevsm ] show , the stellar masses of the lbgs are significantly smaller than that of the passive ones .
can then the progenitors of the compact etgs be found among them ?
the lbgs are star
forming galaxies and thus continue to increase their stellar mass until they quench .
thus , the question can be reformulated as whether there is enough time between @xmath6 and @xmath9 for the lbgs to cessate their star formation activity , reach a @xmath52 low enough to satisfy the definition of passivity given above , and build up enough stellar mass to reproduce the distribution of the etgs . in the appendix
we discuss possible quenching paths that the lbgs need to follow to be classified as passive at @xmath9 according to our definition . to summarize , from the knowledge of the stellar mass and the star
formation rate of the individual galaxies at the time of the observation , and assuming a functional form for the star - formation history during the quenching phase , we can predict the final stellar mass and @xmath52 of the former lbg once they have quenched .
this calculation shows that there are indeed plausible quenching scenarios that could evolve some of the lbgs in our sample into passive galaxies as defined above . since the quenching phase is believed to be fairly rapid ( e.g. * ? ? ?
* ) , in the calculation we model it with a decreasing exponential function @xmath63 , with the time
scale @xmath42 equal to 100 myr for all galaxies .
this timescale was estimated using the sound crossing time in compact galaxies ( we used @xmath64 km s@xmath28 for the sound speed in the ism of lbgs , as discussed in more detail in the appendix ) .
we also assumed that the declined phase of the sfr began at the time of observation , i.e. @xmath65 .
as discussed in the appendix , the time difference between @xmath6 and @xmath66 is sufficiently large that from the point of view of the selection of the candidate progenitors , these are conservative assumptions since both a later quenching time @xmath67 and a longer @xmath42 would still result in galaxies that are passive according to our criterion while yielding larger stellar mass and hence increases the number of candidates .
but just comparing the two stellar mass distributions of descendants and candidate progenitors ( after they have finished forming stars ) is not sufficient to set up physically motivated selection criteria . a key property to consider when trying to identify the progenitors of the compact etgs is the morphology of these systems , and specifically their very high stellar density .
the question we are posing is : can star forming galaxies at @xmath68 with any morphology be the progenitors of the compact passive ones ? or only galaxies with certain morphology can evolve into such systems , either via merging or via in situ star formation ? based on n body simulations of binary merging events , @xcite suggested that the compact etg progenitors are to be searched for among star
forming galaxies that are compact themselves and also have large gas fractions , i.e. @xmath69% . these progenitors in turn can form most of the stellar mass of the final descendant in a highly concentrated merger triggered burst .
@xcite also argue that a powerful nuclear starburst in a merger remnant producing most of the stellar mass of the remnant itself will result in a galaxy that , once passive , will resemble the compact ones .
as we will discuss later , however , it is reasonable to believe that , whether most of the final stellar mass is formed during a merger induced nuclear burst or in situ star formation , the progenitors must be compact themselves , and must have stellar densities comparable to or higher than the descendants , and with similar stellar mass profiles .
in other words , it is not reasonable to expect that star forming galaxies at @xmath18 whose light profile is more diffuse than the @xmath9 compact passive ones can be their progenitors , even if most of the stellar mass of the final descendant is produced after the epoch of observation in a compact region .
this is true both for the case of a single galaxy that forms stars and quenches star formation , in situ , or for galaxies that merge .
the physical reason is that an existing diffuse stellar component can not be shrunk into a compact one and it can not be hidden from observations either , regardless if a new , more massive compact stellar component is subsequently created ( note that mergers puff up compact structures into diffuse ones , not vice versa ) . to first order , if a diffuse component is observed in the candels h band images in a star forming galaxy at @xmath6 , the same diffuse component will also be observed in the same image if the same galaxy were placed at @xmath9 after quenching its star formation activity .
this is in part because of the lower redshift and in part because of the increase in stellar mass between the time when the galaxy is observed and when the quenching process is complete . given that the quenching phase is not instantaneous , candidate progenitors must have 1 ) smaller stellar mass than the passive ones , but star
formation rate such that after they quench the final stellar mass density reproduces that of the passive ones ; 2 ) morphology and size similar to that of the etgs , since the ongoing formation of stars does not change the dynamics of the pre existing stellar orbits and hence the appearance of the galaxies . here
we will explore a scenario where these progenitors , therefore , essentially evolve at constant size by converting gas into stars while their stellar mass and density increases , therefore maintaining the compact light profile observed in the compact etg .
gyr@xmath28 ( purple and red points ) .
stellar mass of lbgs increases according to the assumed sfh and no size evolution ( i.e. each projected purple and red point has a corresponding black cross at the same half - light radius
. see text for details ) .
candidate etg progenitors are selected from this sample ( red ) as those lbgs whose projected mass puts them in the compact etg selection window . ]
it is important to understand that we are not postulating that all galaxies that satisfy the two general requirements above will evolve by growing their stellar mass at essentially constant size . on average ,
galaxies evolve by growing their stellar mass and enlarging their size , as shown by the existence of a mass size relationship and the overall size evolution of star forming galaxies ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
but not all galaxies must do so , or must all follow the same mass size growth path , as evidenced by the scatter in the mass size relationship itself . nor
are we saying that those lbg that satisfy the two points above quench their star formation after the epoch of observation and appear passive ( as per our definition ) at @xmath23 .
what we are saying is that if lbgs at @xmath6 include progenitors of the compact etgs at @xmath23 , then the morphological properties of the latter imply that the former must grow in mass at essentially constant size , as well as have stellar mass and sfr such that under general assumptions about their star formation history subsequent to the epoch of observation , they quench and are passive at the epoch when the compact etg are observed . to identify such lbgs ,
if they exist , we must model the quenching of their star formation activity .
thus , we need to assume 1 ) the time when they begin quenching , and 2 ) how they quench , i.e. the form of the declining star formation history
. the quenching phase can start at any time after the epoch of observation of the lbg , or even slightly before , since a galaxy in the early phase of declining star formation would still be classified as a `` lyman break galaxy '' as long as this phase is not too advanced .
the details of star - formation history of galaxies during the quenching phase are not known @xcite , and there are suggestions that star - forming galaxies can have bursty and chaotic sfhs @xcite . therefore a simple function such as an exponential decay is almost certainly an over simplification , especially on short time scales . if the goal is to model the formation of etgs , however , all that is relevant is that the star formation rate overall decreases on a relatively long time scale , namely long compared to the time scale of rapid bursts ( i.e. a few @xmath70 yr ) , since their red colors ( at the time of observation ) , high masses , and very low ssfrs imply that they must have been actively star forming at least 1 gyr earlier @xcite .
previous studies investigating the sfhs of the compact etgs , have supported this interpretation @xcite [ !
t ] it is important to keep in mind that the decaying exponential function that we used to model the quenching phase of the lbg s star
forming activity has nothing to do with the function adopted to describe their star formation history up to the time of the observation .
the latter is used to model the assembly of the galaxies that we observe and _ only _ to estimate their stellar mass at the time of observation ( recall that we tested both an exponentially declining and a delayed increasing function when measuring the observed stellar mass , with similar results ) . to describe the sfh of the quenching phase , which begins at or after the time of observation , we use only the exponentially declining sfh with @xmath42 = 100myr as described in the appendix , and this is used to estimate the additional stellar mass produced after the time of observation and during the quenching phase ( we do not make an attempt to incorporate mass loss from stars ) . this stellar mass
is then added to the mass of the lbg formed up to the time of observation to derive the projected mass distribution of the candidate progenitors of the compact etgs and compare with the observed one for compact etgs ( see figures [ sizevsm ] and [ dmass_hist ] ) . in any case , the combination of the time when they start quenching and the duration of the quenching phase must be such that by redshift @xmath71 the galaxies would be observed as passive , according to the criteria discussed above , and that the stellar mass distribution must reproduce that of the compact etgs . thus , once we assume the starting time of the quenching phase and the time scale @xmath42 ( e.g. the limiting case of the sound crossing time , @xmath72 myr ) , from the measured ( photometric ) redshift , sfr and stellar mass of the lbg , we can estimate the redshift at which the galaxy will satisfy the condition for passivity , as well as its stellar mass at that time .
additionally , from the measured size of the lbg , and under the assumption that the galaxy evolves at constant size , we can estimate the stellar density when the galaxy is passive and see whether or not it matches the projected stellar density of the compact etgs , i.e. @xmath73x@xmath74kpc@xmath30 .
we present our compact progenitor selection in figure [ sizevsm ] , which in the top panel shows the mass - size relationship for the lbgs as observed , and also the @xcite etg sample and its selection criteria . in the bottom panel ,
we additionally plot where the same lbgs will lie in the mass - size diagram assuming their projected stellar mass , meaning , their expected mass after they satisfy the condition of passivity , namely @xmath75 gyr@xmath28 .
each lbg accumulates stellar mass according to its observed sfr and assumed sfh .
the distribution of final projected mass compared to the observed mass is shown in figure [ dmass ] , and it should be noted that a large fraction of the projected final mass of most galaxies must be extrapolated using the observed sfr and assumed sfh . those 44 lbgs with projected properties which meet our compact etg selection criteria are our sample of candidate compact etg progenitors . the rest of the lbgs ( 136 ) , which end up either less massive or with lower stellar density , are non - candidates .
we additionally note in this figure the existence of 11 lbgs from our sample who are _ already _ compact in stellar density , as observed , at @xmath76 . as might be expected , the candidate plausible progenitors tend to have higher sfrs than the non - candidates .
this trend can be seen in the sfr - m * relation for the two samples , which is shown in figure [ massvssfr ] .
there is one galaxy in the candidate sample for which we have estimated a rather large sfr ( @xmath2210@xmath77 ) .
we have investigated the sed of this galaxy for any signature of agn , and found that it is not detected at 24@xmath32 m , nor does it have an x - ray detection in the _ chandra _ 4ms image .
this galaxy appears to simply be one of the redder galaxies in our candidate lbg sample , hence its high sfr .
even if our dereddening procedure overestimates the sfr of this object , which we think is quite likely , this galaxy will regardless end up in the candidate sample due to its high stellar mass and compact size already placing it in the candidate selection window .
its exclusion , or inclusion , does not significantly change the results presented in the following sections .
we also note that the candidates do not differ in their observed mass distribution from the non - candidates .
this can be further seen in their color - mass diagram for the samples considered here , shown in figure [ etgcolor ] .
we additionally present in figure [ etgcolor ] the rest - frame u - v vs v - j color - color diagram , showing the colors of the three samples of star - forming and passive galaxies .
the color distributions are important for understanding inherent differences in the two samples of lbgs , as we will show in section 4 . [ ! t ]
as mentioned in section 2 , we have fit stellar population synthesis models @xcite to each galaxy .
the best fitting sed template to each galaxy , found as explained in section 2 , has been used to generate an average sed template for candidate and non - candidate samples .
we restrained ourselves to those lbgs in both samples which have spectroscopically confirmed redshifts ( @xmath78 of all lbgs , @xmath79 of candidates , @xmath80 of non - candidates ) , as this will minimize the diluting effects of including objects with only photometric redshifts which have larger uncertainties .
we have also derived average empirical seds directly from the observed photometry , k - corrected to common restframe wavelength using the best fit templates .
the errors on the average empirical sed are estimated using the following procedure . with each galaxy
s observed photometry , we produce a gaussian deviate of each photometric measurement , given that point s photometric error .
we then re - calculate the average empirical sed .
we do this procedure 10,000 times , and the standard deviation of the average empirical seds of the gaussian deviates is the error on the average . [ !
t ] 25 objects .
top panel : ssfr vs mass for the same galaxies . ]
t ] these two measures of the average seds for the two samples of spectroscopic lbgs are shown in the top panel of figure [ seds ] .
we have chosen to normalize the average seds at @xmath81 , the vicinity of the observed @xmath82-band , so as to emphasize differences in the uv and optical parts of the sed . at redshift @xmath83 , there are four prominent emission lines characteristic of star - forming galaxies which may enter the @xmath82 band ( or also the h - band , @xmath84,@xmath85,oii , and oiii ) , and in fact in a large fraction of our galaxies we do see an enhancement of the flux density in the @xmath82 and/or h - band , relative to the best fitting sed template .
therefore , we choose to normalize by the value of the best fitting template at 5000a , rather than bias the normalization high by using the observed flux density in the @xmath82 band .
the excess flux density of these contaminated photometric points with respect to other photometric points , and also the best - fit sed , suggests the photometry of some galaxies is affected by the presence of emission lines .
we investigate the extent to which these lines may affect the average seds in the bottom panel of figure [ seds ] by repeating the analysis after removing the individual affected photometric point from the galaxy s observed sed if one of the four lines listed above enters any bandpass at greater than 1% the maximum transmittance of the band .
the result of this test is shown in the bottom panel of figure [ seds ] .
it is clear that removing contaminated photometry brings the points in the average empirical sed that are based on the observed @xmath82 and h - band photometry into better agreement with the average sed template .
the difference between the seds of the compact candidates and non - candidates is still clear when removing the affected photometry . [ ! !
t ] the results shown in figure [ seds ] indicate that on average , the compact candidates with spectroscopic redshifts have a redder restframe - uv sed than the non - candidates , but an otherwise apparently identical optical one .
there are two explanations for a redder uv slope : an older population of stars ( for example , a more evolved burst of star - formation ) , or , a larger amount of dust obscuration ( on average ) .
the effects of age or dust on galaxy seds are generally degenerate and are notorious for confusing the measurement of physical properties of galaxies .
however , we argue here that the flatness and consistency of the average seds of candidates and non - candidates red - ward of 4000a argues in favor of the interpretation that the difference is due to a difference in average stellar ages ( since peak of sf activity ) , and not average dust properties .
figure [ modelseds ] illustrates how the uv and optical parts of @xcite sed templates of star forming galaxies vary with varying age , varying dust obscuration , and varying both age and tau for constant t/@xmath42 ( @xmath86 ) .
the model seds are normalized in the same way as the observed data , and the averages for candidates and non - candidates are included for comparison .
varying dust does not reproduce well the difference in composite seds , because of the similarities in the optical part of the seds of both candidate and non - candidate samples .
varying age is a better description of the observed trends
. a better look at the photometric differences between candidates and non - candidates and how they may vary with trends in age and dust can be seen in figure [ obscc ] , which shows the observed v - z and irac channels 1 - 3 colors of lbgs .
this compares colors blue - ward of the 4000a break with colors red - ward , which are the key features of the differing average seds .
there is overlap in color distributions of the candidates and non - candidate samples , but the mean of the candidates are offset red - ward in v - z from the non - candidates , but not offset red - ward in irac colors .
the same color distribution is present among the lbgs which have photometric redshifts ( bottom panel ) , indicating this difference in color persists among the entire lbg sample .
tracks of a sample @xcite template sed with varying age and one with varying dust obscuration show that varying age changes v - z color but not irac colors , while varying dust changes both colors .
this is consistent with , although does not prove , a difference in the average stellar age ( since peak of the starburst ) between the two samples , with the candidate etg progenitors appearing older than non - candidates .
we note that similar trends are also seen in the u - v vs v - j color - color diagram presented in figure [ etgcolor ] , where candidates tend to be redder in u - v , but do not appear redder in v - j color , than the non - candidates .
although the dispersion in colors between candidates and non - candidates in figure [ obscc ] overlap , we note that the difference in average seds are significant , as determined by the more robustly determined errorbars from the simulations presented in in figure [ seds ] . here
we present further investigation of the potential contribution of dust to the redder restframe - uv sed of the candidates as compared to that of the non - candidates , using the _
spitzer_/mips 24@xmath32 m
* dickinson et al in prep ) and _
100@xmath32 m data @xcite .
the 3@xmath29 detection limit of the 24@xmath32 m catalog is 20 @xmath32jy , and for the 100@xmath32 m catalog ( based on prior information of the 24@xmath32 m catalog ) is 0.8 mjy @xcite .
generally the lbgs have a low detection rate ( @xmath87 ) at both 24@xmath32 m , and 100@xmath32 m . of the candidates , 4@xmath41 are within 1 arc - second of a 24@xmath32 m detection , and 2@xmath41 are within 1 arc - second of a 100@xmath32 m detection . of the non - candidates , 8@xmath41 are within 1 arc - second of a 24@xmath32 m detection , and 3@xmath41 are within 1 arc - second of a 100@xmath32 m detection .
although the detection rate is low , we checked to see if these galaxies should have infrared fluxes above the detection limits of those surveys , given the amount of dust obscuration that can be inferred from their measured uv - corrected sfrs . to see
if our lbgs should be detected given this estimated amount of dust - obscured sf , we first estimate the infrared luminosity ( @xmath88(total)=@xmath89(8 - 1000@xmath32 m ) from the amount of obscured sfr in each galaxy .
the obscured sfr is derived from the uv - corrected sfr and the uv - uncorrected ( i.e. measured directly from the restframe - uv sed ) sfr . to estimate the expected flux densities at 24 and 100-@xmath32 m
, we use the infrared template of @xcite whose total infrared luminosity best matches that estimated for each lbg , convolved with each bandpass .
we find that less than 3(2)@xmath41 of the non - candidates are expected to be detected at 24(100)@xmath32 m , and less than 7(7)@xmath41 of the candidates would be detected .
we also test the inferred flux densities using the updated templates of @xcite for main sequence galaxies , and find comparable results .
these are generally consistent with our findings listed above for the actual number of lbgs with infrared counterparts . since the overwhelming majority of these lbgs are below the detection limit of the _ spitzer_/mips and _ herschel_/pacs surveys , we study the average dust properties with a stacking analysis . because the detection rate within a 1 arc - second search radius is so low ,
we adopt the following procedure for both wavelengths to carry out the stacking . for objects which are within one arc - second of a detection
, we use a 42x42 pixel image [ 50x50 arc - seconds ] from the real observed 24@xmath32 m or 100@xmath32 m image . for objects which are not detected ( more than 90@xmath41 )
we use a 42x42 pixel image from the residual map at 24@xmath32 m or 100@xmath32 m , where the flux from formally detected objects has been removed using the psf , following the methods of @xcite at 24@xmath32 m , and @xcite at 100@xmath32 m .
the residual maps thus includes low - level infrared emission from non - detected sources , and also noise , but no flux from neighboring detections .
using the residual map for non - detections minimizes the flux of nearby bright , but unrelated , infrared sources contaminating the lbg stacked flux .
we then stack the images of the candidates and non - candidates using a weighted averaging based on the rms maps in the case of the 24@xmath32 m stack and the weight maps in the case of the 100@xmath32 m maps .
stacked fluxes are determined by performing aperture photometry on the stacked images , and the published aperture corrections from @xcite and the _ herschel_/pacs technical documentation . the uncertainty on the stacked fluxes for candidates and non - candidates are determined by the following procedure .
we repeat the above stacking for the same number of random positions in the maps as galaxies in the candidate and non - candidate samples .
we generate 1000 sets of these random stacks , and use the standard deviation of the stacked fluxes of these random positions as the uncertainties of each sample s stacked flux .
the results from the stacking analysis are shown in figure [ stack100 ] .
we find that candidates and non - candidates have statistically indistinguishable , and non - significant stacked flux at 100@xmath90 . for candidates we find @xmath91jy and for non - candidates we find @xmath92jy .
the candidates have no significant stacked emission at 24@xmath32 m , while the non - candidates do have some significant stacked flux . for candidates we find @xmath93jy , and for non - candidates we find @xmath94jy , an approximately 4@xmath29 detection .
this supports our hypothesis that the candidates are redder because of older ages rather than dust , because the candidates do not show evidence of higher dust emission , and in any case , the non - candidates on average appear to have more dust emission than the candidates . [ ! t ]
100myr , e(b - v ) = 0 , varying age of the burst .
bottom left : model with @xmath95100myr and age = 100myr , varying e(b - v ) .
bottom right : model with e(b - v ) = 0 , varying both age and @xmath42 such that the ratio remains constant .
composite seds of candidates ( red ) and non - candidates ( blue ) are included .
varying dust does not reproduce well the variation of the uv , while keeping the optical - nir part flat.,title="fig : " ] 100myr , e(b - v ) = 0 , varying age of the burst .
bottom left : model with @xmath95100myr and age = 100myr , varying e(b - v ) .
bottom right : model with e(b - v ) = 0 , varying both age and @xmath42 such that the ratio remains constant .
composite seds of candidates ( red ) and non - candidates ( blue ) are included .
varying dust does not reproduce well the variation of the uv , while keeping the optical - nir part flat.,title="fig : " ] [ !
ht ] [ ! !
t ] of particular importance to studies of star - forming and compact galaxies , a class to which our candidate lbg sample belongs , is the contribution of an active galactic nucleus ( agn ) .
some galaxies whose flux is dominated by that of an agn may in effect _ appear _ compact in terms of bolometric output , but are not in fact compact in terms of their stellar density . rather , they may simply be out - shined by the agn .
x - ray detections amongst our two lbg samples thus play an important role in understanding the contribution ( if any ) to contamination in our sample from bolometrically dominating agn . to assess the fraction of agn present among the candidate sample , relative to the lbg sample as a whole , we have matched the lbg samples to x - ray detections from the 4 ms observations with the _ chandra _ x - ray observatory @xcite .
the overwhelming majority of the lbgs are not detected in the 4 ms observations , with the exception of 3 sources from the non - candidate sample .
similarly low detection rates for lbgs have been found by @xcite .
nevertheless , since the non - compact sample is larger than the compact one we can not infer anything about the x - ray emission frequency in the two samples .
we further assess differences in x - ray properties between the two samples using stacking of the 4 ms imaging . in both the candidate and non - candidate samples ,
the x - ray stack showed no statistically significant x - ray emission , with an upper limit on x - ray luminosity of 10@xmath96 erg / sec .
this result suggests that our candidate sample is not contaminated by agn any more than the non - candidates . in this section
, we study the morphology and size of the young and old stellar populations of the galaxies in each sample .
the aims of this analysis are 1 ) to explore faint surface - brightness features ( e.g. * ? ? ?
* ) and 2 ) to study if there are morphological differences in the two samples that can help us assess if their merging histories differ . since our objects are faint ( h@xmath3825 ) , and because we are interested in low surface - brightness features which may be too faint to observe on an individual galaxy basis , we study the average distribution of the stellar populations by stacking the two samples .
we stack the @xmath97/wfc3 h band , which is in rest - frame optical at @xmath2 and therefore probes the older stellar population .
we also stack the @xmath97/acs ( z band ) images , which is at the rest - frame uv and therefore probes the young stars and star - forming regions . to produce the stacked images , we generate images in each band , in which we have masked out any neighboring galaxy
which is not @xmath98 selected as a u
band dropout by our color selection . to mask the emission from these interlopers
, we use the corresponding segmentation map from the sextractor detection process in each band .
the images are then shifted so that each image is centered at the position of the peak of the h - band emission . to perform the stack ,
we do an inverse variance weighted mean of each pixel , using the map rms plus poisson noise to compute the weights .
this weighting scheme ensures that possible non - azimuthally symmetric low
surface brightness structures will be preserved in the final stack .
we then measure average structural properties of the galaxies using galfit , and make azimuthally averaged light profiles using the iraf function _
ellipse_. as mentioned in section 4.1 , given their redshift distribution , many of the lbgs may have optical emission lines entering the h - band ( primarily the [ oii]@xmath99 , but also h-@xmath100 ) , and this may bias the h - band light distribution towards the star - forming regions and therefore would not trace the older stars .
the j - band is largely unaffected by these emission lines .
we therefore compare the h - band stacks and j - band stacks to check for discrepancies , which may indicate contamination from this line emission .
the light profiles for the h - band and j - band , for candidates and non - candidates , is shown in figure [ jhprofile ] .
this figure shows that the h - band and j - band stacks , on average , have very similar profiles for each sample , showing that any contamination to the h - band from emission lines is negligible . [ !
t ] in figure [ hprofilesall ] , we compare the average , peak - flux normalized h - band light profiles for the candidates and non - candidates .
we additionally repeat the stacking using the @xcite compact etg sample between @xmath48 , where we mask out all neighboring galaxies prior to stacking .
this comparison between compact candidates ( red ) and non - candidates ( blue ) in this figure essentially reflects our selection criteria : candidates must be compact ( thus have smaller radii ) and tend to be more star - forming , resulting in a higher peak surface brightness than non - candidates .
( due to the flux normalization in figure [ hprofilesall ] , the peaks are coincident , but this fact can be seen from the absolute difference between peak and normalized noise level , and is also visible by comparing the un - normalized peak flux for candidates and non - candidates in the two panels of figure [ jhprofile ] ) .
we include in figure [ hprofilesall ] the stacked average light profile of ultra - compact , compact , and non - compact etgs as defined in section 3 from the @xcite sample , which compares the peak - flux - normalized shape of the light profiles .
the figure shows that not only do the ultra - compact etg sample and our lbg candidate sample have , on average , the same half light radius ( due to the way they have been selected ) , but also nearly identically steep light profile overall .
this presentation with the peak - flux normalization highlights the actual similarity in steepness of the light profiles .
it is important to note here that simply comparing sersic parameters does not adequately highlight the similarity because of the covariance of half - light radius and sersic index .
finally , we note that the average light profile of all the samples considered here shows no evidence of excess flux above a sersic profile , a signature which could imply the presence of tidal debris from recent merging . [ !
t ] as mentioned previously , of particular importance to understanding the mechanism by which these galaxies build up their stellar mass over time is the relative _ spatial _ distribution of young and old stars . in figure
[ hzprofile ] , we compare the average stacked light profiles of the restframe uv ( z - band ) with the restframe optical ( h - band ) for our candidates and non - candidates . in this figure ,
the differing psfs have been fully taken into account by galfit when estimating the sersic parameters , and are plotted after convolution with their respective psfs . in both candidates and non - candidates ,
it is clear that the uv flux is more centrally concentrated than the optical flux , as indicated by the fact that the uv half - light radius of each is smaller than that of the optical , lending support to our assumption in section 3 that these galaxies gain stellar mass through centralized , in situ star - formation , rather than in the outskirts for example , and therefore may build the stellar masses of galaxies at constant ( or modestly increasing ) size , consistent with the evolutionary path we assume in figure [ sizevsm ] ( for a similar discussion on how galaxies grow in mass and size see also @xcite , who , however , do not consider the case of compact galaxies as we do here ) .
if major merging has not been a driving mechanism , then the co - moving number densities of our candidate progenitors and that of the compact etgs should be comparable . in the literature , co - moving number densities have been used as arguments favoring massive dusty starbursts , such as ultra - luminous infrared galaxies ( ulirgs ) and sub - millimeter galaxies ( smgs ) , as being the progenitors of local massive ellipticals in the centers of galaxy clusters @xcite .
these galaxies are relatively rare ( @xmath101 mpc@xmath102 , @xcite ) like the high - redshift etgs ( @xmath103x@xmath104 for non - compact etgs and @xmath105x@xmath106 for compact etgs , @xcite ) .
see table [ table1 ] for a comparison of number densities .
due to their extreme nature , smgs and ulirgs may be capable of forming the requisite stellar mass at high - redshift @xcite , as well as have sufficient gas content to produce compact remnants @xcite .
( we will discuss the similarities between smgs and compact etgs further in section 5.2 . ) evolutionary scenarios for smgs can in principle be tested using galaxy clustering ( e.g. * ? ? ?
* ) , however current samples of high - redshift compact etgs is prohibitively small for this measure and comparison .
larger samples utilizing the full candels survey data may contain enough compact etgs for this purpose in the future .
for now , co - moving number densities provide a first order consistency check .
it is important to keep in mind , however , that an accurate comparison of the spatial abundances of the two populations would require not just quantifying the effective amount of major merging that takes place both before and after the star
forming galaxies quench , but also the fact the lbg selection does not recover all the star forming galaxies at @xmath6 ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in any case
, if our sample of compact candidates truly contains the progenitors of compact etgs , of which the majority grow their stellar mass in situ ( i.e. steady growth independent of merging ) and evolve , by quenching , into compact etgs , then the co - moving number densities of compact candidate lbgs must at least be large enough to account for the observed number densities of compact etgs .
we measure the co - moving number density of our lbg samples by directly integrating the redshift distribution to get the cosmological volume sampled by the galaxies : @xmath107 where @xmath108 is the redshift distribution of each lbg sample .
we then use this to get the observed number density of lbgs in each sample : @xmath109 where @xmath110 are the number of lbgs in each sample .
the co - moving number density of the @xcite compact etgs ( stellar density @xmath21 3x@xmath111 @xmath50kpc@xmath30 ) is @xmath112x@xmath106 mpc@xmath102 between @xmath60 .
co - moving number densities of other samples of compact etgs from other samples have similar values within this redshift and mass range @xcite .
we find that our candidate compact etg progenitors have a co - moving volume density of @xmath113x@xmath114 mpc@xmath102 , consistent within the uncertainty with the volume density of the @xcite sample .
more specifically , the compact candidate sample can account for @xmath2292@xmath41 of the compact etgs by number density found at lower redshifts .
although within the uncertainty this can account for all detected compact etgs , this does not account for the destruction of compact galaxies through merging or rejuvenation of star - formation between @xmath2 and @xmath9 . in section 4.4 , we will discuss a fraction of compact z@xmath223 galaxies that we find are missed by our lbg selection criteria , and how this fraction affects the above estimate of number density .
we have chosen to use the lbg color selection for our sample in the above analysis , because of its efficiency , lack of interloper contamination ( with u band dropouts the only modest source of contamination is that by galactic stars , which are easily identified in the _ hst _ images and removed ) and it is largely model independent ( see * ? ? ?
it also avoids the potential bias due to the degeneracy between dust obscuration and age , which could result in including galaxies with low specific star
formation rate among the progenitors . in any case
, a more general search for candidate progenitors can be done using a sample selection based on photometric redshifts and subsequent sed fitting to spectral libraries to derive stellar mass .
in fact , a fraction of @xmath2 galaxies with spectroscopic redshifts are missed by our color selection , primarily due to the fact that they reside in crowded fields where their tfit photometry may be affected by nearby u band detected galaxies , causing them to be excluded from our sample of u band dropouts .
therefore , we briefly present again here our main results for candidates and non - candidates selected in an _ identical _ way as the lbg candidates and non - candidates , but this time from all galaxies with photometric or spectroscopic redshifts between @xmath115 .
we will call this sample our sed - selected sample . [ !
t ] generally , the sed - selected samples are larger , with the number of non - candidates increasing by @xmath116 and the number of candidates almost doubling , with an increase by @xmath117 . not surprisingly , there is some significant overlap between the lbg and sed - selected samples . of the 136 lbg - selected non - candidates , 113 are in the sed - selected non - candidates . of the 44 lbg - selected candidates , 42 of them are in the sed - selected candidates .
( the two lbgs not present in the sed sample are galaxies with photometric redshifts between 2 @xmath38z@xmath38 3 discussed in section 3 ) .
therefore we stress these are not independent samples , and therefore this is not an independent test of our results .
rather , adding the results from the sed - selected sample serves to augment our analysis with an increased sample size that is not biased by the fact that the lbg selection selects galaxies on the basis of their uv emission , and therefore can not be too dusty .
we additionally note that while the lbg samples studied in section 4 have similar 24@xmath32 m detection rates ( @xmath87 ) , the sed - selected non - candidates remain with a low - ir detection ( @xmath118 ) and the sed - selected candidates jump to a @xmath119 ir detection rate . lccccl sub - millimeter galaxies & as observed & @xmath120 x @xmath104 & e.g. @xcite & + sub - millimeter galaxies & incl .
duty cycle & @xmath120-@xmath121 x @xmath122 & @xcite , & + & & & @xcite & + ulirgs & as observed & @xmath123 x @xmath104 & e.g. @xcite & + ultra - compact etgs & @xmath48 & @xmath124 x @xmath104 & @xcite & + compact etgs & @xmath48 & @xmath112 x @xmath106 & @xcite & + non - compact etgs & @xmath48 & @xmath125 x @xmath104 & @xcite & + compact candidate lbgs & @xmath115 & @xmath126 x @xmath106 & this study & + non - candidate lbgs & @xmath115 & @xmath127 x @xmath106 & this study & + all h @xmath38 25 lbgs & @xmath115 & @xmath128 x @xmath106 & this study & + we present in figure [ sedsamplesed ] the same analysis of average seds presented in section 4.1 , this time using the sed - selected candidate and non - candidate samples .
the top panel shows that candidates are redder , in both the restframe - uv and the optical parts of the spectrum , suggesting a dustier average sed .
this is no doubt reflecting the increased fraction of ir detections among the candidates using the sed - selection .
however since the number of ir detections in each sample is still a small fraction of the total , we have repeated the analysis without including the ir - detected galaxies .
the bottom panel shows this comparison of non - ir detected candidates and non - candidates , and shows the same signature of older stellar ages that was seen for the lbg sample in figure [ seds ] .
this figure , combined with the results of section 4 , are evidence that @xmath115 galaxies which are selected to be compact in stellar density on average have a redder sed that is best explained by an aging stellar burst .
this supports our key result found for the lbgs in section 4 , namely that there is a correlation between the compactness of galaxies , and the average age of their forming stellar populations , the sense being that more compact galaxies have older forming populations , i.e. the starburst is older .
this sed - selected sample of compact candidates has a larger co - moving number density of 2.6x10@xmath129 mpc@xmath102 , compared to 1.2x10@xmath130 mpc@xmath102 for the lbg - selected compact candidates .
this number density of sed - selected compact candidates ( which includes @xmath117 of the lbg candidates ) is a factor of @xmath222 larger than the compact etgs .
the key point of this paper is the identification of physically motivated candidate progenitors of the massive , compact etgs observed at @xmath131 . both individual images and very deep stacks show that the light profile of the massive compact etgs does not have any diffuse light in excess of their extremely compact core , which is approximated by a steep sersic profile with typical parameters @xmath132 and @xmath133 kpc .
simulations show that merging events rearrange the stellar mass profile of the merging partners in a way that the profile of the merger remnant is more diffuse than that of the initial partners ( e.g. * ? ? ?
* ; * ? ? ?
dissipative processes in a wet merger might channel gas to a nuclear region and produce a massive , compact component in a starburst episode , however the pre - existing diffuse component would remain visible and none is observed .
this implies that star forming progenitors of the compact etgs must be at least as compact themselves , and thus we hypothesized that their progenitors may be found among compact star - forming galaxies at higher redshifts .
we looked for such progenitors and found some plausible candidates among lbg at @xmath6 .
the candidate progenitors have been chosen to have similar size and morphology to the etgs at @xmath134 , and mass and star formation rate such that after they quench their star
formation activity and would be classified as `` passive '' , e.g. according the definition given by @xcite , their stellar mass and projected stellar density are in line with the analogous properties of the etgs .
it is important to understand that the selection of the candidate progenitors depends only mildly on the details of the assumed time evolution of the quenching phase of the lbgs ; much more important parameters in determining if a given lbg is a candidate progenitor or not are its physical size , its star formation rate and , to a lesser extent , its stellar mass at the time of observation , i.e. at @xmath135 .
all of the observed @xmath134 massive passive galaxies with stellar mass @xmath14 m@xmath5 can be accounted for with progenitors selected from lbg ; we found that a factor of 2 more candidate progenitors are found if more general selection criteria for star forming galaxies are adopted . with the addition of these new candidates ,
the evolutionary constraints between these two populations is relaxed somewhat ; up to half of them may merge and increase size , rejuvenate their star - formation or fail to quench , and there are still sufficient compact candidates to account for the formation of compact etgs .
the recent study by @xcite , following a similar methodology to that presented here , found relatively few galaxies ( 5 ) which may be considered progenitors of compact etgs , ( although with higher mass limits of m@xmath19 m@xmath5 for etgs , and initial z@xmath213 star - forming samples with @xmath136 ) .
their conclusion for these more massive samples differs from ours , in that a significant fraction of progenitors of these more massive compact etgs must be created between 2 @xmath38 z @xmath38 3 ( they suggest through merging ) . their sample is considered compact according to our criteria , and therefore likely overlaps with our sample on the massive end .
we note that we do find a similar number ( 9 ) of sed - selected massive ( m@xmath20 m@xmath5 ) galaxies that are already compact at z@xmath213 . to quantify any differences in the buildup of high - redshift etgs as a function of stellar mass , for example if merging must contribute to the massive end as suggested by @xcite , relative to the more general etg samples studied here where merging is not required
, it will be necessary to study larger samples including other candels fields in the future . [ ! !
t ] we subsequently studied the properties of the candidate progenitors and compared them to those of the non - candidate lbg .
the most remarkable difference is that the candidate progenitors have significantly redder rest frame far
uv colors than the non - candidates but essentially identical optical sed .
the mid ir properties of both types of galaxies show that larger dust obscuration of the candidates vs. the non candidates is unlikely to be responsible for the difference .
if anything , the non candidates have larger dust luminosity given their activity of star formation .
this forces us to conclude that the redder uv is explained by an older burst ( or period ) of star formation , namely the phase of star formation of the candidates has progressed more toward quenching it .
this is consistent with the general finding that a compact stellar morphology is the best predictor of passivity , i.e. early star
formation quenching , among massive galaxies at @xmath137 ( e.g. * ? ? ?
the fact that position of the candidate progenitors on the main sequence is higher than that of the non candidates is particularly interesting in the context of the discussion of the main sequence of galaxies presented in @xcite .
the arguments presented there suggest that galaxies on the main sequence with above average sfr must quench rapidly and early , resulting in a generally shorter lifetime of star - formation .
it is therefore interesting that not only do we find that our candidate sample follows such an elevated distribution of star - formation rates on the main sequence relative to the non - candidates , as seen in figure [ massvssfr ] , but we also find that they appear to show signs their star - formation will shut down . qualitatively speaking , the candidates are consistent with this scenario outlined in @xcite for a more rapid evolution for galaxies which follow the main sequence with higher sfrs . the rest
frame uv and optical morphology of both the etgs and of the lbg candidate progenitors show the lack of any diffuse component in excess of the very compact main body of the galaxies .
this is observed both in individual galaxies and in deep stacks .
additionally , lbg candidate progenitors and ultra - compact etgs show similarly steep profiles ( see figure [ hprofilesall ] ) .
the fact that star formation is observed to take place only in very compact regions in the lbg progenitor candidates and that no `` halo '' stars ( down to a limit of @xmath2229 magnitudes per square arcsec ) , i.e. no stellar component in excess of the steep light profile of the compact body of the galaxies , are observed in the passive ones is consistent with our suggestion that the growth of stellar mass in these galaxies takes place at essentially constant size , i.e. the stellar density increases with the stellar mass . in turn
this is indicative that star formation takes place in these systems via a highly dissipative accretion of gas .
we speculate in the following sections that this might be consistent with some ideas on how cold accretion proceeds in massive galaxies ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the existence and the spatial abundance of compact candidate progenitors shows that there are sufficient ordinary star - forming galaxies with small size and large stellar mass and star formation rate that can evolve through in situ star - formation to form compact etgs , without having to invoke special mechanisms that in a relatively short timescale can both quench a star - forming galaxy , while at the same time change its morphology into a compact one , e.g. gas - rich mergers with differential dust obscuration to hide the extended halos ( see * ? ? ?
* ) , to explain the emergence of compact etgs .
rather , simple secular evolution through in situ star - formation is sufficient , provided a quenching mechanism exists that can sufficiently remove the cold gas supply rapidly from these galaxies .
we discuss various arguments in favor of this scenario below .
ulirgs and smgs are often identified as the progenitors of local etgs , as well as the compact etgs at high - redshift , based on number densities @xcite , velocity - size relations @xcite , consistency with the faber - jackson relation @xcite for local ellipticals @xcite , and clustering @xcite ( although see * ? ? ?
* for the limitations of smg clustering measurements ) .
the connection between these ir - luminous galaxies and etgs is based primarily on the high sfrs and stellar masses that are typical of ulirgs and smgs , which are capable of producing the requisite stellar mass of compact etgs in short timescales .
since these galaxies are commonly believed to be the result of major mergers , the quenching of star - formation and morphological transformations can be easily tied to the merger . however , studying the stellar distributions of these galaxies is difficult due to their large dust obscurations .
this means only small numbers of smgs have been mapped with hst at near - ir wavelengths to measure the spatial extent of the bulk of the stellar mass .
the results from near - ir mapping indicates the morphologies of smgs are rather heterogeneous , including smgs that are large and irregular with multiple components , @xcite , dominant components which are disk - like @xcite , and are on average larger in size than compact etgs @xcite .
similar results have been found for ulirgs @xcite .
while a large fraction of these objects appear to be gas - rich mergers @xcite , collectively these studies suggest smgs and ulirgs display a large variety of morphologies and sizes , and as a population are not similar in morphology to compact etgs . while the smg or ulirg merger phase is likely a plausible avenue for quenching massive galaxies ( e.g. * ? ? ?
* ) , its unclear how the average smg can decrease in half - light radius by a factor of 2 or more after the star - formation has been quenched in order to form a compact etg with the average properties of the @xcite samples .
the more ordinary lbgs are generally smaller in radius compared to star forming galaxies at lower redshift with similar mass @xcite , but are larger than the compact etgs , as seen in figure [ sizevsm ] . despite this , our sample of compact candidate progenitors make up the low end of the size distribution of @xmath68 lbgs . how do such small but massive galaxies form ? according to major merger simulations ( see e.g. * ? ? ?
* ; * ? ? ?
* ) , with mass ratio from unity to @xmath138 , compact star - forming remnants may result from merging of galaxies with very large gas fractions , e.g. larger than @xmath139% .
a compact remnant is formed when a sizable mass of new stars is formed at the center of the new structure by the highly dissipative gas .
however , an extended stellar halo made by the older stellar populations will remain .
partners that are compact at the onset of the merging event , as well as a continuous accretion of gas as the merging progresses , increase the mass fraction of the remnant in the compact structure relative to that in the halo . but according to the simulations , however , a sizable fraction of the stellar mass of the remnant will be found in the halo , in general disagreement with observations of compact etgs ( some have resorted to hide the extended halo with dust obscuration to bring the simulations in better agreement with the observations ( see e.g. * ? ? ?
in a simplified sense , gas rich mergers produce two spatially segregated stellar populations : a centralized starburst , embedded in an extended older population .
in fact , a key feature of gas rich merger remnants seen generally among simulations is the need for a two - component fit to the light profile : a centrally steeper one to account for the starburst component and a shallower more extended one to account for the remnant stellar component @xcite . when fitting single sersic profiles to simulated gas - rich mergers , @xcite find that the segregation of older and younger stellar populations , primarily driven by the combination of light excess of the dispersed older stellar populations at large radii and central cusp from the starburst , causes the sersic fits
to be better approximated with large values of sersic index , @xmath140 . as a result
, they find the following distribution of sersic indices and effective radii from gas - rich merger simulations : merger remnants are best fit by cuspy high sersic index fits ( @xmath141 10 ) , along with large radii ( driven by the extended component ) too large to be consistent with compact etgs ( * ? ? ?
* their figure 12b ) . compared with observed properties of the etg sample of @xcite
, there appears to be some disagreement between the observed measures , and those measured from gas - rich merger simulations . for comparison
, we have constructed a similar plot ( figure [ nvsre ] ) showing the distribution of sersic indices , as a function of effective radius for the candidate sample and also the etgs of @xcite .
we additionally plot the measured properties from the h - band stacks for the candidate sample and compact etgs shown in figure [ hprofilesall ] , as the stacks are more sensitive to the presence ( if any ) of extended low - surface brightness structures that would increase effective radius .
these measures of the light profiles of compact etgs represent an improvement in sample size , and in the case of the stacks , sensitivity to deviations from a sersic profile and low surface - brightness features , relative to the comparison made with @xcite .
middle panel : same plot with the @xcite sample , split according to the compactness criteria .
bottom panel : the distribution of effective radii of the @xcite models presented above ( black and grey ) , as modeled by sersic profiles , along with direct measures of half - mass radii , half - light radii of the stellar component , and half - light radii after incorporating dust attenuation .
, title="fig : " ] .
middle panel : same plot with the @xcite sample , split according to the compactness criteria .
bottom panel : the distribution of effective radii of the @xcite models presented above ( black and grey ) , as modeled by sersic profiles , along with direct measures of half - mass radii , half - light radii of the stellar component , and half - light radii after incorporating dust attenuation .
, title="fig : " ] the distribution of our measures also do not coincide with the simulated gas - rich merger measures , in that our galaxies are well - fit by single sersic profiles , and do not show cuspy centers or extended light at large radii .
we also see no evidence of these features in any of our stacks , with all light profiles being well fit by single sersic profiles with @xmath1422 - 4 , down to the noise limit in our stacks , and in general agreement with the distribution of measurements from individual galaxies . to the extent
that the simulations represent real gas - rich mergers , our data of compact etgs and compact candidates ( including both individual sersic fits as well as those of the stacks ) appear inconsistent with the simulations .
we note that the distribution of radii of the models in the top and middle panels of figure [ nvsre ] is mainly driven by the fact that sersic models are poor fits to the merger simulations
. however , the actual simulated half - mass and half - light radii , without any assumption on the shape of the light profile , are still in agreement with the observations .
this can be seen more clearly in the bottom panel , which shows that typical half - mass and half - light radii of the simulations are of order @xmath221 kpc .
therefore , we can not rule out gas - rich merging based on this analysis .
major mergers are additionally characterized by tidal tails , debris , and other general disruptions of pre - existing galaxy components .
the timescales for dissipating these tidal features can be very long ( 1 gyr ) , with longer timescales correlating with larger gas fractions @xcite .
therefore if gas - rich mergers are the primary producer of compact star - forming galaxies at high - redshift , these features should still be present around our compact candidate sample .
since the stellar mass in tidal debris may make up a small fraction of total stellar mass ( especially in the case of tidal debris ) we estimate here exactly how much stellar mass in extended distributions our stacks account for , given the depth of the stack .
this thus provides an upper limit to the amount of residual stellar mass that can be lurking undetected around our galaxy samples . to estimate this , we take advantage of the relationship between our measured total h - band magnitude and total stellar mass , shown in the top panel of figure [ hmag ] .
this figure shows that the two are strongly correlated , for all three samples ( candidate lbgs , non - candidate lbgs , and compact etgs ) considered here , albeit with large scatter in the case of the total quantities for lbgs ( top panel ) .
nevertheless we can translate this relationship to magnitude per unit area and stellar mass per unit area , to relate surface brightness with surface density of stellar mass .
this relationship ( for both samples of galaxies together ) is shown in the bottom panel of figure [ hmag ] .
given the depths of our stacks , which go down to @xmath22 29 magnitudes per square arcsec , we estimate that the stacks are sensitive to surface densities of at least @xmath143 @xmath50-kpc@xmath30 .
this implies that , if our candidate galaxy samples are compact because of major mergers , the progenitor(s ) ( i.e. the non - dissipational old stellar component ) must contain less than this surface density of old stars in tidal debris or extended stellar halos .
[ ! t ]
@xcite present the range of surface brightness profiles of the extended stellar halos of gas - rich merger remnants , which are more extended than observed compact etgs ( their figure 13 ) .
we present in figure [ lastfig ] a comparison of these simulated merger remnants , along with the intrinsic sersic profile ( i.e. unconvolved with the psf ) fit to the stacks of the lbg samples ( top panel ) as well as the etg samples ( bottom panel ) .
we find that in all cases the intrinsic shape of the compact samples ( candidate lbgs , compact and ultra - compact etgs ) do not match the extended distributions of the simulated merger remnants .
while the absolute normalization of the observed light profiles should depend on galaxy flux , distance , and mass , and are not necessarily expected to match the simulations in magnitude , the steepness and shape of the profiles on large scales ( i.e. @xmath144r@xmath145 kpc ) do not agree , independent of this normalization .
in other words , the light profile of the compact galaxies is very often too steep relative to the extended distributions of the simulated merger remnants , suggesting that if this extended stellar halo were present in our samples , we would have detected it .
this is in agreement with the comparisons made in @xcite to the observed etgs of @xcite and @xcite , and we show that similar conclusions can be drawn for our candidate lbg sample .
the extended stellar halo characteristic of simulated gas - rich merger remnants appears to be absent from our observed light profiles for compact galaxies .
the steepness of the light profiles of these galaxies set important constraints that any mechanism of compact etg creation must satisfy , and we have argued here that this excludes large disky galaxies with shallower profiles from being compact etg progenitors .
the above discussion proposes that gas - rich major merging of galaxies with a sizable stellar component , i.e. observable in the _ hst _ candels images , is not the mechanism responsible for the formation of the massive , compact etg observed at @xmath146 .
we suggest that a more natural explanation of our observations may be that the star - formation in these compact galaxies is being driven primarily by accretion of cold gas , which efficiently forms stars centrally rather than forming stars in an extended disk .
the exact details of how the gas accretes have been discussed elsewhere by means of simulations and analytical calculations .
it has been suggested that the main physical mechanism is one where the cold gas dissipates angular momentum in a compact disk @xcite , and as more gas accretes the disk develops vdi @xcite that are very effective in driving the gas further down the bottom of the potential well , giving rise to a very compact structure .
direct cold mode accretion ( cma ) of the cold gas into the compact structure is also another mechanism @xcite that can give rise to very compact star forming galaxies @xcite .
both the vdi and the cma predict the formation of very compact star forming galaxies , with the vdi - driven wet inflow predicting a mixture of a perturbed disk and a rotating compact bulge .
it is important to keep in mind that current spectroscopic observations of both compact star
forming galaxies and passive galaxies @xcite at @xmath147 do not have sufficient spatial resolution to distinguish between a compact disk and a spheroid ( especially a rotating one ) , the kinematical signature of both structures simply being that of broadened emission and absorption lines .
regardless of the details of how the cold gas is funneled into very compact regions , the morphology of compact galaxies seem to require a highly dissipative mechanism for the assembly of their stellar mass as opposed to the merger of sub galactic structures with a sizable pre - existing stellar component .
t ] star formation in the compact galaxies is then subsequently quenched and they evolve passively since then .
recent studies have shown that star - formation can be quenched solely due to feedback from the highly concentrated stellar distribution .
two examples are stellar winds driven by intense starbursts @xcite , and internal ram pressure on dust grains @xcite . in fact
these feedback mechanisms imprint a maximum possible surface density of star - formation ( eddington limited ) @xcite , and evidence of this has been seen in extremely rare compact starbursts at lower redshifts @xcite .
other studies have shown that at high redshift compactness is the most sensitive statistical predictor of passivity among massive galaxies @xcite , a fact which is in broad general agreement with the fact that compact and ultra compact galaxies dominate the population of passive galaxies at @xmath18 , and with the finding that we have reported here that compact star
forming galaxies appear to have more evolved bursts compared to non compact ones . in conclusion
, we speculate here that the high - redshift compact etgs are the direct descendants of compact , star - forming galaxies , which themselves are compact because their star - formation is primarily driven by the accretion of cold gas to the central regions of the galaxy .
their star - formation is quenched due to their compactness because of stellar feedback ( e.g. * ? ? ?
* ) , halo quenching ( in the most massive cases ) @xcite , or some combination of both , and evolve passively after . if they undergo merging and or accretion their compactness is altered and they may end up forming a more diffuse light profile , and if compact star forming galaxies do not form anymore , then the number of compact passive galaxies keeps decreasing with cosmic time ( see cassata et al .
this scenario is generally supported by the distribution of stellar populations we present in figure [ hprofilesall ] , as well as studies of the main drivers of high - redshift star - formation @xcite . with this study
we have shown there are sufficient galaxies to supply the observed abundance of compact etgs this way , and that it is not implausible that compact etgs may be the descendants of compact star - forming lbgs .
the high - redshift etg sample of @xcite also contain some fraction of etgs which are non - compact , ( i.e. of size similar to local ellipticals ) . at the highest redshifts ( [email protected] ) the fraction is tiny , but
the number density of non - compact etgs increases dramatically to the present @xcite .
detailed high - resolution studies of local ellipticals have shown they are best described by multiple morphological components @xcite , even up to three and four sersic components @xcite , suggesting episodic periods of structural buildup .
other studies have proposed that the compact etgs are the cores of local ellipticals , with stellar mass buildup occurring in an inside - out fashion @xcite .
what is clear is that the sizes of etgs increase dramatically over time , in part because newly - formed etgs appear with progressively larger sizes as the universe evolves @xcite .
we suggest that these non - compact z@xmath211 etgs may form from an independent evolutionary track to the compact etgs , with non - compact etgs the result of high - redshift ( major ) merging activity . at z@xmath31 ,
the number density of non - compact etgs in the @xcite sample steadily increases with time , and has increased sufficiently to make up half of all m @xmath21 10@xmath4 m@xmath5 etgs by z@xmath221 .
incidentally , the number density of galaxies which are likely to be gas - rich mergers is similar to that of non - compact etgs at [email protected] ( see table [ table1 ] ) . with a constant merger rate with redshift , and assuming each merger quenches the star - formation ( at least for the next gyr or two before another merger rejuvenates star - formation ) , qualitatively it is plausible that this steady increase in non - compact etgs could be explained by a constant supply of mergers per unit time .
in fact , at z@xmath381 , during the buildup of these non - compact etgs , primarily dry mergers are found to account for both the assembly of the massive end ( m@xmath2110@xmath148 ) of the red sequence @xcite , as well as explain their morphologies as measured at low - redshift @xcite .
direct comparison of number densities to assess whether or not the actual number densities of mergers is sufficient to explain the increase in non - compact etgs at z@xmath211 will need to await accurate counts of mergers out to higher redshifts .
we have demonstrated the existence of a significant population of compact lbgs , which have consistent star - formation histories , stellar mass densities , and co - moving volume densities with high - redshift compact etgs ( the sample of * ? ? ?
we find that : * these candidate progenitors of compact etgs show distinct sed properties from the non - candidates , consistent with an older burst of sf , i.e. the burst appears to show evidence of fading . *
stacking from infrared images are consistent with this interpretation , and favor an older burst over an increased contribution from dust . *
the average x - ray properties of the compact and non - compact ones are consistent with each other .
one interpretation is agn activity has not influenced the selection of the candidates .
* structural properties of candidates and compact etgs differ from predictions of gas - rich merger simulations , suggesting this is not the dominant mechanism producing compact star - forming galaxies and compact etgs at high - redshift . *
we suggest the compact etgs are formed primarily through the quenching of compact star - forming galaxies whose in - situ star - formation is driven by cold accretion from the igm , via vdi and cma .
we speculate that merger driven evolution may contribute to the non - compact etg population at high - redshift .
we thank the anonymous referee whose valuable suggestions have improved this paper significantly .
mg and ccw wish to thank alvio renzini for many fruitful and inspiring conversations on this work .
we thank benjamin magnelli and ranga - ram chary for graciously providing 24@xmath32 m residual maps for our stacking analysis .
the study also benefitted from illustrative discussions with stacey alberts , alexandra pope , and arjen van der wel .
ccw thanks joseph meiring and dan popowich for technical consulting .
this work is based on observations taken by the candels multi - cycle treasury program with the nasa / esa hst .
we acknowledge support from grant program nsf ast 08 - 8133 , and support for _ hst _ program go 12060.10-a was provided by nasa through grants from the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas5 - 26555 .
we investigate first the general range of acceptable exponentially declining sfhs ( @xmath42-models ) that agree with the red colors and low ssfrs of the compact etgs .
the low ssfrs of etgs generally sets a limit to values of the decline - timescale , @xmath42 , and the time since the last starburst activity in the galaxy .
too recent a burst , or too long a value of the decline timescale , will leave too high of a ssfr in the galaxy . as the ssfr is a function of both sfr and also stellar mass ,
we investigate how the limiting value of @xmath42 changes as a function of both sfr and stellar mass . in figure
[ maxtau ] , we show how large @xmath42 can be for a @xmath76 galaxy , as a function of sfr and stellar mass measured at @xmath76 , and still have a measured ssfr @xmath149 gyr@xmath28 by [email protected] .
this figure shows that , for most lbgs in our sample ( median m * of @xmath150 , median sfr of @xmath2260 @xmath50/yr ) , the value of @xmath42 must be small ( @xmath151myr ) .
a larger @xmath42 would result in non - negligible ssfr , and exclusion of the galaxy by the compact etg selection .
not surprisingly , galaxies with lower initial ssfrs ( bottom right corner ) can tolerate higher values of @xmath42 and still be considered passive by [email protected] .
( in gyr , contours ) for exponentially declining sfhs .
if the value of @xmath42 is larger than this limiting @xmath42 for a given stellar mass and sfr , then a galaxy at @xmath2 will not meet the ssfr criteria of the etg selection , by @xmath1 . the majority of the parameter space occupied by our galaxies must have @xmath152 gyr . ]
we estimate a lower - limit to @xmath42 according to the following assumption . for the star - formation to decline in a galaxy , regardless of the quenching mechanism ,
the cold gas must be removed or depleted , and the shortest time in which this can happen is limited by the sound speed , @xmath153 , in the ism of a star - forming galaxy .
generally , this assumption would imply that each lbg has its own quenching timescale related to the diameter of the galaxy , such that sound crossing - time ( and hence quenching time ) is larger for larger - sized galaxies . for small galaxies such as the more compact ones we consider here
, this quenching time is relatively fast . for a typical sized lbg , diameter d@xmath154kpc
, we estimate the timescale @xmath155kpc@xmath156km / s@xmath157myr for cold gas depletion . in the analysis presented in this paper ,
we adopt this as our lower limit to @xmath42 .
however , to test the robustness against this assumption of a fixed minimum timescale for our sample , we note here that repeating the analysis using a sample selected using a different value of @xmath42 for each galaxy that depends on size as outlined above does not significantly change our results . with these constraints in hand ,
our strategy here is to use what we do know about the sfhs of etgs , and work backwards to gain some insight into which lbgs have consistent properties .
we acknowledge that any given lbg may or may not follow an
etg - consistent sfh between z=3 and z=1.6 , but note that any progenitors of etgs among the lbgs _ must _ follow an
etg - consistent one .
therefore , the real etg progenitors , if any exist among lbgs , will be contained in a sample selected this way .
we use the lower limit to @xmath42 calculated above , to select our sample , as it will result in the most conservative sample of plausible progenitors . by conservative
we mean it results in the smallest number of candidates , with the smallest increase in their stellar masses over the course of @xmath222 gyr .
repeating the analysis with a longer @xmath42 will only cause a net increase in stellar mass over this time , thus resulting in more massive galaxies , and additionally adding more candidates .
but , as figure [ maxtau ] suggests , alternative sfhs with slightly larger @xmath42 s are consistent with the observed compact etg properties , and so the effect of this assumption should be considered ( although the range in allowable values of @xmath42 is small )
. phrased in another way , this range in @xmath42 introduces a scatter in the _ actual _ amount of stellar mass that lbgs will form ( compared to that which we _ assume _ ) , between @xmath2 and @xmath1 . in figure
[ deltam ] , we assess exactly how much this scatter in stellar mass buildup is between our upper and lower limits in @xmath42 . to quantify the scatter , we calculate @xmath158 for a range of initial values of sfr and @xmath159 measured at @xmath2 , analogous to figure [ maxtau ] . for each region of the figure we use as the maximum value of @xmath42 the value derived in figure [ maxtau ] .
we find that the assumption of different @xmath42 values within the limits have less than a factor of two effect on the change in mass , compared to final mass , of the lbgs .
we can compare these values to the actual uncertainty in the estimate of stellar mass in lbgs , for example , according to the simulations of stellar mass estimates for lbgs in the analysis of @xcite .
the estimated error in stellar mass of lbgs is magnitude dependent , but may vary up to a factor of 2 for u band dropouts .
the scatter we find in final mass , depending on assumption of the sfh , is less than that scatter associated with the initial lbg mass estimate . therefore , the differences in extrapolated mass buildup from varying the assumed value of @xmath42 are not the dominant source of uncertainty in weather or not an lbg is included in the candidate sample .
we additionally note that we do not make an attempt to incorporate mass - loss from stars that is returned to the ism during the cycle of star - formation . in section 3 and the maximum @xmath42 allowable in figure [ maxtau ] ,
compared to the maximum amount .
the difference between the two is somewhat comparable to the uncertainties in stellar mass measurements in @xcite . ] , g. 2007 , in astronomical society of the pacific conference series , vol .
374 , from stars to galaxies : building the pieces to build up the universe , ed .
a. vallenari , r. tantalo , l. portinari , & a. moretti , 303 | we use goods and candels images to identify progenitors of massive ( @xmath0 ) compact `` early type '' galaxies ( etgs ) at @xmath1 .
since merging and accretion increase the size of the stellar component of galaxies , if the progenitors are among known star - forming galaxies , these must be compact themselves .
we select candidate progenitors among compact lyman break galaxies at @xmath2 based on their mass , sfr and central stellar density and find that these account for a large fraction of , and possibly all , compact etgs at @xmath1 .
we find that the average far uv sed of the candidates is redder than that of the non - candidates , but the optical and mid
ir sed are the same , implying that the redder uv of the candidates is inconsistent with larger dust obscuration , and consistent with more evolved ( aging ) star - formation .
this is in line with other evidence that compactness is a sensitive predictor of passivity among high redshift massive galaxies .
we also find that the light distribution of both the compact etgs and their candidate progenitors does not show any extended `` halos '' surrounding the compact `` core '' , both in individual images and in stacks .
we argue that this is generally inconsistent with the morphology of merger remnants , even if gas rich , as predicted by n body simulations .
this suggests that the compact etgs formed via highly dissipative , mostly gaseous accretion of units whose stellar components are very small and undetected in the hst images , with their stellar mass assembling in
situ , and that they have not experienced any major merging until the epoch of observations at @xmath1 . |
radiation from sase xfel consists of many independent spikes in both the temporal and spectral domains .
self - seeding is a promising approach to significantly narrow the sase bandwidth to produce nearly transform - limited pulses @xcite-@xcite .
we discussed the implementation of a single - crystal self - seeding scheme in the hard x - ray lines of european xfel in @xcite .
for this facility , transform - limited pulses are particularly valuable , since they naturally support the extraction of more fel power than at saturation by exploiting tapering in the tunable - gap baseline undulators @xcite-@xcite .
tapering is implemented as a stepwise change of the undulator gap from segment to segment .
simulation results presented in @xcite show that the fel power of the transform - limited x - ray pulses may be increased up to @xmath8 tw by operating with the tapered baseline undulator sase1 ( or sase2 ) .
in particular , it is possible to create a source capable of delivering fully - coherent , @xmath9 fs ( fwhm)-long x - ray pulses with @xmath10 photons per pulse at a wavelength of @xmath11 nm , fig .
[ f1 ] .
the method exploits a combination of cascade self - seeding scheme with single crystal monochromator and undulator tapering technique . the amplification- monochromatization cascade scheme is distinguished , in performance , by spectral purity of the output radiation and smaller heat loading of crystals .
, scaledwidth=100.0% ] we can apply the same scheme to harder x - rays , and obtain @xmath1 gw fully - coherent x - ray pulses at a wavelength of @xmath12 nm . in this paper
we propose to perform monochromatization at @xmath11 nm with the help of self - seeding , and amplify the seed in a first part of the output undulator .
the amplification process can be stopped at some position well before the fel reaches saturation , where the electron beam gets considerable bunching at the @xmath13nd harmonic of the coherent radiation .
a second part of the output undulator tuned to the @xmath13nd harmonic frequency , follows beginning at that position , and is used to obtain @xmath13nd harmonic radiation at saturation .
one can prolong the exchange of energy to the advantage of the photon beam by tapering the last part of the output undulator on a segment by segment basis .
[ f2 ] shows the design principle of our self - seeding setup for harder photon energy mode of operation .
two self - seeding cascades , identical to those considered in @xcite ( see fig . [ f1 ] ) , are followed by the same output undulator with changed gap configuration , compared to fig .
[ f1 ] .
kev.,scaledwidth=100.0% ] an advantage of the proposed scheme is the possibility to increase user capacity . in this paper
we describe a photon beam distribution system , which may allow to switch the hard x - ray beam quickly among many experiments in order to make a more effective use of the facility .
monochromaticity is the key for implementing multi - user operation in the hard x - ray range , which can be granted by using crystal deflectors and small absorption of the radiation in crystals at photon energies larger than @xmath14 kev .
in contrast to the broadband sase bandwidth , transform - limited hard x - ray bandwidths are of order of @xmath15 , and match the bragg width of crystals .
thus , using @xmath16 mm thick diamond crystals one may obtain two beams , one transmitted and one bragg reflected , with minimum intensity loss .
we suggest to flip crystals for switching reflectivity similarly as for polarization switching techniques with x - ray phase retarders at synchrotron radiation facilities , that is based on the use of piezo - electric components .
crystal deflectors can be repeated a number of times to form an almost complete ring .
it is then possible to distribute monochromatic hard x - rays among ten independent experiments .
in this section we describe a concept for a photon beam distribution system , which may allow to switch the fel beam quickly among many instruments in order to make a more effective use of the facility .
the high photon energy , and monochromaticity of the output radiation are the key for reaching such result .
the deflector can be simply switched off by tilting it , which results in a change of the angle of incidence of the radiation .
the angular change necessary for switching on and off the reflected beam is less than @xmath17 mrad .
only @xmath18 of incoming @xmath0 kev radiation is absorbed outside the reflection range . , scaledwidth=100.0% ] in fact , dealing with @xmath0 kev monochromatic radiation one may use crystals as deflectors , fig .
the deflector is constituted by a diamond plate with a thickness of @xmath16 mm .
the crystal is used in bragg reflection geometry and exploits the c(111 ) reflection plane . for the c(111 )
reflection , the angular acceptance of the crystal deflector is of the order of @xmath19rad , and the spectral bandwidth is about @xmath20 . as a result ,
the angular acceptance of the deflector is much wider compared to the photon beam divergence , which is of the order of a microrad .
the bandwidth of the hard x - ray pulse is of order of @xmath15 and matches the bragg width of the crystal . in this case ,
more than @xmath21 of the peak reflectivity can be achieved and only @xmath18 of the incoming @xmath0 kev radiation is absorbed outside of the reflection range .
thus , by use of a @xmath16 mm - thick diamond crystal one may obtain two beams , a transmitted as well as a bragg reflected beam , with minimal intensity loss .
one of the biggest advantages of using a crystal deflector in bragg geometry is that the reflected beam can be switched on and off by changing the crystal angle only .
the typical angular change necessary for the switching is less than @xmath17 mrad .
this opens a new possibility of fast switching of the reflectivity , which is necessary for many - user operation . in order to achieve a stable photon beam deflection
, the rotation error must be less than @xmath22 mrad .
existing technology enables rotating crystals to satisfy this requirements .
for example , at synchrotron radiation facilities , x - ray phase retarder crystals are driven by piezo - electric devices operated at hundred @xmath23 repetition rate , which flip the crystals with a rotation error of about a fraction of a micro - radian @xcite .
kev constitutes the key for reaching such result.,scaledwidth=100.0% ] a photon beam distribution system based on flipping crystals can provide an efficient way to obtain a many - user facility . a possible layout is shown in fig . [ f4 ] .
the output radiation passes through the distribution system , consisting of a series of crystals in bragg geometry .
photon macro - pulses at @xmath0 kev photon energy can then be fed into @xmath3 separate beamlines .
the switching crystals need to flip at frequency @xmath24 hz , so that each user receives one macropulse per second .
it should be noted that the single crystal provides a sufficiently large deflection angle ( of order of @xmath2 degrees ) , so that the problem of separation of neighboring beamlines does not exist .
the layout of the user instruments follows a similar approach as for synchrotron radiation sources .
, scaledwidth=100.0% ] a second possible layout of a future hard x - ray laboratory based on a hard - photon `` ring '' distribution system is shown in fig . [ f5 ] .
one can use a number of crystal reflectors to form an almost complete ring . the photon beam transport line guiding photons from sase1 ( or sase2 ) to the experimental hall
is connected tangentially to one of the straight section of the photon ring , and the beam is injected by the reflecting crystal .
using flipping crystals in each photon ring cell it is possible to quickly switch the photon beam from one instrument to the other , thus providing many - user capability .
this layout of the laboratory follows a similar approach as for synchrotron light sources .
finally , it should be remarked that the proposed beam distribution system operates at fixed wavelength .
however as reported in @xcite : `` for many scattering experiments it is not necessary to continuously vary the x - ray wavelength or fine - tune to a core shell resonance .
generally , however it is desirable to have a large photon energy exceeding @xmath3 kev .
this optimizes probing condensed matter systems on the atomic length scale by minimizing deleterious absorption , while preserving scattering cross sections '' .
obviously , our idea to increase the user access by means of a `` photon ring '' has advantages for such kind of experiments and can be applied to the european xfel as well as to the lcls - ii design .
in this section we report on a feasibility study performed with the help of the fel code genesis 1.3 @xcite running on a parallel machine . we will present a feasibility study for a short - pulse mode of operation of the sase1 and sase2 fel lines of the european xfel , based on a statistical analysis consisting of @xmath1 runs .
the overall beam parameters used in the simulations are presented in table [ tt1 ] .
we refer to the setup in fig .
[ f2 ] . up to the output undulator
, simulations are identical to those already presented in @xcite .
the choice of the fodo lattice parameters is also kept identical .
this is in agreement with the present concept to use the self - seeding setup at the european xfel in @xcite to produce monochromatic beams using harmonic generation simply by acting on the gap of the output undulator .
.parameters for the low - charge mode of operation at the european xfel used in this paper . [ cols="<,^,^",options="header " , ] [ tt1 ] the expected beam parameters at the entrance of the sase1 and sase2 undulators are shown in fig .
[ s2e ] , @xcite .
wakes inside the undulators are also accounted for and expected to obey the dependence in fig .
[ wake ] , @xcite .
cells ) tuned at @xmath11 nm .
grey lines refer to single shot realizations , the black line refers to the average over a hundred realizations.,title="fig:",scaledwidth=50.0% ] cells ) tuned at @xmath11 nm .
grey lines refer to single shot realizations , the black line refers to the average over a hundred realizations.,title="fig:",scaledwidth=50.0% ] the output power and spectrum after the first part of the output undulator ( @xmath25 cells ) tuned at @xmath11 nm is shown in fig .
[ out1 ] .
nm , scaledwidth=100.0% ] the tapering law used in the last @xmath26 cells is shown in fig . [
tap ] .
cells ) tuned at @xmath12 nm .
grey lines refer to single shot realizations , the black line refers to the average over a hundred realizations.,title="fig:",scaledwidth=50.0% ] cells ) tuned at @xmath12 nm .
grey lines refer to single shot realizations , the black line refers to the average over a hundred realizations.,title="fig:",scaledwidth=50.0% ] nm .
grey lines refer to single shot realizations , the black line refers to the average over a hundred realizations.,title="fig:",scaledwidth=50.0% ] nm .
grey lines refer to single shot realizations , the black line refers to the average over a hundred realizations.,title="fig:",scaledwidth=50.0% ] nm.,title="fig:",scaledwidth=50.0% ] nm.,title="fig:",scaledwidth=50.0% ] the output power and spectrum of the entire setup , that is after the second part of the output undulator ( @xmath26 cells ) tuned at @xmath12 nm , is shown in fig .
the evolution of the output energy and of the variance of the energy fluctuations in the photon pulse are reported in fig .
[ envarout ] .
finally , the expected transverse size and divergence of the electron beam can be seen in fig .
[ enout ] .
in this work we addressed the potential for enhancing the capabilities of the european xfel . in the hard x - ray regime , a high longitudinal coherence , in addition to full transverse coherence , will be the key to a performance upgrade .
high longitudinal coherence is achievable based on a single - crystal self - seeding scheme , which has been studied for the european xfel parameters in @xcite .
this scheme is compact and can be straightforwardly installed in the baseline undulator with minimal modifications and virtually no operational risk . with the radiation beam
monochromatized down to the transform limit , the output power of the european xfel could be increased by tapering the tunable - gap baseline undulator . in particular , in @xcite we proposed a scheme suitable for the european xfel , to generate tw - level , fully coherent pulses at photon energies of @xmath27 kev .
in this paper we proposed a study of the performance of a self - seeding scheme with single - crystal monochromators for the european xfel at x - rays energies higher than @xmath27 kev . by combining the two techniques of self - seeding and undulator tapering
, we found that @xmath1 gw x - ray transform -limited pulses at the photon energy of @xmath0 kev can be generated without modification to the tw mode of operation at photon energy of @xmath27 kev .
this paper also describes an efficient way for obtaining a many - user facility , based on the high photon energy and high monochromaticity of the output radiation expected from our concept .
we propose a photon beam distribution system based on the use of crystals in the bragg reflection geometry as deflectors . about @xmath21 reflectivity
can be achieved for monochromatic x - rays .
angular and bandwidth acceptances of crystal deflectors match bandwidth and divergence of the x - ray beam .
therefore , it should be possible to deflect the full radiation pulse of an angle of order of a radian without perturbations .
the proposed photon beam distribution system would allow to switch the hard x - ray beam quickly between many instruments in order to make a more efficient use of the european xfel source .
we are grateful to massimo altarelli , reinhard brinkmann , serguei molodtsov and edgar weckert for their support and their interest during the compilation of this work . | we propose to use the self - seeding scheme with single crystal monochromator at the european x - ray fel to produce monochromatic , high - power radiation at @xmath0 kev . based on start to end simulations
we show that the fel power of the transform - limited pulses can reach about @xmath1 gw by exploiting tapering in the tunable - gap baseline undulator .
the combination of high photon energy , high peak power , and very narrow bandwidth opens a vast new range of applications , and includes the possibility to considerably increase the user capacity and fully exploit the high repetition rate of the european xfel .
in fact , dealing with monochromatic hard x - ray radiation one may use crystals as deflectors with minimum beam loss . to this end
, a photon beam distribution system based on the use of crystals in the bragg reflection geometry is proposed for future study and possible extension of the baseline facility .
they can be repeated a number of times to form an almost complete ( one meter scale ) ring with an angle of @xmath2 degrees between two neighboring lines .
the reflectivity of crystal deflectors can be switched fast enough by flipping the crystals with piezo - electric devices similar to those for x - ray phase retarders at synchrotron radiation facilities .
it is then possible to distribute monochromatic hard x - rays among @xmath3 independent instruments , thereby enabling @xmath3 users to work in parallel .
the unmatched repetition rate of the european xfel would be therefore fully exploited .
* deutsches elektronen - synchrotron * * + * desy 11 - 224 november 2011 @xmath4 @xmath5 * extension of self - seeding to hard x - rays * @xmath6 * kev as a way to increase user access at the european xfel * @xmath7 gianluca geloni , _ +
european xfel gmbh , hamburg _ vitali kocharyan and evgeni saldin _ + deutsches elektronen - synchrotron desy , hamburg _
@xmath5 issn 0418 - 9833 @xmath5 * notkestrasse 85 - 22607 hamburg * |
in vivo neurons in cortical and other neural circuits experience a large background of synaptic inputs , acting as a source of noise .
noisy inputs have an important impact on the dynamics of neurons , making neural responses highly variable and affecting many of their response characteristics @xcite .
a fundamental problem is thus to determine the output statistics of the neuronal activity given an input noise statistics .
in addition , the knowledge of the neuronal firing properties can be used to explore large - scale networks using a mean - field approach . in this framework
the stationary states of populations of interacting neurons are self - consistently obtained from their firing responses @xcite .
this allows the efficient exploration of the parameters space and the characterization of the various regime of functioning of these networks . for these reasons it appears crucial to have an accurate estimation of the input - ouput relationship of neurons , especially , in presence of realistic synaptic currents . in this direction
we investigate the firing rates of integrate - and - fire ( if ) neurons .
the firing frequency of neurons with instantaneous synaptic inputs was first obtained in ref .
the effect of synaptic dynamics has been studied under various assumptions @xcite . here
we derive exact formulae that include additional important features .
first we take into account the presence of a finite refractory time , which leads to current correlations between spikes affecting the firing rate of the neuron .
indeed , neurons at their reset potential will evolve with a synaptic current that is still correlated with the _ positive _ going current that made them cross the threshold .
furthermore , we also consider the case of multiple inputs from different synaptic receptors , as it occurs in the vast majority of cortical circuits .
the obtention of the firing rate can be recast into the form of a mean first passage time ( mfpt ) calculation .
since the synaptic dynamics here plays a central role we have to consider a multi - dimensional fokker - planck equation .
the nonequilibrium distribution of the synaptic currents at the reset potential implies that the proper phase space boundary conditions must be found self - consistently from the neuronal and synaptic dynamics . to cope with these issues
we extend the recently developed tools by doering and coworkers @xcite . in this approach
the relevant parameter is the ratio between the synaptic current and membrane potential time constants .
similar considerations also appear in various areas of science when studying the escape rate from a metastable state .
the present results are of special interest in the context of stochastic resonance or stochastic activation @xcite , where considerable attention has been paid to the coexistence of several colored noises with nonequilibrium distributions .
the sub - threshold neuronal dynamics of integrate - and - fire neurons is described by the depolarization @xmath0 of the soma , which evolves according to an evolution equation of the form @xcite @xmath1 with @xmath2 the membrane time constant .
the function @xmath3 governs the dynamics of the membrane voltage when no synaptic currents are present .
for @xmath4 we have a perfect integrate - and - fire neuron ; for @xmath5 we have a leaky integrate and fire neuron .
@xmath6 denotes the synaptic current and @xmath7 its associated input conductance .
the synaptic current evolves according to @xmath8 where @xmath9 denotes a gaussian white noise with zero mean and unit variance : @xmath10 .
the synaptic current is thus exponentially correlated in time with a correlation time @xmath11 : @xmath12 in terms of @xmath13 , where @xmath14 is the average current , and the noise intensity @xmath15 .
this form of the synaptic dynamics arises when the neuron receives a large number of inputs during its characteristic time , as it is typically the case in vivo in the cortex @xcite . in the limit of instantaneous synaptic current , @xmath16 , or for times
much larger than the correlation time @xmath11 , the dynamics simplifies to @xmath17 .
note that the adjunction of the synaptic dynamics renders the voltage dynamics by itself non - markovian .
the neuron fires an action potential when the voltage reaches the threshold potential @xmath18 .
after an absolute refractory period of duration @xmath19 it is reset at the value @xmath20 . during the refractory period no further firing can occur .
by contrast , even when the neuron is in the refractory state , the synaptic current continues to evolve under eq .
, leading to current correlations of order @xmath21 between spikes .
this source of correlations must be taken into account in order to obtain a self - consistent input - output relationship . in the following
we will work in the reduced variables @xmath22 where we introduced the parameters @xmath23 using the adimensional time @xmath24 we obtain the system of equations [ evol.reduced ] @xmath25 the potential @xmath26 is such that @xmath27 , where the prime denotes a derivative with respect to @xmath28 .
the threshold and reset potentials become @xmath29 in these new variables .
henceforth we will also refer to the variables @xmath28 and @xmath30 as the voltage and the synaptic current , respectively .
the process obeys the fokker - planck equation @xcite @xmath31 p \end{aligned}\ ] ] for the joint probability distribution @xmath32 . we will treat @xmath33 as a small parameter , considering the situation where the synaptic time scale is smaller than the membrane time constant
this is for example the case for ampa receptors , which constitute the main fast excitatory inputs in the brain and have a time constant @xmath34 ms smaller than the time constant of pyramidal cells , @xmath35 ms @xcite .
note that the variable @xmath30 keeps a finite variance irrespective of the value of the parameter @xmath33 .
following doering et al @xcite we have a singular perturbation problem for the quantity @xmath36 the latter gives the mean time a neuron spends at points @xmath37 before crossing the threshold .
alternatively , this stationary problem corresponds to the situation where we perform a time average over a long trajectory where the neuron restarts its time evolution at the reset potential after firing and after its absolute refractory period .
after some transients the time spent at each point in phase space will reach the stationary value @xmath38 @xcite .
the mean first passage time before firing then reads @xmath39 while the firing rate of the neuron is given by @xmath40 the factor @xmath19 accounts for the lowering of the firing rate due to the time spent in the refractory state .
importantly , we must consider the additional dependance on the refractory time that appears through the mfpt itself . to obtain the mean first passage time we first generalize the calculation of doering et al @xcite to the case where the synaptic current at reset potential follows an arbitrary distribution @xmath41 .
this distribution will have to be determined self - consistently from the neuronal dynamics .
the function @xmath38 obeys the equation @xmath42 q(v , z ) = - \delta(v-\eta ) \mu(z ) \
, , \ , \label{g.eq}\end{aligned}\ ] ] where we introduced the operator @xmath43 .
furthermore , it must satisfy the half - line absorbing boundary condition @xcite @xmath44 this condition stems from the observation that the probability current @xmath45 must be positive at the threshold potential for firing to occur .
intuitively this corresponds to the fact that the potential can not cross the threshold from above . for further reference
we here introduce the eigenfunctions of the operator @xmath46 : @xmath47 where @xmath48 are the hermite polynomials @xcite . the latter form a complete orthogonal basis in the inner product @xmath49 .
hence we may introduce the quantities @xmath50 characterizing the distribution @xmath14 . note that @xmath51 and that @xmath52 is the mean of the distribution @xmath14 .
we now insert the ansatz @xmath53 into eq . and collect terms according to powers of @xmath33 .
the zeroth order corresponds to the white noise limit of eqs .
, i.e. , to the limit where synaptic events are instantaneous .
we have @xmath54 whose solution is @xmath55 with a yet undetermined function @xmath56 . at the next order of perturbation we find @xmath57 yielding @xmath58 with the yet unknown function @xmath59 .
the second - order terms give @xmath60 \ , , \ , \ , \end{aligned}\ ] ] where we used that @xmath61 .
the operator @xmath46 is invertible only in the subspace of functions spanned by @xmath62 , @xmath63 .
hence the term proportional to @xmath64 must vanish , leading to the following equation for @xmath56 : @xmath65 with @xmath66 .
similarly , the integrability condition for @xmath67 leads to the following equation for @xmath59 : @xmath68 the solutions of these equations that are integrable read @xmath69 and @xmath70/\sigma^2 } \qquad \qquad \nonumber \\ & + & c_1 \sigma^{-1 } e^{[u(\eta)-u(v)]/\sigma^2 } \theta(\eta - v ) \
, , \label{r1}\end{aligned}\ ] ] where @xmath71 is the heaviside function .
remarkably the first - order correction to the mfpt , and hence to the firing rate , only depends on @xmath72 , that is on the mean of the current distribution @xmath73 .
note that the mfpt can be expressed as @xmath74 dv$ ] as the functions @xmath75 are the reduced densities , @xmath76 .
we now have to determine the factor @xmath72 and the constants @xmath77 and @xmath78 self - consistently . in this regard ,
we first consider the phase space boundary condition .
the latter was solved in ref .
@xcite and can be summarized by the condition @xmath79 where @xmath80 with @xmath81 the riemann zeta function .
the crucial point to observe is that this boundary condition remains valid in presence of an arbitrary current distribution @xmath14 since , as shown by eq .
, the presence of a non - vanishing mean current , @xmath82 , only affects the residence time in the region @xmath83 , leaving the region @xmath84 unchanged .
hence the constants @xmath77 and @xmath78 take the values @xmath85 and @xmath86 .
the factor @xmath87 is referred to as the milne extrapolation length as it characterizes the non - vanishing value of the probability distribution at the threshold potential .
we now have to evaluate the value of the mean current @xmath72 at the reset potential .
it must be determined self - consistently as the synaptic current remains correlated between spikes due to its finite time constant .
accordingly , the probability distributions at the threshold and reset potentials are related as follows : @xmath88 where @xmath89 is the green function of the synaptic current , which is given by a normal distribution for @xmath30 of mean @xmath90 and variance @xmath91 .
consequently we have that @xmath92 where we used eq . with @xmath93 .
the mean reset current is thus related to the mean threshold current .
again , we can verify that , at first - order in @xmath33 , the expression for the probability flux given in ref .
@xcite remains unchanged in the present situation , as already suggested by eq . .
hence we find @xmath94 at first - order in @xmath33 .
when the refractory time is large compared to the synaptic time constant the current will relax to its stationary distribution @xmath95 so that @xmath96 for @xmath97 .
when the refractory time vanishes correlations are maximal since the current distribution @xmath14 at the reset potential exactly matches the current distribution at the threshold , resulting in a contribution @xmath98 .
the same method can be applied to obtain the full distribution of the synaptic current . combining these results we obtain the firing rate of if neurons .
an important case is the leaky integrate - and fire neuron for which the mean first passage time takes the compact form @xmath99 dw \label{mfptif}\end{aligned}\ ] ] where @xmath100 , @xmath101 , and erf is the error function @xcite .
this formula interpolates between the two - limiting cases of zero and infinite refraction time previously considered in refs @xcite and @xcite , respectively .
similar expressions can be obtained for nonlinear if neurons as well . the firing rate as a function of the refractory period
is depicted in fig .
[ fig1 ] along with the result of monte - carlo simulations .
the firing rate is compared to the situation where the neuron is unresponsive during the refractory period but the corresponding current correlations are neglected [ i.e. , formula but with @xmath102 .
the corrections due to the effect of the refractory state on the mfpt are significant , even at moderate firing rates .
to investigate the validity of the expansion ( [ ansatz ] ) , we numerically obtained the firing rate as a function of the ratio @xmath103 , which is plotted in fig .
the agreement with the first - order formula holds for the region @xmath104 . for ratios @xmath103 up to @xmath105
, we found that the firing rate could be well fitted by considering the effective threshold @xmath106 as seen in fig .
the effective reset potential @xmath107 appears unaffected by the second order corrections , as observed in ref .
in particular , the dependance on the refractory period remains identical .
we further consider the case where several synaptic inputs are present , i.e. , @xmath108 where the currents @xmath109 obey evolution equations of the form albeit with different time constants @xmath110 , means @xmath111 , and variances @xmath112 .
we here assume that the synaptic time constants satisfy @xmath113 with @xmath114 of order unity , @xmath115 , so that @xmath110 are of order @xmath33 : @xmath116 .
the calculations of the previous section can be extended to this case by considering the three - dimensional phase space fokker - planck equation . inserting the eigenfunctions @xmath117 into the expansion in powers of @xmath33
, we obtain that the boundary condition now reads @xmath118 , where @xmath119 .
we could not obtain the milne extrapolation length analytically , which was determined numerically to be @xmath120 .
it replaces the factor @xmath121 in the present situation .
the white noise limit is obtained from eq . with @xmath122 and @xmath123 , and where the potential @xmath124 incorporates both input currents , @xmath125 .
the first - order correction reads @xmath126/\sigma^2 } \ , , \end{aligned}\ ] ] with the corresponding correction to the mfpt given by @xmath127 ( for simplicity we did not include the refractory state in this analysis ) .
remarkably , this first - order correction only depends on the total noise strength @xmath128 and not on the ratio @xmath129 of the synaptic time constants . accordingly , as regards the firing rate , we may treat different synaptic currents as an effective single input with parameters @xmath125 and @xmath123 , even when they present different time courses .
this effect is valid at first - order in the parameter @xmath33 ; the case of very long synaptic time constants compared to the membrane integration time was studied in ref .
the generalization to @xmath130 synaptic inputs is straightforward and does not change this result .
we have obtained analytical expressions for the stationary firing rate of integrate - and - fire neurons endowed with realistic synaptic dynamics .
precisely we have included two salient features into the description : the presence of a finite refraction time on the one hand , and the presence of multiple synaptic inputs on the other hand .
the finiteness of the refractory period of the neuron provides a new source of correlations interacting with the synaptic dynamics .
since the synaptic dynamics has a finite relaxation time , neurons at the reset potential may still be correlated with the positive current that made them cross the firing threshold , resulting in an increased firing rate .
this mechanism is important for fast excitatory and inhibitory synapses such as ampa and @xmath131 synapses as their time constants is in the range of @xmath132 to @xmath133 ms @xcite , comparable to the absolute refractory period of cortical neurons , @xmath134 ms .
the presence of multiple synaptic inputs has been considered as well . in this case the milne extrapolation length must be determined numerically to complete the theoretical analysis .
importantly the firing rate only depends on the variances of the synaptic inputs and not on their relative time constants .
this allows us to consider the combined effect of excitatory and inhibitory synaptic inputs from , say , ampa and @xmath131 receptors .
these analytical results are crucial to assess the effect of noise on neuronal dynamics .
for instance they provide accurate expressions that can be used in the mean - field exploration of large - scale neuronal networks @xcite . in particular , the network collective properties , such as the stability of the low activity spontaneous state or of the persistent
`` memory '' states , will depend on the transfer function .
more generally , the study of nonequilibrium , multiple colored noises appears in a wide range of situations of concern , from chemical networks to biology .
d. andrieux thanks professor x - j wang for support and encouragement in this research and acknowledges financial support from the f.r.s .- fnrs belgium .
t. monnai acknowleges financial support from a waseda university grant for special projects . | we derive analytical formulae for the firing rate of integrate - and - fire neurons endowed with realistic synaptic dynamics .
in particular we include the possibility of multiple synaptic inputs as well as the effect of an absolute refractory period into the description . |
the conduction of heat or the diffusion of matter through a solid body is of importance in the physical and engineering sciences .
the classic reference of carslaw and jaeger , @xcite , analyses many examples and applications .
the mathematical tools used in @xcite are centred around separation of variables and laplace transforms and , in many cases , require properties of special functions . from a mathematical point of view ,
the heat equation , heat content and heat trace link the underlying geometry of the manifold and its boundary and boundary conditions to the spectral resolution of the laplace operator . over the last few decades
, a considerable amount of progress has been made in understanding the asymptotic behaviour of the heat content for small time @xmath5 , see @xcite .
it was discovered by preunkert , @xcite , that even in the absence of boundary conditions the heat content of a ball @xmath6 in euclidean space @xmath7 which is at initial temperature @xmath2 , while @xmath8 has initial temperature @xmath3 , has non - trivial asymptotic behaviour as @xmath9 . for small @xmath5 , the initial condition on the complement of @xmath6 acts in a similar way to a dirichlet @xmath3 boundary condition .
this was subsequently stated for bounded , open sets with @xmath10 boundary in @xcite , and proved in @xcite .
the discussion is simplified by the fact that the heat kernel on @xmath7 is known explicitly .
the general situation for the heat content of a compact subdomain @xmath11 in a compact riemannian manifold @xmath12 was examined in @xcite .
the tools of pseudo - differential calculus used there rely heavily upon the smoothness assumptions on the boundary .
two - sided estimates for the heat content of non - compact sets in @xmath7 were obtained in @xcite .
these estimates are very different from the ones where dirichlet @xmath3 boundary conditions are imposed .
see @xcite and @xcite . in this paper
we denote the fundamental solution of the heat equation on @xmath13 by @xmath14 and for an open set @xmath15 , we define @xmath16 then @xmath17 satisfies the heat equation on @xmath13 @xmath18 ( see chapter 2 in @xcite ) and @xmath19 where @xmath20 is the boundary of @xmath0 .
we define the _ heat content of @xmath0 in @xmath13 at @xmath5 _ by @xmath21 so by , @xmath22 we denote the lebesgue measure of a measurable set @xmath23 by @xmath24 , its perimeter by @xmath25 , and its @xmath26-dimensional hausdorff measure by @xmath27 . if @xmath0 is a bounded , open set in @xmath13 , @xmath28 , with @xmath10 boundary @xmath20 , then theorem 2.4 of @xcite implies that @xmath29 in @xcite , we obtained explicit bounds for @xmath30 for bounded , open sets @xmath0 in euclidean space with @xmath10 boundary which are uniform in @xmath5 and in the geometric data of @xmath0 .
these bounds imply that @xmath31 we observe that if @xmath32 is a closed set in @xmath7 with @xmath33 then by the definition of the perimeter ( see @xcite ) and , @xmath34 the observations in suggest that holds for all open sets @xmath0 with finite perimeter and finite lebesgue measure .
the proof of such a statement is well beyond the scope of this paper . in this paper , we focus on the heat content of a bounded , connected , open set @xmath35 with polygonal boundary .
we introduce some notation and then present the main result : theorem [ t1 ] .
let @xmath36 denote the interior angles of @xmath20 .
each such angle @xmath37 is supported by two edges provided @xmath38 by , we may exclude angles @xmath39 we label the corresponding vertices by @xmath40 and note that these @xmath41 vertices need not be pairwise disjoint .
let @xmath42 denote the infinite wedge of angle @xmath37 with vertex @xmath43 such that @xmath44 and the boundary of the wedge contains the two edges which are adjacent to @xmath43 and have an angle @xmath37 .
let @xmath45 for @xmath46 , we also define the open sector @xmath47 and @xmath48 [ t1 ] let @xmath35 be a bounded , connected , open set with polygonal boundary @xmath20 with @xmath49 and @xmath50 as defined in to . then as @xmath4 , @xmath51 where @xmath52 is given by @xmath53 and @xmath54 is given by @xmath55 for @xmath56 , where @xmath57 denotes the smallest angle between @xmath58 and @xmath59 . in any of the remaining cases , such as @xmath60 , we define @xmath61 by taking appropriate limits using lhpital s rule .
the terms which involve the area and perimeter are as expected and agree with those in .
we see that the heat content has a non - trivial dependence on the interior angles of the polygonal boundary .
we observe that @xmath62 is continuous on @xmath63 , decreasing on @xmath64 $ ] and symmetric with respect to @xmath65 .
that is @xmath66 by and , we conclude that @xmath67 is non - negative .
we remark that @xmath61 is symmetric with respect to @xmath68 and @xmath69 and that @xmath70 . by lemma [ l3.2 ] ,
section [ s3 ] , it follows that @xmath71 is non - negative .
in addition , if @xmath0 is a regular @xmath41-gon in @xmath1 , then @xmath72 , and the angular contribution to the heat content is @xmath73 we observe that the coefficient of @xmath74 in the expansion of the heat content of a compact domain @xmath11 with smooth boundary @xmath75 is also equal to @xmath3 , see theorem 1.6 in @xcite . here
@xmath76 is the trace of the second fundamental form when @xmath77 is oriented with a smooth inward - pointing unit normal vector field . the expansion for the heat content of a polygon with dirichlet @xmath3 boundary conditions was obtained in @xcite .
there it was shown that if @xmath78 solves the heat equation @xmath79 with @xmath80 and satisfies a dirichlet boundary condition @xmath81 for any @xmath82 then @xmath83 where @xmath84 we note that both and have angular contributions which are additive . however , in theorem [ t1 ] there is an additional term in the case where vertices have multiplicity larger than @xmath2 .
no such term is present in since sectors based at the same vertex do not feel each other s presence due to the dirichlet @xmath3 boundary condition .
the strategy to prove is inspired by @xcite , and relies on some model computations .
we use an analogous strategy to prove . for points
@xmath85 close to a vertex , say @xmath86 for some @xmath87 , @xmath88 , @xmath89 is approximated by @xmath90 . for points
@xmath85 which have a distance at least @xmath91 to @xmath20 , for some @xmath92 , @xmath89 is approximated by @xmath2 .
for the remaining points in @xmath0 , @xmath89 is approximated by @xmath93 , where @xmath94 is the half - plane which contains @xmath0 and whose boundary contains the edge of @xmath20 nearest to @xmath95 . as was the case in @xcite ,
the model computations involving the infinite wedge @xmath42 are the most difficult to carry out . however , in contrast to the case with dirichlet 0 boundary conditions , we must also consider the contribution to the heat content from a vertex which belongs to the boundaries of more than one wedge .
we deal with these computations in lemma [ l3.1 ] and lemma [ l3.2 ] in section [ s3 ] . in section [ s4 ]
we carry out the half - plane computations .
it is quite remarkable that , in contrast to the smooth case , the asymptotic expansion in half powers of @xmath5 of @xmath96 in theorem [ t1 ] terminates after the term of order @xmath5 , leaving an exponentially small remainder as @xmath4 .
this agrees with the fact that there are no further locally computable invariants of @xmath0 and @xmath20 available from which non - trivial quantities could be built .
a similar phenomenon has been observed for the asymptotic expansion of the heat trace .
see for example @xcite .
the precise form of the exponential remainder remains an open problem .
the extension of the results in this paper to general polyhedra in @xmath97 is another challenge beyond the scope of this paper .
in section [ s7 ] , we use theorem [ t1 ] to compute the heat content of a fractal polyhedron which is constructed as follows .
let @xmath98 be an open cube of side - length 1 .
let @xmath99 .
attach a regular open cube @xmath100 of side - length @xmath101 to the centre @xmath102 of each face of @xmath103 , and such that all the faces are pairwise - parallel .
now proceed by induction . for
@xmath104 attach @xmath105 open cubes @xmath106 of side - length @xmath107 to the centres of the boundary faces of the cubes @xmath108 , again with pairwise - parallel faces .
we define the fractal polyhedron @xmath109 as @xmath110}\right\}\ ] ] for @xmath111 ( see figure [ fig1 ] ) .
we note that for this range of @xmath101 , no cubes in the construction of @xmath109 overlap .
for the two - dimensional construction , see theorem 4 in @xcite . in that paper , the critical value @xmath112 , where the squares just touch , is allowed
this is due to the fact that the dirichlet @xmath3 boundary conditions guarantee the independence of the heat flow in these touching squares . in this paper
, we do not impose dirichlet @xmath3 boundary conditions on @xmath113 and if the cubes touch , then this could give rise to an extra term .
for this reason , we only allow @xmath111 . with @xmath114 . ]
we have that @xmath115 and that the two - dimensional hausdorff measure of the boundary is given by @xmath116 both quantities are finite for @xmath111 . moreover , the total length of the edges of @xmath117 is finite if and only if @xmath118 and equals @xmath119 .
in addition to theorem [ t1 ] , in order to compute the heat content of @xmath109 , we require the heat content of a sector of angle @xmath65 in an infinite wedge of angle @xmath120 where this sector and wedge share one common edge and vertex .
this will be computed in section [ s6 ] . in section [ s7 ] ,
we prove the following theorems which give the asymptotic expansion for the heat content of @xmath109 as @xmath4 .
[ t2 ] let @xmath121 .
fix @xmath111 , @xmath122 .
there exists a periodic , continuous function @xmath123 with period @xmath124 such that @xmath125 it is easy to see that if we write @xmath126 , in theorem [ t2 ] then @xmath127 is the interior minkowski dimension of the vertices of @xmath117 for @xmath118 , whereas for @xmath128 , it is the interior minkowski dimension of the edges of @xmath117 .
below we state the corresponding result for the critical case @xmath129 .
[ t3 ] for @xmath130 , there exists a periodic , continuous function @xmath131 with period @xmath132 such that , @xmath133 in section [ s2 ] below we introduce some further notation and state and prove several lemmas .
let @xmath0 be as given in theorem [ t1 ] . in the proofs of lemma [ l1 ] to lemma [ l4 ]
we use a variant of kac s principle of not feeling the boundary to reduce the computation of the heat content of @xmath0 to a collection of model computations . for @xmath88 , @xmath134 , @xmath135 , @xmath136
we define @xmath137 and @xmath138 and @xmath139 choose @xmath50 as in and let @xmath140 .
[ l1 ] @xmath141 by ( * ? ? ?
* proposition 9(i ) ) , we have that @xmath142 where @xmath143 . since @xmath144 , @xmath145 .
hence @xmath146 as required .
we partition the region @xmath147 into @xmath41 sectors , @xmath148 , of radius @xmath50 , @xmath41 rectangles , @xmath149 , of height @xmath150 and @xmath151 cusps of height @xmath150 .
the contributions to @xmath30 from these regions will be computed in sections [ s3 ] , [ ss4 ] and [ ss5 ] respectively .
each sector has two neighbouring cusps .
each cusp is adjacent to a rectangle and a sector .
the corresponding @xmath152-dimensional result to ( * ? ? ?
* lemma 7 ) is the following .
[ l ] let @xmath153 be non - empty , open subsets of @xmath154 such that @xmath155 and @xmath156 .
let @xmath157 be a bounded , measurable subset of @xmath158 .
then @xmath159 where @xmath160 we write @xmath161 by , we have that @xmath162 and @xmath163 this completes the proof .
[ l2 ] let @xmath164 and @xmath165 such that @xmath166 .
suppose @xmath167 are interior angles of @xmath20 which are supported by edges which meet at the same vertex @xmath168
. then @xmath169 by lemma [ l ] with @xmath170 , @xmath171 , @xmath172 and @xmath173 we have that @xmath174 we also have that @xmath175 combining and gives .
[ l3 ] let @xmath176 denote the half - plane such that @xmath177 and @xmath178
. then @xmath179 using lemma [ l ] with @xmath170 , @xmath180 , @xmath181 and @xmath182 the result follows .
[ l4 ] let @xmath183 denote the cusp which is adjacent to @xmath149 and @xmath148 .
let @xmath94 be the half - plane as in lemma [ l3 ]
. then @xmath184 using lemma [ l ] with @xmath170 , @xmath180 , @xmath185 and @xmath186 the result follows .
in this section we approximate @xmath187 by @xmath90 .
we then compute the contribution to the heat content of @xmath0 from a sector , @xmath188 with corresponding angle @xmath189 .
we also compute the contribution to the heat content of @xmath0 from two disjoint wedges whose boundaries intersect in a vertex of @xmath20 .
firstly , we define @xmath190 [ l3.1 ] for @xmath191 , @xmath192 changing to polar coordinates with @xmath193 and @xmath194 in gives that @xmath195 where @xmath196 .
the change of variable @xmath197 gives that @xmath198 we have that @xmath199 and @xmath200 via the change of variables @xmath201 and integrating by parts with respect to @xmath68 , we obtain @xmath202 similarly , @xmath203 we first compute @xmath204 and then we deal with @xmath205 . by and , we have that @xmath206 where @xmath207 \cap ( [ 0,\frac{\pi}{2 } ] \cup [ \frac{3\pi}{2},2\pi])$ ] .
first suppose @xmath208 $ ] , then by we have that @xmath209 @xmath210 similarly , for @xmath211 $ ] , we have that @xmath212 in order to have in the same format as , we replaced 1 by @xmath213 in . for @xmath214 $ ]
, we have that @xmath215 now by , @xmath216 via the change of variables @xmath217 , we also have that @xmath218 hence by and , we see that for @xmath214 $ ] and @xmath4 , @xmath219 it remains to compute @xmath205 . via the change of variables @xmath220
, we see that @xmath221 by 2.173 ( 1 ) and 2.172 in @xcite , we see that @xmath222 therefore by , and the change of variables @xmath223 , we obtain that @xmath224 thus @xmath225 we note that @xmath226 where @xmath227 hence it is necessary to compute @xmath228 if @xmath229 , then @xmath230 and if @xmath231 , then @xmath232 hence @xmath233 therefore , by combining , and , , respectively , we obtain that , as @xmath4 , @xmath234 via the change of variables @xmath235 and integrating by parts with respect to @xmath236 , we have @xmath237 again integrating by parts with respect to @xmath236 , we see that @xmath238 hence by , and the change of variables @xmath239 , we have that @xmath240 this completes the proof of lemma [ l3.1 ] . in section [ ss6 ]
, we deal with the remaining integral in .
[ l3.2 ] let @xmath241 be two disjoint wedges in @xmath1 with corresponding angles @xmath242 respectively such that @xmath243 for some @xmath244 .
let @xmath57 denote the angle between @xmath245 and @xmath246 such that @xmath247 .
then @xmath248 where @xmath249 is as defined in . by changing coordinates to polar coordinates , as in the proof of lemma [ l3.1 ]
, we have that @xmath250 we see that @xmath251 hence @xmath252 therefore becomes @xmath253 by considering the function @xmath254 for @xmath255 in each of the four quadrants , we have that @xmath256 is a primitive for @xmath257 for all @xmath258 .
integrating by parts with respect to @xmath255 , we obtain @xmath259 hence , by , and integrating by parts with respect to @xmath69 , we have that @xmath260 as required .
in this section , we consider points @xmath261 .
we partition this region into @xmath41 rectangles @xmath149 and @xmath151 cusps @xmath183 , each of height @xmath262 .
we approximate @xmath187 by @xmath93 , where @xmath176 is the half - plane such that @xmath177 and @xmath178 , as in lemma [ l3 ] .
we first compute the contribution to the heat content of @xmath0 from a rectangle , @xmath149 , of height @xmath262 and length @xmath263 , where @xmath264 .
we have that @xmath265 let @xmath266 and @xmath267 .
then , by integrating by parts with respect to @xmath268 , we obtain @xmath269 we now compute the contribution to the heat content of @xmath0 from a cusp .
let @xmath183 denote the cusp which is adjacent to @xmath149 and @xmath148 .
then @xmath270 integrating by parts with respect to @xmath95 , we obtain @xmath271 hence the contribution from each cusp is @xmath272 finally , we deal with the remaining integrals from the sector and cusp contributions .
each sector has two neighbouring cusps so we are interested in the following integral @xmath274 by definition of @xmath49 , in , we can write the remaining integral from the sector contribution , , as @xmath275 adding and , we obtain @xmath276 this completes the proof of theorem [ t1 ] .
in this section , we compute the heat content of a @xmath65-sector in a @xmath120-wedge which share one common edge ( and vertex ) .
this is a crucial ingredient in the computation of the heat content of the fractal polyhedron @xmath109 , which was constructed in section [ s1 ] , and will be used in section [ s7 ] .
[ l6.1 ] let @xmath277 such that @xmath278 and @xmath279 share one common edge ( and vertex ) .
then , for @xmath4 , @xmath280 we remark that the coefficient of @xmath5 is equal to @xmath3 in this case . similarly to the proof of lemma [ l3.1 ] ,
the left - hand side of equals @xmath281 now @xmath282 and letting @xmath197 , we have that @xmath283 we also have that @xmath284 @xmath285 now @xmath286 equals @xmath287 by integrating by parts with respect to @xmath68 .
in addition , as for the computation of @xmath205 ( see ) , we have that @xmath288 note that @xmath289 hence @xmath290 thus , by , , and , we have that , as @xmath4 , @xmath291 therefore , by , and , as @xmath4 , @xmath292 we note that @xmath293 by theorem [ t1 ] and lemma [ l6.1 ] , this implies that , as @xmath4 , @xmath294 the result of lemma [ l6.1 ] and formula will be used in section [ s7 ] below .
in this section , we use theorem [ t1 ] to compute the heat content of the fractal polyhedron @xmath109 which was constructed in section [ s1 ] . to do this ,
we adapt the scheme of @xcite to the three - dimensional setting below .
the key step in @xcite was to obtain a renewal equation by making a suitable ansatz for the heat content .
the corresponding ansatz has been made here in and for @xmath295 , and in and for @xmath130 . in order to derive the required renewal equation
, we need to compute the contribution to the heat content @xmath296 from @xmath297 and @xmath298 .
we do this below in lemma [ l7.1 ] and lemma [ l7.2 ] respectively . in what follows , for @xmath299
, @xmath300 is the 3-dimensional analogue of .
we make the following approximations for @xmath301 .
let @xmath302 and let @xmath303 . if @xmath304 , then we have that @xmath305 by the principle of not feeling the boundary , ( * ? ? ?
* proposition 9(i ) ) .
we define @xmath306 if @xmath307 , then we have that @xmath308 where @xmath309 i.e. @xmath94 is a half - space whose boundary contains the face of @xmath117 nearest to @xmath95 .
let @xmath310 if @xmath311 , then we have that @xmath312 where @xmath313 is the infinite wedge @xmath314 for entrant edges and @xmath313 is the infinite wedge @xmath279 for re - entrant edges .
( see the proof of lemma [ l7.1 ] for further details ) .
the estimates , follow by similar arguments to those given in the proof of lemma [ l ] with @xmath315 , @xmath316 , @xmath317 respectively and @xmath318 .
it remains to approximate @xmath301 for @xmath95 near a vertex of @xmath319 , @xmath320 respectively .
we only require the contribution to the heat content @xmath296 from these vertices to derive the required renewal equation .
the relevant approximation to make here is via a one - sided infinite cone @xmath321 with vertex @xmath322 such that @xmath323 .
definition aside , no viable expressions are known for @xmath324 in this 3-dimensional setting . for our purposes , it is sufficient to approximate the neighbourhood of each vertex @xmath325 by a cube @xmath326 .
each cube @xmath326 centred at @xmath327 has side - length @xmath328 and is chosen such that the faces of @xmath329 are pairwise - parallel to those of @xmath103 .
we are interested in the contribution to the heat content @xmath296 from the region @xmath330 .
there are two cases to consider .
either the vertex @xmath327 is entrant and @xmath330 is @xmath331 of @xmath326 , or the vertex @xmath327 is re - entrant and @xmath330 is @xmath332 of @xmath326 .
if @xmath327 is entrant , then the coefficient of @xmath333 in the expansion for @xmath296 is equal to @xmath334 by separation of variables .
unfortunately , we were unable to compute the coefficient of @xmath333 for a re - entrant vertex .
however , the contribution to the heat content @xmath296 from each region @xmath330 is of order : @xmath335 from the volume , @xmath336 from the surface area of the adjoining faces of the vertex , and @xmath337 from the length of the adjoining edges of the vertex .
thus , if we choose @xmath91 as follows : @xmath338 then the contribution to the heat content @xmath296 from each region @xmath330 is @xmath339 .
we note that this choice of @xmath91 gives @xmath340 .
partition @xmath297 into the following sets .
( i ) @xmath319 has 32 vertices ; @xmath342 . at each vertex @xmath343 , consider a cube @xmath344 of side - length @xmath328 centred at @xmath343 .
let @xmath345 and @xmath346 .
( ii ) @xmath319 has 36 edges ; @xmath347 .
let @xmath348 .
( iii ) @xmath349 .
( iv ) the interior of @xmath297 minus ( i ) , ( ii ) and ( iii ) ; an open polygon @xmath350 with distance at least @xmath91 to @xmath117 . ( v )
the remainder , which has measure zero .
the contribution to the heat content from ( iv ) is @xmath351 . to compute the contribution from ( ii ) , there are two types of edges to consider .
each edge @xmath352 is the intersection of two faces of @xmath117 . for fixed @xmath353 ,
let @xmath354 denote the plane which is orthogonal to these faces and intersects @xmath352 in exactly one point .
apply lemma [ l ] with @xmath355 , @xmath356 respectively , @xmath357 and @xmath358 so that either ; ( i ) the contribution from @xmath359 can be approximated by that from a sector @xmath360 in a wedge @xmath314 , or ( ii ) the contribution from @xmath359 can be approximated by that from a sector @xmath361 in a wedge @xmath279 .
each cross - section of @xmath364 is a union of rectangles and cuspidal regions .
thus by section [ s4 ] , we have that the contribution from @xmath364 equals @xmath365 for edges of type ( i ) , the sector @xmath366 has two cuspidal neighbours . for edges of type ( ii ) ,
the sector @xmath366 has one cuspidal neighbour but by lemma [ l6.1 ] , the term of order @xmath333 is half that of lemma [ l ] . even though the sector and cusp terms of order @xmath333 cancel out , the remainder is dominated by the contribution from ( i ) , which is @xmath367 this completes the proof of lemma [ l7.1 ] .
following the strategy of @xcite , in order to compute @xmath368 , we introduce a model solution which approximates @xmath369 in one of the six components of @xmath370 .
consider the half - space @xmath371 and attach one of the six components of @xmath370 to @xmath94 .
the resulting set is @xmath372}\right\}.\ ] ] let @xmath373 be the unique solution of and with @xmath374 .
define @xmath375 applying lemma [ l ] with @xmath315 , @xmath376 , @xmath377 and @xmath378 , where @xmath379 , we have @xmath380 in contrast to @xcite , where the temperature of the boundary is fixed for all @xmath381 , we must account for the fact that @xmath382 and @xmath383 feel each other s presence .
the choice of @xmath384 above is a lower bound for the distance between @xmath382 and @xmath383 .
the proof of lemma [ l7.2 ] is analogous to that of lemma [ l7.1 ] .
we note that in this case there is an additional type of edge to consider .
namely , the edges where the contribution from ( @xmath389 can be approximated by that from a sector @xmath390 in a wedge @xmath279 .
there are 4 such edges so the contribution to the heat content is @xmath391 by .
this completes the proof of lemma [ l7.2 ] by our choice of @xmath91 , .
re - write @xmath394 as @xmath395 then @xmath396 consists of 5 copies of @xmath397 scaled by a factor @xmath101 , say @xmath398 .
each of these copies has a face @xmath399 connecting it to @xmath298 .
let @xmath400 be the half - space such that @xmath401 and @xmath402 .
put @xmath403 .
then @xmath404 is a copy of @xmath405 scaled by a factor @xmath101 .
thus , by scaling , we have @xmath406 define @xmath407 , where @xmath408 .
applying lemma [ l ] with @xmath409 , @xmath410 , @xmath411 and @xmath412 , we have @xmath413 hence @xmath414 combining this with lemma [ l7.2 ] gives the result .
define @xmath418 substitute into to obtain @xmath419 from , we also have @xmath420 which implies @xmath421 or equivalently , @xmath422 similarly to ( * ? ? ?
* proposition 5 ) , define @xmath423 by @xmath424 then @xmath425 hence @xmath426 since there is a constant @xmath388 such that @xmath393 for @xmath387 , there is a constant @xmath427 such that @xmath428 combining with and gives . for @xmath130 , @xmath431 .
define @xmath432 substitute with @xmath130 into to obtain @xmath433 by considering with @xmath5 replaced by @xmath434 , we obtain that @xmath435 or equivalently , @xmath436 define @xmath437 by @xmath438 then , as in the proof of lemma [ l7.4 ] , we have @xmath439 via .
since holds for @xmath130 , we obtain the same remainder estimate as in lemma [ l7.4 ] . combining lemma [ l7.1 ] , and lemma [ l7.4 ] , lemma [ l7.5 ] we obtain , respectively .
using the fact that @xmath440 is continuous , it can be shown that @xmath441 is continuous for @xmath111 , see ( * ? ? ?
* section 5.1 ) . | let @xmath0 be a bounded , connected , open set in euclidean space @xmath1 with polygonal boundary .
suppose @xmath0 has initial temperature @xmath2 and the complement of @xmath0 has initial temperature @xmath3 .
we obtain the asymptotic behaviour of the heat content of @xmath0 as time @xmath4 .
we then apply this result to compute the heat content of a particular fractal polyhedron as @xmath4 . |
a polygon @xmath6 is called rational if all angles of @xmath7 are rational multiples of @xmath8 .
let @xmath9 denote the number of ( cylinders of ) periodic billiard trajectories of euclidean length at most @xmath10 .
it is a theorem of h. masur @xcite that there exist constants @xmath11 and @xmath12 such that for @xmath13 , @xmath14 a natural question is whether ( [ eq : masur ] ) can be converted to an asymptotic formula as @xmath15 . a well known construction associates a `` translation surface '' @xmath16 to each rational polygon @xmath7 .
essentially the algorithm `` unfolds '' the billiard trajectories , by reflecting the polygon instead of reflecting the trajectory .
more precisely , let @xmath17 denote the group generated by reflections in the sides of the polygon @xmath7 . since @xmath7 is rational
, @xmath18 is finite .
the `` translation surface '' consists of @xmath18 copies of @xmath7 , with each copy glued to each of its mirror images along the reflecting side .
for example , if @xmath7 is the unit square , then @xmath16 is the torus @xmath19 , and if @xmath7 is the isosceles triangle with angles @xmath20 , @xmath20 , @xmath21 , and @xmath5 is even , then @xmath16 is the regular @xmath5-gon with opposite sides identified .
a translation surface can be defined in one of the following equivalent ways : 1 .
[ transsurfdefn - polygons ] a union of polygons @xmath22 where each @xmath23 , and the @xmath24 are glued along parallel sides , such that each side is glued to exactly one other , and the total angle in each vertex is an integer multiple of @xmath25 .
2 . an orientable surface with a flat metric and isolated conical singularities that has trivial rotational holonomy .
( note that trivial rotational holonomy means in particular that parallel transport of a vector along a small loop going around a conical point brings a vector back to itself .
this implies that all cone angles are integer multiples of @xmath25 . )
a pair @xmath26 , where @xmath27 is an ( orientable ) riemann surface , and @xmath28 is a holomorphic 1-form on @xmath27 .
( note that away from the zeroes of @xmath28 , there is a local coordinate @xmath29 such that @xmath30 , and this coordinate is unique up to translation .
then one can define the metric on @xmath27 as @xmath31 .
this metric is flat , with conical singularities appearing at the zeroes of @xmath28 .
the term `` translation surface '' comes from the fact that away from the cone points the surface can be covered by charts so that the transition functions are translations ( @xmath32 ) .
if @xmath33 is an @xmath5-tuple of positive integers such that the sum of the @xmath34 is even , we denote by @xmath35 the moduli space of translation surfaces @xmath26 such that the multiplicities of the zeroes of @xmath28 are given by @xmath36 ( or equivalently such that the orders of the conical singularities are @xmath37 ) .
( actually , for technical reasons , the singularities of @xmath26 should be labeled ; thus , an element of @xmath35 is a tuple @xmath38 , where @xmath39 are the singularities of @xmath27 , and the multiplicity of @xmath40 is @xmath34 . ) the moduli space of translation surfaces is naturally stratified by the spaces @xmath35 ; each is called a _
stratum_. by construction , billiard trajectories on @xmath7 correspond to `` straight lines '' on @xmath16 , which are geodesics not passing through singularities .
it is easy to see that any such geodesic is part of a family of freely homotopic parallel geodesics of the same length .
such a family is called a cylinder .
let @xmath41 denote the number of cylinders on @xmath16 of length at most @xmath10 .
( by the length of a cylinder we mean the length of any of the closed geodesics that comprise it ) .
there is an action of @xmath0 on the moduli space of translation surfaces that preserves the stratification . for our purpose
, it is easiest to see this using definition ( [ transsurfdefn - polygons ] ) : since @xmath0 acts on @xmath42 , for @xmath43 , we can define @xmath44 , where all identifications between the sides of the polygons for @xmath45 are the same as for @xmath16 .
this action generalizes the action of @xmath0 on the space of flat tori @xmath46 .
we can visualize this as a composition of `` the usual linear action '' with `` cut and paste . ''
we note that `` cut and paste '' is an isometry on the surface ( and in fact preserves the horizontal and vertical directions as well ) .
note that if @xmath16 is a union of triangles , and @xmath47 is a large element of @xmath0 then @xmath45 is a union of long and thin triangles .
we may if we wish `` cut and paste '' @xmath45 and retriangulate to try to present @xmath45 as a union of triangles with bounded side lengths . for @xmath48 ,
let @xmath49 denote the stabilizer of @xmath16 .
the group @xmath50 is called the _ veech group _ of @xmath16 . if @xmath50 is a lattice in @xmath0 then @xmath16 is called a _
veech surface_. it is a theorem of veech @xcite that if @xmath16 is a veech surface , then there exists @xmath51 such that @xmath52 as @xmath15 .
one has the formula @xcite , ( reproduced in @xcite ) @xmath53 where @xmath54 , @xmath55 , and @xmath56 .
the left hand side counts ( cylinders ) of closed geodesics in an annulus , and the right hand side is an integral over part of the @xmath0 orbit of @xmath16 .
thus , the @xmath0 action can be used to count closed geodesics ( and thus periodic billiard trajectories ) .
a closer examination of ( [ eq : count ] ) shows that the integral is over large circles inside the @xmath0 orbit .
these large circles can be approximated by horocycles , which are orbits of @xmath57 .
thus the ergodic properties of the action of @xmath58 play a key role .
ratner s theorem @xcite is the classification of the invariant measures for the action of a unipotent subgroup on the homogeneous space @xmath59 , where @xmath60 is a lie group and @xmath61 is a lattice in @xmath60 .
an important question is whether a similar theorem holds for the @xmath2-action on a stratum @xmath35 .
one can also ask this question when one restricts the action to any @xmath0 invariant submanifold of a stratum . in this paper
, we will classify the @xmath2-invariant measures on a certain family of @xmath0-invariant manifolds .
another result in this direction was obtained by mcmullen @xcite who , in genus @xmath62 , classified the measures invariant under all of @xmath0 .
we say that a translation surface @xmath16 is a branched cover of a translation surface @xmath27 if the covering map @xmath8 respects the translation structure ( i.e. if we identify @xmath63 and @xmath64 where the @xmath65 are riemann surfaces and the @xmath66 are holomorphic 1-forms on @xmath65 then we require that @xmath67 is holomorphic and @xmath68 . ) now let @xmath69 be a veech surface .
then the @xmath0 orbit of @xmath27 is a closed subset @xmath70 of @xmath35 .
let @xmath71 be another stratum , and let @xmath72 denote the set of all translations surfaces @xmath73 that are branched covers of @xmath74 .
we will always assume that @xmath75 is such that @xmath72 is not - empty .
then @xmath72 is @xmath0 invariant .
there are two types of veech surfaces : arithmetic and non - arithmetic .
a surface @xmath76 is an arithmetic veech surface if and only if @xmath27 is a ( holomorphic ) branched cover of a torus , @xmath28 is the pullback by the covering map of the standard differential @xmath77 on the torus , and the branch points project to points of finite order ( under the additive group of the torus ) .
equivalently ( see @xcite ) @xmath16 is an arithmetic veech surface if and only if @xmath50 is commensurable to @xmath78 .
all other veech surfaces are called non - arithmetic ( and their veech groups , which are always non - uniform lattices , are non - arithmetic lattices in @xmath0 ) . the case where @xmath27 is arithmetic was analyzed in @xcite . in this paper , we assume that @xmath27 is not arithmetic , which implies that the genus of @xmath27 is greater then 1 . then considering the euler characteristic , it is easy to see that the degree of @xmath8 is determined by @xmath70 and @xmath75 .
this implies that @xmath72 is closed .
( in the case where the genus of @xmath27 is 1 , one also has to fix the degree of the cover ; see @xcite for the details . )
the main result of this paper is a classification of the @xmath2-invariant ergodic measures on @xmath72 .
this allows us to prove asymptotic formulas of the form ( [ eq : asymp ] ) for @xmath79 ( see theorem [ theorem : branched : cover : count ] ) . in particular we prove the following : [ theorem : triangle : count ] let @xmath80 be a triangle with angles @xmath81 where @xmath4 , @xmath5 odd .
then , as @xmath15 , @xmath82 the fact that the surface @xmath83 associated to @xmath80 is not veech but is a branched cover of degree 2 of a veech surface is due to p. hubert and t. schmidt ( see proposition 4 in @xcite and its proof ) .
we should also note that if @xmath84 then the veech group of @xmath83 is infinitely generated ( see @xcite ) .
however , the veech group of @xmath83 plays no direct role in our analysis .
here is an outline of the paper .
section [ measclasssect ] states our main theorem .
section [ preliminariessect ] establishes notation and presents a few basic lemmas .
section [ shearingsect ] explains shearing , " the foundation of our study of invariant measures .
section [ mainpfsect ] proves our main theorem ( [ theorem : classification ] ) that classifies @xmath2-invariant measures .
section [ countabilitysection ] proves that there are only countably many closed orbits of a certain type .
section [ avgcirclesect ] uses our main theorem ( and the countability result of [ countabilitysection ] ) to prove that large circles in @xmath0-orbits become uniformly distributed with respect to certain natural measures .
section [ application ] applies the equidistribution result of [ avgcirclesect ] to derive asymptotic estimates for the number of periodic trajectories in branched covers of veech surfaces .
let @xmath85 . let @xmath27 be a veech surface , which means that @xmath86 is a lattice in @xmath87 . here , we use @xmath27 to also denote the isometry class of @xmath27 ; this is a single point in the moduli space . for @xmath88 , we define @xmath89 to be the natural fiber bundle over @xmath90 whose fiber over @xmath27 is @xmath91 .
thus , a point of @xmath89 is represented by @xmath92 , where @xmath93 and @xmath94 . in other words ,
a point in @xmath89 represents a surface in @xmath95 together with @xmath96 marked points on @xmath97 .
we note that the space @xmath72 parameterizing branched covers is itself a finite branched cover of the space @xmath89 for a suitable @xmath96 .
( the covering map just maps @xmath98 to the surface in @xmath70 it covers , and notes the locations of the branch points . )
thus , to classify the @xmath2-invariant measures on @xmath72 it is enough to classify @xmath2-invariant measures on @xmath89 ( see [ fincovclassify ] ) .
if @xmath27 is a torus , then @xmath89 can be identified with the homogeneous space @xmath99 .
in this situation , a special case of ratner s theorem @xcite classifies all the ergodic @xmath2-invariant probability measures on @xmath89 .
we generalize this to allow @xmath27 to be any veech surface . the proof is based heavily on ideas of ratner @xcite and margulis - tomanov @xcite .
an introduction to these ideas can be found in @xcite .
let @xmath100 be the singular set of @xmath27 .
then for @xmath101 , @xmath102 is the singular set of @xmath103 .
let @xmath104 , and let @xmath105 denote the set @xmath106 where @xmath101 and @xmath107 .
then @xmath108 is isomorphic to the natural fiber bundle over @xmath109 whose fiber over @xmath27 is @xmath110 .
we have a natural embedding of @xmath42 in the space @xmath111 of smooth vector fields on @xmath112 , so , for each @xmath113 and @xmath114 , we have a trajectory @xmath115 in @xmath112 that is defined for @xmath116 in a certain open interval containing @xmath117 ( until the trajectory hits the singular set ) .
we are interested only in the forward trajectory , that is , for @xmath118 . by including the singular points of @xmath27
, we extend @xmath119 to a continuous curve @xmath120 in @xmath27 that is defined for @xmath116 in a closed interval ( and for all points in @xmath27 ) : * let @xmath121 for all @xmath113 and @xmath122 ; and * if @xmath123 and @xmath116 is in the closure of the domain of @xmath119 , let @xmath124 then each @xmath113 defines a function @xmath125 , defined by @xmath126 , where @xmath127 is a dense , open subset of @xmath27 . note that @xmath128 is a local isometry ( hence continuous ) . on the other hand , @xmath128 is usually not invertible , because a singular point will typically have several preimages .
in addition , @xmath128 is usually not uniformly continuous , because of branch cuts . for @xmath129
, we have a continuous map @xmath130 ( where @xmath131 is a certain subset of @xmath89 ) , defined by @xmath132 ( thus , @xmath133 does not change the surface @xmath27 , but moves the marked points in the directions specified by @xmath134 . )
let @xmath135 be the pseudosemigroup generated by @xmath136 .
( the prefix pseudo " simply refers to the fact that these maps are not defined on the entire space @xmath89 , but only on a subset . )
although the maps in @xmath135 may not be one - to - one , they are always finite - to - one . for @xmath129 ,
let @xmath137 be the restriction of @xmath133 to @xmath138 .
then @xmath137 is a diffeomorphism ( and local isometry ) from a dense open subset of @xmath139 to a dense open subset of @xmath139 .
let @xmath140 be the pseudogroup that is generated by @xmath141 .
we remark that @xmath140 is transitive on @xmath142 .
note that each of @xmath143 and @xmath144 is normalized by the action of @xmath145 on @xmath89 , so we have corresponding semidirect products @xmath146 and @xmath147 .
let @xmath148 and @xmath149 note that , for @xmath150 , we have @xmath151 on the intersection of their domains , so @xmath152 is a pseudosemigroup .
also , @xmath152 commutes with the action of @xmath2 .
let @xmath153 be an ergodic @xmath2-invariant probability measure on @xmath89 .
the projection of @xmath153 to @xmath154 is @xmath2-invariant , so it must be either lebesgue measure or the arc - length on a closed @xmath2-orbit @xcite .
the interesting case is when the projection is lebesgue .
a weak statement of our results is simply to say that , in this case , some horizontal translate of @xmath153 must be @xmath87-invariant : [ theorem : u->g ] suppose @xmath153 is any ergodic @xmath2-invariant probability measure on @xmath89 , such that the projection of @xmath153 to @xmath154 is lebesgue
. then there exists @xmath155 , such that @xmath156 is @xmath87-invariant ( and the domain of @xmath157 has full measure ) . to obtain a more precise description of the @xmath2-invariant measures , one need only describe the @xmath87-invariant measures on @xmath89 .
[ remark : beforethm ] 1 .
[ remark : beforethm - gamma = g ] it is easy to see that the @xmath87-invariant probability measures on @xmath89 are in natural one - to - one correspondence with the @xmath61-invariant probability measures on @xmath91 ( cf . , e.g. , ( * ? ?
* pf . of cor .
[ remark : beforethm - nosing ] it is the @xmath61-invariant measures on @xmath158 that are the most important to understand , because it is easy to see that every ergodic measure on @xmath91 arises from the following construction .
choose some @xmath159 and some probability measure @xmath160 on @xmath161 that is invariant under a finite - index subgroup of @xmath61 .
the corresponding measure on @xmath162 is invariant under a finite - index subgroup @xmath163 of @xmath61 . by averaging over @xmath164 ,
this yields a @xmath61-invariant measure supported on the subset @xmath165 of @xmath91 .
we will show that every ergodic measure is carried by a nice subspace of @xmath91 .
in particular , any ergodic measure carried by @xmath158 is the lebesgue measure on a flat submanifold of @xmath158 .
the natural lebesgue measure on the diagonal @xmath166 of @xmath167 is a @xmath61-invariant probability measure on @xmath167 .
note that @xmath168 is a @xmath87-invariant subspace of @xmath169 , and that the pseudogroup @xmath170 of diffeomorphisms it generates is transitive on @xmath18 .
[ theorem : gammainvt ] suppose @xmath153 is an ergodic @xmath61-invariant probability measure on @xmath91
. then there exist * a point @xmath171 , and * a @xmath87-invariant linear subspace @xmath172 of @xmath173 , such that 1 . the orbit @xmath174 of @xmath175 under @xmath176 is a closed subset of @xmath91 whose dimension is @xmath177 , 2 .
[ theorem : gammainvt - finiteindex ] some finite - index subgroup of @xmath61 fixes @xmath174 setwise , and 3 .
@xmath153 is the @xmath174-invariant lebesgue measure on @xmath178 . 1 . conversely ,
if @xmath172 is @xmath87-invariant , @xmath174 is closed , and some finite - index subgroup of @xmath61 fixes @xmath174 , then the @xmath176-invariant lebesgue measure on @xmath178 is a @xmath61-invariant probability measure .
however , it may not be ergodic .
we wish to emphasize that conclusion [ theorem : gammainvt]([theorem : gammainvt - finiteindex ] ) implies the set @xmath178 is a finite union of translates of @xmath174 .
the theorem can be stated in the following equivalent form ( see [ remark : beforethm]([remark : beforethm - gamma = g ] ) ) : theorem : gammainvt [ theorem : ginvtmeas ] suppose @xmath153 is an ergodic @xmath87-invariant probability measure on @xmath142
. then there exist * a point @xmath179 and * a @xmath87-invariant linear subspace @xmath172 of @xmath173 , such that 1 . the orbit @xmath180 of @xmath175 under @xmath181 is a closed subset of @xmath89 whose dimension is @xmath182 , and 2 .
@xmath153 is the @xmath183-invariant lebesgue measure on this orbit .
this results in the following explicit version of theorem [ theorem : u->g ] : [ theorem : classification ] suppose @xmath153 is an ergodic @xmath2-invariant probability measure on @xmath142
. then there exist * a point @xmath179 , * a @xmath87-invariant subspace @xmath172 of @xmath173 , and * some @xmath184 , such that 1 .
@xmath185 , 2 .
the orbit @xmath180 of @xmath175 under @xmath181 is a closed subset of @xmath89 whose dimension is @xmath186 , and 3 .
@xmath187 is the @xmath188-invariant lebesgue measure on this orbit .
we will give an application to counting the number of periodic trajectories on @xmath27 ( see [ application ] ) .
theorems [ theorem : u->g ] , [ theorem : gammainvt ] , and [ theorem : ginvtmeas ] have been stated only for expository purposes they are not a part of the logical development .
we prove only theorem [ theorem : classification ] , and the interested reader can easily derive the other theorems as corollaries .
our results imply that the closure of every @xmath61-orbit in @xmath91 is of a nice geometric form .
since @xmath189 is a @xmath61-invariant finite set , it suffices to describe the orbits of points in @xmath190 : corollary : gorbit [ corollary : gammaorbit ] suppose @xmath191 . then there exists a @xmath87-invariant linear subspace @xmath172 of @xmath173 , such that 1 . the orbit @xmath174 of @xmath175 under @xmath176 is a closed subset of @xmath91 ( and its dimension is @xmath177 ) , 2 .
some finite - index subgroup of @xmath61 fixes @xmath174 setwise , and 3 .
@xmath178 is the closure of the @xmath61-orbit of @xmath175 .
we collect all the notation in this section .
some of this repeats the definitions given in the previous sections .
[ prelimnotation ] * let @xmath192 .
* there is a natural action of @xmath87 on the moduli space of translation surfaces .
we can visualize this as a composition of `` the usual linear action '' with `` cut and paste . ''
we note that `` cut and paste '' is an isometry on the surface ( and in fact preserves the horizontal and vertical directions as well ) .
* let @xmath27 be a veech surface , which means that @xmath193 is a lattice in @xmath87 . here
, we use @xmath27 to also denote the isometry class of @xmath27 ; this is a single point in the moduli space .
* let @xmath88 .
* we define @xmath89 to be the natural fiber bundle over @xmath194 whose fiber over @xmath195 is @xmath91 .
thus , a point of @xmath89 is represented by @xmath92 , where @xmath196 and @xmath94 . *
the metric on @xmath89 is defined by @xmath197 , [ ( m '' , ( q_i)_{i=1}^k ] \bigr ) \\ & \hskip0.5 in = \min_{\begin{matrix } g \in g , \\
g m ' = m '' \end{matrix } } \left ( \|g - \operatorname{id}\| + \sum_{i=1}^k d_{m''}(g p_i , q_i ) \right ) .
\end{aligned}\ ] ] * note that @xmath89 is @xmath87-equivariantly homeomorphic to @xmath198 , where * * @xmath61 acts on @xmath87 by right multiplication , * * @xmath61 acts on @xmath91 componentwise , and * * @xmath87 acts on @xmath89 via @xmath199 .
* let @xmath100 be the singular set of @xmath27 .
* let @xmath104 .
* let @xmath200 .
* any @xmath129 naturally defines a vector field on @xmath139 . by taking the time - one map of the corresponding flow ( where it is defined ) , we obtain a diffeomorphism @xmath137 between two dense open subsets of @xmath139 .
the collection @xmath201 generates a transitive pseudogroup @xmath140 of local diffeomorphisms of @xmath142 .
+ we extend @xmath137 to a ( continuous ) transformation @xmath133 that is defined on a slightly larger subset of @xmath202 , by letting @xmath203 if the limit exists .
( see [ measclasssect ] for a more concrete definition of @xmath133 , in terms of the flow corresponding to @xmath134 . )
we let @xmath144 be the pseudosemigroup generated by these maps .
+ because the action of @xmath87 on @xmath89 normalizes @xmath140 and @xmath140 , we have semidirect products @xmath204 and @xmath205 .
note that @xmath204 is transitive on @xmath142 .
+ it is important to note that , because of the singularities and resulting branch cuts , @xmath137 is usually not uniformly continuous ( even though it is a local isometry ) .
furthermore , @xmath206 is not a uniformly continuous function of @xmath134 .
see figure [ fig : opposite ] . abusing notation , we may sometimes write @xmath207 instead of @xmath208 .
* let @xmath209 , where @xmath210 .
* let @xmath211 , where @xmath212 .
* let @xmath213 , where @xmath214 .
* let @xmath153 be a @xmath2-invariant probability measure on @xmath89 , such that @xmath153 projects to the lebesgue measure on @xmath215 .
* let @xmath216 and @xmath217 .
then @xmath152 is a pseudosemigroup .
* let @xmath218 and @xmath219 * for @xmath220 , we define @xmath221 by @xmath222 .
thus , @xmath223 * the set @xmath224 is @xmath2-invariant .
thus , it is either null or conull
. let us assume it is conull .
( if not , then by ergodicity , there exists @xmath225 such that @xmath156 is supported on @xmath226 . so @xmath156 can be described by a construction similar to remark [ remark : beforethm]([remark : beforethm - nosing ] ) .
the conclusion of theorem [ theorem : classification ] is therefore obtained by induction on @xmath96 . ) + note that @xmath227 acts on @xmath228 , by @xmath229 .
therefore , the group @xmath230 acts on @xmath228 .
* let @xmath231 note that @xmath232 acts on @xmath233 , but we do not yet know that @xmath233 is conull . * let @xmath234 . because @xmath227 acts on @xmath228 , we know that @xmath235 is a closed subgroup of @xmath227 .
* let @xmath236 .
equivalently , @xmath237 * let @xmath238 .
note that @xmath172 is a @xmath87-invariant subspace of @xmath173 , so @xmath239 is a pseudosemigroup .
* let @xmath240 .
* let @xmath241 . by permuting coordinates
, we may assume @xmath242 .
* let @xmath243 ( the first @xmath244 coordinates ) be the natural projection .
* for @xmath245 , we use @xmath246 to denote the fiber measure of @xmath153 over the point @xmath247 of @xmath248 .
the following is obtained by applying the pointwise ergodic theorem to the action of @xmath2 on @xmath89 .
[ unifgeneric ] for any @xmath249 , there is a uniformly generic set " @xmath250 in @xmath89 , such that 1 .
@xmath251 , 2 . for every @xmath252 and every compact subset @xmath253 of @xmath89 , with @xmath254 , there exists @xmath255 ,
such that , for all @xmath256 and all @xmath257 , we have @xmath258 \mid d(u^s \omega , k ) < \epsilon \,\ } > ( 1 - \epsilon ) ( 2l ) , \ ] ] where @xmath259 is the lebesgue measure on @xmath260 .
[ simultaneous ] suppose a lie group @xmath60 acts continuously on a borel subset @xmath27 of a locally compact metric space .
if * @xmath2 is a one - parameter , normal subgroup of @xmath60 , and * @xmath153 is an ergodic @xmath2-invariant probability measure on @xmath27 , then 1 .
[ simultaneous - conull ] there is a @xmath2-invariant , borel subset @xmath261 of @xmath27 , such that 1 .
@xmath262 , and 2 .
@xmath263 for all @xmath264 , + and 2 .
[ simultaneous - cpct ] for any @xmath252 , there is a compact subset @xmath253 of @xmath27 , such that 1 .
@xmath254 , and 2 .
@xmath265 for all @xmath266 .
ratner s argument in ( * ? ? ?
2.2 ) shows , for each @xmath267 , that there is a neighborhood @xmath268 of @xmath269 in @xmath270 and a conull @xmath2-invariant subset @xmath271 of @xmath27 , such that @xmath272 for the reader s convenience , we sketch the proof of this fact . because @xmath269 normalizes @xmath2 but does not belong to @xmath273 , we know that @xmath274 is @xmath2-invariant and ergodic , but is not equal to @xmath153 . therefore @xmath274 and @xmath153 are mutually singular , which implies there is a compact subset @xmath275 of @xmath27 , such that @xmath276 and @xmath277 . by continuity and compactness
, there are open neighborhoods @xmath278 and @xmath279 of @xmath275 , and a symmetric neighborhood @xmath280 of @xmath281 in @xmath60 , such that @xmath282 and @xmath283 . from the pointwise ergodic theorem ,
we know there is a conull @xmath2-invariant subset @xmath271 of @xmath27 , such that the @xmath2-orbit of every point in @xmath271 spends 99% of its life in @xmath284 .
now suppose there exists @xmath285 , such that @xmath286 .
then there exists @xmath287 , @xmath288 , and @xmath289 , such that @xmath290 and @xmath291 both belong to @xmath284 .
this implies that @xmath292 and @xmath293 both belong to @xmath279 .
this contradicts the fact that @xmath294 .
( [ simultaneous - conull ] ) cover @xmath295 with countably many balls @xmath296 , and let @xmath297 .
( [ simultaneous - cpct ] ) let @xmath253 be any compact subset of @xmath261 with @xmath254 .
[ kms : uniqerg ] for a.e .
@xmath113 , the foliation by orbits of @xmath298 is uniquely ergodic on @xmath112 .
[ uniqerghoriz ] suppose @xmath153 is a @xmath2-invariant probability measure on @xmath228 whose projection to @xmath154 is lebesgue . if @xmath153 is @xmath227-invariant , then @xmath153 is the lebesgue measure .
theorem [ kms : uniqerg ] implies that the foliation by orbits of @xmath227 is uniquely ergodic on @xmath299 , for a.e
. @xmath101 .
thus , almost every fiber of @xmath153 over @xmath154 is the lebesgue measure .
in this section , we prove the crucial fact that the direction of fastest transverse divergence between two nearby @xmath2-orbits is always along the stabilizer of @xmath153 .
the analogous statement for unipotent flows is a cornerstone of the proof of ratner s theorem ( * ? ? ?
3.3 ) , ( * ? ? ?
7.5 ) , ( * ? ? ?
5.2.4@xmath300 ) . [ divnotn ] * for any @xmath101
, we may write @xmath301 with @xmath302 . for a sequence @xmath303
, we have @xmath304 if and only if @xmath305 . *
suppose @xmath306 , say .
for @xmath220 with @xmath307 , let * * @xmath308 , * * @xmath309 , and * * @xmath310 .
+ note that @xmath311 if @xmath312 .
* suppose @xmath313 and @xmath314 are two sequences in a metric space . if @xmath315 , we may write @xmath316 . [ ugucalc ] a simple calculation shows that @xmath317 for a sequence @xmath304 , we denote @xmath318 , and @xmath319 .
then @xmath320 if @xmath321 ( and @xmath322 ) .
( shearing " ) let us discuss the action of @xmath2 on @xmath173 . for any @xmath323 and @xmath129 , we have @xmath324 assume , now that @xmath325 there is some @xmath326 , such that @xmath327
. then @xmath328 thus , under the @xmath2-flow , @xmath329 and @xmath330 move apart along a leaf of the @xmath227-foliation .
in other words , the direction in which two nearby points move apart fastest is along @xmath227 .
we use the notation of ( [ divnotn ] ) to state the main result of this section .
[ divstab ] for every @xmath249 , there is a compact subset @xmath250 of @xmath228 , with @xmath331 , such that , if * @xmath332 are convergent sequences in @xmath250 , * @xmath333 for some @xmath334 and @xmath335 , * @xmath304 and @xmath336 , * @xmath337 with @xmath338 and * @xmath339 converges , then @xmath340 .
define @xmath341 \to a \operatorname{horiz}$ ] by @xmath342 where , letting @xmath343 and @xmath344 , we have @xmath345 it is clear that @xmath346 is continuous
. we will show @xmath347 for all @xmath116 .
then @xmath348 as desired .
let @xmath250 be a uniformly generic set for the action of @xmath2 on @xmath89 with @xmath331 ( see [ unifgeneric ] ) . by passing to a subset
, we may assume that @xmath349 and that @xmath250 is compact .
for any @xmath252 , we know , from lemma [ simultaneous ] ( with @xmath350 ) , that there is a compact subset @xmath253 of @xmath228 , such that @xmath351 and @xmath352 , for all @xmath353 . when @xmath5 is large , the definition of @xmath250 implies that @xmath354 for all but @xmath355% of the values of @xmath356 in the interval @xmath357 $ ] ( or longer intervals ) ( see [ unifgeneric ] ) .
note that the jacobian of @xmath358 is uniformly bounded on @xmath359 $ ] .
more precisely , @xmath360 , so @xmath361 .
therefore , @xmath362 for all but @xmath363% of the values of @xmath356 in the interval @xmath364 $ ] .
thus , ( [ divstabpf - uclose ] ) and ( [ divstabpf - uclose ] ) hold _ simultaneously _ for all but @xmath365% of the values of @xmath356 in the interval @xmath364 $ ] .
let @xmath366 .
because @xmath367 , we know that translating @xmath368 by a vector in @xmath227 can not move it into @xmath100 . hence @xmath369 for any compact subset @xmath370 of @xmath227
. therefore , if @xmath5 is sufficiently large , and , for convenience , we let @xmath371 , then * @xmath372 has no singularities in @xmath373 so * @xmath374 has no singularities in @xmath375 this implies that @xmath376 therefore @xmath377 when ( [ divstabpf - uclose ] ) and ( [ divstabpf - uclose ] ) hold simultaneously , we conclude that @xmath378 from the definition of @xmath253 , we conclude that @xmath379 . that is , @xmath380 belongs to @xmath381 for all but @xmath365% of the values of @xmath116 in @xmath382 $ ] .
because @xmath355 is arbitrary , @xmath346 is continuous , and @xmath381 is closed subgroup , we conclude that @xmath383 must actually belong to the stabilizer for all values of @xmath116 , as desired .
we assume the notation of ( [ prelimnotation ] ) .
recall , in particular , that @xmath153 is carried by @xmath228 , and that the group @xmath384 acts on @xmath228 .
[ fibersonhoriz ] almost every fiber of @xmath153 over @xmath385 is supported on finitely many orbits of @xmath386 .
because @xmath153 is an ergodic probability measure , it suffices to show that almost every fiber is supported on countably many such orbits . for @xmath250 as in proposition [ divstab ] , we know @xmath387 is conull , so it suffices to show , for each @xmath249 , that each fiber of @xmath250 is contained in the union of countably many orbits of @xmath386 .
suppose not .
( this will lead to a contradiction . ) because any uncountable set contains one if its accumulation points , there exist @xmath388 and a sequence @xmath313 in @xmath97 , such that * @xmath389 , * @xmath390 , * @xmath391 , and * @xmath392 . because @xmath390 and @xmath391 , we may write @xmath393 for some @xmath394 with @xmath395 . by assumption , we know @xmath396 , so @xmath397 is a nonzero element of @xmath398 .
because @xmath399 and proposition [ divstab ] implies that @xmath400 converges to an element of @xmath401 , we conclude that @xmath402 is nontrivial .
this contradicts the definition of @xmath235 .
[ finfiboverxd ] after restricting to an appropriate conull subset @xmath403 of @xmath228 , each fiber of @xmath404 is finite . we know
, from proposition [ fibersonhoriz ] , that almost every fiber of @xmath404 is carried by only finitely many orbits of @xmath405 .
( from theorem [ kms : uniqerg ] , we may assume that each of these is an embedded copy of @xmath386 . ) letting @xmath406 we may define a measurable function @xmath407 $ ] by @xmath408 .
this function is essentially @xmath2-invariant , so it must be essentially constant . because this is true for all @xmath244 , we conclude that @xmath404 is carried by a single point in each orbit of @xmath386 .
since there are only finitely many such orbits to consider , we conclude that a.e .
fiber consists of a finite number of atoms , as desired .
[ ainvt ] we may assume @xmath153 is @xmath409-invariant . choose @xmath250 as in proposition [ divstab ] , with @xmath410 . from corollary [ uniqerghoriz ] , we know that @xmath153 projects to the lebesgue measure on @xmath385 .
furthermore , by passing to a conull subset , we may assume @xmath250 has finite fibers over @xmath385 ( see [ finfiboverxd ] ) .
thus , it is easy to see that there exist @xmath411 , @xmath412 , and @xmath413 , such that @xmath414 , @xmath415 , and @xmath395 . then , in the notation of ( [ divnotn ] ) , with @xmath416 , and choosing @xmath417 appropriately , we have @xmath418 ( cf .
pf . of [ fibersonhoriz ] to see that @xmath419 )
. we conclude , from proposition [ divstab ] , that the identity component of @xmath420 is nontrivial . because the identity component of @xmath421 is trivial ( by definition of @xmath235 ) , we conclude that @xmath422 contains a one - parameter subgroup that is not contained in @xmath227 .
any such subgroup is conjugate to @xmath409 ( via an element of @xmath227 ) .
thus , by replacing @xmath153 with a translate under @xmath227 , we may assume @xmath153 is @xmath409-invariant .
@xmath233 is conull . by passing to a quotient
, we may assume @xmath423 .
for each nonzero vector @xmath424 , let @xmath425 note that @xmath426 .
suppose there is a subset @xmath427 of positive measure in @xmath87 , such that @xmath428 for @xmath429 .
then the pointwise ergodic theorem implies , for a.e .
@xmath430 , that we have @xmath431 for all @xmath432 in a non - null subset @xmath433 of @xmath2 . furthermore , because @xmath153 is @xmath2-invariant , we may assume @xmath434 for all @xmath435 .
therefore @xmath436 for all @xmath435 .
this contradicts the fact that , because @xmath437 is countable whenever @xmath438 , we have @xmath439 for all but countably many choices of the line @xmath440 .
[ allinvvert ] there is a conull subset @xmath261 of @xmath233 , such that @xmath441 for all @xmath442 .
let @xmath261 be a generic set for for the action of @xmath409 on @xmath233 ; thus , @xmath261 is conull and , for each @xmath442 , @xmath443 given @xmath444 , such that @xmath445 with @xmath446 and @xmath447 , we wish to show @xmath448 . choose a sequence @xmath449 , such that @xmath450 and @xmath451 each belong to @xmath250 . because @xmath452 and @xmath453 is the foliation that is contracted by @xmath454 , we know that @xmath455 .
furthermore , because @xmath409 acts on the lie algebra of @xmath456 with twice the weight that it acts on the lie algebra of @xmath457 , we see that @xmath458 .
thus , letting @xmath356 be within a constant multiple of @xmath459 , we see , in the notation of ( [ divnotn ] ) , with @xmath460 and @xmath461 , that @xmath462 , but @xmath463 . thus , proposition [ divstab ] asserts that @xmath464 converges to a nontrivial element of @xmath465 . since @xmath466 is a scalar multiple of @xmath467 ,
we conclude that @xmath468 .
therefore @xmath469 , so @xmath448 .
we require the following entropy estimate .
[ entlem ] suppose @xmath172 is a closed connected subgroup of @xmath453 that is normalized by @xmath470 , and let @xmath471 be the jacobian of @xmath472 on @xmath172 . 1
. [ entlem - inv- > ] if @xmath153 is @xmath172-invariant , then @xmath473 .
[ entlem - gmininw ] if there is a conull , borel subset @xmath261 of @xmath89 , such that @xmath474 , for every @xmath442 , then @xmath475 .
[ entlem - equal ] if the hypotheses of ( [ entlem - gmininw ] ) are satisfied , and equality holds in its conclusion , then @xmath153 is @xmath172-invariant .
[ vyinvt ] @xmath153 is @xmath476-invariant . from lemma [ entlem]([entlem - inv- > ] ) , with @xmath472 in the role of @xmath477 , we have @xmath478 from proposition [ allinvvert ] and lemma [ entlem]([entlem - gmininw ] ) , we have @xmath479 combining these two inequalities with the facts that * @xmath480 and * @xmath481 , we have @xmath482 thus , we must have equality throughout , so the desired conclusion follows from lemma [ entlem]([entlem - equal ] ) .
@xmath153 is the lebesgue measure on a single orbit on @xmath142 of the pseudogroup @xmath483 .
we know : * @xmath2 preserves @xmath153 ( by assumption ) , * @xmath235 preserves @xmath153 ( by definition ) , * @xmath409 preserves @xmath153 ( see [ ainvt ] ) , and * @xmath476 preserves @xmath153 ( see [ vyinvt ] ) .
therefore , @xmath153 is preserved by the pseudogroup @xmath484 generated by these maps .
because * this pseudogroup is transitive on the quotient @xmath485 , and * @xmath153 has finite fibers over @xmath485 ( see [ finfiboverxd ] ) , this implies that some orbit of the pseudogroup has positive measure . by ergodicity of @xmath2 , then this orbit is conull . to obtain the conclusions of theorem [ theorem : classification ] , we let @xmath238 . then @xmath153 is supported on the @xmath486-orbit of some point @xmath487 in @xmath89 . note that , by choosing @xmath177 to be minimal , we can guarantee that whenever @xmath40 is a singular point of @xmath27 , the subspace @xmath172 projects to @xmath117 in the @xmath488 coordinate of @xmath173 .
therefore , the dimension of the orbit is equal to the dimension of the pseudosemigroup .
for our application , we need the following analogue of ( * ? ? ? * cor .
a(2 ) ) .
[ prop : countable ] the set of subspaces @xmath172 occurring in theorem [ theorem : classification ] is countable . for each such @xmath172 ,
the set of closed orbits of @xmath489 is countable .
[ lemma : ws : countable ] the set of @xmath87-invariant subspaces @xmath172 of @xmath173 such that there exists @xmath171 with @xmath490 closed is countable .
let @xmath491 be the dimension of @xmath172 .
after possibly renumbering the factors , we may assume that @xmath492 then if we denote elements of @xmath173 by @xmath493 where each @xmath494 , then @xmath172 is given by the following equations : for @xmath495 , @xmath496 recall that the linear holonomy map @xmath497 is given by @xmath498 , where @xmath499 is the holomorphic @xmath500-form that determines the flat structure on @xmath27 .
let @xmath501 denote the image of @xmath502 , and let @xmath503 denote the set of real numbers @xmath504 such that there exist nonzero @xmath505 , @xmath506 with @xmath507 .
then @xmath503 is clearly a countable set .
we will show that each @xmath508 belongs to @xmath509 .
let @xmath510 denote projection onto the first @xmath511 factors .
note that the intersection of @xmath512 with each fiber of @xmath404 is finite .
now pick @xmath244 , @xmath513 , and @xmath514 , @xmath495 .
we may assume that @xmath515 .
choose @xmath516 such that @xmath40 and @xmath517 are non - singular .
let @xmath518 be any element of @xmath519 with @xmath520 .
we represent @xmath518 by a piecewise linear closed curve on @xmath27 beginning and ending at @xmath40 and not passing through any singularities ; we will also denote this representative by @xmath518 .
we obtain a closed curve @xmath521 by keeping @xmath522 fixed for @xmath523 , @xmath524 .
because @xmath512 is a branched cover of @xmath525 , @xmath526 lifts to a closed curve @xmath527 in @xmath512 .
let @xmath528 denote the projection of @xmath527 to the @xmath529 factor .
we wish to calculate @xmath530 , so let us describe @xmath528 more precisely .
the curve @xmath518 is a collection of segments connecting points @xmath531 , with @xmath532 , @xmath533 .
then @xmath528 is a collection of segments connecting the points @xmath534 , with @xmath535 . by perturbing the @xmath536 ,
we can make sure that @xmath528 is well defined and is not passing through any singularities . by construction
, the endpoint @xmath537 of @xmath528 belongs to the finite set @xmath538 . after replacing @xmath518 by an integer multiple
, we may assume that @xmath528 is closed .
but , in view of the explicit description of @xmath528 , @xmath539 , so @xmath540 .
in the rest of this section we will abuse notation by writing @xmath541 for @xmath542 .
[ lemma : periodic : countable ] let @xmath27 be a veech surface , and let @xmath61 be the veech group of @xmath27 .
a point @xmath175 is called a periodic point if the @xmath61 orbit of @xmath175 is finite .
then the set of periodic points is countable . when @xmath27 is non - arithmetic , which is the only case that we need to discuss , it is proven in @xcite that the number of periodic points is countable ( in fact , _ finite _ ) .
the following generalization of this statement also follows from the results of @xcite , but we include a short proof of as a warm up to the proof of proposition [ prop : countable ] .
it is clearly enough to show that for each @xmath543 , the set @xmath544 of points of period @xmath5 is countable . to do this
it is enough to show that for each point @xmath545 , there exists a neighborhood @xmath2 of @xmath175 that does not contain any other points of @xmath544 .
suppose the last statement is false .
then there exists a sequence of points @xmath546 such that @xmath547 .
we may assume after passing to a subsequence that the @xmath517 approach @xmath175 from some given direction @xmath134 ( i.e. that @xmath548 ) .
let @xmath163 denote the intersection of all the index @xmath5 subgroups of @xmath61 . then , since @xmath61 is finitely generated , @xmath163 is of finite index in @xmath61 and for each @xmath549 , and all @xmath514 , @xmath550 .
then each element of @xmath163 must fix @xmath134 .
this contradicts the fact that @xmath163 , being a finite index subgroup of @xmath61 , is zariski dense in @xmath87 .
it remains to prove the following assertion : let @xmath551 be an @xmath87-invariant subspace
. then the set @xmath552 of closed orbits of @xmath553 is countable .
we triangulate @xmath27 , with the vertices at the singular points .
this yields a cell decomposition of @xmath91 in which the cells @xmath554 of maximal dimension are products of triangles .
let @xmath555 denote the interior of @xmath556 , and let @xmath190 denote the union of the @xmath555 .
for @xmath171 , let @xmath557 denote the distance between @xmath175 and the complement of @xmath190 ( i.e. the distance to the boundary of the cell containing @xmath175 ) .
let @xmath558 , and let @xmath559 be any @xmath87-invariant complement to @xmath172 .
we may assume that @xmath172 is given by the equations ( [ eq : w ] ) .
in view of lemma [ lemma : periodic : countable ] we may also also assume that @xmath172 has dense projection onto any of the @xmath42 factors ( i.e. for a fixed @xmath514 , not all @xmath508 are @xmath117 ) .
then , for any @xmath560 , @xmath561 is dense in @xmath512 .
let @xmath562 be an @xmath563-tuple of non - negative integers , and let @xmath564 denote the set of of @xmath560 such that @xmath565 has exactly @xmath566 connected components
. now suppose @xmath552 is uncountable .
then there exist @xmath562 such that @xmath564 is uncountable . then by compactness , there exist @xmath512 in @xmath564 such that for every @xmath567 there exists @xmath568 such that the hausdorff distance between @xmath512 and @xmath569 is less then @xmath355 . let @xmath570 be the minimum over @xmath244 of the minimal distance between connected components of @xmath565 .
let @xmath571 , and number all the connected components of the intersection of @xmath512 with the interiors of the cells as @xmath572 , @xmath573 .
let @xmath574 denote the generators of @xmath61 .
we may choose a point @xmath40 in each @xmath572 such that for all @xmath514 , @xmath575 , @xmath576 is in the interior of some component @xmath577 , where @xmath578 depends on @xmath244 and @xmath514 .
let @xmath579 .
now choose @xmath252 so that : * @xmath580 .
* for any @xmath244 , @xmath581 , we have @xmath582 . * for each @xmath244 , @xmath581 and each @xmath514 , @xmath583 , we have @xmath584
. now choose @xmath585 so that the hausdorff distance between @xmath569 and @xmath512 is less then @xmath355 .
note that if @xmath586 with @xmath587 there exists a unique @xmath588 , such that @xmath589 and @xmath590 .
also @xmath591 does not depend on the choice of @xmath592 , and @xmath593 .
furthermore @xmath594 since @xmath512 and @xmath569 can not share a point in @xmath190 .
let @xmath456 denote the finite set @xmath595 .
we now claim that each generator @xmath596 preserves the set @xmath456 .
indeed consider the points @xmath597 and @xmath598 .
since both @xmath512 and @xmath569 are @xmath61-invariant , we must have @xmath599 and @xmath600 . by construction , @xmath601 , and @xmath602 .
recall that @xmath603 is the only vector in @xmath559 of norm at most @xmath604 such that @xmath605 .
but @xmath606 , and @xmath607 .
also @xmath608 , since @xmath559 is @xmath87-invariant .
thus @xmath609 .
we have proved that for each generator @xmath596 , we have @xmath610 .
this immediately implies that @xmath611 .
then a finite index subgroup of @xmath61 will fix a single vector in @xmath456 , which contradicts the fact that @xmath61 is zariski dense in @xmath87 .
let @xmath612 denote the haar measure on @xmath613 .
for @xmath614 and @xmath123 , let @xmath615 where @xmath616 is the atomic probability measure supported at @xmath617 , and @xmath618 .
then each @xmath619 is a probability measure on @xmath89 .
we can think of @xmath619 as the measure supported on a circle of radius @xmath116 inside the @xmath87-orbit through @xmath617 . in this section
we prove the following theorem : [ theorem : largecircles ] suppose @xmath620 .
then there exists a @xmath87-invariant subspace @xmath172 of @xmath173 such that 1 .
the @xmath621 orbit through @xmath617 is closed , and 2 .
@xmath622 , where @xmath153 is lebesgue measure on this orbit .
[ remark : largecircles ] if @xmath623 , then @xmath153 is the lebesgue measure on @xmath89 .
[ lemma : circles : unipotent ] suppose @xmath624 .
then there is a subsequence @xmath625 such that the measures @xmath626 converge to a probability measure @xmath627 that is invariant under the unipotent element @xmath628 of @xmath87 .
it follows from ( * ? ? ?
* corollary 5.3 ) that there is a subsequence @xmath625 such that the measures @xmath626 converge to a probability measure @xmath627 .
we can find @xmath629 such that @xmath630 converges to @xmath432 .
( recall that @xmath631 is the @xmath632 matrix representing rotation by @xmath633 ) .
now the measures @xmath634 are @xmath630 invariant , hence @xmath627 is @xmath432 invariant .
assume @xmath627 is not the lebesgue measure on @xmath89 .
note that we do not know at this point whether @xmath627 is ergodic .
however , standard results ( using @xmath432-invariance ) imply that @xmath627 projects to lebesgue measure in @xmath154 . for convenience ,
if @xmath635 and @xmath636 , let @xmath637 by theorem [ theorem : classification ] , and by proposition [ prop : countable ] , there exists a @xmath87-invariant proper subspace @xmath172 of @xmath173 and an orbit @xmath512 of @xmath621 , such that @xmath638 we will show that this implies that @xmath639 . in that case , the entire @xmath87-orbit of @xmath617 lies in @xmath512 , so @xmath640 .
furthermore , we show that as long as @xmath172 was chosen as small as possible , @xmath627 must be lebesgue measure on @xmath512 . we choose @xmath172 to be of minimal dimension . from the structure of the @xmath87-invariant subspaces on @xmath173
, we see that @xmath641 , @xmath642 , and after renumbering the factors , we can make sure that @xmath643 thus , @xmath644 is complementary to @xmath172 , and @xmath645 is complementary to @xmath646 in @xmath227 .
[ lemma : support : in : interval ] there exists @xmath252 and a box @xmath647 \times \dots \times [ \alpha_k , \beta_k ] \subset \operatorname{horiz}\ ] ] such that @xmath648 .
let @xmath649 be the ( unique ) orbit of @xmath650 that is open and dense in @xmath512 .
( in other words , @xmath649 consists of the elements in @xmath512 of which as few coordinates as possible are singular points . ) note that @xmath651 by the minimality of @xmath177 , we see that @xmath652 .
hence @xmath653 . by combining this with ( [ lemma : support : in : intervalpf - winvt ] ) and the fact that @xmath654 , we conclude that there is a box @xmath655 , such that @xmath656 .
since @xmath657 , then @xmath658 , as desired . as in the previous sections ,
let @xmath659 be the natural projection onto the first @xmath511 coordinates . for @xmath660 ,
we let @xmath661 .
note that @xmath662 is a finite set .
we claim that @xmath663 to see this , note that , because @xmath172 is a proper subspace of @xmath173 , we have @xmath664 . hence , by induction on @xmath96 , we may assume there is a @xmath87-invariant subspace @xmath665 of @xmath666 , such that the projection of @xmath627 to @xmath385 is the lebesgue measure on the @xmath667 orbit @xmath668 through @xmath669 .
then @xmath670 is conull for @xmath627 , so @xmath671 from the minimality of @xmath177 , we conclude that @xmath672 .
therefore @xmath673 , which establishes the claim .
[ assumexnotino ] we may assume @xmath674 .
( otherwise , from the fact that @xmath512 is a branched cover of @xmath385 ( and lemma [ fincovclassify ] below ) , we would immediately conclude that @xmath627 is the lebesgue measure on @xmath512 , as desired . )
this will lead to a contradiction . for @xmath675 ,
@xmath676 let @xmath677 [ lemma : trig : polynomial ] suppose @xmath678 , where @xmath679 is as defined in lemma [ lemma : support : in : interval ] .
suppose also that @xmath249 , @xmath680 , @xmath681 , and @xmath682 .
then there exists @xmath683 depending only on @xmath684 such that for any @xmath685 and any @xmath686 with @xmath687 we have @xmath688 where @xmath689 .
if we write @xmath690 , with @xmath494 , and also write @xmath691 then the condition ( [ eq : far : from : hor ] ) implies that there exists at least one @xmath514 , @xmath692 with @xmath693 .
the rest of the argument will take place in the @xmath529 factor ( see figure [ fig : boxes ] ) .
we note that the components of the map @xmath694 are trigonometric polynomials of degree @xmath500 .
in other words , the path @xmath695 parametrizes an ellipse .
let @xmath696 .
then if @xmath685 and @xmath697 then @xmath698 . thus , the ellipse @xmath695 leaves @xmath699 .
then in view of the dimensions of the boxes , the portion of the ellipse in @xmath700 is at most @xmath701 times the portion of the ellipse in @xmath699 .
[ lemma : neighborhoods ] for any @xmath702 , there exists @xmath703 and a compact subset @xmath427 of @xmath385 with @xmath704 , such that 1
. @xmath705 does not contain any singular points , other than perhaps points in @xmath512 , and 2 .
[ lemma : neighborhoods - unique ] for each @xmath706 , there is a unique @xmath707 , such that @xmath708 . for @xmath660 ,
let @xmath709 be the surface corresponding to @xmath29 ( so we may write @xmath710 as @xmath711 ) . let @xmath712 denote the singular set of @xmath709 .
for any @xmath702 on the fixed surface @xmath709 , there exist only finitely many horizontal trajectories of length at most @xmath713 connecting points of @xmath714 to points of @xmath714 .
therefore we can find a large compact subset @xmath427 of @xmath385 such that for any @xmath715 , @xmath709 has no horizontal trajectories of length at most @xmath716 connecting points of @xmath714 to other points of @xmath714 . since @xmath627 projects to lebesgue measure on @xmath385 ( see [ projtolebonxd ] ) , we may choose @xmath427 so that @xmath717 .
now we can choose @xmath703 by compactness .
note that , because @xmath718 , we have @xmath719 . therefore @xmath720 . because @xmath721 ( see [ assumexnotino ] ) , we may choose @xmath249 so that @xmath722 .
we may also assume that on the surface corresponding to @xmath617 , the distance between any two singular points is at least @xmath723 .
let @xmath679 , @xmath724 $ ] and @xmath355 be as in lemma [ lemma : support : in : interval ] .
choose @xmath725 so that for all @xmath726 , we have @xmath724 \subset [ -l_1 , l_1]$ ] .
let @xmath727 . now choose @xmath728 and @xmath703 so that lemma [ lemma : neighborhoods ] holds . finally , choose @xmath729 .
assume @xmath730 .
we will abuse notation by writing @xmath541 for @xmath731 .
we claim that if @xmath732 , with @xmath733 , then ( [ eq : far : from : hor ] ) holds .
indeed , we have then @xmath734 so @xmath735 also , @xmath736 therefore @xmath737 now let @xmath738 suppose @xmath739 .
let @xmath740 be the unique element of @xmath741 with @xmath742 , and let @xmath743 note that @xmath744 , so we may let @xmath745 be the component of @xmath746 that contains @xmath633 . by ( the proof of ) lemma [ lemma : trig : polynomial ] , @xmath747 .
we claim that if @xmath748 , then @xmath749 is disjoint from @xmath750 . to see this , note that if @xmath751 , then there exist @xmath752 , such that , letting @xmath753 we have @xmath754 and @xmath755 now if @xmath748 , then @xmath756 , so @xmath757 . lemma [ lemma : neighborhoods]([lemma : neighborhoods - unique ] ) therefore implies that @xmath758 , so @xmath759 .
since each point of @xmath760 is contained in a unique interval , the circle is covered at most twice by the intervals @xmath745 .
it follows that @xmath761 .
equivalently , this means that @xmath762 since this holds for all sufficiently large @xmath116 , we get @xmath763 since @xmath627 projects to lebesgue measure , we know that @xmath764 . hence @xmath765 .
this contradicts lemma [ lemma : support : in : interval ] .
[ corollary : gorbit ] suppose @xmath620 .
then there exists a @xmath87-invariant subspace @xmath172 of @xmath173 , such that the closure of @xmath766 is @xmath767 .
let @xmath172 be as in the conclusion of theorem [ theorem : largecircles ] .
because @xmath768 is closed and @xmath87-invariant , it contains the closure of @xmath766 . on the other hand ,
the support of @xmath619 is a subset of @xmath766 , so @xmath766 is dense in the support of @xmath769 ; that is , @xmath766 is dense in @xmath770 .
corollary [ corollary : gammaorbit ] ( stated at the end of [ measclasssect ] ) follows from ( [ corollary : gorbit ] ) by a standard argument ( inducing the action of @xmath61 to an action of
@xmath87 ) .
we now give the general setup for the counting problems we are considering . for additional background and more detailed definitions , see the introduction to @xcite .
[ notation : counting ] * let @xmath16 be a translation surface .
a _ saddle connection _ on @xmath16 is a straight line segment connecting two singularities .
since a saddle connection has a well defined length and direction , each saddle connection is associated with a non - zero vector in @xmath42 .
let @xmath771 denote the set of vectors in @xmath42 that are associated to saddle connections in @xmath16 .
_ regular _ closed geodesic on @xmath16 , we mean a closed geodesic that does not pass through singularities . * as mentioned in the introduction , any regular closed geodesic is part of a family of freely homotopic parallel closed geodesics of the same length .
such a family is called a _
cylinder_. all the geodesics comprising a cylinder have the same length and direction ; thus we can associate to a cylinder a non - zero vector in @xmath42 .
note that each boundary component of a cylinder is a union of saddle connections .
let @xmath772 denote the set ( with multiplicity ) of vectors in @xmath42 that are associated to cylinders in @xmath16 .
in particular , if @xmath16 is a standard torus , then @xmath773 is the set of primitive vectors in @xmath774 . * for any @xmath775 , let @xmath776 denote the ball in @xmath42 of radius @xmath10 centered at @xmath117 .
* let @xmath777 be a subset of @xmath778 with multiplicity ; i.e. a set of vectors with positive weights .
the weights are usually positive integers ( e.g. , we may consider saddle connections with multiplicity ) , but need not be ( e.g. , we may weight each cylinder by the reciprocal of its area ) . *
let @xmath779 denote the cardinality ( with weights ) of @xmath780 .
we are interested in the asymptotics of @xmath779 as @xmath15 .
if @xmath781 , we will denote @xmath779 by @xmath782 , and if @xmath783 then , as in the introduction , we will denote @xmath779 simply by @xmath41 .
* recall from the introduction that @xmath71 denotes a _ stratum _ of translation surfaces .
* let @xmath784 denote the subset of @xmath71 consisting of the surfaces of area @xmath500 ( where area is taken using the associated translation metric ) .
* as in [ avgcirclesect ] , let @xmath612 denote the haar measure on @xmath785 . * for @xmath786 and @xmath123 , let @xmath787 where @xmath788 is the atomic probability measure supported at @xmath16 , and @xmath618 .
then @xmath789 is a probability measure on @xmath784 . *
finally , for a bounded compactly supported function @xmath790 , let @xmath791 the function @xmath792 is called the _ siegel - veech _ transform of @xmath793 .
we now summarize the relevant results from @xcite , @xcite and @xcite that will be used in [ tribillsect ] .
[ theorem : counting : ems : em ] let @xmath794 be a translation surface , and suppose the following hold ( using notation [ notation : counting ] ) : 1 .
[ ( a ) ] @xmath795 varies linearly under the @xmath0 action ; i.e. , for all @xmath796 and all @xmath797 , we have @xmath798 .
[ ( b ) ] there exists a constant @xmath370 , such that for all @xmath799 , we have @xmath800 .
[ ( c ) ] as @xmath801 , the measures @xmath789 converge to an @xmath0-invariant ( probability ) measure @xmath153 .
[ ( d ) ] let @xmath802 denote the characteristic function of the trapezoid whose vertices are at @xmath803 , @xmath804 , @xmath805 and @xmath806 .
let @xmath807 denote the closure of the @xmath0 orbit of @xmath16 .
then for any @xmath252 and any compact subset @xmath253 of @xmath784 , there exist continuous functions @xmath808 and @xmath809 such that for all @xmath810 , we have @xmath811 then , the following hold : 1 . [
( i ) ] there exists a constant @xmath812 , such that as @xmath813 , @xmath814 2 .
[ ( ii ) ] we have the siegel - veech formula : there exists a constant @xmath815 such that for any continuous compactly supported @xmath816 , @xmath817 3 . [ ( iii ) ] the constant @xmath815 in ( [ ( i ) ] ) is the same as the constant @xmath815 in ( [ ( ii ) ] ) . conclusion ( [ ( ii ) ] ) depends only on assumptions ( [ ( a ) ] ) and ( some version of ) ( [ ( b ) ] ) .
it was proved by w. veech in @xcite , where this approach to counting on translation surfaces was originated .
the proof is reproduced in ( * ? ? ? * theorem 2.2 ) .
assumption ( [ ( b ) ] ) may be replaced by : 1 . [ ( b ) ] there exist constants @xmath818 and @xmath819 such that for all @xmath794 , @xmath820 , where @xmath821 is the length of the shortest saddle connection on @xmath16 .
in fact , ( [ ( b ) ] ) is used in the proof of theorem [ theorem : counting : ems : em ] instead of ( [ ( b ) ] ) .
the assertion that ( [ ( b ) ] ) implies ( [ ( b ) ] ) follows from ( * ? ? ?
* theorem 5.1 ) .
it follows from ( [ ( b ) ] ) and ( * ? ? ?
* theorem 5.2 ) that any limit measure of the probability measures @xmath789 must be a probability measure ( see ( * ? ? ?
* corollary 5.3 ) ) .
thus , the measure @xmath153 of ( [ ( c ) ] ) is automatically a probability measure .
the assertion ( [ ( d ) ] ) is a technical assumption needed since the siegel - veech transform @xmath822 may not be continuous even if @xmath793 is .
let @xmath157 be the characteristic function of the trapezoid as in ( [ ( d ) ] ) .
we have the following lemma from calculus ( cf .
* lemma 3.4 ) ) : for any @xmath113 , @xmath823 if we multiply both sides of ( [ eq : calculus ] ) by @xmath824 and sum over all @xmath825 , we get , under assumption ( [ ( a ) ] ) , @xmath826 or equivalently , @xmath827 ( the fact that we only have approximate equality and not equality in ( [ eq : count : integral ] ) does not affect the asymptotics . see @xcite for the details ) .
the assumption ( [ ( c ) ] ) means that for any bounded continuous function @xmath828 on @xmath784 , @xmath829 we would like to apply ( [ eq : circles : bounded : func ] ) to @xmath830 , which is neither bounded nor continuous . the fact that @xmath830 is not continuous
is handled by assumption ( [ ( d ) ] ) .
to handle the fact that @xmath830 is not bounded , we decompose @xmath831 , where @xmath832 is bounded and @xmath833 is supported outside of a large compact set . then the contribution of @xmath833 can be shown to be negligible using ( * ? ?
* theorem 5.2 ) , in view of assumption ( [ ( b ) ] ) .
the details of this argument are given in @xcite .
now applying ( [ eq : circles : bounded : func ] ) with @xmath834 and substituting into ( [ eq : count : integral ] ) , we get @xmath835 by iterating ( replacing @xmath10 with @xmath836 ) , and summing the resulting geometric series , we get @xmath837 this implies ( [ ( i ) ] ) .
now by ( [ eq : siegel : veech : general ] ) , @xmath838 this , together with ( [ eq : temp : asymp ] ) , implies ( [ ( iii ) ] ) . as a corollary of theorem [ theorem : counting : ems : em ] and theorem [ theorem : largecircles ] we have the following : [ theorem : branched : cover : count ] suppose @xmath16 is a branched cover of a veech surface @xmath27 .
let @xmath41 denote the number of cylinders of periodic trajectories in @xmath16 of length at most @xmath10 .
then there exists a constant @xmath51 such that as @xmath15 , @xmath839 we use theorem [ theorem : counting : ems : em ] , with @xmath840 .
assumption ( [ ( a ) ] ) clearly holds , and ( [ ( b ) ] ) also holds since the boundary of every cylinder contains a saddle connection .
now let @xmath841 be the connected component of @xmath72 that contains @xmath16 ( where @xmath72 is as in the introduction ) .
since @xmath842 and @xmath841 is closed and @xmath0-invariant , the support of any of the measures @xmath789 is contained in @xmath841 . also , since @xmath841 is a branched cover of the space @xmath89 , a measure classification theorem on @xmath89 automatically yields a measure classification theorem on @xmath841 ( see lemma [ fincovclassify ] below ) . thus assumption ( [ ( c ) ] ) of theorem [ theorem :
counting : ems : em ] follows from theorem [ theorem : largecircles ] . finally
, in our setting ( [ ( d ) ] ) is automatically satisfied , since the orbit closure @xmath807 is a proper submanifold of @xmath784 , the measure @xmath153 is lebesgue measure on @xmath807 , and ( after intersecting with any compact set ) the set of discontinuities of @xmath830 is contained in a finite union of submanifolds of positive codimension in @xmath807 . thus theorem [ theorem : branched : cover : count ] follows from ( [ ( i ) ] ) of theorem [ theorem : counting : ems : em ] .
[ fincovclassify ] suppose @xmath172 is a @xmath87-invariant subspace of @xmath173 , @xmath512 is a closed orbit of @xmath239 in @xmath89 , and @xmath807 is a ( connected ) branched cover of @xmath512 , such that the action of @xmath87 on @xmath512 lifts to @xmath807 . if @xmath160 is any @xmath432-invariant probability measure on @xmath807 that projects to the lebesgue measure on @xmath512 , then @xmath160 is the lebesgue measure on @xmath807 .
let @xmath153 and @xmath843 be the lebesgue measures on @xmath512 and @xmath807 , respectively . then , because it projects to @xmath153 , the measure @xmath160 must be absolutely continuous with respect to @xmath843 ; thus , we may write @xmath844 , for some borel function @xmath793 on @xmath807 .
it is not difficult to see that @xmath843 is ergodic for @xmath87 , so ( by decay of matrix coefficients ( * ? ? ?
* thm . 2.4.2 ,
p. 29 ) , or by the the mautner phenomenon ( * ? ?
2.2.15 , p. 21 ) ) it is ergodic for @xmath432 .
this implies that @xmath793 is constant .
so @xmath845 ( up to a normalizing scalar multiple ) .
let @xmath4 be an odd integer . as in the introduction , let @xmath846 and let @xmath83 denote the corresponding translation surface . in the rest of this section , we complete the proof of theorem [ theorem : triangle : count ] by computing the constant @xmath815 in theorem [ theorem : branched : cover : count ] for the case of the surface @xmath83 .
our general strategy is to use ( [ ( ii ) ] ) and ( [ ( iii ) ] ) of theorem [ theorem : counting : ems : em ] . to pass from @xmath83 to @xmath80 ,
note that @xmath847 , and since @xmath83 consists of @xmath848 triangles , @xmath849 .
the surface @xmath83 can be drawn as in figure [ fig : pentagons ] .
as shown in @xcite and as one can see from the figure , @xmath83 is a double cover of a surface @xmath850 consisting of a double @xmath5-gon with opposite sides identified . the surface @xmath850 is a veech surface ( see @xcite ) , but @xmath83 is not ( see @xcite ) .
most of the information in this section comes from @xcite .
let @xmath851 ( realized with the two equal sides having length @xmath500 , and one of the equal sides horizontal ) .
then the surface corresponding to @xmath852 can easily be seen to be ( isomorphic to ) @xmath850 .
the cylinder decomposition in the vertical direction consists of @xmath853 cylinders @xmath854 , and for @xmath855 , we have : @xmath856 ( the closed trajectories in the cylinder @xmath854 have length @xmath857 ) . since for all @xmath855 , @xmath858
, the unipotent @xmath859 belongs to the veech group @xmath860 of @xmath850 . note that @xmath861 the unipotent @xmath862 , together with the rotation by @xmath3 generate @xmath860 .
it is shown in @xcite that @xmath863 where @xmath864 denotes the poincar volume on the hyperbolic plane @xmath865 .
the following lemma is from @xcite : [ lemma : gj ] suppose @xmath866 is a lattice , and suppose that @xmath61 intersects non - trivially the stabilizer @xmath867 in @xmath0 of @xmath740 .
( the above condition is equivalent to the discreteness of the orbit @xmath868 . )
let @xmath869 be either of the two generators of @xmath870 .
let @xmath776 denote the ball in @xmath42 of radius @xmath10 centered at the origin .
then , as @xmath871 , @xmath872 where @xmath873 is any vector perpendicular to @xmath740
. we also record the following trivial consequence ( of the existence of the asymptotics ) : suppose @xmath740 is as in lemma [ lemma : gj ] , and suppose @xmath874 is a scalar multiple of @xmath740 . then , as @xmath15 .
@xmath875 we now apply lemma [ lemma : gj ] with @xmath876 , @xmath877 , @xmath878 , @xmath879 , and using ( [ eq : move : perp ] ) we get that the number of cylinders in the ball of radius @xmath10 that are in the orbit of the cylinder @xmath880 is asymptotic to @xmath881 we now use ( [ eq : trivial ] ) to see that for @xmath855 , the number of cylinders in the ball of radius @xmath10 that are in the @xmath860 orbit of @xmath854 is asymptotic to @xmath882 ( in the above , we used the identity @xmath883 ) . since every cylinder is in the orbit of some @xmath854 , we get ( after summing over @xmath514 ) , @xmath884 using the identity ( * ? ? ?
* lemma 6.3 ) @xmath885 and using the expressions ( [ eq : heights : widths ] ) , we get @xmath886 since @xmath887 we have @xmath888 this is the formula in @xcite .
it is useful for the sequel to compare the result of ( [ eq : count : j : one : orbit ] ) with the result of the corresponding siegel - veech formula .
let @xmath889 denote the orbit of @xmath850 .
this is a closed submanifold of the stratum , which is also called a teichmller curve . for @xmath855 , define @xmath890 by the formula @xmath891 ( where @xmath796 ) .
let @xmath892 denote the characteristic function of the @xmath355-ball in @xmath42 centered at the origin , and for @xmath893 , define the siegel - veech transform @xmath894 .
[ lemma : disjoint : cusps ] if @xmath355 is sufficiently small , then @xmath895 takes on only the values @xmath117 and @xmath500 ; we have @xmath896 if and only if @xmath27 has a cylinder decomposition such that the @xmath529 cylinder from the left has height at most @xmath355 . given @xmath27 ,
such a cylinder decomposition is unique if it exists .
this is straightforward .
( the uniqueness of the decomposition follows from the fact , proved by veech @xcite , that @xmath897 has only one cusp . )
let @xmath160 denote the normalized @xmath0 invariant measure on @xmath898 .
then , we have the siegel - veech formula : @xmath899 we now apply theorem [ theorem : counting : ems : em ] with @xmath900 , and @xmath901 . the validity of assumption ( [ ( c ) ] ) can be deduced from the mixing property of the geodesic flow , see @xcite for a general proof in variable negative curvature , or @xcite for a simplified exposition in the constant curvature case .
we obtain that @xmath902 where @xmath903 is as in ( [ eq : xn : v1:siegel : veech ] ) . comparing with ( [ eq : count : j : one : orbit ] ) , we see that @xmath904 substituting into ( [ eq : xn : v1:siegel : veech ] ) we get @xmath905 it is possible to prove ( [ eq : measure : small : v1 ] ) directly , and thus to compute the asymptotics in ( [ eq : count : one : orbit ] ) without using lemma [ lemma : gj ] .
we chose this indirect derivation of ( [ eq : measure : small : v1 ] ) to minimize the amount of computation .
we now return to our surface @xmath83 , which is a branched cover of @xmath850 ( see figure [ fig : pentagons ] ) .
@xmath850 is a union of two @xmath5-gons , and the two branch points @xmath175 and @xmath906 are at the centers of the @xmath5-gons .
we now wish to apply theorem [ theorem : largecircles ] to the point @xmath907 .
it is important to note that @xmath850 is hyperelliptic , and that our two branch points are interchanged by the hyperelliptic involution .
since the hyperelliptic involution commutes with the @xmath0 action , this it true for any point in the orbit of @xmath908 .
thus , the @xmath0 orbit of @xmath908 is _ not _ dense in the space @xmath909 , and indeed we have in theorem [ theorem : largecircles ] a proper @xmath910 of real dimension @xmath62 .
let @xmath713 denote the subspace @xmath911 .
the above argument shows that @xmath912 .
but since we know that @xmath83 is not veech , @xmath913
. hence @xmath914 and @xmath915 .
let @xmath916 .
then @xmath917 consists of points of the form @xmath918 where @xmath893 , @xmath919 , @xmath920 and @xmath592 and @xmath921 are interchanged by the hyperelliptic involution of @xmath27 . by theorem [ theorem : largecircles ] ,
@xmath922 where @xmath153 is lebesgue measure on @xmath512 .
now let @xmath923 denote the orbit closure @xmath924 . since @xmath83 is a double cover of @xmath850 , branched over @xmath175 and @xmath906 , for any @xmath796
, @xmath925 is a double cover of @xmath926 branched over @xmath927 and @xmath928 , and @xmath929 .
thus , in particular , every surface in @xmath923 is a double cover of a surface in @xmath898 .
thus we have a natural map @xmath930 that maps each surface @xmath931 to the surface in @xmath898 of which it is a double cover , and notes the locations of the branch points . now in view of ( [ eq : marked : points ] ) and lemma [ fincovclassify ] , @xmath932 where @xmath933 is normalized lebesgue measure on @xmath923 .
hence , by theorem [ theorem : counting : ems : em ] , we have a quadratic asymptotic formula @xmath934 with the constant @xmath815 given by @xmath935 where as above , @xmath936 is the characteristic function of the ball of radius @xmath355 centered at the origin , and @xmath937 .
let @xmath740 be some periodic direction for @xmath83 , hence for @xmath850 .
we may use an element @xmath518 of the veech group @xmath860 of @xmath850 to map @xmath740 to the vertical direction . note that @xmath938 is a double cover of @xmath939 . in the vertical direction
, @xmath850 has the cylinder decomposition @xmath940 described above .
[ lemma : cyl : decomp ] for any @xmath941 , the branch points of @xmath938 will project to two points in the same cylinder , say @xmath942 .
the cylinder decomposition of @xmath938 in the vertical direction is the following : 1 .
[ lemma : cyl : decomp - nobranch ] for each @xmath943 , there are two cylinders on @xmath938 of the same length as @xmath854 ( one on each `` sheet '' ) .
[ lemma : cyl : decomp - branch ] on @xmath938 there are two cylinders of the same length as @xmath942 and two cylinders of twice the length of @xmath942 . the fact that both branch points project to the same cylinder of @xmath850 follows from the fact that each cylinder of @xmath850 is preserved by the hyperelliptic involution @xmath944 of @xmath850 ( since different cylinders have different lengths ) and the fact that the branch points are interchanged by @xmath944 . from figure
[ fig : pentagons ] , the cover @xmath83 is determined by two slits ( drawn as the thick lines in the figure ) , which are interchanged by @xmath944 . since @xmath944 commutes with the @xmath0 action , the cover @xmath938 of @xmath850 is also determined by two slits , which are interchanged by @xmath944 . for each cylinder @xmath854 of @xmath850 , let @xmath945 denote the closed trajectory in the center of @xmath854 .
note that for any @xmath514 , @xmath945 is mapped to itself under @xmath944 .
also , since @xmath944 exchanges the slits , we see that @xmath945 intersects each slit the same number of times .
thus , @xmath945 breaks up into two closed paths of the same length when lifted from @xmath850 to @xmath938 .
this proves ( [ lemma : cyl : decomp - nobranch ] ) and the first assertion of ( [ lemma : cyl : decomp - branch ] ) .
it is easy to see that the closed vertical trajectories on @xmath942 between the boundary of @xmath942 and one of the branch points double in length when lifted from @xmath850 to @xmath938 .
this proves the second assertion of ( [ lemma : cyl : decomp - branch ] ) .
[ cor : descent ] the function @xmath946 is constant on the fibers of @xmath947 ( a.e . ) , and thus descends to a function @xmath948 . the latter function , for @xmath355 sufficiently small , is given ( a.e . ) by the formula @xmath949 where for @xmath855 , @xmath950 is as in lemma [ lemma : disjoint : cusps ] , and @xmath96 is such that @xmath592 ( and @xmath921 ) belong to the @xmath951 cylinder from the left in the unique cylinder decomposition that contains a cylinder of height at most @xmath355 .
choose a fundamental domain for @xmath860 in the upper half plane .
since @xmath850 has a vertical cylinder decomposition , we may assume that the cusp of the fundamental domain approaches @xmath952 , rather than approaching a point on the real axis .
this means that as @xmath157 goes to infinity in the fundamental domain , the unique short cylinder decomposition of @xmath953 is the image under @xmath157 of the vertical cylinder decomposition of @xmath850 .
we first prove that ( [ cor : descent : eqn ] ) is correct for all @xmath27 in the @xmath0 orbit of @xmath83 . to do this , let @xmath796 , and write @xmath954 ,
where @xmath941 , and @xmath157 is in the fundamental domain .
note that @xmath955 is zero unless @xmath27 has a short cylinder .
thus , if @xmath954 , and @xmath157 is in a compact part of the fundamental domain , then ( in view of lemma [ lemma : cyl : decomp ] ) , we have @xmath956 . therefore , we may assume that @xmath157 is in the cusp , and hence the unique short cylinder decomposition of @xmath957 is the image , under the linear action of @xmath157 , of the vertical cylinder decomposition of @xmath938 .
then it is clear from lemma [ lemma : cyl : decomp ] that ( [ cor : descent : eqn ] ) holds for @xmath958 . to complete the proof , note that both sides of ( [ cor : descent : eqn ] ) are continuous off a closed set of measure @xmath117 , namely , the set where a branch point projects to an edge of a cylinder ( in a cylinder decomposition containing a cylinder of height at most @xmath355 ) .
then use the fact that the @xmath0 orbit of @xmath83 is dense .
the analogue of the first assertion of corollary [ cor : descent ] fails in the context of @xcite , in part since there we are dealing with covers of high degree .
this is responsible for most of the combinatorial complexity of the argument in @xcite . in view of corollary [ cor :
descent ] , ( [ eq : c : tilde : o ] ) becomes @xmath959 where @xmath960 is lebesgue measure on the translation surface @xmath27 , and @xmath944 denotes the hyperelliptic involution . performing the integral over @xmath27 , we get @xmath961 where @xmath962 and @xmath963 ( so that @xmath964 denotes the relative area of the @xmath951 cylinder from the left in any cylinder decomposition ) .
now , using ( [ eq : measure : small : v1 ] ) , we get @xmath965 since the second term in the parenthesis is independent of @xmath96 , and @xmath966 , this can be rewritten as @xmath967 the first term in the parenthesis is in view of ( [ eq : count : veech : prelim ] ) twice the limit of @xmath968 .
the second term in the parenthesis is , since @xmath969 , equal to @xmath970 in view of ( [ eq : count : veech : final ] ) , we get @xmath971 simplifying , we get @xmath972 alternatively , @xmath973 to pass from @xmath83 to @xmath80 , note that @xmath847 , and since @xmath83 consists of @xmath848 triangles , @xmath849 .
this research was partially supported by grants from the national science foundation ( dms-0244542 and dms-0100438 ) , the packard foundation , and an epsrc advanced research fellowship .
the authors are grateful to the swiss federal technical institute ( eth ) of zurich , the university of chicago , the newton institute ( cambridge , england ) , and the american institute of mathematics for their hospitality , and to kariane calta , franois ledrappier , howard masur , martin schmoll , and anonymous referees for helpful comments on a preliminary version of the manuscript .
a. eskin , h. masur , and a. zorich .
moduli spaces of abelian differentials : the principal boundary , counting problems , and the siegel - veech constants .
hautes tudes sci .
* 97 * ( 2003 ) , 61179 .
h. masur .
lower bounds for the number of saddle connections and closed trajectories of a quadratic differential . in _ holomorphic functions and moduli _ ,
, d. drasin , ed . , springer - verlag : new york , 1988 , pp . | there is a natural action of @xmath0 on the moduli space of translation surfaces , and this yields an action of the unipotent subgroup @xmath1 .
we classify the @xmath2-invariant ergodic measures on certain special submanifolds of the moduli space .
( each submanifold is the @xmath0-orbit of the set of branched covers of a fixed veech surface . ) for the @xmath2-action on these submanifolds , this is an analogue of ratner s theorem on unipotent flows .
the result yields an asymptotic estimate of the number of periodic trajectories for billiards in a certain family of non - veech rational triangles , namely , the isosceles triangles in which exactly one angle is @xmath3 , with @xmath4 and @xmath5 odd . |
studies of cp violation in the @xmath12-meson system and the determination of the three angles @xmath13 , @xmath8 and @xmath5 of the usual non - squashed unitarity triangle @xcite of the cabibbo kobayashi maskawa matrix ( ckm matrix ) @xcite are among the central targets of future @xmath12-physics experiments . during the recent years
, several strategies were proposed to accomplish this task @xcite . in this context , also quasi - two - body modes @xmath14 of neutral @xmath15-mesons ( @xmath16 ) , where both @xmath17 and @xmath18 carry spin and continue to decay through cp - conserving interactions , are of particular interest @xcite . in this case
, the time - dependent angular distribution of the decay products of @xmath17 and @xmath18 provides valuable information . for an initially , i.e. at time @xmath19 , present @xmath20-meson ,
it can be written as @xmath21 where we have denoted the angles describing the kinematics of the decay products of @xmath17 and @xmath18 generically by @xmath22 , @xmath23 and @xmath24 .
note that we have to deal , in general , with an arbitrary number of such angles .
the observables @xmath25 describing the time evolution of the angular distribution ( [ ang ] ) can be expressed in terms of real or imaginary parts of certain bilinear combinations of decay amplitudes . in the applications discussed in this paper , we will focus on @xmath26_f$ ]
decays , where @xmath17 and @xmath18 are both vector mesons , and @xmath27 denotes a final - state configuration with cp eigenvalue @xmath28 .
it is convenient to analyse such modes in terms of the linear polarization amplitudes @xmath29 , @xmath30 and @xmath31 @xcite .
whereas @xmath31 describes a cp - odd final - state configuration , both @xmath29 and @xmath30 correspond to cp - even final - state configurations , i.e. to the cp eigenvalues @xmath32 and @xmath33 , respectively .
the observables of the corresponding angular distribution are given by @xmath34 as well as by the interference terms @xmath35 this formalism is discussed in more detail in @xcite , where several explicit angular distributions can be found and appropriate weighting functions to extract their observables in an efficient way from the experimental data are given . in the following considerations ,
the main role is played by neutral @xmath26_f$ ] decays , where the `` unevolved '' decay amplitudes can be expressed as @xmath36\label{ampl}\\ \overline{a}_f&=&\eta_f\,{\cal n}_f\left[1-b_f\ , e^{i\rho_f } e^{-i\omega}\right],\label{ampl - cp}\end{aligned}\ ] ] where @xmath37 denotes a cp - violating weak phase and @xmath38 are cp - conserving strong phases . in this case , the observables ( [ obs1 ] ) and ( [ obs2 ] ) allow us to probe the @xmath39@xmath40 mixing phase @xmath41 and the weak phase @xmath37 , as we will show in this paper . concerning practical applications ,
@xmath37 is given by one of the angles of the unitarity triangle .
however , the observables specified in ( [ obs1 ] ) and ( [ obs2 ] ) are not independent from one another and do not provide sufficient information to extract @xmath41 and @xmath37 , as well as the corresponding hadronic parameters , simultaneously . to this end
, we have to use an additional input . usually , the weak phase @xmath37 is of central interest .
if we fix the mixing phase @xmath41 separately , it is possible to determine @xmath37 and interesting hadronic quantities as a function of a _ single _ hadronic parameter in a _ theoretically clean _ way . if we determine this quantity , for instance , by comparing @xmath14 with an @xmath9-related mode , all remaining parameters , including @xmath37 , can be extracted .
if we are willing to make more extensive use of flavour - symmetry arguments , it is in principle possible to determine the @xmath39@xmath40 mixing phase @xmath41 as well .
an example for such a strategy is given by the decay @xmath1 , which can be combined with @xmath2 to extract the @xmath3@xmath4 mixing phase @xmath42 and if penguin effects in the former mode should be sizeable also the angle @xmath5 of the unitarity triangle .
as an interesting by - product , this strategy allows us to take into account also the penguin effects in the extraction of the @xmath6@xmath7 mixing phase from @xmath2 , which is an important issue for `` second - generation '' @xmath12-physics experiments at hadron machines .
moreover , we may resolve a discrete ambiguity in the extraction of the ckm angle @xmath8 , and may obtain valuable insights into @xmath9-breaking effects .
other interesting applications of the general formalism presented in this paper , involving @xmath10 and @xmath43 decays , are also briefly noted .
as the extraction of @xmath37 with the help of these modes involves `` penguin '' , i.e. flavour - changing neutral - current ( fcnc ) processes and relies moreover on the unitarity of the ckm matrix , it may well be affected by new physics .
in such a case , discrepancies would show up with other strategies to determine this phase , for example with the theoretically clean extractions of @xmath5 making use of pure `` tree '' decays such as @xmath44 . since no fcnc processes contribute to the decay amplitudes of these modes , it is quite unlikely that they and the extracted value of @xmath5 are significantly affected by new physics . the outline of this paper is as follows : in section [ evol ] , the time - dependent observables of the @xmath14 angular distribution are given .
the strategies to extract ckm phases , as well as interesting hadronic parameters , with the help of these observables are discussed in section [ strat ] . in section [ bpsi ] ,
we focus on the extraction of @xmath8 and @xmath5 from @xmath1 and @xmath2 .
further applications of the formalism developed in sections [ evol ] and [ strat ] are discussed in section [ appl ] , and the conclusions are summarized in section [ concl ] .
in this section , we consider the general case of a neutral quasi - two - body decay @xmath45_f$ ] into a final - state configuration @xmath27 with cp eigenvalue @xmath28 that exhibits `` unmixed '' decay amplitudes of the same structure as those given in ( [ ampl ] ) and ( [ ampl - cp ] ) .
if we use linear polarization states to characterize the final - state configurations as , for example , in @xcite , we have @xmath46 . at this point a comment on the angular distribution of the cp - conjugate decay @xmath47 , which is given by @xmath48 is in order .
since the meson content of the @xmath49 states is the same whether these result from the @xmath20 or @xmath50 decays , we may use the same generic angles @xmath22 , @xmath23 and @xmath24 to describe the angular distribution of their decay products . within this formalism
, the effects of cp transformations relating @xmath51_f$ ] to @xmath52_f$ ] are taken into account by the cp eigenvalue @xmath28 appearing in ( [ ampl - cp ] ) , and do not affect the form of @xmath53 .
therefore the same functions @xmath53 are present in ( [ ang ] ) and ( [ ang - cp ] ) ( see also @xcite ) . in view of applications to @xmath54
decays , we allow for a non - vanishing width difference @xmath55 between the @xmath15 mass eigenstates @xmath56 ( `` heavy '' ) and @xmath57 ( `` light '' ) .
in contrast to the @xmath58 case , this width difference may be sizeable in the @xmath54 system @xcite ; it may allow studies of cp violation with `` untagged '' @xmath54 data samples , where one does not distinguish between initially , i.e. at time @xmath19 , present @xmath59 or @xmath7 mesons @xcite .
the time evolution of the observables corresponding to ( [ obs1 ] ) takes the following form : @xmath60\ ] ] @xmath61,\ ] ] where @xmath62 denotes the mass difference between the @xmath15 mass eigenstates , and @xmath63/2 $ ] .
the quantities @xmath64 , @xmath65 , @xmath66 and @xmath67 , which are not independent from one another and satisfy the relation @xmath68 are given by @xmath69\\ r_{\rm h}^f&=&|{\cal n}_f|^2\bigl[(1-\eta_f\cos\phi_q)\nonumber\\ & & -2\,b_f\cos\rho_f\left\{\cos\omega-\eta_f\cos(\phi_q+\omega)\right\}+ b_f^2\left\{1-\eta_f\cos(\phi_q+2\,\omega)\right\}\bigr]\\ a_{\rm d}^f&\equiv&=2\,|{\cal n}_f|^2\,b_f\sin\rho_f\,\sin\omega \label{ad - expr}\\ a_{\rm m}^f&=&\eta_f\,|{\cal n}_f|^2 \left[\sin\phi_q-2\,b_f\cos\rho_f\,\sin(\phi_q+\omega)+ b_f^2\sin(\phi_q+2\,\omega)\right].\label{am - expr}\end{aligned}\ ] ] here the phase @xmath41 denotes the cp - violating weak @xmath39@xmath40 mixing phase : @xmath70 where @xmath71 is tiny in the standard model because of a cabibbo suppression of @xmath72 .
this phase cancels in @xmath73 it is also interesting to note that there are no @xmath74 terms present in the `` untagged '' combination @xmath75 whereas @xmath76.\ ] ] because of ( [ obs - rel ] ) , each of the @xmath77 or @xmath78 ( @xmath46 ) terms of the @xmath14 angular distribution provides three independent observables , which we may choose as @xmath66 , @xmath67 and @xmath79 .
the time evolution of the interference terms ( [ obs2 ] ) is analogous to ( [ sum ] ) and ( [ diff ] ) .
let us first give the expressions for the observables corresponding to ( [ rn - def ] ) : @xmath80= & & \mbox{}\quad\,-\left\{b_0\cos(\rho_0-\delta_{0,\parallel})+b_\parallel \cos(\rho_\parallel+\delta_{0,\parallel})\right\}\cos\omega+ b_0\,b_\parallel\cos(\rho_0-\rho_\parallel-\delta_{0,\parallel } ) \bigr]\label{re - sum}\\ \lefteqn{i^f_{\rm d}\equiv\frac{1}{2}\left[\im\{a_f^\ast(0)a_\perp(0)\}+ \im\{\overline{a}_f^\ast(0)\overline{a}_\perp(0)\}\right]}\nonumber\\ & & \mbox{}\quad\,=|{\cal n}_f||{\cal n}_\perp|\bigl[\,b_f\cos(\rho_f- \delta_{f,\perp})-b_\perp\cos(\rho_\perp+\delta_{f,\perp})\bigr ] \sin\omega,\label{im - sum}\end{aligned}\ ] ] where the @xmath81 denote the differences of the cp - conserving strong phases of the amplitudes @xmath82 and @xmath83 . on the other hand
, the rate differences corresponding to ( [ diff ] ) take the following form : @xmath84\qquad\mbox{}\label{re - diff}\\ \im\{a_f^\ast(t)a_\perp(t)\}-\im\{\overline{a}_f^\ast(t ) \overline{a}_\perp(t)\}&=&2\,e^{-\gamma_q
t}\left[i_f \cos(\delta m_qt)-i_{\rm m}^f \sin(\delta m_qt)\right ] , \quad\mbox{}\label{im - diff}\end{aligned}\ ] ] where @xmath85\sin\omega\label{rd - expr}\\ r_{\rm m}&=&|{\cal n}_0||{\cal n}_\parallel|\bigl [ \cos\delta_{0,\parallel}\sin\phi_q-\left\{b_0 \cos(\rho_0-\delta_{0,\parallel})+b_\parallel\cos(\rho_\parallel+ \delta_{0,\parallel})\right\}\sin(\phi_q+\omega)\nonumber\\ & & + \,\,b_0\,b_\parallel\cos(\rho_0-\rho_\parallel-\delta_{0,\parallel } ) \sin(\phi_q+2\,\omega)\bigr]\label{rm - expr}\end{aligned}\ ] ] and @xmath86\label{i - expr}\\ i_{\rm m}^f&=&|{\cal n}_f||{\cal n}_\perp|\bigl [ \cos\delta_{f,\perp}\cos\phi_q-\left\{b_f \cos(\rho_f-\delta_{f,\perp})+b_\perp\cos(\rho_\perp+\delta_{f,\perp } ) \right\}\cos(\phi_q+\omega)\nonumber\\ & & + \,\,b_f\,b_\perp\cos(\rho_f-\rho_\perp-\delta_{f,\perp } ) \cos(\phi_q+2\,\omega)\bigr].\end{aligned}\ ] ] note that @xmath87 in ( [ im - sum ] ) and ( [ im - diff ] ) .
the minus sign in the latter expression is due to the different cp eigenvalues of @xmath87 and @xmath88 .
if we set `` @xmath89 '' in ( [ rd - expr ] ) and ( [ rm - expr ] ) , we get expressions taking the same form as ( [ ad - expr ] ) and ( [ am - expr ] ) , which provides a nice cross check .
the expressions given above generalize those derived in @xcite in two respects : they take into account penguin contributions , and they allow for a sizeable value of the @xmath39@xmath40 mixing phase @xmath41 . in the discussion of @xmath2 in @xcite , it was assumed that @xmath90 is a small phase , and terms of @xmath91 were neglected .
= 4.3truecm unfortunately , not all of the observables @xmath79 , @xmath66 and @xmath67 are independent from those of the interference terms ( [ obs2 ] )
. this can be seen by considering two different final - state configurations @xmath27 and @xmath92 . in this case , the time - dependent angular distribution provides nine observables . to be definite ,
let us consider the case @xmath93 and @xmath94 .
then we have six observables , corresponding to @xmath79 , @xmath66 and @xmath67 ( @xmath87 ) , as well as the three observables @xmath95 , @xmath96 and @xmath97 , which are due to the real parts in ( [ obs2 ] ) .
the measurement of @xmath79 and @xmath66 allows us to fix the magnitudes @xmath98 , @xmath99 and @xmath100 , @xmath101 . using in addition the observables @xmath95 and @xmath96
, we can determine the angle @xmath102 between the unmixed amplitudes @xmath103 and @xmath104 , as well as the angle @xmath105 between @xmath106 , @xmath107 ( see fig .
[ fig : ampls ] ) .
so far , the relative orientation of the amplitudes @xmath108 , @xmath109 and @xmath110 , @xmath111 is not determined . however
, if we use , in addition , the mixing - induced cp asymmetry @xmath112 , we are in a position to fix @xmath113 , where @xmath114 denotes the angle between the amplitudes @xmath103 and @xmath106 : @xmath115 since the relative orientation of the amplitudes @xmath116 and @xmath104 is also fixed this way , we can predict the values of the two remaining mixing - induced cp - violating observables @xmath117 and @xmath97 .
consequently , only seven of the nine observables are independent from one another .
it is convenient to introduce the following `` normalized '' observables : @xmath118 @xmath119 @xmath120 which have the advantage that they do not depend on the overall normalization factors @xmath121 .
the observables @xmath122 and @xmath123 allow us to determine the hadronic parameters @xmath124 and @xmath125 as functions of @xmath37 and @xmath41 : @xmath126}\ ] ] @xmath127 @xmath128\left(\frac{\hat a_{\rm d}^f}{\sin\omega}\right),\ ] ] where @xmath129 with @xmath130 and @xmath131 it should be emphasized that no approximations were made in order to derive these expressions .
if we consider , in addition to @xmath122 and @xmath123 , either the observables specified in ( [ r - hat ] ) or those given in ( [ i - hat ] ) , we obtain seven normalized observables , which depend on five hadronic parameters ( @xmath124 , @xmath125 , @xmath132 , @xmath133 and @xmath134 ) , as well as on the two cp - violating weak phases @xmath41 and @xmath37 . however , only five of the seven observables are independent from one another , so that we have not sufficient observables at our disposal to extract these parameters simultaneously .
to accomplish this task , we have to make use , for example , of another decay that can be related to @xmath14 through flavour - symmetry arguments . on the other hand , if we use @xmath41 and @xmath37 as an input , all hadronic parameters describing the decay @xmath14 can be extracted _ without _ any additional assumption , thereby providing valuable insights into hadronic physics and a very fertile testing ground for model calculations of @xmath14 .
the measurement of the angular distributions discussed in this paper requires high statistics and can probably only be performed at `` second - generation '' @xmath12-physics experiments at hadron machines , such as lhcb or btev , where also decays of @xmath54-mesons can be studied . since several promising strategies to extract the weak phases @xmath41 and @xmath37 at such experiments were already proposed ( see , for example , @xcite ) , it may indeed be an interesting alternative to use measurements of angular distributions not to extract ckm phases , but to explore hadronic physics . in practical applications , the parameters @xmath124 typically measure the ratio of `` penguin '' to `` tree '' contributions .
applying the bander silverman
soni mechanism @xcite , and following the formalism developed in @xcite , which makes use among other things of the `` factorization hypothesis '' , we obtain for various classes of @xmath12 decays @xmath135 the main reason for these relations is that the form factors , which depend on the final - state configuration @xmath27 , cancel in the ratios @xmath124 of `` penguin '' to `` tree '' contributions .
although non - factorizable contributions are expected to play an important role , thereby affecting ( [ bss - rel ] ) , it is interesting to investigate the implications of these relations on the observables of the angular distributions in more detail
. if we introduce @xmath136 we obtain @xmath137 @xmath138 @xmath139 @xmath140,\ ] ] where @xmath141 these relations provide an interesting test of whether ( [ bss - rel ] ) is realized in the decay @xmath14 .
note that @xmath142 does not
in contrast to ( [ i - bss ] ) vanish for trivial values of @xmath143 .
let us now focus on the extraction of ckm phases from the observables of the @xmath14 angular distribution .
as we have already noted , to this end , we have to employ an additional input , since we have only five independent normalized observables at our disposal , which depend on seven `` unknowns '' .
although it would be desirable to determine @xmath41 and @xmath37 simultaneously , usually only the ckm phase @xmath37 is of central interest .
the @xmath6@xmath7 mixing phase @xmath144 is negligibly small in the standard model . it can be probed and in principle even determined with the help of the decay @xmath2 ( see , for example , @xcite ) .
large cp - violating effects in this decay would signal that @xmath145 is not tiny , and would be a strong indication for new - physics contributions to @xmath6@xmath7 mixing . on the other hand , the @xmath3@xmath4 mixing phase @xmath42 can be fixed in a reliable way through the `` gold - plated '' mode @xmath146 @xcite . strictly speaking ,
mixing - induced cp violation in @xmath147 probes @xmath148 , where @xmath149 is related to the weak @xmath150@xmath151 mixing phase and is negligibly small in the standard model . because of the small value of the cp - violating parameter @xmath152 of the neutral kaon system , @xmath149 can only be affected by very contrived models of new physics @xcite .
a measurement of mixing - induced cp violation in @xmath153 allows us to fix @xmath42 only up to a twofold ambiguity .
several strategies to resolve this ambiguity were proposed in the literature @xcite , which should be feasible for `` second - generation '' @xmath12-physics experiments . as we will see in the following section , also the decay @xmath1 , in combination with @xmath2 , allows us to accomplish this task .
if we use @xmath41 thus determined as an input and consider , in addition to @xmath122 and @xmath123 , either the observables specified in ( [ r - hat ] ) or those given in ( [ i - hat ] ) , we can determine @xmath37 as a function of a _ single _ hadronic parameter .
let us , for the moment , focus on the latter case , i.e. on the observables @xmath122 , @xmath123 , @xmath154 , @xmath155 and @xmath156 , @xmath157 , @xmath142 for a given final - state configuration @xmath87 . since @xmath121 and @xmath158 cancel in these quantities , they depend only on the hadronic parameters @xmath124 , @xmath125 , @xmath159 , @xmath160 , @xmath143 , as well as on the weak phases @xmath37 and @xmath41 . consequently , we have seven observables at our disposal , which depend on seven `` unknowns '' .
however , only five of the seven observables are independent from one another , as we have discussed in the previous section .
if we use @xmath41 as an input , we can , for instance , obtain @xmath37 and @xmath124 , @xmath125 , @xmath159 , @xmath160 as functions of the strong phase difference @xmath143 in a _ theoretically clean _ way .
although the following discussion deals with @xmath143 , we can also replace this quantity by another hadronic parameter of our choice . if we fix @xmath143 , for example , by comparing @xmath14 with an @xmath9-related mode , all parameters can be extracted . using in addition the observables @xmath79
, we can also determine the normalization factors @xmath161 . comparing them with those of the @xmath9-related mode used to fix @xmath143
, we can obtain valuable insights into @xmath9-breaking corrections .
the observables that we have not used so far can be used to resolve discrete ambiguities , arising typically in the extraction of these parameters .
let us now give the formulae to implement this approach in a mathematical way .
the general expression for the observable @xmath156 ( see ( [ i - expr ] ) and ( [ i - hat ] ) ) leads to the equation @xmath162 where @xmath163\\ b_f&=&\frac{1}{n_f}\bigl[\left(b_f\sin\rho_f - b_\perp\sin\rho_\perp \right)\cos\omega - b_fb_\perp\left(\sin\rho_f\cos\rho_\perp- \cos\rho_f\sin\rho_\perp\right)\bigr]\\ c_f&=&\hat i_f,\end{aligned}\ ] ] with @xmath164 the solution of ( [ del - eq ] ) is straightforward , and is given as follows : @xmath165 if we insert @xmath124 and @xmath125 , determined as functions of @xmath37 and @xmath41 with the help of ( [ bdet])([bsin ] ) , into the expressions given above , we can for a given value of @xmath41 determine @xmath143 as a function of @xmath37 .
it should be emphasized that the relation between @xmath143 , @xmath37 and @xmath41 obtained this way is valid _ exactly_. using @xmath157 or @xmath142 instead of @xmath156 would lead to the same relation , since these observables are not independent from @xmath156 .
alternatively , we may use the observables ( [ r - hat ] ) instead of ( [ i - hat ] ) . the general expression for @xmath166 ( see ( [ re - sum ] ) and ( [ r - hat ] ) ) implies an equation similar to ( [ del - eq ] ) , where @xmath167 , @xmath168 and @xmath169 have to be replaced through @xmath170\\ b&=&\frac{1}{n}\bigl[1-\left(b_0\cos\rho_0+b_\parallel\cos\rho_\parallel \right)\cos\omega+b_0b_\parallel\left(\cos\rho_0\cos\rho_\parallel+ \sin\rho_0\sin\rho_\parallel\right)\bigr]\\ c&=&\hat r,\end{aligned}\ ] ] where @xmath171 obviously , the most efficient strategy of combining the observables provided by the @xmath14 angular distribution depends on their actually measured values .
if we are willing to make more extensive use of flavour - symmetry arguments than just to fix the strong phase difference @xmath134 , it is in principle possible to determine also the @xmath39@xmath40 mixing phase @xmath41 . in the following section
, we will have a closer look at the decay @xmath1 , which can be related to @xmath2 through @xmath9 arguments and a certain dynamical assumption concerning `` exchange '' and `` penguin annihilation '' topologies . however , before we turn to these modes , which allow the simultaneous extraction of @xmath42 and @xmath5 , let us first give two useful expressions for the observables @xmath166 and @xmath156 .
since the parameters @xmath124 measure typically the importance of `` penguin '' topologies in comparison with current current contributions , they may not be too large .
if we eliminate the hadronic parameters @xmath124 and @xmath125 in @xmath166 and @xmath156 with the help of the observables @xmath122 and @xmath123 and keep only the leading - order terms in @xmath124 , we obtain @xmath172 allowing us to determine @xmath37 if the strong phase differences @xmath173 or @xmath143 are known .
interestingly , the leading - order expressions ( [ r - approx ] ) and ( [ id - approx ] ) do not depend on the @xmath39@xmath40 mixing phase @xmath41 .
a possible disadvantage of @xmath166 is that @xmath37 enters in combination with @xmath174 . since @xmath173 is a difference of cp - conserving strong phases ( see ( [ delta - def ] ) )
, it may be small , thereby weakening the sensitivity of these observables on @xmath37 .
the situation concerning this point is very different in the case of the observables @xmath156 , which allow us to determine @xmath37 even in the case of @xmath175 .
if we combine the observables describing the time - dependent angular distribution of the decay @xmath176\ , \rho^0[\to\pi^+\pi^-]$ ] with those of @xmath177\ , \phi[\to k^+k^-]$ ] , we may extract the @xmath3@xmath4 mixing phase @xmath42 and the angle @xmath5 of the unitarity triangle .
the @xmath1 angular distribution can be obtained straightforwardly from the @xmath2 case , which has been discussed in detail in @xcite , by performing appropriate replacements of kinematical variables .
the decay @xmath178 originates from @xmath179 quark - level transitions ; the structure of its decay amplitude is completely analogous to the one of @xmath180 ( see @xcite ) . for a given final - state configuration @xmath27 with cp eigenvalue @xmath28
, we have @xmath181_f)=\lambda_c^{(d)}\left[a_{\rm cc}^{(c)f}+ a_{\rm pen}^{(c)f}\right]+\lambda_u^{(d)}a_{\rm pen}^{(u)f } + \lambda_t^{(d)}a_{\rm pen}^{(t)f}\,,\ ] ] where @xmath182 is due to current current contributions , and the amplitudes @xmath183 describe penguin topologies with internal @xmath184 quarks ( @xmath185 .
these penguin amplitudes take into account both qcd and electroweak penguin contributions , and @xmath186 are the usual ckm factors . employing the unitarity of the ckm matrix and the wolfenstein parametrization @xcite , generalized to include non - leading terms in @xmath187 @xcite
, we obtain @xmath188_f)=-\lambda\,{\cal a}_f\left[1-a_f\,e^{i\theta_f } e^{i\gamma}\right],\ ] ] where @xmath189,\ ] ] with @xmath190 , and @xmath191.\ ] ] the quantity @xmath192 is defined in analogy to @xmath193 , and the relevant ckm factors are given as follows : @xmath194 it should be emphasized that ( [ bd - ampl ] ) is a completely general parametrization of the @xmath178 decay amplitude within the standard model , relying only on the unitarity of the ckm matrix .
in particular , this expression takes also into account final - state - interaction effects , which can be considered as long - distance penguin topologies with internal up- and charm - quark exchanges @xcite . comparing ( [ bd - ampl ] ) with ( [ ampl ] )
, we observe that @xmath195 let us now turn to @xmath196 .
using the same notation as in ( [ bd - ampl ] ) , we have @xmath197_f)=\left(1-\frac{\lambda^2}{2}\right){\cal a}_f ' \left[1+\epsilon\,a_f'\,e^{i\theta_f ' } e^{i\gamma}\right],\ ] ] where @xmath198 and @xmath199 take the same form as ( [ aa - def ] ) and ( [ a - def ] ) , respectively , and @xmath200 the primes remind us that we are dealing with a @xmath201 transition .
consequently , if we compare ( [ bd - ampl2 ] ) with ( [ ampl ] ) , we obtain @xmath202 the @xmath2 and @xmath1 observables can be related to each other through @xmath203 @xmath204 where the factor of @xmath205 is due to the @xmath206 wave function .
these relations rely both on the @xmath9 flavour symmetry of strong interactions and on the neglect of certain `` exchange '' and `` penguin annihilation '' topologies .
although such topologies , which can be probed , for example , through @xmath207 , @xmath208 decays , are usually expected to play a very minor role , they may in principle be enhanced through final - state - interaction effects @xcite . for the following considerations , it is useful to introduce the quantities @xmath209 which can be fixed through the `` untagged '' @xmath1 and @xmath2 observables with the help of ( [ rel-1 ] ) .
consequently , each of the linear polarization states @xmath46 provides the following three observables : @xmath210 applying ( [ rel-2 ] ) to ( [ h - def ] ) , these observables depend only on the hadronic parameters @xmath211 and @xmath212 , as well as on the @xmath3@xmath4 mixing phase @xmath42 and the angle @xmath5 of the unitarity triangle .
if we choose two different linear polarization states , the observables ( [ u - obs ] ) allow us to determine the corresponding hadronic parameters and @xmath8 and @xmath5 simultaneously .
this approach can be implemented in a mathematical way as follows : if we consider a given final - state configuration @xmath27 and combine the observables @xmath213 and @xmath122 , which do not depend on @xmath214 , with each other , we can determine @xmath211 and @xmath212 as functions of @xmath5 : @xmath215 @xmath216 @xmath217\left(\frac{\hat a_{\rm d}^f}{\sin\gamma}\right),\ ] ] where @xmath218\sin^2\gamma-\epsilon_f e_f}{\left(1- \epsilon^2h_f\right)^2\sin^2\gamma+\epsilon^2 e_f}\ ] ] @xmath219 with @xmath220 ^ 2.\ ] ] these expressions allow us to eliminate the hadronic parameters @xmath211 and @xmath212 in the mixing - induced cp asymmetry @xmath123 , thereby fixing a contour in the @xmath5@xmath214 plane , which is related to @xmath221 with @xmath222 the solution of ( [ phi - eq ] ) has already been given in ( [ sol ] ) . if we consider two different final - state configuratons @xmath27 and @xmath92 , we obtain two different contours in the @xmath5@xmath214 plane ; their intersection allows us to determine both @xmath5 and @xmath42 . using , in addition , the observables ( [ r - hat ] ) or ( [ i - hat ] ) depending on which final - state configurations @xmath27 and @xmath92 we consider we may resolve discrete ambiguities , arising typically in the extraction of @xmath214 and @xmath5 . because of the strong suppression of @xmath223 through @xmath224 in ( [ h - def ] ) , this approach is essentially unaffected by possible corrections to ( [ rel-2 ] ) , and relies predominantly on the relation ( [ rel-1 ] ) .
if we insert the values of @xmath214 and @xmath5 thus determined into the expressions for the observables of the third linear polarization state @xmath225 , which has not been used so far , its hadronic parameters @xmath226 , @xmath227 and @xmath228 can also be determined . comparing @xmath226 with the @xmath2 parameter @xmath229
, we can obtain valuable insights into the validity of ( [ rel-1 ] ) .
moreover , several other interesting cross checks can be performed with the many observables of the angular distributions .
because of our poor understanding of the hadronization dynamics of non - leptonic @xmath12 decays , only the `` factorization '' approximation can be used for the time being to estimate factorizable @xmath9-breaking corrections to ( [ rel-1 ] ) .
explicit expressions for the @xmath2 observables can be found in @xcite , and @xmath9-breaking effects in the corresponding form factors were studied in @xcite . however , also non - factorizable effects are expected to play an important role , and experimental insights into these issues would be very helpful to find a better theoretical description .
the simultaneous extraction of @xmath214 and @xmath5 discussed above works only if the hadronic parameters @xmath211 and @xmath212 are sufficiently different from each other for two different final - state configurations @xmath27 . if , for example , ( [ bss - rel ] ) should apply to @xmath1 which seems to be quite unlikely
the @xmath3@xmath4 mixing phase has to be fixed separately in order to determine @xmath5 . in this case
, each linear polarization state @xmath46 provides a strategy to extract @xmath5 that is completely analogous to the one proposed in @xcite , which makes use of @xmath230 decays . if we combine @xmath213 with @xmath123 , we obtain @xmath231 the intersection of the contours in the @xmath5@xmath211 plane described by this expression with those related to ( [ bdet ] ) allows us to determine @xmath5 and @xmath211 .
if we use @xmath214 as an input in order to extract @xmath5 from @xmath1 , it is , however , more favourable to follow the approach discussed in the previous section , i.e. to use ( [ sol ] ) , and to fix @xmath143 ( or @xmath173 ) through the the @xmath2 observables with the help of ( [ rel-1p ] ) and ( [ rel-2 ] ) . using in addition the observables involving the third linear polarization state @xmath225 that we have not employed so far
, we can also fix its hadronic parameters @xmath227 and @xmath228 , as well as the strong phase difference @xmath232 . comparing @xmath232 with its @xmath2 counterpart @xmath233
, we may obtain valuable insights into possible corrections to ( [ rel-1p ] ) . as an interesting by - product
, this strategy allows us to take into account also the penguin effects in the extraction of the @xmath6@xmath7 mixing phase @xmath90 from @xmath2 .
although the penguin contributions are strongly suppressed in this mode because of the tiny parameter @xmath224 ( see ( [ bd - ampl2 ] ) ) , they may well lead to uncertainties of the extracted value of @xmath90 at the level of @xmath234 , since @xmath235 within the standard model .
a measurement of @xmath236 would allow us to determine the wolfenstein parameter @xmath237 @xcite , thereby fixing the height of the unitarity triangle .
since the decay @xmath2 is very accessible at `` second - generation '' @xmath12-physics experiments performed at hadron machines , for instance at lhcb , it is an important issue to think about the hadronic uncertainties affecting the determination of @xmath90 from the corresponding angular distribution .
the approach discussed above allows us to control these uncertainties with the help of @xmath238 .
the experimental feasibility of the determination of @xmath5 from the @xmath1 angular distribution depends strongly on the `` penguin parameters '' @xmath211 .
it is very difficult to estimate these quantities theoretically .
in contrast to the `` usual '' qcd penguin topologies , the qcd penguins contributing to @xmath238 require a colour - singlet exchange , i.e. are `` zweig - suppressed '' .
such a comment does not apply to the electroweak penguins , which contribute in `` colour - allowed '' form .
the current current amplitude @xmath182 originates from `` colour - suppressed '' topologies , and the ratio @xmath239 $ ] , which governs @xmath211 , may be sizeable
. it would be very important to have a better theoretical understanding of the quantities @xmath240 .
however , such analyses are far beyond the scope of this paper , and are left for further studies .
if the parameters @xmath211 should all be very small , which would be indicated by @xmath241 , we could still determine the @xmath3@xmath4 mixing phase from the observables @xmath242 .
if we use , in addition , @xmath243 and fix @xmath244 through the corresponding @xmath2 observable , @xmath245 can be determined as well .
consequently , the @xmath3@xmath4 mixing phase @xmath214 can be fixed _
unambiguously _ this way , thereby resolving a twofold ambiguity , which arises in the extraction of @xmath214 from @xmath153 .
this mode probes only @xmath246 .
since @xmath42 , we are left with a twofold ambiguity for @xmath247 $ ] .
if we assume that @xmath248 $ ] , as implied by the measured value of @xmath152 , we can fix @xmath8 unambiguously . for alternative methods to deal with ambiguities of this kind , see @xcite .
before we turn to @xmath10 and @xmath249 decays , let us note that the approach presented in this section can also be applied to the angular distributions of the decay products of @xmath250\,k^\ast[\to\pi^0k_{\rm s}]$ ] and @xmath251 .
for the @xmath230 and @xmath252 variants of these strategies , see @xcite .
in this section , we discuss further applications of the general strategies presented in section [ strat ] . all of the methods discussed below have counterparts using @xmath0 decays into two pseudoscalar mesons .
if we replace the pseudoscalars by higher resonances , for example , by vector mesons , as in the following discussion , the angular distributions of their decay products provide interesting alternative ways to extract ckm phases and hadronic parameters , going beyond the @xmath253 strategies . because of the many observables provided by the angular distributions , we can
, moreover , perform many interesting cross checks , for example , of certain flavour - symmetry relations .
the decay @xmath256 originates from @xmath257 quark - level processes . using the same notation as in ( [ bd - ampl ] )
, we have @xmath258_f)=\left(1-\frac{\lambda^2}{2}\right ) { \cal c}_f\,e^{i\gamma}\left[1-d_f\,e^{i\theta_f}e^{-i\gamma}\right],\ ] ] where @xmath259\ ] ] and @xmath260.\ ] ] in order to distinguish the @xmath256 amplitudes from the @xmath178 case discussed in the previous section , we have introduced the tildes .
the phase structure of the @xmath256 decay amplitude given in ( [ bdrhorho - ampl ] ) , which is an exact parametrization within the standard model , is completely analogous to the one for the @xmath261 amplitude given in @xcite , where a more detailed discussion can be found . the expressions for the observables describing the time evolution of the angular distribution of the decay products of @xmath262\,\,\rho^-[\to\pi^-\pi^0]$ ] can be obtained straightforwardly from the formulae given in section [ evol ] , by performing the following substitutions : @xmath263 because of the factor of @xmath264 in front of the square brackets on the right - hand side of ( [ bdrhorho - ampl ] ) , we have to deal with a small complication .
since the observables are governed by @xmath265,\ ] ] we have to do the following replacement , in addition : @xmath266 let us now turn to the decay @xmath267 , which is due to @xmath268 quark - level transitions .
for a given final - state configuration @xmath27 of the @xmath269 pair , its decay amplitude can be parametrized as follows : @xmath270_f)=\lambda\,{\cal c}_f'\,e^{i\gamma } \left[1+\left(\frac{1-\lambda^2}{\lambda^2}\right)d_f'\,e^{i\theta_f ' } e^{-i\gamma}\right],\ ] ] where @xmath271 and @xmath272 take the same form as ( [ cf - def ] ) and ( [ df - def ] ) , respectively , and the primes have been introduced to remind us that we are dealing with a @xmath201 mode .
the phase structure of ( [ bs - ampl ] ) is completely analogous to the @xmath273 decay amplitude @xcite .
the observables of the time - dependent @xmath274\,k^{\ast-}[\to\overline{\pi } \overline{k}]$ ] angular distribution can be obtained straightforwardly from the formulae given in section [ evol ] by simply using the replacements : @xmath275 moreover , we have to perform the substitution @xmath276 because of the factor of @xmath264 in front of the square brackets in ( [ bs - ampl ] ) .
explicit expressions for the @xmath254 and @xmath255 angular distributions in terms of helicity amplitudes can be found in @xcite .
since @xmath256 and @xmath267 are related to each other by interchanging all down and strange quarks , the @xmath277-spin flavour symmetry of strong interactions implies @xmath278 as well as @xmath279 in contrast to ( [ rel-1])([rel-2 ] ) , these relations do not rely on any dynamical assumption just on the @xmath277-spin flavour symmetry .
they can be used to combine the @xmath254 and @xmath255 observables with each other , thereby allowing the extraction of the ckm angle @xmath5 and of the @xmath280@xmath281 mixing phases @xmath42 and @xmath282 .
in contrast to the @xmath283 , @xmath284 variant of this approach proposed in @xcite , _ both _ mixing phases and the ckm angle @xmath5 can in principle be determined simultaneously . however , for the extraction of @xmath5 , it is more favourable to fix @xmath214 and @xmath90 separately .
then we are in a position to determine two contours in the @xmath5@xmath134 and @xmath5@xmath285 planes in a _ theoretically clean _ way with the help of ( [ sol ] ) .
using now the @xmath277-spin relation ( [ surel-1p ] ) , @xmath5 and all hadronic parameters describing the decays @xmath254 and @xmath255 can be determined .
as we have already noted , the hadronic parameters provide a very fertile testing ground for model calculations of the decays @xmath254 and @xmath255 . in particular , the penguin parameters @xmath286 and
@xmath272 would be very interesting ; comparing their values with each other , we could obtain valuable insights into @xmath277-spin - breaking corrections .
moreover , there is one strong phase difference @xmath232 left , which can be compared with its @xmath277-spin counterpart @xmath287 .
if we should find a small difference between these phases , it would be quite convincing to assume that our @xmath277-spin input ( [ surel-1p ] ) is also not affected by large corrections .
let us finally note that there is another interesting way to parametrize the @xmath254 decay amplitudes ( see also @xcite ) . if we eliminate @xmath288 through the unitarity of the ckm matrix instead of @xmath289
, as done in ( [ bdrhorho - ampl ] ) we obtain @xmath290_f)=\left(1-\frac{\lambda^2}{2}\right)\lambda^3 a\,r_b\,e^{i\gamma}\left[\tilde a_{\rm cc}^{(u)f}+\tilde a_{\rm pen}^{(uc)f } \right]\left[1+r_f\,e^{i\sigma_f}e^{-i(\beta+\gamma)}\right],\ ] ] where @xmath291,\ ] ] with @xmath292 taking into account that we have @xmath42 and @xmath293 within the standard model , we arrive at @xmath294 and at the `` effective '' mixing phase @xmath295 .
consequently , using the strategy presented in section [ strat ] , the @xmath254 angular distribution allows us to probe also the combination @xmath296 directly , i.e. to determine @xmath13 as a function of a cp - conserving strong phase difference @xmath134 ( see also ( [ r - approx ] ) and ( [ id - approx ] ) ) . needless to note that the decay @xmath297 may also be interesting in this respect .
since the normalization factors @xmath298 of the parametrization ( [ bdrhorho - ampl2 ] ) are proportional to @xmath299 which is governed by `` colour - allowed tree - diagram - like '' topologies , it may well be that @xmath300 .
this relation would allow us to extract @xmath13 , as well as the @xmath254 hadronic parameters , which include also another strong phase difference @xmath232 , providing an important cross check .
the decays @xmath303 and @xmath304 are pure `` penguin '' modes , originating from @xmath305 and @xmath306 quark - level transitions , respectively .
they do not receive contributions from current current operators at the `` tree '' level , and can be parametrized within the standard model in complete analogy to ( [ bd - ampl ] ) and ( [ bd - ampl2 ] ) .
we have just to set the current current amplitudes equal to zero in these expressions .
the decays @xmath303 and @xmath304 are related to each other by interchanging all down and strange quarks , i.e. through the @xmath277-spin flavour symmetry of strong interactions , and the strategies to probe @xmath5 and the @xmath280@xmath281 mixing phases are analogous to those discussed in section [ bpsi ] .
since the @xmath307 decays are pure `` penguin '' modes , they represent a particularly sensitive probe for new physics . an interesting alternative to parametrize the @xmath308_f$ ] decay amplitudes within the standard model
is given as follows : @xmath309_f)=\lambda^3a\,r_t\ , { \cal a}_{\rm pen}^{(tu)f}\,e^{-i\beta}\left[1-g_f\,e^{i\varphi_f } e^{i\beta}\right],\ ] ] where @xmath310 may well be sizeable due to the presence of final - state - interaction effects @xcite .
consequently , we have @xmath311 because of the factor of @xmath312 in front of the square brackets on the right - hand side of ( [ bdkk - ampl ] ) , the `` effective '' mixing phase is given by @xmath313 .
consequently , the strategy presented in section [ strat ] allows us to probe the cp - violating weak phase @xmath8 of the ckm element @xmath314 . within the standard model
, we have @xmath315 .
however , this relation may well be affected by new physics , and represents a powerful test of the standard - model description of cp violation ( for a recent discussion , see @xcite )
. therefore it would be very important to determine this combination of ckm phases experimentally .
the observables of the @xmath316\,\overline{k^{\ast0}}[\to\pi^+k^-]$ ] angular distribution may provide an important step towards this goal .
the angular distributions of certain quasi - two - body modes @xmath317 , where both @xmath17 and @xmath18 carry spin and continue to decay through cp - conserving interactions , provide valuable information about ckm phases and hadronic parameters .
we have presented the general formalism to accomplish this task , taking into account also penguin contributions , and have illustrated it by having a closer look at a few specific decay modes . in comparison with strategies using non - leptonic @xmath0 decays into two pseudoscalar mesons , an important advantage of the angular distributions is that they provide much more information , thereby allowing various interesting cross checks , for instance , of certain flavour - symmetry relations .
moreover , they provide a very fertile testing ground for model calculations of the @xmath318 modes .
we have pointed out that the decay @xmath1 can be combined with @xmath2 to extract the @xmath3@xmath4 mixing phase @xmath42 and if penguin effects in the former mode should be sizeable also the angle @xmath5 of the unitarity triangle . as an interesting by - product
, this strategy allows us to take into account also the penguin effects in the extraction of the @xmath6@xmath7 mixing phase from @xmath2 .
if penguin effects should be very small in @xmath1 , @xmath214 could still be determined and it would even be possible to resolve a twofold ambiguity , arising in the extraction of this ckm phase from @xmath153 . other interesting applications , involving @xmath10 and @xmath11 decays , were also noted . within the standard model
, these modes are expected to exhibit branching ratios at the @xmath319 level ; also the one for @xmath320 may well be enhanced , from its `` short - distance '' expectation of @xmath321 to this level , by final - state - interaction effects . since the formalism presented in this paper is very general , it can of course
be applied to many other decays .
detailed studies are required to explore which channels are most promising from an experimental point of view . although the @xmath58 modes listed above may already be accessible at the asymmetric @xmath322@xmath323 @xmath12-factories operating at the @xmath324 resonance , which will start taking data very soon , the strategies presented in this paper appear to be particularly interesting for `` second - generation '' experiments at hadron machines , such as lhcb or btev , where also the very powerful physics potential of the @xmath54 system can be exploited . for reviews , see , for instance , _ the babar physics book _ , eds . p.f . harrison and h.r .
quinn ( slac report 504 , october 1998 ) ; y. nir , published in the proceedings of the 18th international symposium on lepton photon interactions ( lp 97 ) , hamburg , germany , 28 july1 august 1997 , eds . a. de roeck and a. wagner ( world scientific , singapore , 1998 ) , p. 295
[ hep - ph/9709301 ] ; m. gronau , _ nucl .
phys.proc .
suppl . _ * 65 * ( 1998 ) 245 ; r. fleischer , _ int . j. mod .
_ * a12 * ( 1997 ) 2459 .
see , for example , y. grossman and h.r .
quinn , _ phys .
_ * d56 * ( 1997 ) 7259 ; j. charles , a. le yaouanc , l. oliver , o. pne and j .- c .
raynal , _ phys .
_ * b425 * ( 1998 ) 375 and _ phys .
rev . _ * d58 * ( 1998 ) 114021 ; a.s .
dighe , i. dunietz and r. , _ phys .
_ * b433 * ( 1998 ) 147 .
see , for instance , l. wolfenstein , _ phys .
_ * d52 * ( 1995 ) 537 ; b. blok and i. halperin , _ phys .
_ * b385 * ( 1996 ) 324 ; b. blok , m. gronau and j.l .
rosner , _ phys .
_ * 78 * ( 1997 ) 3999 ; j .- m .
grard and j. weyers , _ eur .
j. _ * c7 * ( 1999 ) 1 ; m. neubert , _ phys .
_ * b424 * ( 1998 ) 152 ; a.f .
falk , a.l .
kagan , y. nir and a.a .
petrov , _ phys .
_ * d57 * ( 1998 ) 4290 ; d. atwood and a. soni , _ phys . rev .
_ * d58 * ( 1998 ) 036005 . | the time - dependent angular distributions of certain @xmath0 decays into final states that are admixtures of cp - even and cp - odd configurations provide valuable information about ckm phases and hadronic parameters .
we present the general formalism to accomplish this task , taking also into account penguin contributions , and illustrate it by considering a few specific decay modes .
we give particular emphasis to the decay @xmath1 , which can be combined with @xmath2 to extract the @xmath3@xmath4 mixing phase and if penguin effects in the former mode should be sizeable also the angle @xmath5 of the unitarity triangle . as an interesting by - product
, this strategy allows us to take into account also the penguin effects in the extraction of the @xmath6@xmath7 mixing phase from @xmath2 .
moreover , a discrete ambiguity in the extraction of the ckm angle @xmath8 can be resolved , and valuable insights into @xmath9-breaking effects can be obtained .
other interesting applications of the general formalism presented in this paper , involving @xmath10 and @xmath11 decays , are also briefly noted .
cern - th/99 - 92 + hep - ph/9903540 * extracting ckm phases from angular distributions of + @xmath0 decays into admixtures of cp eigenstates * robert fleischer + _ theory division , cern , ch-1211 geneva 23 , switzerland _
cern - th/99 - 92 + april 1999 |
the interplay between network structure and search dynamics has emerged as a busy sub - field of statistical network studies ( see e.g. refs .
@xcite ) . consider a simple graph @xmath0 ( where @xmath1 is a set of @xmath2 vertices and @xmath3 is a set of @xmath4 edges unordered pairs of vertices ) .
assume information packets travel from a source vertex @xmath5 to a destination @xmath6 .
we assume the packages are myopic agents ( at a given timestep they have access to information about the vertices in their neighborhood , but not more ) , have memory ( so they can e.g. perform a depth - first search ) but no previous knowledge of the network . let @xmath7 be the time for a packet @xmath8 to travel between its source and destination .
one commonly studied quantity of search efficiency is the expectation value of @xmath9 , @xmath10 , for randomly chosen @xmath5 and @xmath6 . in this work we attempt to find efficient ways to index @xmath1 and
utilize these indices for packet navigation .
we propose two schemes of indexing the vertices , and corresponding methods for packet navigation .
these schemes , along with two depth - first search methods ( not using vertex indices for more than remembering the path ) are examined on four network models .
we will first present the indexing and search schemes , then the network models for testing the algorithms , and last numerical results . to @xmath6 ( with @xmath11 ) . on the way from @xmath6 to @xmath12 the packet chooses the neighbor ( of the current vertex ) with lowest index , which here gives a longer route than the optimal @xmath13 .
( d ) shows a possible partition of branches of non - root vertices into classes of as similar size as possible ( as done in the asu indexing scheme ) .
( e ) shows a possible indexing based on the partition in ( d ) .
panel ( f ) displays a search from @xmath5 to @xmath6 with @xmath14 .
the shortest path from @xmath6 to @xmath12 is accurately found , but a detour to @xmath15 makes the search from @xmath12 to @xmath6 sub - optimal . ]
now we turn to the schemes for assigning indices to the vertices and using them in search processes .
our two main schemes are both inspired by search trees .
packets first moves towards a root vertex @xmath12 , then towards the destination .
unless the network really is a tree , this approach can not be exact a packet is not guaranteed to find the shortest way both from @xmath5 to @xmath12 and from @xmath12 to @xmath6 ) .
however , as we will see , one can assign indices such that the search either from @xmath5 to @xmath12 , or from @xmath12 to @xmath6 is certain to be as short as possible .
one of our schemes , asd ( accurate search up ) , will be such that the shortest upward search is guaranteed , the other , asd ( accurate search down ) , will have the shortest possible @xmath12 to @xmath6 search . on a technical note ,
@xmath1 is a set of distinct elements and an indexing scheme is a bijection @xmath16 $ ] . in the remainder of the text we will not explicitly distinguish @xmath17 from @xmath18 .
the numbers @xmath19 can be arranged into a search tree @xcite such that the expected value of @xmath9 scales like @xmath20 . in fig .
[ fig : ill](a ) we give an example of a search tree . to go from source @xmath5 to destination @xmath6
a packet first moves to the root @xmath12 by going to the neighbor with lowest index value . from the root to the destination , the package moves to the neighbor with the largest index smaller than , or equal to , @xmath6 .
our strategy for the asd indexing and search scheme is to construct a spanning tree @xmath21 for the network ; index the tree to make it a search tree ; and use the algorithm above to navigate from @xmath5 to @xmath6 .
the problem is , however that real networks are not trees .
imagine adding edges between vertices of the same heights and branches to the tree in fig .
[ fig : ill](a)the tree will still be a spanning tree , but the packets may not take the same path from @xmath5 to @xmath6 any more . as we will see , with certain ways of constructing the tree and indexing the vertices the search , either from @xmath5 to @xmath12 or @xmath12 to
@xmath6 will be optimal .
we construct @xmath21 in the following way : 1 .
let the root @xmath12 be a vertex of smallest eccentricity ( maximal distance to an other vertex ) .
2 . construct the tree such that the distances to the root is the same in @xmath21 and @xmath22 .
in other words , such that all edges in @xmath23 go between different neighborhoods @xmath24 and @xmath25 for some level @xmath26 , where @xmath27 is the _ height _ of the tree ( by the choice of @xmath12 , @xmath27 is also the radius of the graph ) .
such a tree can be constructed by finding the set of followed edges in a breadth - first search @xcite starting from @xmath12 . when it is not clear which vertex , or edge , to choose in the above construction , we choose one at random from all the possible candidates .
when @xmath23 is constructed , let the indices be a preordering of the vertices in @xmath23 ( i.e. the order of first - occurrence of the vertex in a depth - first search of the graph ) @xcite .
now we prove that this indexing and search algorithm always gives the shortest paths from the root to a vertex @xmath6 .
let @xmath28 be the edges of @xmath23 and let @xmath29 be the maximal subtree with @xmath30 as root . by construction ,
all vertices in @xmath29 have indices in @xmath31 $ ] ( where @xmath32 denotes the cardinality of a subgraph ) .
let @xmath33 be the largest index in @xmath30 s neighborhood smaller than @xmath6 .
assume there is an edge @xmath34 that the search will follow , i.e. that @xmath35 .
this means that @xmath36 . by construction
, @xmath33 is the only vertex in @xmath37 at a distance @xmath38 ( the distance from the rest of @xmath37 to the root is at least @xmath39 ) . since @xmath40 , we have @xmath41 which contradicts the existence of an edge @xmath42 . thus searches from @xmath12 to @xmath6
will always follow the edges of @xmath23 , which also means the @xmath12@xmath6-searches will be as short as possible .
searching upwards , from @xmath30 to @xmath12 , in a graph indexed as above is harder .
we know that one shortest path goes via a vertex @xmath43 with smaller index than @xmath30 , but there might sub - optimal paths via indices @xmath33 in the intervals @xmath44 and @xmath45 , and there might also be paths via vertices of index larger than @xmath43 , that is optimal .
for example , assume the search tree in fig .
[ fig : ill](a ) comes from a graph with the additional edges @xmath46 , @xmath47 and @xmath48 ( see fig .
[ fig : ill](b ) ) .
then , the shortest path from @xmath49 to @xmath12 via a vertex of lower index is @xmath50 , but there is an equally long path via a vertex of larger index , @xmath13 , and longer paths via vertices both smaller and larger than @xmath51 but smaller than @xmath49 .
there thus no general way of finding the shortest way from @xmath5 to @xmath12 . instead , we always choose the vertex with the smallest index in the neighborhood . by this strategy
a packet will come closer to @xmath12 , in index space , for every step .
furthermore , in tree - like parts of the graph , the search will follow the shortest paths .
an illustration of the asd search can be found in fig .
[ fig : ill](c ) .
consider a tree @xmath21 constructed as in the previous section and an indexing such that @xmath52 implies @xmath53 ( i.e. , all indices of a level further from the root is larger than in levels closer to @xmath12 ) .
with such an indexing , since the neighbor of a vertex with the smallest index necessarily is one step closer to the root , a packet can always find one shortest way too the root .
but once the package is at the root the indices is not of so much help .
the search from @xmath12 to @xmath6 has to be , essentially , a depth - first search .
there are , however , a few tricks to speed up the search . first , there is no need to search deeper than @xmath6if @xmath54 , then @xmath55 .
second , one can choose the indices @xmath56 of one level in the tree in a way to narrow down the search .
for example , one can divide the vertices into @xmath57 classes ( defined by e.g. the remainder when the index is divided by @xmath57 ) and index vertices of connected regions of the graph with indices of the same class .
the search can then be restricted to the same class as the destination .
we will pursue this idea with @xmath58 . to derive the asu indexing scheme ,
the first goal is to divide the vertices into classes of connected subgraphs .
furthermore , we require all classes to be connected to the root vertex . another aim is to make the classes of as similar sizes as possible .
our first step is to make @xmath59 ( the degree , or number of neighbors , of @xmath12 ) parallel depth - first searches .
second we group the @xmath59 search trees into @xmath57 groups with maximally similar sizes . in our case , we seek a partition of the search trees into two classes such that the sums of vertices in the respective classes are as close as possible
. then we go through the levels , starting from the root , and assign numbers such that vertices of one partition have even indices , while the other has odd numbers ( this assignment might not always work ) . to avoid systematic errors we sample the elements of levels randomly .
this construction scheme is illustrated in fig .
[ fig : ill](d ) and ( e ) . as a reference
, we also run simulations for two depth - first search methods that do not utilize indices @xcite .
one of them , rnd , is regular depth - first search where the neighbors are traversed in random order . in the other , deg ,
the neighbors are chosen in order from high to low degree . just like for asu and asd methods ,
a packet is assumed to have knowledge about its neighborhood if the destination is in the neighborhood of a vertex , then the search will be over the next time step .
the efficiency of our indexing and search schemes are more or less directly affected by the network structure . to investigate this relationship we test the search schemes on four different types of network models : modified erds
rnyi ( er ) graphs @xcite , square lattices , barabsi
albert ( ba ) @xcite and holme kim ( hk ) @xcite networks . to facilitate comparison
, we have the same average degree , four ( dictated by the square grid ) , in all networks .
the er model is the simplest model for randomly generating simple graphs with @xmath2 vertices and @xmath4 edges .
the edges are added one by one to randomly chosen vertex pairs ( the only restriction being that loops or multiple edges are not allowed ) .
a problem for our purpose is that er graphs are not necessarily connected ( something required to measure @xmath10 ) . to remedy this
we propose a scheme to make networks connected . 1 .
[ step : components ] detect the connected components
[ step : seq ] go through the connected components sequentially .
denote the current component @xmath60 .
[ step : random_component ] pick a component @xmath61 randomly .
2 . pick a random edge @xmath62 whose removal would not fragment @xmath61 . if no such edge exist , go to step [ step : seq ] .
[ step : rnd_vertex ] pick a random vertex @xmath33 of @xmath60 .
[ step : repl ] replace @xmath62 by @xmath63 . if the edge @xmath63 would exist already ( an unlikely event ) , go to step [ step : repl ] . if there is no vertex @xmath64 such that @xmath63 does not already exist , then go to [ step : seq ] .
3 . if the network is disconnected still , go to step [ step : components ] . in practice , even for our largest system sizes , the above algorithm converges in a few iterations .
the number of edges needed to be added never exceed a few percent of @xmath4 , and this addition is made with greatest possible randomness ; hence we believe the essential network structure of the er model is conserved .
we use square lattices with periodic boundary conditions .
@xmath2 vertices spread out regularly on a @xmath65-grid such that the vertex with coordinates @xmath66 , @xmath67 , is connected to @xmath68 , @xmath69 , @xmath70 , @xmath71 ( if @xmath72 , we formally let @xmath73 , if @xmath74 we let @xmath75 represent @xmath76 ; and correspondingly for @xmath77 ) .
the popular ba model @xcite of networks with a power - law degree distribution works as follows ( with our parameter settings ) .
start with one vertex connected to two degree - one vertices .
iteratively add vertices connected to two other vertices .
let the probability of connecting the new vertex to a vertex @xmath30 already present in the network is proportional to @xmath78 ( so called _ preferential attachment _ ) .
the hk model @xcite is a modification of the ba model to give the network higher number of triangles .
when edges are added from the new vertex to already present vertices , the first edge is added by preferential attachment .
the second edge is added to one of @xmath30 s neighbors , forming a triangle . as a function of the graph
sizes @xmath2 . in all panels , we display data for the different indexing and search schemes .
the shaded areas are unreachable ( corresponding to @xmath10 values smaller than the theoretical minimum the average distance @xmath79 ) .
the different panels correspond to the modified er model , square grid , ba model and hk model networks respectively .
error bars would have been smaller than the symbol sizes . ]
we study the search schemes on the four different network topologies numerically .
we use @xmath80 independent networks and @xmath80 different @xmath5@xmath6-pairs for every network .
the network sizes range from @xmath81 to @xmath82 .
to @xmath12 with the asd indexing and search scheme . a packet from @xmath83 to @xmath76
will travel along the perimeter to @xmath84 and then move towards the center . ] in fig .
[ fig : sca ] we display the average search times as a function of system size for our simulations .
the most conspicuous feature is that the asd scheme is always , by far , the most efficient .
while asu and deg are close to the least efficient method ( rnd ) , asd is rather close to the theoretical limit ( equal to the average distances @xmath10the upper border of the shaded areas in fig .
[ fig : sca ] ) . to be more precise
, @xmath10 is quite constant , about two times larger than the average distance . the other search schemes ( asu , deg and rnd ) follow faster increasing functional forms .
for the square lattice , these three schemes increase , approximately proportional to @xmath2 ( the analytical value for two - dimensional random walks ) whereas for asd , @xmath10 scale like distances in square grids , @xmath85 .
one way of interpreting this result is that while asd manages to find the root as fast as it finds the destination from the root , asu fails to find @xmath6 faster than a random search .
the slow downward performance of asu is not unexpected
the @xmath12@xmath6-search in asu only differs from a random depth - first search in that it does not search further than the level of the destination , and that it restricts the search - space to half its original size by dividing the vertices into odd and even indices .
the fast upward search of asd is more surprising . in fig .
[ fig : wcs ] we show a network where asd performs badly .
the average time to search upwards is @xmath86 as @xmath87 .
the downward search takes @xmath88 giving a total expected value of @xmath89 .
this can be compared to the average distance @xmath90 .
for this example , @xmath10 and @xmath79 diverge in a way not seen in the network models .
why is the search so much faster in the model networks ?
one point is that the worst - case indexing seen in fig . [
fig : wcs ] is very unlikely .
since the spokes would be sampled randomly , the chance that a vertex at the perimeter not finds @xmath12 in two steps is @xmath91 , the probability of a perimeter vertex to find @xmath12 in @xmath84 steps is @xmath92 , and so on . carrying on this calculation , a vertex at the perimeter reaches @xmath12 in @xmath93 timesteps giving @xmath94not too far from the observed @xmath95 .
we note however that for the model networks many other factors that are not present in the wheel - graph of fig .
[ fig : wcs ] affect @xmath96 .
for example , the high density of short triangles in the hk model networks will introduce many edges between vertices of the same level in @xmath21 which will affect the search efficiency .
@xmath10 is approximately linear for the asu , deg and rnd on all network models .
the slopes of these curves are , however , a little different .
first , the deg method is more efficient ( compared to asu and rnd ) for ba networks , than for the modified er model .
this observation ( also made in ref .
@xcite ) can be explained by the skewed degree distribution in the ba - network the packet reaches high - degree vertices fast .
the packet can see a large part of the network from these hubs , and is therefore more likely to see @xmath6 .
more interesting , perhaps , is that asu is more efficient for the networks with a higher density of short cycles ( the square lattice and hk models ) .
a rough explanation is that the partition procedure of asu cuts off many edges between vertices at the same distance from @xmath12 .
since there are many such edges in these network models , the network will effectively be sparser ( without changing @xmath22 s diameter ) , which results in a better performance .
we have investigated navigation in valued graphs , more specifically in indexed graphs
graphs where every vertex is associated with a unique number in the interval @xmath97 $ ] .
these indices can be assigned to facilitate the packet navigation .
the packets are assumed to have no _ a priori _ knowledge about the network , except the neighborhoods of their current positions , but memory enough to perform a depth - first search .
we find that one of our investigated methods , asd , is very efficient for four topologically very different network models .
the searches with the asd scheme are roughly twice as long as the shortest paths ( scaling in the same way as the average distance ) .
navigation on indexed graphs has applications in distributed information systems . if , specifically , the amount of information that can be stored at the vertices were limited , search strategies such as ours would be useful .
one such system is the autonomous system level internet where the information stored at each vertex ( with the current protocols ) increase at least as fast as the networks themselves . for most real - world applications ( other examples being _ ad hoc _ networks@xcite or peer - to - peer networks @xcite ) there are additional constraints so that the algorithms of this paper can not immediately be applied .
such networks are typically changing over time , so the indexing should ideally be possible to be extended on the fly as vertices and edges are added and deleted from the network . apart from this ,
a future direction for research on indexed graphs is to improve the performance of the algorithms presented in this work .
there might be search - tree based algorithm that neither finds the shortest path to the root , nor finds the shortest way to the destination .
for some network topologies there might be faster algorithms that are not based on constructing a spanning tree .
consider , for example , modular networks @xcite ( i.e. networks with tightly connected subgraphs that are only sparsely interconnected ) in such networks the search can be divided into two stages first find the cluster of the destination , then the destination .
these two stages should be reflected in a fast navigation algorithm .
n. ganguly , l. brusch , and a. deutsch .
design and analysis of a bio - inspired search algorithm for peer to peer networks . in o.
babaoglu , m. jelasity , a. montresor , c. fetzer , and s. leonardi , editors , _ self - star properties in complex information systems _ , pages 358372 , new york , 2007 .
springer - verlag .
n. sarshar , p. o. boykin , and v. p. roychowdhury .
percolation search in power law networks : making unstructured peer - to - peer networks scalable . in _ proceedings of fourth international conference on peer - to - peer computing _ , pages 29 .
ieee , 2004 . | we investigate efficient methods for packets to navigate in complex networks .
the packets are assumed to have memory , but no previous knowledge of the graph .
we assume the graph to be indexed , i.e. every vertex is associated with a number ( accessible to the packets ) between one and the size of the graph .
we test different schemes to assign indices and utilize them in packet navigation .
four different network models with very different topological characteristics are used for testing the schemes .
we find that one scheme outperform the others , and has an efficiency close to the theoretical optimum .
we discuss the use of indexed - graph navigation in peer - to - peer networking and other distributed information systems . |
this research was supported in part by the polish committee for scientific research ( kbn ) under contract no . 5 p03b 014 21 .
numerical calculations were performed at the interdisciplinary centre for mathematical and computational modelling ( icm ) at warsaw university .
author would like to thank aurel bulgac for discussions and valuable comments .
999 g. baym , h.a .
bethe , c.j .
pethick , _ nucl .
phys . _ * a175 * , 225 ( 1971 ) .
negele and d. vautherin , _ nucl .
phys . _ * a207 * , 298 ( 1973 ) ; p. bonche and d. vautherin , _ nucl .
phys . _ * a372 * , 496 ( 1981 ) ; _ astron .
astrophys . _ * 112 * , 268 ( 1982 ) .
pethick and d.g .
ravenhall , _ annu .
nucl . part .
* 45 * , 429 ( 1995 ) .
f. douchin , p. haensel , j. meyer , _ nucl .
phys . _ * a 665 * , 419 ( 2000 ) .
f. douchin , p. haensel , _ phys .
_ * b485 * , 107 ( 2000 ) .
p. magierski , p - h .
heenen , _ phys .
rev . _ * c65 * , 045804 ( 2002 ) .
p. magierski , a. bulgac , p .- h .
heenen , _ int .
* a17 * , 1059 ( 2002 ) .
a. bulgac , p. magierski , _ nucl .
phys . _ * a683 * , 695 ( 2001 ) ; _ nucl .
* a703 * , 892 ( 2002 ) ; _ phys . scripta _ * t90 * , 150 ( 2001 ) ; _ acta phys . pol . _ * b32 * , 1099 ( 2001 ) .
baiko , a.y .
potekhin , d.g .
yakovlev , _ phys .
rev . _ * e64 * , 057402 ( 2001 ) .
a. rosch , _ quantum coherent transport of a heavy particle in a fermionic bath _ ( shaker verlag , aachen , 1997 ) ; a. rosch , t. kopp , _ phys .
lett . _ * 75 * , 1988 ( 1995 ) ; _ phys .
lett . _ * 80 * , 4705 ( 1998 ) ; p. haensel , j.l .
zdunik , j. dobaczewski , _ astron .
astroph . _ * 222 * , 353 ( 1989 ) . | the inner crust of a neutron star consists of nuclei immersed in a superfluid neutron liquid . as these nuclei move through the fermionic medium they bring it into motion as well .
as a result their mass is strongly renormalized and the spectrum of the ion lattice vibrations is significantly affected .
consequently we predict that the specific heat and the lattice thermal energy of the coulomb crystal at these densities are noticeably modified . in the standard picture
the inner crust of a neutron star consists of neutrons , protons and electrons . at low temperatures ( @xmath0 mev )
one expects the presence of atomic nuclei ( ions ) immersed in a neutron gas . at the densities : @xmath1 g/@xmath2 g/@xmath3 they form a crystal lattice stabilized by the coulomb interaction .
the electrons which are ultrarelativistic at these densities form a strongly degenerate , uniform gas .
the detailed structure of this part of the star has been the subject of a considerable theoretical effort ( .@xcite and references therein ) . at the bottom of the crust
the appearance of non - spherical nuclei , forming exotic structures , has been predicted . in this region besides the coulomb interaction also the quantum effects , associated with a neutron scattering on nuclear inhomogeneities , play an important role , leading eventually to a disordered phase .@xcite . in order to understand various processes associated with thermal evolution of a neutron star ,
the properties of coulomb crystals in the inner crust have been studied by many authors ( see e.g. .@xcite and references therein ) .
the plasma frequency of the system reads : @xmath4 , where @xmath5 denotes the ion density , @xmath6 is the proton number of an ion , and @xmath7 is the ion mass . since @xmath8 , where @xmath9 is the wigner - seitz cell radius then :
@xmath10 the quantities : @xmath6 and @xmath9 , in the above equation , can be obtained from calculations based on a density functional method .
however , the determination of the ion mass @xmath7 requires a certain caution , since the nucleus is immersed in the fermionic environment and its bare mass has to be renormalized to take into account the interaction with surrounding neutrons .
if the neutrons outside a nucleus have formed a normal fermi gas the problem would have been rather complicated.@xcite . in the inner crust
, however , it simplifies due to the fact that at low temperatures , @xmath11 mev , the neutron gas is predicted to be strongly paired . in this case
one may apply the hydrodynamic approximation which describe the system by an irrotational velocity field .
although at finite temperatures there always exists a gas of quasiparticle excitations which form a normal flow , at temperatures much lower than the critical temperature its influence can be neglected .
we assume in the following that the nucleus has a uniform nucleon density @xmath12 and fills a sphere of radius @xmath13 .
the neutron gas is characterized by the density @xmath14 .
we assume moreover that the flow velocity is below the critical velocity for loss of superfluidity . in order to calculate the mass @xmath7 of an ion ,
the poisson equation for the velocity field has to be solved , both inside and outside a nucleus .
the boundary conditions read : @xmath15 where @xmath16 and @xmath17 stand for the velocity field inside and outside a nucleus , respectively .
the vector @xmath18 is an outward normal to the surface of a spherical nucleus and @xmath19 is a radial coordinate ( the center of a nucleus is placed at @xmath20 ) .
the velocity of the surface element was denoted by @xmath21 .
the first equation stems from the requirement that the phase of the wave function of the superfluid system is continuous , the second one is a conservation of mass in radial flow , and finally the third one ensures that the correct asymptotic behavior for the velocity field is obtained .
the knowledge of the velocity field allows us to determine the kinetic energy and corresponding effective ion mass .
consequently , the following expression for the mass of an ion being a subject of a translational motion can be derived : @xmath22 where @xmath23 , and @xmath24 is a nucleon mass .
the above expression relates the mass of an ion to the density of the neutron gas .
note , that for a uniform system , i.e. @xmath25 the ion mass is equal to zero .
since the nucleon density vary substantially across the inner crust , one expects that the properties of the coulomb crystal will be noticeably modified . indeed , the plasma frequency in the inner crust changes as compared to the value calculated for the case : @xmath26 ( ion in the vacuum ) .
namely , @xmath27 , where @xmath28 corresponds to the plasma frequency for the system with renormalized ion masses . in the fig .
1 we have plotted @xmath29 and @xmath30 as a function of the nucleon density .
one can notice large discrepancies between the value obtained using the bare nuclear mass and the renormalized one .
it clearly indicates that the coulomb crystal becomes more stiff at the bottom of the crust .
consequently the density of the phonon spectrum is also modified as compared to the case of bare ion masses .
the thermodynamic functions : the lattice thermal energy and the specific heat have been shown in the fig .
2 . one can notice that the renormalized values at a given density and temperature are smaller .
in the limit of @xmath31 the analytic expressions for the ratios of lattice thermal energy @xmath32 and specific heat @xmath33 can be derived . applying the phenomenological relations from the ref.@xcite one
gets : @xmath34 where @xmath35 and @xmath36 denote the renormalized values of lattice energy and specific heat , respectively .
the following conclusions can be drawn from the presented results : 1 .
there is a substantial renormalization effect of a nuclear / ion mass in the inner crust of a neutron star , due to the presence of a superfluid neutron liquid .
2 . consequently , the phonon spectrum and thermodynamic functions of the coulomb crystal are significantly altered .
3 .
thermal and electric conductivities of the inner crust , governed by the electron - phonon scattering are expected to be modified . in particular , the contributions coming from umklapp processes have to be recalculated using the renormalized ion masses .
4 . there is clearly a need for consideration of other nuclear degrees of freedom like e.g. shape vibrations , which may significantly modify the electron transport properties . |
a simple theoretical way to consider inclusive decays of hadrons containing one heavy quark is the _ spectator model _ , which states that only the heavy quark participates in the transition , while the light degrees of freedom are unaffected by the process . as a result
, the model predicts that all the hadrons with the same heavy quark should have the same lifetime . on the other hand ,
the experimental ratios of beauty hadron lifetimes are @xcite : @xmath4 @xmath5 eq.([ratiolam ] ) shows that the ratio @xmath6 significantly differs from the spectator model prediction .
one can employ a more refined theoretical approach , based on qcd , to compute inclusive decay widths of heavy hadrons ; the method consists in an expansion in powers of @xmath7 @xcite . by using the optical theorem ,
the inclusive width of a hadron containing the heavy quark @xmath8 into a final state with assigned quantum numbers @xmath9 can be written as : @xmath10 where @xmath11 is the transition operator describing @xmath8 with the same momentum in the initial and final state : @xmath12 \hskip 3 pt .
\label{trans}\ ] ] @xmath13 is the effective weak lagrangian for the decay @xmath14 .
the time ordered product in ( [ trans ] ) is computed by an operator product expansion ( ope ) : @xmath15 where @xmath16 are local operators ordered by increasing dimension , and @xmath17 are coefficients ordered by increasing powers of @xmath18 . as a result , the inclusive width of a heavy hadron @xmath19 ( from now on we will refer to the beauty sector ) is represented by a sum of expectation values over @xmath19 of operators with increasing dimension : @xmath20 . & & \label{ris}\end{aligned}\ ] ] @xmath21 is the ckm matrix element relevant in the considered decay .
the first operator in ( [ ris ] ) is @xmath22 , which has dimension @xmath23 ; next , the chromomagnetic operator @xmath24 , responsible of the mass splitting between hadrons which differ only for the heavy quark spin orientation , has @xmath25 ; finally , @xmath26 have @xmath27 and contribute to @xmath2 in the expansion .
@xmath28 can be further reduced by using the heavy quark equation of motion in the limit @xmath29 : @xmath30 @xmath31 is the heavy quark kinetic energy operator due to its residual motion inside the hadron .
some important features should be noticed in ( [ ris ] ) .
the first term of the expansion reproduces the spectator model result ; @xmath32 terms are absent @xcite since @xmath33 operators can not contribute , being reducible , by equation of motion , to the @xmath23 operator @xmath22 ; finally , @xmath34 and @xmath35 are spectator blind , i.e. they are not sensitive to the light flavour .
@xmath2 terms come from four quark operators , which can be classified as follows @xcite : @xmath36 where @xmath37 , @xmath38 being the gell - mann matrices .
their contribution accounts for the presence of the spectator quark , which participates in the decay through the mechanisms known as weak annihilation and pauli interference ( the latter depends on the presence of two identical quarks in the final state ) . in order to use the result ( [ ris ] )
one has to know the matrix elements of all the previously mentioned operators .
let us define : @xmath39 @xmath40 @xmath41 depends on the spin @xmath42 of @xmath19 : @xmath43 \lambda_2 $ ] .
it enters with @xmath44 in the mass formula for a heavy hadron in the @xmath29 limit : @xmath45 the parameters @xmath46 , @xmath44 , @xmath47 are independent of the heavy quark mass . for @xmath0 mesons
, @xmath48 can be derived experimentally , since it is related to the mass splitting : @xmath49 . on the other hand : @xmath50 , since the light degrees of freedom have zero total angular momentum relative to the heavy quark .
various theoretical determinations exist for @xmath51 . from the mass formula ( [ mass ] ) and using experimental data on the mass of the @xmath52 baryon , one obtains : @xmath53 and hence it can be assumed : @xmath54 . in order to evaluate the matrix elements of @xmath27 operators , different approaches must be used for mesons and baryons . for @xmath0 mesons , factorization approximation gives : @xmath55 @xmath56 being the @xmath57 decay constant .
the same technique can not be applied to baryons , so that one must use other approaches . in the case of @xmath52 ,
heavy quark symmetries reduce the relevant operators to two @xcite : @xmath58 one can parameterize the matrix elements as follows : @xmath59 in the valence quark approximation : @xmath60 . it should be stressed that only large values of the parameter @xmath61 in ( [ otilde ] ) @xmath62 could explain the observed difference between @xmath3 and @xmath63 .
calculations based on quark model @xcite or exploiting the experimental splitting : @xmath64 @xcite give values of @xmath61 smaller than required . in order to have a different theoretical estimate , we computed by qcd sum
rules the matrix elements of the relevant @xmath27 operators over @xmath52 .
this calculation will be described in the next section .
the starting point to evaluate @xmath65 by qcd sum rules within the heavy quark effective theory ( hqet ) is the correlation function : @xmath66|0 > & & \label{corr}\end{aligned}\ ] ] where @xmath67 , @xmath68 are dirac indices ; the variable @xmath69 @xmath70 is related to the residual momentum of the incoming ( outgoing ) baryonic current @xmath71 with @xmath72 .
a suitable interpolating field for @xmath52 , in the limit @xmath29 , is given by @xcite : @xmath73 where @xmath74 means the transpose , @xmath75 , @xmath76 , @xmath77 are colour indices , @xmath78 is the effective heavy quark field , @xmath79 and @xmath67 is the charge conjugation operator .
@xmath80 is the @xmath52 light flavour matrix : @xmath81 the parameter @xmath82 is fixed to @xmath83 from the qcd sum rules analysis of @xmath84 @xcite , defined by : @xmath85 ( @xmath86 is the spinor for a @xmath52 of four - velocity @xmath87 ) . the hadronic representation of the correlator ( [ corr ] ) can be obtained saturating it by baryonic states , and considering the double pole contribution in the variables @xmath69 and @xmath88 : @xmath89 at the value @xmath90 .
@xmath91 represents the binding energy of the light degrees of freedom in @xmath52 : @xmath92 and must be derived within the same qcd sum rule framework . on the other hand , in the euclidean region , for negative values of @xmath69 , @xmath88 , the correlator ( [ corr ] ) can be computed in qcd , in terms of a perturbative contribution and of vacuum condensates .
the result can be written in a dispersive form : @xmath93 where possible subtraction terms have been omitted .
the explicit expression of the computed spectral function @xmath94 can be found in @xcite .
a sum rule for @xmath95 can be derived by equating @xmath96 and @xmath97 . moreover , invoking a global duality ansatz , the contribution of the higher resonances and of the continuum in @xmath96 in ( [ phad ] ) can be modeled as the qcd term outside the region @xmath98 , @xmath99 , with @xmath100 an effective threshold . finally
, the application of a double borel transform to both @xmath97 and @xmath96 in the momenta @xmath101 : @xmath102 ( and similar for @xmath88 with the borel parameter @xmath103 ) allows us to remove the subtraction terms in ( [ pope ] ) , that are polynomials in @xmath69 or @xmath88 ( the borel transform of a polynomial vanishes ) .
moreover , the convergence of the ope is factorially improved by the transform , and the contribution of the low - lying resonances in @xmath96 is enhanced for low values of the borel variables .
the symmetry of the spectral function in @xmath104 , @xmath105 suggests the choice @xmath106 .
the final sum rule reads : @xmath107 -0.3 cm the result of the numerical analysis of ( [ sr ] ) is depicted in fig .
[ th271_fig1 ] .
a stability window is observed , starting at @xmath108 and continuing towards large values of @xmath109 ; in this range the result for @xmath110 is independent of the external parameter @xmath109 .
the variation of the result with @xmath109 and with @xmath100 provides us with an estimate of the accuracy of the numerical outcome .
we find @xcite : @xmath111 a result corresponding to @xmath112 .
as for @xmath113 in ( [ otilde ] ) , since our computational scheme considers only valence quark processes , a sum rule for the matrix element in ( [ otilde ] ) would give @xmath114 .
the explicit calculation confirms this conclusion . using the computed @xmath61 and @xmath115 in the formulae in ref .
@xcite , we get : @xmath116 .
we presented a method based on the @xmath117 expansion to compute inclusive widths of heavy hadrons .
we also described the qcd sum rule calculation of the matrix element @xmath65 contributing to @xmath2 to the @xmath52 lifetime .
the result : @xmath116 implies that such correction can not explain the observed difference between @xmath3 and @xmath63 .
if the experimental data will be confirmed , a theoretical reanalysis of the problem would be required , as suggested for example in @xcite .
9 j. alcavaz et al .
( lep @xmath0 lifetime group ) , http://wwwcn.cern.ch/ claires / lepblife.html .
i. bigi et al . , in s. stone ( ed . ) @xmath0 decays , world scientific .
j. chay et al . , phys .
b 247 ( 1990 ) 399 . i. bigi et al . , phys .
b 293 ( 1992 ) 430 ; 297 ( 1993 ) 477 ( e ) .
m. neubert and c.t .
sachrajda , nucl .
b 483 ( 1997 ) 339 .
p. ball et al .
, phys . rev .
d 49 ( 1994 ) 2472 .
p. colangelo et al . , phys . rev .
d 54 ( 1996 ) 4622 .
b. guberina et al . , z. phys .
c33 ( 1986 ) 297 .
rosner , phys .
b 379 ( 1996 ) 267 . e.v .
shuryak , nucl .
b 198 ( 1982 ) 83 .
p. colangelo and f. de fazio , phys .
b 387 ( 1996 ) 371 .
g. altarelli et al . , phys .
b 382 ( 1996 ) 409 . p. colangelo et al . , hep - ph/9705390 . | we describe a theoretical approach to compute inclusive widths of heavy hadrons , based on a systematic expansion in the inverse powers of the heavy quark mass .
the method reproduces quite well the experimental ratios of @xmath0 mesons lifetimes . as for @xmath1
, we present a qcd sum rule calculation of the @xmath2 contribution to @xmath3 ; we conclude that such correction is not responsible of the small experimental value of @xmath1 . |
the concept of topological order @xcite was initiated in the 80 s with the discovery of the fractional quantum hall effect and high temperature superconductivity @xcite . since then it has also been seen in short range resonating valence bond states @xcite and in quantum spin liquids @xcite .
once it s usefulness was realized in solid state quantum computation @xcite several exactly solvable models have been constructed achieving this .
the classic example emerged when kitaev wrote down the toric code hamiltonian in two dimensions @xcite .
these systems are examples of lattice models which host anyons @xcite as part of their low energy excitations .
they can also be thought of as particular phases of the @xmath2 lattice gauge theory which host these deconfined excitations with anyonic statistics .
they posses ground states with degeneracies which are topological invariants .
this degeneracy is stable up to the addition of weak perturbations .
this feature makes this model a probable candidate for fault tolerant quantum computation @xcite .
these models have been extended to other discrete groups and involutary hopf algebras as well @xcite . in these cases
they can be thought of as arising from lattice gauge theories based on these discrete groups or involutary hopf algebras @xcite .
earlier works showing the existence of anyons in two dimensional discrete gauge theories can be found in @xcite .
discrete gauge theories emerge in these models when a continuous gauge group breaks down to one via spontaneous symmetry breaking @xcite .
these systems are usually perturbed by adding qudit terms to act on the edges of the lattice which carry the gauge degrees of freedom .
they drive the system out of the topological phase @xcite .
the resulting models are rendered unsolvable analytically and are thus subject to study using numerical methods .
however by considering restricted plaquette and vertex operators they can be made solvable .
such studies were carried out in @xcite .
this resulted in condensed phases of the quantum double model .
these works were crucial in understanding the stability of these topological phase represented by the quantum double model .
it is thus an important problem to find exactly solvable models which remain in topological phases in the presence of these perturbations . in this spirit
we introduce exactly solvable models which are constructed by looking at possible hamiltonians that can be generated using the transfer matrices of such systems .
we show exactly solvable models which include the single qudit perturbations . the cases of abelian and non - abelian groups are studied separately .
it is shown that in the abelian case the system remains in the topological phase corresponding to the quantum double model .
the situation turns out different for the non - abelian cases .
we find that the model remains in a topological phase but it is in a modified version with respect to the corresponding topological phase of the quantum double model .
these are seen by studying the examples of the group algebras of @xmath0 and @xmath1 denoted by @xmath3 and @xmath4 respectively .
we then go beyond the transfer matrices of lattice gauge theories by introducing more parameters in the transfer matrix of lattice gauge theories to find other topological phases .
this transfer matrix is made up of two - qudit operators apart from the usual operators making up the transfer matrix of lattice gauge theories .
we work with the @xmath5 case to show how one can obtain the double semion phase from such considerations . for more general groups
these transfer matrices contain the twisted quantum double models as defined in @xcite .
such models were also defined in @xcite while considering trivial global symmetry groups .
the paper is organized as follows .
section 2 describes the transfer matrices of two dimensional lattice gauge theories .
the local operators used to build these transfer matrices and the parameters used to study phase transitions between different phases is explained here .
hamiltonians of two dimensional lattice models are obtained from these transfer matrices by taking their logarithms .
this can be done in different ways resulting in several models .
these include exactly solvable models and models which can not be solved analytically .
the models of interest are presented in section 3 .
their excitations along with their braiding statistics and fusion rules are studied in section 4 .
section 5 shows the construction of the double semion model from the transfer matrix picture .
section 6 makes up our concluding remarks .
the system is defined on a two dimensional lattice @xmath6 with gauge degrees of freedom located on the links of this lattice .
the gauge degrees of freedom belong to the group algebra of a gauge group @xmath7 , denoted by @xmath8 , with basis elements @xmath9 and multiplication rule given by @xmath10 .
thus the states on the links are given by linear combinations of @xmath11 . in the case of @xmath5
this is nothing but the familiar two state system given by @xmath12 and @xmath13 .
the system is evolved in time resulting in a three dimensional manifold @xmath14 which is more precisely given by @xmath15 $ ] where @xmath16 $ ] is the unit time interval .
@xmath14 is also discretized with elements of the algebra @xmath8 living on the links .
this can be used to construct partition functions for this system by associating weights to links and faces of the closed triangulated 3-manifold @xmath14 in the spirit of state - sum models @xcite .
however when @xmath14 is not closed the tensors associated to the weights have non contracted indices at the ends of the unit time interval giving us transfer matrices . such a procedure was used to construct transfer matrices of lattice gauge theories in @xcite . in
@xcite the partition function , and hence the transfer matrix , was built out of structure constants of the algebra @xmath17 , which in this case is given by @xmath8 , as the weights associated to links and plaquettes .
apart from these structure constants the partition function is also parametrized by a pair of elements @xmath18 belonging to the center of the algebra @xmath17 and @xmath19 belonging to the center of the dual algebra @xmath20 . @xmath21 and @xmath22 denote space and time directions . thus the transfer matrix in our case is given by @xmath23 .
we do not go into the details of constructing this transfer matrix in this paper .
the details can be found in @xcite .
for the purposes of this paper we will start with the most general transfer matrix for @xmath24 with particular choices of @xmath7 and show how we can construct various models of physical interest from them .
for a general group algebra based on a group @xmath7 the transfer matrix is written as [ tg ] u((g ) , z_s , z_t , z_s^ * , z_t^ * ) & = & _ p ( _ c _ cb_p^c ) + & & _
j(_gg b_gt_j^g ) _
i ( _ r a_rx_i^r ) _ v a_v where @xmath25 , @xmath26 , @xmath27 and @xmath28 denotes plaquettes , links and vertices respectively .
the operators in eq.([tg ] ) are given by [ operatorsg ] a_v & = & _ gg_g l_i_1^gr_i_2^g^-1l_i_3^gr_i_4^g^-1 + [ operatorsbg ] b_p^c & = & _ \{g_i}(g_1 g_2 g_3 g_4,c ) t_j_1^g_1t_j_2^g_2t_j_3^g_3
^ -1t_j_4^g_4 ^ -1 + [ operatorsxg ] x_i^r & = & _ gg_r(g ) l_i^g where @xmath29 is the character of the group element @xmath30 in the irr @xmath31 , @xmath32 is the number of elements with non - zero character in the irr @xmath31 , @xmath33 $ ] is the set of conjugacy classes in @xmath7 , and the operators @xmath34 act on the states @xmath35 on the links as follows t_i^g|k & = & _ g , k|k ; t_i^g^-1|k= _ g^-1 , k|k , + l_i^g|k & = & |gk ; r_i^g^-1|k= |kg^-1 .
the operator in eq.([operatorsxg ] ) can also be defined using the @xmath36 operators instead of the @xmath37 operators .
however this does not matter as it can be shown that they are the same .
this is because the elements of a given conjugacy class have the same coefficients .
so the orientation of the lattice does not matter for the definition of these operators .
however in the operator @xmath38 the coefficients of the elements in a given conjugacy class are not the same and hence the orientation of the lattice does matter in this case .
the convention is that if the orientation of the link matches with the orientation of the plaquette then we use @xmath39 otherwise we use @xmath40 . the parameters on the two sides of eq.([tg ] )
are related by the following z_s & = & _ c_c_gc_g + z_t & = & _ r a_r _ gg_r(g ) _
g + z_s^ * & = & _ ggb_g_g + z_t^ * & = & _ gg _ g_g where @xmath41 is the basis of the algebra dual to @xmath8 that is it is dual to the basis of the algebra given by @xmath42 .
these are the most general form for the central elements of a group algebra and its dual . in this paper
we are not using @xmath43 directly .
instead , we describe the models in terms of @xmath44 and @xmath45 .
the parametrization by algebra and dual algebra elements is important in the context of @xcite but it will not play a major role here . the action of the operators in eq.([operatorsg]-[operatorsxg ] )
is shown in fig.([lattop ] ) . and plaquette operator @xmath46
are shown along with the orientations of the lattice . ]
note from the form of the terms in the transfer matrix in eq.([tg ] ) that the operators on the plaquettes @xmath25 , are the same as the plaquette operators of the quantum double model @xcite .
however the operators on the vertices @xmath28 are not the same as the star operators of the quantum double models . in order to obtain precisely the star operators of the quantum double case we need to choose @xmath47 to be z_t^
* = _ r g_r_gg_r(g)_g .
this results in the following vertex operator a_v = _ r g _
r a_v^r with a_v^r = _
gg_r(g ) l_i_1^gr_i_2^g^-1l_i_3^gr_i_4^g^-1 .
the operators @xmath48 commute with @xmath46 . when the irr @xmath31 is the identity representation we obtain the star operator projecting to the vacuum sector .
this is the operator which appears in the hamiltonian of the quantum double model @xcite .
note that these subtleties do not matter for the abelian models .
they only make a difference in the non - abelian instances .
the operators acting on the links @xmath27 , @xmath26 in eq.([tg ] ) can be thought of as magnetic field terms .
such terms were considered in @xcite to induce condensations of quasiparticle excitations leading to phase transitions while remaining in the exactly solvable regime .
we now write down the transfer matrices for the examples we consider in this paper .
the hamiltonian is given by [ sidesg ] h = _ v a_v^g + _
c_c_p b_p^c where @xmath46 is the projector to a conjugacy class @xmath81 of @xmath7 and is given by b_p^ c = _ gc b_p^g and @xmath82 is a real parameter .
the modified vertex operator @xmath83 is labeled by an element @xmath84 .
it is given by a_v^g & = & l_i_1^gr_i_2^g^-1l_i_3^gr_i_4^g^-1 + & + & l_i_1^g^-1r_i_2^gl_i_3^g^-1r_i_4^gwith the complex parameters @xmath85 and @xmath86 satisfying a_h_1a_h_1^ * & = & a_h_2a_h_2^ * , h_1 , h_2 g h_1h_2 + b_h_1b_h_1^ * & = & b_h_2b_h_2^ * , h_1 , h_2 g h_1h_2 + b_gu & = & b_u , ug . the choice of parameters depend on the orientation of the @xmath87 lattice . the oriented lattice is shown in fig([lattop ] ) .
every time the arrow goes into the vertex we use the @xmath86 set of parameters .
the @xmath85 set of parameters is used when the arrow goes away from the vertex . with this definition
the model can be defined on a @xmath87 lattice with an arbitrary triangulation . for these choice of parameters
the vertex operators commute for adjacent vertices .
the commutation between the modified vertex operator @xmath83 and the plaquette operator @xmath46 projecting to the conjugacy class @xmath81 is non - trivial . for this proof
we need not consider the @xmath88 as it trivially commutes with the plaquette operator .
the proof for the remaining part of the vertex operator goes as follows . consider the plaquette @xmath25 and the vertex @xmath28 shown in oriented square lattice of fig([proof ] ) . ) and eq.([rhs ] ) . ]
they share two links labeled @xmath27 and @xmath26 in the fig([proof ] ) . according to the orientation of the links shown in the fig([proof ] )
we have [ lhs ] a_v^gb_p^c & = & ( l_i^gl_j^g r_j_1^g^-1r_j_2^g^-1 ) + & = & ( r_j_1^g^-1r_j_2^g^-1 ) and b_p^ca_v^g[rhs ] & = & ( l_i^gl_j^g r_j_1^g^-1r_j_2^g^-1 ) + & = & ( r_j_1^g^-1r_j_2^g^-1 ) + & = & ( r_j_1^g^-1r_j_2^g^-1 ) where we have used the relation t^h_il^g_i = l^g_it^g^-1h_i .
thus the shifted summation in eq.([rhs ] ) only shuffles the terms in the projector .
this shows that the two summations in eq.([lhs ] ) and eq.([rhs ] ) are the same implying @xmath83 commutes with @xmath46 .
we can use a similar proof to show that the other three vertex operators commute with the plaquette operator .
thus the hamiltonian in eq.([sidesg ] ) is exactly solvable .
note that we have not used the full projectors for the modified vertex operators .
this can be simply understood by looking at the identity term in the expression for the projectors .
the identity term coupled with the perturbations is like adding magnetic fields to the system
. this will take us out of the exactly solvable regime we are interested in and also we will no longer remain in a topological phase .
thus we avoid full projectors for the modified vertex operators .
we can write down a similar model where we modify the plaquette operator with the @xmath89 operators defined in eq.([operatorsxg ] ) . in the examples to follow
we will write down these models for the @xmath3 case alone .
the corresponding models for the non - abelian cases are more cumbersome and so we do not show them here . before we write down the specific examples
let us look at the deformed quantum double algebra .
consider the product of two deformed vertex operators @xmath90 and @xmath91 .
it is given by a_v^g_1a_v^g_2 = a_v^g_1g_2 m(g_1 , g_2 ) , with m(g_1 , g_2 ) = .
we can now check associativity of the product by comparing @xmath92 and @xmath93 which requires [ mc]m(g_1g_2 , g_3)m^g_3(g_1 , g_2 ) = m(g_1 , g_2g_3)m(g_2 , g_3 ) , which is seen to be true by using the expression for @xmath94 and from m^g_3(g_1 , g_2 ) = . the quantum double algebra given by a^g_vb_p^h = b^ghg^-1_pa_v^g , is unchanged but the product in the algebra is changed to ( a_v^g_1b_p^h_1)(a_v^g_2b_p^h_2 ) = _ h_2 , g_2 ^ -1h_1g_2a_v^g_1g_2m(g_1 , g_2 ) b_p^h_2 .
this algebra is similar to quasi quantum doubles as seen in @xcite where the deformation occurs through a 2-cocycle .
such algebras also have a non - coassociative coproduct with an associator given by a 3-cocycle related to the 2-cocycle by a slant product @xcite .
the condition in eq.([mc ] ) is similar to the one obtained in studying obstructions to implementations of global symmetries on physical systems as used in analysis of symmetry protected topological phases @xcite .
note that we have not studied the co - algebra structure of these operators here .
it is not necessary for what is to follow and we stop with the above crucial remarks .
we now look at the simple example of @xmath49 to illustrate the form of the transfer matrix defined in eq.([tg ] ) .
[ tz2 ] u((_2 ) , z_s , z_t , z_s^ * , z_t^ * ) & = & _ p ( _ 1 b_p^1 + _ -1 b_p^-1)_j ( b_1 t_j^1 + b_-1 t_j^-1 ) + & & _ i ( a_1 x_i^1 + a_-1 x_i^-1 ) _ v(_1 a_v + _
-1 a_v^-1 ) where @xmath28 , @xmath25 , @xmath27 and @xmath26 label vertices , plaquettes and links respectively .
the operators @xmath50 , @xmath51 , @xmath52 and @xmath53 are all projectors and are given by a_v^1 & = & + b_p^1 & = & + x_i^1 & = & + t_j^1 & = & .
these are the operators appearing in the toric code @xcite system and their action is shown in the figure([lattop ] ) . the parameters on both sides of eq.([tz2 ] )
are related in the following way z_s & = & _ 1 _ 1 + _ -1_-1 + z_t & = & ( ) _ 1 + ( ) _ -1 + z_s^ * & = & b_1 _ 1 + b_-1 _ -1 + z_t^ * & = & ( ) _ 1 + ( ) _ -1 where @xmath54 and @xmath55 are the basis elements of @xmath5 and its dual respectively .
the hamiltonian in this case is given by [ sides2v ] h = _ v a_v + _
p(_1 b_p^1 + _ -1b_p^-1 ) with @xmath95 given by a_v & = & where the complex parameters @xmath96 have the same modulus .
it is easy to check that the model is exactly solvable as the terms commute with each other . to compute the spectrum of the hamiltonian we need to find the eigenvalues of the modified vertex operator @xmath95 and the plaquette operators .
each plaquette operator is a sum of two projectors and so the eigenvalues of that operator depends on the magnitude of @xmath97 .
the eigenvlaues of @xmath95 are given by @xmath98 .
the dual of the above model in eq.([sides2v ] ) is given by the following hamiltonian [ sides2 ] h = _
v(_1 a_v^1 + _
-1a_v^-1 ) + _ p b_p where b_p = ( b_p^1 - b_p^-1)_i_kp , k=1 ^ 4(a_1 x_i_k^1 + a_-1x_i_k^-1 ) + h.c . with @xmath99 being the perimeter of the plaquette @xmath25 .
this model can be defined on any two dimensional lattice for an arbitrary triangulation .
on the square lattice this plaquette operator continues to be a four spin interaction term as in the usual quantum double model .
it is easy to check that for two adjacent plaquettes @xmath100 and @xmath101 , these operators commute and that @xmath102 .
this operator clearly commutes with the vertex operators as can be easily seen from the expression for @xmath103 .
thus the hamiltonian in eq.([sides2 ] ) continues to be an exactly solvable model .
the hamiltonian in eq.([sides2 ] ) is got by the following choice of the parameters in the transfer matrix of eq.([tz2 ] ) : z_s^ * & = & _ 1 + _ -1 + z_t^ * & = & ( ) _ 1 + ( ) _ -1 .
the values of the other two parameters @xmath104 and @xmath105 can be found but after a cumbersome calculation involving ten variables .
we do not show this computation here but just remark that it is possible to find these parameters as well . though this computation may seem irrelevant it is an important calculation to show that the semion model can not be obtained from the transfer matrix in eq.([tz2 ] ) .
we will make more comments about this in the remarks section .
in general it is important to consider the orientation of the two dimensional triangulated lattice .
there are orientations for the links and the plaquettes of this lattice .
these are shown in fig([lattop ] ) .
for the @xmath5 case the orientations are not important as the inverse of the group elements in this case is the same as the group element itself
. the ground states @xmath131 for the hamiltonian in eq.([sides2 ] ) is given by the following conditions a_v^1|g= |g , a_v^-1|g= 0 , b_p|g= -|g ; v , p.to validate these conditions we assume the relative magnitudes of the parameters in the hamiltonian in eq.([sides2 ] ) according to the conditions stated .
for the @xmath5 case we have the following operators create the flux ( violations of the plaquette operator conditions ) and charge ( violation of the vertex operator conditions ) .
[ excite2 ] v_^*^f & = & _ i^ * _ i^x + v_^c & = & _ i
_ i^z_j(n_1x_j^1 + n_-1x_j^-1 ) where @xmath132 and @xmath133 are strings in the direct and dual lattice respectively . the parameters @xmath134 and @xmath135 satisfy @xmath136 .
this condition makes the operator creating the flux independent of @xmath137 .
the dyonic excitations is created by applying both these strings on the dual and direct lattice respectively .
the statistics of these excitations of these particles can easily be found out using the operators in eq.([excite2 ] ) .
the self statistics of fluxes are trivial as the operator creating the flux is the same as in the toric code case
. hence they remain bosonic .
the self statistics of charges is also seen to be bosonic as ^z_i(n_1 x_i^1 + n_-1x_i^-1)^z_i(n_1 x_i^1 + n_-1x_i^-1 ) = 1 .
the mutual statistics between the charges and the fluxes is clearly fermionic as can be trivially seen from the expressions of the operators creating the charges in eq.([excite2 ] ) . for convenience
we now label the particles in this model as 1 , @xmath138 , @xmath139 and @xmath140 .
1 is the vacuum , @xmath138 is the charge , @xmath139 is the flux and @xmath140 is the dyonic excitation .
the fusion rules can be obtained from the algebra of the orators creating them in eq.([excite2 ] ) , and they are found to be [ fusion2 ] & = & mm= = 1 + m & = & and the remaining fusion rules involving the fusion of the vacuum are all trivial . thus we find the fusion rules to be the same as the ones for the @xmath5 anyon model .
the elements of the group @xmath0 are denoted by @xmath56 with @xmath57 , the elements of the group algebra @xmath3 and its dual are denoted by @xmath58 and @xmath59 respectively . in this case
the transfer matrix is written as [ tzn ] u((_n ) , z_s , z_t , z_s^ * , z_t^ * ) & = & _ p(_k=0^n-1_^k b_p^^k)_j(_k=0^n-1b_^kt_j^^k ) + & & _ i(_k=0^n-1a_^kx_i^^k)_v(_k=0^n-1_^k
a_v^^k ) where the index @xmath60 labels the irr of @xmath0 .
the operators @xmath61 , @xmath62 , @xmath63 and @xmath64 are given by a_v^^k & = & _ l=0^n-1_^k(^l)x_i_1^lx_i_2 ^ -lx_i_3^lx_i_4 ^ -l + b_p^^k & = & _ l=0^n-1_^k(^l)z_j_1^lz_j_2^lz_j_3 ^ -lz_j_4 ^ -l + x_i^^k & = & _ l=0^n-1_^k(^l)x_i^l + t_j^^k & = & _ l=0^n-1 _ ^k(^l)z_j^l where @xmath65 and @xmath66 are single qudit operators generating @xmath0 and are defined by [ genxn ] x_i |^k & = & [ genzn ] z_j|^k & = & ^k|^k .
@xmath67 is the character of the element @xmath56 in the irr labeled by @xmath60 . the parameters on the two sides of eq.([tzn ] )
are related in the following way z_s & = & _ k=0^n-1_^k_^k + z_t & = & _ k=0^n-1(_l=0^n-1_^k(^l)a_^l)_^k + z_s^ * & = & _ k=0^n-1b_^k_^k + z_t^ * & = & _ k=0^n-1(_l=0^n-1_^k(^l)_^l)_^k .
the hamiltonian in this case is given by [ sidesnv ] h = _ v ( a_v + h.c . )
+ _ p ( _ k=0^n-1_^kb_p^^k ) with @xmath95 given by a_v & = & & & _ i_k , kp , k=(1,2 ) , k=(3,4)(_j=0^n-1a_^jt_i_k^^j)(_j=0^n-1b_^jt_i_k^^j ) where the parameters @xmath106 and @xmath107 satisfy @xmath108 for @xmath109 and @xmath110 , @xmath111 for @xmath112 .
these operators commute for adjacent vertices when the parameters satisfy these conditions .
it is also easy to check that @xmath113 and hence it has the nth roots of unity as its eigenvalues .
as the hermitian conjugate is added to @xmath95 the eigenvalues are real . the dual version of the model in eq.([sidesnv ] )
is given by the hamiltonian [ sidesn ] h = _ v(_k=0^n-1 _ ^ka_v^^k ) + _ p ( b_p+h.c . ) where @xmath103 is given by b_p & = & ( _ k=0^n-1 ^n - k b_p^^k ) + & & _ i_k , kp , k=(1,2 ) , k=(3,4 ) ( _ j=0^n-1 a_^jx_i_k^^j)(_j=0^n-1 b_^jx_i_k^^j ) with the solutions for the @xmath114 s given by the relations @xmath115 .
the modulus of these parameters are the same as in their dual version .
these operators commute for adjacent plaquettes and @xmath116 . for the @xmath3 case ,
the hamiltonian is given by eq.([sidesn ] ) .
the conditions for the ground states are similar to the @xmath141 case .
again we assume the parameters to be such that these conditions hold .
as in the @xmath141 case the fluxes do not change but the operators creating the charges have to be modified .
there are @xmath142 charges and we denote them by @xmath143 with @xmath144 .
the operators creating them are given by [ chargen]v_^e^n - k = z^n - k(_j=0^n-1m_(n - k)^jx^^j ) where @xmath132 is a direct triangle @xcite and the parameters @xmath145 satisfy [ recurn]a_^j - km_(n - k)^j = m_(n - k)^j - n+1a_^j for @xmath146 .
at this point there seems to be one free parameter among the @xmath145 s for a given charge @xmath143 .
the remaining are found in terms of this single `` free '' parameter using the recursion relations in eq.([recurn ] ) .
however this is not there once we consider the fusion rules .
they consistently fix all the parameters @xmath145 in the string .
the string operators for the fluxes are the same as in the @xmath3 quantum double model and are given by [ fluxn ] v_^*^f^k = x_j^k where @xmath133 is a dual triangle and @xmath26 is the label for this edge .
it is easy to check using the string operators in eq.([chargen ] ) and eq.([fluxn ] ) that the statistics of these particles are the same as in a @xmath0 anyon model .
the charges and fluxes have bosonic self - statistics .
the mutual statistics are not trivial and they are given by e^n - k_1f^k_2 = ^(n - k_1)k_2f^k_2e^n - k_1 .
the fusion rules of the fluxes do not change .
the fusion rules of the charges are the same when we constrain the free parameters of the parameters @xmath145 .
this constrain is given by m_(n - k_1)^j - k_2m_(n - k_2)^j = m_(n-(k_1+k_2))^j for @xmath147 , @xmath148 belong to @xmath149 and @xmath146 .
this relation comes when we impose the @xmath0 fusion rules for the charges for the operators creating the charges in eq.([chargen ] ) which is e^n - k_1e^n - k_2 = e^n-(k_1+k_2 ) .
these fix all the @xmath145 parameters unambiguously .
as we have shown the existence of all the charges and fluxes of the @xmath3 quantum double model exist in our model they are in the same phase
. a similar construction can be carried out when the vertex operator is modified and the flux operator is left unchanged as given in eq.([sidesnv ] ) .
this model is also in the same phase as the @xmath3 quantum double phase .
the group @xmath68 .
it has three conjugacy classes , @xmath69= \left([1 ] , [ r ] , [ \tau]\right)$ ] , where @xmath70=\{1\}$ ] , @xmath71 = \{r , r^2\}$ ] and @xmath72 = \{\tau , \tau r , \tau r^2\}$ ] .
we denote the group elements by @xmath30 , the elements of the group algebra and its dual by @xmath73 and @xmath74 respectively .
the transfer matrix is given by [ ts3 ] u((s_3 ) , z_s , z_t , z_s^ * , z_t^ * ) & = & _ p ( _ c _ cb_p^c ) + & & _
j(_gs_3 b_gt_j^g ) _
i ( _ r a_rx_i^r ) _ v a_v with [ operatorss3]a_v & = & _ gs_3_g l_i_1^gr_i_2^g^-1l_i_3^gr_i_4^g^-1 + b_p^c & = & _ \{g_i}(g_1 g_2 g_3 g_4,c ) t_j_1^g_1t_j_2^g_2t_j_3^g_3 ^ -1t_j_4^g_4 ^ -1 + [ operatorsxs3 ] x_i^r & = & _ gs_3_r(g ) l_i^g .
the operators @xmath75 are projectors to the various elements @xmath76 .
they are given by the following relations t^1 & = & + a + z_11 + ( z^2)_11 + t^r & = & + a + ^2z_11 + ( z^2)_11 + t^r^2 & = & + a + z_11 + ^2 ( z^2)_11 + t^ & = & - a + z_22 + ( z^2)_22 + t^r & = & - a + ^2z_22 + ( z^2)_22 + t^r^2 & = & - a + z_22 + ^2 ( z^2)_22 where & = & ( cc 1 & 0 + 0 & 1 ) ; a = ( cc 1 & 0 + 0 & -1 ) + z_11 & = & ( cc z & 0 + 0 & 0 ) ; ( z^2)_11 = ( cc z^2 & 0 + 0 & 0 ) + z_22 & = & ( cc 0 & 0 + 0 & z ) ; ( z^2)_22 = ( cc 0 & 0 + 0 & z^2 ) where each of the matrices is divided in blocks of three by three matrices . 1 denotes the three by three identity and @xmath77 is the generator of @xmath78 defined in eq.([genzn ] ) . the parameters on the two sides of eq.([ts3 ] )
are related by z_s & = & _ [ 1]_1 + _ [ r](_r + _
r^2 ) + _ [ ] ( _ + _ r + _ r^2 ) + z_t & = & _ rs_3 a_r _ gs_3_r(g ) _
g + z_s^ * & = & _ gs_3b_g_g + z_t^ * & = & _ gs_3 _ g_g . the transfer matrix in eq.([tg ] ) can be used to obtain the hamiltonians through u((g ) , z_s , z_t , z_s^ * , z_t^ * ) = e^-h where @xmath79 is the hamiltonian . in the next section
we will see that there are several ways of grouping terms in the transfer matrices while taking their logarithms which result in different hamiltonians for a given set of parameters .
the group @xmath1 has six elements given by @xmath117 with @xmath118 and @xmath119 .
we choose @xmath30 to belong to a normal subgroup @xmath120 of @xmath1 .
the only normal subgroup of @xmath1 is given by @xmath121 .
thus the modified vertex operator has either @xmath122 or @xmath123 .
there are three conjugacy classes for @xmath1 given by @xmath69 = \left(\{1\ } , \{r , r^2\ } , \{\tau , \tau r , \tau r^2\}\right)$ ]
. the hamiltonian is given by [ sides3v ] h = _ v ( a_v^r + h.c . ) + _ p_c_c(_kcb_p^k ) with @xmath124 given by a_v^r = l^r_i_1(_hs_3 a_ht^h)r^r^2_i_2(_hs_3 b_ht_i_2^h)l^r_i_3(_hs_3 a_ht_i_3^h)r^r^2_i_4(_hs_3 b_ht_i_4^h ) with the parameters satisfying the following b_r & = & b_1 ; b_r^2 = b_1 + b_r^2 & = & b _ ; b_r = b _ and they have the same modulus .
the eigenvalues of @xmath124 are found by computing its cube as @xmath125 .
this gives the eigenvalues as @xmath126 and @xmath127 .
for the chosen parameters the model is solvable as all the terms commute with each other .
we can define exactly solvable models by choosing @xmath30 to be either @xmath128 , @xmath129 , or @xmath130 .
these models are in another phase which we do not discuss here . in the next section we find out the quantum phases in which these models live in by studying their excitations .
the quantum phase of this case is not in the phase of the corresponding quantum double model .
this is because some of the particles in the quantum double phase condense and some others do not exist .
we will write down which ones condense and which ones do not exist for the @xmath1 example in what follows .
we will work with the vertex version of the model given by eq.([sides3v ] ) .
the @xmath1 anyon model consists of eight excitations including the vacuum @xcite .
they are given by @xmath150 1 is the vacuum , @xmath151 and @xmath152 are the abelian and non - abelian charges respectively . @xmath153 and @xmath154 are the non - abelian fluxes in the model .
@xmath155 , @xmath156 and @xmath157 are the dyonic excitations . in the notation of @xcite these excitations
are given by @xmath158 we will switch to this notation when we discuss the fusion rules in this model .
the operators creating these excitations , called ribbon operators , can be found in @xcite .
we do not write all these operators here . we will only write down operators which will be modified in our model .
the abelian charge @xmath151 condenses as the ribbon operator creating this excitation commutes with both the plaquette and the modified vertex operators .
the ribbon operator creating the non - abelian charge @xmath152 is not modified . the non - abelian flux @xmath153 is created by the following modified operators . [ modfl3 ] _
; id(r)^(1,1)(1,1 ) & = & f _ ; id(r)^(1,1)(1,1)(_hs_3 n_ht^h ) + _ ; id(r)^(2,1)(2,1 ) & = & f _ ; id(r)^(2,1)(2,1)(_hs_3 m_ht^h ) where the parameters are given by n_r & = & n_1 ; n_r^2 = n_1 + n_r & = & n _ ; n_r^2 = n_and m_r^2 & = & m_1 ; m_r = m_1 + m_r^2 & = & m _ ; m_r = m_. at this point @xmath134 , @xmath159 , @xmath160 and @xmath161 seem to be free parameters .
however the ribbon operators should not depend on these free particles as this will indefinitely increase the number of particles .
we will see that the fusion rules fix these parameters unambiguously .
the non - abelian flux @xmath154 does not exist in this case as if we attempt to modify the corresponding ribbon operator to commute with the modified vertex operator we will be forced to constrain the parameters @xmath85 which eventually makes the model trivial in the sense they remove the additional qudit terms on the sides .
hence this changes the model and the phase .
so we conclude that in this phase this flux excitation does not exist or they create states which can not be expanded in the basis of the excitations of the condensed model .
this is not surprising as even in the @xmath4 quantum double model we can write operators which after acting on the ground states create states that can not be expanded in the basis of the eight excitations of the @xmath1 anyon model .
one such operator is given by the following matrix o = ( cc 0 & z + 0 & 0 ) .
it can be easily checked that this matrix does not commute with both the plaquette and vertex operators of the quantum double model of @xmath4 .
it can also be checked that this matrix can not be expanded in terms of the matrices creating the eight excitations of the @xmath4 quantum double model .
the operator creating the flux excitation @xmath154 faces a similar situation in our model .
the non - existence of this flux also means that the dyonic excitation @xmath157 does not exist in this model .
on the other hand the dyonic excitations given by @xmath155 and @xmath156 do exist in this model and they are created by the modified operators creating flux excitations and the unchanged charge operators .
we now fix the remaining `` free '' parameters @xmath134 , @xmath159 , @xmath160 , @xmath161 using appropriate fusion rules .
the fusion rules of the @xmath1 anyon model can be found in @xcite .
as we do not have the excitations denoted by @xmath162 and @xmath163 we can ignore their fusion rules .
they are decoupled from the fusion rules of the other particles . the fusion rule which will be useful to fix the remaining parameters
is given by j^xj^x = 1 + a + j^x.the operators creating the excitation @xmath164 are not modified in our model and hence we can not use this channel to fix the remaining parameters .
we use the other two channels instead .
so we have _ ;
id(r)^(1,1)(1,1 ) _ ; id(r)^(2,1)(2,1 ) = 1.this gives us the conditions m_1n_1 = ; m_n_= .
the second channel is obtained by _ ; id(r)^(1,1)(1,1 ) _ ; id(r)^(1,1)(1,1 ) = _ ; id(r)^(2,1)(2,1 ) which gives us the conditions m_1 = n_1 ^ 2 ; m_= n_^2.these conditions unambiguously fix all the parameters in the ribbon operators of eq.([modfl3 ] ) .
thus we have seen that this model is in a quantum phase different from the @xmath4 quantum double model .
the hamiltonians can be got from the transfer matrix @xmath80 by taking its logarithm .
there are several ways to do this and in general we can obtain many complicated models with several spins interacting .
however for simplicity we will only consider four spin couplings . unless specified we will work only on the square lattice .
however the models we define are not sensitive to the details of the lattice in the sense that the phase they describe is the same irrespective of the choice of the lattice .
the first two operators on the vertices and plaquettes are four spin interaction terms by construction .
the operators on the links @xmath27 , @xmath26 are not interaction terms between spins but magnetic field terms or perturbations @xcite as noted previously .
these terms do not commute with the vertex and plaquette operators in general making the procedure for taking logarithms in their presence cumbersome .
we now write down exactly solvable models with vertex operators modified by the perturbations and which continue to remain in a topological phase .
this model is first written for a general group algebra @xmath8 and then we discuss the abelian and non - abelian cases separately .
we find out the quantum phase for the three examples considered .
the transfer matrices considered so far were made up of two kinds of operators , the plaquette and vertex operators , given by @xmath46 and @xmath165 , which coupled four qudits and the single qudit operators , given by @xmath166 and @xmath167 .
such operators arise in a transfer matrix parametrized by the centers of the algebra @xmath17 and its dual @xmath20 .
we denoted these transfer matrices as @xmath168 where @xmath17 was taken to be the group algebra @xmath8 .
the parameters @xmath104 , @xmath105 , @xmath169 and @xmath47 are the centers of the algebra and its dual respectively . in the pictorial language of
@xcite these parameters were associated to the spacelike plaquettes , timelike plaquettes , spacelike links and timelike links respectively .
these are the operators which result in the transfer matrix of a lattice gauge theory and we studied the quantum phases we can produce using the hamiltonians from these transfer matrices .
in particular we obtained the quantum double phase of abelian gauge groups from non - trivial hamiltonians constructed by modifying these transfer matrices .
for example in the @xmath2 case we obtained the toric code phase .
however it is not possible to obtain the double semion phase from this modified transfer matrix . in order to obtain this
we need to consider more general transfer matrices made up of other kinds of operators .
such operators can be obtained by considering more general transfer matrices parametrized by the centers of not just the algebra @xmath17 but by two copies of the algebra @xmath170 and its dual which result in operators coupling two qudits .
a transfer matrix of this sort can be formally written as @xmath171 where @xmath172 and @xmath173 are the centers of @xmath170 and its dual respectively . in the language of the pictorial formalism introduced in @xcite
, these parameters are associated to the volumes and the vertices of the triangulated three dimensional lattice respectively .
we will not go into the details of this pictorial formalism in this paper and only show the hamiltonians that we can obtain from such a transfer matrix .
it is possible to build such transfer matrices made up of the usual four qudit plaquette and vertex operators as well as two and three qudit operators which are combinations of @xmath166 and @xmath167 , acting on different links , coupled together .
such considerations will lead to both the toric code phase and the double semion phase as we shall see now .
we note that such transfer matrices are not the transfer matrices of the usual lattice gauge theory based on discrete groups .
we only work with the case of @xmath49 in what follows .
we will work on a lattice where there are vertices that have valency three .
the modified vertex operator acting on this vertex is shown in fig([vertex ] ) .
in general the lattice is oriented but as we are working with the @xmath2 case in this section we ignore the orientation .
the operators coupling two links together are given by @xmath174 , @xmath175 and @xmath176 using these three two link operators we can write down the modified vertex operator as [ z2d ] a_v^-1 & = & x_l_1x_l_2x_l_3 + & & ( a_1 t_l_1 ^ 1t_l_2 ^ 1t_l_3 ^ 1 + a_2 t_l_1 ^ 1t_l_2 ^ 1t_l_3 ^ -1 + a_3 t_l_1 ^ 1t_l_2 ^ -1t_l_3 ^ 1 + a_4 t_l_1 ^ 1t_l_2 ^ -1t_l_3 ^ -1 .
a_5 t_l_1 ^ -1t_l_2 ^ 1t_l_3 ^ 1 + a_6 t_l_1 ^ -1t_l_2 ^ 1t_l_3 ^ -1 + a_7 t_l_1 ^ -1t_l_2 ^ -1t_l_3 ^ 1 + a_8 t_l_1 ^ -1t_l_2 ^ -1t_l_3 ^ -1 ) . the relation between @xmath177 s and @xmath114 s are given by a_1 & = & b_11b_21b_31 , a_2 = b_11b_22b_33 , a_3 = b_12b_23b_31 , a_4 = b_12b_24b_33 + a_5 & = & b_13b_21b_32 , a_6 = b_13b_22b_34 , a_7 = b_14b_23b_32 , a_8 = b_14b_24b_34 .
for adjacent vertex operators to commute the parameters have to satisfy the following relation = = = .thus there are four free parameters given by @xmath178 , @xmath179 , @xmath180 and @xmath181 . in order to have @xmath182 we further restrict them with the relation @xmath183 .
thus there are just three free parameters now .
the plaquette operator is unchanged and is that of the toric code .
we now obtain the excitations in this model .
the operator creating the vertex excitations that is the charges is unchanged with respect to the usual toric code .
the change is seen in the operators creating the fluxes .
they have to be modified in order for them to commute with the modified vertex operator given in eq.([z2d ] ) .
consider the following operator creating the flux on the two plaquettes adjacent to link @xmath184 as shown in the fig([triangle ] ) . and
the two plaquettes adjacent to it . ]
[ f2d ] f_l_1 & = & x_l_1 + & & ( n_1 t_l_1 ^ 1t_l_2 ^ 1t_l_3 ^ 1 + n_2 t_l_1 ^ 1t_l_2 ^ 1t_l_3 ^ -1 + n_3 t_l_1 ^ 1t_l_2 ^ -1t_l_3 ^ 1 + n_4 t_l_1 ^ 1t_l_2 ^ -1t_l_3 ^ -1 .
n_5 t_l_1 ^ -1t_l_2 ^ 1t_l_3 ^ 1 + n_6 t_l_1 ^ -1t_l_2 ^ 1t_l_3 ^ -1 + n_7 t_l_1 ^ -1t_l_2 ^
-1t_l_3 ^ 1 + n_8 t_l_1 ^ -1t_l_2 ^ -1t_l_3 ^ -1 ) + & & ( n_1 t_l_1 ^ 1t_l_2^1t_l_3^1 + n_2 t_l_1 ^ 1t_l_2^1t_l_3^-1 + n_3 t_l_1 ^ 1t_l_2^-1t_l_3^1 + n_4 t_l_1 ^ 1t_l_2^-1t_l_3^-1 .
n_5 t_l_1 ^ -1t_l_2^1t_l_3^1 + n_6 t_l_1 ^ -1t_l_2^1t_l_3^-1 + n_7 t_l_1 ^ -1t_l_2^-1t_l_3^1 + n_8 t_l_1 ^ -1t_l_2^-1t_l_3^-1 ) . for this operator to commute with the vertices @xmath185 and @xmath186 we have the following relations
n_5 = n_1 k , n_6 = n_2 k , n_7 = n_2 k , n_8 = n_1 k , with @xmath187 .
the relation between @xmath134 and @xmath188 is fixed by the condition that @xmath189 commutes with the vertex operators associated to @xmath28 and @xmath190 .
this is given by n_1 ^ 2
= n_2 ^ 2.we now study the two cases separately .
when @xmath191 we have the following form for the operator @xmath189 given by [ f1 ] f_l_1 = n_1 ^ 2 x_l_1 where @xmath192 and @xmath77 are pauli matrices acting on the link @xmath184 .
this string operator can not create the semion phase and hence remains in the toric code phase .
this can be seen by computing the fusion rules and the braiding statistics .
the fusion rule for this operator is given by f_l_1f_l_1 = n_1 ^ 4k^2.by choosing @xmath193 we obtain the fusion rules for the toric code phase .
let us work on the hexagonal lattice as shown in fig([braiding ] ) for computing the braiding rules .
we consider two sets of fluxes created by the operators @xmath194 and @xmath195 respectively .
the braiding is carried out as shown in fig([braiding ] ) and is given by [ gb ] f_f_ & = & x_12x_13x_43x_23x_13x_14 + & & nt^23_0,2,3 nt^13_0,1,3 nt^23 , 13_2 , 3 , 4 nt^13 , 14_2 , 1 , 4 nt^13 , 14_2 , 1 , 4 nt^13_1 , 3 , v nt^14_1 , 4 , v + & & nt_1 , 3 , v^
nt_1 , 2 , v^
nt_2 , 1 , 4 nt_2 , 3 , 4 nt_v^ , 1 , 3 nt_v^ , 4 , 3 where @xmath196 is the part of @xmath195 that intersects @xmath194 .
[ bg ] f_ f _ & = & x_23x_13x_14x_12x_13x_43 + & & nt^13_1,3 , v^
nt^12_1,2 , v^
nt^12 , 13_2 , 1 , 4 nt^13 , 34_2 , 3 , 4 nt^13_v^ , 1 , 3 nt^34_v^ , 4 , 3 + & & nt_0 , 2 , 3 nt_0 , 1 , 3 nt_2 , 3 , 4 nt_2 , 1 , 4 nt_1 , 3 , v nt_1 , 4 , v. the notation requires explanation .
the vertices are numbered as shown in fig([braiding ] ) . with this numbering @xmath197 , for example , denotes the @xmath192 pauli matrix acting on the link between vertices numbered 1 and 2 .
we also have nt_0 , 2 , 3 nt_0 , 1 , 3 & = & + & & .
and we have nt^23_0,2,3 nt^13_0,1,3 & = & + & & .
it is easy to see from the expression for the operator creating the flux in eq.([f1 ] ) that the two equations eq.([gb ] ) and eq.([bg ] ) are the same .
hence the flux excitation is bosonic . for the case when @xmath198 we have the following operator that creates the flux excitation .
[ f2 ] f_l_1 = n_1 ^ 2 x_l_1z_l_2z_l_3z_l_2z_l_3. the fusion rule for this operator is the same as the fusion rule of the operator in eq.([f1 ] ) .
it is easy to check using the expression for the flux operator in eq.([f2 ] ) in eq.([bg ] ) and eq.([gb ] ) that the flux operator is bosonic in this case as well . in order to obtain the double semion phase we need to use the following form for the modified vertex operator [ asem ] a_v = _ l_1 , l_2 , l_3\{1 , -1 } x_l_1x_l_2x_l_3 ( l_2 ^ -1l_1 ^ -1 , -l_2 , -1)(-l_2 , -1 , l_3)(-l_1 , -1 , l_3)^-1t_l_1^l_1 t_l_2^l_2 t_l_3^l_3where @xmath199 is an element of the cohomology group @xmath200
the cohomology group @xmath200 is given by @xmath2 .
hence there are two cocycles one of which is the trivial cocycle .
it is easy to see that for the trivial cocycle the operator @xmath165 in eq.([asem ] ) reduces to the operator of the toric code . for the non - trivial cocycle @xmath201 , the rest are equal to the identity element .
the operator in eq.([asem ] ) is precisely the same as the one used to define the twisted quantum double model in @xcite for the @xmath5 case .
the two - link operators that are used to obtain the modified vertex operator in eq.([asem ] ) are given by @xmath202,\ ] ] @xmath203\ ] ] and @xmath204.\ ] ] the expressions for the modified flux operator on a single link @xmath184 is given by f_l_1 = x_l_1where @xmath205 and @xmath206 are functions of the variables on the links @xmath184 , @xmath207 , @xmath208 , @xmath209 and @xmath210 .
the action of this string is shown in fig([triangle ] ) .
using the convention of @xcite we order the vertices as follows @xmath211 .
the operator creating the flux , @xmath189 has to commute with @xmath165 , @xmath212 , @xmath213 and @xmath214 .
this imposes the following conditions on the coefficients @xmath205 and @xmath206 & = & + & = & + & = & 1 . using these conditions it is easy to check that the fusion and braiding rules for the operator creating the fluxes is that of the double semion phase .
we have used modified transfer matrices of 3d lattice gauge theories with discrete gauge groups to find exactly solvable quantum models in two dimensions .
the abelian cases were studied with the example of the @xmath0 groups and were shown to generate a quantum phase identical to the quantum double phase based on the group algebra of these abelian groups .
this was seen by studying the excitations of these systems .
the excitations were the same as the @xmath0 anyon models .
the only change was in the operator creating these excitations . in the case
when the vertex operators were modified the flux excitations were created by new operators whereas in the case where the plaquette operators were modified the operators creating the charge excitations were modified .
thus these models present another example of a system hosting @xmath0 anyons on the two dimensional lattice .
there have been other examples of exactly solvable lattice models which host the @xmath0 anyons @xcite .
the important point about the models presented in this paper is that they involve terms which are usually added as perturbations on the links to induce quantum phase transitions of the @xmath0 quantum double models .
the models remain solvable in spite of including these terms .
the only difference is that the modified vertex and plaquette operators obey a deformed version of the quantum double algebra .
it is interesting question to find out if a suitable choice of a representation category can generate these models using the string - net formalism @xcite .
the story changes when the discrete groups are non - abelian .
the example of the dihedral group @xmath1 was studied and the phase was found to be different from that of the corresponding quantum double phase . in particular the phase was found to be a modified version of the @xmath1 quantum double phase .
the abelian charge was no longer an excitation of the system instead it condensed to the ground state .
the flux excitations corresponding to the @xmath72 $ ] conjugacy class is not existent that is can not be expanded in the set of excitations of the system .
similar models were studied in @xcite where they add the magnetic terms to the links just like perturbations .
the plaquette operators were considered to be projectors to a normal subgroup of the full gauge group @xmath7 .
the resulting phases in @xcite are condensed versions of the @xmath1 quantum double phase .
the construction of the transfer matrices for two dimensional lattice gauge theories was shown in @xcite .
in particular it was shown to contain the quantum double models for any involutory hopf algebra .
the models in @xcite were found by trivializing the @xmath105 and @xmath169 parameters .
this implies there were no perturbation terms included in the transfer matrices considered in @xcite .
the models considered in this paper can be considered as an extension of the formalism presented in @xcite where we make non - trivial choices for the parameters @xmath105 and @xmath169 .
condensed phases of abelian quantum double models were exhibited in @xcite .
the models considered in this paper can be regarded as another way to obtain condensed phases along lines similar to @xcite .
the models considered in this paper included only four spin couplings .
we could include more couplings which will result in more complicated models .
for example if we couple a plaquette with the legs of the plaquette we can obtain models similar in spirit to the double semion model @xcite . however we can not use the transfer matrix in eq.([tz2 ] ) to obtain the double semion model .
the reason being that there are no choices of parameters @xmath104 , @xmath105 which will result in this model .
this calls for the construction of more general transfer matrices which can lead to these kinds of models . in section 5
we showed how we can go about finding the double semion phase by including two - qudit operators in the transfer matrix of lattice the gauge theory .
we only worked out the case of @xmath5 in this section . for other groups
this reduces to the twisted quantum double models constructed in @xcite .
these models can be thought of as arising due to the introduction of new parameters to the vertices which we denoted as @xmath173 .
hence it is natural to use @xmath172 to obtain the models dual to the twisted quantum double models .
we will explore these in future works .
it is interesting to note that the transfer matrix approach provides a variety of lattice gauge models including topological and non - topological ones .
this approach may very well be appropriate for studying quantum phase transitions in these lattice models .
pp , mjbf and pts thank fapesp for support of this work .
jpij thanks cnpq for support during this work .
d. v. else , c. nayak , _ classifying symmetry - protected topological phases through the anomalous action of the symmetry on the edge _ , phys .
b 90 , 235137 ( 2014 ) and arxiv:1409.5436 [ cond-mat.str-el ] .
m. koch - janusz , m. levin , a. stern , _ exactly soluble lattice models for non - abelian states of matter in 2 dimensions _
b 87 , 104406 ( 2013 ) and arxiv:1212.0593 [ cond-mat.str-el ] . phys . rev .
b 88 , 115133 ( 2013 ) and arxiv:1306.2789 [ cond-mat.str-el ] . | two dimensional lattice models such as the quantum double models , which includes the toric code , can be constructed from transfer matrices of lattice gauge theories with discrete gauge groups .
these transfer matrices are built out of local operators acting on links , vertices and plaquettes and are parametrized by the center of the gauge group algebra and its dual .
for general choices of these parameters the transfer matrix contains operators acting on links which can also be thought of as perturbations to the quantum double model driving it out of its topological phase towards a paramagnetic phase .
these perturbations can be thought of as magnetic fields added to the system which destroy the exact solvability of the quantum double model .
we modify these transfer matrices with perturbations and extract exactly solvable models which remain in a quantum phase , thus nullifying the effect of the perturbation .
the algebra of the modified vertex and plaquette operators now obey a deformed version of the quantum double algebra .
the abelian cases are shown to be in the quantum double phase whereas the non - abelian phases are shown to be in a modified phase of the corresponding quantum double phase .
this is shown by working with the groups @xmath0 and @xmath1 for the abelian and non - abelian cases respectively .
the quantum phases are determined by studying the excitations of these systems .
the fusion rules and the statistics of these anyons indicate the quantum phases of these models .
the implementation of these models can possibly improve the use of quantum double models for fault tolerant quantum computation .
we then construct theories which arise from transfer matrices that are not the transfer matrices of lattice gauge theories .
in particular we show that for the @xmath2 case this contains the double semion model .
more generally for other discrete groups these transfer matrices contain the twisted quantum double models .
these transfer matrices can be thought of as being obtained by introducing extra parameters into the transfer matrix of lattice gauge theories .
these parameters are central elements belonging to the tensor products of the algebra and its dual and are associated to vertices and volumes of the three dimensional lattice . as in the case of the lattice gauge theories we construct the operators creating the excitations in this case and study their braiding and fusion properties . |
a significant feature in the mott transition is the emergence of states in the mott gap in the single - particle spectrum due to the doping of a mott insulator .
@xcite since such a feature does not appear when a noninteracting band insulator is doped , this phenomenon must be related to strong electronic correlations intrinsic in the mott transition . however , although these states , which are called doping - induced states or in - gap states , have been observed in materials such as cuprate high - temperature ( high-@xmath2 ) superconductors and in various theoretical calculations , @xcite their interpretations are controversial . in particular , the question of whether the doping - induced states in the two - dimensional ( 2d ) hubbard and @xmath0-@xmath1 models in the parameter regime relevant to high-@xmath2 cuprates near the mott transition are essentially disconnected from the low - energy states by an energy gap or not has attracted considerable interest .
this is because this issue is related not only to the nature of the mott transition but also to the fermi surface near the mott transition ( hole pockets or a free - electron - like fermi surface ) @xcite and the anomalous features observed in high-@xmath2 cuprates .
@xcite the presence of the energy gap has been suggested by numerical calculations @xcite and interpreted in terms of various concepts , such as the binding between double occupancy and vacancy , @xcite the effect of a charge 2@xmath3 boson , @xcite hybridization between the quasiparticle and cofermion , @xcite and a spin - polaron shakeoff .
@xcite the absence of the energy gap has also been suggested by numerical calculations , @xcite where the doping - induced states have been interpreted as a quasiparticle coupled with antiferromagnetic correlations @xcite and as essentially the magnetic excitation of the mott insulator which emerges in the single - particle spectrum because charge character is added by doping .
@xcite in this paper , we address this issue from a general perspective , by investigating how the dispersion relation of the doping - induced states should generally behave in the small - doping limit of a continuous mott transition , based on an analysis of the quantum numbers and the overlaps of relevant states .
the validity of the argument is demonstrated using the two - leg ladder and bilayer @xmath0-@xmath1 models .
an intuitive real - space picture of the doping - induced states is also presented .
we consider zero - temperature properties of repulsively interacting systems where the ground states are mott insulators at half - filling and metals in the small - doping limit .
the numbers of sites , unit cells , electrons , and doped holes are denoted by @xmath4 , @xmath5 , @xmath6 , and @xmath7 , respectively .
the doping concentration is denoted by @xmath8 .
we study the single - particle spectral function defined as @xmath9 where @xmath10 and @xmath11 denote the ground state at @xmath12 and the excitation energy of the eigenstate @xmath13 from @xmath10 , respectively . here
, @xmath14 denotes the creation operator of an electron with momentum @xmath15 and magnetization @xmath16 .
if @xmath10 has spin @xmath17 , magnetization @xmath18 , and momentum @xmath19 , @xmath20 can be nonzero only for @xmath13 with @xmath21 holes , spin @xmath22 , magnetization @xmath23 , and momentum @xmath24 .
@xcite in addition , if the ground - state energy ( including the chemical - potential term ) at @xmath12 is lower than that of @xmath25 by @xmath26 , @xmath6-conserving excited states with excitation energy @xmath27 from @xmath28 can contribute to @xmath29 at @xmath12 , where @xmath30 for @xmath31 in a metallic phase ( including @xmath32 ) . in interacting systems , it is generally expected that low - energy eigenstates ( particularly those of dominant excitations ) having the same quantum numbers as those of components of @xmath33 have nonzero overlap with @xmath33 for @xmath15 where the lower hubbard band ( lhb ) is not completely filled .
thus , if @xmath34 has spin 1/2 and momentum @xmath19 , @xmath35 with spin 0 and momentum @xmath36 should contribute to @xmath29 at @xmath37 and @xmath38 for the top of the hole - like band [ lhb at half - filling ( @xmath39 ) ] in the small - doping limit , which causes hole - pocket - like behavior as in a doped band insulator .
@xcite in addition , the magnetically excited states with spin 1 at half - filling @xcite having the dispersion relation @xmath40 should emerge following doping along @xmath41 in @xmath29 for @xmath15 where the lhb is not completely filled . according to this argument ,
if the magnetic excitation of the mott insulator is gapless at momenta @xmath36 and @xmath42 , the doping - induced states in the small - doping limit should also be gapless at @xmath43 and @xmath44 ( @xmath45 if @xmath46 ) and exhibit the momentum - shifted magnetic dispersion relation in the momentum region where the lhb is not completely filled ( primarily outside the fermi surface ) .
this behavior essentially agrees with the results for the one - dimensional ( 1d ) and 2d hubbard models in refs . and those for the 2d @xmath0-@xmath1 model in ref . and suggests the coexistence of hole - like behavior and electron - like behavior for @xmath47 .
although it has been pointed out that the spin-0 and spin-1 states at half - filling can appear in the electron - addition spectrum at @xmath48 , the relationship of the momenta ( dispersion relations ) has not been recognized in ref . ;
the doping - induced states in the 2d @xmath0-@xmath1 and hubbard models have been considered to be disconnected by an energy gap from the low - energy states relevant to hole - like fermi surfaces and have been interpreted as spin - polaron shakeoff bands in refs . .
however , according to the above argument , it is natural to interpret the doping - induced states in the small - doping limit as essentially the states of the spin - wave mode of the mott insulator and to consider that they exhibit the gapless spin - wave dispersion relation @xcite shifted by @xmath19 in the 2d @xmath0-@xmath1 and hubbard models .
to strengthen the above argument , we consider the @xmath0-@xmath1 model defined by the following hamiltonian : @xmath49 where @xmath50 and @xmath51 for the annihilation operator @xmath52 and number operator @xmath53 of an electron with magnetization @xmath16 at site @xmath54 . here , @xmath55 denotes the spin operator at site @xmath54 . at each site , a vacant state @xmath56 or a singly - occupied magnetization-@xmath16 state @xmath57 is allowed , but double occupancy is forbidden . here
, we assume that @xmath35 has spin 0 and momentum @xmath36 and that @xmath34 has spin 1/2 , magnetization @xmath58 , and momentum @xmath19 .
for instance , @xmath35 on a bipartite lattice with the same number of sites in the two sublattices has spin 0 in a finite - size system with @xmath59 for neighboring sites between the two sublattices and @xmath60 otherwise .
@xcite because @xmath61 can be expressed as @xmath62 and @xmath63 using @xmath64 where @xmath65 denotes the state at site @xmath54 ( @xmath66 for @xmath67 , and 0 otherwise ) , @xmath68 where @xmath69 using the site @xmath70 coordinate @xmath71 .
equation ( 1 ) means that the overlaps between the electron - addition states at @xmath48 ( @xmath72 and @xmath73 ) and the single - spin excited states at half - filling ( @xmath74 and @xmath75 ) are related to the overlap between the electron - removal state for the momentum at the top of the lhb at half - filling ( @xmath76 ) and @xmath34 . from eq .
( [ eq : overlap0 ] ) , if @xmath77 is @xmath78 , @xmath79 and @xmath80 are expected to be @xmath81 at @xmath82 @xmath15-points . then
, for these @xmath15-points , the normalized electron - addition states should normally have @xmath78 overlap with the normalized single - spin excited states because @xmath83 and @xmath84 .
@xcite in addition , because all sites are singly occupied at half - filling , @xmath85 and @xmath86 .
hence , @xmath87 thus , the overlap can be roughly estimated in a decoupling approximation as @xmath88 where @xmath89 . at @xmath37 , @xmath90 because @xmath35 is assumed to be a spin-0 state .
the above results imply that the single - spin excited states of the mott insulator contribute significantly to the doping - induced states in the small - doping limit for @xmath15 where the lhb is not completely filled ( @xmath91 ) and for @xmath92 if @xmath77 is @xmath78 . in 1d systems , even if @xmath93 for @xmath94 , @xcite exact - solution analyses and numerical calculations strongly suggest that the mode of the doping - induced states in the small - doping limit primarily reflects the dominant part of the magnetic excitation .
to demonstrate the validity of this argument , we consider the two - leg ladder and square - lattice bilayer @xmath0-@xmath1 models for @xmath95 , @xmath96 , @xmath97 , and @xmath98 ( @xmath99 is not too small for the ground state to have spin 0 or 1/2 [ ] )
. here , @xmath100 and @xmath101 denote @xmath102 and @xmath103 for neighboring sites between chains ( planes ) , respectively , and @xmath104 and @xmath105 denote @xmath102 and @xmath103 for neighboring sites in chains ( planes ) , respectively , in the ladder ( bilayer ) model .
the other @xmath106 and @xmath107 are set to zero .
although some aspects of the doping - induced states in similar models have been discussed , @xcite this paper clarifies the nature of the doping - induced states through explicit calculations of the spectral weights , dispersion relation , and overlaps with magnetically excited states , using the effective eigenstates up to @xmath108 and @xmath109 . the ground state at half - filling ( spin 0 and momentum @xmath36 ) , the low - energy spin-1 eigenstate at half - filling , and the low - energy eigenstate at @xmath48 are expressed , respectively , as @xmath110 where @xmath111 , @xmath112 , and @xmath113 represent the two - electron spin-0 state , spin-1 state , and one - electron bonding state , respectively , at the @xmath70-th two - site unit cell with coordinate @xmath114 for @xmath115 .
here , @xmath18 and @xmath116 denote the magnetization and inter - unit - cell momentum , respectively .
the magnetic dispersion relation at half - filling is obtained by calculating the excitation energy of @xmath117 from @xmath35 : @xmath118 the ground state at @xmath48 with magnetization @xmath58 ( @xmath34 ) is @xmath119 if @xmath116 can be set to @xmath120 , where @xmath120 indicates @xmath121 and @xmath122 for 1d and 2d systems , respectively . for the small - doping limit , the value of @xmath123 is set such that @xmath124 with @xmath30 .
we calculate the spectral function at @xmath48 . by adding an electron to the ground state at @xmath48 , spin-0 and 1 states
are obtained @xcite as follows : @xmath125 ) for spin @xmath126 , and @xmath127 ) and @xmath128 ) for spin @xmath129 . here , @xmath130 indicates the momentum with inter- and intra - unit - cell momenta @xmath116 and @xmath131 , respectively , and @xmath132 .
thus , there are spectral weights at @xmath133 and @xmath38 in @xmath134 [ @xmath135 and essentially along the dispersion relation of the magnetic excitation shifted by @xmath136 , @xmath137 , in @xmath138 [ @xmath139 , where @xmath134 and @xmath138 denote the spectral functions for @xmath140 and @xmath141 , respectively .
the shift of @xmath120 is due to the momentum difference between @xmath34 and @xmath35 .
the spectral weight for @xmath142 averaged over the momentum is @xmath143 , which is consistent with the spectral - weight sum rule .
@xcite one quarter of this spectral weight ( @xmath144 ) is from the ground state at half - filling , which contributes to the mode causing hole - pocket - like behavior , whereas three quarters ( @xmath145 ) is from the magnetically excited states ( @xmath117 ) , which spread over the entire @xmath116 region for @xmath146 and can be identified as the doping - induced states .
in addition , eq .
( [ eq : overlap0 ] ) and the validity of the decoupling approximation [ eq .
( [ eq : overlap2 ] ) ] can be confirmed using @xmath147 a similar analysis is applicable to multi - hole - doped cases in the small - doping regime .
the ground state at @xmath12 is effectively expressed as a linear combination of direct products of @xmath148 @xmath149 and @xmath150 @xmath151 states . because the doping - induced states are obtained by replacing one of the @xmath148 @xmath149 states with @xmath152 , the spectral weights increase as @xmath153 increases .
in addition , because a magnetic - excitation mode similar to that of the mott insulator remains for @xmath154 and because the fermi momenta remain near @xmath155 , significant spectral weights should continue to be located almost along the momentum - shifted magnetic dispersion relation of the mott insulator in @xmath138 .
the states obtained by replacing one of the @xmath148 @xmath149 states with @xmath151 form a mode as in a doped band insulator ( @xmath156 the upper edge of the lhb at half - filling ) in @xmath134 .
@xcite more simply and generally , the behavior of the spectral function at finite @xmath157 can be understood as being deformed continuously from that of the small - doping limit if phase transition does not occur for @xmath158 .
this change should be small in the small - doping regime .
for confirmation , fig .
[ fig : akw ] shows the results for the @xmath0-@xmath1 ladder obtained using the non - abelian dynamical density - matrix renormalization group ( ddmrg ) method @xcite [ the ddmrg method @xcite in the u(1)@xmath159su(2 ) basis @xcite ] , where 120 density - matrix eigenstates are retained .
the spin dynamical structure factors @xmath160 and @xmath161 are defined , respectively , in the same way as @xmath162 and @xmath163 for @xmath142 but for spin operators .
the mode for @xmath164 in @xmath162 at @xmath165 corresponds to the doping - induced states and exhibits almost the same dispersion relation as that of the magnetic excitation at @xmath39 , but shifted by approximately @xmath121 in @xmath166 [ figs .
[ fig : akw](a)[fig : akw](c ) ] . the states for @xmath167 and @xmath168 in @xmath163 form a mode as in a doped band insulator [ figs .
[ fig : akw](d ) and [ fig : akw](e ) ] . thus , the results shown in fig .
[ fig : akw ] are consistent with the above analysis .
the results shown in this section ( sec .
[ sec:2siteclusters ] ) demonstrate that the magnetically excited states of the mott insulator emerge as the doping - induced states in the small - doping limit in 1d and 2d systems and that the dispersion relation of the doping - induced states is essentially the same as that of the magnetic excitation , but shifted by the fermi momentum . in the above ladder and bilayer @xmath0-@xmath1 models , the doping - induced states in the small - doping limit are disconnected from the low - energy states because the magnetic excitation at half - filling has an energy gap ; in models having a gapless magnetic excitation at half - filling , the doping - induced states should exhibit a gapless dispersion relation in the small - doping limit .
the difference in the electron - addition spectra between a mott insulator and the small - doping limit is explained as follows . in a mott insulator , @xmath169 yields no spectral weights in the mott gap because there are no eigenstates with @xmath170 in this energy regime . in the small - doping limit , in contrast
, @xmath171 can yield spectral weights in the mott gap because low - energy eigenstates for the magnetic excitation of the mott insulator exist .
@xcite the emergence of the electron - addition states in the mott gap reflects the spin - charge separation of the mott insulator .
@xcite the spin - charge separation means that the low - energy properties are described in terms of independent spin and/or charge degrees of freedom rather than in terms of an electron - like quasiparticle . in 1d interacting electron systems ,
the spin - charge separation occurs even in a metallic phase in the low - energy limit .
@xcite on the other hand , in a mott insulator , the spin - charge separation occurs @xcite in any dimension : the low - energy properties are described in terms of the spin excitation which is gapless or has an energy gap of the order of the ( effective ) spin - spin interaction , whereas the charge excitation has an energy gap of the order of the coulomb interaction . because of this spin - charge separation , although there are low - energy magnetically excited states , the electron - addition excitation does not have spectral weights in the mott gap in a mott insulator ; the charge quantum number of the magnetically excited states ( @xmath172 ) differs from that of @xmath169 ( @xmath170 ) .
however , the magnetically excited states can emerge in the electron - addition spectrum with the dispersion relation shifted by the fermi momentum following the doping of the mott insulator as discussed above .
@xcite this is in contrast with the noninteracting band insulator case , where the lowest spin excitation energy is the same as the band gap . in the ladder and bilayer @xmath0-@xmath1 models considered above ,
although the ground - state behavior is similar to that of a ( doped ) band insulator , the characteristic of the mott transition ( the emergence of the states ) occurs for @xmath142 , reflecting the spin - charge separation of the mott insulator ( the presence of the low - energy spin excitation and the absence of the low - energy charge excitation ) .
the situation in the small - doping regime can be intuitively explained as follows .
the added electron , which is located at the position of a vacancy at low energies , has almost no chance to hop to another vacant site if vacancies are distributed homogeneously .
however , the added electron can hop with a small energy to a neighboring site occupied by an electron having the opposite magnetization , at the same time as that electron hops to the site that was occupied by the added electron .
this is nothing but the spin exchange process .
thus , in general , the electron - addition excitation in the small - doping regime exhibits a dispersion relation similar to that of the magnetic excitation of the mott insulator , but shifted by the fermi momentum .
this shift is due to the momentum difference between the ground states before and after the electron addition .
the spectral weight in the large - repulsion regime should be @xmath173 because the probability that the added electron occupies a vacant site is essentially @xmath174 in the small - doping regime .
even if the ground state at half - filling has an antiferromagnetic long - range order , @xmath171 should have nonzero overlap with the spin - wave mode at half - filling because both involve spin-1 components if @xmath35 and @xmath34 have spin-0 and spin-1/2 components , respectively .
thus , the doping - induced states in the small - doping limit are expected to exhibit essentially the same dispersion relation as that of the spin - wave mode , but shifted by the fermi momentum , in the momentum region where the lhb is not completely filled ( primarily outside the fermi surface ) .
this picture is consistent with the results for the 2d hubbard and @xmath0-@xmath1 models in refs . .
the continuous transformation of the single - particle dispersion relation for @xmath142 into the momentum - shifted magnetic dispersion relation of the mott insulator does not necessarily mean that the magnetic excitation of the mott insulator is interpreted in terms of spinons ( spin-1/2 and charge-0 quasiparticles ) , because this characteristic is independent of quasiparticle pictures .
the momentum shift is simply due to the momentum difference between the ground states before and after the electron addition . in the above ladder and bilayer @xmath0-@xmath1 models
, the magnetic excitation is typically interpreted in terms of a triplon ( spin-1 and charge-0 quasiparticle ) .
the doping - induced states in the small - doping limit can involve not only the magnetically excited states but also other low - energy excited states ( spin-0 excited states ) at half - filling if @xmath171 contains components with these quantum numbers .
nevertheless , the doping - induced states are expected to be primarily due to the magnetic excitation because this is generally the most dominant low - energy excitation in a mott insulator ( fig .
[ fig : akw ] ) .
@xcite in a first - order mott transition where phase separation occurs for @xmath175 , the states obtained by adding an electron in the @xmath176 limit are mixtures of the electron - addition states from the ground state at @xmath177 and those at @xmath178 , which are not exactly the magnetically excited states at half - filling .
nevertheless , the dispersing mode for @xmath142 in the lhb before the mott transition in the large - repulsion regime should more or less reflect the magnetic excitation from a mott insulating state ( an eigenstate at half - filling ) similar to the ground state at @xmath177 .
thus , the indications that a spin - charge - separated mott insulator is being approached , such as a reduction in spectral weight from the dispersing mode , @xcite are expected to generally appear near the mott transition in the large - repulsion regime , provided the value of @xmath179 is relatively small .
we consider the issue of the doping - induced states in the 2d hubbard and @xmath0-@xmath1 models mentioned in sec .
[ sec : intro ] . because the low - energy magnetic excitation at half - filling is well described in terms of the gapless spin - wave mode , @xcite according to the present results , a mode exhibiting the spin - wave dispersion relation shifted by the fermi momenta should emerge in the electron - addition spectrum in the small - doping limit
; the mode should be gapless at the fermi momenta because the spin - wave mode is gapless .
if the lhb around @xmath180 is essentially completely filled , the mode appears primarily outside this momentum region .
this picture is consistent with the results in refs . ;
interpretations suggesting that the doping - induced states are essentially disconnected from the low - energy states by an energy gap @xcite are not consistent with this picture in the small - doping limit . for the 1d hubbard model
, this picture essentially agrees with the results in refs . . in summary ,
the general and direct relationships of the doping - induced states with the magnetically excited states of the mott insulator were clarified based on an analysis of the quantum numbers and the overlaps , as well as through a demonstration using the ladder and bilayer @xmath0-@xmath1 models .
the results imply that the doping - induced states in the small - doping limit exhibit essentially the same dispersion relation as that of the magnetic excitation of the mott insulator , but shifted by the fermi momentum , in the momentum region where the lhb is not completely filled in a continuous mott transition .
this characteristic should be general and independent of the dimensionality and quasiparticle pictures . with the reverse use of this characteristic , the magnetic dispersion relation of the mott insulator
can be inferred from the dispersion relation of the electron - addition excitation in the small - doping limit by shifting this dispersion relation by the fermi momentum . in particular ,
if the electron - addition excitation has a gap in an energy regime , the magnetic excitation is also expected to have a gap in this energy regime .
this work was supported by kakenhi ( grants no .
23540428 and no .
26400372 ) and the world premier international research center initiative ( wpi ) , mext , japan .
the numerical calculations were partly performed on the supercomputer at the national institute for materials science .
5 h. eskes , m. b. j. meinders , and g. a. sawatzky , phys .
lett . * 67 * , 1035 ( 1991 ) .
h. romberg , m. alexander , n. ncker , p. adelmann , and j. fink , phys .
b * 42 * , 8768 ( 1990 ) . c. t. chen , f. sette , y. ma , m. s. hybertsen , e. b. stechel , w. m. c. foulkes , m. schluter , s - w .
cheong , a. s. cooper , l. w. rupp , jr . ,
b. batlogg , y. l. soo , z. h. ming , a. krol , and y. h. kao , phys .
lett . * 66 * , 104 ( 1991 ) .
t. takahashi , t. watanabe , t. kusunoki , and h. katayama - yoshida , j. phys .
solids * 52 * , 1427 ( 1991 ) .
m. taguchi , a. chainani , k. horiba , y. takata , m. yabashi , k. tamasaku , y. nishino , d. miwa , t. ishikawa , t. takeuchi , k. yamamoto , m. matsunami , s. shin , t. yokoya , e. ikenaga , k. kobayashi , t. mochiku , k. hirata , j. hori , k. ishii , f. nakamura , and t. suzuki , phys .
lett . * 95 * , 177002 ( 2005 ) .
e. dagotto , rev .
mod . phys . *
66 * , 763 ( 1994 ) .
m. imada , a. fujimori , and y. tokura , rev .
phys . * 70 * , 1039 ( 1998 ) .
n. p. armitage , p. fournier , and r. l. greene , rev .
mod . phys . * 82 * , 2421 ( 2010 ) . m. b. j. meinders , h. eskes , and g. a. sawatzky , phys .
b * 48 * , 3916 ( 1993 ) .
h. eskes and a. m. ole , phys .
lett . * 73 * , 1279 ( 1994 ) .
h. eskes , a. m. ole , m. b. j. meinders , and w. stephan , phys .
b * 50 * , 17980 ( 1994 ) .
e. dagotto , a. moreo , f. ortolani , j. riera , and d. j. scalapino , phys .
* 67 * , 1918 ( 1991 ) .
e. dagotto , a. moreo , f. ortolani , d. poilblanc , and j. riera , phys . rev .
b * 45 * , 10741 ( 1992 ) .
e. dagotto , f. ortolani , and d. scalapino , phys .
b * 46 * , 3183 ( 1992 ) .
r. preuss , w. hanke , c. grber , and h. g. evertz , phys .
79 * , 1122 ( 1997 ) .
r. preuss , w. hanke , and w. von der linden , phys .
rev . lett . * 75 * , 1344 ( 1995 ) .
n. bulut , d. j. scalapino , and s. r. white , phys .
b * 50 * , 7215 ( 1994 ) . w. stephan and p. horsch , phys .
lett . * 66 * , 2258 ( 1991 ) . y. ohta , k. tsutsui , w. koshibae , t. shimozato , and s. maekawa , phys . rev .
b * 46 * , 14022 ( 1992 ) .
a. moreo , s. haas , a. w. sandvik , and e. dagotto , phys .
b * 51 * , 12045 ( 1995 ) .
p. phillips , t .-
choy , and r. g. leigh , rep .
. phys . * 72 * , 036501 ( 2009 ) .
p. phillips , rev .
phys . * 82 * , 1719 ( 2010 ) .
t. tohyama , phys .
b * 70 * , 174517 ( 2004 )
. s. sakai , y. motome , and m. imada , phys .
lett . * 102 * , 056404 ( 2009 ) .
s. sakai , y. motome , and m. imada , phys .
b * 82 * , 134505 ( 2010 ) . y. yamaji and m. imada , phys . rev
* 106 * , 016404 ( 2011 ) .
y. yamaji and m. imada , phys .
b * 83 * , 214522 ( 2011 )
. s. sakai , s. blanc , m. civelli , y. gallais , m. cazayous , m .- a .
m ' easson , j. s. wen , z. j. xu , g. d. gu , g. sangiovanni , y. motome , k. held , a. sacuto , a. georges , and m. imada , phys .
lett . * 111 * , 107001 ( 2013 ) .
t. d. stanescu , m. civelli , k. haule , and g. kotliar , ann . phys .
1682 ( 2006 ) . b. kyung , s. s. kancharla , d. snchal , a .- m .
s. tremblay , m. civelli , and g. kotliar , phys .
b * 73 * , 165114 ( 2006 ) .
a. liebsch , phys .
b * 81 * , 235133 ( 2010 ) .
m. kohno , phys .
105 * , 106402 ( 2010 ) .
m. kohno , phys .
lett . * 108 * , 076401 ( 2012 ) .
m. kohno , jps conf .
proc . * 3 * , 013020 ( 2014 ) .
m. kohno , phys .
b * 92 * , 085128 ( 2015 ) .
m. kohno , phys . rev .
b * 90 * , 035111 ( 2014 ) .
r. eder and y. ohta , phys .
b * 54 * , 3576 ( 1996 ) .
r. eder , k. seki , and y. ohta , phys .
b * 83 * , 205137 ( 2011 ) .
r. eder , y. ohta , and t. shimozato , phys .
b * 50 * , 3350 ( 1994 ) .
r. eder and y. ohta , phys .
b * 69 * , 100502 ( 2004 ) . t. m. rice , s. haas , m. sigrist , and f .- c .
zhang , phys .
b * 56 * , 14655 ( 1997 ) .
s. haas and e. dagotto , phys .
b * 54 * , 3718 ( 1996 ) . c. lavalle , m. arikawa , s. capponi , f. f. assaad , and a. muramatsu , phys . rev .
90 * , 216401 ( 2003 ) .
m. m. zemlji , p. prelovek , and t. tohyama , phys .
* 100 * , 036402 ( 2008 ) .
p. prelovek and a. ramak , physica c * 341348 * , 893 ( 2000 ) .
p. nozires , _ theory of interacting fermi systems _ ( w. a. benjamin , new york , 1964 ) .
p. w. anderson , phys . rev . * 86 * , 694 ( 1952 ) .
e. manousakis , rev .
phys . * 63 * , 1 ( 1991 ) .
j. e. hirsch , phys .
b * 31 * , 4403 ( 1985 ) .
e. lieb and d. mattis , j. math .
phys . * 3 * , 749 ( 1962 ) .
s. sorella and a. parola , j. phys . : condens .
matter * 4 * , 3589 ( 1992 ) .
s. sorella and a. parola , phys .
76 * , 4604 ( 1996 ) .
k. penc , k. hallberg , f. mila , and h. shiba , phys .
b * 55 * , 15475 ( 1997 ) . m. kohno , phys .
b * 56 * , 15015 ( 1997 ) . v. ya .
krivnov , a. a. ovchinnikov , and v. o. cheranovskii , synth . met .
* 33 * , 65 ( 1989 ) .
e. jeckelmann , phys .
b * 66 * , 045114 ( 2002 ) .
i. p. mcculloch , a. r. bishop , and m. gulacsi , philos . mag .
b * 81 * , 1603 ( 2001 ) .
i. p. mcculloch , ph.d .
thesis , australian national university , 2001 .
f. d. m. haldane , j. phys .
c : solid state phys .
* 14 * , 2585 ( 1981 ) .
s. tomonaga , prog .
* 5 * , 544 ( 1950 )
. j. m. luttinger , j. math .
phys . * 4 * , 1154 ( 1963 ) .
p. w. anderson , physica c * 341348 * , 9 ( 2000 ) . | in strongly correlated electron systems , the emergence of states in the mott gap in the single - particle spectrum following the doping of the mott insulator is a remarkable feature that can not be explained in a conventional rigid - band picture . here ,
based on an analysis of the quantum numbers and the overlaps of relevant states , as well as through a demonstration using the ladder and bilayer @xmath0-@xmath1 models , it is shown that in a continuous mott transition due to hole doping , the magnetically excited states of the mott insulator generally emerge in the electron - addition spectrum with the dispersion relation shifted by the fermi momentum in the momentum region where the lower hubbard band is not completely filled .
this implies that the dispersion relation of a free - electron - like mode in the electron - addition spectrum eventually transforms into essentially the momentum - shifted magnetic dispersion relation of the mott insulator , while its spectral weight gradually disappears toward the mott transition .
this feature reflects the spin - charge separation of the mott insulator . |
during the last decade , significant progress on the electromagnetic excitation of nucleon resonances has been obtained . for pion and eta photoproduction very precise data of unpolarized cross sections and photon asymmetries
were measured at mami@mainz , elsa@bonn , legs@brookhaven and graal@grenoble . for electroproduction , at mainz , bonn and bates
measurements for the @xmath1 excitation were performed at low @xmath2 , and for higher @xmath2 up to about @xmath3 gev@xmath4 , at jlab a very large amount of data was collected in an energy range up to the third nucleon resonance region . in parallel with the ongoing experiments , several theoretical groups developed models and analysis techniques , which were applied to the data .
the model - independent gwu / said analysis @xcite mostly analyzed the pion photoproduction data and improved the values of the photon couplings over the years .
coupled channels analyses were performed by the giessen group @xcite and by the bonn - gatchina group @xcite .
transition amplitudes were also determined in the framework of dynamically generated resonances by coupling to meson - baryon channels by the jlich @xcite and ebac @xcite groups . however , most successful concerning the general applicability to the higher resonances , were the unitary isobar models of the mainz group ( maid model ) @xcite and of the jlab group @xcite who used dispersion relations as an additional constraint to reduce the model dependence due to incomplete experimental input .
with our unitary isobar model maid , we analyzed all available electroproduction data in order to determine the transition form factors for all four - star resonances below @xmath5 gev .
in most cases we could obtain both single-@xmath2 and @xmath2 dependent transition form factors for the proton target . in the case of the neutron ,
the parametrization of the @xmath2 dependence had to take a simpler form because of the much smaller world database .
single-@xmath2 data points for longitudinal and transverse form factors have been obtained for transitions from the proton to the @xmath6 @xmath7 and @xmath8 , which can be downloaded from the maid website @xcite . here only results for the @xmath9 , and the @xmath10 are shown .
full results and details of the parameterizations are given in our recent review article , ref .
@xcite .
the main motivation for exploring the nucleon transition form factors is to obtain a precise knowledge of the nucleon excitation spectrum , which provides together with the elastic form factors a complete description of the nucleon s electromagnetic structure .
this structure can be compared with qcd inspired quark models and , in recent years , more and more also with lattice qcd calculations .
the @xmath1 is the only nucleon resonance with a well - defined breit - wigner resonance position , @xmath11 mev .
it is an ideal single - channel resonance , the watson theorem applies , and the breit - wigner position coincides with the k - matrix pole position . for these reasons
, the @xmath12 form factors can be determined in an essentially model independent way . the magnetic form factor shown in fig .
[ fig : p33ffs ] is very well known already from inelastic electron scattering up to high momentum transfer , @xmath13 gev@xmath4 , and can be parameterized in a surprisingly simple form found in our previous maid analysis , @xmath14 with @xmath15 the standard dipole form factor of the nucleon .
electromagnetic form factors of the @xmath16 to @xmath1 transition .
the upper panel shows the magnetic form factor @xmath17 in the definition of ash and the middle and lower panels show the @xmath18 and @xmath19 ratios , respectively .
the curves are the result of our maid2007 fit to the full dataset of pion electroproduction .
the data points are obtained in analyses at fixed @xmath2 .
the open symbols show the results of the jlab analysis @xcite , all others are obtained in our maid analysis .
the two largest @xmath2 values ( @xmath20 ) show the maid and the jlab analyses of the cross sections by villano et al . @xcite . for further details , see our review , ref .
@xcite.,width=453 ] the electric and coulomb form factors are much smaller and are usually given as ratios of @xmath21 and @xmath17 to the magnetic form factor in percent .
figure [ fig : p33ffs ] also compares the maid2007 solutions ( solid lines ) for the ratios @xmath22 and @xmath23 with other analyses .
the ratio @xmath22 from maid2007 stays always below the zero line , in agreement with the original analysis of the data @xcite and also with the dynamical model of sato and lee @xcite who concluded that @xmath22 remains negative and tends towards more negative values with increasing @xmath2 instead of an uprise towards unity .
this indicates that the predicted helicity conservation at the quark level is irrelevant for the present experimental @xmath2 range . for the ratio @xmath23
our @xmath2 dependent fit approaches a negative constant for large @xmath2 in good agreement with the predictions of ji et al .
@xcite , buchmann @xcite and ramalho et al .
@xcite , who use a relation between the ratio @xmath23 and the ratio of the electric and magnetic neutron form factors . for @xmath24 our maid2007 single-@xmath2 analysis for @xmath23 disagrees with the jlab analysis of aznauryan
whereas our analysis stays almost constant , the jlab analysis suggests a much larger negative slope . by repeating our data analysis at @xmath25 and @xmath3 gev@xmath4 in different energy ranges
, we found a strong dependence of the fit on the energy interval used .
our results of 2007 were obtained with the full energy range of the measured data , @xmath26 mev .
figure [ fig : p33ffs ] also shows an analysis ( blue open circles ) in the energy range of @xmath27 mev , much closer to the resonance position . if we choose the energy interval even closer to resonance , @xmath28mev , the errors increase further by a factor of 2 and the @xmath18 ratio becomes large and positive , while the @xmath19 ratio remains the same .
we conclude that the analysis in this @xmath2 range strongly depends on the energy interval and the parametrization of the background used in the analysis and requires further studies .
the helicity amplitudes for the roper resonance @xmath29 are displayed in the upper panel of fig .
[ fig : p11s11ffs ] . our latest @xmath2 dependent solution ( solid lines )
is in reasonable agreement with the single-@xmath2 analysis ( red circles ) .
the figure shows a zero crossing of the transverse helicity amplitude @xmath30 at @xmath31 gev@xmath4 and a maximum at the relatively large momentum transfer @xmath32 gev@xmath4 .
the longitudinal roper excitation @xmath33 rises to large values around @xmath34 gev@xmath4 and produces one of the strongest longitudinal amplitudes that we find in our analyses .
transverse ( @xmath35 ) and longitudinal ( @xmath36 ) electromagnetic form factors of the @xmath37 to @xmath29 ( upper panel ) and @xmath37 to @xmath10 ( lower panel ) transitions .
the curves are the result of our maid2007 fit to the full dataset of pion electroproduction .
the data points are obtained in analyses at fixed @xmath2 .
the open symbols show the results of the jlab analysis @xcite , all others are obtained in our maid analysis . for further details ,
see ref .
@xcite.,width=529 ] the lower panel in fig .
[ fig : p11s11ffs ] displays the results for the @xmath10 resonance .
the red single-@xmath2 data points show our results of 2007 , which are in good agreement with our @xmath2 dependent analysis ( solid lines ) .
the black triangles are the 2009 results of ref .
the blue data point at @xmath38 represents the pdg average over several @xmath39 and @xmath40 analyses .
while we find values around 65 in all maid analyses , the jlab analysis obtains values around 90 for @xmath39 and 110 for @xmath40 @xcite .
also the said and bonn - gatchina groups extract values around 100 , but in a very recent analysis , based on pion photoproduction , shrestha and manley report , however , also a small coupling of @xmath41 @xcite . from the @xmath2
dependent parametrization of the transition form factors we can calculate transverse charge transition densities , as viewed from a light front moving towards the baryon @xcite . for
that we first have to transform the helicity form factors @xmath42 to the dirac - like form factors @xmath43 and @xmath44 , which results in simple linear relations .
the transverse densities are then obtained by a 2-dim fourier - bessel transformation .
the densities relating to @xmath45 appear as fully symmetrical monopole patterns , while the @xmath46 form factors transform to additional dipole patterns ( further details in ref .
@xcite ) .
as an example , in fig .
[ fig : p11s11dens1p ] we show the polarized quark transverse charge densities from the proton to the roper and to the @xmath10 resonances . comparing these results , we find that the dipole contribution of the up and down quarks to the polarized densities is much less pronounced for the @xmath0 due to the much smaller @xmath47 form factor ratio .
comparison of the polarized quark
transverse transition charge densities corresponding to the e.m . transitions @xmath48
( @xmath49 ) ( left panel ) and @xmath50 ( @xmath51 ) ( right panel ) .
the light ( dark ) regions correspond to positive up quark ( negative down quark ) densities . , title="fig:",width=260 ] comparison of the polarized quark transverse transition charge densities corresponding to the e.m . transitions @xmath48 ( @xmath49 ) ( left panel ) and @xmath50 ( @xmath51 ) ( right panel ) .
the light ( dark ) regions correspond to positive up quark ( negative down quark ) densities .
, title="fig:",width=260 ] x .- d .
ma and f. yuan , phys .
* 90 * , 241601 ( 2003 ) .
a. j. buchmann , phys .
lett . * 93 * , 212301 ( 2004 ) .
g. ramalho , m. t. pena and f. gross , phys .
d * 78 * , 114017 ( 2008 ) .
a. n. villano et al . , phys .
rev . * c80
* , 035203 ( 2009 ) . i. g. aznauryan et al .
[ clas collaboration ] , phys .
rev . * c80 * , 055203 ( 2009 ) . m. shrestha and d. m. manley , arxiv:1208.2710 [ hep - ph ] . | longitudinal and transverse transition form factors for most of the four - star nucleon resonances have been obtained from high - quality cross section data and polarization observables measured at mami , elsa , bates , graal and cebaf . as an application
, we further show how the transition form factors can be used to obtain empirical transverse charge densities .
contour plots of the thus derived densities are shown and compared for the roper and @xmath0 nucleon resonances .
address = institut fr kernphysik , johannes gutenberg - universitt , d-55099 mainz , germany |
dynamical friction of scs in galaxies have been discussed in many different contexts , such as formation of stellar galactic nuclei via merging of old gcs ( e.g. , tremaine et al . 1975 ) , transformation from non - nucleated dwarfs into nucleated ones ( e.g. , oh & lin 2000 ) , physical meanings for the presence of the gc system ( gcs ) in the fornax dwarf galaxy ( e.g. , oh et al .
2001 ) , and dark matter distributions of dwarf galaxies ( e.g. , hernandez & gilmore 1998 ; goerdt et al . 2006 ; inoue 2009 ) .
orbital decay of gcs due to dynamical friction can significantly change spatial distribution of gcs and thus increase possibilities of them to be destroyed by strong galactic tidal fields ( e.g. , vesperini 2000 ) .
dynamical friction of scs ( including gcs ) in galaxies is thus considered to be important for better understanding the evolution of gc luminosity functions ( gclfs ) and that of half - number radii of the gcss in galaxies ( e.g. , vesperini 2000 ) .
recently , van de ven & chang ( 2009 ) have suggested the importance of dynamical friction in the dynamical evolution of scs around nuclear rings in galaxies . [ cols="^,^,^,^,^,^,^,^ " , ]
1 shows the time evolution of the projected distributions of scs for the last 0.64 gyr in the standard model with @xmath9 and with @xmath10 .
the projected spatial distribution of scs can change with time not only because of the dynamical evolution of the disk but also because of dynamical friction of scs against the disk field stars .
the half - number radius of the scs evolves from @xmath11 ( corresponding to @xmath12 kpc ) to @xmath13 ( @xmath14 kpc ) , which reflects that dynamical friction of scs can cause orbital decay of massive scs within the disk .
the total number of scs within @xmath15 in the disk can gradually increase so that a multiple sc system with the total mass ( @xmath16 ) of @xmath17 can form in the nuclear region of the disk .
the @xmath18 , @xmath19 , and @xmath20components of velocity dispersions of the multiple scs are 17.7 km s@xmath21 , 23.6 km s@xmath21 , and 3.2 km s@xmath21 .
the velocity dispersions comparable to internal velocity dispersions of their nuclear massive scs suggest that the nuclear scs can soon merge to form a single nucleus .
2 and 3 show orbital evolution projected onto the @xmath22-@xmath23 and @xmath22-@xmath24 planes for scs with two different masses . in both case , the distances from the center of the stellar disk ( @xmath25 ) become significantly smaller owing to orbital decay caused by dynamical friction against disk field stars .
the extent to which a sc can decay its orbit owing to dynamical friction depends on its initial mass and location with respect to the center of its host galaxy : the scs with masses of @xmath26 and @xmath27 can decrease @xmath25 by 31% and 63% of their initial @xmath25 , respectively .
the orbit of the more massive sc , which is initially in the inner region of the disk , can be influenced also by the inner spirals and bar developed during the disk evolution so that the orbital evolution can become more chaotic .
these orbital decay seen in these two scs can be more clearly seen in more massive scs owing to more effective dynamical friction for them ( see the appendix a for details of other scs orbits in the standard model ) .
4 shows that about 60 % of the scs have final @xmath25 ( @xmath28 ) significantly smaller than initial @xmath25 ( @xmath29 ) owing to dynamical friction of the sc against disk field stars .
spiral arms and bar in the disk can drive the angular momentum redistribution of the disk so that @xmath30 scs have @xmath31 : these are transferred to the outer regions of the disk by _ global galactic dynamics_. fig .
4 also shows that more massive scs with @xmath32 are more likely to have smaller @xmath33 ( i.e. , more significant orbital decay due to dynamical friction ) .
the mean @xmath33 for the 20 scs at @xmath34 gyr in this model is @xmath35 .
5 shows that all scs can significantly decrease their @xmath36 ( corresponding to apo - center distances from the galaxy center ) owing to ( i ) dynamical friction and ( ii ) radial motions caused by non - axisymmetric structures ( bar and spiral arms ) .
5 also shows that scs with smaller @xmath29 can experience more significant orbital decay ( i.e. , smaller @xmath37 ) due to dynamical friction , because dynamical friction is more effective in higher densities of background field stars .
the mean @xmath37 for the 20 scs at @xmath34 gyr in this model is @xmath38 .
6 shows the time evolution of the total mass of scs ( in units of @xmath1 ) within @xmath39 , which is referred to as @xmath40 for convenience .
this @xmath40 does not literally correspond to the mass of a single stellar nucleus and thus goes up and down rapidly according to time evolution of their orbits .
clearly , around @xmath41 gyr , @xmath42 can exceed the value of 0.003 observed for nucleated dwarf galaxies in the virgo cluster ( ct et al .
this relatively rapid formation of a multiple sc system is due to very efficient dynamical friction of massive scs with the mean mass of @xmath43 in the disk .
therefore , the time evolution of @xmath42 depends strongly on the adopted initial mass function of scs ( this point is discussed later ) .
dynamical friction processes of scs in disk galaxies depend strongly on model parameters such as @xmath1 , @xmath44 ( @xmath45 ) , @xmath46 ( = @xmath47 ) , central stellar densities of disks , and @xmath48 . the derived @xmath1-dependence is the most important result in the present study so that it is briefly summarized in the table 2 .
we illustrate the derived dependences on model parameters as follows : \(i ) orbital decay of scs due to dynamical friction is more important in less massive host galaxies , mainly because relative velocities ( @xmath49 ) between scs and disk field stars ( which are comparable to stellar velocity dispersions of the disks ) are smaller in the galaxies : the frictional drag force can be proportional to @xmath49 ( see 7.18 in binney & tremaine 1987 ) .
7 shows that @xmath33 is systematically smaller in disk galaxies with smaller @xmath1 for a fixed @xmath50 .
7 also shows that for disks with @xmath51 , no / few scs can show significant orbital decay ( @xmath52 ) , which means that dynamical friction of scs is not so important in their orbital evolution .
\(ii ) dynamical friction of scs by disk field stars is less effective in galaxies with smaller @xmath50 , mainly because stellar densities of disk field stars are lower so that frictional drag force can be weaker .
8 shows that for disks with @xmath53 , only a few scs can show significant orbital decay ( @xmath52 ) , which implies that dynamical friction of scs can not be so important in orbital evolution of scs in the early phase of disk evolution of galaxies when the mass fractions of disk stars can be very small .
8 also shows that a smaller fraction of scs have @xmath54 , which reflects that radial redistribution of scs due to the presence of non - axisymmetric structures ( spirals and bars ) does not happen in galaxies with smaller degrees of self - gravitating .
\(iii ) galactic bulges appear to be less important in orbital decay of scs within disks in comparison with other galactic properties such as @xmath1 and @xmath50 .
8 shows no remarkable differences between models with different @xmath47 ( = 0.17 and 1.0 ) , though @xmath47 can be an important parameter which determines whether stellar bars can form and thus affect orbital evolution of scs .
\(iv ) the time scale of dynamical friction of scs by disk field stars is longer in lsbs than in hsbs , because stellar densities of background field stars can be key factors for the effectiveness of dynamical friction . as a result of this , formation of stellar nuclei via sc merging
can proceed more slowly in lsbs .
for example , @xmath40 can become @xmath55 only after 2.4gyr orbital evolution of scs in the lsb model in which @xmath56 is by a factor of 2.5 larger than that in the standard model ( and other model parameters are the same as those in the standard model ) .
\(v ) final properties of scss ( e.g. , @xmath57 ) and the formation processes of nuclear multiple scs ( thus nucleus formation processes ) in galaxies depend on initial mfs ( e.g. , @xmath58 ) .
for example , @xmath57 can evolve more rapidly and have smaller final values in models with smaller @xmath58 ( i.e. , larger @xmath16 for the most massive scs ) in the present study ( e.g. , @xmath59 for models with @xmath60 mag ) .
also final @xmath40 is larger in models with smaller @xmath61 ( e.g. , @xmath62 for @xmath63 mag ) .
this result implies that formation processes of stellar galactic nuclei ( e.g. , whether dwarfs are nucleated or non - nucleated ) depend on original mfs of scs in galaxies .
the present study has first shown that more massive scs can spiral - in to the nuclear regions of galactic disks owing to dynamical friction within less than @xmath64 gyr in low - mass disk galaxies .
it is highly likely that these multiple scs in nuclear regions of disk galaxies can merge with one another to form stellar galactic nuclei ( or single massive nuclear scs ) .
if this is the case , then the merged scs ( i.e. , stellar galactic nuclei or nuclear scs ) would not be identified as ordinary scs in observations .
therefore , it is meaningful for the present study to investigate scmfs located within @xmath65 for comparing the simulated scmfs with the observed ones . in the followings ,
we assume that mfs are for scs located within @xmath65 ( i.e. , excluding the nuclear scs in the simulations ) .
9 shows that no scs with @xmath66 can be seen in the final mf for the standard model , because more massive scs spiral - in to nuclear regions of disk galaxies more rapidly owing to dynamical friction of scs against disk field stars .
this result implies that mfs of scs for @xmath67 ( i.e. , higher - mass end ) can steepen significantly in comparison of their original mfs ( at the epoch of sc formation ) owing to dynamical friction of scs .
furthermore , given that dynamical friction of scs against disk field stars is more effective in less massive disk galaxies , the mfs can more significantly steepen in less massive disk galaxies .
thus , it is highly likely that the scmfs can be steeper in less massive disk galaxies . based on observational studies of gclfs for the early - type galaxies in the acs virgo cluster survey , jordn et al .
( 2006 ) showed that the gcmfs for relatively high masses ( @xmath68 ) are steeper in less luminous galaxies in the virgo cluster .
they suggested that since gcmfs are not strongly affected by dynamical evolution of gcs ( e.g. , dynamical friction ) over a hubble time , the steeper gcmfs arise from variations in initial conditions .
our results however imply that dynamical friction of scs against disk field stars is so effective as to change the shapes of gcmfs within a relatively short time scale .
our results also imply that if these less luminous early - type galaxies were formed from disk galaxies either by galaxy interactions or by major mergers , their gcmfs can be steeper owing to changes of gcmfs in _ their progenitor disks _ caused by dynamical friction . since tremaine et al .
( 1975 ) first investigated dynamical evolution of gcs interacting with background stars through dynamical friction in m31 to discuss nucleus formation in m31 , many authors discussed formation of stellar galactic nuclei via mergers of scs ( or gcs ) in galaxies with different physical properties ( e.g. , lotz et al .
2001 ; capuzzo - dolcetta & vicari 2005 ) .
recent fully self - consistent n - body simulations have shown that merging of scs in the inner regions of galaxies lead to the formation of stellar galactic nuclei with their projected radial density profiles similar to the observed ones ( e.g. , capuzzo - dolcetta & miocchi 2008 ) . although these works did not discuss potential importance of dissipative gas dynamics and star formation in the formation of stellar galactic nuclei ( e.g. , bekki et al .
2006 ; bekki 2007 ) , they significantly advanced our understanding dynamical evolution of multiple sc systems in galaxies .
one of remaining questions in the sc merger scenario of nucleus formation is whether the scenario can explain the observed relation between stellar nucleus masses and their host galaxy ones ( @xmath69 , where @xmath70 is the galaxy mass ; ct et al .
2006 ) . since the observational study by ct et al .
( 2006 ) derived total mass of the luminous components ( i.e. , not those for entire galaxies including the extended dark matter halos ) of galaxies by using reasonable @xmath71 , it is reasonable to use @xmath1 rather than @xmath72 for the purpose of discussing the observed @xmath73 relation .
10 shows @xmath42 as a function of @xmath1 for the five models in which only @xmath1 are different between them .
the simulated @xmath42 is close to the observed value of 0.003 for less luminous disk galaxies with @xmath74 .
scs can lose a significant fraction of their original masses during their orbital evolution in galaxies ( e.g. , fujii et al .
2008 ) , which is not considered at all in estimating @xmath42 in the present study .
thus the simulated @xmath42 for less luminous disks can be regarded as fairly consistent with observations .
10 also shows that formation of stellar galactic nuclei with @xmath75 by sc merging is not possible in more luminous disks owing to ineffective dynamical friction .
therefore , the result implies that if the observed @xmath76 relation can be hold for all galaxies , the sc merger scenario may well have a serious problem in explaining the observation .
however , galaxies are more likely to have massive black holes ( mbhs ) rather than stellar nuclei in more luminous galaxies with @xmath77 ( e.g. , ferrarese et al .
2006 ) , though it is not so clear whether these galaxies with mbhs also have stellar nuclei ( i.e. , mbhs and stellar nuclei coexist in galaxies ) .
thus it would be currently reasonable to claim that the sc merger scenario is a promising one for the formation of stellar galactic nuclei in galaxies .
we have first shown that dynamical friction of scs against disk field stars can be so effective in low - mass disk galaxies so that orbital decay of scs can be as significant as to change physical properties ( e.g. , half - number radii and scmfs ) of their scss .
previous theoretical and numerical works focused exclusively on dynamical friction of halo scs ( e.g. , old gcs ) against dark matter halos and showed that the time scale of orbital decay of scs due to dynamical friction against dark matter halos is relatively long ( a few to several gyr ) even for low - mass dwarfs like the fornax ( e.g. , goerdt et al .
2006 for the cuspy nfw profile of its dark matter halo ) . since they did not investigate the influences of stellar disks on orbital evolution of halo scs through dynamical friction , it is worthwhile for the present study to investigate how halo scs evolve in disk galaxies if they can experience dynamical friction against both dark matter halos and against disk field stars .
11 shows that orbital decay of halo scs due to dynamical friction against halos and disks is much less significant owing to ( i ) lower mass densities of the halos and ( ii ) larger velocity differences in the scs and dark matter halos for the standard model .
the mean @xmath33 for the halo scs is 1.53 , which is by a factor of @xmath78 larger than that for the disk scs . only some more massive scs with @xmath79
can show significant orbital decay ( i.e. , decreasing apocenter distances of their orbits ) for halo scs .
this less effective dynamical friction in halo scs is a common feature in the present models , which suggests that the presence of disks in galaxies does not significantly change dynamical friction processes of halo scs .
it should be however stressed that the halo scs initially in the inner region of the disk galaxy in the standard model shows a slightly flattened final spatial distribution : this can be due to relatively stong dynamical friction between halo scs and disk field stars in the inner regions of galaxies .
we have investigated orbital evolution of scs being influenced by dynamical friction of disk field stars based on galaxy - scale numerical simulations on dynamical evolution of disk galaxies .
we have investigated models with variously different model parameters in order to understand dependences of the effectiveness of dynamical friction of scs on physical properties of disk galaxies .
we summarize our principle results as follows .
\(1 ) dynamical friction of scs against disk field stars is much more effective in orbital decay of scs in comparison with that against galactic halos in disk galaxies .
for example , the half - number radius of the scs ( @xmath3 ) in a galaxy with @xmath80 can decrease by @xmath81% owing to orbital decay of scs caused by dynamical friction within well less than @xmath82 yr for a reasonable mf of scs .
however orbital decay of scs due to dynamical friction can be more clearly seen only for more massive scs ( e.g. , those with @xmath83 ) .
\(2 ) dynamical friction processes of scs against disk field stars depend on physical properties of their host galaxies .
for example , the dynamical friction is much more effective in disks with lower @xmath1 owing to smaller stellar velocity dispersions . as a result of this
, @xmath3 can more rapidly evolve with time in disks with lower @xmath1 .
evolution of scss in more massive disks ( with @xmath84 ) can not be clearly seen in the present study , which implies dynamical friction is important only for evolution of scss ( e.g. , @xmath3 ) in less massive disk galaxies .
\(3 ) dynamical friction of scs by disk field stars is more effective in disks with higher degrees of self - gravitating of the disks ( e.g. , higher @xmath50 ) , which implies that higher stellar densities and non - axisymmetric structures such as bars and spiral arms can promote dynamical friction .
orbital decay of scs due to dynamical friction in disks does not depend so strongly on bulge masses of the disks for @xmath85 in the present study .
evolution of scss due to orbital decay of scs caused by dynamical friction is more important in hsbs than in lsbs .
\(4 ) more massive scs can spiral - in to the nuclear regions ( @xmath86 ) of disk galaxies owing to dynamical friction so that they can form multiple sc systems there in less massive disk galaxies with @xmath87 .
these systems show velocity dispersions comparable to or less than internal stellar velocity dispersions of scs so that they are likely to merge quickly with one another to form single massive scs there .
thus formation of stellar galactic nuclei via merging of more massive scs is possible in less massive disk galaxies .
\(5 ) the observed @xmath88 relation ( i.e. , @xmath89 ) for less luminous galaxies can be closely associated with formation of stellar galactic nuclei by sc merging .
the lf / mfs for more massive scs in scss of disk galaxies can steepen owing to the formation of stellar galactic nuclei through merging of more massive scs inwardly transferred to the nuclear region of galaxies .
this steeping of sclf / mfs due to dynamical friction can be more significant in less massive galaxies .
i am grateful to the anonymous referee for valuable comments , which contribute to improve the present paper .
kb acknowledges the financial support of the australian research council throughout the course of this work .
numerical computations reported here were carried out both on the grape system at the university of new south wales and on those kindly made available by the center for computational astrophysics ( cfca ) of the national astronomical observatory of japan .
bekki , k. , 2007 , pasa , 24 , 77 bekki , k. , peng , e. w. , 2006 , mnras , 370 , 1737 bekki , k. , couch , w. j. , drinkwater , m. j. , shioya , y. , 2004 , apj , 610 , l13 bekki , k. , couch , w. j. , shioya , y. , 2006 , apj , 642 , l133 bekki , k. , yahagi , h. , forbes , d. a. , 2006 , apj , 645 , l29 binney , j. , tremaine , s. 1987 in galactic dynamics , princeton ; princeton univ . press .
capuzzo - dolcetta , r. , miocchi , p. , 2008 ,
mnras , 388 , l69 capuzzo - dolcetta , r. , vicari , a. , 2005 , mnras , 356 , 899 ct et al . , 2006 , apjs , 165 , 57 ferrarese , l. , et al .
, 2006 , apj , 644 , l21 freeman , k. c. , 1970 , apj , 160 , 811 freeman , k. c. , illingworth , g. , & oemler , a. , jr . , 1983 , apj , 272 , 488 fujii , m. , iwasawa , m. , funato , y. , makino , j. , 2008 , apj , 686 , 1082 goerdt , t. , moore , b. , read , j. i. , stadel , j. , zemp , m. , 2006 , mnras , 368 , 1073 harris , w. e. , 1991 , ara&a , 29 , 543 harris , w. e. , & van den bergh , s. , 1981 , aj , 86 , 1627 hernandez , x. , gilmore , g. , 1998 , mnras , 297 , 517 hurley , j. r. , bekki , k. , 2008 , mnras , 389 , l61 inoue , s. , 2009 , mnras , in press jordn et al . , 2006 , apj , 651 , l25 kravtsov , a. v. , gnedin , o. y. , 2005 , apj , 623 , 650 lada , c. j. , lada , e. a. 2003 , ara&a , 41 , 57 lotz , j. m. , telford , r. , ferguson , h. c. , miller , b. w. , stiavelli , m. , mack , j. , 2001 , apj , 552 , 572 mastropietro , c. , moore , b. , mayer , l. , debattista , v. p. , piffaretti , r. , stadel , j. , 2005 , mnras , 364 , 607 miller , b. w. , lotz , j. m. , 2007 , apj , 670 , 1074 navarro , j. f. , frenk , c. s. , white , s. d. m. , 1996 , apj , 462 , 563 ( nfw ) neto , a. f. , et al . 2007 mnras , 381 , 1450 oh , k. s. , lin , d. n. c. , 2000 , apj , 543 , 620 oh , k. s. , lin , d. n. c. , richer , h. b. , 2000 , apj , 531 , 727 olsen , k. a. g. , miller , b. w. , suntzeff , n. b. ; schommer , robert a. , bright , j. , 2004 , aj , 127 , 2674 stiavelli , m. , miller , b. w. , ferguson , h. c. , mack , j. , whitmore , b. c. , lotz , j. m. , 2001 , aj , 121 , 1385 sugimoto , d. , chikada , y. , makino , j. , ito , t. , ebisuzaki , t. , umemura , m. , 1990 , nat , 345 , 33 tremaine , s. d. , ostriker , j. p. , spitzer , l. , jr . , 1975 , apj , 196 , 407 van den bergh , s. , 2000 , the galaxies of the local group , cambridge : cambridge univ . press . van de ven , g. , chang , p. , 2009 ,
apj , 697 , 619 vesperini , e. , 2000 , mnras , 318 , 841 wielen r. , 1977 , a&a , 60 , 263
fig . a1 shows time evolution of distances ( @xmath25 in units of @xmath56 ) of ten scs from the center of their host disk galaxy in the standard model . in order to show more clearly orbital evolution of each individual sc ,
only 5 scs are selected and shown for less massive scs ( @xmath90 ) and for more massive ones ( @xmath91 ) .
the orbital eccentricity of a sc in the present study is described as follows : @xmath92 where @xmath36 and @xmath93 are the minimum and maximum @xmath25 for the sc .
it is found that @xmath36 of the selected ten scs range from 0.006 to 0.51 and the mean @xmath36 is 0.17 . it is also found that @xmath94 of the scs range from 0.19 to 0.97 and the mean @xmath94 is 0.67 . owing to dynamical interaction between the non - axisymmetric structures ( e.g. , bar and spiral arms ) developed in the later phase of the disk s dynamical evolution , orbital evolution of these scs can be quite complicated .
one example of this is that not all of the scs show significant declines in @xmath25 owing to dynamical friction within @xmath95 gyr .
some of scs show orbital decay before the formation of a stellar bar , their @xmath25 can again become larger after the bar formation owing to redistribution of angular momentum by the bar .
it appears that sc - sc interaction does not play a vital role in the orbital evolution of scs owing to the small number ( @xmath96 ) of scs in this model . | we numerically investigate orbital evolution of star clusters ( scs ) under the influence of dynamical friction by field stars of their host disk galaxies embedded in dark matter halos .
we find that scs with masses larger than @xmath0 can show significant orbital decay within less than 1 gyr due to dynamical friction by disk field stars in galaxies with disk masses ( @xmath1 ) less than @xmath2 .
we also find that orbital decay of scs due to dynamical friction is more remarkable in disk galaxies with smaller @xmath1 and higher mass - ratios of disks to dark matter halos .
the half - number radii ( @xmath3 ) and mean masses within @xmath4 of the sc systems ( scss ) in low - mass disk galaxies with @xmath5 are found to evolve significantly with time owing to dynamical friction of scs .
more massive scs that can spiral - in to the central regions of disks can form multiple sc systems with smaller velocity dispersions so that they can merge with one another to form single stellar nuclei with their masses comparable to @xmath6% of their host disk masses .
based on these results , we suggest that luminosity functions ( lfs ) for more massive globular clusters ( gcs ) with masses larger than @xmath7 can steepen owing to transformation of the more massive gcs into single stellar nuclei through gc merging in less luminous galaxies .
we also suggest that the half - number radii of gc systems can evolve owing to dynamical friction only for galaxies with their total masses smaller than @xmath8 .
[ firstpage ] globular clusters : general galaxies : formation galaxies : kinematics and dynamics galaxies : halos galaxies : star clusters |
the post - linac collimation system should simultaneously fulfill several different functions . in first place , during routine operations , it should remove with high efficiency off - momentum and large amplitude halo particles , which could be lost inside undulator modules and become the source of radiation - induced demagnetization of undulator permanent magnets .
the system also must protect the undulator modules and other downstream equipment against miss - steered and off - energy beams in the case of machine failure without being destroyed in the process . because the collimation system is designed as passive
, the collimators must be able to withstand a direct impact of such number of bunches which can be delivered to their locations until a failure will be detected and the beam production in the rf gun is switched off . from the beam dynamics point of view
, the collimation section , as a part of the beam transport line from linac to undulator , must meet a very tight set of performance specifications .
it should be able to accept bunches with different energies ( up to @xmath0 from nominal energy ) and transport them without any noticeable deterioration not only of transverse , but also longitudinal beam parameters , i.e. it must be sufficiently achromatic and sufficiently isochronous .
this will not only reduce jitter of transverse beam parameters and time of flight jitter due to an energy jitter , but also will allow to fine - tune the fel wavelength by changing the electron beam energy without adjusting magnet strengths ( an energy change of @xmath1 corresponds to @xmath2 change in the fel wavelength ) and , even more , will make possible to scan the fel wavelength within a bunch train by appropriate programming of the low level rf system . some of above requirements are not in good agreement with one another and , as often , the basic problem is to find a balance among all competing factors so as to have at the end a system which still satisfactory fulfills design goals .
for example , relatively large betatron functions , which are needed at the collimator locations to guarantee their survival during occasional beam impacts , lead , as a rule , to unacceptable chromatic aberrations and , therefore , chromaticity correcting sextupoles are essential in preventing the dependence of linear optical parameters on the energy deviation .
chromatic - aberration correction with sextupoles , in the next turn , requires a beamline with dispersion , which makes separate regulation of transverse and energy collimation depths difficult and thus reduces flexibility of a system . in this article
we give an overview of the optics solution which fulfills all listed above requirements , and more details can be found in @xcite .
the part of the beam transport from linac to undulator , which we call the * post - linac collimation section * and which is shown in fig .
[ figoverview ] , consists of two arcs separated by a straight section ( phase shifter ) and includes matching modules at both ends to adapt the optic to the desired upstream and downstream beam behavior .
the collimation section bends the beam in the vertical plane and its length measured in the projection on the linac axis is about @xmath3 .
the outgoing beam axis points slightly downward with an angle of about @xmath4 and the vertical offset of the center of the last quadrupole of the second arc from the linac axis is @xmath5 .
the arcs are almost identical except that in the second arc the polarity of dipole and sextupole magnets is reversed and dipole bending angles are slightly smaller in absolute values in order to produce the net downward beam deflection .
each arc consists of four @xmath6 cells , constitutes a second - order achromat and is first - order isochronous .
three different types of collimators are foreseen in the xfel post - linac collimation system : main primary collimators , supplementing primary collimators and secondary collimators ( absorbers ) .
the principal purpose of the main primary collimators is to intercept trajectories of all incoming particles which would otherwise appear outside of the downstream dynamic aperture .
the main collimators will also shade a part of the beam pipe in the collimation section ( but not all ) from uncontrolled beam impacts . supplementing primary collimators assist to accomplish this work or , at least , reduce the probability of such events .
simultaneously , the supplementing collimators will not affect the transverse and energy collimation depths set by the main collimators . to improve the overall cleaning efficiency at the exit and better
localize losses inside of the collimation section several absorbers placed in the shadow of the primary collimators will be used .
the first arc will collimate transverse positions of incoming particles and the second , after a shift of vertical and horizontal betatron phases by odd multiples of @xmath6 , their transverse momenta .
the energy and vertical plane collimation will be done simultaneously , and therefore the ratio of dispersion to vertical betatron function at collimator locations has to be properly adjusted in order to achieve the required transverse and energy collimation depths . because , according to the optics design , dispersion can not be varied during machine operations , the rough preliminary adjustment was made already during design stage by appropriate selection of the arc parameters , and the operational flexibility will be provided by usage of collimators with exchangeable apertures and by tuning betatron functions at the collimator locations .
large beam spot size requested at the collimator locations and , in the same time , the possibility to transport bunches with different energies ( up to @xmath0 from nominal energy ) while preserving with good accuracy energy independent input and output matching conditions , make the control of chromatic effects one of the main issues in the design of the optics in the collimation section . without correction
the chromatic aberrations are unacceptable and , therefore , introduction of chromaticity correcting sextupoles becomes essential in improving overall system performance .
there are different approaches to the problem of compensation of chromatic effects , and the solution , which we found to be most adequate to the design requirements , is as follows : we compensate the arc chromatic effects by tuning arcs to become second - order achromats
. reduction of chromatic aberrations in the system straight sections is done for the particular betatron functions transported through these parts and without involving sextupole fields , simply by an accurate drift - quadrupole optics design ( a straight drift - quadrupole system can not be made an achromat , but it can be made a second - order apochromat with respect to certain incoming beam ellipses , i.e. it can transport these beam ellipses without introducing first - order chromatic distortions ) . , which are required to bring linear momentum compaction of the collimation section to @xmath7 ( red curve ) and to @xmath8 ( green curve).,width=283 ] in the first design , which is described in the tdr @xcite , the linear momentum compaction of the whole collimation system ( @xmath9 matrix coefficient ) was approximately equal to @xmath10 , which at that time was considered as acceptable value . in later studies of the microbunching instability it was found that even such a small value can not be neglected in the calculation of the gain of this instability and that , in order to reduce this gain , it is desirable to have @xmath11 of the collimation section equal to zero or , even better , to bring it to a positive value of about @xmath12 @xcite . because some other reasons for the choice of the linear momentum compaction could appear and the exact value of @xmath13 is not clear yet , we made system modifications in such a way , that though @xmath11 could not be varied dynamically during machine operations , the linear momentum compaction can be adjusted within about @xmath14 limits by system realignment while keeping space positions not only of the system end point and the system straight sections but also of the arc centers unchanged , as can be seen in fig . [ figdiffrefpath ] . the design which we describe in this paper ( *
baseline design * ) is the first - order isochronous beam line with @xmath15 . according to the system
design the arcs are tuned to become second - order achromats and this , together with the fixed system geometry and the requirement of the first - order isochronicity , completely determines the setting of the arc magnets and also the behavior of linear dispersion function .
thus the modifications of the betatron functions along the collimation section can be provided only by retuning of quadrupoles in the matching sections and in the phase shifter .
this is not a limitation in our case , because the design of these sections is done in such a way that with appropriate adjustment of their quadrupoles we are not only able to provide betatron functions of about @xmath16 at the points where the collimators are located ( * standard collimation optics * ) , but also able to translate smoothly the standard collimation optics into an optics with regular fodo - like transport through the entire collimation section .
this flexibility is an important property of the designed system and will be extensively used during machine commissioning and/or during measurements of beam parameters .
for example , commissioning starts with optics set to provide regular fodo - like transport and with sextupoles switched off and then , with experience gained , this optics can be translated step by step into the standard collimation optics .
betatron functions corresponding to the standard collimation optics can be seen in fig .
[ figbeta ] .
[ figdisp ] shows the linear dispersion , which is independent on setting of quadrupoles in the matching sections and in the phase shifter .
) along the collimation section .
beam energy 17.5 @xmath17 .
normalized emittances 1.4 @xmath18 .
red curve shows the design spot size ( linear theory ) .
all other curves are results extracted from the tracking simulations .
a matched gaussian beam at the entrance with @xmath19 ( blue ) and @xmath20 ( green ) energy offsets , with @xmath21 transverse offset ( black ) , and with both @xmath19 energy and @xmath21 offsets ( magenta ) .
sextupoles are switched on.,width=283 ] ellipses ( matched at the entrance ) after tracking through the entire collimation section .
the relative energy deviations are equal to @xmath22 , @xmath23 and @xmath24 ( red , green and blue ellipses , respectively ) .
sextupoles are switched on.,width=283 ] the optics solution for the xfel post - linac collimation section described in this paper meets all design specifications .
it is capable of providing simultaneously a large beam spot size at the collimator locations ( fig .
[ figspots ] ) and , in the same time , to transport bunches with different energies ( up to @xmath0 from nominal energy ) while preserving with good accuracy energy independent input and output matching conditions ( fig .
[ figell ] ) .
these criteria are met by designing a magnetic system whose second - order chromatic and geometric aberrations are controlled by the symmetry of the first - order optics and sextupole fields .
the system uses four main primary collimators and the studies presented in @xcite show that these collimators are able to confine all particles which passed the collimation section freely ( without touching collimator apertures ) into a volume in the phase space , that can be safely transported through all downstream beamlines ( including undulator modules ) to the beam dumps . | in this article we give an overview of the the post - linac collimation system for the european x - ray free - electron laser ( xfel ) facility @xcite with main emphasis on lattice and optics design . |
in the standard model ( sm ) , tau leptons are produced either in pairs or in association with a tau anti
neutrino , as expected from lepton flavour conservation . in electron
proton collisions , pairs of tau leptons are produced via photon
photon interaction @xmath53 ( figure [ diag]a ) , in which a photon from the electron interacts with a photon emitted by the proton @xcite .
tau leptons and tau anti
neutrinos are produced in @xmath54 boson decays , as illustrated in figure [ diag]b @xcite .
the signature of these events is a high transverse momentum ( @xmath55 ) isolated tau lepton , large missing transverse momentum @xmath56 due to the undetected neutrinos , and a hadronic system , typically of low @xmath55 .
this paper presents a measurement of the production of tau lepton pairs ( @xmath57 ) and a search for events with an isolated tau lepton accompanied by large missing transverse momentum ( @xmath58 ) .
the measurement of @xmath57 production is performed at low transverse momentum considering both leptonic and hadronic tau decays .
this measurement complements that of electron and muon pair production previously performed by the h1 collaboration @xcite . in the search for @xmath58 events tau
decays are identified in the hadronic channel only .
this search complements the previous h1 measurements of the production of events with an isolated electron or muon and large missing transverse momentum , which have revealed an excess over the sm expectation of events containing in addition a high @xmath55 hadronic system @xcite .
the zeus collaboration has also studied the production of events with an isolated lepton and large missing transverse momentum @xcite .
the analysis is based on data from electron
proton collisions at a centre of mass energy of 301 or 319 gev , recorded by the h1 experiment at hera in the period @xmath59@xmath60 .
the total integrated luminosity amounts to @xmath61 pb@xmath62 for the measurement of @xmath57 production and @xmath63 pb@xmath62 for the search for @xmath58 events .
this paper is organised as follows .
the physics processes leading to tau lepton production at hera are described in section @xmath64 together with the relevant background processes . in section
@xmath65 the h1 detector and the experimental conditions are briefly described .
particle identification is presented in section [ sec : partid ] .
the selection of events with tau pairs and the resulting measurements are described in section @xmath66 .
the search for events with an isolated high @xmath55 tau lepton and large missing transverse momentum is presented in section @xmath67 .
a summary is given in section @xmath68 .
the standard model processes leading to tau lepton production in electron
proton collisions are briefly outlined in this section , together with the dominant background processes . for each process , @xmath69 is used to label the hadronic final state , excluding the tau decay products .
the following processes , denoted henceforth as signal , lead to events containing genuine tau leptons in the final state : * * tau pair production : @xmath70 * + the production of tau pairs proceeds mainly via photon
photon collisions @xcite , as shown in figure [ diag]a .
the proton can remain intact ( elastic production , which dominates ) or be dissociated in the interaction ( inelastic production ) .
the incident electron is usually scattered at small angles and is often not observed in the main detector . only a small fraction of the total cross section is visible in the detector , as the cross section steeply falls with the transverse momentum @xmath71 of the tau leptons .
the cross section is about 20 pb for @xmath71 @xmath72 @xmath64 gev .
this process is modelled using the grape @xcite monte carlo ( mc ) generator .
* * production of @xmath54 bosons : @xmath73 * + the dominant production mechanism for @xmath54 bosons in @xmath0 collisions @xcite is depicted in figure [ diag]b .
the cross section is largest in the photoproduction regime ( photon proton collisions ) , in which the @xmath54 boson usually has small transverse momentum and the scattered electron is not observed in the main detector .
the @xmath54 production cross section at hera is about 1 pb , resulting in a cross section of about 0.1 pb for the process @xmath74 .
this process is modelled using the epvec generator @xcite .
the next to leading order qcd corrections to @xmath54 production @xcite are taken into account by weighting the events as a function of the rapidity and transverse momentum of the @xmath54 boson @xcite .
the following processes , denoted henceforth as background , do not contain genuine tau leptons but contribute to the selected samples through misidentification or mismeasurement : * * neutral current deep inelastic scattering ( nc dis ) : @xmath75 * + the scattered electron , or a quark or gluon that hadronises into a collimated jet of low particle multiplicity , may fake the hadronic tau decay signature . missing transverse momentum may occur in nc dis events because of fluctuations in the detector response and limited geometrical acceptance .
the rapgap @xcite generator is used to calculate this contribution to the background , including diffractive processes . * * photoproduction of jets : @xmath76 * + photoproduction processes may contribute to the background if a jet in the hadronic final state is misidentified as a hadronic tau decay . as in nc
dis events , missing transverse momentum may be measured due to fluctuations in the detector response and limited geometrical acceptance .
the pythia generator @xcite is used to calculate the contribution from non - diffractive hard scattering photoproduction processes .
the contribution from diffractive photoproduction processes is simulated using the rapgap generator . *
* electron or muon pair production : @xmath77 * + events with two leptons ( @xmath78 , @xmath79 or @xmath80 ) may mimic events containing @xmath81 pairs in which both tau leptons decay leptonically .
the production of electron or muon pairs may also contribute as background if one electron or muon is not identified and fakes the signature of a hadronic tau decay .
this contribution is calculated using the grape generator . * * charged current deep inelastic scattering ( cc dis ) : @xmath82 * + cc dis events contain genuine large missing transverse momentum due to the presence of a neutrino in the final state .
a jet originating from the struck quark may fake the hadronic tau decay signature .
this process constitutes the main background contribution to the @xmath58 analysis .
the django @xcite generator is used to calculate the background contribution from cc dis processes .
all generated events are passed through the full geant @xcite based simulation of the h1 apparatus and are reconstructed using the same procedure that is applied to the data .
at hera , electrons of energy @xmath83 gev collide with protons of energy @xmath84 or @xmath85 gev , corresponding to a centre of mass energy @xmath86 of @xmath87 or @xmath88 gev , respectively .
the analysed datasets consist of @xmath89 pb@xmath62 of @xmath90 collisions at @xmath91 gev ( taken in the period 1994 - 1997 ) , @xmath92 pb@xmath62 of @xmath90 collisions at @xmath93 gev ( 1999 - 2000 ) and @xmath94 pb@xmath62 of @xmath95 collisions at @xmath93 gev ( 1998 - 1999 ) , corresponding to a total integrated luminosity of @xmath96 pb@xmath62 .
the data recorded in the period 1994 - 1995 are not included for the measurement of tau pair production , reducing the integrated luminosity to @xmath97 pb@xmath62 for this analysis .
a detailed description of the h1 detector can be found in @xcite .
only the components essential to the present analysis are described here .
a right handed cartesian coordinate system is used with the origin at the nominal primary @xmath0 interaction vertex .
the proton beam direction defines the @xmath98 axis .
the polar angles @xmath99 and transverse momenta @xmath55 of all particles are defined with respect to this axis .
the azimuthal angle @xmath100 defines the particle direction in the transverse plane .
the pseudorapidity is defined as @xmath101 .
the inner tracking system contains the central ( @xmath102 ) and forward ( @xmath103 @xmath104 @xmath99 @xmath104 @xmath105 ) drift chambers , and the proportional chambers which are employed for triggering purposes .
it is used to determine the position of the interaction vertex and to measure the trajectories of charged particles .
particle transverse momenta and charges are determined from the curvature of the trajectories in a solenoidal magnetic field of @xmath106 t. hadronic final state particles as well as electrons and photons are absorbed in the highly segmented liquid argon ( lar ) calorimeter @xcite ( @xmath107 ) which is @xmath108 to @xmath109 radiation lengths deep depending on the polar angle .
the hadronic section of the lar calorimeter is @xmath66 to @xmath110 hadronic interaction lengths deep .
electromagnetic shower energies are measured with a precision of @xmath111 and hadronic shower energies with @xmath112 , as determined in test beam measurements @xcite . in the backward region ( @xmath113 ) , the lar calorimeter is complemented by a lead scintillating fibre spaghetti calorimeter ( spacal ) @xcite .
the lar and spacal calorimeters are enclosed within the superconducting coil and are surrounded by an iron return yoke which is instrumented with streamer tubes to allow for the detection of muons in the range @xmath114 . dissociated proton states may be measured at small polar angles by a set of detectors in the forward direction : the plug , a sandwich calorimeter constructed from copper plates and silicon counters ( @xmath115 ) , the proton remnant tagger ( prt ) , an array of scintillators ( @xmath116 @xmath104 @xmath99 @xmath104 @xmath117 ) , and the drift chambers of the forward muon detector ( fmd ) @xcite ( @xmath118 @xmath104 @xmath99 @xmath104 @xmath119 ) .
the luminosity is determined from the rate of bethe - heitler events @xmath120 , where the photon is detected in a calorimeter located downstream of the interaction point .
the events selected in this analysis are triggered by detecting electromagnetic clusters in the lar or spacal calorimeter ( electron trigger ) , by measuring a large missing transverse momentum in the lar calorimeter ( @xmath56 trigger ) , or by using hits in the muon detectors combined with central tracker signals ( muon trigger ) . in the kinematic range of this analysis , events containing a pair of tau leptons
are triggered with an efficiency of about @xmath121 ( @xmath122 ) if one tau lepton decays leptonically ( both tau leptons decay hadronically ) .
events containing single tau leptons and missing transverse momentum above @xmath123 gev ( @xmath124 gev ) are triggered with an efficiency of about @xmath125 ( @xmath126 ) .
an electron candidate is defined by the presence of a compact and isolated electromagnetic cluster of energy deposits in the lar or spacal calorimeter .
the kinematics of the electron candidate are measured from the calorimeter cluster . among the charged tracks reconstructed in the event
, the track with the lowest extrapolated distance of closest approach to the cluster ( @xmath127 ) is associated to the electron if it satisfies the condition @xmath128 cm . in this case , the azimuthal angle @xmath100 and the charge of the electron are measured from the associated track .
additional energy within a cone of radius @xmath129 in the pseudorapidity azimuth ( @xmath130@xmath100 ) plane around the electron candidate is required to be less than @xmath131 of the energy attributed to the electron candidate .
the electron identification efficiency is estimated using nc dis events and is greater than @xmath132 for an electron energy above @xmath66 gev .
a muon candidate is identified by a track in the inner tracking system associated with a track segment or an energy deposit in the instrumented iron .
the muon momentum is measured from the track curvature in the solenoidal magnetic field .
the rejection of hadrons traversing the calorimeter and reaching the muon detectors is improved by requiring that the muon candidate deposits less than @xmath66 gev around its extrapolated track in the lar calorimeter within a cylinder of radius @xmath133 cm in the electromagnetic and @xmath134 cm in the hadronic section . the efficiency to identify muons
is estimated using elastic @xmath135 events and is greater than @xmath126 in the energy range considered in this analysis .
the hadronic final state ( hfs ) is measured @xcite by combining energy measurements from the calorimeter with charged particle momenta measured by the inner tracker .
identified isolated electrons or muons are excluded from the hfs .
the hadronic energy scale is calibrated by comparing the transverse momentum of well measured electrons to that of the hfs in nc dis events @xcite .
jets with @xmath136 gev are reconstructed from hfs particles using an inclusive @xmath137 algorithm @xcite in the @xmath55 recombination scheme with a separation parameter set to one .
leptonic tau decays ( branching ratio @xmath122 @xcite ) are identified by detecting an electron or a muon as described above .
hadronic tau decays typically produce low multiplicity , collimated , hadronic jets , henceforth denoted as @xmath138-jets .
depending on the number of charged hadrons produced , the hadronic decay modes are summed up in two categories , referred to as `` 1-prong '' ( one charged hadron , branching ratio @xmath139 ) and `` 3-prong '' ( three charged hadrons , branching ratio @xmath140 ) .
the branching ratio for decays into more than three charged hadrons is small ( about @xmath141 ) and such decays are not considered in the present analysis .
two different algorithms to identify hadronic tau decays , applied to jets reconstructed in the angular range @xmath142 , have been developed and are described below .
the measurement of tau lepton pair production , in which the tau leptons generally have low momentum , requires an optimal background rejection .
in contrast , the search for tau leptons produced in @xmath54 decays at high @xmath55 uses an algorithm that maximises the identification efficiency , since the signal cross section is low and the background is less severe . for the measurement of the tau lepton pair production
, an algorithm has been developed that uses multiple neural networks to discriminate between @xmath138jets and the background from electrons , muons or hadronic jets .
a detailed description of the algorithm can be found in @xcite .
the algorithm is implemented in two steps . in the first step , 1prong ( 3prong ) candidates for hadronic tau decays
are preselected by requiring exactly one ( three ) well reconstructed track(s ) in the drift chambers within a cone ( `` @xmath138-cone '' ) around the jet axis .
the opening angle of the @xmath138cone varies between 5@xmath143 and 30@xmath143 depending on the jet momentum , with smaller angles at higher momentum due to the larger lorentz boost in the direction of the tau candidate .
the tracks are required to be not associated to identified electrons and muons and the scalar sum of their transverse momenta is required to be larger than @xmath64 gev .
the fine granularity of the lar calorimeter is used to match extrapolated tracks with energy deposits in the calorimeter and to reconstruct additional neutral particles associated to the tau candidate from unmatched energy deposits in the @xmath138cone .
the sum of the four vectors of the tracks and of the neutral particles defines the @xmath138jet four vector .
if all associated tracks have a well measured charge , the charge of the tau candidate is reconstructed as the sum of the charges of the associated tracks . in the second step of the algorithm ,
various variables related to the particle multiplicity and collimation of the @xmath138jet candidate are used .
the set of variables includes : the multiplicity of the neutral clusters within the @xmath138cone ; the invariant masses calculated from clusters , from tracks and from charged and neutral particles in the @xmath138cone ; the difference in energy measured from tracks and from clusters ; the distance in @xmath144 between tracks and clusters ; the first and second transverse moments of the distribution of energy deposits in the calorimeter with respect to the jet axis ; the sum of energy deposits detected in an extended cone of radius @xmath145 in @xmath130-@xmath100 around the @xmath138cone . two neural networks ( nn ) using these variables are separately trained using mc simulations to identify 1-prong and 3-prong tau decay modes , the output of which is denoted by @xmath146 and @xmath147 , respectively .
their numerical value varies between zero and one and is used to discriminate between tau candidates ( close to one ) and hadronic jets ( close to zero ) .
depending on the track multiplicity of the jet , the output of one of these nns is used to select tau candidates . by requiring @xmath148 ( for 1prong candidates ) or @xmath149 ( for 3prong candidates )
, the efficiency of this algorithm to identify hadronic tau lepton decays in @xmath57 events is about @xmath125 , as calculated from mc simulations of tau decays with visible energy in the range considered in this analysis .
the probability for hadronic jets to be misidentified as hadronic 1prong ( 3prong ) tau decays is @xmath150 ( @xmath151 ) .
the signature of hadronic 1prong tau decays may also be faked by unidentified electrons and muons .
two additional neural networks are trained to veto such cases .
the output of these nns , denoted by @xmath152 and @xmath153 respectively , is expected to be close to one for tau decays and close to zero for electrons or muons that fail the identification criteria described in sections 4.1 and 4.2 . a different approach to tau identification
is used to search for events with a high momentum tau lepton and large @xmath56 .
a high identification efficiency and a sufficient level of background rejection are achieved by requiring a collimated jet , containing only one charged particle and isolated with respect to other tracks and jets within a cone of radius @xmath145 in @xmath130@xmath100 .
the identification of hadronic @xmath138-decays is therefore restricted to hadronic 1prong decay modes .
the collimation of a jet is measured by the jet radius @xmath154 , defined as the energy weighted average distance in @xmath130@xmath100 between the jet axis and all hfs particles @xmath155 ( neutral and charged ) in the jet : @xmath156 jets with one track and @xmath157 are selected as tau candidates .
the four vector of the tau candidate is taken to be that of the jet . for jets with @xmath158 gev , this identification procedure results in an efficiency of about @xmath159 to identify 1prong hadronic decays of tau leptons resulting from decays of @xmath54 bosons .
the misidentification probability for hadronic jets is less than @xmath160 .
the decay modes investigated in the present analysis are classified as _ leptonic _
when both taus decay leptonically ( branching ratio @xmath161 , excluding decays to same flavour leptons ) , when one tau decays hadronically and the second leptonically ( @xmath162 ) and _ hadronic _ when both taus decay hadronically ( @xmath163 ) .
the case where both tau leptons decay to charged leptons of the same flavour ( @xmath78 or @xmath164 final state ) is not included in the present analysis , as the separation of the @xmath81 signal from electron or muon pair production is difficult .
leptonic and hadronic tau decays are identified as described in the previous section and by applying in addition the following selection criteria .
the isolation of electrons and muons is measured by the distance in the @xmath144 plane to the closest hadronic jet ( @xmath165 ) and to the closest track ( @xmath166 ) . *
@xmath167 candidates are reconstructed in the polar angular region @xmath168 and are required to have an energy above 5 gev and a transverse momentum above 3 gev .
they must be isolated from jets with @xmath169 .
* @xmath170 candidates are reconstructed in the polar angular region @xmath171 and are required to have a transverse momentum above 2 gev .
they must be isolated from other tracks and jets with @xmath172 and @xmath169 , respectively .
* @xmath138jet candidates of transverse momenta above 2 gev are reconstructed in the polar angular region @xmath173 , using algorithm * a * as described in section [ sec : tauid ] .
the output of the neural network is required to satisfy @xmath148 or @xmath149 .
events with two tau candidates are selected .
if the charge of both tau candidates is measured , events with two candidates of the same charge are rejected . in order to avoid significant background contributions from nc dis and @xmath174 processes ,
the analysis is restricted to elastic @xmath175 production .
inelastic events are vetoed by requiring no extra track or energy deposit above the noise level in the main detector in addition to those associated to the decay products of both tau leptons and a possible additional electron .
furthermore , no significant activity should be observed in the forward detectors ( plug , fmd and prt ) .
the requirements applied @xcite ensure that the proton remains intact or dissociates into a low mass state .
remaining background originates mainly from elastic lepton pair production and from diffractive nc dis or @xmath174 processes .
electron and muon pair production processes ( @xmath176 and @xmath177 ) constitute a background to tau pair production in semi
leptonic decay modes when one of the leptons is correctly identified as @xmath167 or @xmath170 while the second lepton fails to be identified by the algorithms described in sections [ sec : elid ] and [ sec : muid ] and fakes the signature of a 1prong hadronic tau decay . in order to reject this background , @xmath178 ( @xmath179 )
is required for 1prong tau jet candidates selected in the @xmath180jet ( @xmath181jet ) channel . to further reduce the background from nc dis processes , where the scattered electron is selected as a candidate for a @xmath182 decay and the struck quark fragments into a collimated jet of low particle multiplicity that fakes the @xmath138jet signature , the longitudinal momentum balance calculated as @xmath183 is employed , where @xmath184 and @xmath185 denote the energy and polar angle of each particle detected in the event . for events in which only momentum in the proton direction is undetected , @xmath186 is equal to twice the energy @xmath187 of the electron beam , i.e. @xmath188 gev . for events containing tau leptons , @xmath186 values well below @xmath188 gev
are expected since the neutrinos produced in the tau decays are not detected and , most of the time , the scattered beam electron escapes down the beam pipe . by requiring @xmath189 gev when the detected electron has the same charge as the beam lepton , the nc dis background is rejected to a large extent .
the @xmath186 cut is also applied to the leptonic channel and rejects muon pair production events , for which the scattered electron is detected together with one produced muon while the second muon escapes in the forward direction .
if a second electron is detected in the event , this condition is not applied . * decay channel * & * leptonic * & & * hadronic * + & @xmath167 & @xmath167 & @xmath170 & @xmath138jet + & @xmath170 & @xmath138jet & @xmath138jet & @xmath138jet + & + & + & + & & @xmath190 & @xmath191 & + & & & + + + + the selection criteria for @xmath81 events are summarised in table [ tabletaupaireventselection ] . with these selection criteria , @xmath192 of the elastic @xmath193 events in which both tau leptons satisfy @xmath194 gev and @xmath195 are selected .
the efficiency is limited by the fact that the energy of the detected tau decay products is significantly lower than the tau lepton energy , since the neutrinos escape detection .
due to the aforementioned elasticity requirements , the remaining background from nc dis and @xmath174 processes consists mainly of exclusive diffractive events , for which the validity of the resolved pomeron model @xcite as implemented in the rapgap program is questionable .
hence the rapgap prediction for nc dis ( @xmath174 ) is normalised to the number of events observed in a control sample in which an electron and a jet ( two jets ) are selected in the @xmath55 and @xmath99 ranges of the analysis , and where the elastic requirements are applied .
it has been verified that the shapes of the observed kinematic distributions are reasonably well described by rapgap .
furthermore , to check that the probability for an electron , a muon or a hadronic jet to be misidentified as a @xmath138jet candidate is well described by the mc simulation , several event samples are studied in which the contribution of individual background processes is enhanced .
these event samples are selected in a phase space similar to that of the @xmath81 event sample , requiring the conditions for elastic production to be fullfilled . in order to enhance the background component in the control samples ,
the condition @xmath148 or @xmath149 on the nn outputs for @xmath138jet candidates is not applied .
the following samples are employed : * * control sample * + a @xmath196 dominated event sample is defined by selecting events with one electron and one 1prong @xmath138jet candidate .
no veto condition on @xmath152 is applied .
events with two same charge tau candidates are rejected .
* * control sample * + a @xmath135 enriched event sample is defined by selecting events with one muon and one 1prong @xmath138jet candidate .
no veto condition on @xmath153 is applied .
events with two same charge tau candidates are rejected .
* * control sample * + a @xmath174 enriched event sample is defined by selecting events containing two jets with one , two or three tracks .
if both jets are 1prong tau candidates , the requirements @xmath197 and @xmath198 are applied to suppress the contributions from @xmath196 and @xmath135 processes . in order to reduce the contribution from nc dis process , only events with @xmath199 gev are accepted . * * nc dis control sample * + a nc dis enriched event sample is defined by selecting events with one electron and one @xmath138jet candidate . if the @xmath138jet is a 1prong candidate , @xmath200 is required in order to suppress the contribution from @xmath196 processes .
the selection criteria of the background control samples are summarised in table [ tab : ttcontrol ] , in which the numbers of events obtained from the data and the mc simulation are also given . & & & & * nc dis * + & & & & + & @xmath167 & @xmath170 & two jets & @xmath167 + & 1prong @xmath138jet & 1prong @xmath138jet & with @xmath201 tracks & @xmath138jet + & & & + & & & + & + & + & & & @xmath190 & @xmath202 + & & & @xmath191 & + & & & @xmath199 gev & + h1 data & @xmath63 & @xmath108 & @xmath203 & @xmath203 + total sm & @xmath204 & @xmath205 & @xmath206 & @xmath207 + & ( @xmath208% @xmath209 ) & ( @xmath210% @xmath211 ) & ( @xmath212% @xmath174 ) & ( @xmath213% nc dis ) + + + the distributions of all quantities used in the selection of the @xmath81 event sample are well described by the mc simulation both in shape and normalisation in the control samples .
examples of these distributions are shown in figure [ figuretaupairbackgroundenrichedsamples ] .
the distributions of the electron and muon rejection discriminators @xmath214 and @xmath215 are shown in figures [ figuretaupairbackgroundenrichedsamples]a and [ figuretaupairbackgroundenrichedsamples]b for the control samples where most of the tau candidates are unidentified electrons and muons , respectively .
the distribution of the nn output @xmath146 ( @xmath147 ) is shown in figure [ figuretaupairbackgroundenrichedsamples]c ( [ figuretaupairbackgroundenrichedsamples]d ) for 1prong ( 3prong ) jets in the @xmath174 control sample which is enriched in hadronic jets . in figures
[ figuretaupairbackgroundenrichedsamples]a d the contribution from tau pair production populates the region close to one , while the background accumulates at values close to zero , as expected .
finally , the distribution of the @xmath186 variable is shown in figure [ figuretaupairbackgroundenrichedsamples]e for the nc dis control sample .
the agreement between data and simulation in the control samples shows that the background contribution as well as the experimental efficiencies are modelled by the mc simulation within the attributed systematic uncertainties described in the next section .
( 150,189)(0,0 ) ( 0.,0 . )
( 70.,10 . )
( 0.,63 . )
( 70.,63 . )
( 0.,126 . )
( 70.,126 . ) in this section , the systematic uncertainties associated with the measurement of elastic @xmath81 production are discussed .
the uncertainties on the signal expectation are determined by varying the following experimental quantities by @xmath216 standard deviation . * * electron identification and reconstruction * + in the kinematic region considered in the @xmath81 analysis , the uncertainty on the electron identification efficiency is @xmath217 ( @xmath141 when the electron energy is above @xmath218 gev ) .
the electron energy scale uncertainty is estimated to be @xmath131 .
the uncertainties on the measurement of the electron angles @xmath99 and @xmath100 are @xmath65 mrad and @xmath219 mrad , respectively . *
* muon identification and reconstruction * + the uncertainty on the muon identification efficiency is @xmath217 . the relative scale uncertainty on the muon momentum is conservatively taken to be @xmath217 .
the uncertainties on the measurements of the muon angles @xmath99 and @xmath100 are @xmath65 mrad and @xmath219 mrad , respectively . *
* identification and reconstruction of hadronic tau decays * + for each of the charged tracks associated to the @xmath138jet , a reconstruction efficiency uncertainty of @xmath131 is assigned . the efficiency to identify hadronic
@xmath138 decays with the nn algorithm * a * has an additional uncertainty of @xmath220 @xcite , estimated by comparing different simulations of shower development in the lar calorimeter .
it has been verified , using a large statistics sample of hadronic jets from inclusive cc dis and nc dis samples , that the output of both nns is well described by the mc simulation .
the uncertainties arising from the modelling of tau decays are negligible , as estimated by comparing the results obtained using either the pythia @xcite or tauola @xcite programs to simulate the decays of the tau leptons generated with grape .
the energy of the neutral clusters of @xmath138jet candidates has a relative uncertainty of @xmath151 , corresponding to the hadronic energy scale uncertainty .
the uncertainties on the measurements of the @xmath138jet angles @xmath99 and @xmath100 are estimated to be @xmath218 mrad and @xmath66 mrad , respectively . * * elastic event selection * + the efficiency with which elastic events are selected and proton dissociative events are rejected depends on the noise level in the lar and on the performance of the forward detectors @xcite .
its uncertainty does not exceed @xmath131 . * * triggering * + the trigger efficiency is studied using elastic @xmath221 and @xmath222 events and diffractive @xmath174 di - jet events in a phase space similar to that of the @xmath81 analysis .
the uncertainty on the trigger efficiency depends on the region in @xmath99 in which the tau candidates are detected : central ( @xmath223 ) or forward ( @xmath224 ) .
the uncertainty is @xmath220 if both tau candidates are in the central region , @xmath225 if one tau candidate is detected in the forward and the other in the central region and @xmath226 if both tau candidates are detected in the forward region . * * luminosity * + the luminosity of the analysed datasets is measured with an uncertainty of @xmath227 .
the individual effects of the above experimental uncertainties are combined in quadrature , yielding a total uncertainty of @xmath228 on the signal expectation .
the largest contributions to this uncertainty arise from systematics attributed to the tau identification procedure ( @xmath229 ) and to the trigger efficiency ( @xmath230 ) .
contributions from background processes , modelled using the generators described in section @xmath64 , are attributed relative systematic uncertainties of @xmath125 ( @xmath174 ) , @xmath226 ( nc dis ) , @xmath229 ( @xmath231 , @xmath232 ) , estimated from the level of agreement observed between the mc simulation and the data in the background enhanced control samples described in section [ sec : taupbgstudy ] .
( 150,135)(0,0 ) ( 0.,0 . )
( 85.,15 . )
( 0.,70 . )
( 75.,70 . ) in total , @xmath109 @xmath81 candidate events are selected , in agreement with the sm expectation of @xmath233 events , of which @xmath234 are expected from @xmath193 signal processes . the signal expectation is dominated ( @xmath126 ) by the elastic production component .
the lepton pair production , nc dis and @xmath174 processes contribute with similar rates to the background expectation .
the numbers of observed and expected events in the four analysed channels are shown in table [ tab : taupaires ] .
an event selected in the semi
leptonic channel is displayed in figure [ event ] .
& * leptonic * & & * hadronic * & * total * + & @xmath167 @xmath170 & @xmath167 @xmath138jet & @xmath170 @xmath138jet & @xmath138jet @xmath138jet & + h1 data & @xmath68 & @xmath64 & @xmath218 & @xmath235 & @xmath109 + sm & @xmath236 & @xmath237 & @xmath238 & @xmath239 & @xmath240 + @xmath57 & @xmath241 & @xmath242 & @xmath126 & @xmath125 & @xmath243 + the distributions of the polar angle and of the transverse momentum of the tau candidates , together with the visible invariant mass , are shown in figure [ figuretaupairresults1 ] .
the measured distributions are compatible with the sm expectations .
as expected for @xmath193 processes , most tau decay products are detected with relatively small transverse momenta .
[ sec : tpxsection ] using the selected sample , a measurement of the cross section for the elastic production of @xmath81 pairs is performed in the kinematic region defined by @xmath244 and @xmath245 gev . for this measurement ,
the data samples collected at @xmath246 gev and @xmath88 gev are combined , taking into account their respective luminosities . assuming a linear dependence of the cross section on the proton beam energy , as predicted by the sm , the measured cross section corresponds to an effective centre of mass energy of @xmath247 gev .
the cross section is calculated using the formula : @xmath248 where @xmath249 is the number of observed events , @xmath250 the expected contribution from background processes , @xmath251 the total integrated luminosity and @xmath252 the signal acceptance .
the contribution from inelastic @xmath193 processes is included in the background expectation .
the signal acceptance @xmath252 is calculated using the grape generator , as the ratio of the number of events accepted at reconstructed level to the number of events generated in the defined phase space .
it accounts for the selection and trigger efficiencies and for the differences in momentum between the original @xmath138 leptons and the detected decay products .
the measured cross section for elastic tau pair production @xmath253 integrated over the phase space defined above is @xmath254 where the first error is statistical and the second systematic .
the result is in agreement with the sm expectation of @xmath255 , calculated using the grape generator .
events containing an isolated tau lepton and large missing transverse momentum are selected with a procedure similar to that used in the analysis of events with an isolated electron or muon and large missing transverse momentum @xcite .
the tau leptons are identified using hadronic decays only , as the leptonic tau decays lead to final states which can not be distinguished from those studied in @xcite .
the event selection is performed in two steps . in the first step ,
the net transverse momentum reconstructed from all particles ( electrons , muons and hadrons ) @xmath256 , is required to be above @xmath123 gev . in order to ensure uniform trigger conditions , the net transverse momentum measured from all energy deposits detected in the calorimeter , @xmath257 , is required to be above @xmath123 gev .
the reconstructed @xmath256 is approximately equal to @xmath257 except for events containing muons in the final state . in order to exploit further the event topology in the transverse plane ,
the variable @xmath258 is employed , defined as the ratio of the anti parallel to parallel projections of all energy deposits in the calorimeter with respect to the direction of @xmath257 @xcite .
events with genuine missing transverse momentum are in general reconstructed with @xmath258 values close to zero and large values of @xmath257 , whereas background events from nc dis and @xmath174 processes are intrinsically balanced , with larger values of @xmath258 and low values of @xmath257
. only events with @xmath258 below @xmath129 are accepted . for events with @xmath259 below @xmath124
gev , a stricter condition of @xmath260 is applied .
the background from nc dis is further reduced by requiring @xmath186 @xmath104 @xmath210 gev .
additionally , the events are required to contain at least one isolated hadronic jet with transverse momentum above 7 gev in the central region of the detector@xmath261 @xmath104 @xmath262 @xmath104 @xmath263 .
the isolation is characterised by the distance of the jet in @xmath144 to the nearest hadronic jet ( @xmath264 ) and the nearest track not belonging to the jet ( @xmath265 ) . in the second selection step ,
the isolated jets are required to pass the tau identification criteria of algorithm * b * described in section [ sec : tauid ] .
a jet is accepted if it is narrow ( @xmath266 ) and contains exactly one charged track ( @xmath267 ) . in order to remove background from hadronic jets containing mostly neutral particles ,
this track is required to have a transverse momentum @xmath268 greater than @xmath66 gev .
if more than one isolated jet satisfies these requirements , the one with the highest @xmath55 is considered as the tau candidate . in order to further reduce the background from intrinsically balanced events , in which the @xmath269 is due to a mismeasurement , the acoplanarity @xmath270 , defined as the angle in the transverse plane between the @xmath138jet candidate and the hadronic system excluding the tau candidate ( @xmath69 ) , is required to be below @xmath271 .
this criterium removes events in which the @xmath138jet candidate and the rest of the hadronic system are back to back , as is typical for nc dis and @xmath174 events .
a summary of all selection criteria is presented in table [ taucuts ] . using these selection criteria , sm @xmath272 events are selected with an overall efficiency of @xmath273 . in comparison to @xmath54 decays into electrons or muons @xcite , @xmath274 decays are selected with a significantly lower efficiency , mainly due to the branching ratio ( @xmath139 ) for hadronic 1prong tau decays and to the more restricted polar angular range of this analysis . &
@xmath275 & @xmath276 @xmath123 gev + & @xmath269 & @xmath276 @xmath123 gev + & @xmath186 & @xmath277 @xmath210 gev + & @xmath258 & @xmath278 ( @xmath279 gev ) + & @xmath280 & @xmath276 @xmath219 + & @xmath281 & @xmath276 @xmath68 gev + & @xmath261 & @xmath282 @xmath283 + & @xmath284 & @xmath285 + & @xmath286 & @xmath285 + & @xmath287 & @xmath288 + & @xmath268 & @xmath289 gev + & @xmath154 & @xmath290 + & @xmath270 & @xmath291 + after applying the selection criteria , the main background is expected to occur from events with genuine missing transverse momentum produced by cc dis processes , in which a narrow jet with low track multiplicity fakes the tau signature .
additional background arises from nc dis and @xmath174 processes , which have a much larger cross section than the cc dis process and lead to events that contain hadronic jets but no genuine @xmath269 .
the modelling of the cc dis , nc dis and @xmath174 backgrounds is verified using samples in which the contribution of each background process is enhanced .
the selection criteria defining the background control samples are listed in table [ backgrcut ] , together with the observed number of events and the sm expectation .
the cc dis background control sample is selected using the `` @xmath269 + isolated jet '' preselection described in section [ sec : eventsel ] .
the nc dis and @xmath174 samples are selected in a complementary phase space at large @xmath186 and large @xmath258 , respectively . in these background samples
the tight selection criteria of tau candidates are not applied and only isolated jets , as defined in table [ taucuts ] , are considered . as shown in figure [ backgr]a c ,
the distributions of @xmath154 , @xmath292 , and @xmath268 in the cc dis control sample are well described by the mc simulation , both in shape and normalisation .
the distribution of the acoplanarity angle @xmath270 in the control sample enriched in @xmath174 events , shown in figure [ backgr]d , shows a clear peak towards 180@xmath293 , corresponding to back to back events . in figure [ backgr]e the distribution of the hadronic transverse momentum @xmath294 for the nc dis control sample
the good agreement between data and mc simulation observed in all control samples confirms the good understanding of background rates and of the properties of the jets ( shape and multiplicity ) used in the tau identification procedure .
& & & + & & & + & + & @xmath295 gev & @xmath296 gev & @xmath297 gev + & @xmath298 & @xmath299 gev & @xmath300 + & @xmath301 gev & @xmath302 & no electrons + data & @xmath303 & @xmath304 & @xmath305 + total sm & @xmath306 & @xmath307 & @xmath308 + & ( 93% cc ) & ( 98% nc ) & ( 80% @xmath309 ) + ( 150,165)(0,0 ) ( 0,110 ) ( 70,110 ) ( 0,55 ) ( 70,55 ) ( 0,0 ) ( 70,10 ) in this section , the systematic uncertainties associated with the search for events containing an isolated tau lepton and large missing transverse momentum are discussed .
the effect of those uncertainties on the expectations from sm w production and from background processes is determined by varying the experimental quantities by @xmath216 standard deviation . * * tau identification procedure * + the main experimental uncertainty on the signal and background expectations arises from the requirement of exactly one track within the jet and from the condition on the jet radius @xmath157 .
the energy of charged tracks is reconstructed with an uncertainty of @xmath217 .
the measurements of the polar and azimuthal angles are attributed a systematic error of 3 mrad and 1 mrad respectively .
the uncertainty on the track reconstruction efficiency is @xmath131 and has a relatively large effect on the background expectation , due to migrations of hadronic jets with multiple charged particles into the single track category .
the modelling of @xmath154 is studied with high statistical precision using jets in a dedicated inclusive nc dis sample .
the measured and simulated jet radius distributions are compared and an uncertainty depending on the polar angle of the @xmath138jet is attributed to @xmath154 .
the uncertainties associated with the tau identification are @xmath229 for the expected @xmath54 signal and @xmath310 for the background predictions . * * hadronic final state reconstruction * + the hadronic energy measurement has a relative uncertainty of @xmath151 .
the polar angle measurement has an uncertainty varying from 3 mrad for jets reconstructed in the central region to 10 mrad for forward jets .
the topological variable @xmath258 is attributed a relative uncertainty of @xmath220 . * * trigger * + the uncertainty on the @xmath269 trigger efficiency is @xmath217 deduced from a large statistics nc dis sample for which the trigger information is reconstructed offline , ignoring the signal from the electron @xcite . * * luminosity * + the luminosity of the analysed datasets is measured with an uncertainty of @xmath227 . * * theoretical errors on signal and background contributions * + the uncertainty on the @xmath54 production signal cross section is estimated to be @xmath229 @xcite .
the expectations from nc dis and @xmath174 processes are each attributed an additional uncertainty of @xmath225 , a value which has been estimated from the control samples described in section [ sec : backgstudy ] .
this uncertainty covers the sensitivity of the jet radius and multiplicity ( used in the tau identification algorithm ) to the modelling of parton showers in nc dis and @xmath174 mc samples .
the individual effects of the experimental uncertainties are combined in quadrature to give the total experimental systematic uncertainty .
the total uncertainty on the sm w signal is @xmath225 and that on the sm background is @xmath311 . for both signal and background processes ,
the total uncertainty is dominated by the uncertainty arising from the tau identification procedure and the theoretical uncertainties . in the final event sample @xmath67 events
are observed in the data , compared to a total sm expectation of @xmath312 @xmath313 events , of which @xmath314 are expected from sm @xmath54 production .
the @xmath315 spectrum and other properties of the events in the final sample are shown in figure [ isotaufinal ] .
table [ isotauresults ] summarises the results .
the events observed in the data are concentrated in the region of very low @xmath316 , where the contribution from cc dis background processes dominates the sm expectation . in the region
@xmath317 gev , where an excess of events containing isolated electrons or muons is observed @xcite , no event is found for a sm prediction of @xmath318 , of which @xmath319 are expected from sm @xmath54 production . in the absence of a signal ,
a model independent upper limit on the cross section for the production of events containing an isolated tau leptons and large missing transverse momentum is set in the kinematic region : @xmath320 @xmath104 @xmath321 @xmath104 @xmath322 , @xmath323 gev and @xmath324 gev . the limit is calculated using a modified frequentist approach based on likelihood ratios @xcite and taking into account the systematic uncertainties discussed in section [ sec : syserr ] . the acceptance for processes producing isolated tau leptons in events with large missing transverse momentum in the given kinematic region is estimated using the mc simulation for sm @xmath54 production , implemented in the epvec generator .
an additional model uncertainty of @xmath220 is attributed to the acceptance .
this uncertainty is estimated by comparing the acceptance predicted by epvec with that obtained using the generator anotop , which simulates the anomalous production of single top quarks in @xmath0 collisions at hera @xcite and produces @xmath54 bosons with different kinematic distributions distribution for top events is peaked at about @xmath325 gev , while events from sm w production are concentrated at low @xmath316 values . ] .
an upper limit of @xmath326 pb at @xmath132 confidence level is obtained for the production cross section of events containing an isolated tau lepton and large missing transverse momentum in the phase space defined above . in the region
@xmath327 gev , the upper limit is @xmath328 pb at @xmath132 confidence level . these limits are higher than the cross sections measured in the electron and muon channels @xcite .
the present measurement is therefore compatible with the previous measurement of events with an electron or muon and @xmath269 , as expected if lepton universality is assumed .
.number of events observed in the data compared to the sm expectation for signal ( w decay into @xmath138 ) and background processes in the final @xmath329 sample . [ cols="^,^,^,^,^",options="header " , ] ( 150,140)(0,0 ) ( -10,70 ) ( 73,70 ) ( -10,0 ) ( 73,0 )
in this paper , the production of tau leptons in @xmath0 collisions at hera is investigated in events containing a @xmath81 pair and events containing an isolated tau lepton and large missing transverse momentum .
the production cross section of @xmath81 pairs is measured in elastic processes , in a combination of leptonic , semi leptonic and hadronic decay modes of the two tau leptons . in a data sample corresponding to an integrated luminosity of @xmath61 pb@xmath62 , @xmath109 events
are observed , in agreement with a standard model expectation of @xmath233 events , of which @xmath234 are expected from @xmath193 signal processes .
this is the first observation of tau pair production in @xmath0 collisions .
a search for the production of isolated tau leptons in events with large missing transverse momentum is performed in a data sample corresponding to an integrated luminosity of @xmath63 pb@xmath62 .
the selection yields @xmath67 candidate events , compatible with a standard model expectation of @xmath312 @xmath313 events .
no event is observed in the region @xmath330 , where @xmath318 events are expected , including @xmath331 events from @xmath332 decays .
an upper limit on the production cross section of @xmath333 @xmath104 @xmath334 pb at @xmath132 confidence level is set .
we are grateful to the hera machine group whose outstanding efforts have made this experiment possible .
we thank the engineers and technicians for their work in constructing and maintaining the h1 detector , our funding agencies for financial support , the desy technical staff for continual assistance and the desy directorate for support and for the hospitality which they extend to the non desy members of the collaboration .
grape 1.1 , t. abe , comput .
. commun .
* 136 * ( 2001 ) 126 [ hep - ph/0012029 ] .
diener , c. schwanenberger and m. spira , eur .
j. c * 25 * ( 2002 ) 405 ; p. nason , r. rckl and m. spira , j. phys .
g * 25 * ( 1999 ) 1434 [ hep - ph/9902296 ] ; + m. spira , proc . of the workshop `` monte carlo generators for hera physics '' ( 1991 ) , eds .
a. t. doyle , g. grindhammer , g. ingelman , h. jung , p. 623
[ hep - ph/9905469 ] .
pythia 5.7 , t. sjstrand , comput .
commun .
* 82 * ( 1994 ) 74 .
django 2.1 , g.a .
schler and h. spiesberger , proc . of the workshop `` physics at hera '' ( 1991 ) , eds .
w. buchmller and g. ingelman , vol .
3 , p. 1419 . geant 3 , r. brun _ et al .
_ , cern - dd / ee/84 - 1 .
i. abt _ et al . _
[ h1 collaboration ] , nucl .
instrum .
meth . a * 386 * ( 1997 ) 310 and 348 .
b. andrieu _ et al .
_ [ h1 calorimeter group ] , nucl . instrum .
a * 336 * ( 1993 ) 460 .
b. andrieu _ et al .
_ [ h1 calorimeter group ] , nucl . instrum
. a * 336 * ( 1993 ) 499 ; + b. andrieu _ et al . _
[ h1 calorimeter group ] , nucl .
instrum .
a * 344 * ( 1994 ) 492 ; + b. andrieu _ et al . _
[ h1 calorimeter group ] , nucl .
instrum .
a * 350 * ( 1994 ) 57 .
r. d. appuhn
_ [ h1 spacal group ] , nucl .
instrum .
a * 386 * ( 1997 ) 397 .
j. ban _ et al .
_ [ h1 bemc group ] , nucl .
instrum .
a * 372 * ( 1996 ) 399 .
i. r. kenyon _ et al .
_ [ h1 fmd group ] , nucl .
instrum .
a * 340 * ( 1994 ) 304 .
m. peez , `` recherche de dviations au
del du modle standard dans les processus de grande nergie transverse sur le collisionneur lectron proton hera '' , phd thesis , aix - marseille university , 2003 , + available at http://www-h1.desy.de/publications/[email protected] .
b. portheault , `` premire mesure des sections efficaces de courant charg et neutre avec le faisceau de positrons polaris hera ii et analyses qcd lectrofaibles '' , phd thesis , paris - sud xi university , 2005 , + available at http://www-h1.desy.de/publications/[email protected] .
s. d. ellis and d. e. soper , phys .
d * 48 * ( 1993 ) 3160 [ hep - ph/9305266 ] ; + s. catani , y. l. dokshitzer , m. h. seymour and b. r. webber , nucl .
b * 406 * ( 1993 ) 187 ; + c. adloff _ et al . _
[ h1 collaboration ] , nucl .
b * 545 * ( 1999 ) 3 [ hep - ex/9901010 ] .
s. eidelman _ et al . _
[ particle data group ] , phys .
b * 592 * ( 2004 ) 1 .
c. veelken , `` search for events with isolated leptons and large missing transverse momentum in @xmath0 collisions at hera '' , phd thesis , university of liverpool , 2006 , + available at http://www-h1.desy.de/publications/[email protected] .
p. laycock , `` a measurement of the diffractive reduced cross - section @xmath337 at high @xmath338 with the h1 detector at hera '' , phd thesis , university of liverpool , 2003 , + available at http://www-h1.desy.de/publications/[email protected] .
g. ingelman and p. e. schlein , phys .
b * 152 * ( 1985 ) 256 .
pythia 6.3 , t. sjstrand , l. lnnblad , s. mrenna and p. skands , hep - ph/0308153 .
tauola 2.7 , m. jezabek , z. was , s. jadach and j. h. khn , comp .
. commun .
* 64 * ( 1991 ) 275 . | the production of tau leptons in @xmath0 collisions is investigated using data recorded by the h1 detector at hera in the period 1994 - 2000 .
tau leptons are identified by detecting their decay products , using leptonic and hadronic decay modes .
the cross section for the production of tau lepton pairs is measured for the first time at hera .
furthermore , a search for events with an energetic isolated tau lepton and with large missing transverse momentum is performed .
the results are found to be in agreement with the standard model predictions .
# 1#2#3#4#1 * # 2 * ( # 3 ) # 4 desy 06 - 029issn 0418 - 9833 + april 2006 * tau lepton production in collisions at hera * h1 collaboration submitted to eur .
phys . j. c a. aktas@xmath1 , v. andreev@xmath2 , t. anthonis@xmath3 , b. antunovic@xmath4 , s. aplin@xmath1 , a. asmone@xmath5 , a. astvatsatourov@xmath3 , a. babaev@xmath6 , s. backovic@xmath7 , a. baghdasaryan@xmath8 , p. baranov@xmath2 , e. barrelet@xmath9 , w. bartel@xmath1 , s. baudrand@xmath10 , s. baumgartner@xmath11 , j. becker@xmath12 , m. beckingham@xmath1 , o. behnke@xmath13 , o. behrendt@xmath14 , a. belousov@xmath2 , n. berger@xmath11 , j.c .
bizot@xmath10 , m .- o .
boenig@xmath14 , v. boudry@xmath15 , j. bracinik@xmath4 , g. brandt@xmath13 , v. brisson@xmath10 , d. bruncko@xmath16 , f.w .
bsser@xmath17 , a. bunyatyan@xmath18 , g. buschhorn@xmath4 , l. bystritskaya@xmath19 , a.j .
campbell@xmath1 , f. cassol - brunner@xmath20 , k. cerny@xmath21 , v. cerny@xmath22 , v. chekelian@xmath4 , j.g .
contreras@xmath23 , j.a .
coughlan@xmath24 , b.e .
cox@xmath25 , g. cozzika@xmath26 , j. cvach@xmath27 , j.b .
dainton@xmath28 , w.d .
dau@xmath29 , k. daum@xmath30 , y. de boer@xmath19 , b. delcourt@xmath10 , m. del degan@xmath11 , a. de roeck@xmath31 , e.a .
de wolf@xmath3 , c. diaconu@xmath20 , v. dodonov@xmath32 , a. dubak@xmath33 , g. eckerlin@xmath1 , v. efremenko@xmath19 , s. egli@xmath34 , r. eichler@xmath34 , f. eisele@xmath13 , a. eliseev@xmath2 , e. elsen@xmath1 , s. essenov@xmath19 , a. falkewicz@xmath35 , p.j.w .
faulkner@xmath36 , l. favart@xmath3 , a. fedotov@xmath19 , r. felst@xmath1 , j. feltesse@xmath26 , j. ferencei@xmath16 , l. finke@xmath17 , m. fleischer@xmath1 , g. flucke@xmath5 , a. fomenko@xmath2 , g. franke@xmath1 , t. frisson@xmath15 , e. gabathuler@xmath28 , e. garutti@xmath1 , j. gayler@xmath1 , c. gerlich@xmath13 , s. ghazaryan@xmath8 , s. ginzburgskaya@xmath19 , a. glazov@xmath1 , i. glushkov@xmath37 , l. goerlich@xmath35 , m. goettlich@xmath1 , n. gogitidze@xmath2 , s. gorbounov@xmath37 , c. grab@xmath11 , t. greenshaw@xmath28 , m. gregori@xmath38 , b.r .
grell@xmath1 , g. grindhammer@xmath4 , c. gwilliam@xmath25 , d. haidt@xmath1 , l. hajduk@xmath35 , m. hansson@xmath39 , g. heinzelmann@xmath17 , r.c.w .
henderson@xmath40 , h. henschel@xmath37 , g. herrera@xmath41 , m. hildebrandt@xmath34 , k.h .
hiller@xmath37 , d. hoffmann@xmath20 , r. horisberger@xmath34 , a. hovhannisyan@xmath8 , t. hreus@xmath42 , s. hussain@xmath38 , m. ibbotson@xmath25 , m. ismail@xmath25 , m. jacquet@xmath10 , l. janauschek@xmath4 , x. janssen@xmath3 , v. jemanov@xmath17 , l. jnsson@xmath39 , d.p .
johnson@xmath3 , a.w .
jung@xmath43 , h. jung@xmath44 , m. kapichine@xmath45 , j. katzy@xmath1 , i.r .
kenyon@xmath36 , c. kiesling@xmath4 , m. klein@xmath37 , c. kleinwort@xmath1 , t. klimkovich@xmath1 , t. kluge@xmath1 , g. knies@xmath1 , a. knutsson@xmath39 , v. korbel@xmath1 , p. kostka@xmath37 , k. krastev@xmath1 , j. kretzschmar@xmath37 , a. kropivnitskaya@xmath19 , k. krger@xmath43 , m.p.j .
landon@xmath38 , w. lange@xmath37 , g. latovika - medin@xmath7 , p. laycock@xmath28 , a. lebedev@xmath2 , g. leibenguth@xmath11 , v. lendermann@xmath43 , s. levonian@xmath1 , l. lindfeld@xmath12 , k. lipka@xmath37 , a. liptaj@xmath4 , b. list@xmath11 , j. list@xmath17 , e. lobodzinska@xmath46 , n. loktionova@xmath2 , r. lopez - fernandez@xmath41 , v. lubimov@xmath19 , a .-
i .
lucaci - timoce@xmath1 , h. lueders@xmath17 , d. lke@xmath47 , t. lux@xmath17 , l. lytkin@xmath32 , a. makankine@xmath45 , n. malden@xmath25 , e. malinovski@xmath2 , s. mangano@xmath11 , p. marage@xmath3 , r. marshall@xmath25 , l. marti@xmath1 , m. martisikova@xmath1 , h .- u .
martyn@xmath48 , s.j .
maxfield@xmath28 , a. mehta@xmath28 , k. meier@xmath43 , a.b .
meyer@xmath1 , h. meyer@xmath49 , j. meyer@xmath1 , v. michels@xmath1 , s. mikocki@xmath35 , i. milcewicz - mika@xmath35 , d. milstead@xmath28 , d. mladenov@xmath50 , a. mohamed@xmath28 , f. moreau@xmath15 , a. morozov@xmath45 , j.v .
morris@xmath24 , m.u .
mozer@xmath13 , k. mller@xmath12 , p. murn@xmath51 , k. nankov@xmath50 , b. naroska@xmath17 , th .
naumann@xmath37 , p.r .
newman@xmath36 , c. niebuhr@xmath1 , a. nikiforov@xmath4 , g. nowak@xmath35 , k. nowak@xmath12 , m. nozicka@xmath21 , r. oganezov@xmath8 , b. olivier@xmath4 , j.e .
olsson@xmath1 , s. osman@xmath39 , d. ozerov@xmath19 , v. palichik@xmath45 , i. panagoulias@xmath1 , t. papadopoulou@xmath1 , c. pascaud@xmath10 , g.d .
patel@xmath28 , h. peng@xmath1 , e. perez@xmath26 , d. perez - astudillo@xmath23 , a. perieanu@xmath1 , a. petrukhin@xmath19 , d. pitzl@xmath1 , r. plaakyt@xmath4 , b. portheault@xmath10 , b. povh@xmath32 , p. prideaux@xmath28 , a.j .
rahmat@xmath28 , n. raicevic@xmath7 , p. reimer@xmath27 , a. rimmer@xmath28 , c. risler@xmath1 , e. rizvi@xmath38 , p. robmann@xmath12 , b. roland@xmath3 , r. roosen@xmath3 , a. rostovtsev@xmath19 , z. rurikova@xmath4 , s. rusakov@xmath2 , f. salvaire@xmath17 , d.p.c .
sankey@xmath24 , e. sauvan@xmath20 , s. schtzel@xmath1 , s. schmidt@xmath1 , s. schmitt@xmath1 , c. schmitz@xmath12 , l. schoeffel@xmath26 , a. schning@xmath11 , h .- c .
schultz - coulon@xmath43 , f. sefkow@xmath1 , r.n .
shaw - west@xmath36 , i. sheviakov@xmath2 , l.n .
shtarkov@xmath2 , t. sloan@xmath40 , p. smirnov@xmath2 , y. soloviev@xmath2 , d. south@xmath1 , v. spaskov@xmath45 , a. specka@xmath15 , m. steder@xmath1 , b. stella@xmath5 , j. stiewe@xmath43 , a. stoilov@xmath50 , u. straumann@xmath12 , d. sunar@xmath3 , v. tchoulakov@xmath45 , g. thompson@xmath38 , p.d .
thompson@xmath36 , t. toll@xmath1 , f. tomasz@xmath16 , d. traynor@xmath38 , p. trul@xmath12 , i. tsakov@xmath50 , g. tsipolitis@xmath52 , i. tsurin@xmath1 , j. turnau@xmath35 , e. tzamariudaki@xmath4 , k. urban@xmath43 , m. urban@xmath12 , a. usik@xmath2 , d. utkin@xmath19 , a. valkrov@xmath21 , c. valle@xmath20 , p. van mechelen@xmath3 , a. vargas trevino@xmath14 , y. vazdik@xmath2 , c. veelken@xmath28 , s. vinokurova@xmath1 , v. volchinski@xmath8 , k. wacker@xmath14 , g. weber@xmath17 , r. weber@xmath11 , d. wegener@xmath14 , c. werner@xmath13 , m. wessels@xmath1 , b. wessling@xmath1 , ch .
wissing@xmath14 , r. wolf@xmath13 , e. wnsch@xmath1 , s. xella@xmath12 , w. yan@xmath1 , v. yeganov@xmath8 , j. ek@xmath21 , j. zlek@xmath27 , z. zhang@xmath10 , a. zhelezov@xmath19 , a. zhokin@xmath19 , y.c . zhu@xmath1 , j. zimmermann@xmath4 , t. zimmermann@xmath11 , h. zohrabyan@xmath8 , and f. zomer@xmath10 |
the cone @xmath22 and its spherical angle @xmath38 have much symmetry . first observe that @xmath22 contains the line @xmath42 spanned by @xmath43 ; this is the line where all the inequalities @xmath44 , @xmath45 are equalities .
thus we can split @xmath22 as a sum of a line and lower dimensional cone @xmath46 where @xmath47 .
notice that @xmath48 , @xmath45 is orthogonal to @xmath43 , so @xmath49 provide a basis of @xmath50 .
it is convenient to define @xmath51 where @xmath52 is the unit sphere in @xmath53 .
the domain @xmath38 is a double cone over @xmath54 .
more precisely , let @xmath55 be one of the intersection points in @xmath56 ( there are two such points ) , and let @xmath57 be polar coordinates in @xmath58 , centered at @xmath55 .
then @xmath59 where we have identified @xmath52 with the unit sphere in the tangent space @xmath60 . in figure [ fig - domains ] we show @xmath61 and @xmath62 . and @xmath62.,height=192 ] in later sections
, we will use a generalization of this type of spherical cone . in general , let @xmath63 be a domain in the equatorial @xmath52 of @xmath58 , and let @xmath64 $ ] .
then we define the truncated cone @xmath65 we abbreviate @xmath66 . in this notation , @xmath67 .
the domain @xmath54 has symmetry .
if we let @xmath68 then we see @xmath69 and @xmath70 are pairwise congruent .
thus @xmath54 is a face of the regular @xmath71-hedral tesselation of the standard @xmath52 one obtains by connecting the vertices of a regular @xmath71-simplex with great circle arcs . in particular , one can compute the diameter of @xmath54 as the distance from a vertex to the center of the opposite face , which is @xmath72 moreover , the spherical angle of @xmath54 at a vertex is @xmath73 .
so we can construct a succession of comparison domains for @xmath74 starting with @xmath61 and using the coning construction described above .
to this end , we let @xmath75 = { \mathcal{t}}_1 , \qquad \hat { \mathcal{t}}_n : = { \mathcal{t}}{\mathcal{c}}(\hat { \mathcal{t}}_{n-1 } , \delta(n)).\ ] ] by induction , @xmath76 , and so @xmath77 .
we discuss two types of separation of variables in this section .
the first type is the separation of variables used in @xcite to estimate expected exit times of brownian motion from euclidean cones , and the second is the separation of variables one performs to estimate eigenvalues of a spherical domain with a conical structure .
one needs the domains in question to have a certain amount of regularity in order to separate variables .
in particular , these domains are piecewise @xmath78 , satisfy an exterior cone condition , and are well approximated ( in terms of hausdorff distance ) by finite volume , @xmath78 domains .
see the introduction to @xcite for a precise statement .
it is straight - forward to verify that all of the domains we consider satisfy deblassie s hypotheses .
we will refer to these domains as _
nice_. we first review deblassie s @xcite argument .
consider the cone @xmath79 over a domain @xmath25 : @xmath80 let @xmath26 be the time it takes for a brownian particle to exit @xmath27 , with starting position @xmath28 , and let @xmath81 be the probability that @xmath30 .
then @xmath82 satisfies the heat equation @xmath83 in polar coordinates @xmath84 , the pde becomes @xmath85 moreover , the solution scales as @xmath86 , so we can separate variables and look for a solution of the form @xmath87 , where @xmath88 .
then we get a positive separation constant @xmath89 , and @xmath90 letting @xmath91 , this ode becomes @xmath92 which has solutions of the form @xmath93 , where @xmath94 is the confluent hypergeometric function ( see _ e.g. _ , @xcite ) .
using @xmath94 and restricting to @xmath95 compact and @xmath96 , one obtains ( see @xcite ) an expansion for @xmath82 of the form @xmath97 where @xmath98 is the @xmath15th dirichlet eigenfunction of @xmath99 on @xmath34 and @xmath100^{1/2 } - \frac{n-1}{2}.\ ] ] this sum converges uniformly in @xmath101 .
the leading term in the expansion is @xmath102 , which yields the conclusion of inequality ( [ eigen - est1 ] ) .
next we separate variables to relate the eigenvalues @xmath103 and @xmath104 .
first recall that we can write the laplacian for @xmath58 in polar coordinates as @xmath105 where @xmath106 is the laplacian on the equatorial @xmath52 .
this lemma gives us a dimension reduction .
let @xmath63 be a nice domain in an equatorial @xmath52 with first eigenvalue @xmath107 , and let @xmath108 be the double cone over @xmath63 .
then the first dirichlet eigenvalues of @xmath109 and @xmath63 are related by @xmath110 in particular , @xmath111 whenever @xmath112 . * proof * : set @xmath113 , where @xmath114 and @xmath115 for @xmath116
. then @xmath82 is an eigenfunction on @xmath109 with eigenvalue @xmath117 precisely when @xmath118 separating variables with a positive separation constant @xmath119 yields @xmath120 choosing @xmath121 to be the first eigenfunction of @xmath63 , we obtain the ode @xmath122 which has regular singular points at @xmath123 .
if we try a solution of the form @xmath124 , for some power @xmath125 , we find @xmath126.\end{aligned}\ ] ] both coefficients must vanish , so we have @xmath127 and @xmath128 . solving for @xmath125 , we find @xmath129 next , observe that if @xmath130 then @xmath131 . finally , the formula for @xmath132 is monotone increasing in @xmath133 , and so @xmath111 whenever @xmath112 .
@xmath134 a second solution to equation ( [ sep - var2 ] ) has the form @xmath135 , where @xmath125 is again given by equation ( [ rel - eigenval ] ) but @xmath136 this eigenfunction vanishes on @xmath137 , so it corresponds to a higher eigenvalue . at this point
, we can prove that the expected capture time for @xmath138 predators is infinite . to prove the expected capture time is infinite , by inequality ( [ eigen - est1 ] )
we need to show @xmath139 , or , equivalently , that @xmath140 . in the case of @xmath141
, we have @xmath142 $ ] , and so @xmath143 . in the case
@xmath144 , we have @xmath145 $ ] , and so @xmath146 . we can not compute @xmath147 so easily , but we can find a test function to show that @xmath148 .
we shall find @xmath147 numerically in section [ numer - sec ] .
recall the rayleigh characterization of the first eigenvalue of a domain @xmath63 : @xmath149 to show that @xmath148 , it suffices to find @xmath150 so that @xmath151 .
let @xmath152 and observe that @xmath153 off the set of focal points of @xmath154 , which is a set of measure zero .
a computation shows @xmath155 one can generalize the eigenvalue relationship ( [ eigen - rel1 ] ) to spherical cones of the form @xmath156 , for @xmath157 , using the hypergeometric function @xmath158 later we will use the following lemma to relate the eigenvalues of @xmath159 and @xmath160 , which we will then estimate to bound the asymptotic decay rates of probabalistic exit times .
let @xmath63 be a nice domain in a equatorial @xmath161 with eigenvalue @xmath107 .
then the first dirichlet eigenvalues of @xmath63 and @xmath156 are related by @xmath162 where @xmath117 is the first dirichlet eigenvalue of the ode ( [ sep - var2 ] ) on @xmath163 $ ] . if @xmath164 then @xmath117 is the unique zero of @xmath165 in @xmath166 , where @xmath125 is defined in equation ( [ rel - eigenval ] ) and @xmath167 if @xmath168 then @xmath117 is the unique zero of @xmath169 in @xmath170 , where @xmath171 with @xmath94 defined by equation ( [ confluent ] ) .
* proof * : we separate variables and look for a solution of the form @xmath172 , with @xmath173 on @xmath174 , but @xmath175 .
then equation ( [ sep - var2 ] ) becomes @xmath176 \sin^m r\ , u.\ ] ] now let @xmath177 , where @xmath178 , which transforms the ode above into @xmath179 the solution to this ode is the hypergeometric function @xmath180 , with @xmath181 the lemma follows from taking @xmath182 , where we choose @xmath117 so that @xmath183 .
@xmath134 one can use this lemma to compute @xmath184 iteratively . in this case
we start with @xmath185 and apply equation ( [ eigen - rel2 ] ) . for the reader s reference
, we include a table of @xmath184 and the corresponding lower bound for @xmath186 ( see equation ( [ decay - exp ] ) ) . in this table
we use equation ( [ eigen - rel1 ] ) and the fact that @xmath187 so @xmath188 .
@xmath189 from this table , one can see that @xmath190 is slightly less than @xmath191 , so @xmath192 comes close to , but is not quite , an effective comparison domain for proving theorem [ main - thm ] .
in this section , we complete the proof of theorem [ main - thm ] .
first observe that it suffices to show @xmath193 .
in fact , it suffices to find an effective lower bound for @xmath147 using the following scheme .
we define @xmath194 by @xmath195 if we construct a domain @xmath196 such that @xmath197 and @xmath198 , then by domain monotonicity and equation ( [ eigen - rel2 ] ) , @xmath199 completing the proof of theorem [ main - thm ] . as the last step in our proof ,
we construct a domain @xmath200 as a perturbation of @xmath201 , such that @xmath202 and @xmath203 .
figure [ fig - arcs ] shows sketches of the domains @xmath201 , @xmath204 and @xmath205 , after stereographic projection to the plane .
, solid curve is @xmath204 , and the outer dashed curve is @xmath205 .
we generated this figure using the computer program maple.,width=393 ] previous work of rayleigh @xcite and plya
szeg @xcite motivates us to consider this type of domain perturbation .
they studied the eigenvalue of a planar domain which has the form @xmath206 in polar coordinates , for some small @xmath207 , and gave an expression for the eigenvalue @xmath208 in terms of @xmath209 . in our case
, we fix @xmath210 and find a domain @xmath204 with @xmath119 as its first eigenvalue .
suppose the functions @xmath211 and @xmath212 satisfy @xmath213 then @xmath214 is the first eigenfunction of @xmath215 on its nodal domain @xmath204 . by construction , @xmath216
, so @xmath82 is an eigenfunction .
also , @xmath82 does not change sign on its nodal domain , so it must be the first eigenfunction . in polar coordinates , @xmath217 .
let @xmath218 and set @xmath219 , so that equation ( [ sep - var2 ] ) becomes @xmath220 next we take @xmath221 , corresponding to the @xmath222 mode of the interval @xmath223 $ ] , and write the solution in terms of the hypergeometric function : @xmath224 finally , we take the @xmath225 superposition of the @xmath226 modes to define @xmath227 and let @xmath204 be the nodal domain of @xmath228 . by construction , @xmath229 . the domains
@xmath201 , @xmath204 and @xmath205 all have three boundary curves , and two of these boundary curves for each domain lie along the great circle arcs @xmath230 . for convenience , we convert to a planar domain using stereographic projection , with the south pole corresponding to @xmath231 in polar coordinates on @xmath232 .
the relationship between polar coordinates @xmath57 on @xmath232 and polar coordinates @xmath233 on the plane is @xmath234 . under stereographic projection
, the great circle arcs @xmath235 correspond to rays emanating from the origin at angles @xmath236 .
the third boundary curve of @xmath201 is given by the circular arc @xmath237 which we can rewrite as @xmath238 thus showing @xmath197 is equivalent to showing @xmath239 along the arc given by @xmath240 for @xmath241 .
one can see from figure [ fig - arcs ] that @xmath242 along this third boundary component of @xmath201 .
one strategy for a rigorous proof that @xmath242 is the following .
we first evaluate @xmath243 at a point @xmath244 on the curve , checking @xmath242 at @xmath245 , and bound the derivative @xmath246 on an interval containing @xmath244 .
our bound @xmath247 gives us a lower bound @xmath248 .
thus @xmath249 on a possibly smaller neighborhood of @xmath244 .
we then repeat this procedure with each endpoint of this ( smaller ) interval .
one can simplify the computations by observing @xmath201 and @xmath204 are symmetric about the ray @xmath250 .
then it suffices to evaluate @xmath243 and its derivative at @xmath251 .
@xmath134 . given our lower bound @xmath252
, one can use equation ( [ eigen - rel2 ] ) to show @xmath253 .
this in turn gives @xmath254 and @xmath255 .
in the same way , the lower bound @xmath256 gives @xmath257 .
in this section we describe a numerical computation approximating the eigenvalue @xmath147 , which relies on stenger s sinc - galerkin scheme @xcite . we will show @xmath258 , and so @xmath259 .
also , using equation ( [ eigen - rel2 ] ) , we have @xmath260 and so @xmath261 . given @xmath262 and a positive integer @xmath263 , one defines the @xmath263th cardinal sinc function of stepsize @xmath264 as @xmath265 and if @xmath266 , @xmath267 following stenger@xcite , for @xmath268 we shall let the basis functions be @xmath269 , where @xmath270 for @xmath271 with sinc points @xmath272 and @xmath273 for @xmath274 with sinc points @xmath275 .
we let @xmath276 and @xmath277 .
the dimension of the space of functions @xmath278 is @xmath279 .
also @xmath280 and @xmath281 , but both are zero at the other sinc points because their interior sinc approximations are subtracted . the functions @xmath282 form a basis of sinc functions when we are working on a rectangular strip .
the most important property of sinc functions is that they are well - suited to approximating integrals in a strip . in particular
, one can increase accuracy in a numerical computation by increasing the number of evaluation points , without recomputing grids ( as in finite elements ) . in order to take advantage of this simplicity
, we will conformally map a subdomain of @xmath201 to a half - infinite strip .
recall that @xmath201 is an equilateral triangle on @xmath232 with all its interior angles equal to @xmath283 .
let @xmath284 be the vertices of @xmath201 , let @xmath285 be the midpoint of the side opposite @xmath286 , and let @xmath287 be the center of mass of @xmath201 . observe that @xmath201 is invariant under reflection through the lines @xmath288 , which divide @xmath201 into six congruent subtriangles ( see figure [ fig - triangles ] ) .
thus the first eigenfunction is invariant under these reflections , and we can recover it by restricting to the smaller triangle @xmath63 , which has vertices @xmath289 .
the triangle @xmath63 has a right angle at the vertex @xmath290 , and an angle of @xmath291 at the vertices @xmath287 and @xmath292 .
the first dirichlet eigenfunction of @xmath201 , restricted to @xmath63 , will have dirichlet data on the edge @xmath293 and neumann data on the edges @xmath294 and @xmath295 .
we first transform @xmath63 to a planar domain ( which we again denote as @xmath63 ) using stereographic project , sending @xmath290 to @xmath296 .
this transformation sends @xmath287 to @xmath297 and @xmath292 to @xmath298 .
now @xmath63 is bounded by the two straight line segments joining @xmath296 to @xmath297 and @xmath296 to @xmath298 and the circular arc joining @xmath299 to @xmath298 which makes an angle of @xmath291 with the axes .
next we find a conformal transformation @xmath94 which takes @xmath63 to the half - infinite strip @xmath300 such that @xmath301 , @xmath302 , and @xmath303 . under this transformation ,
the first eigenfunction @xmath82 satisfies @xmath304 where @xmath305 is the pulled back laplacian .
there are formul for conformally mapping domains bounded by finitely many circular arcs @xcite generalizing the schwarz - christoffel formula .
following forsythe @xcite , the schwarz triangle mapping @xmath306 or from @xmath307 of the upper halfplane to @xmath308 is given by @xmath309 the group generated by reflections of @xmath63 along its edges in fact tiles the sphere .
the inverse function @xmath310 extends to a single sheeted cover and is invariant under the symmetry group , thus is a rational function of the plane .
thus we may compute @xmath94 .
writing @xmath311 , @xmath312 pulling back under @xmath313 , the conformal weight takes the form @xmath314 the branch cuts for the square and cube roots may be taken above the negative real axis .
thus @xmath315 lies in the fourth quadrant so the denominator in @xmath94 is nonvanishing .
the green s function may be found by the method of images .
thus , denoting @xmath318 , @xmath319 and @xmath320 , we get @xmath321 .
the green s function is @xmath322 pulling back by @xmath94 , the eigenvalue problem for the triangle may be restated as finding an eigenvalue for the integral operator @xmath323 the key point is to notice that the operator has logarithmic and algebraic singularities at the points @xmath324 and @xmath325 .
thus we need to handle these singularities .
the solution is zero along the imaginary axis , but free along the other two sides , which may be extended to functions to the plane which have odd reflection symmetry along the imaginary axis and even reflection symmetry along the other sides .
we shall approximate @xmath326 in an @xmath125-dimensional space @xmath327 with the same symmetries .
also , noticing that the eigenfunction on @xmath328 behaves like @xmath329 at the vertex @xmath287 , we actually have @xmath330 and @xmath82 decays algebraically at @xmath331 . choosing a basis @xmath332 of @xmath327 , we shall compute the matrix of the transformation @xmath333 , whose largest eigenvalue @xmath334 as @xmath335 and which is an upper bound @xmath336 @xcite .
the integral operator shall be computed numerically via sinc quadrature , which can handle such mild singularities .
we make the approximation @xmath337 where the sums are over @xmath338 and @xmath339 . here
@xmath340 , where @xmath341 and @xmath342 are defined in equations ( [ sinc - basis1 ] ) and ( [ sinc - basis2 ] ) , respectively .
we define @xmath343 .
thus the approximation @xmath344 is a collocation , as it equals @xmath82 at the sinc points .
thus , the matrix is approximated by @xmath345 where @xmath346 .
the approximating sum is carried over @xmath347 terms , corresponding to sinc quadratures in the four regions bounded by singularities ( _ e.g. _ in case @xmath348 and @xmath349 ) : @xmath350 , @xmath351 , @xmath352 and @xmath353 and where @xmath354 and @xmath355 are corresponding weights and @xmath356 and @xmath357 are corresponding sinc points .
the computation shows that @xmath358 which is above the critical value .
this is computational evidence that the exit time has finite expectation .
we present a table of a few the computed eigenvalues for approximations in spaces of given dimension , coming from the sinc - galerkin collocation scheme described . | we resolve a question of bramson and griffeath by showing that the expected capture time of four independent brownian predators pursuing one brownian prey on a line is finite .
our main tool is an eigenvalue estimate for a particular spherical domain , which we obtain by a coning construction and domain perturbation . in this paper , we examine the expected capture time of a single brownian prey pursued by @xmath0 independent brownian predators . all motion is restricted to a line .
bramson and griffeath @xcite first considered this problem , and estimated the capture time in various circumstances . in particular , they showed that if at time @xmath1 there are predators on both sides of the prey then the expected capture time is finite .
for this reason , we will assume that the initial position of the prey is @xmath2 and the initial positions of the predators are @xmath3 .
in this case , @xcite showed that the expected capture time is infinite for @xmath4 , and conjectured that it is finite for @xmath5 ( as indicated by simulations ) .
li and shao @xcite showed that the expected capture time is finite for @xmath6 . using a similar method , we resolve the remaining case by showing that the expected capture time is finite for @xmath7 .
[ main - thm ] let @xmath8 be pursued by @xmath9 , where @xmath10 are independent , standard brownian motions on a line , @xmath11 , and @xmath12 .
then the expected capture time is finite if and only if @xmath5 .
one key difference between li and shao s method in @xcite and our method here is that they consider a difference process , while we do a geometric splitting in spherical polar coordinates to reduce the dimension of the problem .
see section [ separate - sec ] for more details on our dimension reduction .
one can reformulate the capture problem described above in terms of exit of a brownian particle in @xmath13 from a specific cone .
we denote the position of the prey at time @xmath14 as @xmath8 and the position of the @xmath15th predator at time @xmath14 as @xmath16 . by our choice of initial conditions ,
the initial position of the brownian particle @xmath17 is @xmath18 .
the event of capture is then equivalent to the brownian particle @xmath19 first leaving the cone @xmath20 with @xmath21 , and so we must estimate the expected first exit time of a brownian particle from the cone @xmath22 , with the starting position @xmath23 .
deblassie @xcite developed the theory of estimating exit times for brownian motion from cones in euclidean space .
let @xmath24 be the cone over a domain @xmath25 .
also let @xmath26 be the exit time from @xmath27 of a brownian particle with starting position @xmath28 , and let @xmath29 be the probability that @xmath30 .
deblassie showed @xmath31 , where @xmath32^{1/2 } - \frac{n-1}{2}.\ ] ] here @xmath33 is the first dirichlet eigenvalue of @xmath34 . in the particular case
we are interested in , the expected exit time of a brownian particle from @xmath22 is finite if and only if @xmath35 , which reduces to @xmath36 where @xmath37 .
our method for proving theorem [ main - thm ] is to estimate the first eigenvalue of @xmath38 using the monotonicity property of eigenvalues , a coning construction , and domain perturbation . the rest of the paper proceeds as follows . in section [ geometry - sec ]
we discuss the geometry of @xmath22 , @xmath38 , and related regions .
section [ separate - sec ] contains the separation of variables background one needs to estimate the expected capture time . for the reader s convenience , we also include the proof that the expected capture time is infinite for @xmath4 predators in this section .
we prove the @xmath7 eigenvalue estimate in section [ eigen - est - sec ] .
finally , we describe a numerical computation of @xmath39 and a lower bound for @xmath40 in section [ numer - sec ] .
davar khoshnevisan told us about this problem .
we thank him and pedro mendez for lending their ears and expertise in this project .
t. thanks frank stenger for suggesting the algorithm of section [ numer - sec ] . |
flat cosmological models with a cosmological constant are currently serious candidates to describe the dynamics of the universe .
there are three main reasons for the present interest on these models .
first , they can reconcile inflation with dynamic estimates for the density parameter ( @xmath15 ) .
observations indicate @xmath16 for matter that clumps on scales @xmath17 mpc , while inflationary models usually predict @xmath18 .
second , these models , when normalized by data from the cosmic background explorer ( cobe ) predict less power in the perturbation spectrum at small scales than standard cold dark matter ( cdm ) , in accordance with observations @xcite . the third motivation for introducing a cosmological constant is the `` age crisis . '' in flat models with @xmath19 only if @xmath20 ( @xmath21 is the present value of the hubble parameter in units of @xmath22 km / s mpc@xmath23 ) is it possible to get theoretical ages for the universe that are higher than the lowest values ( @xmath24 gyear ) estimated for the galactic globular cluster system @xcite .
however , current estimates based on observations of cepheids stars in the virgo cluster and type ia supernovae indicate higher values for @xmath21 @xcite . in the past
the cosmological constant was introduced and , with the improvement of the observational data , later discarded .
now , however , the situation may change . by taking into account the vacuum contribution to the energy - momentum tensor
we can define an effective cosmological constant ( @xmath25 ) that is the sum of two terms , the bare cosmological term and @xmath26 ( @xmath27 is the vacuum energy density ) . from quantum field theory
we should expect @xmath28 ( @xmath29 is the planck mass ) , or perhaps another energy density related to some spontaneous symmetry - breaking scale such as @xmath30 or @xmath31 to the fourth power .
the problem lies in that these values are enormous when compared with astronomical bounds for @xmath32 .
extreme fine - tuning between @xmath33 and the bare @xmath0 is necessary to make theory compatible with observations .
one possible explanation for a small @xmath0 term is to assume that it is dynamically evolving and not constant , that is , as the universe evolves from an earlier hotter and denser epoch , the effective cosmological term also evolves and decreases to its present value @xcite .
there are also strong observational motivations for considering cosmological models with a decreasing @xmath0 term instead of a constant one .
usually in a dynamical-@xmath0 cosmological model , the distance to an object with redshift @xmath34 is smaller than the distance to the same object in a constant-@xmath0 model with the same value of the density parameter . as we shall discuss in sec .
iv , this implies that constraints coming from lensing statistics and from high redshift supernovae can be considerably weaker in these models @xcite .
recently @xcite we suggested a class of models in which @xmath0 decreases as @xmath1 [ here @xmath2 is the scale factor of the friedman - robertson - walker ( frw ) metric and @xmath3 is a constant ( @xmath35 ) ] .
although chen and wu @xcite gave some interesting arguments favoring the special value @xmath36 , it is clear that the above functional dependence with the scale factor is only phenomenological and does not come from particle physics first principles .
however , we believe it deserves further investigation for the following reasons .
first , these models generalize several other models present in the literature .
so , by investigating their properties we are studying at once the models they generalize .
second , since @xmath37 the universe age in these models is always larger than the age obtained in the standard einstein
de sitter cosmology and if @xmath38 the age is larger than the one we get in an open model with the same @xmath39 .
this is an important aspect if we are interested in solving the `` age problem . ''
further , since @xmath40 the @xmath41 term is generally not important during the radiation - dominated phase and nucleosynthesis proceed as in the standard model @xcite .
finally , the models are mathematically simple and in most cases can be treated analytically .
this paper is organized as follows . in sec .
ii the basic equations of the models are obtained and our main assumptions presented . in sec .
iii we investigate the effect of the @xmath0 term on the cosmic microwave background ( cmb ) anisotropy and compute the angular power spectrum for small values of @xmath42 for different values of the parameter @xmath3 and @xmath7 .
constraints on the models from high redshift sne ia and from lensing statistics are obtained in sec .
in this paper we consider spatially flat , homogeneous , and isotropic cosmologies with a time - dependent @xmath0 term : @xmath43 here the parameters @xmath44 and @xmath3 are restricted to the range @xmath4 , @xmath45 and the factor 3 was only introduced for mathematical convenience .
we consider the cosmic fluid to be a mixture of nonrelativistic matter and radiation ( @xmath46 ) with a perfect fluid energy momentum tensor , @xmath47 where @xmath48 is the total energy density ( radiation plus nonrelativistic matter ) and @xmath49 is the thermodynamic pressure . as in ref .
@xcite we assume that vacuum decays only into relativistic particles , such that the nonrelativistic matter energy momentum tensor is conserved ( @xmath50 ) .
the radiation energy density has two parts : one conserved , @xmath51 , @xmath52 , and a second one , @xmath53 , which arises due to the vacuum decay .
here @xmath54 is the present value of the scale factor and @xmath55 is the present value of the hubble parameter . in the following , subscripts
@xmath56 will always indicate present values .
the einstein equations for the models we are considering reduce to two equations : namely , @xmath57 and @xmath58 where @xmath59 is the matter density parameter and @xmath60 .
the above equations are quite general and apply for a broad spectrum of models . for instance , if @xmath61 , the usual flat frw model with a cosmological constant is obtained . if we take @xmath36 , eqs .
( 2.3 ) and ( 2.4 ) assume the same form of the einstein equations for open models and also appear in some string - dominated cosmologies @xcite .
further , we would obtain the _ same _ einstein equations if , instead of a cosmological term , we would have considered ( beside conserved matter and radiation ) an exotic x fluid with equation of state , @xmath62 @xcite .
all we discuss here also applies for these cosmologies and to emphasize this point in the following we shall use the expression `` x component '' to interchangeably designate the @xmath0 term or the x fluid .
the evolution of perturbations in the models we are considering may be studied considering two different phases @xcite . for @xmath63 , the energy density of the universe
is dominated by radiation and/or matter , and the contribution from the x component may be neglected . during the second phase , characterized by @xmath64 , the energy density of the universe
is dominated by nonrelativistic matter and/or the x component . in this phase ,
the x component contribution becomes more and more important and the deviations from the usual matter - dominated cosmological model increase as the universe expands . to describe the perturbation growth in these two phases , we consider the usual linear perturbation theory . each fourier component of the perturbation , @xmath65 , grows independently of the other modes and
may be related to the primordial power spectrum with the help of the transfer function : @xmath66 where @xmath67 is the primordial spectrum , usually taken to be @xmath68 in ( 3.1 ) @xmath69 is the transfer function , which incorporates all deformations undergone by the @xmath70-mode perturbation and @xmath71 may be any time @xmath72 .
if the initial perturbation is adiabatic and @xmath73 we have a primordial scale invariant adiabatic perturbation that is usually predicted by inflationary models .
in fact , for flat models without a cosmological constant it can be shown that , in this case , the quantity @xmath74 at @xmath75 , is independent of @xmath70 for any @xmath70 . here
, @xmath76 is the instant when the perturbation crosses the hubble radius .
however , this is not necessarily true if we have a cosmological term . in @xcite
we adopted a true scale invariant harrison - zeldovich spectrum for the models we are analyzing . here
we shall use the primordial scale invariant spectrum as defined above .
the power spectrum is defined in the standard way by @xmath77 where @xmath78 is the present time .
for flat models with @xmath79 , and after radiation - matter equality @xmath80 , all modes , inside and outside the hubble radius , evolve in the very same way .
so , @xmath65 , in fact , does not experience any deformation for @xmath81 and @xmath82 reflects the power spectrum for any @xmath81 . since in our model @xmath83 , the evolution during the first phase , dominated by radiation and matter ,
may be described by simply taking the standard power spectrum at @xmath84 , with the transfer function as computed by bond and efstathiou @xcite : @xmath85 where @xmath86 mpc , @xmath87 mpc , @xmath88 mpc , and @xmath89 . after radiation and x component equality , @xmath64 , the increasingly importance of the x component energy density
must be taken into account . to describe the perturbations after @xmath90
, we consider the evolution of matter perturbations in the background dominated by matter and the x component .
we shall use the approximation that the x component is smooth . in this case
it can be shown @xcite that all modes , inside and outside the hubble radius , grow at the same rate as in a flat universe dominated by nonrelativistic particles .
so again no extra deformation is introduced in the power spectrum by the presence of the x component and , to study the multipole expansion of the power spectrum , we are allowed to use the transfer function given by eq .
( 3.3 ) . since one of our goals
is to understand the effect of the decaying cosmological term ( x component ) on the cmb anisotropy , we consider the relation between the mass density perturbation and the fractional perturbation to the cmb temperature @xmath91 .
almost 30 years ago , sachs and wolfe @xcite obtained the expression relating fluctuations in the gravitational potential on the last scattering surface with cmb temperature anisotropies on large angular scales ( @xmath92 ) . for a flat universe
their formula can be written as @xcite @xmath93^{t_{ob}}_{t_{em } } - \left[\frac{a}{2}\frac{d}{dt}\left(a\frac{dd}{dt}\right)k \right ] ^{t_{ob}}_{t_{em } } \nonumber \\ & + & \frac12 \int^{t_{ob}}_{t_{em } } dt k \frac{d}{dt}a \frac{d}{dt}a\frac{dd}{dt } , \end{aligned}\ ] ] where @xmath94 describes the time dependence of the growing mode of @xmath95 , @xmath96 , @xmath97 is the unit vector pointing along a null ray from the observer to the source , and where @xmath98 is given by @xmath99 the first term in the right - hand side ( rhs ) of ( 3.4 ) describes the anisotropy caused by the relative motion source observer .
the effect of the observer s motion may be systematically removed from the experimental data , and will not be considered any further .
the effect of the source motion is a doppler contribution related to velocity perturbations on the last scattering surface and can not be eliminated .
the second term in eq .
( 3.4 ) includes the usual sachs - wolfe effect ( sw ) and a constant part evaluated at the present epoch which does not contribute to the observed anisotropy .
finally , there is the last term , usually called the integrated sachs - wolfe effect ( isw ) , which represents the redshift or blueshift of the photon energy caused by its traveling through regions of space with time - varying gravitational potential .
note that during the radiation - matter domination phase the growing modes behave as @xcite @xmath100 where @xmath101 is the scale factor at matter - radiation energy density equality .
so , since in flat models , during the radiation domination , we have @xmath102 , and during matter domination the scale factor grows as @xmath103 , it is easy to see that the contribution of the last term in eq . ( 3.4 ) vanishes in both cases .
this reflects the fact that , during these eras , the gravitational potential is time independent .
however , during the x component domination , the gravitational potential will not be constant and the isw term will no longer be zero .
in fact its contribution may be of the same order of the usual sw term for the lower modes in the harmonic expansion . in order to study the cmb anisotropies in the models described above , we keep three contributions in the rhs of ( 3.4 ) : sw , isw , and doppler .
the full temperature correlation function is defined by @xmath104 where @xmath105 ...
@xmath106 denotes an average over all positions @xmath107 and all directions @xmath108 , @xmath109 separated by an angle @xmath110 . to compute the average value in eq .
( 3.7 ) , we first analyze the isw contribution @xcite .
we have @xmath111 here @xmath112and @xmath113 where @xmath114 and @xmath115 .
note that if @xmath61 and @xmath116 we obtain @xmath117 .
equation ( 3.8 ) was obtained by making use of the field eqs .
( 2.3 ) and ( 2.4 ) , with @xmath118 , and the time evolution differential equation for @xmath119 , @xmath120 by using that , @xmath121 and @xmath122 , \end{aligned}\ ] ] where @xmath123 , we reduce ( 3.8 ) to @xmath124 , \end{aligned}\ ] ] where @xmath125 .
analogously we obtain @xmath126 , \end{aligned}\]]where @xmath127 , with @xmath128 the doppler term is given by @xmath129\right)_{x = y = i(a_{dec},a_0)}.\nonumber \\\end{aligned}\ ] ] the cross terms reduce to @xmath130\end{aligned}\ ] ] @xmath131\right)_{x = i(a_{dec},a_0 ) } \end{aligned}\ ] ] @xmath132\right)_{x = i(a_{dec},a_0)}\end{aligned}\ ] ] the total contribution to ( 3.7 ) may be written as @xmath133 it is convenient to expand the temperature correlation function in angular multipoles , using legendre polynomials @xmath134 : @xmath135 and @xmath136 . with the help of the relation @xmath137 = \sum_{n=0}^{\infty } ( 2n+1 ) p_{n}(\cos(\theta ) ) j_{n}[kr ] j_{n}[kq]$ ] , where @xmath138 $ ] is a spherical bessel function , the integrals in the contributions for @xmath139 decouple and the angular power spectrum becomes @xmath140 with @xmath141 where @xmath142,\ ] ] @xmath143,\end{aligned}\ ] ] @xmath144\right)_{x = i(a_{dec},a_0)}\right)^{2 } , \end{aligned}\ ] ] @xmath145 x_{l}(k ) , \end{aligned}\ ] ] @xmath146\frac{d}{dx}\left(j_{l}[k x ] \right)_{x = i(a_{dec},a_0)},\end{aligned}\ ] ] and @xmath147\right)_{x = i(a_{dec},a_0)}. \end{aligned}\ ] ] using the expressions above , we compute the angular power spectrum for different values of @xmath3 and @xmath7 .
calculations are done in a semianalytical way through the following steps . to make the integration in @xmath148 [ eq . ( 3.25 ) ] , we change variables . by defining @xmath149 , and by using an interpolated polynomial ,
we rewrite the integral in the new @xmath150 variable as @xmath151.\ ] ] here @xmath152 is an interpolated polynomial , which replaces the function @xmath153 , where @xmath154 stands for the inverse of @xmath149 .
the precision of this interpolation may be estimated by comparing @xmath152 and the original function @xmath155 . in the models we consider
, this comparison shows that using the seventh - order interpolated polynomial , @xmath156 , introduces an error always smaller than @xmath157 . with the help of this approximation ,
the resulting seven integrals in @xmath158 can be analytically computed .
the form of these expressions remains the same for all models that we want to consider .
changing among models only changes the numerical values of @xmath159 together with the upper limit in the integral , @xmath160 .
once we have calculated these auxiliary functions @xmath161 , we may compute the coefficients in eqs .
( 3.24 ) , ( 3.28 ) , and ( 3.30 ) by direct integration in @xmath70 . in the limit @xmath162 ,
the integrand is well behaved . for large values of @xmath70 ,
the integrand decreases very fast .
however , while decreasing , these integrands oscillate wildly , making the integral procedure slower and slower .
to avoid this technical difficulty , we truncate the integral at some value of @xmath70 , and we certify ourselves that the truncation is affecting our results in a controlled way . for values of @xmath70 larger than the truncation value , we compute the integrals replacing the oscillating @xmath163 by an envelope function , and the result of these integrals is taken as an upper bound of the error introduced by the truncation in @xmath70 .
this error grows steadily with @xmath42 , starting with @xmath164 for @xmath165 and growing up to the range @xmath166 for @xmath6 and different values of @xmath3 .
the numerical results obtained with the method described above are plotted in figs . @xmath167 and @xmath168 . in fig .
@xmath168 we plot for @xmath36 and @xmath169 the separate contributions from the sachs - wolfe , doppler and integrated sachs - wolfe effects . crossed terms are not shown .
basically , in the range @xmath170 , the sachs - wolfe contribution remains always important , with a slight decrease .
the integrated sachs - wolfe is large for small values of @xmath42 , but soon it becomes unimportant , reducing to just @xmath171 of @xmath172 at @xmath6 . on the other hand ,
the doppler contribution starts very small and grows fast to become the dominant contribution at @xmath6 .
for the analyzed models , the curves @xmath173
@xmath174 @xmath42 in fig .
@xmath167 , show a common feature , a minimum for small values of @xmath42 .
we see that , for fixed @xmath7 and @xmath21 , the position of this minimum changes with @xmath3 , becoming deeper as @xmath3 grows larger .
this feature reflects the fact that the contribution from the isw effect is larger for larger values of the parameter @xmath3 .
we also observe that for fixed @xmath3 and @xmath21 and different values of the matter content , the minimum grows deeper for smaller values of @xmath7 .
again this feature agrees with the fact that it is the x component term which brings up the nonzero isw effect , which , by its turn , makes the minimum deeper .
so , since @xmath175 , less nonrelativistic matter means more x component contribution , more isw effect , and consequently deeper minimum .
the supernovae cosmology project is an ongoing program to systematically search and study high - z supernovae . in a recent report @xcite perlmutter _
et al . _ analyzed seven sne ( with redshift @xmath176 ) , of more than 28 supernovae discovered , and obtained constraints on cosmological parameters , specially on the cosmological constant .
their preliminary result , @xmath177 at the @xmath178 confidence level , strongly constraints models for the universe with a cosmological constant . in this section
we use their observational results and adapt their procedure to constraint the class of models described in sec .
the basic idea is to use type ia supernovae as standard candles for the classic magnitude - redshift test . as in @xcite
we express the apparent bolometric magnitude @xmath179 as @xmath180 where in our case the luminosity distance ( in units of @xmath181 ) is given by , @xmath182 in eq .
( 4.1 ) @xmath183 is the `` zero point '' magnitude ( or hubble intercept magnitude ) , which is estimated from the apparent magnitude and redshift of low - redshift ( @xmath184 ) sne ia .
the nearby supernovae data set used by perlmutter _
et al . _ in the determination of @xmath185 were those 18 sne ia , discovered by the calan / tololo supernovae search @xcite for which the first observations were made no later than 5 days after maximum .
although sne ia are very similar explosion events , it is now known that they do not constitute a completely homogeneous class .
recently progress was made in the study of their inhomogeneities .
phillips @xcite showed that there is a correlation between absolute magnitudes ( m ) at maximum light and the initial decline parameter @xmath186 , the b - magnitude decline in the first 15 days after maximum . by studying the calan / tololo type ia supernovae , hamuy _
_ @xcite confirmed the existence of the correlation suggested by phillips and obtained a prescription for correcting observed b - magnitudes to make them comparable to an arbitrary `` standard '' sne ia light curve of width @xmath187 . after adding the correction to the sne ia low - z set they reduced the magnitude dispersion from @xmath188 to @xmath189 .
_ did the same for the seven high - z supernovae and reduced the dispersion from @xmath190 to @xmath191 mag . in our computations
we follow @xcite and use the corrected b - magnitude intercept at @xmath192 mag , @xmath193 , and consider only those 5 supernovae that have @xmath186 values in the range @xmath194 mag , which is the range of values investigated in the 18 low - z supernovae dataset . to construct the @xmath195 we used the data points outer error bars of perlmutter _
et al . _ , which are obtained by adding in quadrature the inner error bars of @xmath196 ( the apparent b - magnitude after width - luminosity correction ) to @xmath197 . in fig .
@xmath198 we plot the @xmath8 , @xmath9 , and @xmath10 confidence levels for the parameters @xmath3 and @xmath7 .
we see that for the interesting range @xmath199 , models with @xmath200 are in good agreement with the data .
the goodness - of - fit ( as defined in @xcite ) for the considered case is @xmath201 .
if we fix @xmath61 we recover the result of perlmutter _
et al . _ for constant @xmath0 , @xmath202 , with @xmath203 at the @xmath178 confidence level ( one tail ) .
another interesting case is @xmath36 . in this case
we again have to consider one degree of freedom in @xmath204 , and we find @xmath205 ( @xmath206 ) with goodness - of - fit equal to @xmath207 .
so , models with @xmath208 and @xmath209 are also in good agreement with the data .
it is interesting to compare the above constraints with those arising from lensing statistics . in fig .
@xmath210 we plot contours of constant likelihood ( @xmath8 , @xmath9 , and @xmath10 ) for the models , arising from lensing statistics .
we used data from the hst snapshot survey , the crampton survey , the yee survey , the eso / liege survey , the hst go observations , the cfa survey , and the not survey @xcite .
we considered 859 ( @xmath211 ) quasars plus 5 lenses and to obtain the contours we adapted from @xcite the procedure denoted there as `` a2 '' . by comparing fig . @xmath198 with fig .
@xmath210 we see that the sne ia constraints are stronger for low values of @xmath3 and @xmath7 ( @xmath212 and @xmath213 ) while the lensing ones are more important for larger values of the parameter @xmath3 ( @xmath214 ) . from fig .
@xmath210 we observe that models with @xmath12 and @xmath13 are in good agreement with the data .
it is a pleasure to thank scott dodelson , josh frieman , and albert stebbins for several useful discussions .
special thanks are due to saul perlmutter for kindly answering our questions and chris kochanek for sending us data on lensing surveys .
this work was supported in part by the brazilian agency cnpq and by the doe and nasa at fermilab through the grant nagw-2381 .
g.efstathiou , w. j. sutherland , and s. j. maddox , nature ( london)*348 * , 705 ( 1990 ) .
d. a. vandenberg , m. bolte , and p. b. stetson , annu .
. astrophys . * 34 * , 461 ( 1996 ) ; see also , b. chaboyer _ et al . _ , astro - ph 9706128 , with new results based on recent hipparcos parallax measurements which indicate younger ages ( @xmath215 @xmath216 smaller ) for the oldest globular clusters . if these results are confirmed the age problem will possibly be weakened .
however , it is still early to say that the age crisis has gone away ( see , e.g. , w. l. freedman , in astro - ph 9706072 for a recent discussion on this and related topics ) .
n. r. tanvir _ et al .
_ , nature ( london ) * 377 * , 27 ( 1995 ) ; w. freedman , astro - ph/9612204 ; a. g. riess , w. h. press , and r. kirshner , ap .
j. * 438 * , l17 ( 1995 ) ; m. hamuy _
. j. * 112 * , 2398 ( 1996 ) .
see , e.g. , m. ozer and m. o. taha , nucl .
phys . * b287 * , 776 ( 1987 ) ; k. freese _ et al .
_ , nucl . phys .
* b 287 * , 797 ( 1987 ) ; m. reuter and c. wetterich , phys . lett .
* b 188 * , 38 ( 1987 ) ; b. ratra and p. j. e. peebles , phys . rev . * d 37 * , 3407 ( 1988 ) ; i. waga , ap
. j. * 414 * , 436 ( 1993 ) ; j. a. frieman _
lett . * 75 * , 2077 ( 1995 ) ; k. coble , s. dodelson , and j. a. frieman , phys . rev . *
d 55 * , 1851 ( 1997 ) . v. silveira and i. waga , phys . rev . *
d 50 * , 4890 ( 1994 ) .
w. chen and y. s. wu , phys .
rev . * d 41 * , 695 ( 1990 ) . for the models considered in this paper @xmath217 is negligibly small during nucleosynthesis .
however , this is not necessarily true for all @xmath0-decaying models present in the literature .
for instance , for a class of models in which @xmath218 = constant , freese _ et al . _
@xcite showed that standard primordial nucleosynthesis constraints the parameter @xmath219 to be @xmath220 .
a. vilenkin , phys .
* 53 * , 1016 ( 1984 ) ; d. n. spergel and u. l. pen , astro - ph/9611198 ; l. m. a. bettencourt , p. laguna and , r. a. matzner , astro - ph/9612350 . j. n. fry , phys .
* b 158 * , 211 ( 1985 ) ; v. sahni , h. a. feldman and a. stebbins , ap . j. * 385 * , 1 ( 1992 ) ; h. a. feldman and a. e. evrard , int . j. mod .
* d 2 * , 113 ( 1993 ) ; j. stelmach and m. p. dabrowski , nucl .
* b 406 * , 471 ( 1993 ) ; h. martel , ap .
j. * 445 * , 537 ( 1995 ) ; p. j. steinhardt , nature ( london ) , * 382 * , 768 ( 1996 ) ; m. s. turner and m. white , astro - ph/9701138 .
j. r. bond and g. efstathiou , ap .
j. * 285 * , l45 ( 1987 ) ; g. efstathiou , proc . natl .
usa * 90 * , 4859 ( 1993 ) .
the evolution of modes bigger than the hubble radius is described by the general relativistic equations [ see , e. g. , t. padmanabhan , _ structure formation in the universe _
, ( cambridge : cambridge university press ) , p. 141 - 146
( 1993 ) ] . since nonrelativistic matter is conserved , and assuming that the x component is smooth the relativistic equation for @xmath221 may be written as an equation for @xmath222 , giving us the same newtonian equation used to describe the evolution of modes inside the horizon .
r. g. sachs and a. m. wolfe , ap .
j. * 147 * , 73 ( 1967 ) .
j. e. peebbles , _ principles of cosmology _
( princeton : princeton university press , 1993 ) .
a. h. jaffe , a. stebbins , and j. a. frieman , ap .
j. * 420 * , 9 ( 1994 ) .
s. perlmutter _
et al . _ ,
j. * 483 * , 565 , ( 1997 ) .
et al . _ , a. j. * 109 * , 1 ( 1995 ) ; m. hamuy _
_ , a. j. * 112 * , 2391 ( 1996 ) ; m. m. phillips , ap .
j. * 413 * , l105 ( 1993 ) .
l. f. bloomfield torres and i. waga , mon . not .
. soc . * 279 * , 712 ( 1996 ) .
d. maoz _ et al .
j. * 409 * , 28 ( 1993 ) ; d. crampton , r. d. mcclure , and j. m. fletcher , ap .
j. * 392 * , 23 ( 1992 ) ; h. k. c. yee , a. v. filipenko , and d. h. tang , a. j. * 105 * , 7 ( 1993 ) ; a. j. surdej _
* 105 * , 2064 ( 1993 ) ; e. e. falco , in _ gravitational lenses in the universe _ , edited by j. surdej , d. fraipont - caro , e. gosset , s. refsdal , and m. remy ( liege : univ .
liege ) , 127 ( 1994 ) ; c. s. kochanek , e. e. falco , and r. shild , ap .
j. * 452 * , 109 ( 1995 ) ; a. o. jaunsen _
_ , a & a * 300 * , 323 ( 1995 ) . | we investigate some properties of flat cosmological models with a @xmath0 term that decreases with time as @xmath1 ( @xmath2 is the scale factor and @xmath3 is a parameter @xmath4 ) .
the models are equivalent to standard cosmology with matter and radiation plus an exotic fluid with the equation of state @xmath5 .
we study the effect of the decaying @xmath0 term on the cosmic microwave background anisotropy and by using a seminumeric method we compute the angular power spectrum ( up to @xmath6 ) for different values of @xmath3 and @xmath7 .
we also investigate the constraints imposed on the models by the magnitude - redshift test in which high - redshift type ia supernovae ( sne ia ) are used as standard candles .
we obtain the @xmath8 , @xmath9 , and @xmath10 confidence levels on the parameters @xmath3 and @xmath7 and compare them with those arising from lensing statistics .
our analysis reveals that the sne ia constraints are stronger for low values of @xmath3 and @xmath7 , while those from lensing statistics are more important for @xmath11 .
models with @xmath12 and @xmath13 are in good agreement with the data .
psfig # 1@xmath14#1 |
summation to the least term ( optimal truncation of series ) can be traced back to cauchy s study of the gamma function ( see cauchy ( 1882))[cauchy ] .
the method was applied by stokes ( see stokes ( 1904))[stokes ] to solutions of differential equations and was instrumental in his theory of stokes phenomenon .
optimal truncation techniques have been greatly improved by dingle ( 1973)[dingle a ] , and in recent years as a result of the new ideas and methods of hyperasymptotics introduced by berry ( 1989 ) , who also discovered a surprisingly universal transition in hyperasymptotics , often called berry smoothing [ berry - smooth2 ] .
rigorous results on optimal truncation and smoothing under various assumptions were proved by mcleod ( 1992 ) ( [ mcleod]second order linear equations ) , paris ( 1992 ) [ paris1](linear systems with power coefficients ) , olver ( 1995 ) [ olver1 a](gamma function ) , olver and olde daalhuis ( 1995a , b ) [ olver - daalh1 ] ( linear second order systems ) [ olver - daalh2 ] , olde daalhuis ( 1995 , 1996 ) [ daalhuis ] ( linear second order systems ) , the authors ( 1996 ) ( [ ck1 a]first order nonlinear and second order linear ) and others .
a wide range of interesting applications were considered by berry and howls [ berry - howls ] [ berry - hyp ] ( 1990 , 1993 ) ( saddle point integrals ) , berry and keating ( 1992 ) ( asymptotics of the zeta function ) and berry ( 1991 , 1994 , 1995)[berry - zet ] [ berry - eva ] [ berry - gamma ] ( asymptotics of the gamma function , applications to dynamical systems , and pde s ) . other techniques of exponentially improved asymptotics have found interesting applications in dynamical systems ( see neishtadt 1984 , [ neishtadt ] ramis and schfke 1996 ) .
[ ramis ] in this paper we prove that asymptotic series of solutions of generic analytic ( linear or nonlinear ) differential systems near irregular singularities are summable to the least term , with errors of the order of magnitude of the least term with respect to the functions associated to the series by a specific generalized borel summation ( the balanced average ) .
in fact we show that the same is true for all series components of the transseries solutions of such differential systems .
we show that berry smoothing extends to this class of asymptotic series
. we deal with equations of the form @xmath5 for large @xmath6 in some direction , and study solutions that go to zero in this limit .
the assumptions on ( [ eqor1 ] ) are ( a1 ) the function @xmath7 is analytic at @xmath8 .
( a2 ) the eigenvalues @xmath9 of the linearization @xmath10 are different from zero and nonresonnant ( cf .
[ nonres ] ) .
it is worth mentioning that a large class of differential systems are not presented in the form described above but can be brought to it by suitable changes of variables .
for example , under the change of variables @xmath11 , the airy equation @xmath12 becomes @xmath13 also , the painlev i equation @xmath14 becomes @xmath15 by taking @xmath16 $ ] and @xmath17 . in this form , written as systems , both equations satisfy our assumptions ( see [ nonres ] and the notes therein ) .
the transformations themselves , when they exist , are readily obtained by comparing the complete formal solution ( transseries ) of the original equation with the transseries of the normalized equation , ( [ eqformgen , n ] ) below .
it is convenient to pull out the inhomogeneous and the linear terms ( relevant to leading order asymptotics ) and rewrite the system in the form @xmath18 under the assumptions ( a1 ) and ( a2 ) equation ( [ eqor ] ) admits , in normalized form , an @xmath19parameter family of formal exponential series solutions @xmath20 ( see [ sec : normal ] ) where @xmath21 are formal power series .
we study those formal solutions ( [ eqformgen , n ] ) which are at the same time _
asymptotic expansions _
( or proper transseries , in the sense of calle 1993[ecalle - book a][ecalle a ] ) . in our context , only _
finitely _ many powers can be present , of any exponential which is _ not small _ in the given direction .
note however that any exponential will decrease in _ some _ direction , and our analysis eventually covers an @xmath22 parameter family of formal solutions . for _ linear _ equations ,
any formal solutions ( [ eqformgen , n ] ) contain finitely many exponentials and are proper transseries .
for nonlinear equations we require , for large @xmath6 in a given direction @xmath23 in @xmath24 , that @xmath25 for any @xmath26 such that @xmath27 ( more details in [ sec : normal ] ) .
we start by illustrating the various concepts on a very simple example . the equation @xmath28 has the general solution @xmath29 ( @xmath30 ) .
the general _ formal _ solution for large @xmath6 is the elementary transseries @xmath31 with @xmath32 arbitrary . if @xmath33 , the exponential term is large and classical asymptotics gives full meaning to the family @xmath34 : for each @xmath34 there is a unique actual solution @xmath35 of the equation such that @xmath36 for large @xmath6 .
but as @xmath37 crosses an antistokes line @xmath38 into a sector where @xmath39 decays , @xmath40 becomes small _ beyond all orders _ of the divergent series and is undefined ( as part of @xmath34 ) in classical asymptotics .
classical asymptotics gives up the exponential and retains the information that all solutions are asymptotic to @xmath41 . in this framework
there is no natural way for asymptotically choosing any privileged solution as corresponding to @xmath41 ; @xmath41 becomes more of a property of the differential equation as a whole , shared by all its particular solutions . in terms of estimating solutions out of @xmath41 , classical asymptotics
provides _ polynomial _ precision : @xmath42}|\le \mbox{const.}|x|^{-n-2}$ ] for large @xmath6 , where @xmath43}:=\sum_{k=0}^{n}k!x^{-k-1}$ ] .
one of the most convenient and frequently used techniques going beyond poincar asymptotics is summation to the least term .
when @xmath6 is very large , the terms of @xmath41 start by decreasing rapidly ( @xmath44 if @xmath45 ) .
it is tempting to keep adding these terms as long as they continue to decrease .
the least term , for @xmath46 , evaluates to @xmath47 . even taking into account the slight ambiguity as to where to stop adding the terms , the numerical precision attained at this stage appears to be high enough to measure the small exponential beyond all orders .
it turns out that in any direction towards infinity there exists ( see e.g. costin and kruskal 1996[ck1 b ] ) an actual solution of the equation which lies within @xmath48 of the optimally truncated @xmath41 ( a nearest solution ) ; this solution , @xmath49 , is then necessarily unique for any fixed @xmath37 , since different solutions are separated by @xmath50 . as a manifestation of the stokes phenomenon , the association between @xmath41 and @xmath49 changes with @xmath51 . in our example @xmath52 if @xmath53 and @xmath54 when @xmath55 . since the solutions @xmath49 do not have singularities for large @xmath6
, nothing abrupt should happen to the terms beyond all orders , either .
michael berry discovered that there is indeed an intermediate range @xmath56 where the constant @xmath57 changes smoothly , in a surprisingly universal way .
we prove that berry smoothing applies to all nonlinear systems satisfying ( a1 ) and ( a2 ) , and the transition function is the same for all these systems .
technically , we show that berry smoothing is governed by the dominant behavior at the singularities in borel space nearest to the origin , and these singularities have the same essential features for all equations within our class . going beyond nonresonance , the picture changes . however
, properly interpreted , berry smoothing might persist in many resonant cases .
see appendix 2 for a discussion . as in the toy model above , in some cases ( linear homogeneous second order and first order equations notably , see costin and kruskal 1996[ck1 c ] ) there is _ only one _ solution of the original differential equation nearest to the divergent series . in general
there is a restricted family of such solutions .
olver s ( 1964 ) example illustrates very well that these special solutions might not be the obvious ones .
at least for this reason it is important to identify a method of recovering them from their transseries .
hyperasymptotics as introduced by berry leads to a substantial improvement in precision over optimal truncation but nevertheless ends up with nonzero errors .
recent very interesting developments of the original ideas of berry due to olde daalhuis ( 1996)[olver1 b ] , [ daalhuis b ] suggest that for some linear differential equations hyperasymptotics may eliminate all errors , but unfortunately at the price of giving up the convenience of optimal truncation by adding up vastly more terms beyond the least one in an extended family of expansions . at present
the only technique which at the same time gives full account of the complete formal solutions and is widely applicable is generalized borel summation ( see calle 1981 and 1993[ecalle - book b ] [ ecalle b ] , and also costin 1995 and 1997[costin][costin++a ] for rigorous results in the context of ( [ eqor ] ) ) . in this paper
we show that there is a natural connection between borel summation and optimal truncation , and that optimal truncation is compatible to a unique summation procedure , the balanced averaging introduced in costin ( 1997)[costin++b ] .
let @xmath58 and @xmath59 where @xmath60 ( those eigenvalues contained in the _ open _ half - plane @xmath61 centered along @xmath62 ) .
we require that for _ any _ @xmath63 : \(1 ) @xmath64 are @xmath65-linearly independent .
\(2 ) the complex numbers in the set @xmath66 ( note : the set is _ finite _ ) have _ distinct _ directions .
these are the stokes lines @xmath67 . that the set of @xmath68 which satisfy ( 1 ) and ( 2 ) has full measure follows from the fact that ( 1 ) and ( 2 ) follow from the condition : @xmath69 indeed , if ( [ strongnonr ] ) fails , one of @xmath70 is a rational function with rational coefficients of the other @xmath71 and @xmath72 , corresponding to a zero measure set in @xmath73 . * notes*. ( i ) since only small exponentials are allowed in a proper transseries , our conditions only restrict those ( sets of ) eigenvalues which are associated , through ( [ eqformgen , n ] ) , to exponentials that can be simultaneously _ small _ in some direction .
\(ii ) if condition ( 1 ) fails , then higher terms in the transseries ( which are present for nonlinear systems ) may _ resonate _ , that is @xmath74 for some @xmath75 ( @xmath76 ) .
this affects the equations characterizing @xmath77 .
however , we think that our results should not change in any substantial way unless in fact @xmath78 for some @xmath79 .
see appendix 2 for examples of this latter situation .
\(iii ) in the airy and p1 equations discussed in the introduction we have , after normalization , @xmath80 and the assumptions of nonresonnance are satisfied .
\(iv ) for linear systems , a weaker nonresonnance condition would suffice for our analysis , namely that @xmath81 for @xmath79 ( see costin 1995 ) .
this condition , as well as ( 2 ) above , insures that the stokes directions are distinct . by means of normal form calculations ( changes of variables ) , assuming ( a1 ) and ( a2 ) , it is possible to prepare ( [ eqor ] ) so that ( see wasow 1968 and tovbis 1992[wasow a][to1 ] ) ( n1 ) @xmath82 and ( n2 ) @xmath83 ( to this end
, it is _ not _ necessary that the original matrix @xmath84 in ( [ eqor ] ) is diagonalizable ) . for convenience , we rescale @xmath6 and reorder the components of @xmath85 so that : ( n3 ) @xmath86 for @xmath87 , @xmath88 , and @xmath89 ( cf .
( a2 ) ) if @xmath90 , where @xmath91 .
( to simplify notations , we will formulate some of our results relative to @xmath92 ; they can be easily adapted to any other @xmath9 , @xmath93 . ) to unify the treatment , by taking @xmath94 for some @xmath95 , we ensure that ( n4 ) @xmath96 .
( there is an asymmetry at this point : the opposite inequality can not be achieved , in general , as simply and without violating analyticity at infinity . ) finally , through a transformation of the form @xmath97 we arrange that ( n5 ) @xmath98 .
we choose @xmath99 . under the assumptions ( a1 ) and ( a2 ) and after the preparation ( n1 ) through ( n5 ) , the system ( [ eqor ] ) admits an @xmath22parameter family of formal exponential series solutions , ( [ eqformgen , n ] ) ( see [ iwano ] iwano 1957 [ wasow c]and wasow 1968 ; in costin 1997[costin++c ] , a brief derivation in the context of ( [ eqor ] ) is given ) . in ( [ eqformgen ,
n ] ) , @xmath100 is an arbitrary vector of constants , and @xmath101 .
the terms of an
_ asymptotic _ expansion are well - ordered with respect to @xmath102 .
thus , ( [ eqformgen , n ] ) is asymptotic in a given direction iff any _ ascending _ chain @xmath103 , @xmath104 , in ( [ eqformgen , n ] ) is _ finite _ ( agreeing to omit the terms with @xmath25 ) . for ( [ eqor ] )
the condition reads : ( n6 ) @xmath105 for all @xmath26 such that @xmath106 .
this selects the eigenvalues that lie in a half - plane centered at @xmath107 . without loss of generality ,
we let @xmath37 vary around @xmath108 . from now on , @xmath109 , @xmath110 , @xmath111 and @xmath112 where the index is selected by ( n6 )
. we will henceforth consider that ( [ eqor ] ) is presented in prepared form , and use the designation transseries only for those formal solutions satisfying ( n6 ) .
with respect to borel summation we are using the results and notation of costin ( 1995 and 1997)[costin b ] .
( a summary is presented in appendix 1 . )
all the analysis in this paper is based on the structure of singularities in borel space of the solutions of ( [ eqor1 ] ) and does not otherwise use ( [ eqor1 ] ) .
the results and proofs can be easily adapted to other functions than solutions of differential systems , having a similar singularity structure . to simplify the notation , as mentioned in the introduction ,
we formulate all results relative to @xmath88 ; they can be reformulated without any difficulty relative to any other @xmath9 of modulus one .
we can reduce the discussion to the case when @xmath113 ( cf .
theorem [ as ] ; see [ sec : asder ] for a discussion ) .
behavior for large @xmath114 of the coefficients @xmath115 in the asymptotic series @xmath116 .
let @xmath117 [ aco ] for @xmath118 with @xmath114 we have @xmath119 where @xmath120 for large @xmath121 .
the minimum of @xmath122 is reached for @xmath123 where the constant does not depend on @xmath124 .
we can write @xmath125 } $ ] with @xmath126}\sim \mathbf{e}_j\gamma_{r , j}$ ] and thus , defining @xmath127}=\sum_{r}\mathbf{a}_{r,0;[j ] } x^{-r}$ ] we have @xmath128}\ ] ] the terms of each series @xmath127}$ ] have the same argument ( phase ) iff @xmath6 belongs to the @xmath129-th stokes line , i.e. iff @xmath130 .
this is a precise formulation of _ dingle s rule of signs _
dingle 1973[dingle b ] , pp.7 ) . for resonant differential systems the rule needs to be reinterpreted .
see also the appendix for some further comments . difference between solutions and their optimally truncated asymptotic series .
[ mt ] \(i ) let @xmath85 be a solution of ( [ eqor ] ) and @xmath131 , with @xmath114 and @xmath132 .
then the difference between @xmath85 and the optimal truncation of @xmath116 is of the order of magnitude of the least term , @xmath133}(x)=o(\mathbf{a}_{0,r}r^{-r})\ \ \ ( \mbox{for fixed $ \xi$ , as } r\rightarrow\infty)\end{aligned}\ ] ] if and only if : ( 1 ) @xmath85 is the balanced borel sum ( see eq .
( [ defmedc ] ) and theorem [ ceq ] ) of the formal solution @xmath116 , i.e. , @xmath134 and ( 2 ) @xmath135 for all @xmath129 such that @xmath136 .
\(ii ) a similar estimate holds for the higher series in the transseries : for any @xmath137 and large @xmath121 we have @xmath138}=o(\mathbf{a}_{\bfk , r}r^{-r } ) \ \ \ ( \mbox{as } r\rightarrow\infty)\end{aligned}\ ] ] berry smoothing .
we now focus on the thin parabolic region ( the berry scale ) @xmath139 , latexmath:[$\arg(x)-\arg(\lambda_j)=\omega line for @xmath92 ) .
theorem [ bs ] shows that on the berry scale near a stokes line , the constant @xmath57 ( `` beyond all orders '' ) of a given solution changes , and the transition is given , as predicted by berry , by an error function .
[ bs ] let @xmath141 , with @xmath142 fixed and @xmath114 ; as @xmath143 we have @xmath144}=\frac{1}{2}s_1\,\mathrm{erf}\left(\frac{\omega}{\sqrt{2}}\right)\mathbf{y}_{\mathbf{e}_1}+o(e^{-x}x^{-\beta'})=\cr \frac{1}{2}s_1\,\mathrm{erf}\left(\frac{\omega}{\sqrt{2}}\right)e^{-x}x^{-\beta'}\mathbf{e}_1+o(e^{-x}x^{-\beta'})\end{gathered}\ ] ] we note here that there always exist directions along which the smoothing is visible ( namely the directions @xmath145 such that @xmath93 in our normalization ) . for other stokes directions , @xmath146 may be much larger than the contribution of erfc . for the directions of @xmath147 with @xmath148
the smallest term of the series is too large for the berry transition to be seen .
the change in the terms beyond all orders when crossing the stokes line is always measurable by borel summation ( cf .
theorem [ re ] ( ii ) and costin 1997[costin++f ] ) but might be too small for optimal truncation .
connection with borel summation . in the context of generic nonlinear systems
there is a one parameter family of _ distinct _ generalized borel summation formulas which are compatible with all usual operations ( addition , multiplication , differentiation , composition and their inverses , when meaningful ) and which preserve reality and asymptotic inequalities ( see costin 1997)[costin++g ] . in other words ,
the summation operators can not be intrinsically distinguished , in terms of their algebraic and/or classical asymptotics properties .
nevertheless , proposition [ bormed ] shows that only the generalized borel summation with @xmath149 , the balanced average , is compatible with least term truncation .
balanced average is thus the only borel summation method that guarantees optimal least term truncation properties ( incidentally , the balanced average is also the most symmetric and `` natural '' in its family ) .
we note that in our context it can be shown that the balanced average equals the ( apparently ) more complicated median average of ecalle .
[ bormed ] let @xmath150
. then @xmath151}(r\mathrm{e}^{\mathrm{i}\xi})=o(\mathbf{a}_{0,r}r^{-r } ) \end{aligned}\ ] ] for large @xmath121 and all @xmath152 if and only if @xmath149 . _ note _ : with @xmath116 fixed , @xmath85 is a different solution in different stokes sectors , see theorem [ re ] . to simplify the exposition ,
detailed proofs are provided for the principal series @xmath116 .
there are only minor adjustments necessary to cover all other components @xmath153 of the transseries ; we mention them at the end .
the notation is the same as in costin ( 1997)[costin++h ] and is explained in
[ sec : anset ] .
we let @xmath154 . taking @xmath155 where @xmath156 is a contour in @xmath157 such that @xmath158 ( or a linear combination of such contours as in the balanced average )
, we get by successive integrations by parts @xmath159 taking @xmath160 we get @xmath161 equation ( [ estimnth ] ) represents the error in summation near the least term ; we will show that for large @xmath121 @xmath162 \(1 ) if , exceptionally , all stokes constants are zero ( i.e. @xmath164 ( cf . ( [ defsj ] ) ) for @xmath165 ) , then @xmath166 is entire and @xmath167 for some @xmath168 ( theorems [ as ] and contour , @xmath169 for large @xmath121 .
this implies at once that @xmath170 is _ convergent _ in @xmath171 .
this case is thus trivial : summation to the least term is nothing else than convergent summation , and the sum is a true solution of the equation .
\(2 ) by theorem [ as ] for all @xmath26 such that @xmath172 , @xmath166 is analytic along @xmath173 .
we thus only count the @xmath9 with @xmath174 and rescale variables and change notations so that @xmath175 and @xmath176 .
( 0,0 ) # 1#2#3#4#5 @font ( 11839,9372)(720,-8725 ) fig
. 1 [ asymp1 ] for large @xmath114 we have @xmath177 _ proof_. let @xmath178 .
then @xmath179 is continuous in @xmath180 in @xmath181\backslash \{\phi_i , i=1 ... n\}$ ] and has at most jump discontinuities at @xmath182 ( theorem [ ceq ] ) . writing @xmath183 , pushing the contour of integration towards the circle of radius @xmath184 ( fig .
1 ) , and letting @xmath185 and @xmath186 we have @xmath187 using theorem [ re ] we have , with @xmath188 , @xmath189 where the @xmath129-th term has been integrated by parts @xmath190 times and we used theorem [ as ] to show that the boundary terms vanish at @xmath191 , and are @xmath192 at @xmath193 . by theorem [ as ] ( ii ) we have for small @xmath194 @xmath195 and @xmath196 . changing variables to @xmath197 in ( [ pushc ] ) we get , by laplace
s method , @xmath198 where @xmath199 are analytic and @xmath200 for small @xmath201 and thus the @xmath202 have an asymptotic series in @xmath203 , and @xmath204 . combining all these estimates
we get @xmath205 so that , with @xmath206 , @xmath207 because of the @xmath208 , the factors inside the sum can not cancel each other .
the rest of proposition [ asymp1 ] is immediate .
we distinguish two cases : if @xmath209 then , by convention , the direction of the laplace transform integral is @xmath210 . if @xmath211 then @xmath166 is analytic and exponentially bounded along @xmath212 ( theorems [ as ] and [ ceq ] ) .
we may assume @xmath213 ( cf .
[ sec : asder ] ) .
let @xmath214 be large , subject to the following conditions : ( @xmath215 ) @xmath216 .
( @xmath217 ) dist@xmath218 ( @xmath219 ) @xmath220 .
let @xmath221 be such that @xmath222 ( in particular , there are no singular points of @xmath166 on the circle of radius @xmath223 ) .
let @xmath224 and @xmath225 .
let @xmath226 . by theorem
[ as ] , @xmath227 is analytic in @xmath228 and by theorem [ ceq ] ( i ) , @xmath229 is uniformly continuous in the interior of @xmath230 ; let @xmath231 . by theorems [ as ] and [ ceq ] ( ii ) we can choose @xmath168 large enough so that for @xmath232 @xmath233 are chosen independently of @xmath121 and of @xmath6 . by @xmath234 and ( [ cddy ] ) for @xmath235
we have @xmath236 we have @xmath237 where @xmath238 depends only on @xmath168 and @xmath239 . by construction , for @xmath240 we have @xmath241 if @xmath242 .
we then have ( see fig 1 . )
, @xmath243 where @xmath244 @xmath245 contains the contribution of the outer circle , and in the last equality we integrated by parts @xmath246 times and included the boundary terms in @xmath247 .
we have @xmath248 for some @xmath249 depending only on @xmath250 and @xmath251 . in conclusion ,
@xmath252 for large enough @xmath253 where @xmath254 is independent of @xmath253 .
we have , with @xmath255 , @xmath256 : @xmath257 the behavior for large @xmath121 of the last integral is obtained by standard steepest descent : the function @xmath258 has a saddle point at @xmath259 and , for @xmath260 $ ] the steepest descent path originating at zero starts in a direction opposite to the pole @xmath261 and continues to @xmath262 . for @xmath263 , the contour in ( [ maininte ] )
can thus be deformed to a steepest descent curve without crossing the pole or the saddle .
the main contribution to the integral comes then from a region where @xmath264 is small .
we have the estimate for ( [ maininte ] ) : @xmath265^{-1}(1+o(r^{-1}))\end{gathered}\ ] ] as @xmath266 are continuous along @xmath267 , we obtain @xmath268 for some @xmath269 independent of @xmath121 and @xmath6 . since @xmath270
are continuous for small @xmath271 on the segment between @xmath147 and @xmath193 we have , using ( [ mainresur ] ) , ( [ intsup ] ) and integration by parts , with @xmath272 being the part of @xmath267 within @xmath271 distance of @xmath147 , @xmath273 with @xmath274 the last integral is very similar to ( [ pushc ] ) and is estimated in the same way , giving @xmath275 where , in view of ( [ estimnth ] ) we need to take @xmath276 instead of @xmath121 . noting that @xmath277 and comparing ( [ pushc2 ] ) to ( [ finalasympt ] ) , ( [ aim2 ] ) is proven .
this section deals with the special and important case of exponential asymptotics on and near the stokes line corresponding to one of the @xmath9 of largest module ( @xmath93 in our normalization ; without loss of generality we take @xmath278 ) .
berry s smoothing formula is proved , and we show that the balanced average is the only borel summation process compatible with optimal truncation .
the generalized borel summations that have good algebraic and analytic properties are given by ( [ defmedc ] ) .
using theorem [ as ] ( i ) which shows that for @xmath279 we have @xmath280 , formula ( [ mainresur ] ) , and the estimates in theorem [ ceq ] ( i ) for @xmath281 we see that @xmath282 choosing @xmath283 , relation ( [ estimnth ] ) reads , in view of theorem [ ceq ] , @xmath284 where @xmath285 means @xmath286 in order to prove theorem [ bs ] and proposition [ bormed ] , we need to estimate ( [ estimnthmed ] ) near the stokes line . replacing everywhere @xmath287 by @xmath288 and taking @xmath289 , the calculations leading to ( [ conc1 ] ) work without any other change and we get : @xmath290 for @xmath291 , @xmath292 does not vanish for @xmath293 , @xmath294 , and thus @xmath295 and the estimates leading to ( [ pushc2 ] ) are valid if we take @xmath289 .
we get , for @xmath291 , @xmath296 and @xmath297 we are therefore left with the problem of estimating @xmath298 we consider the case @xmath299 , the other case being treated symmetrically . [ estimei ]
let @xmath300 , @xmath301 , @xmath302 , @xmath299 , @xmath303 and @xmath304 .
\i ) for large @xmath114 @xmath305t\right\ } \drm t+o(r^{-1})\right)\cr-(1-\alpha)\frac{2\pi\irm}{(r-1)!}x^{r-1}\erm^{-xs}\end{gathered}\ ] ] with @xmath306 \ii ) when @xmath307 is small we have @xmath308 where @xmath309 , @xmath310 . _ proof . _
@xmath311 furthermore @xmath312 taking @xmath313 for small @xmath271 we have @xmath314 and @xmath315 part ( i ) of proposition [ estimei ] follows by combining ( [ defomega] ... [rotcont3 ] ) ) .
part ( ii ) is a straightforward calculation from ( i ) , using stirling s formula and laplace method .
we note that with @xmath316 held constant the integral in ( [ eqei ] ) is bounded for @xmath303 uniformly in @xmath114 .
thus , for @xmath317 we have @xmath318 for some @xmath319 and thus we obtain from ( [ conc1medp ] ) , proceeding as for ( [ subsegment ] ) , @xmath320 where @xmath321 , and @xmath322 is smooth on @xmath323 $ ] .
taking @xmath324 small and @xmath121 correspondingly large we have , in view of proposition [ estimei ] ( i ) , @xmath325 where @xmath326 .
using now proposition [ estimei ] ( ii ) to evaluate the inner integral we obtain @xmath327 where @xmath328 , where , as in ( [ pushc2 ] ) we need to take @xmath276 instead of @xmath121 ( cf .
( [ aim2 ] ) ) .
taking @xmath329 in the expression ( [ pushc2 ] ) , which was shown to be of the order of the least term , we see that the least term is @xmath330 _ smaller _ than ( [ cut3 ] ) unless @xmath331 \(a ) asymptotics along straight lines : if @xmath332 ( @xmath333 ) we see that only @xmath334 ( corresponding to the balanced average , which on the interval @xmath335 is the half sum of the upper and lower continuations ) ensures errors in optimal truncation of the order of the least term .
\(b ) asymptotics along parabolas and berry smoothing .
theorem [ bs ] is straightforward application of stirling s formula to ( [ cut3 ] ) .
( note also that by theorem [ ceq ] ( ii ) and theorem [ as ] ( ii ) @xmath336 . )
convergence of the puiseux series near the singularities of @xmath337 in borel space was the key in proving universality of the berry transition , since using this convergence , the calculation reduced in effect to the case of one particular function , @xmath338 . * the same line of proof as for @xmath116 works for @xmath153 as well .
indeed , we note that @xmath339 have the same singularity structure in borel space as @xmath116 .
this follows from theorem [ as ] ( ii ) and the fact that convolution with a locally analytic function preserves the convergence of puiseux series ( cf .
example ( 3a ) in appendix 1 ) .
the directions @xmath342 are the _ stokes lines _ ( note : sometimes known as _ anti-_stokes lines ! ) . we construct over @xmath343 a surface @xmath157 , consisting of homotopy classes of smooth curves in @xmath343 starting at the origin , moving away from it , and
crossing at most one stokes line , at most once ( fig .
2 ) : the laplace transform along a direction @xmath180 of a function @xmath345 @xmath346 will depend in general on @xmath180 ; the usual convention is to choose @xmath180 so that @xmath347 .
thus , the borel sum of @xmath2 in the direction @xmath6 , if it exists , is defined as @xmath348 with @xmath349 . by ( n6 ) and the agreed association between @xmath194 and @xmath6 , and since laplace integrals will not depend on the direction of @xmath194 until one of the stokes lines is crossed , we may as well assume that the direction of integration is either @xmath350 or is arbitrarily close to it .
define @xmath351 as the restriction of @xmath157 to @xmath352 where @xmath353 and @xmath354 .
we denote the analytic continuation of @xmath356 along a curve @xmath357 by @xmath358 . for the analytic continuations near a stokes line
@xmath359 we use notations similar to calle s : @xmath360 is the branch of @xmath356 along a path @xmath357 with @xmath361 , while @xmath362 denotes the branch along a path that crosses the stokes line between @xmath363 and @xmath364 .
we use the notations @xmath365 for @xmath366 and @xmath367 if integration is along a curve @xmath357 .
we write @xmath368 if @xmath369 for all @xmath26 and @xmath370 if @xmath368 and @xmath76 .
the relation @xmath371 is a well ordering on @xmath372 .
we let @xmath208 be the unit vector in the @xmath373 direction .
formal expansions are denoted with a tilde and capital letters @xmath374 will usually denote borel transforms or functions otherwise associated to borel space .
for notational convenience , we will not however distinguish between @xmath375 , which turn out to be convergent series , and the sums of these series @xmath376 as germs of ramified analytic functions .
the sum above coincides with the one in which @xmath415 is exchanged with @xmath416 , accounting for the reality - preserving property .
clearly , if @xmath376 is analytic along @xmath350 , then the terms in the infinite sum vanish and @xmath417 ; we also let @xmath417 if @xmath418 , where again @xmath419 is analytic .
it follows from ( [ defmed ] ) and theorem [ ceq ] below that the laplace integral of @xmath420 along @xmath421 can deformed into contours as those depicted in fig .
2 , with weight @xmath422 for a contour turning around @xmath423 . more generally , we consider the averages with @xmath426 , this represents the most general family of averages of borel summation formulas which commute with complex conjugation , with the algebraic and analytic operations and have good continuity properties ( see costin 1995[costin c ] ) .
the value @xmath427 is special in that it is the only one compatible with optimal truncation .
[ ceq ] ( i ) the branches of @xmath428 in @xmath351 have limits in a @xmath429-algebra of distributions , @xmath430 ( cf . [ sec : nc ] ) their laplace transforms in @xmath431 @xmath432 exist simultaneously and with @xmath433 and for any @xmath434 there is a constant @xmath435 and an @xmath436 large enough , so that for @xmath437 we have @xmath438 . if @xmath443 and @xmath444 then @xmath445 is continuous in @xmath446\backslash\{\phi_i : i\le n\}$ ] in the @xmath447 topology and has ( at most ) jump discontinuities for @xmath448 . for each @xmath137 , @xmath449 and any @xmath435
there is an @xmath450 and an @xmath451 such that @xmath452 are continuous in @xmath446\backslash\{\phi_i ;
-\bfk'\cdot\bflam+\lambda_i : i\le n , \bfk'\le\bfk\}$ ] in the @xmath453 topology and have ( at most ) jump discontinuities on the boundary .
\(ii ) the sum ( [ defmed ] ) converges in @xmath440 ( and coincides with the analytic continuation of @xmath376 when @xmath376 is analytic along @xmath421 ) . for any @xmath324
there is a large enough @xmath436 _ independent of @xmath137 _ so that @xmath454 with @xmath455 are laplace transformable for @xmath456 and furthermore @xmath457 .
in addition , if @xmath458 , then for large @xmath168 , @xmath459 . converges uniformly for @xmath464 , and @xmath85 is a solution of ( [ eqor ] ) .
when the direction of @xmath194 is not the real axis then , by definition , @xmath465 , @xmath466 is the usual laplace transform and ( [ soleqn ] ) becomes for some @xmath471 ( depending on the solution and chosen form ) . with the convention binding the directions of @xmath6 and @xmath194 and
the representation form being fixed , ( cf .
the beginning of [ sec : anset ] ) ) the representation of a solution is unique .
consider the expression of a fixed solution @xmath85 of ( [ eqor ] ) as a borel summed transseries ( [ soleqn ] ) .
as @xmath37 varies , ( [ soleqn ] ) changes only through @xmath471 , and that change occurs when the stokes lines are crossed .
we have , in the neighborhood of @xmath421 , with @xmath475 defined in ( [ defsj ] ) : [ r1 ] in view of ( [ mainresur ] ) the different analytic continuations of @xmath337 along paths crossing @xmath421 at most once can be expressed in terms of @xmath477 .
the most general formal solution of ( [ eqor ] ) that can be formed in terms of @xmath478 with @xmath479 is ( [ eqformgen , n ] ) with @xmath480 arbitrary and @xmath481 for @xmath482 . any true solution of ( [ eqor ] ) based on such a transseries
is given in ( [ soleqnp ] ) with @xmath471 as above .
any average @xmath483 along paths going forward in @xmath421 such that @xmath484 is thus of the form ( [ defmedc ] ) . [ stokestr ]
assume only @xmath92 lies in the right half plane .
let @xmath485 be two paths in the right half plane , near the positive/ negative imaginary axis such that @xmath486 as @xmath487 along @xmath485 .
consider the solution @xmath85 of ( [ eqor ] ) given in ( [ soleqn ] ) with @xmath488 and where the path of integration is @xmath489 .
then [ asymptrick ] i ) let @xmath491 and @xmath492 be solutions of ( [ eqor ] ) so that @xmath493 for large @xmath6 in an open sector @xmath494 ( or in some direction @xmath23 ) ; then @xmath495 for some constants @xmath267 , where the indices run over the eigenvalues @xmath496 with the property @xmath497 in @xmath494 ( or @xmath23 ) .
if @xmath498 for all @xmath129 , then @xmath499 .
\ii ) let @xmath491 and @xmath492 be solutions of ( [ eqor1 ] ) and assume that @xmath500 has differentiable asymptotics of the form @xmath501 with @xmath502 and @xmath503 , for large @xmath6 .
then @xmath504 for some @xmath26 .
let @xmath517 be a linear space and @xmath518 a family of norms satisfying ( [ focus - pocus ] ) .
for each @xmath168 we define a banach space @xmath519 as the completion of @xmath520 . enlarging @xmath517
if needed , we may assume that @xmath521 .
for @xmath522 , ( [ focus - pocus ] ) shows that the identity is an embedding of @xmath523 in @xmath524 .
let @xmath525 be the projective limit of the @xmath519 .
that is to say consider now the case when @xmath528 are commutative banach algebras .
then @xmath527 inherits a structure of a commutative algebra , in which @xmath529 ( `` convolution '' ) is continuous .
we say that @xmath530 is a * focusing algebra*. for more details see costin ( 1997)[costin++k ] .
let @xmath531 and @xmath532\subset ( -\pi/2,\pi/2),|p|\le k\}$ ] ( or a finite union of such sectors ) and @xmath533 be a small neighborhood of the origin .
@xmath534 will be the closure of @xmath533 , cut along the negative axis , and together with these upper and lower cuts . *
( 2 ) * if @xmath539 in example @xmath540 , then the norms ( [ norm00 ] ) are not equivalent anymore for different @xmath168 , but convolution is still continuous in ( [ norm00 ] ) and the projective limit of the @xmath541 , @xmath542 , is a focusing algebra . * ( 3a ) * @xmath543 . for @xmath544 and @xmath545 ,
this space is given by @xmath546 , where @xmath345 is analytic in the interior of @xmath547 and continuous in its closure .
we take the family of ( equivalent ) norms it is clear that convergence of @xmath356 in @xmath549 implies uniform convergence of @xmath345 on compact sets in @xmath547 ( for @xmath194 near zero , this follows from cauchy s formula ) .
@xmath550 are thus banach spaces and focusing spaces by ( [ normf1 ] ) .
the spaces @xmath551 are isomorphic to each - other .
the application @xmath552 a natural generalization of @xmath553 is obtained taking @xmath554 with positive real parts , no two of them differing by an integer . if @xmath555 with @xmath556 analytic , then @xmath557 iff @xmath558 for all @xmath26 ( e.g. , by a puiseux series argument ) .
it is then natural to identify the space @xmath559 of functions of the form @xmath560 with @xmath561 .
convolution with analytic functions is defined on @xmath559 while convolution of two functions in @xmath559 takes values in @xmath562 .
we write @xmath563 when the concrete values of @xmath564 do not matter . * ( 3b ) * a particular case of the preceding example is @xmath565 consisting of analytic functions in the interior of @xmath547 , continuous on its closure , and vanishing at the origin together with the first @xmath566 derivatives .
@xmath567 can be identified with @xmath568 . *
( 4 ) * @xmath440 , the `` staircase distributions '' .
let @xmath569 be the test functions on @xmath570 and @xmath571 .
let @xmath572 be the distributions @xmath356 for which @xmath573 on @xmath574 with @xmath575 $ ] .
there is a uniquely associated staircase decomposition , a sequence @xmath576 such that @xmath577 , @xmath578}$ ] and for @xmath591 , @xmath592 for @xmath593 is a closed ideal in @xmath440 ( isomorphic to the restriction @xmath594 of @xmath440 to @xmath595 ) .
the restrictions @xmath596 of @xmath440 to @xmath597 are for @xmath598 banach spaces with respect to the norm ( [ normdistr00 ] ) restricted to @xmath599 . if we choose a different interval length @xmath450 instead of @xmath601 in the partition associated to ( [ stdec0 ] ) , we then write @xmath602 .
obviously , dilation gives a natural isomorphism between these structures .
if @xmath603 is any ray , @xmath604 and @xmath605 are defined in an analogous way and have the same properties as their real counterpart .
once resonant equations are allowed , berry transitions tend to become more complicated : @xmath614 is a ( least term summable for @xmath615 ) solution of a linear differential equation with rational coefficients : and , in analogy with the nonresonant case we write @xmath630 where @xmath631 .
we first see that there is no curve @xmath632 for which the terms of @xmath633 ( or the terms of any linear combination of @xmath634 ) have the same phase .
there are instead two parabolic curves , @xmath635 along which for @xmath636 we have @xmath637/[a_{n}^+/x^{n } ] = 1+o(1)$ ]
. this does not amount to the terms being in phase , but rather is the discrete equivalent of a stationary point . to reinterpret dingle s rule for this example , we will see that there exist two stokes parabolas , each associated with one degree of freedom in the original equation along which the transitions in the constants beyond all orders , as measured by optimal truncation , are maximal .
@xmath642+\frac{n(n-1)a_{n-1}}{x^{n+1}}+\frac{(n-1)(n-2)a_{n-2}}{x^{n } } \cr+\frac{m^2 - 2n-2}{x^n}a_{n-1}=l_+c_+ + l_- c_-+\frac{n(n-1)a_{n-1}}{x^{n+1 } } -\frac{a_n}{x^n } \cr = l_+c_+ + l_- c_-+\frac{n!}{x^n}b_{n}\left(\frac{1}{n}-\frac{b_{n-1}}{b_n x}\right)\end{gathered}\ ] ] + + [ brbs 0]balser , w. ; braaksma , b.l.j . ; ramis , j - p . ; sibuya , y. 1991 multisummability of formal power series solutions of linear ordinary differential equations .
_ asymptotic anal . _ * 5 * , no . 1 2745 + + [ berry - smooth2 b]berry , m. v. 1989 stokes phenomenon ; smoothing a victorian discontinuity . _ proc .
london ser . a _ * 422 * , 721 . + + [ berry keating]berry , m. v. ; keating , j. p. 1992
a new asymptotic representation for @xmath651 and quantum spectral determinants .
london ser .
a * 437 * ( 1992 ) , no .
1899 , 151173 .
_ + + [ berry - eva b]berry , m. v. 1994 evanescent and real waves in quantum billiards and gaussian beams . _
* 27 * no .
11 , l391l398 .
+ + [ berry - zet b]berry , m. v. 1995 the riemann - siegel expansion for the zeta function : high orders and remainders . _
london ser . a _ * 450 * , no . 1939 , 439462 + + [ berry - hyp b]berry , m. v. ; howls , c. j. 1990 hyperasymptotics .
lond . a _ * 430 * , 653668
. + + [ berry - howls b]berry , m. v. ; howls , c. j. 1993 unfolding the high orders of asymptotic expansions with coalescing saddles : singularity theory , crossover and duality._proc .
london ser .
a _ * 443 * , no .
1917 , 107126 + + [ braaksma 0]braaksma , b. l. j. 1992 multisummability of formal power series solutions of nonlinear meromorphic differential equations .
fourier , grenoble _ * 42 * , 3 517540 .
+ + [ cauchy b]cauchy , a 188290 _ oeuvres completes daugustin cauchy , publies sous la direction scientifique de lacademie de sciences et sous les auspices du m. le ministre de linstruction publique . _
paris : gauthier - villars .
+ + [ cope b]cope , f. t. 1934 _ amer .
j. math_. * 56 * 411437 . + + [ costin d]costin , o. 1995 exponential asymptotics , transseries , and generalized borel summation for analytic , nonlinear , rank - one systems of ordinary differential equations internat . math .
notices no .
8 , 377417 .
+ + [ costin++ l]costin , o. ( 1997 ) on borel summation and stokes phenomena for rank one nonlinear systems of ode s . _ to appear in duke math .
j._costin ] + + [ ck1 d]costin , o. ; kruskal , m.d .
1996 optimal uniform estimates and rigorous asymptotics beyond all orders for a class of ordinary differential equations .
ser . a _ * 452 * 10571085 .
+ + [ dingle c]dingle , r.b . 1973 _ asymptotic expansions : their derivation and interpretation . _
new york : academic press . + + [ ecalle - book c]calle , j. 1981 _ fonctions resurgentes . _
publications mathematiques dorsay 81 . + + [ ecalle2]calle , j. 1990 finitude des cycles - limites et acclro - sommation de lapplication de retour . _ lecture notes in math .
_ * 1455 * , berlin : springer . + + [ ecalle c]calle , j. 1993 six lectures on transseries , analysable functions and the constructive proof of dulac s conjecture . _
bifurcations and periodic orbits of vector fields nato adv .
_ * 408 * , pp .
75184 , dordrecht : kluwer acad .
+ + [ ecalle3 0]calle , j. ; menous f. 1995 well - behaved convolution averages and the non - accumulation theorem for limit - cycles the stokes phenomenon and hilbert s 16th problem .
pp . 71101 , river edge , nj : world sci .
publishing .
+ + [ mcleod b]mcleod , j. b. 1992 smoothing of stokes discontinuities . proc .
london ser .
a * 437 * no .
1900 , 343354 .
+ + [ neishtadt b]neishtadt , a. i. 1984 the separation of motions in systems with rapidly rotating phase . _
mech . _ * 48 * no .
2 , 133139 .
+ + [ daalhuis c]olde daalhuis , a. b. 1996 hyperterminants . i. _ j. comput .
_ * 76 * no . 1 - 2 , 255264 .
+ + [ olver - daalh2 b]olde daalhuis , a. b. ; olver , f. w. j. 1994 exponentially improved asymptotic solutions of ordinary differential equations .
irregular singularities of rank one .
london ser .
a _ * 445 * no .
1923 , 3956 .
+ + [ olver - daalh1 b]olde daalhuis , a. b. ; olver , f. w. j. 1995a hyperasymptotic solutions of second - order linear differential equations . _
i. methods appl .
2 , 173197 . + + [ olver - daalh3 0]olde daalhuis , a. b. ; olver , f. w. j. 1995b on the calculation of stokes multipliers for linear differential equations of the second order .
_ i. methods appl . anal .
_ * 2 * no .
2 , 348367 .
+ + [ olver b]olver , f. w. j. 1964 in _ asymptotic solutions of differential equations and their applications . _ ed . c. h. wilcox , pp . 163183 .
new york : wiley .
+ + [ olver1 c]olver , f. w. j. 1995 on an asymptotic expansion of a ratio of gamma functions .
irish acad . a _ * 95 * no .
1 , 59 . + + [ paris1 b]paris , r. b. 1992 smoothing of the stokes phenomenon for high - order differential equations . _ proc .
london ser .
a _ * 436 * no .
1896 , 165186 .
+ + [ ramis b]ramis , j - p . ;
schfke , r. 1996 gevrey separation of fast and slow variables .
_ nonlinearity _ * 9 * , 353384 .
+ + [ stokes b]stokes , g. g. 1904 reprinted in _ mathematical and physical papers by late sir george gabriel stokes . _ vol .
77109 cambridge university press .
+ + [ to1 b]tovbis , a. 1992 normal forms of holomorphic matrix - valued functions and corresponding forms for singular differential operators . linear algebra and applications
* 162/164 * 389407 .
+ + [ wasow d]wasow , w. 1968 _ asymptotic expansions for ordinary differential equations_. new york : interscience publishers . | we prove that for divergent series solutions of nonlinear ( or linear ) differential systems near a generic irregular singularity , the common prescription of summation to the least term is , if properly interpreted , meaningful and correct , and we extend this method to transseries solutions . in every direction in the complex plane at the singularity ( stokes directions _ not _ excepted ) there exists a nonempty set of solutions whose difference from the `` optimally '' ( i.e. , near the least term ) truncated asymptotic series is of the same ( exponentially small ) order of magnitude as the least term of the series .
there is a family of generalized borel summation formulas @xmath0 which commute with the usual algebraic and analytic operations ( addition , multiplication , differentiation , etc ) .
we show that there is exactly one of them , @xmath1 , such that for any formal series solution @xmath2 , @xmath3 differs from the optimal truncation of @xmath2 by at most the order of the least term of @xmath2 .
we show in addition that the berry ( 1989 ) smoothing phenomenon is universal within this class of differential systems . whenever the terms `` beyond all orders '' _ change _ in crossing a stokes line , these terms vary smoothly on the berry scale @xmath4 and the transition is always given by the error function ; under the same conditions we show that dingle s rule of signs for stokes transitions holds . |
in the companion paper ( i.e. part i ) @xcite , we studied scalar linear network codes over commutative rings .
equivalently , these are linear codes over modules where a commutative ring acts on its own additive group via multiplication in the ring .
in particular , we compared the scalar linear solvability of networks over different types of commutative rings of the same size .
we proved that networks that are scalar linearly solvable over some commutative ring are also scalar linearly solvable over some field , although not necessarily of the same size .
additionally , we characterized all commutative rings with the property that there exists a network with a scalar linear solution over the ring but not over any other commutative ring of the same size .
linear network codes can be advantageous due to their ease of implementation and mathematical tractability .
these properties are due to the algebraic simplicity of linear maps and also to the structured nature of the alphabets used .
fields have the most algebraic constraints among alphabets used for linear network coding , e.g. associativity , distributivity , commutativity , invertibility .
more generally , rings may lack commutativity and/or invertibility , thus providing a broader class of alphabets over which to achieve linear network solvability .
we demonstrated in part i that relaxing only the invertibility constraint ( i.e. restricting to commutative rings ) can lead to linear network solvability that would not otherwise be possible with fields of the same alphabet size . in the present paper ( part ii )
, we additionally relax the commutativity constraint , and we study linear coding over general ring alphabets and , even more generally , over modules .
vector and scalar linear codes over rings and fields are special cases of linear codes over modules .
we focus on the relationship between alphabet commutativity and the scalar and vector linear solvability of networks , and we compare the linear solvability of networks over different modules where the alphabet size is the same .
an _ @xmath7-module _ ( specifically a left @xmath7-module ) is an abelian group @xmath8 together with a ring @xmath9 of _ scalars _ and an action @xmath10 such that for all @xmath11 and all @xmath12 the following hold : @xmath13 for brevity , we will sometimes refer to such an @xmath7-module as @xmath14 or simply @xmath15 .
the _ size of a module _ will refer to @xmath16 .
let @xmath17 denote the ring of all @xmath18 matrices with entries in @xmath7 and let @xmath19 denote the abelian group of all @xmath4-dimensional vectors with entries in @xmath15 with vector addition , where @xmath4 is a positive integer .
then @xmath19 is an @xmath17-module where multiplication of elements of @xmath7 with elements of @xmath15 is given by the action of @xmath14 . for basic network coding definitions ,
see part i ( * ? ? ?
* section 1.1 ) .
we will use the same models as in part i for networks , alphabets , etc .
, except we now study the generalized case of linear codes over modules , as opposed to linear codes over rings .
an edge function on the out - edge of a network node is _ linear with respect to the module @xmath14 _ if can be written in the form @xmath20 where @xmath21 are the inputs of the node and @xmath22 are constants .
that is , the messages and edge symbols are elements of the abelian group @xmath15 , and the linear edge and decoding functions are determined by coefficients of the ring @xmath7 . a decoding function is linear with respect to @xmath14 if it has a form analogous to , and a code is _ linear over a module @xmath14 _ if all edge and decoding functions are linear with respect to @xmath14 . the alphabet size in a linear code over a module is the size of the module , i.e. @xmath16 .
the special case of a module where the finite ring @xmath7 acts on its own abelian group @xmath23 by multiplication in @xmath7 is denoted by @xmath24 , and in this case , is equivalent to the definition of a scalar linear code over a ring that we used in part i. a network is _ linearly solvable over a module @xmath14 _ if there exists a linear solution over @xmath14 .
we will focus on two special types of linear codes : * a _ scalar linear code over a ring @xmath7 _ is a linear code over the module @xmath24 .
a network is _ scalar linearly solvable over @xmath7 _ if it has a linear solution over the module @xmath24 . *
@xmath4-dimensional vector linear code over a ring @xmath7 _ is a linear code over the module @xmath25 .
a network is _ vector linearly solvable over @xmath7 _ if it has a linear solution over the module @xmath25 , for some positive integer @xmath4 .
when referring to a linear code or solution over a ring , we will always specify ( in this paper ) scalar versus vector , or if neither is specified , then we are referring to a linear code over a module . additionally ,
when referring to an @xmath7-module @xmath15 , the ring @xmath7 is not assumed to be finite , unless otherwise specified .
however , when referring to a scalar or vector linear code over a ring @xmath7 , the ring @xmath7 is assumed to be finite .
@xmath4-dimensional vector routing code over an alphabet @xmath26 _ is a code in which messages and edge symbols are elements of @xmath27 and edge and decoding functions copy certain input vector components to the certain output vector components .
a vector routing code over @xmath26 is , in fact , a special case of a vector linear code over @xmath26 where each row of each of the matrices @xmath28 in is either all zero or else has @xmath29 one and @xmath30 zeros , and for each @xmath31 , at most one of the matrices @xmath28 has a non - zero @xmath32th row .
we can similarly define a right @xmath7-module and a linear code over a right @xmath7-module .
it can easily be shown that any linear code over a right module is equivalent to a particular linear code over a left module , so we restrict attention only to left modules . in section [ sec : basicmodules ] , lemmas are given which are used in proofs later in the paper .
section [ ssec : simple ] analyzes the linear solvability of networks over ring alphabets which are not necessarily commutative . in part
i , we proved that whenever a network is scalar linearly solvable over some commutative ring , then the smallest commutative ring over which the network is scalar linearly solvable is a field ( and thus the ring is unique ) ( * ? ? ?
* theorem 2.5 ) . here , we prove ( in theorem [ thm : r_dom_by_simple ] ) that if a network is scalar linearly solvable over some ( not necessarily commutative ) ring , then a smallest such ring is a matrix ring over a field .
it remains unknown , however , whether there can be more than one smallest ( not necessarily commutative ) ring over which a network is linearly solvable , since in general , there can exist multiple matrix rings over fields that are the same size .
we demonstrate ( in corollaries [ cor : choose - two - ring ] and [ cor : dim - n_smallest_ring ] ) that for two infinite classes of networks studied in this paper , the smallest size ring over which each network is linearly solvable is indeed unique .
we prove ( in theorem [ thm : min_module ] ) that if a network is linearly solvable over some module , then a smallest such module ( i.e. with a smallest associated abelian group ) corresponds to a vector linear solution over some finite field .-
dimensional vector linear code over a field @xmath33 , the alphabet size of the module is @xmath34 . ]
we prove ( in theorem [ thm : min_module_not_unique ] ) , in contrast to the commutative ring case , that the minimum size module with respect to linear solvability is not necessarily unique .
thus , for a fixed network , vector linear codes over fields are `` best '' in a certain sense , as these codes can minimize the alphabet size needed for a linear solution .
we also show ( in corollary [ cor : general_vector ] ) that for all networks , the following properties are equivalent : ( i ) vector linear solvability over some field ,
( ii ) scalar linear solvability over some ring , and ( iii ) linear solvability over some module .
similarly , we show ( in corollary [ cor : general_scalar ] ) that for all networks , the following properties are equivalent : ( a ) scalar linear solvability over some field , ( b ) scalar linear solvability over some commutative ring , and ( c ) linear solvability over some module whose ring is commutative . in section [ sec : gen_m ] , we present a family of networks that generalize the m network of @xcite , and we enumerate ( in theorem [ thm : dim - n_vector ] ) the particular vector dimensions over which each of these networks has vector linear solutions .
we prove ( in corollary [ cor : non - comm - ring ] ) that these networks have scalar linear solutions over certain non - commutative matrix rings yet do not have scalar linear solutions over any commutative ring .
we also show ( in theorem [ thm : non - comm_16 ] ) that if a network is scalar linearly solvable over a non - commutative ring @xmath7 and is not scalar linearly solvable over any commutative ring , then @xmath35 .
this lower bound is shown to be achievable ( in corollary [ cor : non - comm - ring ] and example [ ex:16-noncommutative ] ) by exhibiting a network which has a scalar linear solution over a non - commutative ring of size @xmath0 but not over any commutative ring .
section [ sec : mods ] focuses on linear solvability of networks over different modules with the same alphabet size , specifically , @xmath4-dimensional vector linear codes over @xmath5 and rings of size @xmath3 .
we prove ( in theorem [ thm : module - beats - rings ] ) that for each prime power @xmath3 , there exists a network with a linear solution over a module of size @xmath3 but with no scalar linear solutions over any ring of size @xmath3 .
these particular networks have @xmath4-dimensional vector linear solutions over @xmath5 .
we show ( in theorem [ thm : commutative_implies_vector ] ) that any network with a scalar linear solution over a commutative ring of size @xmath3 has a @xmath4-dimensional vector linear solution over @xmath5 .
we prove a similar result ( in theorem [ thm : p6_implies_vector ] ) for general rings of size @xmath3 when @xmath36 .
additionally , we show ( in theorems [ thm : commutative_implies_vector ] and [ thm : p6_implies_vector ] ) that these results generalize in a natural way to rings of non - power - of - prime sizes . finally , section [ sec : complexity ] provides some concluding remarks .
an @xmath7-module @xmath15 is _ faithful _ if for all @xmath37 , there exists @xmath38 such that @xmath39 .
in other words , @xmath40 for all @xmath41 if and only if @xmath42 . for any finite ring @xmath7 and positive integer @xmath4 ,
the @xmath17-module @xmath43 is faithful , so vector and scalar linear codes over rings are special cases of linear codes over faithful modules . for a fixed ring @xmath7 ,
there are generally multiple modules over @xmath7 . for example
, if @xmath7 is a subring of @xmath44 , then @xmath45 is an @xmath7-module where the action is multiplication in @xmath44 , and @xmath23 is also an @xmath7-module where the action is multiplication in @xmath7 .
the following lemma shows that the linear solvability of a network over a faithful @xmath7-module is determined entirely by the ring of scalars @xmath7 and not by the module s underlying abelian group .
however , we note that not every ring and group pair can form a module . for example
, the additive group of @xmath46 can not be a @xmath47-module , since @xmath48 in @xmath46 and @xmath49 in @xmath47 .
let @xmath7 be a fixed ring . if a network is linearly solvable over some faithful @xmath7-module , then it is linearly solvable over every @xmath7-module .
[ lem : same_ring ] let @xmath50 be a network that is linearly solvable over the faithful @xmath7-module @xmath8 , and let @xmath51 denote the messages of @xmath50 .
suppose a node in @xmath50 has inputs @xmath52 in a solution over @xmath14 , where , for each @xmath53 , @xmath54 for some @xmath55 .
then for each out - edge of this node , there exist constants @xmath56 such that the edge carries the symbol @xmath57 then , by induction , every edge and decoding function in a linear code over a module is a linear combination of the network messages .
@xmath15 is a faithful @xmath7-module , so @xmath29 and @xmath58 are the only elements of @xmath7 such that @xmath59 and @xmath60 for all @xmath38
. hence it must be the case that decoding functions in the linear solution over @xmath14 are of the form @xmath61 if @xmath62 is some other @xmath7-module , then a linear solution for @xmath50 over @xmath14 is also a linear solution for @xmath50 over @xmath63 , since every edge will carry the same linear combination of the messages ( i.e. the same elements of @xmath7 are the coefficients in the linear combination ) , so , in particular , the decoding functions will be the same linear combination of the messages .
in contrast to lemma [ lem : same_ring ] , if @xmath15 is both an @xmath7-module and an @xmath44-module , then there may exist a network that is linearly solvable over @xmath64 but not @xmath14 .
for example , @xmath46 is a subfield of @xmath65 , so @xmath66 is both a faithful @xmath46-module and a faithful @xmath65-module .
we demonstrate ( in corollary [ cor : choose - two - ring ] ) a network that is scalar linearly solvable over @xmath65 but not @xmath46 , and by lemma [ lem : same_ring ] , this network is linearly solvable over the @xmath65-module @xmath66 but not the @xmath46-module @xmath66 .
the following corollary is a special case of lemma [ lem : same_ring ] and will be frequently used in later proofs .
it demonstrates an equivalence between scalar linear solutions over matrix rings and vector linear solutions over rings .
let @xmath7 be a finite ring , @xmath4 a positive integer , and @xmath50 a network
. then @xmath50 is scalar linearly solvable over the ring of @xmath18 matrices whose elements are from @xmath7 if and only if @xmath50 has a @xmath4-dimensional vector linear solution over @xmath7 .
[ cor : same_ring ] the `` if '' and the `` only if '' directions are each obtained by separately applying lemma [ lem : same_ring ] , since @xmath17 and @xmath43 are faithful @xmath17-modules with matrix - matrix multiplication and matrix - vector multiplication , respectively . note that in a @xmath4-dimensional vector linear code over a ring @xmath7 , the alphabet size is @xmath67 , whereas in a scalar linear solution over @xmath17 , the alphabet size is @xmath68 .
so any network that is scalar linearly solvable over the matrix ring @xmath17 is also linearly solvable over a smaller module alphabet .
we will generalize this idea in theorem [ thm : min_module ] .
as is common in mathematics literature , it will be assumed throughout this paper that ring homomorphisms preserve both additive and multiplicative identities .
if @xmath69 is a ring homomorphism and network @xmath50 is linearly solvable over some faithful @xmath7-module , then @xmath50 is linearly solvable over every @xmath44-module .
[ lem : modhomomorphism ] let @xmath62 be an @xmath44-module and define a mapping @xmath70 by @xmath71 , where @xmath72 is the action of @xmath73 .
one can verify that @xmath62 is an @xmath7-module under @xmath74 .
now , let @xmath15 be a faithful @xmath7-module , and suppose @xmath50 has a linear solution over @xmath14 . by lemma [ lem : same_ring ] , @xmath50 is linearly solvable over @xmath63 , so every edge function in the solution over @xmath63 is of the form @xmath75 where @xmath76 are the parent node s inputs and @xmath77 are constants . form a linear code for @xmath50 over @xmath73 by replacing each coefficient @xmath78 in by @xmath79 .
let @xmath80 be the edge symbol in the code over @xmath73 corresponding to @xmath81 in the code over @xmath63 .
then @xmath82 thus , whenever an edge function in the solution over @xmath63 outputs the symbol @xmath81 , the corresponding edge function in the code over @xmath73 will output the same symbol @xmath81 . likewise , whenever @xmath83 is an input to an edge function in the solution over @xmath63 , the corresponding input of the corresponding edge function in the code over @xmath73 will be the same symbol @xmath83 .
the same argument holds for the decoding functions in the code over @xmath73 , so each receiver will correctly obtain its corresponding demands in the code over @xmath73 . hence , the code over @xmath73 is a linear solution for @xmath50 .
let @xmath7 and @xmath44 be finite rings .
if there exists a ring homomorphism from @xmath7 to @xmath44 , then every network that is scalar linearly solvable over @xmath7 is also scalar linearly solvable over @xmath44 .
[ cor : modhomomorphism ] @xmath23 is a faithful @xmath7-module for any finite ring @xmath7 , so this is a special case of lemma [ lem : modhomomorphism ] where the modules are @xmath24 and @xmath84 . for finite rings @xmath7 and @xmath44 ,
special cases of corollary [ cor : modhomomorphism ] include : * @xmath44 is a subring of @xmath7 : + the identity mapping is an injective homomorphism from @xmath44 to @xmath7 , so any network that is scalar linearly solvable over @xmath44 is also scalar linearly solvable over @xmath7 . * @xmath7 has a two - sided ideal @xmath85 : + there is a surjective homomorphism from @xmath7 to @xmath86 ( see lemma [ lem : surjective_homomorphism ] ) , so any network that is scalar linearly solvable over @xmath7 is also scalar linearly solvable over @xmath86 . *
@xmath87 is the projection mapping : + @xmath88 is a surjective homomorphism , so any network that is scalar linearly solvable over @xmath89 is also scalar linearly solvable over @xmath7 ( and likewise over @xmath44 ) .
cases ( 1 ) , ( 2 ) , and ( 3 ) agree with lemma 2.8 , corollary 2.4 , and lemma 2.6 , respectively , from part i. in fact , corollary [ cor : modhomomorphism ] is a generalization of these results .
we will focus on linear codes over modules whose ring acts on its own abelian group , i.e. scalar linear codes over rings . as noted after corollary [ cor : modhomomorphism ]
, for any two - sided ideal @xmath85 of a finite ring @xmath7 , every network that is scalar linearly solvable over @xmath7 is also scalar linearly solvable over @xmath86 , so in determining the smallest ring over which a network is scalar linearly solvable , it is natural to focus attention on rings without two - sided ideals .
a ring is _ simple _ if it has no proper two - sided ideals .
that is , its only two - sided ideals are the ring itself and the trivial ideal @xmath90 .
the following lemmas give results related to simple rings and network linear solvability .
a finite ring is simple if and only if it is isomorphic to a matrix ring over a field .
[ lem : simple_rings ] this is a corollary of the artin - wedderburn theorem ( e.g. ( * ? ? ?
* , theorem 3.10 ( 4 ) ) and ( * ? ? ?
* , theorem ii.9 ) ) .
* theorem 7 , p. 243 ) if @xmath85 is a two - sided ideal of ring @xmath7 , then the mapping @xmath91 given by @xmath92 is a surjective homomorphism .
[ lem : surjective_homomorphism ] for each finite ring @xmath7 , there exists a simple ring @xmath44 such that the following hold : * there exists a surjective homomorphism from @xmath7 to @xmath44 , * every network that is scalar linearly solvable over @xmath7 is scalar linearly solvable over @xmath44 , and * @xmath93 divides @xmath94 .
[ lem : hom_to_simple ] if @xmath7 is a simple ring , then each statement is trivially true by taking @xmath95 , so we may assume @xmath7 is not a simple ring .
thus , @xmath7 has a proper maximal two - sided ideal @xmath85 .
let @xmath96 , and note that since @xmath85 is maximal , @xmath44 is simple .
the mapping @xmath91 given by @xmath92 is a surjective homomorphism by lemma [ lem : surjective_homomorphism ] , which proves ( a ) .
hence by corollary [ cor : modhomomorphism ] , any network that is scalar linearly solvable over @xmath7 is also scalar linearly solvable over @xmath44 , which proves ( b ) .
since @xmath7 is finite , we know that @xmath97 divides @xmath94 , which proves ( c ) . if @xmath7 is a finite commutative ring and @xmath44 is a simple ring satisfying ( a)-(c ) in lemma [ lem : hom_to_simple ] , then @xmath44 must also be commutative , since there is a surjective homomorphism from @xmath7 to @xmath44 .
however , as we demonstrate in the following example , if @xmath7 is non - commutative , then such an @xmath44 is not necessarily non - commutative .
the following demonstrates : ( i ) a class of non - commutative rings for which the simple ring in lemma [ lem : hom_to_simple ] is non - commutative , and ( ii ) a class of non - commutative rings for which the simple ring in lemma [ lem : hom_to_simple ] is commutative * let @xmath98 denote the ring of integers mod @xmath99 . for any positive integers @xmath100 , and prime divisor @xmath1 of @xmath99 , there exists a surjective homomorphism from the non - commutative ring @xmath101 to the non - commutative simple ring @xmath102 , given by matrix - component - wise reduction mod @xmath1 . * for each field @xmath33 and integer @xmath103 , there exists a surjective homomorphism from the non - commutative ring of upper triangular @xmath18 matrices with entries in @xmath33 to the commutative simple ring @xmath33 ( see the proof of lemma [ lem : p3_matrices ] ) .
the following theorem demonstrates that any smallest ring over which a network is scalar linearly solvable is simple .
if a network is scalar linearly solvable over a ring @xmath7 but not over any smaller ring , then @xmath7 is a matrix ring over a field .
[ thm : r_dom_by_simple ] suppose a network @xmath50 is scalar linearly solvable over a ring @xmath7 that is not simple .
by lemma [ lem : hom_to_simple ] ( a ) ( b ) , there exists a simple ring @xmath44 and a surjective homomorphism @xmath69 , such that @xmath50 is scalar linearly solvable over @xmath44 . since @xmath88 is surjective , @xmath104 , but since @xmath44 is simple and @xmath7 is not , the two rings can not be isomorphic , so @xmath105 , and therefore @xmath106 .
this proves that every smallest size ring over which @xmath50 is scalar linearly solvable must be simple , which implies that such a ring is a matrix ring over a field by lemma [ lem : simple_rings ] . in part
i ( * ? ? ?
* theorem 2.5 ) , we showed that the smallest - size commutative ring over which a network is scalar linearly solvable is unique .
however , there may exist multiple simple rings of the same size ( e.g. @xmath107 and @xmath108 are non - isomorphic simple rings of size @xmath109 ) .
an interesting open question is whether every network with a scalar linear solution over multiple simple rings of the same size also must have a scalar linear solution over some smaller simple ring .
i.e. is the smallest ring @xmath7 in theorem [ thm : r_dom_by_simple ] unique for a given network ? we demonstrate ( in corollaries [ cor : choose - two - ring ] and [ cor : dim - n_smallest_ring ] ) that for two infinite classes of networks ( one of which is a class of multicast networks ) studied in this paper , the smallest - size ring over which each network is scalar linearly solvable is unique .
the following lemma shows that linear solutions over unfaithful modules admit linear solutions over faithful modules .
let @xmath15 be an @xmath7-module .
there exists a ring @xmath44 such that @xmath15 is a faithful @xmath44-module , and any network that is linearly solvable over @xmath14 is linearly solvable over @xmath64 .
if @xmath7 is commutative , then there exists a commutative such @xmath44 .
[ lem : faithful_module ] we use ideas from @xcite here .
let @xmath110 , which is easily verified to be a two - sided ideal of @xmath7 .
let @xmath111 .
it can also be verified that @xmath15 is an @xmath44-module with action @xmath74 given by @xmath112 . if @xmath113 are such that @xmath114 for all @xmath38 , then @xmath115 , which implies @xmath116 .
hence @xmath117 , so the ring @xmath44 acts faithfully on @xmath15 .
if @xmath7 is commutative , then the ring @xmath118 is also commutative .
suppose a network @xmath50 is linearly solvable over @xmath14 .
every edge function in the solution is of the form @xmath119 where the @xmath120 s are the parent node s inputs and the @xmath78 s are constants from @xmath7 .
form a linear code over @xmath64 replacing each coefficient @xmath78 in by @xmath121 .
let @xmath80 be the edge symbol in the code over @xmath64 corresponding to @xmath81 in the code over @xmath14 .
then @xmath122 thus , whenever an edge function in the solution over @xmath14 outputs the symbol @xmath81 , the corresponding edge function in the code over @xmath64 will output the same symbol @xmath81 .
likewise , whenever @xmath83 is an input to an edge function in the solution over @xmath14 , the corresponding input of the corresponding edge function in the code over @xmath64 will be the same symbol @xmath83 .
the same argument holds for the decoding functions in the code over @xmath64 , so each receiver will correctly obtain its corresponding demands in the code over @xmath64 .
hence , the code over @xmath64 is a linear solution for @xmath50 . in a linear network code over a module @xmath14 , in principle , the ring @xmath7 need not be finite ( although representing linear code coefficients might be problematic ) . for example , any abelian group @xmath8 is a @xmath123-module with action given by @xmath124 however , in a linear network code over a module , the alphabet is finite , so the abelian group @xmath15 must be finite .
the following corollary shows that if a network is linearly solvable over a module where the ring is infinite , then it is also linearly solvable over a faithful module where the ring is finite .
let @xmath7 be an infinite ring and let @xmath15 be a finite @xmath7-module . then there exists a finite ring @xmath44 such that @xmath15 is a faithful @xmath44-module and any network that is linearly solvable over @xmath14 is linearly solvable over @xmath64 .
if @xmath7 is commutative , then there exists a commutative such @xmath44 .
[ cor : infinite_ring ] this follows from lemma [ lem : faithful_module ] , and the fact that the ring of a faithful finite module must also be finite .
a _ submodule _ of an @xmath7-module @xmath15 is a subgroup @xmath62 of @xmath15 such that @xmath62 is closed when acted on by @xmath7 .
that is , both @xmath62 and @xmath15 are @xmath7-modules and @xmath125 .
submodules are of particular interest , since by lemma [ lem : same_ring ] , if @xmath15 and @xmath62 are faithful @xmath7-modules , then the set of networks that are linearly solvable over @xmath14 and the set of networks that are linearly solvable over @xmath63 are equal , yet a linear code over @xmath63 has a smaller alphabet if @xmath62 is a proper submodule of @xmath15 . as an example
, let @xmath85 be a two - sided ideal in the ring @xmath7 . then @xmath126 is a subgroup of @xmath23 that is closed under multiplication in @xmath7 , so @xmath127 is a submodule of the @xmath7-module @xmath7 .
as another example , for each finite field @xmath33 and integer @xmath103 , the @xmath128-module @xmath129 is a proper submodule of the @xmath128-module @xmath128 .
lemmas [ lem : submodules ] and [ lem : smaller_module ] show results related to submodules that will be used to prove theorem [ thm : min_module ] .
* theorem 3.3 ( 2 ) , p. 31 ) let @xmath33 be a finite field and @xmath4 a positive integer .
then @xmath129 is the only @xmath128-module that has no proper submodules .
[ lem : submodules ] by lemma [ lem : same_ring ] , for each ring @xmath7 , if a network is linearly solvable over a faithful @xmath7-module , then it is linearly solvable over every @xmath7-module . when a network is solvable over the @xmath7-modules for a particular ring @xmath7
, it may be desirable for linear network coding to determine the minimum - size @xmath7-modules .
lemma [ lem : smaller_module ] considers this question for rings of matrices over a finite field .
let @xmath33 be a finite field and @xmath4 a positive integer .
if @xmath15 is a finite non - zero @xmath128-module , then @xmath34 divides @xmath16 .
[ lem : smaller_module ] since @xmath15 is finite and non - zero , @xmath15 contains a submodule with no proper submodules . by lemma [ lem : submodules ]
, @xmath129 is the only @xmath128-module with no proper submodules , so @xmath129 is a submodule of @xmath15 .
hence by lagrange s theorem of finite groups ( e.g. ( * ? ? ?
* , theorem 8) ) , @xmath34 divides @xmath16 .
the following theorem is a generalization of theorem [ thm : r_dom_by_simple ] , where we characterize smallest - size modules over which networks are linearly solvable .
theorem [ thm : min_module ] demonstrates that if a network is linearly solvable over some module , then there exists a vector linear code over a field that minimizes the alphabet size needed for a linear solution .
suppose a network @xmath50 is linearly solvable over an @xmath7-module @xmath15 .
then the following hold : * there exists a finite field @xmath33 and positive integer @xmath4 such that @xmath50 has a @xmath4-dimensional vector linear solution over @xmath33 and @xmath34 divides @xmath16 . *
if @xmath7 is commutative , then there exists a finite field @xmath33 such that @xmath50 has a scalar linear solution over @xmath33 and @xmath130 divides @xmath16 .
[ thm : min_module ] if the ring @xmath7 is infinite , then by corollary [ cor : infinite_ring ] , @xmath50 is linearly solvable over some faithful module with a finite ring .
if @xmath7 is commutative , then by corollary [ cor : infinite_ring ] , @xmath50 is linearly solvable over some faithful module with a finite commutative ring .
so without loss of generality , assume @xmath7 is finite and @xmath15 is a faithful @xmath7-module . by lemmas [ lem : simple_rings ] and
[ lem : hom_to_simple ] ( a ) , since @xmath7 is finite , there exists a field @xmath33 , a positive integer @xmath4 , and a surjective homomorphism @xmath131 . by lemma [ lem : modhomomorphism ] any network that is linearly solvable over the faithful @xmath7-module @xmath15
is also linearly solvable over every @xmath128-module , so in particular , @xmath50 has a @xmath4-dimensional vector linear solution over @xmath33 .
since @xmath88 is a homomorphism , any @xmath7-module is also an @xmath128-module ( see the proof of lemma [ lem : modhomomorphism ] ) .
thus , both @xmath15 and @xmath129 are @xmath128-modules , so by lemma [ lem : smaller_module ] , we have @xmath34 divides @xmath16 .
if @xmath7 is commutative , then , since @xmath88 is a surjective homomorphism , @xmath128 must also be commutative , which implies @xmath132 .
hence @xmath50 has a scalar linear solution over @xmath33 and @xmath130 divides @xmath16 .
theorem [ thm : min_module ] demonstrates that , in some sense , vector linear codes over finite fields are optimal for linear network coding , as they can minimize the alphabet size needed for a linear solution .
the following lemmas will be used to show ( in theorem [ thm : min_module_not_unique ] ) that a minimum - size module over which a network is linearly solvable is not necessarily unique .
lemma [ lem : kn - vector ] is a result of sun et .
al @xcite .
* proposition 1 , p. 4513 ) let @xmath133 be a prime power and @xmath4 a positive integer .
if a network has a scalar linear solution over @xmath134 , then it has a @xmath4-dimensional vector linear solution over @xmath135 .
[ lem : kn - vector ] for each integer @xmath136 , the _ @xmath99-choose - two network _ is a multicast network that was described by rasala lehman and lehman @xcite and further studied in our part i ( see figure 1 in @xcite ) .
@xcite let @xmath26 be a network alphabet and let integer @xmath137 . * if the @xmath99-choose - two network has a solution over @xmath26 , then @xmath138 .
* let @xmath26 be a field .
the @xmath99-choose - two network is linearly solvable over @xmath26 if and only if @xmath139 .
[ lem : kchoosetwo_fields ] for each integer @xmath103 and prime @xmath1 , the @xmath140-choose - two network is linearly solvable over at least two distinct modules of size @xmath3 but not over over any smaller modules .
[ thm : min_module_not_unique ] by lemma [ lem : kchoosetwo_fields ] , the @xmath140-choose - two network is scalar linearly solvable over @xmath141 and is not solvable over any alphabet whose size is less than @xmath3 . by lemma [ lem : kn - vector ] ,
any network with a scalar linear solution over @xmath141 has a @xmath4-dimensional vector linear solution over @xmath5 .
hence the @xmath140-choose - two network has a scalar linear solution over @xmath141 and a @xmath4-dimensional vector linear solution over @xmath5 , yet the network has no linear solution over any module whose size is less than @xmath3 . the following corollary generalizes theorem 2.11 from part i , which showed the @xmath140-choose - two network is not scalar linearly solvable over any commutative ring of size @xmath3 other than the field @xmath141 . in fact , as a result of corollary [ cor : choose - two - ring ] , the @xmath140-choose - two network is not scalar linearly solvable over any ring of size @xmath3 other than the field . for each integer @xmath103 and prime @xmath1 ,
the unique smallest - size ring over which the @xmath140-choose - two network is scalar linearly solvable is @xmath141 .
[ cor : choose - two - ring ] by lemma [ lem : kchoosetwo_fields ] , the @xmath140-choose - two network is scalar linearly solvable over @xmath141 and is not solvable over any smaller alphabet .
suppose the @xmath140-choose - two network is scalar linearly solvable over a ring @xmath7 of size @xmath3 . by lemmas [ lem : simple_rings ] and
[ lem : hom_to_simple ] ( a ) ( b ) , there exists a field @xmath33 , a positive integer @xmath99 , and a surjective homomorphism @xmath142 such that the @xmath140-choose - two network is scalar linearly solvable over the ring @xmath143 . since @xmath88 is surjective , @xmath144 . by corollary [ cor : same_ring ] , the @xmath145-choose - two network has an @xmath99-dimensional vector linear solution over @xmath33 , so by lemma [ lem : kchoosetwo_fields ] ( a ) , @xmath146 .
hence @xmath147 , which implies @xmath148 and @xmath149 .
since @xmath150 is a surjective homomorphism and @xmath151 , we have @xmath152 .
the following corollaries summarize our results on the linear solvability of networks using scalar and linear vector codes over fields , scalar linear codes over rings , and linear codes over modules .
corollary [ cor : general_vector ] shows an equivalence between vector linear solvability over fields and linear solvability over rings and modules , while corollary [ cor : general_scalar ] shows an equivalence between scalar linear solvability over fields and linear solvability over commutative rings and modules . for any network @xmath50 , the following three statements are equivalent : * @xmath50 is vector linearly solvable over some finite field .
* @xmath50 is scalar linearly solvable over some ring .
* @xmath50 is linearly solvable over some module .
[ cor : general_vector ] if a network has a @xmath4-dimensional vector linear solution over some field @xmath33 , then by corollary [ cor : same_ring ] it has a scalar linear solution over the ring @xmath128 , hence ( i ) implies ( ii ) .
a scalar linear code over a ring is a special case of a linear code over a module , so ( ii ) implies ( iii ) . by theorem [ thm :
min_module ] ( a ) , ( iii ) implies ( i ) . for any network @xmath50 ,
the following three statements are equivalent : * @xmath50 is scalar linearly solvable over some finite field .
* @xmath50 is scalar linearly solvable over some commutative ring .
* @xmath50 is linearly solvable over some module whose ring is commutative .
[ cor : general_scalar ] a scalar linear code over a finite field is a special case of a scalar linear code over a commutative ring , hence ( i ) implies ( ii ) .
a scalar linear code over a commutative ring is a special case of a linear code over a module where the ring is commutative , so ( ii ) implies ( iii ) . by theorem [ thm : min_module ]
( b ) , ( iii ) implies ( i ) .
we summarize our results on minimizing the alphabet size in linear network coding by : * if a network is scalar linearly solvable over some commutative ring , then the ( unique ) smallest such commutative ring is a field ( * ? ? ?
* theorem 2.5 ) .
* if a network is scalar linearly solvable over some ring , then a smallest such ring is a matrix ring over field ( theorem [ thm : r_dom_by_simple ] ) .
it is not known whether such a smallest ring is unique . *
if a network is linearly solvable over some module , then a smallest such module yields a vector linear solution over a field ( theorem [ thm : min_module ] )
. such a module may not be unique ( theorem [ thm : min_module_not_unique ] ) .
for each integer @xmath153 , the _ dim-@xmath99 network _ is defined in figure [ fig : dim - n ] and is referred to as such because it has vector linear solutions precisely over vector dimensions that are multiples of @xmath99 .
we prove this fact in theorem [ thm : dim - n_vector ] .
this infinite family of networks will be used to demonstrate several theorems related to commutative and non - commutative rings .
the special case of @xmath154 corresponds to the _ m network _ of @xcite , shown later in figure [ fig : m - network ] .
the dim-@xmath99 network has @xmath155 nodes and @xmath156 edges .
[ rem : dim - n_size ] for each integer @xmath153 and alphabet @xmath26 , the dim-@xmath99 network has an @xmath99-dimensional vector routing solution over @xmath26 .
[ lem : dim - n_rout ] each message and edge symbol is an element of @xmath157 .
let @xmath158_i$ ] denote the @xmath32th component of @xmath159 .
define an @xmath99-dimensional routing code over @xmath26 by @xmath160_k & = \left [ { x_{i}^{(k ) } } \right]_j & & ( i , j , k = 1,\dots , n ) .
\end{aligned}\ ] ] that is , the @xmath4th component of the @xmath161th out - edge of the @xmath32th source node carries the @xmath161th component of the @xmath4th message originating at the @xmath32th source node . for each @xmath53 and each @xmath162 , let the set of @xmath163 parallel edges from node @xmath164 to receiver @xmath165 carry the symbols @xmath166 .
then each receiver gets the first @xmath163 components of every message from the edges originating at @xmath167 , so in particular , each receiver can recover the first @xmath163 components of each of the messages it demands .
node @xmath168 receives the @xmath99th component of each message , so each of its out - edges can carry any @xmath99 of these components .
let @xmath169 , suppose @xmath170 are the messages receiver @xmath165 demands , and let @xmath171_k & = \left [ { w_{k}^{(n ) } } \right]_{i_k } = \left [ { x_{k}^{(i_k ) } } \right]_n & & ( k = 1,\dots , n ) .
\end{aligned}\ ] ] then @xmath165 can recover the @xmath99th component of each of the messages it demands . since @xmath161 was chosen arbitrarily , the code is an @xmath99-dimensional vector routing solution .
the following lemmas will be used in later proofs .
let @xmath7 be a finite ring and let @xmath172 be positive integers .
if a network has @xmath172-dimensional vector linear solutions over @xmath7 , then the network has a @xmath173-dimensional vector linear solution over @xmath7 .
[ lem : dim_sum ] assume a network has a @xmath174-dimensional vector linear solution over @xmath7 for each @xmath53 .
in the @xmath174-dimensional vector linear solution over @xmath7 , every edge function is of the form @xmath175 where @xmath176 are the inputs to the node and @xmath177 are @xmath178 matrices over @xmath7 . for any such edge function , define a @xmath173-dimensional vector linear edge function over @xmath7 by letting @xmath179 & = \sum_{j = 1}^m \left [ \begin{array}{ccl } m_j^{(1 ) } & & \text{\large0 } \\ & \ddots \\ \text{\large0 } & & m_j^{(n ) } \end{array } \right ] \ , \left [ \begin{array}{c } { x_{j}^{(1 ) } } \\ \vdots \\ { x_{j}^{(n ) } } \end{array}\right ] .
\end{aligned}\ ] ] it is straightforward to see this provides a vector linear solution for the network .
let @xmath180 and @xmath181 be collections of discrete random variables over alphabet @xmath26 , and let @xmath182 be the probability mass function of @xmath180 .
we denote the ( base @xmath183 ) _ entropy _ of @xmath180 as @xmath184 and the _ conditional entropy _ of @xmath180 given @xmath181 as @xmath185 the proof of theorem [ thm : dim - n_vector ] will make use of lemmas [ lem : entropy_sum ] and [ lem : int_ent ] and the following basic information inequalities : @xmath186 let @xmath187 be collections of discrete random variables .
then @xmath188 [ lem : entropy_sum ] @xmath189}\\ & = ( n-1 ) { h\left(x\right ) } + { h\left(x , y_1,\dots , y_n\right)}. \end{aligned}\ ] ] ( * ? ? ?
* lemma v.9 ) let @xmath190 be a linear map , and let @xmath83 be a uniformly distributed random variable on @xmath191 . then @xmath192 is uniformly distributed on the range of @xmath193 , and the base @xmath130 entropy of @xmath192 is @xmath194 .
[ lem : int_ent ] for each integer @xmath153 and each field @xmath33 , the dim-@xmath99 network has a @xmath4-dimensional vector linear solution over @xmath33 if and only if @xmath195 .
[ thm : dim - n_vector ] suppose @xmath195 .
then @xmath196 for some integer @xmath197 . by lemma [ lem : dim - n_rout ]
, the dim-@xmath99 network has an @xmath99-dimensional vector linear solution over @xmath33 , so by taking @xmath198 in lemma [ lem : dim_sum ] , the dim-@xmath99 network has an @xmath199-dimensional vector linear solution over @xmath33 .
conversely , suppose the dim-@xmath99 network has a @xmath4-dimensional vector linear solution over field @xmath33
. then all messages @xmath200 and edge symbols @xmath201 are @xmath4-vectors over @xmath33 . by viewing the message components as independent uniform random variables over @xmath33 and considering the entropy using logarithms base @xmath130
, we have @xmath202 for each @xmath53 , the edge symbols @xmath203 are linear functions of @xmath204 , so @xmath205 the receiver @xmath206 demands the messages @xmath207 and recovers its demands from its in - edges , so @xmath208 for each @xmath209 , the edge symbol @xmath210 is a linear function of only @xmath211 , and the network s messages are jointly independent .
thus , @xmath212 } \\ & \ ; \leq { h\left ( u_{1 } , { x_{1}^{(1 ) } } , \dots , { x_{n}^{(1 ) } } , { w_{1}^{(1 ) } } , \dots,{w_{1}^{(n-1 ) } } , \dots,{w_{n}^{(1 ) } } , \dots,{w_{n}^{(n-1 ) } } \right ) } & & { \left[\mbox{from \eqref{eq : ent_2 } } \right]}\\ & \ ; = { h\left(u_{1 } , { w_{1}^{(1 ) } } , \dots,{w_{1}^{(n-1 ) } } , \dots,{w_{n}^{(1 ) } } , \dots,{w_{n}^{(n-1)}}\right ) } & & { \left[\mbox{from \eqref{eq : entropy_4 } } \right ] } \\ & \ ; \le { h\left(u_{1}\right ) } + \sum_{i = 1}^n \sum_{j = 1}^{n-1 } { h\left({w_{i}^{(j)}}\right ) } & & { \left[\mbox{from \eqref{eq : ent_3 } } \right]}\\ & \ ; \leq k \ , ( 1 + n ( n-1 ) ) .
\end{aligned}\ ] ] by a similar argument , for any @xmath213 , there exists a receiver which demands the messages @xmath170 , so @xmath214 since @xmath215 is a cut - set for each receiver , we have @xmath216 therefore , @xmath217 } \\ & \leq { h\left ( { x_{1}^{(1 ) } } , \dots , { x_{1}^{(n ) } } , \dots , { x_{n}^{(1 ) } } , \dots , { x_{n}^{(n ) } } , { w_{1}^{(1 ) } } , \dots,{w_{1}^{(n ) } } , \dots,{w_{n}^{(1 ) } } , \dots,{w_{n}^{(n ) } } \right ) } & & { \left[\mbox{from \eqref{eq : ent_2 } } \right]}\\ & = { h\left({w_{1}^{(1 ) } } , \dots,{w_{1}^{(n ) } } , \dots,{w_{n}^{(1 ) } } , \dots,{w_{n}^{(n ) } } \right ) } & & { \left[\mbox{from \eqref{eq : entropy_6 } } \right ] } \\ & \le \sum_{i=1}^n \sum_{j=1}^n { h\left({w_{i}^{(j)}}\right ) } & & { \left[\mbox{from \eqref{eq : ent_3 } } \right]}\\ & \leq k n^2 \end{aligned}\ ] ] which implies @xmath218 but , since @xmath219 , we get @xmath220 also , since @xmath221 are independent , @xmath222 for each @xmath53 , we have @xmath223 } \notag \\ & \ ; = k ( n-1 ) ( n-1 ) + { h\left ( { x_{i}^{(1 ) } } , \dots , { x_{i}^{(n ) } } \right ) } & & { \left[\mbox{from \eqref{eq : entropy_3 } , \eqref{eq : entropy_7 } } \right ] } \notag \\ & \;= k ( n^2 - n + 1 ) & & { \left[\mbox{from \eqref{eq : entropy_2 } } \right ] } \label{eq : entropy_9}. \end{aligned}\ ] ] by fixing @xmath224 and summing over all @xmath225 in , we have @xmath226}\\ & \ ; = n^{n-1 } { h\left({w_{1}^{(1 ) } } , \dots,{w_{1}^{(n-1 ) } } , { x_{1}^{(1 ) } } \right ) } + n^{n-2 } \sum_{j=2}^n \sum_{i=1}^n { h\left({w_{j}^{(1 ) } } , \dots,{w_{j}^{(n-1 ) } } , { x_{j}^{(i ) } } \right ) } \\ & \ ; \geq n^{n-1 } { h\left({w_{1}^{(1 ) } } , \dots,{w_{1}^{(n-1 ) } } , { x_{1}^{(1 ) } } \right ) } + n^{n-2 } \sum_{j=2}^n k ( n^2 - n + 1 ) & & { \left[\mbox{from \eqref{eq : entropy_9 } } \right ] } \\ & \ ; = n^{n-1 } { h\left({w_{1}^{(1 ) } } , \dots,{w_{1}^{(n-1 ) } } , { x_{1}^{(1 ) } } \right ) } + n^{n-2 } \ , k \ , ( n-1 ) ( n^2 - n + 1 ) \end{aligned}\ ] ] and so @xmath227 similarly , for each @xmath228 , we have @xmath229 however , for each @xmath230 we also have @xmath231 } \\ & \leq \sum_{j = 1}^n k \left ( \frac { n^2 - n + 1 } { n } \right ) & & { \left[\mbox{from \eqref{eq : entropy_10 } } \right]}\\ & = k \ , ( n^2 - n + 1 ) \end{aligned}\ ] ] and so for each @xmath228 , @xmath232 the variables @xmath233 are linear functions of the uniformly distributed messages , so by lemma [ lem : int_ent ] , @xmath234 ( with logarithms in base @xmath130 ) is an integer .
however , @xmath235 so if @xmath236 is an integer , then we must have @xmath195 .
the following corollary demonstrates it is possible for a network to be scalar linearly solvable over a non - commutative ring but not over any commutative rings , which is , in fact , equivalent to a network being vector linearly solvable over some field but not scalar linearly solvable over any field , by corollaries [ cor : general_vector ] and [ cor : general_scalar ] . for all integers @xmath153 , @xmath237 , and prime @xmath1
, the dim-@xmath99 network has a scalar linear solution over a non - commutative ring of size @xmath238 but has no scalar linear solution over any commutative ring .
[ cor : non - comm - ring ] if the dim-@xmath99 network were scalar linearly solvable over a commutative ring , then by corollary [ cor : general_scalar ] , the dim-@xmath99 network would also be scalar linearly solvable over some finite field .
however , by theorem [ thm : dim - n_vector ] , the dim-@xmath99 network is not scalar linearly solvable over any finite field . by theorem [ thm : dim - n_vector ]
, the dim-@xmath99 network has an @xmath99-dimensional vector linear solution over @xmath141 , so by corollary [ cor : same_ring ] the dim-@xmath99 network has a linear solution over the ring @xmath239 . for each integer @xmath153 ,
the unique smallest - size ring over which the dim-@xmath99 network is scalar linearly solvable is the ring of all @xmath240 matrices over @xmath46 .
[ cor : dim - n_smallest_ring ] by theorem [ thm : dim - n_vector ] , the dim-@xmath99 network has an @xmath99-dimensional vector linear solution over @xmath46 , and by corollary [ cor : same_ring ] , the dim-@xmath99 network has a linear solution over the ring @xmath241 .
suppose the dim-@xmath99 network is scalar linearly solvable over a ring @xmath7 such that @xmath242 . by lemmas [ lem : simple_rings ] and
[ lem : hom_to_simple ] ( a ) ( b ) there exists a field @xmath33 , a positive integer @xmath4 , and a surjective homomorphism @xmath243 such that the dim-@xmath99 network is scalar linearly solvable over @xmath128 . by corollary [ cor : same_ring ] , this implies the dim-@xmath99 network has a @xmath4-dimensional vector linear solution over @xmath33 , which by theorem [ thm : dim - n_vector ] , implies @xmath99 divides @xmath4 .
since @xmath88 is surjective , @xmath244 .
hence we have @xmath245 therefore @xmath246 and @xmath247 . since @xmath248 and @xmath88 is a surjective homomorphism
, we have @xmath249 setting @xmath250 and @xmath251 in corollary [ cor : non - comm - ring ] results in the m network ( see figure [ fig : m - network ] ) having no scalar linear solution over any commutative ring but having a scalar linear solution over a non - commutative ring of size @xmath0 .
the non - commutative ring @xmath252 consists of all @xmath253 binary matrices under ordinary matrix addition and multiplication mod @xmath254 . denote the @xmath0 ring elements by : @xmath255 \ \ \ \ \ \ \ \ ( q , r , s , t \in \{0,1\}).\end{aligned}\ ] ] a scalar linear solution for the m network over the non - commutative ring @xmath252 + ( i.e. where @xmath256 ) is given by : @xmath257 where the out - edges of nodes with a single in - edge each carry the symbol on the in - edge , that is , each receiver directly receives the edge symbols @xmath258 and @xmath259 from the nodes @xmath260 and @xmath261 , respectively
. we also note that if the messages and edge symbols of the m network are @xmath254-dimensional vectors over @xmath46 , instead of @xmath262 binary matrices , then a small modification of the linear code described above provides the @xmath254-dimensional vector linear solution over @xmath46 given in @xcite .
this agrees with corollary [ cor : same_ring ] .
[ ex:16-noncommutative ] the bound in the following theorem is tight via example [ ex:16-noncommutative ] .
if a network is scalar linearly solvable over some non - commutative ring @xmath7 , but not over any commutative rings , then @xmath35 .
[ thm : non - comm_16 ] suppose network @xmath50 is scalar linearly solvable over some non - commutative ring @xmath7 and is not linearly solvable over any commutative ring . by theorem [ thm : r_dom_by_simple ]
, there exists a positive integer @xmath4 and a field @xmath33 such that @xmath50 has a linear solution over @xmath128 and @xmath263 .
if @xmath132 , then @xmath50 is linearly solvable over a field , which contradicts the assumption that @xmath50 is not linearly solvable over any commutative ring .
so @xmath103 , which implies @xmath264 .
the following theorem demonstrates that there exists a network that is linearly solvable over a module of size @xmath3 but not over any ring of size @xmath3 . for each integer @xmath103 and prime @xmath1
, the dim-@xmath4 network has a @xmath4-dimensional vector linear solution over the field @xmath5 but is not scalar linearly solvable over any ring of size @xmath3 .
[ thm : module - beats - rings ] by theorem [ thm : dim - n_vector ] , the dim-@xmath4 network has a @xmath4-dimensional vector linear solution over @xmath5 .
let @xmath7 be a ring of size @xmath3 and suppose the dim-@xmath4 network has a scalar linear solution over @xmath7 . by lemmas [ lem : simple_rings ] and
[ lem : hom_to_simple ] ( b ) ( c ) , there exists a field @xmath33 and a positive integer @xmath99 such that any network that is scalar linearly solvable over @xmath7 is also scalar linearly solvable over @xmath143 and @xmath265 divides @xmath3 .
hence @xmath33 is a field of characteristic @xmath1 and @xmath266 .
since the dim-@xmath4 network is scalar linearly solvable over @xmath7 , the dim-@xmath4 network is scalar linearly solvable over the ring @xmath143 . by corollary [ cor : same_ring ] , this implies the dim-@xmath4 network has an @xmath99-dimensional vector linear solution over @xmath33 , which by theorem [ thm : dim - n_vector ] implies @xmath267 .
however , this contradicts the fact that @xmath266 .
thus , no such ring @xmath7 exists .
both a scalar linear code over a ring of size @xmath3 and a @xmath4-dimensional vector linear code are linear codes over a module of size @xmath3 .
we have already seen ( in theorem [ thm : module - beats - rings ] ) that there exists a network with a @xmath4-dimensional vector linear solution over @xmath5 yet with no scalar linear solutions over any ring of size @xmath3 .
the main result of this section ( theorem [ thm : commutative_implies_vector ] ) will show that any network that is scalar linearly solvable over a commutative ring of size @xmath3 must also have a @xmath4-dimensional vector linear solution over @xmath5 .
the following lemma was proved in part i ( in ( * ? ? ?
* lemmas 2.6 and 5.4 ) ) and will be used in what follows . for each prime @xmath1 and positive integer @xmath4 ,
if a network @xmath50 has a scalar linear solution over some commutative ring of size @xmath3 , then there exists an integer partition @xmath268 of @xmath4 such that @xmath50 is scalar linearly solvable over each of the fields @xmath269 . [
lem : dom_by_max ] the following standard result on rings will be used in later proofs .
* theorem i.1 ) every finite ring is isomorphic to a direct product of rings of prime power sizes . [
lem : prime - power - sizes ] let @xmath270 be a positive integer with prime factorization @xmath271 . if a network @xmath50 has a scalar linear solution over some commutative ring of size @xmath270 , then the following hold : * for each @xmath272 , network @xmath50 has a @xmath174-dimensional vector linear solution over @xmath273 .
* network @xmath50 has a linear solution over the + @xmath274-module @xmath275 .
[ thm : commutative_implies_vector ] suppose @xmath50 is scalar linearly solvable over a commutative ring @xmath7 of size @xmath270 .
by lemma [ lem : prime - power - sizes ] , there exist rings @xmath276 such that @xmath277 and @xmath278 for all @xmath32 .
let @xmath279 . since the projection mapping from @xmath7 to @xmath280 is a surjective homomorphism , by corollary [ cor : modhomomorphism ] , network @xmath50 is scalar linearly solvable over @xmath280 .
then by lemma [ lem : dom_by_max ] , there exists an integer partition @xmath268 of @xmath174 such that @xmath50 is scalar linearly solvable over each of the fields @xmath281 . by lemma [ lem : kn - vector ] , this implies that @xmath50 has an @xmath282-dimensional vector linear solution over @xmath273 for each @xmath283 .
however , by lemma [ lem : dim_sum ] , this then implies that @xmath50 has a @xmath284-dimensional vector linear solution over @xmath273 . hence , for all @xmath279 , a cartesian product code formed from the @xmath285-dimensional vector linear solutions over @xmath286 gives a linear solution to @xmath50 over the described module . in part
i , we showed ( in ( * ? ? ?
* theorems 5.8 and 5.9 ) ) that with respect to ring domination for scalar linear coding , some ring sizes give rise to multiple maximal commutative rings whereas other ring sizes yield only a single unique maximal commutative ring .
if there is just one maximal commutative ring of size @xmath270 , then every network that is linearly solvable over some commutative ring of size @xmath270 is also linearly solvable over the maximal ring .
in contrast , if there are multiple maximal commutative rings of size @xmath270 , then for any commutative ring @xmath7 of size @xmath270 , there is always a different commutative ring @xmath44 also of size @xmath270 , such that some network is scalar linearly solvable over @xmath44 but not over @xmath7 .
thus , in this sense , there is no `` best '' commutative ring of a given size .
however , by theorem [ thm : commutative_implies_vector ] ( b ) , if a network has a linear solution over some commutative ring of size @xmath271 , then it has a linear solution over the @xmath274-module @xmath275 , which also has size @xmath270 .
in fact , we showed ( in theorem [ thm : module - beats - rings ] ) that when @xmath287 , the converse is not true .
so in this sense , @xmath4-dimensional vector linear codes over @xmath5 are strictly `` better '' than scalar linear codes over commutative rings of size @xmath3 .
this section generalizes the results of theorem [ thm : commutative_implies_vector ] to ( not necessarily commutative ) rings of size @xmath270 with prime factor multiplicity less than or equal to @xmath288 . in order to do so , we first will prove some intermediate results and consider special cases .
the following lemma was proved in part i ( in ( * ? ? ?
* theorem 5.9 ) ) and will be used in what follows . for each @xmath289 and prime @xmath1 ,
if a network is scalar linearly solvable over some commutative ring of size @xmath3 , then it is scalar linearly solvable over @xmath141 .
[ lem : p12346 ] lemma [ lem : ring_size_p123 ] characterizes the non - commutative rings of prime - power size whose multiplicity is at most three .
@xcite for each prime @xmath1 , all rings of size @xmath1 and of size @xmath290 are commutative , and the ring of all upper - triangular @xmath262 matrices over @xmath5 is the only non - commutative ring of size @xmath291 .
[ lem : ring_size_p123 ] we remark that there exist rings of size @xmath1 and @xmath290 without identity .
for example , the set @xmath292 with mod @xmath293 addition and multiplication satisfies all of the properties of a ring except there is no multiplicative identity .
however , such rings ( sometimes called `` rngs '' ) do not appear to be practical for linear network coding , as receivers must recover their demands from linear combinations of their inputs .
for example , consider the trivial network shown in figure [ fig : trivial ] consisting of a single message @xmath83 emitted by a source directly connected by a single edge to a receiver demanding message @xmath83 .
the only possible linear functions that can be carried on the edge are of the form @xmath294 for some fixed @xmath295 .
however , no matter what the choice of @xmath296 is , the messages @xmath58 and @xmath297 always get received as @xmath58 mod @xmath293 , so the receiver can not uniquely determine @xmath83 in general .
thus , there is no linear solution for the network over this ring ( with no multiplicative identity ) .
a similar issue arises for the set @xmath298 with mod @xmath297 addition and multiplication , which also satisfies all of the properties of a ring except there is no multiplicative identity .
for each prime @xmath1 , if a network is scalar linearly solvable over some ring of size @xmath290 , then it is a scalar linearly solvable over @xmath299 .
[ lem : p2 ] by lemma [ lem : ring_size_p123 ] , every ring of size @xmath290 is commutative , and by lemma [ lem : p12346 ] , every network that is scalar linearly solvable over some commutative ring of size @xmath290 has a scalar linear solution over @xmath299 . by lemma [ lem : ring_size_p123 ]
, the smallest non - commutative ring is the ring of the @xmath293 binary upper - triangular @xmath253 matrices . as a special case of the following lemma , any network that is scalar linearly solvable over this ring must also have a scalar linear solution over @xmath46 .
for each finite field @xmath33 and integer @xmath103 , any network that is scalar linearly solvable over the ring of upper - triangular @xmath18 matrices over @xmath33 is also scalar linearly solvable over @xmath33 .
[ lem : p3_matrices ] let @xmath7 be the ring of upper - triangular @xmath18 matrices with entries in @xmath33 and let @xmath300 be given by @xmath301 } \right ) = a_{1,1}.\ ] ] then @xmath88 is clearly surjective and preserves identities , and for any @xmath302 , @xmath303 thus @xmath88 is a surjective homomorphism , so by corollary [ cor : modhomomorphism ] , any network that is scalar linearly solvable over @xmath7 is scalar linearly solvable over @xmath33 . for each prime @xmath1
, if a network is scalar linearly solvable over some ring of size @xmath291 , then it is scalar linearly solvable over @xmath304 .
[ lem : p3 ] by lemma [ lem : ring_size_p123 ] , the only non - commutative ring of size @xmath291 is the ring of upper triangular matrices with entries in @xmath5 , and by lemma [ lem : p3_matrices ] , any network that is scalar linearly solvable over this ring is also scalar linearly solvable over @xmath5 . since @xmath5 is a subring of @xmath304 , any network that is scalar linearly solvable over @xmath5 is scalar linearly solvable over @xmath304 . by lemma [ lem : p12346 ] , every network that is scalar
linearly solvable over some commutative ring of size @xmath291 has a scalar linear solution over @xmath304 .
the following three lemmas are proved in the appendix . for each prime @xmath1 ,
if a network is scalar linearly solvable over some ring of size @xmath109 , then it is scalar linearly solvable over at least one of the rings @xmath107 or @xmath108 .
[ lem : p4 ] for each prime @xmath1 , if a network is scalar linearly solvable over some ring of size @xmath305 , then it is scalar linearly solvable over at least one of the rings @xmath306 or @xmath307 .
[ lem : p5 ] for each prime @xmath1 , if a network is scalar linearly solvable over some ring of size @xmath308 , then it is scalar
linearly solvable over @xmath309 .
[ lem : p6 ] theorem [ thm : p23456 ] is a generalization of lemma [ lem : p12346 ] to scalar linear codes over non - commutative rings . extending theorem [ thm : p23456 ] to @xmath310 for @xmath311
is left as an open problem .
let @xmath1 be a prime , and suppose @xmath50 is scalar linearly solvable over a ring @xmath7 .
* if @xmath312 , then @xmath50 is scalar linearly solvable over @xmath299 . *
if @xmath313 , then @xmath50 is scalar linearly solvable over @xmath304 . *
if @xmath314 , then @xmath50 is scalar linearly solvable over at least one of @xmath107 or @xmath108 . *
if @xmath315 , then @xmath50 is scalar linearly solvable over at least one of @xmath306 or @xmath307 . *
if @xmath316 , then @xmath50 is scalar linearly solvable over @xmath309 .
[ thm : p23456 ] this follows immediately from lemmas [ lem : p2 ] , [ lem : p3 ] , [ lem : p4 ] , [ lem : p5 ] , and [ lem : p6 ] .
we also note that by corollary [ cor : choose - two - ring ] , the @xmath317-choose - two network is scalar linearly solvable over @xmath107 but not over @xmath108 and the @xmath318-choose - two network is scalar
linearly solvable over @xmath306 but not over @xmath307 . by corollary [ cor : non - comm - ring ] , the dim-@xmath254 network is scalar linearly solvable over @xmath108 but not over @xmath107 .
we showed in part i ( * ? ? ? * theorem 3.10 ) that there exists a network that is scalar linearly solvable over @xmath307 but not over @xmath306 .
hence it is necessary to include both rings in ( c ) and ( d ) in theorem [ thm : p23456 ] .
let @xmath1 be a prime and @xmath319 , and suppose @xmath50 is scalar linearly solvable over a ring of size @xmath3
. then @xmath50 has a @xmath4-dimensional vector linear solution over @xmath5 .
[ cor : p23456 ] if @xmath320 , then by theorem [ thm : p23456 ] , @xmath50 has a scalar linear solution over a commutative ring of size @xmath3 , so by theorem [ thm : commutative_implies_vector ] , @xmath50 has a @xmath4-dimensional vector linear solution over @xmath5 .
now suppose @xmath321 .
if @xmath50 is scalar linearly solvable over @xmath107 , then by lemma [ lem : kn - vector ] , @xmath50 has a @xmath297-dimensional vector linear solution over @xmath5 .
if @xmath50 is not scalar linearly solvable over @xmath107 , then by theorem [ thm : p23456 ] ( c ) , @xmath50 must be scalar linearly solvable over @xmath108 , so by corollary [ cor : same_ring ] , @xmath50 has a @xmath254-dimensional vector linear solution over @xmath5 , in which case @xmath50 also has a @xmath297-dimensional vector linear solution over @xmath5 by lemma [ lem : dim_sum ] . theorem [ thm : p6_implies_vector ] generalizes the results of theorem [ thm : commutative_implies_vector ] to rings of size @xmath270 with prime factor multiplicity less than or equal to @xmath288 . let @xmath270 be a positive integer with prime factorization @xmath271
. if a network @xmath50 has a scalar linear solution over a ring of size @xmath270 , then , for each @xmath272 such that @xmath322 , network @xmath50 has a @xmath174-dimensional vector linear solution over @xmath273 .
[ thm : p6_implies_vector ] suppose @xmath50 is scalar linearly solvable over a ring @xmath7 of size @xmath270 . by lemma
[ lem : prime - power - sizes ] , there exists rings @xmath276 such that @xmath277 and @xmath278 for all @xmath32 .
now , let @xmath279 and suppose @xmath322 .
the projection mapping from @xmath7 to @xmath280 is a surjective homomorphism , so by corollary [ cor : modhomomorphism ] , network @xmath50 is scalar linearly solvable over @xmath280 .
since @xmath50 is scalar linearly solvable over a ring of size @xmath323 where @xmath322 , by corollary [ cor : p23456 ] , @xmath50 has a @xmath174-dimensional vector linear solution over @xmath273 .
we leave as an open question whether the restriction that @xmath322 can be removed from the statement of theorem [ thm : p6_implies_vector ] .
if this generalization is false , then for what primes @xmath1 and positive integers @xmath4 is it the case that there exists a network with a scalar linear solution over a ring of size @xmath3 but with no @xmath4-dimensional vector linear solution over @xmath5 ?
if such a ring and such a network do exist , the ring must be non - commutative and @xmath311 .
more generally , does there exist a network with a linear solution over some alphabet of size @xmath3 but with no @xmath4-dimensional vector linear solution over @xmath5 ?
for each positive integer @xmath4 and prime @xmath1 , we have shown the set @xmath324 is properly contained in @xmath325 so in this sense , @xmath4-dimensional vector linear codes over @xmath5 may be advantageous compared to scalar linear codes over commutative rings of the same size @xmath3 . in addition , there are more @xmath4-dimensional linear functions over @xmath5 than there over a commutative ring of size @xmath3 . vector linear codes over fields are also optimal in the sense that they minimize the alphabet size needed for a linear solution over a particular network . on the other hand ,
the complexity of implementing vector linear codes is generally higher than for scalar linear codes over commutative rings of the same size .
the main purpose of this appendix is to prove lemmas [ lem : p4 ] , [ lem : p5 ] , and [ lem : p6 ] , which are used in the proof of theorem [ thm : p23456 ] .
it is an open question whether theorem [ thm : p6_implies_vector ] can be extended to all finite rings .
the techniques presented in this section may additionally be useful for examining such questions .
recall that a finite ring is simple if it has no proper two - sided ideals .
the _ radical _ of a ring @xmath7 is the intersection of all its maximal left ideals .
the radical of a ring is a two - sided ideal .
a finite ring @xmath7 with radical @xmath326 is said to be : * _ _ local _ _ is a local commutative ring , then @xmath7 has a single maximal ideal , which corresponds to our definition of a commutative local ring in part i. ] if @xmath327 is a field . *
_ semi - local _ if @xmath327 is simple , or equivalently @xmath7 is isomorphic to a matrix over a local ring ( e.g. @xcite ) . *
_ semi - simple _ if @xmath7 is isomorphic to a direct product of simple rings ( matrix rings over fields ) or equivalently , @xmath328 ( e.g. @xcite ) .
let @xmath1 be a prime , @xmath4 a positive integer , and @xmath7 a semi - local ring of size @xmath3 .
then there exists a unique local ring @xmath44 and positive integers @xmath329 such that the following hold : as an example , let @xmath1 be a prime and let @xmath334 be positive integers . then @xmath335 is a semi - local ring , since @xmath336 is a local ring .
we also remark that in lemma [ lem : local_rings ] , if @xmath7 is itself local , then @xmath337 .
* proposition iv.6 , theorem viii.4 ) ) let @xmath7 be a finite ring with radical @xmath326 . then there exist fields @xmath338 and positive integers @xmath339 such that @xmath340 [ lem : semisimple_radical ] let @xmath7 be a finite ring with radical @xmath326 , and suppose @xmath341 for some fields @xmath338 and positive integers @xmath339
. if a network is scalar linearly solvable over @xmath7 , then it is also scalar linearly solvable over each of the rings @xmath342 .
[ lem : p6_semi1 ] by lemma [ lem : surjective_homomorphism ] , there exists a surjective homomorphism @xmath343 .
let @xmath344 .
then the projection mapping @xmath345 is a surjective homomorphism .
hence the composition of mappings @xmath346 is a surjective homomorphism .
thus by corollary [ cor : modhomomorphism ] , any network with a scalar linear solution over @xmath7 has a scalar linear solution over the ring @xmath347 .
the following is an enumeration of semi - simple rings that we will reference in upcoming proofs . for each prime @xmath1
, it can be verified that the rings given in
are all of the semi - simple rings of sizes @xmath348 or @xmath308 ( up to isomorphism ) .
let @xmath7 be a ring of size @xmath109 with radical @xmath326 , and suppose @xmath50 is scalar linearly solvable over @xmath7 . then @xmath351 , so by lemma [ lem : semisimple_radical ] , @xmath327 is isomorphic to one of the rings in
. if @xmath327 is isomorphic to any of these rings except those in and , then by lemma [ lem : p6_semi1 ] , @xmath50 is also scalar linearly solvable over at least one of @xmath5 , @xmath299 , or @xmath107 .
since @xmath5 and @xmath299 are both subrings of @xmath107 , in these cases , @xmath50 is also scalar linearly solvable over @xmath107 . if @xmath327 is isomorphic to the ring in , then by lemma [ lem : p6_semi1 ] , @xmath50 is also scalar linearly solvable over @xmath108 .
it follows from lemma [ lem : local_rings ] that @xmath327 is not isomorphic to the ring in .
let @xmath7 be a ring of size @xmath305 with radical @xmath326 , and suppose @xmath50 is scalar linearly solvable over @xmath7 .
then @xmath352 , so by lemma [ lem : semisimple_radical ] , @xmath327 must be isomorphic to one of the rings in . if @xmath327 is isomorphic to one of the rings in
( i.e. @xmath353 ) , then @xmath354 .
since @xmath355 is an @xmath7-module and @xmath50 has a linear solution over the faithful module @xmath24 , by lemma [ lem : same_ring ] , @xmath50 has a linear solution over @xmath356 . by theorem [ thm : min_module ] , this implies @xmath50 has a scalar linear solution over @xmath5 .
since @xmath5 is a subring of @xmath306 , in these cases , @xmath50 also has a scalar linear solution over @xmath306 .
it follows from lemma [ lem : local_rings ] that @xmath327 is not isomorphic to either of the rings in or . if @xmath327 is isomorphic to the ring in , then by lemma [ lem : p6_semi1 ] , @xmath50 is scalar linearly solvable over @xmath307 .
if @xmath327 is isomorphic to any of the remaining cases , then by lemma [ lem : p6_semi1 ] , network @xmath50 is scalar linearly solvable over either @xmath5 or @xmath306 . since @xmath5 is a subring of @xmath306 , in these cases
, @xmath50 also has a scalar linear solution over @xmath306 .
let @xmath7 be a ring of size @xmath308 with radical @xmath326 , and suppose @xmath50 is scalar linearly solvable over @xmath7 . then @xmath357 , so by lemma [ lem : semisimple_radical ] , @xmath327 must be isomorphic to one of the rings in . it follows from lemma [ lem : local_rings ] that @xmath327 is not isomorphic to any of the rings in , , or .
if @xmath327 is isomorphic to any of the remaining cases , then it follows from lemma [ lem : p6_semi1 ] that @xmath50 is scalar linearly solvable over @xmath358 for some @xmath359 . since @xmath360 , @xmath358 is a subring of @xmath309 , which implies @xmath50 is scalar linearly solvable over @xmath309 .
j. connelly and k. zeger , `` linear network coding over rings , part i : scalar codes and commutative alphabets , '' submitted to _ ieee transactions on information theory _ , available on - line at : http://arxiv.org/abs/1608.01738 . | we prove the following results regarding the linear solvability of networks over various alphabets . for any network ,
the following are equivalent : ( i ) vector linear solvability over some finite field , ( ii ) scalar linear solvability over some ring , ( iii ) linear solvability over some module .
analogously , the following are equivalent : ( a ) scalar linear solvability over some finite field , ( b ) scalar linear solvability over some commutative ring ,
( c ) linear solvability over some module whose ring is commutative . whenever any network is linearly solvable over a module ,
a smallest such module arises in a vector linear solution for that network over a field .
if a network is linearly solvable over some non - commutative ring but not over any commutative ring , then such a non - commutative ring must have size at least @xmath0 , and for some networks , this bound is achieved .
an infinite family of networks is demonstrated , each of which is scalar linearly solvable over some non - commutative ring but not over any commutative ring . whenever @xmath1 is prime and @xmath2 , if a network is scalar linearly solvable over some ring of size @xmath3 , then it is also @xmath4-dimensional vector linearly solvable over the field @xmath5 , but the converse does not necessarily hold .
this result is extended to all @xmath6 when the ring is commutative .
[ section ] [ theorem ] lemma [ theorem ] corollary [ theorem ] proposition [ theorem ] remark [ theorem ] conjecture [ theorem ] example [ theorem ] definition |
we study the admission control problem in general graphs with edge capacities .
an online algorithm can receive a sequence of communications requests on a virtual path , that may be accepted or rejected , while staying within the capacity limitations .
this problem has typically been studied as a benefit problem .
this means that the online algorithm has to be competitive with respect to the number of accepted requests .
a problem with this objective function is that in some cases an online algorithm with a good competitive ratio may reject the vast majority of the requests , whereas the optimal solution rejects only a small fraction of them . in this paper
we consider the goal of minimizing the number of rejected requests , which was first studied in @xcite .
this approach is suitable for applications in which rejections are intended to be rare events .
a situation in which a significant fraction of the requests is rejected even by the optimal solution means that the network needs to be upgraded .
we consider preemptive online algorithms for the admission control problem .
allowing preemption is necessary for achieving reasonable bounds for the competitive ratio .
each request arrives together with the path it should be routed on .
the admission control algorithm decides whether to accept or reject it .
an online algorithm for both admission control and routing easily admits a trivial lower bound @xcite . * the admission control to minimize rejections problem .
* we now formally define the admission control problem .
the input consist of the following : * a directed graph @xmath12 , where @xmath13 .
each edge @xmath14 has an integer capacity @xmath15 .
we denote @xmath16 . *
a sequence of requests @xmath17 each of which is a simple path in the graph .
every request @xmath18 has a cost @xmath19 associated with it . a feasible solution for the problem must assure that for every edge @xmath14
, the number of accepted requests whose paths contain @xmath14 is at most its capacity @xmath20 .
the goal is to find a feasible solution of minimum cost of the rejected requests .
the online algorithm is given requests one at a time , and must decide whether to accept or reject each request .
it is also allowed to preempt a request , i.e. to reject it after already accepting it , but it can not accept a request after rejecting it .
let @xmath21 be a feasible solution having minimum cost @xmath22 .
an algorithm is @xmath23-competitive if the total cost of the requests rejected by this algorithm is at most @xmath24 . * previous results for admission control .
* tight bounds were achieved for the admission control problem , where the goal is to maximize the number of accepted requests .
awerbuch , azar and plotkin @xcite provide an @xmath25-competitive algorithm for general graphs .
for the admission control problem on a tree , @xmath26-competitive randomized algorithms appear in @xcite , where @xmath27 is the diameter of the tree .
adler and azar presented a constant - competitive preemptive algorithm for admission control on the line @xcite . the admission control to minimize rejections problem
was studied by blum , kalai and kleinberg in @xcite , where two deterministic algorithms with competitive ratios of @xmath28 and @xmath29 are given ( @xmath1 is the number of edges in the graph and @xmath2 is the maximum capacity ) .
they raised the question of whether an online algorithm with polylogarithmic competitive ratio can be obtained .
we note that one can combine an algorithm for maximizing throughput of accepted requests and an algorithm for minimizing rejections and get one algorithm which achieves both simultaneously with slightly degrading the competitive ratio @xcite . in this paper
we show that the admission control to minimize rejections problem is a generalization of the online set cover with repetitions problem described below : * the online set cover with repetitions problem .
* the online set cover problem is defined as follows : let @xmath30 be a ground set of @xmath4 elements , and let @xmath31 be a family of subsets of @xmath30 , @xmath32 .
each @xmath33 has a non - negative cost associated with it .
an adversary gives elements to the algorithm from @xmath30 one by one .
each element of @xmath30 can be given an arbitrary number of times , not necessarily consecutively .
an element should be covered by a number of sets which is equal to the number of times it arrived .
we assume that the elements of @xmath30 and the members of @xmath31 are known in advance to the algorithm , however , the elements given by the adversary are not known in advance .
the objective is to minimize the cost of the sets chosen by the algorithm . * previous results for online set cover .
* the offline version of the set cover problem is a classic np - hard problem that was studied extensively , and the best approximation factor achievable for it in polynomial time ( assuming @xmath34 ) is @xmath35 @xcite .
the basic online set cover problem , where repetitions are not allowed , was studied in @xcite .
a different variant of the problem , dealing with maximum benefit , is presented in @xcite .
an @xmath36-competitive deterministic algorithm for the online set cover problem was given by @xcite where @xmath4 is the number of elements and @xmath1 is the number of sets .
a lower bound of @xmath37 was also shown for any deterministic online algorithm .
a recent result of feige and korman @xcite establishes a lower bound of @xmath6 for the competitive ratio of any randomized _ polynomial _ time algorithm for the online set cover problem , under the @xmath7 assumption .
they also prove the same lower bound for any deterministic _ polynomial _ time algorithm , under the @xmath34 assumption . * our results . *
the main result we give in this paper is an @xmath0-competitive randomized algorithm for the admission control to minimize rejections problem .
this settles the open question raised by blum et al .
for the unweighted case , where all costs are equal to 1 , we slightly improve this bound and give an @xmath38-competitive randomized algorithm , we present a simple reduction between online set cover with repetitions and the admission control to minimize rejections problem .
this implies an @xmath39-competitive randomized algorithm for the online set cover with repetitions problem .
for the unweighted case ( all costs are equal to @xmath40 ) , we get an @xmath5-competitive randomized algorithm .
this matches the lower bound of @xmath6 given by feige and korman .
their results also imply a lower bound of @xmath41 for the competitive ratio of any randomized _ polynomial _ time algorithm for the admission control to minimize rejections problem ( assuming @xmath7 ) .
the derandomization techniques used in @xcite for the online set cover problem do not seem to apply here .
this is why we also consider the bicriteria version of the online set cover with repetition problem . for a given constant @xmath8 ,
the online algorithm is required to cover each element by a fraction of @xmath42 times the number of its appearances .
specifically , at any point of time , if an element has been requested @xmath11 times so far , then the optimal solution covers it by @xmath11 different sets , whereas the online algorithm covers it by @xmath10 different sets .
we give an @xmath5-competitive deterministic bicriteria algorithm for this problem .
* techniques . * the techniques we use
follow those of @xcite together with some new ideas .
we start with an online fractional solution which is monotone increasing during the algorithm .
then , the fractional solution is converted into a randomized algorithm .
interestingly , to get a deterministic bicriteria algorithm we use a different fractional algorithm than the one used for the randomized algorithm .
in this section we describe a fractional algorithm for the problem .
a fractional algorithm is allowed to reject a fraction of a request @xmath18 .
we use a weight @xmath43 for this fraction .
specifically , if @xmath44 , we reject with percentage of precisely @xmath43 .
if @xmath45 , then the request is completely rejected . at any stage of the fractional algorithm we will use the following notation : * @xmath46 will denote the set of requests that arrived so far whose paths contain the edge @xmath14 .
* @xmath47 will denote @xmath48 .
* @xmath49 will denote the requests from @xmath46 that have not been fully rejected ( requests @xmath18 for which @xmath50 ) .
* @xmath51 will denote the excess of edge @xmath14 caused by the requests in @xmath49 .
the requirement from a fractional algorithm is that for every edge @xmath14 , @xmath53 the cost associated with a fractional algorithm is defined to be @xmath54 .
we will now describe an @xmath55-competitive algorithm for the problem , even versus a fractional optimum .
the cost of the optimal fractional solution , @xmath22 is denoted by @xmath56 .
we may assume , by doubling , that the value of @xmath56 is known up to a factor of @xmath57 . to determine the initial value of @xmath56 we look for the first time in which we must reject a request from an edge @xmath14 .
we can start guessing @xmath58 , and then run the algorithm with this bound on the optimal solution .
if it turns out that the value of the optimal solution is larger than our current guess for it , ( that is , the cost exceeds @xmath59 ) , then we `` forget '' about all the request fractions rejected so far , update the value of @xmath56 by doubling it , and continue .
we note that the cost of the request fractions that we have `` forgotten '' about can increase the cost of our solution by at most a factor of @xmath57 , since the value of @xmath56 was doubled in each step .
we thus assume that @xmath56 is known .
denote by @xmath60 the requests with cost exceeding @xmath61 .
the optimal fractional solution can reject a total fraction of at most @xmath62 out of the requests of @xmath60 .
hence , when an edge is requested more than its capacity , the fractional optimum must reject a total fraction of at least @xmath62 out of the requests not in @xmath60 whose paths contain the edge . by doubling the fraction of rejection for all the requests not in @xmath60 ( keeping fractions to be at most @xmath40 ) and completely accepting all the requests in @xmath60
, we get a feasible fractional solution whose cost is at most twice the optimum .
hence , the online algorithm can always completely accept requests of cost exceeding @xmath61 ( and adjust the edge capacities @xmath20 accordingly ) .
denote by @xmath63 the requests with cost at most @xmath64 .
we claim that we can completely reject all the requests from @xmath63 . for each edge @xmath14 , the optimal solution can accept a total fraction of at most @xmath2 out of the requests whose paths contain the edge @xmath14 , and therefore it can accept a total fraction of at most @xmath65 requests .
the fractions of requests accepted out of @xmath63 have total cost at most @xmath66 .
it follows that the optimal solution pays at least @xmath67 for the fractions of requests out of @xmath63 that it rejected .
therefore , the online algorithm can reject all the requests in @xmath63 and pay @xmath68 .
if @xmath69 , then this adds only @xmath70 to the cost of the online algorithm . if @xmath71 , then @xmath72 , so the online algorithm is @xmath57-competitive with respect to the requests in @xmath63 . by the above arguments ,
all the requests of cost smaller than @xmath64 or greater than @xmath61 are rejected immediately or accepted permanently ( edge capacities are decreased in this case ) , respectively .
an algorithm needs to handle only requests of cost between @xmath64 and @xmath61 .
we normalize the costs so that the minimum cost is @xmath40 and the maximum cost is @xmath73 , and fix @xmath56 appropriately .
the algorithm maintains a weight @xmath43 for each request @xmath18 .
the weights can only increase during the run of the algorithm .
initially @xmath74 for all the requests .
assume now that the algorithm receives a request @xmath18 for a path of cost @xmath75 .
for each edge @xmath14 , we update @xmath46 , @xmath49 and @xmath51 according to the definitions given above . the following is performed for all the edges @xmath14 of the path of @xmath18 , in an arbitrary order . 1 .
if @xmath76 , then do nothing .
2 . else , while @xmath77 , perform a _ weight augmentation _ : 1 . for each @xmath78 , if @xmath74 , then set @xmath79.[alg1 ] 2 . for each @xmath78 , @xmath80.[alg2 ] 3 .
update @xmath49 and @xmath51 .
note that the fractional algorithm starts with all weights equal to zero .
this is necessary , since the online algorithm must reject @xmath81 requests in case the optimal solution rejects @xmath81 requests .
hence , the algorithm is competitive for @xmath82 , and from now on we assume without loss of generality that @xmath83 . in the following
we analyze the performance of the algorithm .
[ frac_steps ] the total number of weight augmentations performed during the algorithm is at most @xmath84 .
consider the following potential function : @xmath85 where @xmath86 is the weight of the request @xmath18 in the optimal fractional solution .
we now show three properties of @xmath87 : * the initial value of the potential function is : @xmath88 . * the potential function never exceeds @xmath89 . * in each weight augmentation step , the potential function is multiplied by at least @xmath57 .
the first two properties follow directly from the initial value and from the fact that no request gets a weight of more than @xmath90 . consider an iteration in which the adversary gives a request @xmath18 with cost @xmath75 .
now suppose that a weight augmentation is performed for an edge @xmath14 .
we must have @xmath91 since the optimal solution must satisfy the capacity constraint .
thus , the potential function is multiplied by at least : @xmath92 the first inequality follows since for all @xmath93 and @xmath94 , @xmath95 and the last inequality follows since @xmath96 .
it follows that the total number of weight augmentation steps is at most : @xmath97 [ frac_comp ] for the weighted case , the fractional algorithm is @xmath98-competitive . in case
all the costs are equal to @xmath40 , the algorithm is @xmath99-competitive .
the cost associated with the online algorithm is @xmath100 , which we will denote by @xmath101 .
consider a weight augmentation step performed for an edge @xmath14 . in step [ alg1 ] of the algorithm ,
the weights of at most @xmath29 requests change from @xmath81 to @xmath102 .
this is because before the current request arrived , there could have been at most @xmath2 requests containing the edge @xmath14 and having @xmath74 ( the maximum capacity is @xmath2 ) .
since the maximum cost is @xmath103 , the total increase of @xmath101 in step [ alg1 ] of the algorithm is at most @xmath104 . if follows that in step [ alg1 ] , the quantity @xmath105 can increase by at most @xmath106 .
a weight augmentation is performed as long as @xmath107 . before step [ alg2 ]
we have that @xmath108 .
thus , the total increase of @xmath101 in step [ alg2 ] of the algorithm does not exceed @xmath109 it follows that the total increase of @xmath101 in a weight augmentation step is at most @xmath110 . using lemma [ frac_steps ] which bounds the number of augmentation steps ,
we conclude that the algorithm is @xmath111-competitive . for the weighted case
, we saw that the input can be transformed so that @xmath112 , which implies that the algorithm is @xmath98-competitive . in case
all the costs are equal to @xmath40 , @xmath103 is also equal to @xmath40 and the algorithm is @xmath99-competitive .
we describe in this section an @xmath0-competitive randomized algorithm for the weighted case and a slightly better @xmath38-competitive randomized algorithm for the unweighted case .
the algorithm maintains a weight @xmath43 for each request @xmath18 , exactly like the fractional algorithm .
assume now that the algorithm receives a request @xmath18 with cost @xmath75 .
the following is performed in this case . 1 .
perform all the weight augmentations according to the fractional algorithm .
2 . reject all requests whose weight is at least @xmath113.[rand2 ] 3 .
for every request @xmath114 , if its weight @xmath115 increased by @xmath116 , then reject the request @xmath114 with probability @xmath117.[rand1 ] 4 .
if the current request @xmath18 can not be accepted ( some edge would be over capacity ) , then reject the request .
else , accept the request @xmath18.[rand3 ] we can assume that @xmath118 , the total number of requests whose paths contain a specific edge @xmath14 , is less than @xmath119 . to see this , note that the fractional algorithm normalizes the costs so that the minimum cost is @xmath40 and the maximum cost is at most @xmath120 . if @xmath121 , then since the optimal solution can accept at most @xmath2 requests from @xmath46 , it must pay a cost of at least @xmath122 for requests rejected out of @xmath46 , where @xmath123 is the total cost of these requests .
the online algorithm can reject all the requests in @xmath46 , pay @xmath123 and it will still be @xmath57-competitive with respect to the requests in @xmath46 , since @xmath124 .
[ rand_comp ] for the weighted case , the randomized algorithm is @xmath125-competitive .
denote by @xmath126 the cost of the fractional algorithm .
the expected cost of requests rejected in step [ rand1 ] of the algorithm is at most @xmath127 .
the cost of requests rejected in step [ rand2 ] has the same upper bound .
we now calculate the probability for a request @xmath114 to be rejected in step [ rand3 ] .
this can happen only if the path of request @xmath114 contains an edge @xmath14 for which @xmath128 but the randomized algorithm rejected less than @xmath51 requests whose paths contain the edge @xmath14 .
all the requests with weight at least @xmath113 are rejected for sure , so we can assume that @xmath129 for all @xmath130 .
suppose that @xmath78 and that during all runs of step [ rand1 ] of the algorithm the request @xmath18 has been rejected with probabilities @xmath131 , where @xmath132 .
the probability that @xmath18 will be rejected is at least @xmath133 the last inequality follows since for all @xmath134 , @xmath135 .
the number of requests in @xmath49 which were rejected by the algorithm is a random variable whose value is the sum of mutually independent @xmath136-valued random variables and its expectation is at least @xmath137 . by chernoff bound ( c.f .
, e.g. , @xcite ) , the probability for this random variable to get a value less than @xmath138 is at most @xmath139 for every @xmath140 .
therefore , the probability to be less than @xmath51 is at most @xmath141 the request costs were normalized , so that the maximum cost is at most @xmath120 .
each edge is contained in the paths of at most @xmath119 requests .
therefore , the expected cost of requests which are rejected in step [ rand3 ] because of this edge is at most @xmath142 .
thus , the total expected cost of requests rejected in step [ rand3 ] is @xmath143 .
the result now follows from theorem [ frac_comp ] .
for the unweighted case we slightly change the algorithm as follows . in step [ rand1 ] of the algorithm we reject a request with probability @xmath144 , and in step [ rand2 ] we reject all the requests whose weight is at least @xmath145 .
[ unweighted_rand_comp ] for the unweighted case , the randomized algorithm is @xmath146-competitive .
following the proof of theorem [ rand_comp ] , we get that the probability for an edge to cause a specific request to be rejected in step [ rand3 ] of the randomized algorithm is at most @xmath147 denote by @xmath148 the quantity @xmath149 .
hence , @xmath148 is the maximum excess capacity in the network .
the total expected cost of requests rejected in step [ rand3 ] is at most @xmath150 .
it is obvious that the optimal solution must reject at least @xmath148 requests .
the result now follows from theorem [ frac_comp ] .
we now describe the reduction between online set cover and admission control .
suppose we are given the following input to the online set cover with repetitions problem : @xmath30 is a ground set of @xmath4 elements and @xmath31 is a family of @xmath1 subsets of @xmath30 , with a positive cost @xmath151 associated with each @xmath33 .
the instance of the admission control to minimize rejections problem is constructed as follows : the graph @xmath12 has an edge @xmath152 for each element @xmath153 .
the capacity of the edge @xmath152 is defined to be the number of sets that contain the element @xmath154 .
the maximum capacity is therefore at most @xmath1 .
the requests are given to the admission control algorithm in two phases .
in the first phase , before any element is given to the online set cover algorithm , we generate @xmath1 requests to the admission control online algorithm . for every @xmath33 ,
the request consists of all the edges @xmath152 such that @xmath155 .
the online algorithm can accept all the requests and this will cause the edges to reach their full capacity . in the second phase , each time the adversary gives an element @xmath154 to the online set cover algorithm , we generate a request which consists of the one edge @xmath152 and give it to the admission control algorithm . in case
the request caused the edge @xmath152 to be over capacity , the algorithm will have to reject one request in order to keep the capacity constraint . in case there is a feasible cover for the input given to the online set cover problem , there is no reason for the admission control algorithm to reject requests that were given in the second phase .
this is because requests in the second phase consist of only one edge .
thus , we can assume that the admission control algorithm rejects only requests given in the first phase , which correspond to subsets of @xmath30 .
it is easy to see that the requests rejected by the admission control algorithm correspond to a legal set cover .
we reduced an online set cover problem with @xmath4 elements and @xmath1 sets to an admission control problem with @xmath4 edges and maximum capacity at most @xmath1 .
the fact that the requests we generated are not simple paths in the graph can be easily fixed by adding extra edges .
in this section we describe , given any constant @xmath8 , an @xmath5-competitive deterministic bicriteria algorithm that covers each element by at least @xmath10 sets , where @xmath11 is the number of times the element has been requested , whereas the optimum covers it @xmath11 times .
we assume for simplicity that all the sets have cost equal to 1
. the result can be easily generalized for the weighted case using techniques from @xcite .
the algorithm maintains a weight @xmath156 for each @xmath33 .
initially @xmath157 for each @xmath33 .
the weight of each element @xmath153 is defined as @xmath158 , where @xmath159 denotes the collection of sets containing element @xmath154 .
initially , the algorithm starts with the empty cover @xmath160 . for each @xmath153
, we define @xmath161 , which is the number of times element @xmath154 is covered so far .
the following potential function is used throughout the algorithm : @xmath162 we give a high level description of a single iteration of the algorithm in which the adversary gives an element @xmath154 and the algorithm chooses sets that cover it .
we denote by @xmath11 the number of times that the element @xmath154 has been requested so far . 1 .
if @xmath163 , then do nothing .
2 . else , while @xmath164 , perform a _ weight augmentation _ : 1 . for each @xmath165 , @xmath166.[detalg1 ] 2 . add to @xmath167 all the subsets for which @xmath168.[detalg2 ] 3 .
choose from @xmath159 at most @xmath169 sets to @xmath167[detalg3 ] so that the value of the potential function @xmath87 does not exceed its value before the weight augmentation . in the following
we analyze the performance of the algorithm and explain which sets to add to the cover @xmath167 in step [ detalg3 ] of the algorithm .
the cost of the optimal solution @xmath170 is denoted by @xmath56 .
[ det_steps ] the total number of weight augmentations performed during the algorithm is at most @xmath171 .
consider the following potential function : @xmath172 we now show three properties of @xmath173 : * the initial value of the potential function is : @xmath174 .
* the potential function never exceeds @xmath175 . * in each weight augmentation step , the potential function is multiplied by at least @xmath176 .
the first two properties follow directly from the initial value and from the fact that no request gets a weight of more than @xmath177 .
consider an iteration in which the adversary gives an element @xmath154 for the @xmath11th time .
now suppose that a weight augmentation is performed for element @xmath154 .
we must have that @xmath164 , which means that the online algorithm has covered element @xmath154 less than @xmath10 times .
the optimal solution @xmath21 covers element @xmath154 at least @xmath11 times , which means that there are at least @xmath178 subsets of @xmath21 containing @xmath154 which were not chosen yet .
thus , in step [ detalg1 ] of the algorithm the potential function is multiplied by at least : @xmath179 it follows that for fixed @xmath180 the total number of weight augmentation steps is at most : @xmath181 [ det_poten ] consider an iteration in which a weight augmentation is performed .
let @xmath182 and @xmath183 be the values of the potential function @xmath87 before and after the iteration , respectively .
then , there exist at most @xmath169 sets that can be added to @xmath167 during the iteration such that @xmath184 .
furthermore , the value of the potential function never exceeds @xmath185 .
the proof is by induction on the iterations of the algorithm .
initially , the value of the potential function @xmath87 is less than @xmath186 .
suppose that in the iteration the adversary gives element @xmath154 for the @xmath11th time . for each set @xmath187 ,
let @xmath188 and @xmath189 denote the weight of @xmath190 before and after the iteration , respectively .
define @xmath191 . by the induction hypothesis , we know that @xmath192 , because otherwise @xmath182 would have been greater than @xmath185 .
thus , @xmath193 .
this means that @xmath194 .
we now explain which sets from @xmath159 are added to @xmath167 .
repeat @xmath169 times : choose at most one set from @xmath159 such that each set @xmath187 is chosen with probability @xmath195 .
( this can be implemented by choosing a number uniformly at random in [ 0,1 ] , since @xmath196 . )
consider an element @xmath197 .
let the weight of @xmath198 before the iteration be @xmath199 and let the weight after the iteration be @xmath200 .
element @xmath198 contributes before the iteration to the potential function the value @xmath201 . in each random choice , the probability that we do not choose a set containing element @xmath198 is @xmath202 .
the probability that this happens in all the @xmath169 random choices is therefore @xmath203 .
note that @xmath204 . in case
we choose a set containing element @xmath198 , then @xmath205 will increase by at least @xmath40 and the contribution of element @xmath198 to the potential function will be at most @xmath206 .
therefore , the expected contribution of element @xmath198 to the potential function after the iteration is at most @xmath207 where to justify the last inequality , we prove that @xmath208 for every @xmath209 . to show this we note that @xmath210 and @xmath211 .
this implies that @xmath212 for every @xmath209 .
we can conclude that @xmath213 for every @xmath209 , as needed . by linearity of expectation
it follows that @xmath214 \leq \phi_s$ ] .
hence , there exists a choice of at most @xmath169 sets such that @xmath184 .
the choices of the various sets @xmath190 to be added to @xmath167 can be done deterministically and efficiently , by the method of conditional probabilities , c.f .
, e.g. , @xcite , chapter 15 .
after each weight augmentation , we can greedily add sets to @xmath167 one by one , making sure that the potential function will decrease as much as possible after each such choice .
[ det_comp ] the deterministic algorithm for online set cover is @xmath9-competitive .
it follows from lemma [ det_steps ] that the number of iterations is at most @xmath171 . by lemma [ det_poten ] , in each iteration we choose at most @xmath169 sets to @xmath167 in step [ detalg3 ] of the algorithm .
the sets chosen is step [ detalg2 ] of the algorithm are those which have weight at least @xmath40 .
the sum of weights of all the sets is initially @xmath62 and it increases by at most @xmath62 in each weight augmentation .
this means that at the end of the algorithm , there can be only @xmath171 sets whose weight is at least @xmath40 .
therefore , the total number of sets chosen by the algorithm is as claimed .
* an interesting open problem is to decide if the algorithm presented here for the admission control problem can be derandomized . * recently , feige and korman established a lower bound of @xmath6 for the competitive ratio of any randomized polynomial time algorithm for the online set cover problem , under the @xmath7 assumption .
it is interesting to decide whether this lower bound applies for superpolynomial time algorithms as well . *
the algorithms we gave for the admission control problem did not use the fact that the requests are simple paths in the graph .
all the algorithms treated a request as an arbitrary subset of edges .
b. awerbuch , y. azar , a. fiat , and t. leighton .
making commitments in the face of uncertainty : how to pick a winner almost every time . in _ proc .
28th acm symp . on theory of computing _ , pages 519530 , 1996 .
b. awerbuch , r. gawlick , t. leighton , and y. rabani . on - line
admission control and circuit routing for high performance computation and communication . in _ proc .
35th ieee symp . on found . of comp . science _ , pages 412423 , 1994 .
y. azar , a. blum , and y. mansour .
combining online algorithms for rejection and acceptance . in _ proceedings of the 15th annual acm symposium on parallelism in algorithms and architectures _
, pages 159163 , 2003 .
d. bunde and y. mansour .
improved combination of online algorithms for acceptance and rejection . in _ proceedings of the 16th annual acm symposium on parallelism in algorithms and architectures _ ,
pages 265266 , 2004 . | we study the admission control problem in general networks .
communication requests arrive over time , and the online algorithm accepts or rejects each request while maintaining the capacity limitations of the network . the admission control problem has been usually analyzed as a benefit problem , where the goal is to devise an online algorithm that accepts the maximum number of requests possible .
the problem with this objective function is that even algorithms with optimal competitive ratios may reject almost all of the requests , when it would have been possible to reject only a few .
this could be inappropriate for settings in which rejections are intended to be rare events . in this paper
, we consider preemptive online algorithms whose goal is to minimize the number of rejected requests .
each request arrives together with the path it should be routed on .
we show an @xmath0-competitive randomized algorithm for the weighted case , where @xmath1 is the number of edges in the graph and @xmath2 is the maximum edge capacity . for the unweighted case
, we give an @xmath3-competitive randomized algorithm .
this settles an open question of blum , kalai and kleinberg raised in @xcite .
we note that allowing preemption and handling requests with given paths are essential for avoiding trivial lower bounds .
the admission control problem is a generalization of the online set cover with repetitions problem , whose input is a family of @xmath1 subsets of a ground set of @xmath4 elements .
elements of the ground set are given to the online algorithm one by one , possibly requesting each element a multiple number of times .
( if each element arrives at most once , this corresponds to the online set cover problem . )
the algorithm must cover each element by different subsets , according to the number of times it has been requested .
we give an @xmath5-competitive randomized algorithm for the online set cover with repetitions problem .
this matches a recent lower bound of @xmath6 given by feige and korman for the competitive ratio of any randomized _ polynomial _ time algorithm , under the @xmath7 assumption .
given any constant @xmath8 , an @xmath9-competitive deterministic bicriteria algorithm is shown that covers each element by at least @xmath10 sets , where @xmath11 is the number of times the element is covered by the optimal solution .
[ routing protocols ] |
suppose that a compact lie group acts on a compact manifold @xmath12 .
the equivariant cohomology ring @xmath13 , with coefficients in a ring @xmath14 , encodes topological information about the manifold and the action . in the case of a hamiltonian action on a symplectic manifold , a variety of techniques has made computing @xmath15 tractable .
the work of goresky - kottwitz - macpherson @xcite describes this ring combinatorially when @xmath16 is a torus , @xmath14 - a field , and the action has very specific form .
we give a more general description that has a similar flavor . a theorem of kirwan @xcite states that the inclusion of the fixed points induces an injective map in equivariant cohomology .
[ kirwaninjectivity ] let torus @xmath17 act on a symplectic manifold @xmath1 in a hamiltonian fashion and let @xmath18 denote the natural inclusion of fixed points into manifold .
then the induced map @xmath19 is injective .
+ if @xmath20 are isolated points then also @xmath21 is injective .
if there are @xmath2 fixed points then @xmath22 $ ] , where @xmath23 is the dimension of the torus . therefore we can think about equivariant cohomology classes in @xmath24 as a @xmath2-tuple of polynomials @xmath9 , with each @xmath25 in @xmath26 $ ] .
the goal of this paper is to give necessary and sufficient conditions for a @xmath2-tuple of polynomials to be in the image of @xmath11 , that is to represent an equiviariant cohomology class on @xmath12 .
the following result of chang and skjelbred @xcite guarantees that we only need to consider the case of an @xmath27 action .
[ cslemma ] the image of @xmath28 is the set @xmath29 where the intersection in @xmath30 is taken over all codimension - one subtori @xmath31 of @xmath17 , and @xmath32 is the inclusion of @xmath20 into @xmath33 .
therefore we will consider a circle acting on a compact symplectic manifold @xmath1 in a hamiltonian fashion with isolated fixed points and moment map @xmath34 .
it turns out that with these assumptions we are in the morse theory setting .
[ mmapmorse ] let a circle act on a symplectic manifold @xmath12 with moment map @xmath35 and isolated fixed points . then the moment map @xmath36 is a perfect morse function on @xmath12 ( for both ordinary and equivariant cohomology ) .
the critical points of @xmath36 are the fixed points of @xmath12 , and the index of a critical point @xmath3 is precisely twice the number of negative weights of the circle action on @xmath37 .
the action of torus of higher dimension also carries a morse function . for @xmath38
we define @xmath39 , the component of moment map along @xmath40 , by @xmath41 .
we call @xmath38 * generic * if @xmath42 for each weight @xmath43 of @xmath17 action on @xmath44 , for every @xmath3 in the fixed set @xmath20 .
for such a generic @xmath40 , @xmath45 is a morse function with critical set @xmath20 .
this map is a moment map for the action of a subcircle @xmath46 generated by @xmath38 .
using morse theory , kirwan constructed equivariant cohomology classes that form a basis for integral equivariant cohomology ring of @xmath12 .
then the existence of a basis for rational equivariant cohomology ring of @xmath12 follows .
we quote this theorem with the integral coeficients , and action of @xmath17 , although in this paper we work mostly with rational coefficients and circle actions . in section [ over integers ]
we describe possible generalizations of our work to integer coefficients .
[ kirwanclasses ] let a torus @xmath17 act on a symplectic compact oriented manifold @xmath12 with isolated fixed points , and let @xmath47 be a component of moment map @xmath48 along generic @xmath38 .
let @xmath3 be any fixed point of index @xmath49 and let @xmath50 be the negative weights of the @xmath17 action on @xmath37
. then there exists a class @xmath51 such that * @xmath52 ; * @xmath53 for all fixed points @xmath54 such that @xmath55 .
moreover , taken together over all fixed points , these classes are a basis for the cohomology @xmath56 as an @xmath57 module .
we will call the above classes * kirwan classes*. these classes may be not unique . goldin and tolman give a different basis for the cohomology ring @xmath58 in [ gt ] .
they require @xmath53 for all fixed points @xmath59 of index less then or equal @xmath49 .
goldin and tolman s classes , if they exist , are unique
. therefore they are called * canonical classes*. for our purposes , it is enough to have some basis for the rational equivariant cohomology ring with respect to circle action , and with the following property * elements of the basis are in such a bijection with the fixed points that a class corresponding to a fixed point of index @xmath49 evaluated at any fixed point is 0 or a homogeneous polynomial of degree @xmath60 .
we will call elements of a basis satisfying condition ( * ) * generating classes*. kirwan classes and goldin - tolman cannonical classes satisfy the above condition .
+ another important ingredient of our proof is the atiyah - bott , berline - vergne ( abbv ) localization theorem . for a fixed point @xmath3 let @xmath61 be the equivariant euler class of tangent bundle @xmath37 , which in this case is equal to the product of weights of the torus action .
[ abbv ] let @xmath12 be a compact manifold equipped with an @xmath27 action with isolated fixed points , and let @xmath62 .
then as elements of @xmath63 $ ] , @xmath64 where the sum is taken over all fixed points . in this paper
, we show how to obtain relations describing the image of @xmath65 by applying the localization theorem .
our main theorem is : [ main ] let a circle act on a manifold @xmath12 in a hamiltonian fashion with isolated fixed points @xmath66 , and let @xmath67=h^*_{s^1}(m^{s^1 } ; { \mathbb{q}})$ ] .
suppose we are given @xmath5 -a basis of @xmath68 satisfying condition ( * ) .
then @xmath69 is an equivariant cohomology class on @xmath12 if and only if for every fixed point @xmath3 of index @xmath49 , @xmath70 have @xmath71,\ ] ] where by @xmath72 we mean @xmath73 , with @xmath74 inclusion of the fixed point @xmath75 into @xmath12 .
note that if @xmath3 is a fixed point of index @xmath23 , this condition is automatically satisfied .
this is because @xmath76 is nonzero only at @xmath3 , and there its value is the euler class @xmath61 .
therefore it is sufficient to check the above condition only for points of index @xmath49 , @xmath70 . using the localization theorem
, we easily see that if @xmath69 is cohomology class , then these conditions must be satisfied .
the interesting part of the theorem is that they are sufficient to describe @xmath77 as a subring of @xmath78 .
consider the standard hamiltonian @xmath27 action on @xmath79 by rotation with weight @xmath80 .
the isolated fixed points are south and north poles which we will denote by @xmath81 and @xmath82 respectively .
kirwan class associated to @xmath81 is @xmath83 .
the theorem then says that @xmath84 represents equivariant cohomology class if and only if @xmath85.\ ] ] this observation , together with the chang - skjelbred lemma , recovers the gkm theorem .
theorem [ main ] is useful only if we know the restrictions to the fixed points of a set of generating classes ( whose existence is guaranteed by theorem [ kirwanclasses ] ) .
it is not surprising that there is a translation from the values of generating classes at fixed points to relations defining @xmath86 .
our translation provides a particularly combinatorial description that is easy to apply in examples .
although we can not compute these classes in general , there are algorithms that work for a wide class of spaces , for example gkm spaces , including symplectic toric manifolds and flag manifolds ( see[t ] ) . for the sake of completeness we will describe an algorithm for obtaining kirwan classes for symplectic toric manifolds in section [ toric ]
. the choice of @xmath76 assigned to fixed point @xmath3 may be not unique , even for toric varieties . choosing a generic subcircle of @xmath17
, we get moment map for this circle action that is a morse function and therefore we can talk about flow up from the fixed point @xmath3 .
if for every fixed point @xmath87 in the flow up at @xmath3 the index of @xmath87 is strictly greater then the index of @xmath3 , then the choice of @xmath76 is unique ( see [ gz ] ) .
we then say that the moment map is index increasing .
+ a particularly interesting application of our theorem is when we want to restrict the action of @xmath17 to an action of a subtorus @xmath88 such that @xmath89 , and compute @xmath90 .
we call this process * specialization * to the action of subtorus @xmath91 . having generating classes for @xmath17 action we can easily compute generating classes for @xmath91 action using the projection @xmath92
. theorem [ main ] gives relations that cut out @xmath93 .
in particular we can use this method to restrict the torus action on a symplectic toric manifold to a generic circle , i.e. such a circle @xmath91 for which @xmath89 ( see example 1a and b ) .
a priori we only require that @xmath94 is finite , as we still want to describe @xmath95 by analyzing the relations on polynomials defining the image @xmath96 $ ] .
however it turns out that this requirement implies @xmath97 .
we can explain this fact using morse theory .
if @xmath98 is a moment map for @xmath17 action and @xmath38 is generic , then @xmath99 - component of @xmath48 along @xmath40 , is a perfect morse function with critical set @xmath20 .
therefore @xmath100 . similarly , taking @xmath101 - the moment map for @xmath91 action , and any generic @xmath102 , we obtainn @xmath103 which is also a perfect morse function for @xmath12 . thus @xmath104 .
as obviously @xmath105 , the sets must actually be equal .
+ consider restriction of the gkm action of @xmath17 to a generic subcircle @xmath91 : [ fig : gkmrelations2 ] gkm relations are sufficient to describe the image of @xmath106 in @xmath107 , but their `` projections '' are not sufficient to describe the image of @xmath108 in @xmath109 .
however projecting generating classes and using main theorem to construct relations from such a basis will give all the relations we need .
+ the gkm theorem is a very powerful tool that allows us to compute the image of @xmath110 .
however this theorem can not be applied if for some codimension @xmath83 subtorus @xmath111 we have dim @xmath112 . goldin and holm in [ gh ]
provide a generalization of this result to the case where dim @xmath113 for all codimension @xmath83 subtori @xmath111 .
an important corollary is that , in the case of hamiltonian circle actions , with isolated fixed points , on manifolds of dimension @xmath114 or @xmath115 , rational equivariant cohomology ring can be computed solely from the weights of the circle action at the fixed points . in dimension
@xmath114 this is given for example by the gkm theorem .
in dimension @xmath115 one can apply the algorithm presented by goldin and holm in @xcite or use the fact that any such action is actually a specialization of a toric action ( see @xcite ) . if one wishes to compute integral equivariant cohomology ring , one will need additional piece of information , so called isotropy skeleton
( @xcite ) .
godinho in @xcite presents such an algorithm .
information encoded in the isotropy skeleton is essential .
there can not exist an algorithm computing the integral equivariant cohomology only from the fixed points data .
karshon in @xcite(example 1 ) , constructs two @xmath27 spaces with the same weights at the fixed points but different integral equivariant cohomology ring .
this suggests that we should not hope for an algorithm computing the rational equivariant cohomology ring from the weights at the fixed points for manifolds of dimension greater then @xmath115 .
more information is needed .
tolman and weitsman used generating classes to compute equivariant cohomology ring in case of semifree action in [ tw ] .
their work gave us the idea for constructing necessary relations described in the present paper using information from generating classes .
our proof was also motivated by the work of goldin and holm @xcite where the localization theorem and dimensional reasoning were used . + * organization*. in section [ proof ] , we prove our main result . in section [ toric ]
, we describe the case of toric symplectic manifolds and construct generating classes for their equivariant cohomology .
we present several examples in section [ examples ] .
we finish our work with possible generalization to integer coefficients in section [ over integers ] . +
* acknowledgments . *
the author is grateful to tara holm for suggesting this problem and for helpful conversations .
let a circle act on a manifold @xmath12 in a hamiltonian way with isolated fixed points @xmath66 and let @xmath67=h^*_{s^1}(m^{s^1})$ ] .
we want to show that @xmath69 is an equivariant cohomology class of @xmath12 if and only if for every fixed point @xmath3 have @xmath116,\ ] ] where @xmath76 denotes generating class assigned to @xmath3 .
the moment map is a morse function .
therefore idex of a fixed point is well defined .
we first compare the equivariant poincar polynomials of @xmath12 and @xmath117 to determine the minimum number of relations we will need . let @xmath118 be the number of fixed points of index @xmath49 . then @xmath119 is the total number of fixed points . using theorem 1.3 of frankel and kirwan , we know that @xmath118 is also the @xmath49-th betti number of @xmath12 .
hamiltonian @xmath27 spaces are always equivariantly formal , that is @xmath120 as modules .
thus the equivariant poincar polynomial for @xmath12 is @xmath121 @xmath122 for the @xmath117-set of fixed points , we have + @xmath123 + therefore we need precisely @xmath124 relations of degree @xmath60 , for each @xmath125 . using poincar duality we see that the number above is equal to @xmath126 if we introduce the notation @xmath127 , then relations of type @xmath128 for some constants @xmath129 s are called relations of degree @xmath60 . for any fixed point @xmath3 of index @xmath130 , its canonical class assigns to each fixed point @xmath131 or a homogeneous polynomial of degree @xmath132 .
denote by @xmath133 rational number satisfying @xmath134 .
if @xmath69 is an equivariant cohomology class of @xmath12 then @xmath135 is also .
the localization theorem gives the relation @xmath116.\ ] ] we may rewrite it in the following form : @xmath136 \end{aligned}\ ] ] the coefficients of negative powers of @xmath137 need to be @xmath131 .
therefore for any fixed point @xmath3 and any @xmath138 we get the following relation of degree @xmath139 : @xmath140 note that in any degree , these relations are independent .
we will show this by explicit computation .
suppose that in some degree @xmath139 these relations in @xmath141 s are not independent .
that is , there are rational numbers @xmath142 , not all zero , such that @xmath143 for all possible @xmath141 .
thus for all @xmath144 , we have @xmath145 . multiplying both sides by @xmath146 we obtain @xmath147
recall the definition of @xmath148 to notice that the above equation is equivalent to @xmath149 that means @xmath150 vanishes on every fixed point and therefore is the @xmath131 class , although it is a nontrivial combination of classes @xmath76 . here
we get contradiction with the independence of the @xmath76 s . + to count the relations we have just constructed ,
notice that a relation of degree @xmath151 is obtained from each fixed point of index @xmath130 or less .
dually we get a relation of degree @xmath60 for each fixed point of index @xmath152 or less .
counting relations of degree @xmath60 over contributions from all fixed points , we have found @xmath153 of them , exactly as many as we need .
this completes the proof .
let @xmath154 be a symplectic toric manifold with a hamiltonian action of @xmath155 and moment polytope @xmath156 . denote by @xmath157 the one - skeleton of @xmath12 , that is , the union of all @xmath17-orbits of dimension @xmath83 .
closures of connected components of @xmath157 are spheres , called the isotropy spheres .
denote by @xmath158 the vertices of @xmath159 , and by @xmath160 the @xmath83-dimensional faces of @xmath159 , also called edges .
vertices correspond to the fixed points of the torus action , while edges correspond to the isotropy spheres .
fix a generic @xmath161 , so that for any @xmath162 we have @xmath163 .
orient the edges so that @xmath164 for any edge @xmath165 , where @xmath166 are initial and terminal points of @xmath165 . for any @xmath167 denote by @xmath168 the primitive integral vector in direction of @xmath169 .
label each edge with the primitive integral weight in the direction of the edge in @xmath170 .
this is the weight of @xmath17 action on the corresponding isotropy sphere . for any @xmath171 let @xmath172 denote the smallest face of positive dimension containing @xmath3 and only such points @xmath173 for which @xmath174 . we will call @xmath172 the flow up face for @xmath3 .
we define the class @xmath175 by @xmath176 where the product is taken over all @xmath177 such that @xmath178 and @xmath179 are connected by an edge of @xmath159 .
we use convention that empty product is @xmath83 .
such classes satisfy the gkm conditions and thus are in the image of the equivariant cohomolgy of @xmath12 .
they are poincar duals to the homology classes defined by the flow up submanifolds from fixed points ( in morse theoretic sense ) .
all fixed points are assigned a value which is @xmath131 or a homogeneous polynomial of degree equal to number of edges terminating at this vertex ( as polytope is smooth , exactly @xmath23 edges meet at each vertex ) .
these two facts can be proved using the notion of the axial function introduced in [ gz ] .
the classes we have just defined are also linearly independent , which follows easily from the fact that @xmath76 is nonzero only at vertices @xmath179 above @xmath3 in the partial order given by the orientation of edges .
our first example is a set of generating classes for @xmath180 presented in figure [ fig : cp2 ] .,title="fig:",scaledwidth=70.0% ] + + next we give an example where the generating classes are not unique .
the above algorithm gives the basis presented on figure [ fig : not unique generating classes ] .
however classes in figure [ fig : differentbasis ] also form a basis .
+ + + notice that this algorithm can be used to compute equivariant cohomology even if we do not have the weights for the full @xmath0 action .
it is enough to know the weights of @xmath91 action ( which are just projections of @xmath17 weights under @xmath181 ) , the fact that this action is a specialization of some toric action and positions of isotropy spheres for that toric action . with this information
, we can find the `` flow up face @xmath172 '' for any fixed point @xmath3 .
the above algorithm gives that @xmath182 where the product is taken over all @xmath177 such that @xmath178 and @xmath179 are connected by an isotropy sphere .
having generating classes for @xmath91 action , we may apply the main theorem to obtain all relations needed to describe @xmath183 .
this gives us a method for computing equivariant cohomology for cirlce actions that happen to be part of a toric action .
+ we expect that also for hamiltonian circle actions with isolated fixed points that are part of a gkm ation , the equivariant cohomology ring can be computed from the weights at fixed points and positions of isotropy spheres for this gkm action .
@xmath184 consider the product of @xmath180 blown up at a point and @xmath185 @xmath186[y_0:y_1:y_2][z_0:z_1])|\,x_1\,y_2-x_2\,y_1=0\},\ ] ] and the following @xmath187 action on this space : @xmath188[y_0:y_1:y_2][z_0:z_1])=([e^{iu}x_1:x_2][e^{iv}y_0:e^{iu}y_1:y_2][e^{iw}z_0:z_1]).\ ] ] this is a symplectic toric manifold with moment map @xmath189[y_0:y_1:y_2][z_0:z_1])= \left ( \frac{|x_1|^2}{||x||^2}+\frac{|y_1|^2}{||y||^2},\frac{|y_0|^2}{||y||^2 } , \frac{|z_0|^2}{||z||^2 } \right)\ ] ] where @xmath190 , and similarly for @xmath191 and @xmath192 .
the moment polytope is shown in figure [ fig : momentpolyblowup ] .
[ fig : momentpolyblowup ] .,title="fig:",scaledwidth=30.0% ] using the algorithm from section [ toric ] we can compute generating classes for the equivariant cohomology with respect to @xmath17 action .
they are presented in the table below .
+ [ cols="^,^,^,^,^,^,^,^,^,^",options="header " , ] * the degree @xmath114 relation : + @xmath193.$ ]
in this section we analyze which parts of the paper would work over integer coefficients . in the case of isolated fixed points ,
injectivity theorem of kirwan [ kirwaninjectivity ] holds over @xmath194 the kirwan basis constructed in [ kirwanclasses ] is a basis for integral equivariant cohomology ring . the main tool in our proof , that is the localization theorem , also holds for integral coefficients ( @xcite ) . our method of transferring the information about the basis into explicit relations works over any ring .
therefore the main theorem holds also for integral coefficients .
there we were considering only circle actions .
the analysis of the action of torus of higher dimension could be simplified to analyzing only subcirles actions thanks to chang skjelbred lemma .
this lemma does not hold over @xmath195 in general .
tolman and weitsman in [ tw2 ] , section 4 , construct explicit example where this lemma fails . in this example
@xmath196 acts on @xmath197 , rotating each sphere with speed @xmath114 .
however , under special assumptions , matthias franz and volker puppe in @xcite proved integral chang skjelbred lemma .
[ chang - skjelbred - integers ] let @xmath198 be a @xmath17-space such that @xmath199 is free over @xmath200 , where the cohomology is with integral coefficients . denote by @xmath201 the @xmath202-skeleton of @xmath198 , that is the union of all orbits of dimension @xmath203
. then @xmath204 is exact .
if in addition the isotropy group of each @xmath205 is contained in a proper subtorus of @xmath17 then the following sequence is exact : @xmath206 the injectivity of the map @xmath207 follows from the localization theorem . in the tolman - weitsman example mentioned above the `` proper subtorus assumption ''
was not satisfied , as in their example all orbits of dimension @xmath114 have isotropy group @xmath208
. + 9999 m. atiyah and r. bott . _ the moment map and equivariant cohomology .
_ topology , * 23 * ( 1984 ) , 128 . c. allday and v. puppe , _ cohomological methods in transformation groups , _ cambridge university press , 1993 n. berline and m. vergne . _
classes caractristiques quivariantes .
formules de localisation en cohomologie quivariante .
paris sr .
, * 295 * ( 1982 ) , 539541 . t. chang and t. skjelbred , _ the topological schur lemma and related results _ , annals of mathematics 100(1974 ) , 307 - 321 .
t. frankel , _ fixed points on kahler manifolds _ , annals of mathematics , vol . 70 , no
. 1 , jul . , 1959 , pages 1 - 8 .
m. goresky , r. kottwitz and r. macpherson , _ equivariant cohomology , koszul duality , and the localization theorems _ , invent .
math 131 ( 1998)25 - 83 .
r. goldin and t. holm , _ the equivariant cohomology of hamiltonian g - spaces from residual @xmath27 actions _ , mathematical research letters 8 , 67 - 77 ( 2001 ) .
_ equivariant cohomology of @xmath27-actions on 4-manifolds _ , canad .
math . bull .
50 ( 2007 ) , 365 - 376 .
r. goldin and s. tolman _ towards generalizing schubert calculus in the symplectic category _
, journal of symplectic geometry volume 7 , number 4 ( 2009 ) , 449 - 473 . v. guillemin and c. zara _ the existence of generating families for the cohomology ring of a graph _ , advances in mathematics 174 ( 2003 ) 115 - 153 .
f.c.kirwan , _ the cohomology of quotients in symplectic and algebraic geometry _ , princeton university press , 1984 .
y. karshon , _ hamiltonian torus actions _ , geometry and physics , lecture notes in pure and applied mathematics series 184 , marcel dekker , 1996 , p.221 - 230 y. karshon _ periodic hamiltonian flows on four dimensional manifolds _ , memoirs amer . math
672 , 1999 j. tymoszko , _ an introduction to equivariant cohomology and homology , following goresky , kottwitz , and macpherson _ , snowbird lectures in algebraic geometry , 169 - 188 , contemp .
388 , amer .
soc . , providence , ri , 2005 .
available at arxiv : math/0503369 .
s. tolman and j. weitsman , _ on the semifree symplectic circle actions with isolated fixed points _ ,
topology 39 ( 2000 ) 299 - 309 .
s. tolman and j. weitsman , _ on the cohomology rings of hamiltonian t - spaces _ , pp.251 - 258 in : northen california symplectic geometry seminar , transl . ,
ser 2 196 ( 45 ) , am .
soc . , providence , ri 1999 .
s. tolman and j. weitsman , _ the cohomology rings of abelian symplectic quotients _ , communications in analysis and geometry , volume 11 , number 4 , 751 - 773 .
matthias franz and volker puppe , _ exact sequences for equivariantly formal spaces _ , http://arxiv.org/abs/math/0307112 | we consider a hamiltonian @xmath0 action on a compact symplectic manifold @xmath1 with @xmath2 isolated fixed points .
for every fixed point @xmath3 there exists a class @xmath4 such that the collection @xmath5 , for all fixed points , forms a basis for @xmath6 as an @xmath7 module .
the map induced by inclusion , @xmath8 $ ] is injective .
we will use such classes @xmath5 to give necessary and sufficient conditions for @xmath9 in @xmath10 $ ] to be in the image of @xmath11 , i.e. to represent an equiviariant cohomology class on @xmath12 .
we may recover the gkm theorem when the one skeleton is 2-dimensional .
moreover , our techniques give combinatorial description when we restrict to a smaller torus , even though we are then no longer in gkm case .
+ [ multiblock footnote omitted ] |
white dwarfs located in a binary system may end their lives in two possible ways .
they may accrete from a companion and achieve the chandrasekhar mass thereby triggering a thermonuclear runaway of the object and ultimately exploding as a type ia supernova . in this case
no remnant is left behind .
alternatively these massive white dwarfs , for sufficiently high mass accretion rates , may allow the formation of o+ne+mg cores and due to the high prevailing central density ( beyond @xmath2 ) experience rapid electron capture that lead to the collapse of the core [ 1 ] .
this is termed as accretion - induced collapse ( aic ) .
the end product is a neutron star ( in some double - degenerate scenario , two white dwarfs in a short - period binary system may eventually coalesce to form a massive white dwarf that exceeds the chandrasekhar mass limit and in this case transition to a black hole is possible if the total proto - neutron star mass exceeds the general - relativistic limit for gravitational stability [ 2 ] ) .
the ultimate fate of these white dwarfs are dependent on many factors , e.g. temperature of the environment , the mass accretion rate on the newly formed white dwarf , and the mass of each partner white dwarf [ 2 ] .
the occurrence rate of the aic of white dwarfs is not determined reliably , and are not expected to occur more than once per 20 50 standard typa ia events ( see for example [ 2 ] ) .
supernova explosions from the collapse of o+ne+mg cores remain a potential research interest for astrophysical reasons .
its smaller core and smaller gravitational potential may allow the star to explode hydrodynamically .
they are also proposed as a probable candidate for the site of r - process as compared to the `` neutrino wind '' mechanism based on the condition that they may explode by the prompt bounce - shock mechanism [ 2 - 5 ] .
a detailed analysis of the core evolution of massive stars may be found in [ 1 ] .
the evolution of the stars in the mass range 8 10 m@xmath0 develops central cores which are composed of @xmath3o , @xmath4ne and @xmath1 mg .
oda et al . [ 6 ] employed shell model wave functions of the @xmath5-shell nuclei developed by wildenthal [ 7 ] and calculated the capture rates which contribute to the collapse of the o+ne+mg core .
oda et al . [ 6 ] pointed out three different series of electron capture in the o+ne+mg core of the 8 10 m@xmath0 stars and placed them in the order of low threshold energy as @xmath1 mg @xmath6 @xmath1na @xmath6 @xmath1ne , @xmath4ne @xmath6 @xmath4f
@xmath6 @xmath4o , and @xmath3o @xmath6 @xmath3n @xmath6 @xmath3c .
they , however , renounced the last series in their calculations because of its high threshold energy which did not contribute significantly to the initiation of the collapse of the o+ne+mg core of the 8 10 m@xmath0 stars .
a reasonably large number of recent observations and accumulation of data regarding type ii - p supernovae lead to a rejuvenation of debate on the fate of o+ne+mg cores .
earlier gutirrez et al .
[ 8 ] argued that the abundance of @xmath1 mg was considerably reduced in updated evolutionary calculations .
however the procedure adopted by them was not fully consistent as they kept the ratio of oxygen to neon constant while parameterizing the abundance of @xmath1 mg .
later kitaura et al . [ 9 ] presented simulation result of these cores ( keeping capture rates on @xmath1 mg as a key ingredient ) using an improved neutrino transport treatment .
their outcome was not a prompt but a delayed explosion .
kitaura and collaborators [ 9 ] used the shell model capture rates of takahara et al . [ 10 ] as a key nuclear input parameter in their simulation codes in the non - nuclear statistical equilibrium regime .
one of the reasons for this spherical core collapse simulations not to explode was the reduced electron capture rates employed in their simulations .
this reduced capture rate slowed the collapse process and resulted in a large shock radius .
electron captures on proton and positron captures on neutron play a very crucial role in the supernovae dynamics . during the collapse and accretion phases , these processes exhaust electrons , thus decreasing the degenerate pressure of electrons in the stellar core .
meanwhile , they produce neutrinos which carry the binding energy away .
therefore , electron and positron captures play key role in the dynamics of the formation of bounce shock of supernova .
the type ii supernovae take place in heavy stars .
the positron captures are of great importance in high temperature and low density locations . in such conditions ,
a rather high concentration of positron can be reached from @xmath7 equilibrium favoring the @xmath8 pairs .
the electron captures on proton and positron captures on neutron are considered important ingredients in the modeling of type ii supernovae [ 11 ] .
the key purpose of reporting this work is the presentation of the newly calculated microscopic calculation of the electron and positron capture rates using the pn - qrpa theory with separable interaction . due to
the extreme conditions prevailing in the cores of o+ne+mg stars , interpolation of calculated rates within large intervals of temperature - density points posed some uncertainty in the values of capture rates for collapse simulators .
as such the calculation is done on a detailed temperature and density grid pertinent to presupernova and supernova environment and should prove more suitable for running on simulation codes .
the pn - qrpa theory is proven to be quite successful in the calculation of stellar weak rates .
earlier nabi and klapdor [ 12 ] used the pn - qrpa theory to calculate the weak interaction rates for 178 @xmath5-shell nuclei [ 12 ] and 650 fp / fpg - shell [ 13 ] nuclide in stellar matter .
the reliability of the pn - qrpa calculations was discussed in detail by nabi and klapdor [ 13 ] . there
the authors compared the measured data ( half lives and gamow - teller ( gt ) strength ) of thousands of nuclide with the pn - qrpa calculations and got fairly good comparison .
later the decay and capture rates of nuclei of astrophysical importance were studied separately in detail and were compared with earlier calculations wherever possible both in sd - shell [ 14 ] and fp - shell ( e.g. [ 15 - 17 ] ) regions .
compared to shell model calculations , the pn - qrpa gives similar accuracy in reproducing beta - decay and capture rates in @xmath5-shell nuclei [ 14 ] .
the deformation parameter is as an important parameter for qrpa calculations as pairing [ 18 ] . for the case of
even - even nuclei , experimental deformations are available [ 19 ] and were employed in this work . in this paper a detailed calculation of electron and positron capture rates on @xmath1 mg is being presented for the first time at temperature - density intervals suitable for simulation purposes .
the presence of @xmath1 mg in the core is a result of the previous phase of carbon burning and its relevance is due to its lower electron capture threshold .
section 2 briefly discusses the formalism of the pn - qrpa calculations .
section 3 presents some of the calculated results .
comparison with earlier calculations is also included in this section .
the conclusions are given in section 4 and at the end table 1 presents the detailed calculation of electron and positron capture rates on @xmath1 mg .
the hamiltonian for the calculations was chosen to be of the form @xmath9 where @xmath10 is the single - particle hamiltonian , @xmath11 is the pairing force , @xmath12 is the particle - hole ( ph ) gamow - teller force , and @xmath13 is the particle - particle ( pp ) gamow - teller force .
single particle energies and wave functions were calculated in the nilsson model , which takes into account nuclear deformations .
pairing was treated in the bcs approximation .
the proton - neutron residual interactions occurred as particle - hole and particle - particle interaction .
the interactions were given separable form and were characterized by two interaction constants @xmath14 and @xmath15 , respectively . in this work ,
the values of @xmath14 and @xmath15 was taken as 0.001 mev and 0.05 mev , respectively .
other parameters required for the calculation of weak rates are the nilsson potential parameters , the deformation , the pairing gaps , and the q - value of the reaction .
nilsson - potential parameters were taken from ref . [ 20 ] and the nilsson oscillator constant was chosen as @xmath16 ( the same for protons and neutrons ) .
the calculated half - lives depend only weakly on the values of the pairing gaps [ 21 ] .
thus , the traditional choice of @xmath17 was applied in the present work .
the deformation parameter is recently argued to be one of the most important parameters in qrpa calculations [ 18 ] and as such rather than using deformations from some theoretical mass models ( as used in earlier calculations of pn - qrpa capture rates ) the experimentally adopted value of the deformation parameters for @xmath1 mg , extracted by relating the measured energy of the first @xmath18 excited state with the quadrupole deformation , was taken from raman et al .
q - values were taken from the recent mass compilation of audi et al .
[ 22 ] .
the electron capture ( ec ) and positron capture ( pc ) rates of a transition from the @xmath19 state of the parent to the @xmath20 state of the daughter nucleus is given by @xmath21\left[f_{ij } ( t,\rho , e_{f } ) \right]\left[b(f)_{ij } + \left({\raise0.7ex\hbox{$ g_{a } $ } \!\mathord{\left/ { \vphantom { g_{a } g_{v } } } \right .
\kern-\nulldelimiterspace}\!\lower0.7ex\hbox{$ g_{v } $ } } \right)^{2 } b(gt)_{ij } \right].\ ] ] the value of d was taken to be 6295s [ 23 ] and the ratio of the axial vector to the vector coupling constant as -1.254 [ 24 ] .
@xmath22 are the sum of reduced transition probabilities of the fermi b(f ) and gt transitions b(gt ) .
the @xmath23 are the phase space integrals .
details of the calculations of phase space integrals , reduced transition probabilities and choice of gamow - teller strength parameters can be found in [ 12,13,25 ] .
the total electron ( positron ) capture rate per unit time per nucleus was then calculated using @xmath24 the summation over all initial and final states was carried out until satisfactory convergence in the rate calculations was achieved . here
@xmath25 is the probability of occupation of parent excited states and follows the normal boltzmann distribution .
it is pertinent to mention that experimental data was incorporated in the calculations wherever possible to further increase the reliability of the calculated capture rates .
in addition to the incorporation of the measured value of the deformation parameter , the calculated excitation energies ( along with their log @xmath26 values ) were replaced with an experimental one when they were within 0.5 mev of each other .
missing measured states were inserted and inverse and mirror transitions were also taken into account .
no theoretical levels were replaced with the experimental ones beyond the excitation energy for which experimental compilations have no definite spin and/or parity .
this means that no theoretical levels were replaced in the parent @xmath1 mg beyond 9.3 mev . for the daughter nuclei no replacement
was done in @xmath1na beyond 1.3 mev ( for the electron capture direction ) and in @xmath1al beyond 1.1 mev ( for the positron capture direction ) .
the gamow - teller strength distributions , b(gt@xmath27 ) , and the associated electron and positron capture rates on @xmath1 mg were calculated using the pn - qrpa theory .
quenching of the gt strength was taken into account and a standard quenching factor of 0.77 was used [ 26 ] .
the calculation was performed for a total of 132 excited states of @xmath1 mg up to excitation energies in the vicinity of 40 mev ( corresponding to first 100 excited states ) in the daughters @xmath1na and @xmath1al . the summed b(gt+ )
strength is shown in fig.1 .
it is to be noted that the experimentally extracted total b(gt+ ) strength for excitation energy region up to 7 mev ( where the density of states is still low enough to analyze single peaks ) is 1.36 [ 27 ] .
the corresponding calculated value reported in this work comes to be 2.53 .
the earlier reported pn - qrpa value was 2.65 [ 14 ] .
the differences in the bgt(+ ) value and in the calculated capture rates compared to those of ref .
[ 14 ] are attributed primarily to the choice of deformation parameter and incorporation of all experimental levels ( along with their @xmath28 values , if available ) as stated in previous section . whereas the authors in [ 14 ] incorporated a deformation parameter of 0.435 according to the mass formula of mller and nix [ 28 ] , in this work the measured value of 0.606 [ 19 ] is taken to be the deformation parameter for @xmath1 mg .
details on the sensitivity of b(gt+ ) strength as a function of deformation parameter can be found in [ 29 ] . regarding the measured value of the gt strength , it is to be noted that rakers and collaborators [ 27 ] employed a @xmath5-model space and the universal @xmath5 residual interaction [ 30,31 ] to calculate the wave functions .
the authors in [ 27 ] argued that the full model space calculation was not possible and they had to truncate the model space which subsequently lead to uncertainties and the authors had some reservations in the interpretation of their data .
the pn - qrpa value reported in this work is close to the reported shell model value of 2.1 of brown and wildenthal [ 30 ] and wildenthal [ 31 ] .
takahara et al .
[ 10 ] reported a value of 1.30 .
the total calculated gt strength in this work is 3.34 ( quenched ) .
fig.2 depicts the summed b(gt- ) strength as a function of excitation energies in the daughter nucleus @xmath1al .
the total strength comes to 3.34 ( quenched ) and the ikeda sum rule is fulfilled in the calculations .
moving on to the calculation of capture rates , fig.3 shows four panels depicting the calculated electron capture rates at selected temperature and density domain .
the upper left panel shows the electron capture rates in low - density region ( @xmath29 = 10^{0.5 } , 10^{1.5}$ ] and @xmath30 ) , the upper right in medium - low density region ( @xmath29 = 10^{3.5 } , 10^{4.5}$ ] and @xmath31 ) , the lower left in medium - high density region ( @xmath29 = 10^{6.5 } , 10^{7.5}$ ] and @xmath32 ) and finally the lower right panel depicts the calculated electron capture rates in high density region ( @xmath29 = 10^{9.5 } , 10^{10.5}$ ] and @xmath33 ) .
the capture rates are given in logarithmic scale .
it can be seen from this figure that in the low density region the capture rates , as a function of temperature , are more or less superimposed on one another .
this means that there is no appreciable change in the rates when increasing the density by an order of magnitude .
however as one moves from the medium low density region to high density region these rates start to peel off from one another .
orders of magnitude difference in rates are observed ( as a function of density ) in high density regions . for a given density
the rates increase monotonically with increasing temperatures .
one also notices that the electron capture rates coincide at t@xmath34 = 30k except for the high density region ( where t@xmath34 gives the stellar temperature in units of @xmath35k ) .
one of the channels for the energy release from the star is the neutrino emission which is mainly from the @xmath36 capture on nucleons and @xmath37 annihilation .
positron capture plays a crucial role in the dynamics of stellar core .
they play an indirect role in the reduction of degeneracy pressure of the electrons in the core .
fig.4 shows four panels depicting the calculated positron capture rates at selected temperature and density domain .
the upper left panel shows the positron capture rates in low - density region ( @xmath29 = 10^{0.5 } , 10^{1.5}$ ] and @xmath30 ) , the upper right in medium - low density region ( @xmath29 = 10^{3.5 } , 10^{4.5}$ ] and @xmath31 ) , the lower left in medium - high density region ( @xmath29 = 10^{6.5 } , 10^{7.5}$ ] and @xmath32 ) and finally the lower right panel depicts the calculated positron capture rates in high density region ( @xmath29 = 10^{9.5 } , 10^{10.5}$ ] and @xmath33 ) .
the positron capture rates are given in logarithmic scales .
it is to be noted that the positron capture rates are very slow as compared to electron capture on @xmath1 mg .
the positron capture rates enhance as temperature of the stellar core increases and decrease with increasing stellar density .
one should note the order of magnitude differences in positron capture rates as the stellar temperature increases .
it can be seen from this figure that in the low density region the positron capture rates , as a function of temperature , are more or less superimposed on one another ( similar to the case of electron capture rates ) .
one also observes that the positron capture rates are almost the same for the densities in the range @xmath38 .
however as one moves from the medium high density region to high density region these rates also start to peel off from one another .
when the densities increase beyond the above stated range a decline in the positron capture rate starts . for a given density
the rates increase monotonically with increasing temperatures .
the reported capture rates on @xmath1 mg were also compared against two earlier calculations .
fuller , fowler and newman [ 32 ] ( hereafter ffn ) compiled the experimental data and calculated electron and positron capture rates ( besides other weak - interaction mediated rates ) for the nuclei in the mass range a = 21 - 60 for an extended grid of temperature and density . the gt strength and excitation energies
were calculated using a zero - order shell model . for the discrete transitions for which the @xmath26 values were not available ffn took log @xmath26 = 5.0 .
later oda et al . [ 6 ] performed an extensive calculation of stellar weak interaction rates of @xmath5-shell nuclei in full ( sd)@xmath39-shell model space .
they used the effective interaction of wildenthal [ 7,33,34 ] and the available experimental compilations for their calculations .
fig.5 compares the reported electron capture rates against those of oda et al .
[ 6 ] and ffn [ 32 ] .
it is to be noted that the first two factors in the calculation of capture rates in eqtn .
( 2 ) are similar in all three calculations .
whereas the first factor is a constant , the second factor constitutes the calculation of phase space integrals .
the formalism for the calculation of these integrals is similar in all calculations and are dependent only on corresponding stellar temperatures , densities and fermi energies .
the main difference in the three calculations arises because of the third factor ( calculation of reduced transition probabilities ) which contains in it the in - built nuclear structure effects . in fig.5 @xmath40 ( in the upper panel )
is the ratio of the reported electron capture rates to those calculated using shell model by oda et al [ 6 ] whereas @xmath41 ( in the lower panel ) is the ratio of the reported electron capture rates to those calculated by ffn [ 32 ] . the shell model electron capture rates are usually in good agreement with the ffn rates .
the reported pn - qrpa electron capture rates are enhanced for all values of temperature and density compared to both shell model and ffn results .
this is one major difference with the earlier reported capture rates on @xmath1 mg [ 14 ] . for densities around @xmath29
= 10^{3}$ ] , the pn - qrpa rates are enhanced by as much as factor of 14 or more as compared to the corresponding shell model rates . as the density of the stellar matter increases , the enhancement ratio decreases .
the pn - qrpa electron capture rates are enhanced by as much as factor of 13 around @xmath29 = 10^{7}$ ] and roughly by a factor of 4 at high density around @xmath29 = 10^{11}$ ] .
core - collapse simulators should take note of these enhanced reported electron capture rates .
these rates might contribute to some interesting simulation results .
a similar comparison for positron capture ( pc ) rates against previous calculations is shown in the two panels of fig.6 .
here one notes that the reported pn - qrpa positron capture rates are relatively in better agreement with the ffn rates although both ffn and shell model rates are enhanced . due to orders of magnitude differences
the ratios are plotted on a logarithmic scale in fig.6 .
the comparisons are fairly good at high temperatures .
it is worth mentioning that the positron capture rates are very small numbers and can change by orders of magnitude by a mere change of 0.5 mev , or less , in parent or daughter excitation energies and are more reflective of the uncertainties in the calculation of energies .
the calculated electron and positron capture rates on @xmath1 mg on a fine scale of stellar temperature - density is given in table 1 .
the calculated rates are tabulated in logarithmic ( to base 10 ) scale . in the table , -100.000 means that the rate is smaller than 10@xmath42 .
the first column gives log(@xmath43 ) in units of @xmath44 , where @xmath45 is the baryon density and @xmath46 is the ratio of the electron number to the baryon number .
stellar temperatures ( @xmath47 ) are measured in @xmath35 k. stated also are the values of the fermi energy of electrons in units of mev .
@xmath48(@xmath49 ) are the electron(positron ) capture rates in units of @xmath50 ( eqtn .
the ascii file of table 1 is also available and can be received from the author upon request .
the pn - qrpa theory with separable interaction was used to calculate the electron and positron capture rates on @xmath1 mg on a much detailed temperature - density grid point suitable for simulation purposes .
deformation parameter , which is believed to be one of the important parameters of pn - qrpa calculations , was taken from the experimental compilation to further increase the reliability of the calculations .
the reported electron capture rates are enhanced , by as much as a factor of 12 , compared to the shell model results of oda and collaborators [ 6 ] .
there exists a wide variety of discrepant results in the supernova simulations of o+ne+mg cores including prompt explosions , delayed explosions and no explosions .
fryer et al . [ 35 ] and kitaura et al .
[ 9 ] pointed that a possible explanation to these varying results could be explained by the different nuclear equation of states employed in the simulations and the treatment of capture rates on a large variety of nuclei both in nuclear statistical equilibrium ( nse ) and in the non - nse regime .
recent simulations of o+ne+mg cores by gutirrez et al .
[ 8 ] and kitaura et al .
[ 9 ] employ the shell model capture rates of oda et al .
[ 6 ] and takahara et al .
[ 10 ] , respectively .
the spherical core collapse simulations do not explode partly because of the reduced electron capture rates in the non - nse regime slowing the collapse and resulting in a large shock radius .
it can be expected that a parameter - free multi - dimensional model , with neutrino transport included consistently throughout the entire mass , including a complete treatment of multidimensional convection and burning phases might lead to a better understanding of the phenomena of supernova explosions .
incorporation of the reported ( enhanced ) capture rates might point toward still lower values of @xmath46 and lower entropy in the stellar core and might give some interesting results for simulators transforming a collapse into an explosion .
the author would like to acknowledge the local hospitality provided by the abdus salam ictp , trieste , where part of this project was completed .
99 arnett d 1996 _ supernovae and nucleosynthesis _ ( new jersey : princeton university press ) dessart l , burrows a , livne e and ott c d 2007 _ astrophys . j. _ * 669 * 585 hillebrandt w 1978 _ space sci .
* 21 * 639 sumiyoshi k , terasawa m , mathews g j , kajino t , yamada s and suzuki h 2001 _ astrophys .
j. _ * 562 * 880 wanajo s , tamamura m , naoki i , nomoto k , ishimaru y , beers t c and nozawa s 2003 _ astrophys . j. _ * 593 * 968 oda t , hino m , muto k , takahara m and sato k 1994 _ at .
data nucl .
data tables _ * 56 * 231 wildenthal b h 1984 in _ progress in particle and nuclear physics , vol 11 , pg 5 _ , ed d h wilkinson ( oxford : pergamon ) gutirrez j , canal r and garca - berro e 2005 _ astron . and astrophys .
_ * 435 * 231 kitaura f s , janka h - th and hillebrandt w 2006 _ astron . and astrophys .
_ * 450 * 345 takahara m , hino m , oda t , muto k , wolters a a , glaudemans p w m and sato k 1989 _ nucl .
phys . _ * a504 * 167 nabi j - un 1999 _ phd thesis _ ,
( germany : heidelberg university ) nabi j - un and klapdor - kleingrothaus h v 1999 _ at .
data nucl .
data tables _ * 71 * 149345 nabi j - un and klapdor - kleingrothaus h v 2004 _ at .
data nucl .
data tables _ * 88 * 237476 nabi j - un and rahman m - ur 2007 _ phys .
c _ * 75 * 035803 nabi j - un and rahman m - ur 2005 _ phys .
lett . _ * b612 * 1906 nabi j - un and sajjad m 2007 _ phys .
c _ * 76 * 055803 nabi j - un , sajjad m and rahman m - ur 2007 _ acta phys .
* 320323 stetcu i and johnson c w 2004 _ phys . rev .
c _ * 69 * 024311 raman s , malarkey c h , milner w t , nestor , jr c w and stelson p h 1987 _ at .
data nucl .
data tables _ * 36 * 1196 nilsson s g 1955 _ mat .
vid . selsk _ * 29 * 16 hirsch m , staudt a , muto k and klapdor - kleingrothaus h v 1991 _ nucl .
phys . _ * a535 * 62 audi g , wapstra a h and thibault c 2003 _ nucl .
_ * a729 * 337 yost g p _ et al .
_ ( particle data group ) 1988
_ phys . lett . _
* b204 * 1 rodin v , faessler a , simkovic f and vogel p 2006 _ czech .
j. phys . _
* 56 * 495 nabi j - un , rahman m - ur and sajjad m 2007 _ braz . j. phys . _
* 37 * 123845 martinez - pinedo g , poves a , caurier e and zuker a p 1996 _ phys .
c _ * 53 * r2602 rakers s , bumer c , frekers d , schmidt r , van den berg a m , hannen v m , harakeh m n , de huu m a , wrtche h j , de frenne d , hagemann m , heyse j , jacobs e and fujita y 2002 _ phys .
c _ * 65 * 044323 mller p and nix j r 1981 _ at .
data nucl .
data tables _ * 26 * 165 hirsch m , staudt a , muto k and klapdor - kleingrothaus h v 1993 _ at
. data nucl .
data tables _ * 53 * 16593 brown b a and wildenthal b h 1983 _ phys .
c _ * 27 * 1296 wildenthal b h 1984 _ prog .
phys . _ * 11 * 5 fuller g m , fowler w a and newman m j 1980 _ astrophys . j. suppl . _
* 42 * 447 ; fuller g m , fowler w a and newman m j 1982 _ astrophys . j. suppl . _ * 48 * 279 ; fuller g m , fowler w a and newman m j 1982 _ astrophys . j. _ * 252 * 715 ; fuller g m , fowler w a and newman m j 1985 _ astrophys . j. _ * 293 * 1 brown b a , chung w and wildenthal b h 1978 _ phys . rev .
_ * 40 * 1631 brown b a and wildenthal b h 1985 _ at .
data nucl .
data tables _ * 33 * 347 fryer c l , benz w , herant m and colgate s 1999 _ astrophys . j. _ * 516 * 892 mg , as function of temperature , for different selected densities .
densities are in units of @xmath52 .
temperatures are measured in @xmath35 k and log@xmath53 represents the log of electron capture rates in units of @xmath50.,scaledwidth=80.0% ] mg , as function of temperature , for different selected densities .
densities are in units of @xmath52 .
temperatures are measured in @xmath35 k and log@xmath54 represents the log of positron capture rates in units of @xmath50.,scaledwidth=80.0% ] * table 1 : * calculated electron and positron capture rates on @xmath1 mg for different selected densities and temperatures in stellar matter .
log(@xmath43 ) has units of @xmath55 , where @xmath45 is the baryon density and @xmath46 is the ratio of the electron number to the baryon number .
temperatures ( @xmath47 ) are measured in @xmath35 k. @xmath56 is the total fermi energy of electrons including the rest mass(mev ) .
@xmath48(@xmath49 ) are the electron(positron ) capture rates @xmath57 .
the calculated rates are tabulated in logarithmic ( to base 10 ) scale . in the table , -100.000 means that the rate is smaller than 10@xmath42 . | few white dwarfs , located in binary systems , may acquire sufficiently high mass accretion rates resulting in the burning of carbon and oxygen under nondegenerate conditions forming a o+ne+mg core .
these o+ne+mg cores are gravitationally less bound than more massive progenitor stars and can release more energy due to the nuclear burning .
they are also amongst the probable candidates for low entropy r - process sites .
recent observations of subluminous type ii - p supernovae ( e.g. , 2005cs , 2003gd , 1999br , 1997d ) were able to rekindle the interest in 8 10 m@xmath0 which develop o+ne+mg cores .
microscopic calculations of capture rates on @xmath1 mg , which may contribute significantly to the collapse of o+ne+mg cores , using shell model and proton - neutron quasiparticle random phase approximation ( pn - qrpa ) theory , were performed earlier and comparisons made .
simulators , however , may require these capture rates on a fine scale . for the first time a detailed microscopic calculation of the electron and positron
capture rates on @xmath1 mg on an extensive temperature - density scale is presented here .
this type of scale is more appropriate for interpolation purposes and of greater utility for simulation codes .
the calculations are done using the pn - qrpa theory using a separable interaction .
the deformation parameter , believed to be a key parameter in qrpa calculations , is adopted from experimental data to further increase the reliability of the qrpa results .
the resulting calculated rates are up to a factor of 14 or more enhanced as compared to shell model rates and may lead to some interesting scenario for core collapse simulators . |
blazars are generally considered as the most extreme class of radio - loud active galactic nuclei(agns ) , in which the two - sided jets with magnetized plasmas are launched nearby the central engine at relativistic velocity aligned with the line of sight .
traditionally , blazars are classified as bl lacs and fsrqs according to the equivalent width(ew ) of their strong emission lines @xcite . the objects with a rest frame @xmath1 are called bl lacs , otherwise called fsrqs .
new , based on the location of synchrotron peak frequency(@xmath2 ) @xcite , blazars with values of @xmath3hz are named low synchrotron peaked blazars(lsp ) , ones with values between @xmath4hz@xmath5hz are named intermediate synchrotron peaked blazars(isp ) , others with values of @xmath6hz@xmath7 are named high synchrotron peaked blazars(hsp ) .
otherwise , from the ratio of luminosity of the blr in units of eddington units , e.g. , @xmath8 , blazars with @xmath9 mainly display the properties of fsrqs @xcite .
actually , these classifications are potentially associated with the blr .
the existence of the blr around central engine is an almost uncontroverted fact in blazars , especially for fsrqs .
the blr is known to emit a number of strong lines and the diffuse continuums at the optical to uv bands , but its structure , for example , with spherical or
flat " geometry , is still under debate .
usually a spherical geometry of the blr is assumed , in which its photons follow a blackbody spectrum peaking at the frequency of @xmath10hz @xcite , and its radius @xmath11 is determined by the reverberation mapping technique @xcite derived from its reprocessing emission of the accretion disk @xcite . if the @xmath0-ray emission region of blazars is located within the cavity of the blr , the photons from the blr will significantly absorb the @xmath0-rays with energies above few tens of gev via positron - electron pair production processes @xcite .
@xcite used a spherical blr with half thickness @xmath12(0.05@xmath13pc ) to investigate the absorption of @xmath0-rays , and demonstrated that the @xmath0-rays with energies between 10 and 200gev can not escape from the diffuse photon field of the blr .
for the geometry of the blr , @xcite found that the spherical blr fails to explain the observed line width and shape in quasi - stellar objects .
the black hole mass deficit " observed in nls1s @xcite disappears provided that nls1s have a disc - like , rather than spherical blr @xcite .
@xcite produced a reasonable hardening of spectrum of quasar 3c 279 in tev bands assuming anisotropic blr structure .
the stability of gev breaks , as well as the @xmath0-ray spectral breaks from a fraction of gev to tens of gev band in blazars can be well produced by the absorption of he ii lyman recombination continuum and emission lines , indicating that @xmath0-rays arise within the high - ionization zone of the blr @xcite .
@xcite studied the spectral properties of 3c 454.3 in gev bands , implying that @xmath0-ray emission region is close to the inner boundary of the blr , which is impossible for the spherical blr .
recently , @xcite studied the properties of the blr for a set of blazars and reference quasars , and supported that the blr could have a flat geometry in blazars . in the paper , we study the effect of the geometrical structure of the blr on the absorption of @xmath0-rays in detail .
section 2 , we characterize our model .
subsection 2.1 , the geometry and emission of the blr are presented , and the @xmath0-ray optical depth is calculated in subsection 2.2 .
our results are presented in section 3 .
discussion and conclusions are given in section 4 .
in the following , we assume that the @xmath0-ray emission region moves at relativistic velocity along the jet and locates at a position @xmath14 above the accretion disk at a certain time . the generalized geometry of the blr is shown in figure [ fig : blr ] .
two shaded areas with axis symmetry represent the blr clouds , and the central strip corresponds to the accretion disk .
the blr is assumed to be flat and characterized by the aperture angle @xmath15 measured from the disk plane . in calculations , @xmath15
are set between @xmath16 and @xmath17 .
we only consider the photons from the blr without being obscured by accretion disk .
the blr is geometrically thick with the scale from inner radius @xmath18 to outer radius @xmath19 .
note that throughout this paper , the @xmath18 is fixed at @xmath20 , while @xmath19 is fixed at @xmath21 ( @xmath22 is the gravitational radius , @xmath23 ) .
this constrain is based on the assumptions of canonical blazars as follows : ( 1 ) the accretion disk is a standard disk .
the inner and outer radius are @xmath24 and @xmath25 , respectively @xcite ; ( 2 ) the accretion disk is within the cavity of the blr , while the @xmath19 is fixed at @xmath21 @xcite .
the blackbody temperature of the inner region of the accretion disk is assumed to be @xmath26k .
the detailed mathematical treatments on the blr are presented in appendix .
the accretion disk emits a total luminosity of @xmath27 uniformly toward all directions , the blr clouds intercept a fraction of illuminating continuum and reprocess them to the emission lines and the diffuse continuum .
moreover , as stated by @xcite , the blr also emits the recombination continua of hydrogen and he ii , they could cause jumps in the @xmath0-ray opacity at @xmath28 and 4.8gev .
however , in this work , we do not consider the contributions from the recombination continua to the @xmath0-ray absorption .
for simplicity , we assume that the fractions @xmath29 and @xmath30 of @xmath27 are reprocessed into the emission lines @xmath31 and the diffuse continuum @xmath32 respectively , @xmath29 and @xmath30 are set to be 0.05 , i.e. , the total reprocessing fraction is fixed at 0.1 . in order to evaluate the total emission line flux from the blr
, we consider the 35 components of emission lines .
the sum of line ratios is given by @xmath33 relative to @xmath34 line ratio ( @xmath35 ) .
therefore the flux contributed by a certain emission line equals @xmath36 where @xmath37 is the total emission line flux contributing to @xmath0-ray emission region . throughout this paper ,
the luminosity of the central disk is fixed at @xmath38 erg @xmath39 , corresponding roughly to the intermediate value of the luminosity of fsrqs . according to the assumptions used by @xcite
, the luminosity emitted by accretion disk are reprocessed into the emission lines and the diffuse continuum .
the emissivity of the emission lines is @xmath40 the emissivity of the diffuse continuum is @xmath41 where the number densities of the clouds and electrons are assumed to follow the power - law distributions , e.g. , @xmath42 and @xmath43 , where @xmath44 and @xmath45 are the number densities of clouds and electrons at @xmath18 , respectively . for the radial distribution of clouds ,
the same assumption is adopted , such as @xmath46 , where @xmath47 is the radius of clouds at @xmath18 . in present paper , we adopt the power - law exponents preferred by @xcite to be @xmath48 and @xmath49 .
the intensity of radiation with the angle @xmath50 respect to the jet axis at position @xmath51 is @xmath52 where @xmath53 , @xmath54 is the total emissivity that has @xmath55 and vanishes when @xmath56 or @xmath57 .
the soft photon energy density of the blr is given by @xmath58 the photon number densities associated with the emission lines and the diffuse continuum are @xmath59}n(r,\theta),\end{aligned}\ ] ] where @xmath60 is the monochromatic intensity , and @xmath61 is the normalization factor . at last ,
the @xmath0-ray optical depth is calculated by @xmath62 where @xmath63 is the dimensionless energy of soft photon , @xmath64 is the soft photon number density at position @xmath51 with the angle @xmath50 of integral path respect to jet axis . in the calculation of @xmath65 attenuation optical depth ,
the soft photon frequencies are from @xmath66hz to @xmath67hz .
the upper limit over @xmath51 is set to be 2pc in all the cases .
@xmath68 is the pair - creation cross section , and the threshold of @xmath68 is determined by @xmath69 in order to increase the threshold of @xmath0-ray photons in collisions , we can reduce the factor @xmath70 . in a
flat " geometry of the blr , the decrease of @xmath15 will reduce the probability of head - on " collisions and increase the threshold of the @xmath0-ray photons .
to reduce the number of parameters , we keep the mass of central supermassive black hole(bh ) to be @xmath71 , which is a typical mass of bh in agns .
the total energy density distributions of the blr are shown in figure [ fig : u_blr ] .
it is obvious that the aperture angle @xmath15 has dramatic impact on the energy density distributions
. actually , with the increase of @xmath15 , the illuminating continuum intercepted by the blr clouds increases , leading to the increase of the flux of the reprocessed emissions and the diffuse continuum .
the energy density is almost constant within the cavity of the blr , but at distance nearly close to @xmath18 , the energy density increases because the anisotropic effect of the blr geometry becomes more apparent .
furthermore , the energy density declines outside @xmath18 , but when @xmath15 is relatively low , the energy density declines even if the distance is less than @xmath72 . in the case of small @xmath15 ,
the blr is disc - like , the energy density is inversely proportional to the distance . in contrast , in larger @xmath15 nearly equaling to @xmath73 , the energy density is contributed by clouds within the blr and will decrease along the radial distance due to the obscure effect of the blr clouds .
overall , the location of the peaks of the energy density shifts toward larger distance with the increase of @xmath15 .
the soft photons from the blr are not only scattered by relativistic electrons in the jet to produce high - energy component in the seds of fsrqs , are but also the absorbing sources of @xmath0-rays via the processes of pair - creation .
@xcite found that the @xmath0-ray photons with energies from 10 to 200gev can not escape from the diffuse field of the blr with a spherical shell geometry . in this paper , we extend a closed " geometry to a flat " geometry for calculating the @xmath0-rays absorption . because the emissivity of diffuse and line emissions is proportional to the disk luminosity @xmath27 , the optical depth @xmath74 is also proportional to @xmath27 . in figure
[ fig : tau_e ] , we plot the optical depths versus @xmath0-ray energies .
the @xmath0-ray absorption are shown at three different distances , e.g. , @xmath75pc .
it is shown that the aperture angle @xmath15 affects the @xmath0-rays optical depth @xmath76 greatly . from the top panel in figure [ fig : tau_e ] .
the emitting region locates at @xmath14=0.01pc , we can see that even if @xmath15 is as low as @xmath16 , the optical depth of the @xmath0-ray photons with energies @xmath77gev approaches unity , the photons with energies lower than this critical energy can escape out the diffuse field . from middle panel in figure
[ fig : tau_e ] , when @xmath14 approaches @xmath18 and @xmath15 equals to @xmath16 , the @xmath0-ray photons with energies lower than critical energy @xmath78gev can escape out the diffuse field . increasing @xmath15 to @xmath79 ,
the critical energy is down to @xmath80gev .
accordingly , @xmath15 grows up to @xmath81 , the critical energy is down to @xmath82gev .
when @xmath14=0.3pc within the blr clouds and @xmath15 is less than @xmath81 , the @xmath0-ray photons with energies up to @xmath83gev can escape out the diffuse field .
we calculate the optical depths of @xmath0-ray photons in the framework of flat " geometry of the blr , in which flatness " is characterized by the aperture angle @xmath15 . for illustrating the dependence of @xmath76 on @xmath15
, we consider three positions as above , @xmath84pc , while the energies of @xmath0-ray photons are considered from 50 to 100gev energy bands , the results are shown in figure [ fig : tau_alpha ] .
it shows that when the @xmath0-ray emitting region locates at position @xmath85pc , the energy density of the diffuse field is large and opaque to all of the @xmath0-ray photons with energies beyond 50gev as shown in the upper panel in figure [ fig : tau_alpha ] . in middle panel ,
the @xmath0-ray emitting region is close to the inner boundary of the blr , the @xmath0-ray photons undergo the short path before they escape out the blr under a certain @xmath15 .
we can observe the @xmath0-ray photons with energy 50gev in the case of @xmath86 , 60gev in the case of @xmath87 , 70gev in the case of @xmath88 .
when @xmath0-ray emitting region locates within the clouds of the blr , as long as the aperture angle @xmath89 , the @xmath0-ray photons from 50gev to 100gev will be observed . because the optical depth depends on the height of the emitting region and the aperture angle , when @xmath90 , we can find a critical height @xmath91 of @xmath0-ray transparency for the aperture angle @xmath15 shown in figure [ fig : rc_alpha ] , for @xmath0-ray photons with the energies from 50gev to 100 gev .
it is shown that the critical aperture angle @xmath92 moves toward smaller angle with the increase of the @xmath0-rays energies , because the decrease of @xmath15 will reduce the probability of head - on " collisions , leading the increase of threshold energy .
for the @xmath0-ray with 100gev , @xmath93pc , the critical aperture angle @xmath92 is nearly @xmath16 , and the geometry of the blr becomes disc - like .
the above results are based on the assumption that the outer radius @xmath94 of accretion disk extends to the inner radius @xmath18 of the blr , e.g. , @xmath95 . in the following ,
we try to consider the effect of the disk size on soft photon fields and @xmath0-ray absorption .
the results are shown in figure [ fig : u_blr_ratio ] for @xmath96 versus @xmath51 , in which @xmath15 is set to @xmath16 , and in figure [ fig : rc_ratio ] for @xmath91 versus @xmath15 , in which the outer radius of disk varys as @xmath97 .
the disk size mainly influences the energy density of the diffuse photons field inside the blr . when @xmath98 , e.g. , the central engine is a point source of radiation that isotropically emits and radially locates behind the jet , and the energy density inside the blr becomes a constant . it is noted that the disk size does not affect the relation of @xmath91 and @xmath15 .
it is evident that blr clouds do not exist near the jet axis under the disk when @xmath15 is small , and @xmath94 does not change the energy density of the diffuse photons field .
furthermore , when @xmath15 is large , the critical position @xmath91 will be within the blr clouds , and the energy density keeps constant due to the obscure effect of blr clouds .
the geometry of the blr is usually assumed to be an isotropic shell , the absorption of the @xmath0-ray photons by the diffuse field is serious when the @xmath0-ray emitting region locates within the cavity of the blr @xcite .
the present work extends an isotropic blr to a flat " blr described by a aperture angle @xmath15 . in the previous work @xcite ,
the blr clouds are assumed to intercept a fraction @xmath99 of the illuminating continuum , where @xmath100 is the solid angle surrounded by the blr clouds , e.g. , @xmath101 .
the minimum aperture angle is then given by @xmath102 , we get @xmath103 . giving this limit
, we can determine the upper limit @xmath104 that increases with the decline of @xmath0-ray energy shown in figure [ fig : rc_alpha ] .
alternatively , the geometry of the blr could be reflected by the deprojection factor _ f_. in the commonly assumed isotropic shell , @xmath105 .
when the geometry is disc - like @xcite @xmath106^{-1/2},\end{aligned}\ ] ] where @xmath107 is the scale ratio of the blr , @xmath108 is the thickness , @xmath51 is the radius , and @xmath50 is the angle between the normal to the disk and the line of sight .
we then find that the relationship between @xmath15 and @xmath109 is given by @xmath110^{1/2}.\end{aligned}\ ] ] when we take @xmath111 according to the result given by @xcite , and @xmath112 for blazars , in which @xmath113 is the doppler factor from 10 to 50 randomly , we obtain @xmath114 , indicating that the geometry of the blr is rather flat . under this case , almost all the @xmath0-ray photons can escape out the diffuse photons field . in summary , we extend the spherical geometry of the blr to the flat geometry characterized by the aperture angle @xmath15 . in calculations ,
we take the central bh mass of @xmath115 , the inner and the outer radius of the blr to be @xmath116 and @xmath21 , and the luminosity of the central disk to be @xmath38 erg s@xmath117 .
the diffuse photon energy density decreases with the decline of @xmath15 , but it keeps constant inside @xmath18 and declines outside @xmath18 for a given @xmath15 .
the optical depth @xmath76 of @xmath0-rays not only depends on @xmath14 and @xmath0-ray energy , but also on @xmath15 strongly . setting @xmath118
, we can obtain the relation between the critical distance @xmath91 and the aperture angle @xmath15 for different energies .
we find that when the geometry of the blr is flat , the @xmath0-rays with specified energies can escape away the cavity of the blr .
we acknowledge the financial supports from the national basic research program of china ( 973 program 2009cb824800 ) , the national natural science foundation of china 11133006 , 11163006 , 11173054 , and the policy research program of chinese academy of sciences ( kjcx2-yw - t24 ) . as shown in figure [ fig : blr ] , we firstly perform integration for the shaded region surrounded by two blue dashed lines , the black solid line is the integral path to @xmath119 that has a angle of @xmath50 with respect to the jet axis , and @xmath120 , @xmath121 are the lower and upper limits , respectively .
the integral limits are given by here @xmath124 is the angle that jet axis makes with the line linking the emitting region and outer boundary of the disk , while @xmath125 is the angle that jet axis makes with the line linking the emitting region and apex
@xmath121 is determined by + if @xmath126 : @xmath127^{1/2},\end{aligned}\ ] ] if @xmath128 : @xmath129 if @xmath130 : @xmath131 if @xmath132 : @xmath133^{1/2},\end{aligned}\ ] ] if @xmath134 : @xmath135 where @xmath136^{1/2 } , \nonumber\\ \theta_{\rm \delta}&=&\angle_{\rm 1}-\angle_{2 } , \nonumber\\ \xi&=&\big[r^{2}+r_{\rm blr , o}^{2}-2rr_{\rm blr , o}\cos(\pi/2+\alpha)\big]^{1/2}.\end{aligned}\ ] ] it is noted that when we integrate the shaded region surrounded by two blue dashed lines , the red region will not be integrated , because with the increase of @xmath15 , the line linking @xmath0-ray emitting region and apex m " becomes more steeper , leading to the loss of the red region . for compensation to above calculation , we consider a critical position that satisfies the condition : @xmath137 , if @xmath138 , @xmath139 , the contributions from red region to the soft seed photon field could be neglected .
inversely , we have to integrate the red region .
the integral limits of the red region are given by @xmath140 , \nonumber\\ \theta_{\rm max}&=&\arccos\big[\frac{\sqrt{r^{2}-r_{\rm blr , o}^{2}}}{r}\big],\end{aligned}\ ] ] @xmath141^{1/2},\end{aligned}\ ] ] while @xmath121 is taken as equation ( [ eq : l_begin])([eq : l_end ] ) . | the broad - line region(blr ) is an important component of blazars , especially for the flat spectrum radio quasars(fsrqs ) .
the soft photons arising from the blr will substantially affect the transparency of the @xmath0-ray photons produced in the relativistic jet . in the paper , we study the effect of the geometrical structure of the blr on the absorption of @xmath0-rays .
we find that the @xmath0-ray optical depth strongly depends on the geometrical structure of the blr .
for a flat " geometry of the blr , the @xmath0-ray photons with specified energies could escape transparently even their emission region locates inside the cavity of the blr . |
the study of neutron - rich nuclei with unusually large @xmath5 ratios is challenging the conventional view of nuclear structure @xcite .
it is established that , when going from the @xmath6-stability line to drip line , the shell structure evolves and new magic numbers may emerge due to the dynamic effects of the nucleon - nucleon interaction .
the influence of the proton - neutron interaction and its higher order term ( namely the tensor force ) has been extensively analyzed in recent publications @xcite .
one may expect that other components of the nucleon - nucleon interaction contribute also to the shell evolution .
however , this issue has not been much covered , which may be related to the fact that the isoscalar channel of the effective interaction is usually much stronger than the isovector part . in ref .
@xcite , it was found that a modification for the isovector @xmath7 channel of the monopole interaction is needed to reproduce the energies and electromagnetic transition properties of low - lying states in nuclei around @xmath8c . in this paper
we analyze in detail contributions from the different terms of monopole interaction to the evolution of shell structure .
in particular , we will show that the neutron - neutron monopole interaction can have significant effect on the evolution of @xmath1 shell in carbon isotopes .
the rapid decrease of the @xmath9 in @xmath1 isotones indicates that the @xmath1 shell gap existing in oxygen isotopes erodes in nitrogen and carbon isotopes @xcite . at @xmath1 , the @xmath4 value measured through the lifetime of the @xmath10 state in @xmath0c
@xcite is much larger than the values in @xmath11c @xcite and @xmath12o @xcite while an inelastic scattering measurement presents a much smaller @xmath4 value in @xmath0c @xcite . in this paper
, we will also analyze @xmath13 transition properties in neutron - rich carbon isotopes from the viewpoint of collectivity . in sec .
[ sec : level2 ] we briefly introduce the theoretical framework .
the influence of different monopole terms of the effective interaction on the evolution of the @xmath1 shell is discussed in sec .
[ sec : level3 ] . in sec .
[ sec : level4 ] we analyze @xmath13 decay properties of the nuclei @xmath14c .
the calculations are summarized in sec .
[ sec : level5 ] .
the model space we choose contains @xmath15 and @xmath16 shells @xcite . to facilitate the calculation , we restrict the maximum number of two for nucleons that can be excited from @xmath15 to @xmath16 shell .
this truncation is denoted as @xmath17 space as usual .
it should be mentioned that the full @xmath17 space should also include contributions such as the excitation of one nucleon from the @xmath15 shell to the @xmath18 shell or from the @xmath19 shell to the @xmath16 shell .
such configurations are neglected in the present study , since they do not show any significant influence on the structure of nuclei of concern .
similarly , in the present work , the @xmath20 truncation implies that no nucleon is excited from the @xmath15 shell to the @xmath16 shell .
the well - established wbt @xcite and mk @xcite effective hamiltonians are used .
the wbt interaction is constructed based on the usd @xcite interaction .
the @xmath15-shell and @xmath21-cross - shell matrices in wbt are determined by fitting experimental data .
the @xmath16 , @xmath15 , @xmath21 and @xmath22 parts in the mk interaction are taken from the pw interaction @xcite , ck interaction @xcite , mk potential @xcite and kuo - brown g matrix @xcite , respectively .
calculations are carried out with a newly - established parallel shell - model code described in ref .
@xcite . in @xmath17 calculations
, one has to remove the spurious states which are caused by the center - of - mass motion . as suggested in ref .
@xcite , the shell model hamiltonian @xmath23 is modified to be @xmath24 where @xmath25 is the center - of - mass hamiltonian .
one can separate the low - lying states from the spurious states by taking a large positive @xmath6 value , which can move the center - of - mass excitations to high excitation energies .
we take @xmath26 mev in the present work , which is used in some previous works in this region @xcite .
the effective single - particle energy ( espe ) @xcite , describes in a simple manner the evolution of shell structure within the shell model framework @xcite .
the espe of a given @xmath27 orbit is written as @xcite @xmath28 where @xmath29 is the single - particle energy with respect to the core , @xmath30 is the occupancy of nucleons in the @xmath31 orbit and @xmath32 is the monopole interaction @xcite . as in ref .
@xcite , one has @xmath33\frac{(2j+1)}{(2j+1)(2j'+1)}\langle jj'|v|jj'\rangle_{jt},\ ] ] where @xmath34 is the two - body matrix element and @xmath35 and @xmath36 are the angular momentum and the isospin of the corresponding two - particle state , respectively . to explore the influence of configuration mixing on the espe
, we will adopt the occupancy in eq .
( [ speequ ] ) obtained from the full shell - model diagonalization .
in the independent particle model , the last neutrons in @xmath37c , @xmath38c and @xmath0c should occupy the @xmath39 orbit as in o isotopes . however , the ground state and the first excited state of @xmath40c are @xmath41 and @xmath42 states , respectively , which indicates that the @xmath43 and @xmath39 orbits are inverse in @xmath40c with respect to @xmath8o . as noted in ref .
@xcite , @xmath37c shows a strong @xmath44 configuration .
the rapid decrease of @xmath9 in @xmath1 isotones from o to c @xcite also indicates that shell structures in o and c isotopes are different . in order to study the structure difference between o and c isotopes , firstly we present @xmath9 and espes in the even - neutron c , n and o isotopes , as shown in fig .
[ spegra ] . for n isotopes ,
the energies are effective @xmath9 which are deduced from the @xmath45 weighted average of @xmath46 and @xmath47 @xcite .
it is seen that the calculated @xmath9 values with wbt and mk interactions are systematically larger than experimental values in c and n isotopes .
it has been pointed out that a reduction of @xmath48 in the @xmath16 part of wbt can improve the results @xcite .
for such an improvement , the modified versions of wbt have been suggested by multiplying the @xmath16 part of @xmath48 in wbt by a factor of @xmath49 for c isotopes ( named wbt * @xcite ) and @xmath50 for n isotopes ( named wbtm @xcite ) .
we can see that the @xmath9 calculations are improved with wbt * for c isotopes and wbtm for n isotopes , shown in fig .
[ spegra ] . from @xmath38o to @xmath37c , the quenching of @xmath48 is explained via the core polarization effect @xcite . in fig .
[ spegra ] , we do not show the @xmath17 results , since they are very similar to the @xmath20 calculations .
both the energy and the @xmath51 gap in n isotopes are almost in the middle of those in c and o isotopes . in c isotopes ,
the calculated energy gap between @xmath43 and @xmath39 orbits is much smaller than what is shown in ref .
@xcite . by calculating espes with shell - model occupancy , the correlation effect
is approximately included .
so the espes are different from those calculated with simple occupancy like assuming six neutrons all occupying @xmath39 orbits in @xmath0c and @xmath12o .
the wbt * and wbtm give smaller @xmath51 gaps for c and n isotopes than those by wbt .
the role that each interaction plays in evolution of @xmath1 shell can be investigated based on present espes .
the energy gap between @xmath43 and @xmath39 is written as @xmath52 the difference between @xmath0c and @xmath12o is written as @xmath53 -[\varepsilon_{\nu0d_{5/2}}(^{22}\mbox{o})-\varepsilon_{\nu0d_{5/2}}(^{20}\mbox{c } ) ] \nonumber\\ & = & \sum_{j'}[\langle\nu1s_{1/2}j'|v|\nu1s_{1/2}j'\rangle_{jt}-\langle\nu0d_{5/2}j'|v|\nu0d_{5/2}j'\rangle_{jt}]\times \nonumber\\ & & [ n_{j'}(^{22}\text{o})-n_{j'}(^{20}\mbox{c})]\label{deltae}.\end{aligned}\ ] ] the first part is the monopole interaction between the @xmath31-th orbit and @xmath43 ( @xmath39 ) orbit , while the second part is the difference of occupancy on @xmath31-th orbit between @xmath12o and @xmath0c .
besides two @xmath54 protons , the numbers of protons and neutrons on other orbits are also different between @xmath12o and @xmath0c .
the contribution from each term of the monopole interaction can be analyzed by eq .
( [ deltae ] ) .
there are totally @xmath55 terms of the monopole interaction which contribute in @xmath17 calculation . in @xmath20 calculation
, the number of the terms reduces to @xmath56 because no protons are in @xmath16 shell and neutrons in @xmath15 shell are fully occupied in @xmath12o and @xmath0c .
[ gaponc ] presents @xmath57 and @xmath58 , respectively .
it is seen that the proton - neutron interaction , @xmath59 , has large contribution to @xmath60 as discussed in ref .
the two terms , @xmath2 and @xmath3 , have also significant contributions to @xmath60 . in @xmath17 calculation
, the influence of other @xmath61 terms is small .
when removing two protons from @xmath12o , the proton - neutron interaction reduces the @xmath1 gap and enlarges the mixing between @xmath43 and @xmath39 orbits . at the same time , the neutron - neutron interaction contributes to the shell evolution because the neutrons in @xmath43 and @xmath39 orbits rearrange as a result of correlation effect from shell - model diagonalization .
although the espes do not entirely include the correlation effect because the multipole interaction is not taken into account , we can approximately study the role of neutron - neutron correlation in shell evolution by applying occupancy obtained from full shell - model diagonalization to eq .
( [ speequ ] ) .
the attractive neutron - neutron interactions are mostly due to the attractive singlet - even ( se ) channel of the central force @xcite .
specially , @xmath3 is completely from this channel .
the study of shell evolution extends our understanding of nucleon - nucleon interaction , especially the tensor force @xcite . from a general view of shell evolution ,
when proton ( neutron ) number changes in isotones ( isotopes ) , the neutron ( proton ) shell structure evolves due to the proton - neutron interaction . here
we present a detailed investigation on @xmath1 shell evolution to point out that the role played by the proton - neutron interaction in shell evolution can be `` enhanced '' by the attractive neutron - neutron interaction due to the many - body correlations .
it is important to notice this `` enhancement '' if one wants to study the nuclear force from shell evolution .
it is interesting to investigate the contribution of @xmath62 and @xmath63 in other cases of shell evolution in order to gain a complete understanding of how nuclear force drives the shell evolution .
the @xmath64 shell existed in o isotopes will dramatically shrink in si isotopes @xcite .
we adopt the similar method which we use in the study of @xmath1 shell for a detailed investigation of the @xmath64 shell evolution from @xmath65o to @xmath66si .
the result shows that , besides the proton - neutron interaction which was discussed in ref .
@xcite , the neutron - neutron interactions , @xmath67 and @xmath3 , contribute to the @xmath64 shell evolution from @xmath65o to @xmath66si . because of the limitation of calculation , it is hard to perform full shell - model diagonalization to study shell evolution in heavier regions , such as that in @xmath68 isotones which was discussed in refs .
one may expect that more information from shell evolution can be gained through full shell - model diagonalization in heavier regions in the future .
contribution of each interaction in @xmath1 gap difference between @xmath12o and @xmath69n ( left panel ) , @xmath12o and @xmath0c ( right panel ) .
the dashed line is for the value obtained from observed levels in @xmath12o and @xmath69n @xcite .
the @xmath17 calculations are indicated , otherwise @xmath20 calculations . ]
existing shell - model interactions reproduce well the observed @xmath4 values for carbon isotopes @xcite including the large @xmath4 value in @xmath0c . in shell model , the @xmath4
is calculated by @xmath70^{2},\ ] ] where @xmath71 and @xmath72 are the angular moments of the initial and final states of the @xmath13 transition . @xmath73 and @xmath74 are proton and neutron @xmath13 transition matrix elements , respectively . in table
[ be2 ] , we present @xmath73 , @xmath74 and @xmath4 values for even - even @xmath75c . in the calculations ,
the oscillator parameter of @xmath76 is employed .
an approximate @xmath77-dependent effective charge is used @xcite . for @xmath75c ,
the effective charges are @xmath78 @xmath79 @xmath80 , @xmath81 , @xmath82 , respectively @xcite .
it is seen that the @xmath73 increases significantly from @xmath37c to @xmath0c .
its contribution in @xmath4 values is multiplied by @xmath83 .
thus the @xmath4 value in @xmath0c is much larger than those in @xmath37c and @xmath38c . as discussed in refs .
@xcite , the reduction in the @xmath84 gap from @xmath37c to @xmath0c would be one reason why the proton excitation probability and the @xmath4 value increase from @xmath37c to @xmath0c .
cccc & @xmath85 & @xmath86 & @xmath87 + @xmath37c & & & + mk-@xmath20 & 1.97 & 8.96 & 27.52 + wbt-@xmath20 & 1.27 & 9.39 & 20.88 + wbt-@xmath17 & 1.12 & 9.40 & 19.40 + wbt*-@xmath20 & 1.16 & 9.31 & 19.55 + sfo & 1.20 & 8.27 & 17.00 + sfo - tls & 1.38 & 8.06 & 18.20 + expt .
& & & [email protected]@xmath883.5 @xcite + & & & [email protected] @xcite + @xmath38c & & & + mk-@xmath20 & 2.25 & 11.22 & 30.55 + wbt-@xmath20 & 1.77 & 11.16 & 24.74 + wbt-@xmath17 & 1.78 & 11.15 & 24.89 + wbt*-@xmath20 & 1.78 & 11.08 & 24.69 + sfo & 1.58 & 9.87 & 19.50 + sfo - tls & 1.60 & 9.69 & 19.90 + expt . & & & [email protected]@xmath885.0 @xcite + + @xmath0c & & & + mk-@xmath20 & 3.64 & 11.82 & 42.19 + wbt-@xmath20 & 3.07 & 11.50 & 33.85 + wbt-@xmath17 & 3.00 & 11.32 & 32.43 + wbt*-@xmath20 & 2.96 & 11.30 & 31.95 + expt . & & & 37.5@xmath89 @xcite + the large @xmath4 value in @xmath0c can also be explained from the view of collectivity . in our previous work
, we claimed that @xmath0c may have large quadrupole deformation @xcite .
the intrinsic quadrupole moment @xmath90 of the @xmath91 state can be deduced from the spectroscopic quadrupole moment @xmath92 .
if the @xmath91 state is a rotational deformed state , the intrinsic quadrupole moment ( written as @xmath93 ) can also be deduced from the @xmath94 value of the @xmath13 transition . in the case of a good deformed rotor , these two definitions of the intrinsic quadrupole moments
should give the similar values .
one has @xmath95 and @xmath96 in our case , we have @xmath97 and @xmath98 .
the values of @xmath4 and @xmath92 can be calculated in shell model .
table [ q ] lists the intrinsic quadrupole moments for @xmath37c , @xmath38c and @xmath0c , and the ratios between @xmath90 and @xmath93 .
cccc & @xmath99&@xmath100&@xmath101 + @xmath37c & & & + mk-@xmath20 & @xmath102 & @xmath103&@xmath104 + wbt-@xmath20 & @xmath105 & @xmath106&@xmath107 + wbt-@xmath17 & @xmath108 & @xmath109&@xmath110 + wbt*-@xmath20 & @xmath111 & @xmath112&@xmath113 + expt . &
& @xmath114 @xcite & + & & @xmath115 @xcite & + @xmath38c & & & + mk-@xmath20 & @xmath116 & @xmath117&@xmath118 + wbt-@xmath20 & @xmath119 & @xmath120&@xmath121 + wbt-@xmath17 & @xmath122 & @xmath123&@xmath124 + wbt*-@xmath20 & @xmath125 & @xmath126&@xmath127 + expt . & & @xmath128 @xcite & + + @xmath0c & & & + mk-@xmath20 & @xmath129 & @xmath130&@xmath131 + wbt-@xmath20 & @xmath132 & @xmath133&@xmath110 + wbt-@xmath17 & @xmath134 & @xmath135&@xmath113 + wbt*-@xmath20 & @xmath136 & @xmath137&@xmath113 + expt .
& & @xmath138 @xcite & + it is seen that @xmath90 and @xmath93 are almost the same in @xmath0c , which , associated with a negative quadrupole moment , would indicate that @xmath0c may have an oblate deformed structure , and thus the large @xmath4 value in @xmath0c might be understood with a consideration of the deformed collectivity .
it may be of interest to understand why such collectivity is found in a nucleus as light as @xmath0c .
the six valence neutrons have strong configuration mixing due to the disappearance of the @xmath1 shell , and , due to the reduction of the proton @xmath139 gap in @xmath0c , the probability of the proton excitation to @xmath54 orbit increases , as discussed above .
the strong mixing of configurations in both protons and neutrons would induce the enhanced collective property in @xmath0c .
the results from amd @xcite and deformed skyrme - hartree - fock @xcite have also shown an oblate deformation for @xmath0c . compared with @xmath0c ,
the ratio between @xmath90 and @xmath93 in @xmath38c is significantly less than the number of @xmath140 in both wbt and mk calculations .
although we can not exclude other collective freedoms in @xmath38c in the present study , the smaller @xmath90 in @xmath38c compared with @xmath0c would indicate that the quadrupole deformation in @xmath38c should be smaller than that in @xmath0c . in case of @xmath37c ,
the wbt and mk give rather different results .
calculations with the wbt interaction show that the nucleus @xmath37c is well prolate deformed . in the shell model
where a spherical basis is used usually , the collectivity ( or deformation ) may be discussed in the language of configuration admixtures . in mean - field models ,
the collectivity ( or deformation ) can be discussed qualitatively by the filling of deformed orbits .
the deformed skyrme - hartree - fock calculation has been performed for @xmath8c @xcite , giving coexisting prolate and oblate minima in each potential energy surface of the ground and lowly - excited states .
it was predicted that the @xmath141 g.s . has the lowest energy minimum at a prolate deformation while the first excited @xmath41 state has the lowest minimum at an oblate shape @xcite .
the authors explained also that the observed hindered @xmath142 transition between the @xmath41 and @xmath141 states would be attributed to the shape difference of the two states @xcite .
if we look at the nilsson diagram of single - particle orbits , the @xmath143 $ ] orbit of the @xmath144 shell has a prolate - driving force and @xmath145 $ ] has an oblate - driving force , while @xmath146 $ ] has a prolate - oblate competing force . for the @xmath141 g.s . in @xmath8c
, the three valence neutrons outside the @xmath147 closed shell would have two on the @xmath143 $ ] orbit and one on @xmath146 $ ] , which would lead to a prolate deformation for the ground state .
the first excited @xmath41 state at the lowest energy should have two neutrons on @xmath145 $ ] and one on @xmath143 $ ] , which would make an oblate deformation . back to the even - even c
isotopes which we are investigating , the similar qualitative analyses of deformations may be made .
the @xmath148c isotopes should have coexisting prolate and oblate shapes , similar to @xmath8c .
the g.s . of @xmath37c would favor a prolate shape with two valence neutrons occupying the prolate @xmath143 $ ] orbit .
@xmath0c would favor an oblate deformation with two neutrons on the oblate @xmath145 $ ] orbit , while @xmath38c might be more shape - competitive between prolate and oblate deformations .
these qualitative analyses are consistent with the shell - model calculations given in table [ q ] .
however , it should be noted that the deformations of the carbon isotopes are rather soft , compared to heavy nuclei .
in summary , the nuclei around @xmath0c have been studied with shell model in @xmath20 and @xmath17 model spaces .
the @xmath1 shell closure appears in o isotopes clearly , but vanishes in c isotopes @xcite . besides the proton - neutron interaction which was discussed in ref .
@xcite , the neutron - neutron interactions , @xmath2 and @xmath3 , play also important role . when the @xmath1 shell gap is reduced , neutrons in the @xmath149 and @xmath144 orbits rearrange and the interactions @xmath2 and @xmath3 can have significant effects on the shell evolution . in @xmath17 calculation ,
other @xmath61 terms of monopole interaction have very small contributions to the shell evolution .
it would be useful to do systematic investigations on the contributions of @xmath62 and @xmath63 in order to understand the nuclear force from shell evolution . as a special case of nuclei near the neutron drip line
, we have analyzed the @xmath4 value for @xmath0c in which experimental data have been available . in refs .
@xcite , it was pointed out that the large @xmath4 in @xmath0c is caused by the reduction of the proton @xmath139 gap , which leads to a large proton transition matrix element .
the present shell - model calculation gives the similar result .
usually , a large @xmath4 value is connected with collectivity ( or deformation ) .
therefore , we have made a further investigation of the problem .
our calculation gives a large quadrupole moment ( or large @xmath4 ) for @xmath0c , indicating an increased collectivity ( or deformation ) . in the shell - model language ,
the enhanced collectivity is caused by strong configuration mixings in both valence protons and neutrons .
this work has been supported by the national natural science foundation of china under grant no .
cq acknowledges the supports by the swedish research council ( vr ) under grant nos . 623 - 2009 - 7340 and 621 - 2010 - 4723 and the computational support provided by the swedish national infrastructure for computing ( snic ) at pdc and nsc .
99 o. sorlin and m .-
porquet , prog .
61 ( 2008 ) 602 .
r. v. f. janssens , nature ( london ) 459 ( 2009 ) 1069 .
t. otsuka , r. fujimoto , y. utsuno , b. a. brown , m. honma , and t. mizusaki , phys .
87 ( 2001 ) 082502 .
t. otsuka , t. suzuki , r. fujimoto , h. grawe , and y. akaishi , phys .
95 ( 2005 ) 232502 .
t. otsuka , t. suzuki , m. honma , y. utsuno , n. tsunoda , k. tsukiyama , and m. hjorth - jensen , phys .
( 2010 ) 012501 .
y. utsuno , t. otsuka , b. a. brown , m. honma , t. mizusaki , n. shimizu , arxiv:1201.4077 .
a. umeya and k. muto , phys .
c 74 ( 2006 ) 034330 .
t. suzuki and t. otsuka , int .
e 18 ( 2009 ) 1992 .
m. stanoiu , f. azaiez , zs .
dombrdi , o. sorlin , b. a. brown , m. belleguic , d. sohler , m .
g. saint laurent , j. lopez - jimenez , y. e. penionzhkevich , g. sletten , n. l. achouri , j. c. anglique , f. becker , c. borcea , c. bourgeois , a. bracco , j. m. daugas , z. dlouhy , c. donzaud , j. duprat , zs .
flp , d. guillemaud - mueller , s. grvy , f. ibrahim , a. kerek , a. krasznahorkay , m. lewitowicz , s. leenhardt , s. lukyanov , p.mayet , s.mandal , h. van der marel , w. mittig , j. mrazek , f. negoita , f. de oliveira - santos , zs .
podolyk , f. pougheon , m. g. porquet , p. roussel - chomaz , h. savajols , y. sobolev , c. stodel , j. timr , and a. yamamoto , phys . rev .
c 69 ( 2004 ) 034312 .
m. stanoiu , d. sohler , o. sorlin , f. azaiez , zs .
dombrdi , b. a. brown , m. belleguic , c. borcea , c. bourgeois , z. dlouhy , z. elekes , zs .
flp , s. grvy , d. guillemaud - mueller , f. ibrahim , a. kerek , a. krasznahorkay , m. lewitowicz , s. m. lukyanov , s. mandal , j. mrzek , f. negoita , yu .- e .
penionzhkevich , zs .
podolyk , p. roussel - chomaz , m. g. saint - laurent , h. savajols , g. sletten , j. timr , c. timis , and a. yamamoto , phys .
c 78 ( 2008 ) 034315 .
d. sohler , m. stanoiu , zs .
dombrdi , f. azaiez , b. a. brown , m. g. saint - laurent , o. sorlin , yu .- e .
penionzhkevich , n. l. achouri , j. c. anglique , m. belleguic , c. borcea , c. bourgeois , j. m. daugas , f. de oliveira - santos , z. dlouhy , c. donzaud , j. duprat , z. elekes , s. grvy , d. guillemaud - mueller , f. ibrahim , s. leenhardt , m. lewitowicz , m. j. lopez - jimenez , s. m. lukyanov , w. mittig , j. mrzek , f. negoita , zs .
podolyk , m. g. porquet , f. pougheon , p. roussel - chomaz , h. savajols , g. sletten , y. sobolev , c. stodel , and j. timr , phys rev .
c 77 ( 2008 ) 044303 .
m. j. strongman , a. spyrou , c. r. hoffman , t. baumann , d. bazin , j. brown , p. a. deyoung , j. e. finck , n. frank , s. mosby , w. f. rogers , g. f. peaslee , w. a. peters , a. schiller , s. l. tabor , and m. thoennessen , phys rev .
c 80 ( 2009 ) 021302(r ) .
z. elekes , z. vajta , zs .
dombrdi , t. aiba , n. aoi , h. baba , d. bemmerer , z. flp , n. iwasa , .
kiss , t. kobayashi , y. kondo , t. motobayashi , t. nakabayashi , t. nannichi , h. sakurai , d. sohler , s. takeuchi , k. tanaka , y. togano , k. yamada , m. yamaguchi , and k. yoneda , phys rev .
c 82 ( 2010 ) 027305 . m. petri , p. fallon , a. o. macchiavelli , s. paschalis , k. starosta , t. baugher , d. bazin , l. cartegni , r. m. clark , h. l. crawford , m. cromaz , a. dewald , a. gade , g. f. grinyer , s. gros , m. hackstein , h. b. jeppesen , i. y. lee , s. mcdaniel , d. miller , m. m. rajabali , a. ratkiewicz , w. rother , p. voss , k. a. walsh , d. weisshaar , m. wiedeking , and b. a. brown , phys .
( 2011 ) 102501 .
m. wiedeking _
100 ( 2008 ) 152501
. h. j. ong , n. imai , d. suzuki , h. iwasaki , h. sakurai , t. k. onishi , m. k. suzuki , s. ota , s. takeuchi , t. nakao , y. togano , y. kondo , n. aoi , h. baba , s. bishop , y. ichikawa , m. ishihara , t. kubo , k. kurita , t. motobayashi , t. nakamura , t. okumura , and y. yanagisawa , phys .
c 78 ( 2008 ) 014308 .
s. raman , c. w. nestor jr . , and
p. tikkanen , at .
data nucl .
data tabl .
78 ( 2001 ) 1 .
z. elekes , zs .
dombrdi , t. aiba , n. aoi , h. baba , d. bemmerer , b.a .
brown , t. furumoto , z. flp , n. iwasa , .
kiss , t. kobayashi , y. kondo , t. motobayashi , t. nakabayashi , t. nannichi , y. sakuragi , h. sakurai , d. sohler , m. takashina , s. takeuchi , k. tanaka , y. togano , k. yamada , m. yamaguchi , and k. yoneda , phys .
c 79 ( 2009 ) 011302(r ) .
b. a. brown , prog . part . nucl
47 ( 2001 ) 517 . e. k. warburton and b. a. brown , phys rev .
c 46 ( 1992 ) 923 .
d. j. millener and d. kurath , nucl .
phys . a 255 ( 1975 ) 315 .
b. h. wildenthal , prog .
11 ( 1984 ) 5 ; b. a. brown and b. h. wildenthal , annu . rev . nucl . part .
38 ( 1988 ) 29 .
b. m. preedom and b. h. wildenthal , phys rev .
c 6 ( 1972 ) 1633
. s. cohen and d. kurath , nucl . phys .
73 ( 1965 ) 1 . t. t. s. kuo and g. e. brown , nucl .
85 ( 1966 ) 40 ; t. t. s. kuo , nucl .
phys . a 103 ( 1967 ) 71 .
a. umeya , g. kaneko , t. haneda , and k. muto , phys .
c 77 ( 2008 ) 044301 .
h. l. ma , b. g. dong , and y. l. yan , phys .
b 688 ( 2010 ) 150 .
y. utsuno and s. chiba , phys .
c 83 ( 2011 ) 021301 .
t. otsuka , m. honma , t. mizusaki , n. shimizu , and y. utsuno prog .
( 2001 ) 319 .
r. k. bansal and j. b. french , phys .
11 ( 1964 ) 145 .
t. zheng , t. yamaguchi , a. ozawa , m. chiba , r. kanungo , t. kato , k. katori , k. morimoto , t. ohnishi , t. suda , i. tanihata , y. yamaguchi , a. yoshida , k. yoshida , h. toki , and n. nakajima , nucl .
phys . a 709 ( 2002 ) 103 . c. r. hoffman , t. baumann , d. bazin , j. brown , g. christian , d. h. denby , p. a. deyoung , j. e. finck , n. frank , j. hinnefeld , s. mosby , w. a. peters , w. f. rogers , a. schiller , a. spyrou , m. j. scott , s. l. tabor , m. thoennessen , and p. voss , phys .
b 672 ( 2009 ) 17 . c. x. yuan , c. qi , and f. r. xu , chin . phys .
c 33(s1 ) ( 2009 ) 55 .
h. sagawa , x. r. zhou , x. z. zhang , and t. suzuki , phys .
c 70 ( 2004 ) 054316 .
y. kanada - enyo , phys .
c 71 ( 2005 ) 014310 .
h. sagawa , x. r. zhou , t. suzuki , and n. yoshida , phys .
c 78 ( 2008 ) 041304(r ) . | full shell - model diagonalization has been performed to study the structure of neutron - rich nuclei around @xmath0c .
we investigate in detail the roles played by the different monopole components of the effective interaction in the evolution of the @xmath1 shell in c , n and o isotopes .
it is found that the relevant neutron - neutron monopole terms , @xmath2 and @xmath3 , contribute significantly to the reduction of the @xmath1 shell gap in c and n isotopes in comparison with that in o isotopes .
the origin of this unexpectedly large effect , which is comparable with ( sometimes even larger than ) that caused by the proton - neutron interaction , is related to the enhanced configuration mixing in those nuclei due to many - body correlations .
such a scheme is also supported by the large @xmath4 value in the nucleus @xmath0c which has been measured recently .
shell model , shell evolution , @xmath1 shell , carbon isotopes |
= 3.6pt cccccccccccc name & @xmath6 & @xmath7 & @xmath8 & @xmath9 & @xmath10 & @xmath10 & @xmath10 & @xmath11 & mge ref . &
@xmath12 & @xmath13 + & ( mpc ) & ( ) & ( kpc ) & & ( 0.11 ) & ( 14 ) & ( 0.14 ) & ( 0.14 ) & & & ( % ) + ( 1 ) & ( 2 ) & ( 3 ) & ( 4 ) & ( 5 ) & ( 6 ) & ( 7 ) & ( 8) & ( 9 ) & ( 10 ) & ( 11 ) & ( 12 ) + ngc 0821 & 23.4 & 1.60 & 13.2 & 2.92 & -2.16 & -2.36 & -2.23 & 0.016 & c06 & 1.09 & 8.5 + ngc 1023 & 11.1 & 1.68 & 11.3 & 4.39 & -2.20 & -2.19 & -2.20 & 0.009 & s09 & 0.98 & 12.3 + ngc 2768 & 21.8 & 1.80 & 19.7 & 2.95 & -2.10 & -1.82 & -2.01 & 0.025 & c06 & 1.02 & 5.1 + ngc 2974 & 20.9 & 1.58 & 10.8 & 2.80 & -2.22 & -2.49 & -2.30 & 0.015 & c06 & 1.10 & 14.9 + ngc 3115 & 9.5 & 1.54 & 10.1 & 6.19 & -2.28 & -2.53 & -2.37 & 0.028 & e99 & 1.10 & 10.7 + ngc 3377 & 10.9 & 1.55 & 8.7 & 4.64 & -2.22 & -1.75 & -2.05 & 0.046 & c06 & 0.99 & 8.4 + ngc 4111 & 14.6 & 1.08 & 3.8 & 4.46 & -2.12 & -2.16 & -2.13 & 0.043 & & 0.85 & 7.7 + ngc 4278 & 15.6 & 1.50 & 9.0 & 3.76 & -2.19 & -2.45 & -2.29 & 0.028 & c06 & 1.09 & 7.8 + ngc 4473 & 15.3 & 1.43 & 8.4 & 4.18 & -2.12 & -2.29 & -2.18 & 0.015 & c06 & 1.20 & 6.4 + ngc 4494 & 16.6 & 1.69 & 8.0 & 2.04 & -2.18 & -2.52 & -2.26 & 0.045 & s13 & 1.06 & 9.2 + ngc 4526 & 16.4 & 1.65 & 18.2 & 5.13 & -2.21 & -2.29 & -2.24 & 0.014 & c06 & 0.97 & 9.0 + ngc 4649 & 17.3 & 1.82 & 21.4 & 3.86 & -2.10 & -2.34 & -2.19 & 0.020 & s13 & 1.05 & 8.9 + ngc 4697 & 11.4 & 1.79 & 12.5 & 3.66 & -2.16 & -2.34 & -2.23 & 0.023 & s13 & 0.99 & 7.1 + ngc 7457 & 12.9 & 1.56 & 6.9 & 3.05 & -1.82 & -2.23 & -1.96 & 0.036 & c06 & 1.33 & 6.1 + [ tab1 ] one of the pillars on which our entire paradigm of galaxy formation rests is the fact that dark matter of an unknown nature dominates the mass budget of the universe @xcite . without dark matter ,
the primordial gas would not be able to collapse sufficiently quickly within the center of dark matter halos to form the galaxies we observe @xcite .
total mass - density profiles of spiral galaxies were obtained decades ago , given the simple geometry of their spiral disks and the availability of ionized @xcite and neutral gas @xcite , which provides a kinematical tracer easy to measure and model ( see review by @xcite ) .
the measurements indicated flat circular rotation curves and consequently nearly - isothermal @xmath14 profiles .
the situation is very different for early - type galaxies ( etgs : ellipticals and lenticulars ) , which by definition lack well - defined spiral disks and are cold - gas poor . for etgs one
usually has to rely on expensive observations of the stellar kinematics and more complex dynamical models .
this has restricted most studies to radii not much larger than the half - light radius .
a general consensus has emerged for the mass distribution of the etgs inner parts .
long - slit observations of two different samples of @xmath15 etgs revealed rotation curves to be nearly flat with nearly - isothermal mass profiles , as in spiral galaxies , within the median radius @xmath16 sampled by the kinematics @xcite .
this agrees with strong gravitational lensing studies finding nearly - isothermal slopes , with small scatter , for the _ total _ galaxy density profile of 73 etgs , at a typical radius of @xmath5 @xcite .
these central slopes are similar to those measured for group / cluster - scale profiles using x - ray modeling ( e.g. * ? ? ?
* ) or stacked weak - lensing technique ( e.g. * ? ? ?
the largest detailed study of dark matter in galaxy centers was based on the integral - field observations of the volume - limited atlas@xmath1 sample of 260 etgs @xcite .
it inferred a median dark matter fraction of just @xmath17 , within a sphere of radius , over the full sample @xcite .
_ this shows that studies restricted to the central regions of etgs mainly measure the stellar mass distribution . _ to explore the region where dark matter dominates , one needs to reach at least @xmath18 .
existing studies at these radii targeted one galaxy at a time .
they used observations of extended hi disks @xcite , deep stellar kinematics at a few sparse locations @xcite , globular cluster kinematics @xcite and planetary nebulae @xcite .
the fact that galaxies were modeled using different techniques or kinematic tracers , and for a mix of etgs with axisymmetric and triaxial shapes , resulted in a still unclear picture of the global trends .
tentative conclusions suggest a trend with massive etgs having nearly isothermal total mass profiles and flat circular velocities ( see review by * ? ? ?
* ) , while lower mass ones have more slowly rising dark matter profiles and correspondingly falling circular velocity profiles ( also * ? ? ?
. the situation is changing with the availability of extended stellar kinematics for significant samples of etgs @xcite . here
we present the first detailed and fully homogeneous analysis using larges - scale stellar dynamics of a statistically significant sample of 14 etgs ( but see @xcite for less detailed models ) .
the key novelty of this work is the availability of _ two - dimensional _ stellar kinematics out to a median radius for the sample of @xmath19 from the sluggs survey @xcite , which we combine with integral - field kinematics in the central regions ( @xmath20 ) from the atlas@xmath1 survey @xcite .
we study a subsample of the sluggs etg sample , with kinematics from @xcite .
we want a homogeneous set of galaxies and we use an axisymmetric dynamical modelling method . for this reason we only select etgs classified as fast rotator in @xcite .
we add the fast rotator ngc 3115 , without atlas@xmath1 kinematics .
our sample of 14 galaxies is given in ( stellar masses @xmath0 from @xcite ) .
fast rotators were shown by @xcite to have kinematical axes , within about 1 , extremely well aligned with the _ photometric _ major axes measured at the much larger radii @xmath21 sampled by the sluggs kinematics .
this alignment , for the statistically - significant , atlas@xmath1 sample , shows that fast rotator etgs , as a class , must be axisymmetric out to the region sampled by our models .
significant deviations from axisymmetry would produce a broad distribution of kinematical misalignment that is clearly ruled out by the atlas@xmath1 data .
the sluggs kinematics of fast rotators is also generally consistent with axisymmetry and shows small kinematical misalignment .
some genuine deviations from a bi - symmetric velocity field do exist in the sluggs data .
notable examples are ngc 4494 in the maps of @xcite and ngc 4473 in @xcite
. however , these deviations are more likely due to unrelaxed substructure in the stellar halo , than to non - axisymmetry in the relaxed stellar distribution .
in fact , the tight alignment between atlas@xmath1 stellar kinematics and photometry , as well as the complete lack of photometric twist , is extremely unlikely in a triaxial configuration ( e.g. * ? ? ? * ) . in this work
we attribute deviations from the axisymmetric assumption to either sub - structure or systematic problems in the data , and simply try to remove their effects from our models .
we use the multi - gaussian expansion ( mge ) @xcite to parametrize both stellar and dark halo distributions .
the mges of 8 galaxies come from @xcite and one from @xcite .
they include deep @xmath22-band photometry out to 510 .
three sdss @xmath23-band mges come from @xcite .
the @xmath24-band mge for ngc 3115 comes from @xcite .
the mge for ngc4111 was redone on the @xmath25-band sdss image using the python version of the mge - fitting method of @xcite , to avoid the @xmath23-band saturation ( see * ? ? ?
* supplementary material ) .
the kinematics for the inner ( @xmath20 ) region of our galaxies was observed with the sauron integral - field spectrograph and was homogeneously extracted for the atlas@xmath1 survey as described in @xcite . for 10/14 galaxies the observations , taken as part of the sauron survey ,
were already presented in @xcite .
the kinematics for the outer regions were observed with the keck / deimos multi - slit spectrograph by the sluggs survey .
they were presented and described in @xcite .
briefly , in both cases the kinematics were extracted in pixel space using the ppxf method @xcite .
the atlas@xmath1 data covered the h@xmath26 and mg @xmath27 spectral region ( 48005380 ) and were extracted using the full miles stellar library @xcite as templates . while the sluggs data covered the triplet spectral region ( 8480 - 8750 ) and were extracted using the stellar library of @xcite .
a key step of this analysis is the removal of potential outliers from the sluggs kinematics .
these may have significant effects on the modeling results , owing to the relatively sparse sampling of the kinematics at large radii .
when fitting kinematics to models , one can enforce symmetry during the kinematic extraction ( e.g. * ? ? ?
we adopt here a conceptually similar idea , which is easier to apply to irregularly sampled or discrete kinematics ( see also @xcite ) . at every observed location @xmath28 on the sky , where the @xmath29-axis coincides with the kinematic major axis pa@xmath30 ( taken from @xcite )
, we calculate the rms velocity @xmath31 from the sluggs data , where @xmath32 is the mean stellar velocity and @xmath33 is the velocity dispersion .
we then generate a new set of bi - symmetric positions by replicating four times the observed values as follows : @xmath34 the measurement uncertainties are increased by @xmath35 . given the bi - symmetry of the model , the @xmath36 statistics , describing the quality of the fit , is unchanged
. applies to the even velocity moments like @xmath37 and @xmath38 . while for the odd velocity moments like @xmath24 ,
the last line is replaced by : @xmath39 the resulting kinematics measurements are then smoothed with the ( python version ) of the two - dimensional loess algorithm of @xcite , implemented@xmath40 by @xcite ( see example in ) .
comparison between the smoothed / cleaned @xmath41 and the original one suggests the sluggs error being underestimated due to systematics , with typical errors of @xmath42 ( ) .
the original errors would indicate non - smooth kinematics , which seems inconsistent with the smooth galaxy images . in the following models we do not fit the original kinematics , but
the symmetrized / smoothed one .
this has the added benefit of making the agreement between data and model easy to visually assess .
we need to verify that our models capture the global features of the data , and that our results are not driven by a few deviant values .
this is important in situations where data systematics may be present and relying entirely on @xmath36 statistics may be misleading .
another issue for the modeling is the fact that the atlas@xmath1 observations consist of many more data points with smaller uncertainties , which completely dominate the @xmath36 estimate .
however here we want our dark halo determinations to be especially constrained by the sluggs data , which sample the region where the halo dominates .
similarly to @xcite , we increase the size of the atlas@xmath1 kinematic uncertainties so that , for a good fit , the two datasets provide an equal contribution to the @xmath36 .
we leave the sluggs uncertainties unchanged , to retain properly normalized confidence levels for our model parameters .
the sluggs @xmath37 at the sauron locations tend to be lower than the measured sauron values .
we find a median offset of 11% , which is larger than the 5% level we consider unavoidable between independent datasets .
this offset was noted by @xcite , but its source is unknown .
the atlas@xmath1 data agree on average with hundreds of independent determinations from the literature ( fig .
8 of @xcite ) , suggesting the sluggs data may be offset with respect to the optical literature . here
we simply multiply the sluggs @xmath37 to fit , for each galaxy , the interpolated sauron data at the same locations .
this is the standard kinematics we fit with our models .
however , importantly , we have also run all our models with the sluggs data alone , and confirmed that the slopes of the total mass profiles agree with those of our standard models .
we model the atlas@xmath1 and sluggs stellar kinematics using the ( python version ) of the axisymmetric jeans anisotropic modeling ( jam ) method@xmath40 @xcite .
the approach is the same used in @xcite for the atlas@xmath1 data alone .
it employs a bayesian method with constant ( i.e. , ` ignorant ' ) priors on all parameters .
the key difference between this work and previous stellar dynamical studies of dark halos in etgs is the fact that we place virtually no constraint on the halo profile parameters .
the halo is assumed to be spherical , but is described by a generalized @xcite profile ( gnfw ) with free normalization , inner slope and break radius : @xmath43 our models have 7 free parameters . some are poorly constrained but are not of interest here .
they are just ` nuisance parameters ' , marginalized over to derive the total mass profiles studied here .
the parameters are : ( i ) the inclination @xmath25 ; ( ii ) the anisotropy @xmath44 , with @xmath45 and @xmath46 the stellar dispersion in cylindrical coordinates , for the mge gaussians with @xmath47 ; ( iii ) the anisotropy for the remaining gaussians at larger radii ; ( iv ) the stellar @xmath48 ; ( v ) the break radius of the dark halo , constrained to be @xmath49 kpc ; ( vi ) the halo density @xmath50 at @xmath51 ; and ( vii ) the dark halo slope @xmath52 for @xmath53 .
the first result is the simple fact that the models provide a good description of the stellar kinematics of all the modeled galaxies over the full field ( , [ fig : jam_fits2 ] and ) .
this was not expected .
it would have been natural if we had employed e.g. the more general orbit - superposition method ( e.g. * ? ? ?
* ) , which is fully described by thousands of parameters . however our models have just six non - linear parameters and one scaling factor @xmath48 .
moreover the fits look similarly good even assuming a power - law halo and a constant - anisotropy stellar body ( four nonlinear parameters and one scaling factor ) .
the fact that four parameters are able to consistently describe all features of the two - dimensional maps , for the full set of 14 galaxies , then indicates that ( i ) the ( cleaned ) sluggs data are reliable and ( ii ) the model assumptions provide a good description of the dynamics and mass distribution of the real galaxies .
an alternative interpretation for the good fits would be that the anisotropy and dark matter variations , and data systematics , conspire to mimic the simple orbital structure and mass distribution assumed by the models .
this would be a realistic possibility when studying a single galaxy , given the non full generality of the jam models , but such a conspiracy is unlikely for such a large and diverse set of galaxies .
the second and main result of this work is that all 14 modeled fast - rotator etgs have a nearly - isothermal @xmath14 _ total _ density distribution from @xmath54 out to the median radius of 4 sampled by this study ( ( a ) ) .
total _ mass - density profiles\ , { \rm erf}\left[r\sqrt{1-q_j^2}/(q_j\sigma_j\sqrt{2})\right]}{4\pi\sigma_j^2 r\sqrt{1-q_j^2}}$ ] ] , marginalized over all nuisance parameters , are tightly constrained by the data . in the whole range @xmath55 the profiles
are well described by a power - law @xmath56 with the largest average deviation of 11% .
the corresponding average logarithmic slope is @xmath3 for the sample , with an rms scatter of just @xmath4 .
we do not detect any significant correlation of the slope with , stellar mass , or stellar velocity dispersion . for @xmath57
the average slope is @xmath58 with @xmath59 , while for @xmath60 we find @xmath61 with @xmath62 .
our inner - profile slope and scatter are in excellent agreement with those @xmath63 with @xmath64 found around @xmath5 using strong lensing @xcite .
\(b ) shows that the observed trend is consistent with what one would predict for the whole atlas@xmath1 sample , for cosmologically - motivated _ uncontracted _ nfw halos . in ( c ) , the stellar profiles are very different from the total ones at the radii we sample : they have slopes @xmath65 around @xmath5 ( fig . 2 of @xcite ) but fall off more steeply than @xmath66 around @xmath19 .
\(d ) compares our total profiles with published ones for ngc 0821 , ngc 2974 , ngc 4494 , ngc 4649 , and ngc 4697 ( from * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* respectively ) .
the @xmath67 was derived from the circular velocities @xmath68 , assuming spherical symmetry . in 4/5 cases
the agreement is excellent , with our _ statistical _ uncertainties enclosing the published profiles over the full radial range .
notable is the tight agreement for ngc 2974 , where the @xmath68 was directly measured from a regular hi disk .
we believe the disagreement for ngc 4494 may be due to the inclusion in @xcite models of the strong asymmetry in the sluggs data at @xmath69 .
we excluded those data from our fits , but including them would improve the agreement .
we combine the integral - field stellar kinematics from the atlas@xmath1 survey , within @xmath20 , with the two - dimensional stellar kinematics from the sluggs survey , out to a median radius of about 4 , and a maximum radius of 2.06.2 , for a sample of fast rotator early - type galaxies consistent with axisymmetry .
we construct the first statistically significant set of detailed axisymmetric dynamical models of the two - dimensional stellar kinematics out to those large radii , where dark matter dominates .
we find that the galaxies dynamics is well represented by few relatively simple assumptions .
the models tightly constrain the _ total _ density profiles , which closely approximates the isothermal form @xmath14 , from /10 out to the median radius of 4 sampled by the data , with remarkably little scatter .
the observed total mass distribution is not a generic prediction of @xmath70cdm and provides constraints on the models ( e.g. * ? ? ?
* ; * ? ? ?
our sample highlights the importance of similar studies on larger samples of galaxies , to provide a much needed benchmark for galaxy formation models . for this
, studies like the present one , using deimos on keck or muse on the vlt , can be complemented with models of shallower data , but for much larger samples like manga @xcite . to be most useful , samples need to include both spiral and early - type galaxies .
these should be modeled in a fully homogeneous way by describing the kinematics of both the stars and the gas within the same gravitational potential .
large radio surveys of the neutral hi gas like apertif @xcite can complement manga at the largest radii .
the future availability of large kinematic datasets indicates a bright future for the systematic investigation of the mass profiles in galaxies .
mc acknowledges support from a royal society university research fellowship .
supported by national science foundation grants ast-0909237 and ast-1211995 .
daf thanks the arc for financial support via dp130100388 . ,
m. a. w. , oosterloo , t. a. , van cappellen , w. a. , et al .
2008 , in american institute of physics conference series , vol .
1035 , the evolution of galaxies through the neutral hydrogen window , ed .
r. minchin & e. momjian , 265271 | we study the total mass - density profile for a sample of 14 fast - rotator early - type galaxies ( stellar masses @xmath0 ) .
we combine observations from the sluggs and atlas@xmath1 surveys to map out the stellar kinematics in two - dimensions , out to a median radius for the sample of four half - light radii ( or 10 kpc ) , and a maximum radius of 2.06.2 ( or 421 kpc ) .
we use axisymmetric dynamical models based on the jeans equations , which allow for a spatially varying anisotropy , and employ quite general profiles for the dark halos , and in particular do not place any restriction on the profile slope .
this is made possible by the availability of spatially extended two - dimensional kinematics .
we find that our relatively simple models provide a remarkably good description of the observed kinematics .
the resulting _ total _
density profiles are well described by a nearly - isothermal power law @xmath2 from /10 to at least 4 , the largest average deviation being 11% .
the average logarithmic slope is @xmath3 with _ observed _ rms scatter of just @xmath4 .
this scatter out to large radii , where dark matter dominates , is as small as previously reported by lensing studies around @xmath5 , where the stars dominate .
our bulge - halo conspiracy places much tighter constraints on galaxy formation models .
it illustrates the power of two - dimensional stellar kinematics observations at large radii .
it would now be important to test the generality of our results for different galaxy types and larger samples . |
this paper concerns the vlasov - maxwell description of collisionless magnetized plasmas @xcite related to axisymmetric discs arising in the combined gravitational and electromagnetic ( em ) fields of astrophysical compact objects .
in particular , in this work an application of the kinetic theory developed by @xcite is considered , to the treatment of equilibrium and possibly non - neutral equatorial as well as off - equatorial plasma tori . remarkably , the existence of equilibrium levitating ( i.e. , off - equatorial ) structures has been recently pointed out by @xcite , based on a fluid non - ideal magnetohydrodynamics ( mhd ) description .
indication of existence of structures of this type occurring in collisionless plasmas is also supported by other recent works based on test - particle approach @xcite .
in addition , in the context of charge - separated pulsar magnetospheres , @xcite found an indication of the possible development of stable off - equatorial tori that can be revealed in his early numerical simulations of magnetized collisionless plasmas .
indeed , from the physical point of view , off - equatorial plasma configurations are intrinsically different from the case of equatorial disc systems . the astrophysical relevance of these structures lies in the possibility of modelling coronal plasmas characterized by low - density and high - temperature conditions @xcite . in this regard , two different issues arise , which deserve a detailed investigation . first , under these conditions fluid descriptions may become inappropriate , requiring in principle the adoption of a kinetic treatment .
however , it still remains to be ascertained whether the levitating structures can be recovered as equilibrium solutions in the framework of a kinetic description .
second , a priori it is not obvious whether the kinetic theory developed previously for equatorial plasmas can be extended to the description of off - equatorial tori , how this can be achieved in practice and what are the physical implications as far as their occurrence in real systems is concerned .
the statistical description of plasma dynamics can be carried out in terms of either fluid or kinetic treatments , with the choice of the appropriate framework depending on the plasma phenomenology to be addressed and the relevant features of the phenomena to be studied @xcite .
the majority of fluid approaches are based on hydrodynamic or mhd treatments @xcite . in the case of collisionless plasmas , when these are formulated independently of an underlying kinetic theory , some limitations can arise .
first , it is well known that the set of fluid equations may not be closed , requiring in principle the prescription of arbitrary higher - order fluid fields and closure conditions , including in particular the equation of state ( eos ) or the pressure tensor @xcite .
second , in these approaches typically no account is given of microscopic phase - space particle dynamics together with phase - space plasma collective phenomena .
on the other hand , only in the context of kinetic theory these difficulties can be consistently overcome , as this treatment permits to obtain well - defined constitutive equations for the relevant fluid fields describing the plasma state and to solve at the same time the closure problem @xcite .
these issues become relevant in the case of collisionless or weakly - collisional multi - species plasmas subject to em and gravitational fields where phase - space particle dynamics is expected to play a dominant role . in particular ,
kinetic theory is essential for studying both stationary configurations and dynamical evolution of plasmas when kinetic effects are relevant , such as ones associated with conservation of particle adiabatic invariants @xcite , temperature and pressure anisotropies , diamagnetic and finite larmor - radius ( flr ) effects as well as energy - correction contributions @xcite . the problem of formulating a kinetic theory appropriate for the description of collisionless plasmas in quasi - stationary ( i.e.
, equilibrium)configurations of astrophysical accreting systems and laboratory scenarios has been presented in series of works , based on the non - relativistic vlasov - maxwell description . in our context , several issues have been treated , ranging from laboratory plasmas occurring in tokamak devices @xcite , axisymmetric accretion disc plasmas characterized by locally - nested magnetic surfaces @xcite and current - carrying magnetic loops @xcite around compact objects as well as spatially non - symmetric systems in astrophysical and laboratory contexts @xcite . it was shown that consistent solutions of the vlasov equation can be determined for the species kinetic distribution function ( kdf ) describing collisionless plasmas , based on the identification of the relevant single - particle invariants .
the equilibrium kdfs were expressed in terms of generalized bi - maxwellian distributions , characterized by temperature anisotropy , non - uniform fluid fields and local plasma flows .
chapman - enskog representations of these equilibria were obtained by developing a suitable perturbative kinetic theory , which in turn made possible the analytical calculation of the fluid fields and the identification of the relevant kinetic effects included in the corresponding mhd description . as a basic consequence , it was shown that these solutions can exhibit non - vanishing current densities which can also support a kinetic dynamo mechanism for the self - generation of em fields in which the plasma is immersed @xcite .
finally , more recently the kinetic theory has been extended to describe axisymmetric plasmas characterized by strong shear - flow and/or supersonic velocities @xcite , while @xcite reports a kinetic analysis of the stability properties of particular equilibrium solutions with respect to axisymmetric em perturbations .
a notable feature of the kinetic equilibria mentioned here is the unique prescription of the functional dependences of an appropriate set of fluid fields , carried by the species kdfs , which are related to physical observables of the system .
these fields are referred to as _ structure functions _ and are denoted as @xmath0 ( see in particular @xcite and the definition below ) , with the subscript s being the species index .
depending on the kinetic regime being considered , according to the classification scheme presented by @xcite , these dependences are expressed in terms of the poloidal flux function @xmath1 of the magnetic field and/or the effective potential @xmath2 defined as@xmath3where @xmath4 and @xmath5 are the gravitational and electrostatic ( es ) potentials respectively , with @xmath6 and @xmath7 denoting the species particle mass and charge .
thus , in the general case kinetic theory requires that @xmath8 .
it must be stressed that , behind the apparent simplicity of the result , for practical applications one ultimately needs obtaining a representation of @xmath9 in terms of spatial coordinates , e.g. cylindrical ones @xmath10 .
assuming an axisymmetric configuration where @xmath1 and @xmath2 are generic functions of both @xmath11 , the representation @xmath12 applies .
hence , the complete solution of the problem actually requires determining the explicit representation of the potentials @xmath13 in terms of the spatial cylindrical coordinates . this can be a very difficult task , since in the general case the plasma itself contributes to the generation of the gravitational filed , through its non - vanishing mass - density , and , more important , of the em fields by means of non - vanishing charge and current densities . in the present discussion we ignore , however , the self - generation of the gravitational field , focusing only on the generation of em fields .
it follows that , in order to obtain explicitly the relationship between the em potentials and the coordinate system , one has necessarily to solve ( numerically ) the coupled set of vlasov - maxwell equations to determine @xmath14 and @xmath15 , where the source terms of the fields are prescribed functions of the potentials . on the other hand
, such a solution is also demanded for the following additional reasons : a ) in order to calculate explicitly the characteristic kinetic effects which enter the equilibrium kdf and are associated with diamagnetic - flr and energy - correction effects ( see @xcite and the discussion in section 7 ) ; b ) in order to establish the diffeomorphism @xmath16 which relates cylindrical ( and similarly , spherical coordinates ) to local magnetic coordinates , where @xmath17 is an angle - like coordinate defined on equipotential magnetic surfaces @xmath18 the complexity of the theory , as far as this issue is concerned , might represent a possible limit for the practical realization of kinetic equilibria of this type for configurations of astrophysical interest .
this may be relevant especially when demanding a comparison between kinetic and fluid treatments based on analytical solutions .
therefore , the question arises of whether such a difficulty can be actually encompassed in some scenarios , possibly by invoking suitable asymptotic kinetic orderings to be imposed on the system .
in particular , this concerns the existence of configurations which allow for the construction of the diffeomorphism @xmath19 based on analytical solutions of the potentials @xmath20 in such a way to afford an explicit treatment of the spatial dependences of the equilibrium fluid fields carried by the kdf according to the constraints posed by the vlasov equation . a second point of crucial
importance concerns the determination of the eos that characterizes disc plasmas in the collisionless state .
as mentioned above , this is usually realized by prescribing the form of the scalar pressure , or more generally the components of the pressure tensor , which represent the closure condition for the euler momentum equation in mhd treatments . since collisionless plasmas
are intrinsically characterized by the occurrence of phase - space anisotropies which cause the equilibrium kdf to generally deviate from a simple isotropic maxwellian distribution , the knowledge of the correct form of the eos is not a trivial task . in fact , one can only prescribe the pressure tensor in consistent way on the base of the kinetic theory ( kinetic closure conditions ) . in this regard ,
the problem consists in the identification of the kinetic effects which must be included in the pressure tensor , the understanding of their physical origin and the way they influence the macroscopic configuration of the plasma system . among the relevant ones , contributions associated with es corrections
can play a central role , since they arise from microscopic charge interactions and carry information about local es fields generated by the validity of quasi - neutrality condition in rotating plasmas or deviations away from it in non - neutral systems .
finally , from the astrophysical point of view a further motivation is represented by @xcite , where the discovery of the possible occurrence of stationary configurations of disc plasmas in off - equatorial tori is reported . in that work ,
a newtonian axisymmetric model of non - conductive , charged and perfect fluid tori orbiting in the combined gravitational and dipolar magnetic fields generated by a central compact object is presented .
the result is obtained in the framework of a non - ideal mhd description and shows that the interplay between gravitational and magnetic fields can effectively enhance vertically extended structures in a plasma torus , which may correspond to localized concentrations of matter above and under the equatorial plane .
the importance of this conclusion lies in the possibility of interpreting these off - equatorial tori as forming the surrounding material usually invoked to explain the spectral emission / absorption features in accretion - disc systems .
possible examples of this type are provided by coronal halos consisting of non - neutral ion - electron plasmas or by obscuring dusty - plasma tori that are believed to be produced in galactic nuclei @xcite .
the complete understanding of the physical properties of these structures is far from being satisfactory and deserves further investigations , both theoretical and observational . in particular , here we consider the possibility of a low - density and high - temperature coronal plasma for which the collisionless state applies .
following the arguments discussed above , under these conditions the proper framework for the description of these plasmas is represented by the kinetic theory . in particular , the question is posed of whether kinetic equilibria can be proved to exist for collisionless magnetized plasmas which exhibit off - equatorial maxima in the matter distribution and how these solutions can be possibly related to fluid mhd ones .
this amounts at identifying the appropriate kinetic regimes which meet these conditions and determining the spatial dependences of the corresponding fluid fields , including the pressure tensor and the intrinsically - kinetic effects contained in the eos . to be successful ,
a program of this type must make possible an analytical approach in accordance with the considerations presented above , to extend the kinetic theory developed in @xcite and to permit its practical application to gain insights into an astrophysical issue of notable importance connected with accretion - disc phenomenology . to conclude ,
we should comment on the fact that the reference works cited above on kinetic theory and off - equatorial tori as well as the present investigation are carried out in the framework of a non - relativistic description ( both with respect to the treatment of the gravitational field and the plasma velocities ) .
however , the problem of general - relativistic solutions could in principle be posed , since there is a number of situations in which the formation of accretion discs is due to strong gravitational fields , e.g. around black holes , and general - relativistic corrections must be considered .
although a complete theory of this type is still missing , non - relativistic treatments can nevertheless provide the reference framework for the inclusion of some relevant features characteristic of general - relativistic theories .
this can be achieved for example by the adoption of pseudo - newtonian potentials for the description of spherically - symmetric gravitational fields @xcite .
it has been shown , in fact , that the precision of the pseudo - newtonian description of stationary general - relativistic phenomena can be very high @xcite .
an alternative consists in the inclusion of post - newtonian corrections . in this reference , a kinetic theory of self - gravitating collisionless gases adopting such a technique can be found in @xcite for spherical solutions , and in @xcite for the case of axially - symmetric solutions .
in view of the considerations presented above , the purpose of this paper is the formulation of a kinetic theory appropriate for the analytical treatment of collisionless disc plasmas in axisymmetric off - equatorial tori ( see fig .
[ fig-1 ] for a schematic view of the configuration geometry ) .
the results of the investigation are the following ones : .[fig-1 ] \1 ) the identification of a physically - realizable astrophysical configuration in which the spatial profiles of the potentials @xmath21 can be analytically prescribed , when suitable plasma orderings apply . in the framework of an asymptotic theory ,
this permits to decouple the vlasov equation from the maxwell equations , at least to leading - order , with the possibility of an explicit treatment of the spatial dependences contained in the equilibrium plasma fluid fields .
2 ) the construction of species equilibrium kdfs which are consistent with the kinetic constraints imposed by microscopic phase - space conservation laws for the single particle dynamics .
these are proved to be expressed in terms of generalized maxwellian kdfs characterized by non - uniform fluid fields and isotropic temperature , both to be associated with a finite set of structure functions .
\3 ) the identification of the relevant kinetic regimes which can in principle arise in the configuration determined at point 1 ) above and the prescription of the corresponding functional dependences on the structure functions .
4 ) the development of a perturbative theory which makes possible a representation of the equilibrium kdfs in terms of a chapman - enskog series . as a result , this permits first the analytical evaluation of the plasma fluid fields , and second to distinguish the characteristic kinetic effects which enter the solution and their main physical properties .
\5 ) the proof that the gravitationally - bound and magnetized plasma regime provides the most general functional dependences in the structure functions carried by the equilibrium kdf , which are consistent both with the analytical treatment of the fluid fields and the occurrence of off - equatorial tori . as an illustration of the technique , explicit calculation of the leading - order species number density and velocity profiles
is provided for two configurations of physical interest .
\6 ) the calculation of the eos for the equilibrium plasma , to be expressed in terms of the species pressure tensor associated with the kdf . by making use of the perturbative treatment ,
this includes also the identification of both the kinetic effects and the es corrections that can effectively contribute to the eos . in particular ,
the latter are shown to determine non - trivial deviations from the assumption of having thermal pressure for the collisionless plasma .
\7 ) to determine the constraints to be imposed on the kinetic solution for the validity of the theory and the physical configuration realized , which arise from the maxwell equations for the self - generated equilibrium em fields .
these represent necessary conditions which must be verified a posteriori for the complete solution of the vlasov - maxwell problem and the explicit treatment of es corrections in comparison with fluid - based approaches .
the scheme of the paper is as follows . in section 3
the model assumptions and the fundamental em orderings are presented .
section 4 deals with the definition of plasma orderings and the introduction of corresponding kinetic regimes which characterize collisionless plasmas treated here . in section 5
the equilibrium species kdfs are explicitly determined and their realizations are given for three different plasma regimes . in section 6 a suitable perturbative theory is developed , which allows for the analytical treatment of the equilibrium kdfs and the corresponding fluid fields .
the conditions for the occurrence of off - equatorial tori are then investigated in the case of collisionless plasmas belonging to the gravitationally - bound and magnetized plasma regime . in section 7 the expression of the kinetic corrections which characterize the kinetic equilibria
is provided , while the corresponding contributions in the eos are computed in section 8 .
section 9 contains analysis of the poisson and the ampere equations and the constraints which they pose on the kinetic solution . finally , concluding remarks are summarized in section 10 .
for the construction of kinetic equilibria , we ignore the possible existence of weakly - dissipative effects ( coulomb collisions and turbulence ) and em radiation effects @xcite .
it is assumed that the kdf and the em fields associated with the plasma obey the system of vlasov - maxwell equations , with maxwell s equations being considered in the quasi - static approximation . for definiteness
, we shall consider here a plasma consisting of @xmath22-species of charged particles which are characterized by proper mass @xmath6 and total charge @xmath7 .
in particular , given a generic species kdf @xmath23 defined in the phase - space @xmath24 , with @xmath25 and @xmath26 being the configuration and velocity space respectively , the vlasov equation determines the dynamical evolution of @xmath27 and is given by@xmath28 the plasma is taken to be : a ) non - relativistic , in the sense that the flow velocities of all species are small compared to the speed of light @xmath29 , that the gravitational field can be treated within the classical newtonian theory , and that the non - relativistic vlasov kinetic equation is used as the dynamical equation for the kdf ; b ) collisionless , so that the mean free path of the plasma particles is much longer than the largest characteristic scale length of the plasma ; c ) axisymmetric , so that the relevant dynamical variables characterizing the plasma ( e.g. , the fluid fields ) are independent of the azimuthal angle @xmath30 when referred to a set of either cylindrical coordinates @xmath31 or spherical coordinates @xmath32 . thanks to the axisymmetry assumption , as a shortway in the following we shall denote with @xmath33 the configuration state vector , where @xmath33 denotes either @xmath34 or @xmath35 .
we are concerned here with quasi - stationary configurations , namely solutions which are slowly - varying in time .
this condition is also referred to as equilibrium configuration . for a generic physical quantity @xmath36 that depends on spatial coordinates @xmath33 and time @xmath37 ,
the quasi - stationarity is expressed by letting in the following @xmath38 , with @xmath39 being a small dimensionless parameter to be suitably defined ( see below ) and @xmath40 an integer .
similar considerations apply for the equilibrium kdf @xmath27 , which is denoted in the following as @xmath41 . from the symmetry properties introduced here
one can immediately derive the fundamental quantities which are conserved for the single - particle dynamics .
in particular , under the assumptions of axisymmetry the canonical momentum conjugate to the azimuthal angle @xmath42 is an integral of motion .
this is given by:@xmath43furthermore , from the condition of quasi - stationarity , the total particle energy@xmath44represents an adiabatic invariant of prescribed order , with @xmath2 being defined in eq .
( [ effective ] ) .
following the discussion in @xcite , here we recall that a generic quantity @xmath45 defined in the phase - space is an adiabatic invariant of order @xmath46 with respect to @xmath47 when it satisfies the condition @xmath48 , where @xmath49 is a suitable integer and @xmath50 is the cyclotron frequency .
this means that adiabatic invariants are conserved in asymptotic sense , namely up to a prescribed order of accuracy determined by the parameters @xmath47 and @xmath46 .
we consider solutions of the equilibrium magnetic field @xmath51 which admit , at least locally , a family of nested axisymmetric toroidal magnetic surfaces @xmath52 , where @xmath1 denotes the poloidal magnetic flux of @xmath51 .
the magnetic surfaces can be either locally closed @xcite or locally open @xcite in the configuration domain occupied by the plasma . in both cases
a set of magnetic coordinates ( @xmath53 ) can be defined locally , where @xmath17 is a curvilinear angle - like coordinate on the magnetic surfaces @xmath54 by construction , magnetic coordinates are related to cylindrical or spherical coordinates by a diffeomorphism @xmath55 which must be consistently determined , as discussed above .
each relevant physical quantity @xmath56 can then be conveniently expressed either in terms of the set @xmath33 or as a function of the magnetic coordinates , i.e. @xmath57 .
consistent with these assumptions , we require the em field to be slowly - varying in time , i.e. , of the form @xmath58 in particular , we assume the magnetic field to be represented as@xmath59where @xmath60 and @xmath61 denote the self - generated magnetic field produced by the plasma and a finite external axisymmetric magnetic field ( vacuum field ) . for definiteness , in this treatment both contributions are assumed to exhibit only non - vanishing poloidal components , to be denoted in the following as @xmath62 .
notice that , for what concerns the self - field , this assumption must be verified a posteriori to be consistent with the constraints placed on the kinetic solution by the maxwell equations .
hence , the two fields are represented as@xmath63so that the total poloidal magnetic field takes the form@xmath64with @xmath65 .
in particular , for the purpose of the present work , the external magnetic field is taken to coincide with a dipolar field .
in such a case , the flux function @xmath66 is written in terms of the spherical coordinates @xmath67 as@xmath68with @xmath69 being the magnitude of the dipole magnetic moment .
charged particles are also assumed to be subject to the effective potential @xmath70 defined by eq .
( [ effective ] ) . in principle
, both the es potential @xmath71 and the gravitational potential @xmath72 can be produced by the plasma itself and by external sources .
however , in the following it is assumed that @xmath71 is uniquely generated by the plasma charge density , while we shall neglect the self contribution of the plasma to @xmath4 .
hence , for an axisymmetric disc , the gravitational potential is taken as being stationary and to coincide identically with the potential generated by the central compact object .
the latter is expressed here in terms of the spherically - symmetric newtonian potential as@xmath73where @xmath74 is the newton gravitational constant and @xmath75 is the mass of the compact object . given validity of these assumptions , in order to address the first issue posed in the section 2 we proceed introducing the following fundamental orderings for the self em fields:1 ) the self component of the equilibrium magnetic field @xmath60 is ordered with respect to the external field @xmath61 as@xmath76with @xmath77 , and in general @xmath78.2 ) the equilibrium es potential @xmath5 satisfies the ordering assumption@xmath79with @xmath77 .
this means that the es potential energy is small with respect to the gravitational potential energy , where @xmath47 will be properly defined below . in the following the case @xmath80 in eqs .
( [ basic ] ) and ( [ eees ] ) will be considered .
it is important to remark that the two conditions ( [ basic ] ) and ( eees ) pose strong constraints from the physical point of view on the realizability of the equilibrium configuration for the collisionless disc plasma .
in particular , they must be verified a posteriori in order to warrant the consistency of the kinetic equilibrium solution for the species kdf with the validity of the maxwell equations .
( [ basic ] ) and ( [ eees ] ) are the fundamental em orderings which permit the analytical treatment of the spatial dependences of the equilibrium fluid fields as prescribed by the kinetic solution to be determined below .
in fact , when these orderings hold , to leading - order the magnetic flux function @xmath1 and the effective potential @xmath2 coincide respectively with the vacuum fields , namely the external dipolar flux function @xmath81 in eq .
( [ dipolo ] ) and with the gravitational potential @xmath4 given by eq .
( newton ) .
hence , at this order it is possible to construct explicitly the diffeomorphism which relates the spherical coordinates @xmath67 with the potentials @xmath82 .
this is expressed by letting@xmath83the corresponding relationship holding for cylindrical coordinates can then be obtained from the spherical ones , giving in particular@xmath84
in this section we introduce the relevant orderings for collisionless plasmas , which permit to identify corresponding kinetic regimes characterized by different equilibrium solutions and to verify their consistency with the existence of off - equatorial tori . to this aim
we determine a classification scheme based on the method outlined in @xcite and suitable for the treatment of the physical setting indicated in the previous section .
the first step is the definition of the dimensionless species - dependent parameters @xmath85 , @xmath86 and @xmath87 .
these are prescribed in such a way to be all independent of single - particle velocity and at the same time to be related to the characteristic species thermal velocities .
both perpendicular and parallel thermal velocities ( defined with respect to the magnetic field local direction ) must be considered .
these are defined respectively by @xmath88 and @xmath89 , with @xmath90 and @xmath91 denoting here the species perpendicular and parallel temperatures . in detail , the first parameter is defined as @xmath92 , where @xmath93 is the species larmor radius , with @xmath94 being the characteristic scale - length of the spatial variations of all of the fluid fields associated with the kdf and of the em fields .
the second parameter @xmath86 is related to the particle canonical momentum @xmath95 . by denoting @xmath96 , @xmath86
is identified with @xmath97 .
hence , @xmath86 effectively measures the ratio between the toroidal angular momentum @xmath98 and the magnetic contribution to the toroidal canonical momentum , for all particles in which @xmath99 is of the order @xmath100 while @xmath1 is assumed as being non - vanishing .
in particular , here the magnetic flux can be estimated as @xmath101 , with @xmath102 denoting the characteristic length - scale of flux variations and @xmath103 the magnitude of the poloidal magnetic field .
note that , by definition , @xmath104 , but , in principle , we can also have @xmath105 locally .
we also notice that the present definition of @xmath86 is not the only possibility , as one could also define the parameter @xmath86 such that it is related to the azimuthal flow velocity , namely letting @xmath106 , with @xmath107 being the corresponding angular frequency .
finally , @xmath87 is related to the particle total energy @xmath108 and is prescribed in this work as @xmath109 .
it follows that @xmath87 measures the ratio between particle kinetic and es potential energy with respect to the gravitational potential energy , for all particles having velocity @xmath110 of the order @xmath111 , with @xmath4 being assumed as non - vanishing .
we notice that the definition of @xmath87 differs from that used in previous works @xcite while it is consistent with the ordering assumption ( [ eees ] ) . in the following
we shall denote as thermal subset of velocity space the subset of the euclidean velocity space in which the asymptotic conditions @xmath112 holds .
a comment is in order regarding the role of the magnetic field in the two parameters @xmath86 and @xmath85 . in the first case , the magnetic field
is represented by means of the poloidal flux @xmath1 which contributes to the toroidal canonical momentum @xmath95 , while @xmath85 depends on the magnitude of the total magnetic field . invoking the definitions for @xmath86 and @xmath85 given above
, it follows that @xmath113 , where @xmath94 and @xmath102 are respectively the characteristic scale - lengths of equilibrium fluid and em fields and of the poloidal flux . in general , the two quantities should be considered as independent , with @xmath104 and @xmath114 ( in the present context one has identically @xmath115from eq .
( [ b field ] ) ) .
indeed , the parameter @xmath86 determines the particle spatial excursions from a magnetic flux surface , while @xmath85 measures the amplitude of the larmor radius with respect to the inhomogeneities of the background fluid fields .
these two effects correspond to different physical magnetic - related processes , due respectively to the larmor - radius and magnetic - flux surface confinement mechanisms . in this work
we assume that the ordering condition@xmath116holds for the collisionless plasma considered here .
this amounts at requiring that the larmor radius remains small with respect to the scale - length @xmath94 , which , as shown in @xcite , represents a condition that is expected to be easily verified in accretion - disc systems .
hence , one can consistently identify the small parameter @xmath47 introduced above with @xmath117 .
the classification that is introduced in this work is based on the magnitude of the two parameters @xmath86 and @xmath87 . in detail ,
plasma species will be distinguished as belonging to the following regimes:1 ) _ gravitationally - bound _ if @xmath118 and @xmath119.2 ) _ magnetized _ if @xmath120 and @xmath121.3 ) _ gravitationally - bound and magnetized _ if both @xmath118 and @xmath120 . in the case of regimes 1 and 3
the following asymptotic expansion holds for the total particle energy @xmath122:@xmath123 .
\label{order - sigma}\]]similarly , for regimes 2 and 3 the particle canonical momentum @xmath124 admits the expansion@xmath125 .
\label{ordering - epsilon}\]]it is instructive to analyze the main features of these regimes and the physical conditions for their occurrence .
the action of some energy non - conserving mechanisms is required for the establishment of the case of gravitationally - bound plasmas . in particular , plausible physical mechanisms that can be responsible for the decrease of the single - particle kinetic energy , in both collisionless and collisional ad plasmas , are em interactions ( e.g. , binary coulomb collisions among particles and particle - wave interactions , such as landau damping ) and/or radiation emission ( radiation - reaction ) .
these can in principle be ascribed also to the occurrence of em instabilities and em turbulence .
for single particles these processes can be dissipative . as a consequence ,
these particles tend to move towards regions with higher gravitational potential ( in absolute value ) .
after multiple interactions of this type , the process can ultimately reach an equilibrium state which corresponds to the gravitationally - bound regimes .
as far as the magnetic - field based classification , we notice that the requirement @xmath120 ( regimes 2 and 3 ) means that a particle trajectory remains close to the same magnetic surface @xmath126 , while satisfying the ordering ( [ eps - emme ] ) .
finally , for greater generality , in the rest of the treatment we shall assume that , in the regimes in which @xmath118 and/or @xmath120 , the orderings @xmath127 and @xmath128 apply .
in this section we proceed with the construction of the species equilibrium kdf and its characterization to the plasma regimes identified above . we consider both exact as well as asymptotic representations for the solution , the latter being expressed in terms of a chapman - enskog series . to reach the goal , here we adopt the solution technique developed in @xcite , which consists in the construction of solutions of the vlasov equation of the form @xmath129 , where @xmath130 is a suitable adiabatic invariant .
this amounts at requiring that @xmath130 is necessarily a function of particle adiabatic invariants . in view of the model assumptions introduced above
, it follows that the general form of the equilibrium kdf in the present context is given by@xmath131with @xmath40 and where slow - time dependences are assumed to be uniquely associated with the particle energy . here
@xmath132 denotes the so - called structure functions , i.e. , functions which depend implicitly on the particle state @xmath133 . in order for @xmath134 to be an adiabatic invariant
, @xmath132 must also be functions of the adiabatic invariants .
this restriction is referred to here as a kinetic constraint .
the precise form of the functional dependences of @xmath132 is characteristic of each plasma regime , as discussed below . in order to determine an explicit representation of @xmath130 according to eq .
( [ fss ] ) , we impose the following requirements : \1 ) the kdf must be characterized by species - dependent non - uniform fluid fields , azimuthal flow velocity and isotropic temperature , to be suitably prescribed in terms of the structure functions .
\2 ) open , locally nested magnetic flux surfaces : the magnetic field is taken to allow quasi - stationary solutions with magnetic flux lines belonging to locally nested and generally open magnetic surfaces .
\3 ) kinetic constraints : suitable functional dependences are imposed on the structure functions @xmath132 which depend on the regime being considered and such to warrant @xmath130 to be an adiabatic invariant .
\4 ) in all regimes , @xmath130 is required to be asymptotically close ( in a suitable sense to be defined below ) to a local maxwellian kdf .
this requires the possibility of determining a posteriori a perturbative representation of the kdf equivalent to the chapman - enskog expansion for the analytical treatment of implicit phase - space dependences contained in the structure functions , with the consistent inclusion of es corrections , flr - diamagnetic and/or energy - corrections contributions .
5 ) the kdf @xmath130 must be a strictly - positive real function and it must be summable , in the sense that the velocity moments of the form@xmath135must exist for a suitable ensemble of weight functions @xmath136 , to be prescribed in terms of polynomials of arbitrary degree defined with respect to components of the velocity vector field @xmath137 . then , following @xcite , we express the equilibrium kdf @xmath130 as@xmath138which is referred to as the _ generalized maxwellian kdf .
_ here the structure functions are identified with the set @xmath139 , where @xmath140 , @xmath141 and @xmath142 are related to the species number density , isotropic temperature and azimuthal angular velocity respectively . invoking the definitions ( [ p_fi ] ) and ( [ total_energy ] ) , eq . ( [ gm ] )
can also be written as@xmath143 } { \left ( 2\pi /m_{\mathrm{s}}\right ) ^{3/2}t_{\ast \mathrm{s}}^{3/2}}\exp \left\ { -\frac{m_{\mathrm{s}}\left ( \mathbf{v}-\mathbf{v}_{\ast \mathrm{s}}\right ) ^{2}}{2t_{\ast \mathrm{s}}}\right\ } , \]]where @xmath144 and@xmath145 it is worth pointing out that the form of the solution ( [ gm ] ) holds for all the plasma kinetic regimes identified in the previous section .
the difference in the three cases concerns the prescription of the kinetic constraints to be imposed on @xmath132 .
in particular , consistent with the requirements listed above , these are assigned as follows : 1 ) for gravitationally - bound plasmas it is required that the functional dependence of @xmath132 is of the type@xmath146for which eq .
( [ order - sigma ] ) applies , while implicit dependences with respect to @xmath147 remain excluded in such a case .
\2 ) for magnetized plasmas , the kinetic constraint is realized by imposing@xmath148for which eq .
( [ ordering - epsilon ] ) applies , while implicit dependences with respect to @xmath149 are excluded .
\3 ) for gravitationally - bound and magnetized plasmas both eqs .
( order - sigma ) and ( [ ordering - epsilon ] ) hold , so that the general form of the kinetic constraint is given by@xmath150 the connection between the realization of these regimes and the occurrence of off - equatorial tori will be investigated in the next section .
here it must be noticed that , because of the constraints ( [ gb - k])@xmath151([gb - m - k ] ) , at this stage the structure functions can not be regarded as fluid fields , since they are defined in the phase - space , namely they depend on the single particle velocity via the particle energy @xmath108 and the canonical momentum @xmath95 . instead
, the fluid fields associated with @xmath130 must be properly computed as velocity moments according to eq .
( [ velomo ] ) and they are unique once the precise form of the structure functions is explicitly prescribed in @xmath130 .
in this section we first proceed determining a chapman - enskog representation for @xmath130 that makes possible the treatment of the implicit phase - space functional dependences carried by the structure functions as well as the analytical evaluation of the equilibrium fluid fields and the associated kinetic contributions .
we then apply the result to prove the validity of the kinetic theory developed here as far as the description of off - equatorial toroidal structures is concerned .
this task can be achieved by implementing an appropriate perturbative theory for @xmath130 which was first developed in @xcite and which is based on a taylor expansion of @xmath132 with respect to the dimensionless parameters @xmath87 and @xmath86 .
it is understood that the basic feature of such a kinetic perturbative technique is that it is strictly applicable only in localized subsets of velocity space ( thermal subsets ) , namely to particles whose velocity satisfies the asymptotic ordering ( [ order - sigma ] ) and/or ( ordering - epsilon ) .
a notable consequence of such an approach is that , for each kinetic regime , quasi - stationary , self - consistent , asymptotic solutions of the vlasov - maxwell equations ( kinetic equilibria ) can be explicitly determined by means of suitable taylor expansions of @xmath130 . in particular , it is found that maxwellian - like kdfs can be obtained locally in phase - space , where the appropriate convergence conditions hold .
this procedure provides also the correct constitutive equations of the leading - order fluid fields as well as the precise form of the es , flr - diamagnetic and/or energy - correction contributions to the kdf . in detail , invoking eqs .
( [ order - sigma ] ) and ( [ ordering - epsilon ] ) , a linear asymptotic expansion for the structure functions can be obtained . in the general case , neglecting corrections of @xmath152 as well as of @xmath153 and @xmath154 , with @xmath155 , this is given by@xmath156
_ { \psi _ { \ast \mathrm{s}}=\psi , \phi _ { \ast \mathrm{s}}=\frac{m_{\mathrm{s}}}{z_{\mathrm{s}}e}\phi _ { \mathrm{g } } } \nonumber \\ & & + \left ( \phi _ { \ast \mathrm{s}}-\frac{m_{\mathrm{s}}}{z_{\mathrm{s}}e}\phi _ { \mathrm{g}}\right ) \left [ \frac{\partial \lambda _ { \ast \mathrm{s}}}{\partial \phi _ { \ast \mathrm{s}}}\right ] _ { \psi _ { \ast \mathrm{s}}=\psi , \phi _ { \ast \mathrm{s}}=\frac{m_{\mathrm{s}}}{z_{\mathrm{s}}e}\phi _ { \mathrm{g } } } , \label{qq}\end{aligned}\]]where @xmath9 is the leading - order term which uniquely follows for each regime ( see below ) . when eq .
( [ qq ] ) is applied to the equilibrium kdf @xmath130 and the ordering ( [ eees ] ) is also invoked , the following chapman - enskog representation is found:@xmath157 , \label{ce}\]]where the leading - order contribution @xmath158 coincides with a drifted isotropic maxwellian kdf carrying non - uniform number density , azimuthal differential flow velocity and isotropic temperature . in detail:@xmath159where @xmath160 is the leading - order drift velocity carried by @xmath158 . here
@xmath161 represents the leading - order species number density and is given by@xmath162 , \label{nde}\]]with @xmath163 being referred to as the pseudo - density .
the leading - order structure functions @xmath9 coincide now with the set of functions @xmath164 which are defined in the configuration space , with @xmath165 and @xmath107 being respectively the leading - order species temperature and azimuthal rotation angular frequency .
in addition , the quantities @xmath166 , @xmath167 and @xmath168 represent first - order kinetic corrections .
in particular , @xmath169 is referred to as flr - diamagnetic contribution , @xmath167 carries energy - correction contributions ( with respect to both kinetic and es potential energies ) , while @xmath170 represents a purely es term .
the precise expression of these functions will be given below in a separate section , where we discuss the relevance of these kinetic effects and their physical meaning in the framework of the present perturbative theory .
for the moment , it is sufficient to say that all the first - order corrections are part of the kinetic equilibrium , and can not be neglected for the consistent formulation of the solution .
it must be also stressed here that eq .
( [ ce ] ) is very general : while @xmath170 is non - vanishing for all the kinetic regimes considered above , the existence of @xmath169 and @xmath167 depends instead on the type of kinetic constraints . in particular , we distinguish the following features : 1 ) for gravitationally - bound plasmas ( regime 1 ) @xmath171 and @xmath9 is subject to the constraint@xmath172 \2 ) for magnetized plasmas ( regime 2 ) @xmath173 and the functional dependence of @xmath9 becomes@xmath174 \3 ) for gravitationally - bound and magnetized plasmas ( regime 3 ) in general both @xmath175 and @xmath176 , while for @xmath9 one has in this case@xmath177 we notice that , to the leading - order , the equilibrium solution determined here does not depend on the es potential , but only on the gravitational potential @xmath4 and the magnetic flux @xmath178 , which are assigned and known functions of the spatial coordinates .
the perturbative theory developed here represents the starting point for the application of the kinetic theory to the modelling of off - equatorial plasma tori in axisymmetric disc systems .
this is based on the analysis of the spatial dependences which characterize the leading - order kinetic solution and that can be dealt with analytically thanks to the fundamental em ordering assumptions introduced in section 3 . in particular , for each of the three regimes considered here , the proof that the kinetic equilibria admit off - equatorial solutions follows by analyzing the spatial profile of the leading - order number density defined by eq .
( [ nde ] ) under the requirement of having maxima out of the equatorial plane , namely for @xmath179 .
this requires preliminary to assign the functional form of the structure functions @xmath9 characterizing eq .
( [ nde ] ) in terms of the potentials @xmath1 and/or @xmath4 .
a detailed discussion of this type for all of the three plasma regimes is beyond the scope of this work and will be addressed separately in future studies .
for the purpose of the present investigation , it is sufficient to consider here the case of regime 3 .
in fact , this provides the most general conditions for the occurrence of levitating structures , while regimes 1 and 2 can be viewed as special realizations of regime 3 .
in particular , we notice that the latter is expected to represent also the most plausible realization in real systems , in which both the gravitational and magnetic fields contribute to determine the profiles of the fluid fields .
let us then discuss the case of plasmas belonging to regime 3 .
the number density profile is prescribed according to eqs .
( [ nde ] ) and ( [ r3 ] ) .
we notice that , thanks to the analytical relationships ( [ d1 ] ) and ( d2 ) and the orderings ( [ basic])@xmath151([eees ] ) , any function of @xmath180 can be equivalently expressed in terms of @xmath181 . because of this , the rhs of eq .
( nde ) becomes now a generic function of @xmath181 , namely of the form @xmath182 .
this represents the most general kind of spatial dependence which is admitted by the kinetic equilibrium .
hence , in the general case and in the absence of other particular restrictions , suitable prescriptions of @xmath183 can be determined for regime 3 plasmas , according to the real system to be studied , which admit maxima out of the equatorial plane .
this conclusion has a general character of validity and assures the consistency of the kinetic theory presented here for collisionless axisymmetric plasmas with the possible occurrence of levitating tori in the external gravitational and magnetic fields of the type prescribed above .
we can now explore in more detail the present conclusion by considering explicitly two possible physical realizations of this type of solution : _ case a : _ in this first example we assume that both @xmath163 and @xmath165 in eq .
( [ nde ] ) are constant . from the physical point of view , the requirement @xmath184 means that the spatial variations of the number density are uniquely determined by the exponential term ( maxwellian factor ) , which in turn depends on the leading - order plasma temperature as well as on the rotational frequency , gravitational potential and magnetic flux @xmath1 .
this choice is consistent with previous literature ( see for example @xcite ) .
concerning the condition @xmath185 , this corresponds to a leading - order plasma isothermal profile that is consistent with the kinetic constraints that characterize the solution ( see section 7 ) .
we remark that in the present framework , the isothermal condition can only be satisfied to the leading - order , while for non - uniform plasmas , the full temperature profile is generally non - constant because of higher - order kinetic effects .
these issues will be discussed in detail in sections 7 and 8 . in validity of the prescription of constant @xmath163 and @xmath165 , the only freedom left concerns the functional dependence of @xmath186 , which is considered of the form ( [ r3 ] ) . under this assumption ,
the number density still remains of the type @xmath183 .
in this situation , it is convenient to prescribe @xmath183 consistent with the requirement of exhibiting maxima out of the equatorial plane , independently of its actual representation given by eq .
( [ nde ] ) .
the prescription of a physically - acceptable profile of @xmath183 must be done in such a way to reproduce observational or experimental data .
once the profile of @xmath183 is set , since also @xmath163 and @xmath165 are constant in this example , then eq . ( [ nde ] ) can be inverted and used to uniquely derive the expression of the corresponding species angular frequency @xmath187 which determines the levitating structure .
in particular , the latter is obtained by solving the quadratic algebraic equation@xmath188 hence , under these conditions , it is possible to introduce a density profile which is in agreement with physical configurations and the existence of off - equatorial tori . for leading - order isothermal systems
this also prescribes the form of the corresponding plasma rotation frequency according to eq .
( [ om ] ) .
the extension of this solution method to the case of a non - isothermal plasma species recquires the additional prescription of the temperature profile , namely @xmath189 , while the frequency @xmath107 can still be obtained from eq .
( [ om ] ) . in both cases
we notice that the kinetic equilibrium thus determined generally allows for the existence of two separate roots for @xmath107 .
if both are real , they should correspond to two different admissible equilibria with opposite direction of plasma rotation with respect to the external dipolar magnetic field orientation . _ case b : _ as a second example , we assume the validity of the kinetic constraint eq .
( [ r3 ] ) for all the three structure functions .
in particular , here we treat the situation in which the condition@xmath190is identically satisfied in the configuration domain occupied by the collisionless plasma species . from the physical point of view , eq .
( req1 ) means that the number density profile @xmath161 is not modified by the maxwellian exponential factor and coincides with the pseudo - density @xmath191 .
this requirement is satisfied when the exponential factor in eq .
( [ nde ] ) is one .
hence , this condition is met for the species angular frequency satisfying the algebraic quadratic equation@xmath192 we notice that again eq . ( [ omg ] ) generates two roots for the frequency @xmath107 , as for the case a discussed above .
in addition , eq .
( [ omg ] ) holds for both isothermal and non - isothermal plasmas . finally , in validity of the @xmath193ordering and the ordering ( [ eees ] ) , when @xmath194 , one can infer that the solutions of @xmath107 from eq .
( omg ) are asymptotically close to those from eq .
( [ om ] ) , although the number density and the temperature are not necessarily so . to conclude this section ,
it is useful to make a qualitative comparison of the results obtained here with those presented in ref .
@xcite , where the existence of off - equatorial structures has been proved on the basis of a fluid non - ideal mhd description .
this involves in particular the inspection of eq .
( 41 ) for the pressure profile given in ref .
@xcite , which can give rise to off - equatorial maxima for suitable choices of the coefficients entering the same equation ( see discussions in sections 3 and 4 in the same reference ) .
indeed , pressure and density profiles are proportional ( at least to the leading - order ) when the condition @xmath185 applies ( see also eq .
( [ ptot ] ) below ) .
in such a case , it is immediate to verify that the rhs of eq .
( 41 ) in ref .
@xcite can be effectively expressed as a function of @xmath1 and @xmath4 only , in agreement with the prescription holding for regime 3 plasmas . although the two treatments ( i.e. , the present one and ref .
@xcite ) consider different physical conditions for the levitating plasma , the consistency pointed out here establishes a notable result .
in fact , first it shows that , as anticipated in the introduction , the kinetic theory developed in this paper and its analytical formulation allow for direct comparisons with previous literature works based on fluid approaches .
second , it proves that fluid results can in principle be reproduced consistently on the basis of a kinetic treatment , thus extending their validity to a wider class of plasma regimes .
third , in turn it reinforces the statement given above concerning the general character of the present kinetic theory for regime 3 plasmas in providing a suitable mathematical and physical framework for the investigation of off - equatorial structures .
in this section we provide the explicit representation of the kinetic corrections @xmath169 , @xmath195 and @xmath170 introduced in the chapman - enskog representation of the equilibrium kdf given by eq .
( [ ce ] ) .
the inclusion of these contributions is necessary for the complete solution of the equilibrium problem in the framework of vlasov - maxwell description of collisionless plasmas . in particular , these terms represent the deviations of the kdf from a maxwellian distribution and arise because of the constraints imposed by single - particle phase - space conservation laws on the solution itself .
the precise definition of the first - order terms @xmath169 , @xmath167 and @xmath170 is also required to distinguish the solution among the three kinetic regimes pointed out above . in detail , the first - order corrections @xmath169 and @xmath167 originate from the perturbative treatment of the implicit phase - space dependences carried by the structure - functions @xmath132 entering the equilibrium kdf @xmath130 .
they are found to be given by:@xmath196 v_{\varphi } \nonumber \\ & & + \frac{cm_{\mathrm{s}}}{z_{\mathrm{s}}e}r\left [ \frac{e_{\mathrm{s}}-\omega _ { \mathrm{s}}p_{\varphi \mathrm{s}}}{t_{\mathrm{s}}}-\frac{3}{2}\right ] a_{3\mathrm{s}}v_{\varphi } , \label{h1}\end{aligned}\]]@xmath197 c_{3\mathrm{s}}\left ( \frac{1}{2}v^{2}+\frac{{z_{\mathrm{s}}e}}{m_{\mathrm{s}}}\phi \right ) \nonumber \\ & & + \left [ c_{1\mathrm{s}}+\frac{p_{\varphi \mathrm{s}}\omega _ { \mathrm{s}}}{t_{\mathrm{s}}}c_{2\mathrm{s}}\right ] \left ( \frac{1}{2}v^{2}+\frac{{z_{\mathrm{s}}e}}{m_{\mathrm{s}}}\phi \right ) , \label{h2}\end{aligned}\]]where the following definitions have been introduced:@xmath198hence , @xmath169 and @xmath195 are polynomial functions of the particle velocity which contain diamagnetic and energy - correction contributions and depend on the so - called thermodynamic forces @xmath199 and @xmath200 . the latter represent the gradients of the structure functions across equipotential magnetic and gravitational surfaces and arise in collisionless plasmas characterized by non - uniform fluid fields .
consistency of these expressions with the @xmath86 and @xmath87 ordering assumptions requires that@xmath201 v_{\varphi } \lesssim o\left ( \varepsilon _ { \mathrm{s}}\right ) , \]]@xmath202 c_{3\mathrm{s}}\left ( \frac{1}{2}v^{2}+\frac{{z_{\mathrm{s}}e}}{m_{\mathrm{s}}}\phi \right ) \lesssim o\left ( \sigma _ { \mathrm{s}}\right ) , \]]which implies that @xmath165 must actually be of the form @xmath203 , with @xmath40 , i.e. at most slowly - dependent on @xmath1 and @xmath4 .
this conclusion motivates the choice done in section 6 to treat isothermal plasmas ( to leading - order ) .
as pointed out above , the contribution @xmath169 is null for gravitationally - bound plasmas , while @xmath167 vanishes for magnetized plasmas . instead
, both terms are present in the equilibrium solution for plasmas belonging to regime 3 .
we also notice that the @xmath193expansion generates contributions in @xmath167 which are proportional to both particle kinetic energy and es potential . in particular , terms which depend on @xmath5
contribute to the occurrence of es corrections to the kinetic solution and the corresponding fluid fields .
finally , the last contribution @xmath170 originates from the validity of the @xmath204ordering ( [ eees ] ) when this is taken into account in the expression for the leading - order number density , and results in the following term@xmath205it must be stressed that the es contributions arising in @xmath167 and @xmath170 originate from different perturbative treatments .
in fact , @xmath206 follows from the @xmath204ordering and is common to all the regimes considered here when eq .
( [ eees ] ) applies .
instead the terms in @xmath167 can only be included when the @xmath207ordering applies ( regimes 1 and 3 ) . to conclude the section
, it is worth pointing out that the treatment of the first - order kinetic corrections displayed here requires the following preliminary steps:1 ) the precise identification of the appropriate plasma collisionless kinetic regime.2 ) the prescription of the leading - order spatial profiles of the structure functions , consistent with the kinetic constraints for each regime.3 ) the evaluation of the thermodynamic forces and the explicit calculation of the es potential.in particular , the existence of the equilibria determined here is subject to the validity of the maxwell equations , i.e. the poisson equation for the es potential @xmath5 and ampere s equation ( see section 9 ) .
in this section we proceed with the calculation of the eos corresponding to the kinetic equilibrium determined here .
this requires in particular to compute the species pressure tensor @xmath208 carried by the kdf @xmath130 and defined as@xmath209where @xmath26 denotes the velocity domain of integration . since @xmath130 is isotropic with respect to quadratic particle velocity dependences , one can immediately infer that each species pressure tensor is isotropic and of the form @xmath210 , where @xmath211 denotes the thermal scalar pressure , with @xmath212 and @xmath213 being respectively the species total number density and temperature associated with @xmath214 .
the calculation of @xmath215 can be carried out analytically for thermal particles when the chapman - enskog representation ( [ ce ] ) applies . in the following
we consider this case .
furthermore , consistent with the @xmath86 and @xmath87 orderings , in the first - order terms @xmath166 and @xmath167 we approximate the canonical momentum @xmath216 and the energy @xmath108 respectively with @xmath217 and @xmath218 .
hence , under these assumptions , one can prove that the scalar pressure can be represented as@xmath219 , \label{ptot}\]]where @xmath220 is the leading - order term , with @xmath161 being defined by eq .
( [ nde ] ) .
in addition we notice that the term @xmath169 associated with the @xmath221expansion does not contribute to the eos because it is odd in the azimuthal component of particle velocity .
instead , @xmath170 does not depend explicitly on particle velocity and therefore it is not affected when the integral ( [ pten ] ) is computed on @xmath26 . hence , its contribution in eq .
( [ ptot ] ) is simply proportional to @xmath222 and represents part of the es corrections which enter the definition of the total pressure @xmath215 . finally , invoking eq .
( [ h2 ] ) , explicit calculation gives for @xmath223 the following result:@xmath224where@xmath225 from this result it is interesting to point out that , although to the leading - order the species pressure coincides with the thermal pressure , the first - order corrections introduce deviations in the eos that are distinctive for collisionless plasmas . in particular , the energy - correction contributions which enter through @xmath226
are associated with the gradients of structure functions across gravitational equipotential surfaces and include also es corrections proportional to @xmath5 .
these terms however vanish in uniform collisionless plasmas . on the other hand ,
the es correction to the eos carried by @xmath170 is independent and follows uniquely from the @xmath204ordering introduced above between es and gravitational potential energy .
clearly , all the first - order contributions in the eos arise as specifically - kinetic effects , which characterize the kinetic treatment of collisionless plasmas .
in this section we analyze the constraints which are posed by the maxwell equations on the kinetic equilibria . these concern in particular the validity of the orderings ( [ basic ] ) and ( [ eees ] ) and for this reason they apply to all the plasma regimes identified above .
we consider first the poisson equation for the es potential , which is written as@xmath227 in the general case the solution is non - trivial , because the total number density @xmath228 depends both implicitly and explicitly on the es potential itself ( see for example @xcite ) .
however , the solution simplifies in validity of the sub - ordering expansion ( [ eees ] ) introduced above .
in fact in this case the es potential enters the kinetic solution only through the first - order corrections to the equilibrium kdf .
therefore , consistent with the orderings introduced in the present work and the perturbative theory developed here , one can obtain an asymptotic solution for @xmath5 by considering only the leading - order contribution to the species number density .
thus , when the said sub - ordering applies , neglecting corrections of @xmath229 and @xmath230 and invoking eq .
( [ nde ] ) , the poisson equation becomes to this accuracy:@xmath231where the source term @xmath232 is defined as@xmath233 .
\label{tesader}\]]here @xmath234 does not depend on @xmath5 , and therefore the es potential can be readily obtained by integrating eq .
( [ po2 ] ) yielding@xmath235with @xmath236 being the corresponding green function . for the consistency of the theory , the solution for @xmath5 given by the previous equation
must be checked a posteriori to satisfy the initial ordering ( [ eees ] ) . in particular
, this can represent a constraint condition for the magnitude of the species number densities which contribute to the es potential through the system charge density ( [ tesader ] ) .
manifestly , the validity of the ordering ( [ eees ] ) is necessary for the present theory to apply , and for this reason the calculation of @xmath5 represents the ultimate step to be done in order to warrant the consistency of the treatment .
similar considerations apply to the ampere equation , which determines the self - generation of magnetic field by the equilibrium collisionless plasma .
the ampere equation is written as@xmath237where @xmath60 is defined in eq .
( [ bself ] ) and @xmath238 is the total current density , which is given by@xmath239with @xmath240 being the species flow velocity .
it is immediate to prove that , in the present formulation , @xmath240 is purely azimuthal at equilibrium ; in fact additional components of the velocity can only arise in the presence of temperature anisotropy , see for example @xcite .
again , consistent with the perturbative treatment presented here , one can retain only the leading - order contributions to @xmath238 in eq .
( [ ampere-2 ] ) . under this assumption eq .
( [ ampere-2 ] ) becomes@xmath241where @xmath161 is given by eq .
( [ nde ] ) and @xmath160 ( see eq .
( [ fm ] ) ) .
( [ gss ] ) represents a generalized grad - shafranov equation for the poloidal magnetic flux @xmath242 in which the source term on the rhs depends only on explicitly known quantities .
the solution for @xmath60 which results from eq .
( [ gss ] ) must then be checked a posteriori to verify the ordering condition ( [ basic ] ) introduced above , which is necessary in order to warrant the validity of the theory and its analytical development . in this case eq .
( [ basic ] ) can represent a constraint for the magnitude of the species rotation angular frequencies which contribute to the system charge current .
in this paper , a theoretical investigation of equilibrium configurations for collisionless non - relativistic and axisymmetric plasmas has been presented , taking into account the role of a central spherically - symmetric gravitational field .
the formulation is based on a multi - species kinetic approach developed in the framework of the vlasov - maxwell description .
the case of astrophysical plasmas arising in the gravitational field of compact objects and in the presence of both an external dipolar magnetic field and self electromagnetic fields has been treated .
three different plasma regimes have been identified which are characetrized by distinctive kinetic orderings .
it has been proved that in all cases consistent kinetic equilibria can be determined , with the kinetic distribution function being expressed in terms of generalized maxwellian functions .
it has been shown that the three regimes differ by the form of the kinetic constraints which are imposed on the equilibrium solutions and which uniquely follow from phase - space single - particle conservation laws . by imposing appropriate orderings on the self electromagnetic fields and by developing a suitable perturbative theory ,
an analytical treatment of the equilibria has been proposed . in terms of this ,
several issues have been addressed .
first , the conditions of existence of equilibrium structures corresponding to off - equatorial tori have been investigated .
it has been shown that these systems can generally arise for the regime which has been referred to here as magnetized and gravitationally - bound plasmas .
this analysis can be important from the astrophysical point of view , since off - equatorial tori may represent a physically - realizable model of magnetized coronal plasmas which are believed to characterize accretion discs .
in addition , the treatment based on kinetic theory can pose the basis for comparison with analogous fluid results carried out in terms of mhd theory . as a second application ,
the plasma equation of state has been determined analytically and expressed in terms of the pressure tensor .
it has been shown that the latter exhibits deviations from the thermal pressure characteristic of collisional plasmas because of the existence of specifically - kinetic effects .
these have been identified with diamagnetic , energy - correction and electrostatic contributions which apply in combination with the occurrence of non - uniform fluid fields .
finally , the validity of the poisson and ampere equations have been addressed , showing that they can introduce non - trivial constraints on the magnitude of the plasma number density and flow velocity for the consistency with the orderings introduced in the theory developed here .
the conclusions established in this work can be relevant for future investigations of astrophysical plasmas in equilibrium configurations , with particular focus on collisionless plasmas in accretion discs and off - equatorial tori associated with compact objects .
financial support by the italian foundation angelo della riccia ( firenze , italy ) is acknowledged by c.c .
, j.k . , p.s . and z.s .
would like to express their acknowledgment for the institutional support of faculty of philosophy and science , silesian university in opava ( czech republic ) and thank the project synergy cz.1.07/2.3.00/20.0071 sustaining the international collaboration .
thanks czech science foundation gar 13 - 00070j . | the possible occurrence of equilibrium off - equatorial tori in the gravitational and electromagnetic fields of astrophysical compact objects has been recently proved based on non - ideal mhd theory .
these stationary structures can represent plausible candidates for the modelling of coronal plasmas expected to arise in association with accretion discs .
however , accretion disc coronae are formed by a highly diluted environment , and so the fluid description may be inappropriate .
the question is posed of whether similar off - equatorial solutions can be determined also in the case of collisionless plasmas for which treatment based on kinetic theory , rather than fluid one , is demanded . in this paper
the issue is addressed in the framework of the vlasov - maxwell description for non - relativistic multi - species axisymmetric plasmas subject to an external dominant spherical gravitational and dipolar magnetic field .
equilibrium configurations are investigated and explicit solutions for the species kinetic distribution function are constructed , which are expressed in terms of generalized maxwellian functions characterized by isotropic temperature and non - uniform fluid fields .
the conditions for the existence of off - equatorial tori are investigated .
it is proved that these levitating systems are admitted under general conditions when both gravitational and magnetic fields contribute to shaping the spatial profiles of equilibrium plasma fluid fields .
then , specifically - kinetic effects carried by the equilibrium solution are explicitly provided and identified here with diamagnetic , energy - correction and electrostatic contributions .
it is shown that these kinetic terms characterize the plasma equation of state by introducing non - vanishing deviations from the assumption of thermal pressure . |
the source content of the x ray sky has been investigated over a broad range of fluxes and solid angles thanks to a large number of deep and wide surveys performed in the last few years using _ rosat _ , _ chandra _ and xmm_newton _ ( see @xcite for a review ) . follow up observations unambiguously indicate that active galactic nuclei ( agn ) , many of which are obscured , dominate the global energy output recorded in the cosmic x ray background .
the impressive amount of x ray and multi - wavelength data obtained to date have opened up the quantitative study of the demography and evolution of accretion driven supermassive black holes ( smbhs ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
at present the two deepest x ray surveys , the _ chandra _ deep field north ( cdfn ; * ? ? ?
* ) and _ chandra _ deep field south ( cdfs ; * ? ? ?
* ) , have extended the sensitivity by about two orders of magnitude in all bands with respect to previous surveys @xcite , detecting a large number of faint x ray sources .
however , deep pencil beam surveys are limited by the area which can be covered to very faint fluxes ( typically of the order of 0.1 deg@xmath10 ) and suffer from significant field to field variance . in order to cope with such limitations ,
shallower surveys over larger areas have been undertaken in the last few years with both _
( e.g. the 9 deg@xmath1 bootes survey @xcite , the extended groth strip egs @xcite , the extended _ chandra _ deep field south e - cdfs , @xcite and the champ @xcite ) and xmm_newton _ ( e.g. the hellas2xmm survey @xcite , the xmm_newton _ bss @xcite and the elais s1 survey @xcite ) . in this context the xmm_newton
_ wide field survey in the cosmos field @xcite , hereinafter xmm
cosmos @xcite , has been conceived and designed to maximize the sensitivity and survey area product , and is expected to provide a major step forward toward a complete characterization of the physical properties of x ray emitting smbhs .
a contiguous area of about 2 deg@xmath1 will be covered by 25 individual pointings , repeated twice , for a total exposure time of about 60 ksec in each field . in the first observing run obtained in ao3 ( phase a ) , the pointings were disposed on a 5x5 grid with the aimpoints shifted of 15 each other , so as to produce a contiguous pattern of coverage . in the second run , to be observed in ao4 ( phase b ) , the same pattern will be repeated with each pointing shifted by 1 with respect to phase a. the above described approach ensures a uniform and relatively deep coverage of more than 1 deg@xmath1 in the central part of the field .
when completed , xmm cosmos will provide an unprecedentedly large sample of about 2000 x ray sources with full multi - wavelength photometric coverage and a high level of spectroscopic completeness . as a consequence , the xmm cosmos survey is particularly well suited to address agn evolution in the context of the large scale structure in which they reside . more specifically , it will be possible to investigate if obscured agn are biased tracers of the cosmic web and whether their space density rises in the proximity of galaxy clusters @xcite .
+ the x ray reduction of phase a data along with a detailed analysis of the source counts in different energy bands are presented in this paper which is organized as follows . in section 2 the data reduction procedure and the relative astrometric corrections
are described . in section 3 the source detection algorithms and technique
are discussed .
monte carlo simulations are presented in section 4 .
the logn logs relations and the analysis of the contribution of the xmm - cosmos sources to the x - ray background are discussed in section 5 .
the study of sample variance is presented in section 6 and a summary of the work is reported in section 7 . the strategy and the log of the observations of xmm cosmos
are presented by @xcite , the optical identifications of x - ray sources by @xcite , the analysis of groups and clusters by @xcite , the spectral analysis of a subsample of bright sources by @xcite and the clustering of x - ray extragalactic sources by @xcite . throughout the paper the concordance wmap @xmath11cdm cosmology @xcite
is adopted with h@xmath12=70 km s@xmath7 mpc@xmath7 , @xmath13=0.7 and @xmath14=0.3
the epic data were processed using the xmm_newton _ standard analysis system ( hereinafter sas )
version 6.5.0 .
the observational data files ( odf , `` raw data '' ) of each of the 25 observations , were calibrated using the sas tools _ epchain _ and _ emchain _ with the most recent calibration data files .
events in bad columns , bad pixels and close to the chip gaps were excluded .
+ both the epic pn and mos event files were searched for high particle background intervals . the distribution of the background counts binned in 100 s intervals was obtained in the 1214 kev band for the pn and in the 1012 kev band for the mos , which are dominated by particle background , and then fitted with a gaussian model . all time intervals with background count rate higher than 3@xmath15 above the average best fit value were discarded .
cut above which the corresponding time intervals have been discarded .
, scaledwidth=50.0% ] in fig .
[ fig : bkg ] an example of the application of this method to field # 6 is shown .
once the high energy flares were removed , the 0.310 kev background counts distribution was processed , with the same 3@xmath15 clipping method , in order to remove times during which low energy particle flares were important .
these flares are not easily detected in the 1214 kev band .
as a result of this selection process the average time lost due to particle flares was @xmath1620@xmath17 and 2 observations were completely lost ( see * ? ? ?
+ an important feature observable in the background spectrum of both mos and pn ccds is the al k@xmath18 ( 1.48 kev ) fluorescent emission . in the pn background
two strong cu lines are also present at @xmath0 7.4 kev and @xmath0 8.0 kev . since these emission lines could affect the scientific results , the 7.27.6 kev and 7.88.2 kev energy bands in the pn and the 1.451.54 kev band ( in pn and mos ) were excluded from the detectors events .
images were then created in the 0.52 kev , 24.5 kev and 4.510 kev energy bands with a pixel size of 4 arcsec .
single and double events were used to construct the pn images , while mos images were created using all valid event patterns . out - of - time ( oot )
events appear when a photon hits the ccd during the read - out process in the imaging mode .
the result is that the x position of the event / photon is known , while the y position is unknown due to the readout and shifting of the charges at this time .
for this reason artificial oot event files were created .
a new y coordinate is simulated by randomly shifting the event along the readout axis and performing the gain and cti ( charge transfer inefficiency ) correction afterwards .
for the pn , in full frame mode the oot events constitute about 6.3@xmath17 of the observing time .
those files were filtered in the same way as the event files and the produced images were subtracted from the event images .
images were then added in order to obtain pn+mos mosaics . for each instrument and for each observation
, spectrally weighted exposure maps were created using the sas task _ eexpmap _ , assuming a power law model with photon index @xmath19=2.0 in the 0.52 kev band and @xmath19=1.7 in the 24.5 and 4.510 kev bands . in order to correct the astrometry of our xmm_newton _
observations for each pointing and for each instrument , the produced x - ray source list ( see next section ) was compared with the megacam catalog of the cosmos field @xcite including all the sources with i magnitudes in the range 18 - 23 . in order to find the shift between the two catalogues ,
an optical - x - ray positional correlation was computed using the likelihood algorithm included in the sas task _
eposcorr_. this task uses in a purely statistical way all possible counterparts of an x - ray source in the field to determine the most likely coordinate displacement .
this method is independent of actual spectroscopic identifications , but all post - facto checks using , for example , secure spectroscopic identifications , have demonstrated its reliability and accuracy . using the magnitude range mentioned above ,
systematic effects introduced by bright stars and faint background objects are minimized . in the majority of the observations
the shift between the three cameras turned out to be @xmath16 1 "
( i.e. much smaller than the pixel size of the images used here ) .
since the shift between the epic cameras is negligible , a correlation between the joint mos+pn source list and the optical catalog was calculated to derive the astrometric correction . for the 23 pointings presented here , the shifts between the optical and x - ray catalog are never larger than 3@xmath20 , with an average shift of the order of @xmath211.4@xmath20 and @xmath22 - 0.17@xmath20 . the average displacement in the two coordinates between the pn+mos mosaic x - ray positions and the megacam catalog sources for each pointing of the xmm - cosmos field is shown in fig [ fig : astro ] .
the appropriate offset was applied to the event file of each pointing and images and exposure maps were then reproduced with the corrected astrometry .
lcccccc[!t ] 0.52 & 2.0 & 10.45 & 7.2@xmath410@xmath5 & 1307 & 1281 & 661 + 24.5 & 1.7 & 1.52 & 4.0@xmath410@xmath8 & 735 & 724 & 89 + 4.510 & 1.7 & 1.21 & 9.7@xmath410@xmath8 & 187 & 186 & 3 in order to perform the source detection a sophisticated background modeling has been developed . in x - ray observations
the background is mainly due to two components , one generated by undetected faint sources contributing to the cosmic x - ray background and one arising from soft protons trapped by the terrestrial magnetic field .
for this reason two background templates were computed for each instrument and for each pointing , one for the sky ( vignetted ) background @xcite and one for instrumental and particles background ( unvignetted ) . to calculate the normalizations of each template of every pointing
, we first performed a wavelet source detection ( see * ? ? ?
* ) without sophisticated background subtraction , then we excised the areas of the detector where a significant signal due to sources was detected .
the residual area is split into two parts depending on the value of the effective exposure ( i.e. higher and lower than the median value ) . using the two templates
, we calculate the coefficients of a system of two linear equations from which we obtain the normalizations of both : @xmath23 where a and b are the normalization factors , m@xmath24 the vignetted templates in the region with effective exposure higher and lower than the median , m@xmath25 the unvignetted templates and c@xmath26 are the background counts in the two regions .
the region with effective exposure lower than the median ( i.e. high vignetting , @xmath277 off - axis ) is dominated by the instrumental background , while the region with higher effective exposure is dominated by the sky background .
therefore with this method we have the advantage of better fitting the two components of the background . the standard method for estimating the background , based on the spline functions used in the xmm_newton _ pipelines , returned in our case significant residuals .
the excellent result of this technique can be seen in the signal - to - noise ( snr ) map in fig [ fig : snr ] : despite the significant variations in exposure time and average background level from pointing to pointing , a rather homogeneous signal to noise ratio is achieved across the whole mosaic .
it is worth noting that also pixels with negative values are shown in the map ; these are located where the background model is higher than the measured background . in each pointing the source detection
was conducted on the combined images of the different instruments in the three energy bands mentioned above using the sas tasks _ eboxdetect _ and _ emldetect_. as a first step , the sliding cell detection algorithm _
eboxdetect _ was run on the images in the three energy bands . in this procedure source counts
were collected in cells of 5@xmath45 pixels adopting a low threshold in the detection likelihood ( i.e. likemin=4 ) . the source list produced by _
eboxdetect _ was then used as input for _
emldetect_. for all the sources detected with the sliding cell method this task performs a maximum likelihood psf fit . in this way
refined positions and fluxes for the sources were determined .
due to the particular pattern of our observations ( see * ? ? ?
* ) , the same source could be detected in up to 4 different pointings .
for this reason both _
eboxdetect _ and _ emldetect _ were run in raster mode .
the source parameters ( position and flux ) were fitted simultaneously on all the observations where the source is observable , taking into account the psf at the source position in each pointing . as likelihood threshold for the detection
, we adopted the value det_ml=6 .
this parameter is related to the probability of a random poissonian fluctuation having caused the observed source counts : @xmath28 in principle , the expected number of spurious sources could be estimated as the product of the probability for a random poisson fluctuation exceeding the likelihood threshold times the number of statistically independent trials , n@xmath29 . for a simple box detection algorithm n@xmath29 would be approximately given by the number of independent source detection cells across the field of view . for the complex multi - stage source detection algorithm ,
like the one applied here , n@xmath29 can not be calculated analytically , but has to be estimated through monte carlo simulations .
these simulations , which are discussed in section 4 , return a number of spurious sources of @xmath02% at the likelihood level chosen .
all the sources were fitted with a psf template convolved with a beta model @xcite .
sources which have a core radius significantly larger than the psf are flagged as extended ( _ ext _ parameter @xmath30 ) .
a total of 1307 , 735 and 187 x - ray sources were detected in the three bands .
of these sources , twenty - six were classified as extended .
the analysis of the x - ray extended sources in the cosmos field is beyond the scope of this work ; these sources are extensively discussed by @xcite . a total of 1281 , 724 and 186 point - like sources were detected in the three bands down to limiting fluxes of 7.2@xmath410@xmath5 erg @xmath6 s@xmath7 , 4.7@xmath410@xmath8 erg @xmath6 s@xmath7 and 9.7@xmath410@xmath8 erg @xmath6 s@xmath7 respectively .
the minimum number of net counts for the detected sources is @xmath021 , 17 and 27 in the three bands , respectively .
a total of 1390 independent point - like sources have been detected by summing the number of sources detected in each band but not in any softer energy band .
the number of sources detected only in the 0.52 kev , 24.5 kev , 4.510 kev bands are 661 , 89 and 3 , respectively . from the count rates in the 0.52 kev , 24.5 kev and 4.510 kev bands
the fluxes were obtained in the 0.52 kev , 210 kev and 510 kev bands respectively using the energy conversion factors ( ecf ) listed in table [ tab : tab1 ] , together with a summary of the source detection .
the ecf values have been computed using the most recent epic response matrices in the corresponding energy ranges . as a model , we assumed power - law spectra with n@xmath31 = 2.6@xmath410@xmath32 @xmath6 , ( corresponding to the average value of n@xmath31 over the whole cosmos field @xcite ) and the same spectral indices used to compute the exposure maps without considering any intrinsic absorption ( see table [ tab : tab1 ] ) .
it is worth noting that the spectral indices and the absorptions of the individual sources can be significantly different from the average values assumed here .
in particular @xcite found that the spectral indices @xmath19 of the xmm - cosmos sources are in the range [email protected] , in the cdfs @xcite measured an average photon index @xmath341.75 and similar values were obtained by @xcite in the champ survey .
the mean spectrum assumed here is therefore consistent with the values measured up to now . by changing the spectral index of @xmath35 the ecfs change of 2% , 12% and 4% in the 0.52 kev , 24.5 kev and 4.510 kev , respectively . the cumulative probability to detect a true source with det@xmath36ml@xmath376 in a circle of a given radius in the 0.52 ( @xmath38 ) , 24.5 ( @xmath39 ) and 4.510 kev ( @xmath40 ) energy bands .
the 68% , 90% and 99% levels are plotted as horizontal lines . ]
in order to properly estimate the source detection efficiency and biases , detailed monte carlo simulations were performed ( see , e.g. , * ? ? ?
* ; * ? ? ?
twenty series of 23 xmm_newton _
images were created with the same pattern , exposure maps and background levels as the real data .
the psf of the simulated sources was constructed from the templates available in the xmm_newton _
calibration database .
the sources were randomly placed in the field of view according to a standard 0.5 - 2 kev logn - logs distribution @xcite .
this was then converted to a 24.5 kev and 4.510 kev logn - logs assuming that all the sources have the same intrinsic spectrum ( a power - law with spectral index @xmath41 ) .
we then applied , to the simulated fields , the same source detection procedure used in the real data .
@xcite showed that with the threshold adopted here for source detection , which corresponds roughly to the gaussian 4.5 - 5@xmath15 , the distortion of the slope of the logn - logs due to poissonian noise is @xmath163@xmath17 for a wide range of slopes .
therefore , the uncertainties introduced by using a single logn - logs as base for the simulations are negligible .
a total of 30626 , 13579 and 3172 simulated sources were detected in the 0.52 kev , 24.5 kev and 4.510 kev bands down to the same limiting fluxes of the observations .
for every possible pair of input - output sources we computed the quantity @xmath42 where @xmath43 and @xmath44 are the position and flux of the detected source and @xmath45 and @xmath46 are the corresponding values for all the simulated sources we then flag as the most likely associations those with the minimum value of @xmath47 . the distribution of the positional offsets is plotted in fig .
[ fig : pos ] for each energy band analyzed .
the ratio s@xmath48/s@xmath49 as a function of the output detected flux in the 0.52 kev , 24.5 kev and 4.510 kev bands , respectively@xmath50.,title="fig:",scaledwidth=45.0% ] the ratio s@xmath48/s@xmath49 as a function of the output detected flux in the 0.52 kev , 24.5 kev and 4.510 kev bands , respectively@xmath50.,title="fig:",scaledwidth=45.0% ] the ratio s@xmath48/s@xmath49 as a function of the output detected flux in the 0.52 kev , 24.5 kev and 4.510 kev bands , respectively@xmath50.,title="fig:",scaledwidth=45.0% ] we find that 68% of the sources are detected within 2.1@xmath20 , 1.3@xmath20 and 0.8@xmath20 in the 0.52 kev , 24.5 kev and 4.510 kev bands , respectively . since the detection software fits the position of the source using the information available for the three bands together , we expect to be able to detect sources with an accuracy of the order of , or somewhat better than , that shown in fig . [ fig : pos ] . as in @xcite
, we then define a cut - off radius @xmath51 of 6@xmath20 .
sources with a displacement larger than @xmath51 from their input counterpart are classified as spurious .
these account for 2.7% , 0.5% , 0.6% of the total number of sources in the 0.52 kev , 24.5 kev and 4.510 kev bands , respectively .
source confusion occurs when two or more sources fall in a single resolution element of the detector and result as a single detected source with an amplified flux . in order to determine the influence of the source confusion we adopted the method described in @xcite .
we define as @xmath52 sources those for which @xmath53 ( where @xmath54 is the 1@xmath15 error on the output flux ) .
the fraction of @xmath52 sources is 0.8@xmath17 , 0.15@xmath17 and @xmath160.1@xmath17 in the 0.52 kev , 24.5 kev and 4.510 kev bands , respectively .
+ the photometry was also tested ; the ratio of output to input fluxes in the simulation is plotted in fig.[fig : phot ] . at bright fluxes
this ratio is consistent with one , while at fainter fluxes the distribution of @xmath55 becomes wider , mainly because of increasing errors , and skewed toward values greater than one .
this skewness of the distribution can be explained mainly by two effects : a ) source confusion and b ) eddington bias @xcite . while source confusion , as defined above , affects only a small fraction of the sources , the eddington bias results in a systematic upward offset of the detected flux .
the magnitude of this effect depends on the shape of the logn - logs distribution and the statistical error on the measured flux .
since there are many more faint than bright sources , uncertainties in the measured flux will result in more sources being up - scattered than down - scattered .
together with this , the fact that in the 4.510 kev band we are sampling a flux region in which the logn - logs is steeper than in the other bands ( see section 5 ) , explains why such an effect is more evident in the 4.510 kev band . besides assessing the reliability of our source detection procedure ,
one of the aims of these simulations is to provide a precise estimation of the completeness function of our survey , known also as sky coverage .
we constructed our sky - coverage ( @xmath56 ) vs. flux relation by dividing the number of detected sources by the number of input sources as a function of the flux and rescaling it to the sky simulated area .
having analyzed the simulations with the same procedure adopted for the analysis of the data , this method ensures that when computing the source counts distribution ( see next section ) all the observational biases are taken into account and corrected .
the @xmath56 vs. flux relation relative to the 0.52 kev , 210 kev and 510 kev bands is plotted in fig .
[ fig : area ] .
the total sky area is 2.03 deg@xmath1 and it is completely observable down to fluxes of @xmath00.3 , 1.3 and 2@xmath410@xmath57 erg @xmath6 s@xmath7 in the three bands , respectively .
the sky coverage drops to 0 at limiting fluxes of @xmath07@xmath410@xmath5 erg @xmath6 s@xmath7 , @xmath04@xmath410@xmath8 erg @xmath6 s@xmath7and @xmath09@xmath410@xmath8 erg @xmath6 s@xmath7 , in the 0.52 kev , 210 kev and 510 kev bands , respectively .
[ fig : area ] ccc[!b ] 0.52 kev & & + -13.0 & 2.03 & 4.5@xmath58 + -13.5 & 2.03 & 18.8@xmath59 + -14.0 & 2.03 & 105.2@xmath60 + -14.5 & 2.03 & 327.0@xmath61 + -15.0 & 0.58 & 790@xmath62 + -15.1 & 0.12 & 931@xmath63 + 210 kev & & + -13.0 & 2.03 & 8.6@xmath64 + -13.5 & 2.03 & 57.0@xmath65 + -14.0 & 1.40 & 258.9@xmath66 + -14.3 & 0.13 & 600.1@xmath67 + 510 kev & & + -13.5 & 2.03 & 21.3@xmath68 + -14.0 & 0.35 & 111@xmath69 once the sky coverage is known , the cumulative source number counts can be easily computed using the following equation : @xmath70 where _
@xmath71 _ is the total number of detected sources in the field with fluxes greater than @xmath44 and @xmath72 is the sky coverage associated with the flux of the i@xmath73 source .
the variance of the source number counts is therefore defined as : @xmath74 source number counts are reported in table [ key ] .
the cumulative number counts , normalized to the euclidean slope ( multiplied by s@xmath75 ) , are shown in figures [ fig : logn1 ] , [ fig : logn2 ] and [ fig : pippo ] , ) sources compared with the rosat medium sensitivity survey @xcite ( @xmath76 ) , combined rosat , xmm_newton _ , _ chandra _ @xcite ( @xmath77 ) , the _ chandra _ deep field south 1@xmath15 error tie @xcite ( @xmath78 ) , the _ chandra _ deep field north 1@xmath15 error tie @xcite ( @xmath79 ) , the 100 ks of the xmm_newton _ lockman hole @xcite ( @xmath80 ) , the hellas2xmm @xcite ( @xmath81 ) and the extended cdfs @xcite ( @xmath82 ) surveys .
the overlayed black - dashed line represents the logn - logs predicted by the model of @xcite .
the source number counts are plotted scaled by s@xmath75 in order to highlight the deviation from the euclidean behavior . ] ) sources compared with _
chandra _ , xmm_newton _ and asca ( @xmath83 ) @xcite , hellas bepposax @xcite ( @xmath84 ) the chandra deep field south 1@xmath15 error tie @xcite ( @xmath78 ) , the hellas2xmm ( @xmath85 ) @xcite , the elais s1 ( @xmath86 ) @xcite , the extended cdfs @xcite 1@xmath15 error tie ( @xmath82 ) and the 100 ks of the lockman hole ( @xmath87 ) @xcite .
the overlayed black - dashed line represents the logn - logs predicted by the model of @xcite .
the source number counts are plotted scaled by s@xmath75 in order to highlight the deviation from the euclidean behavior . ] ) sources compared with the hellas2xmm @xcite ( @xmath85 ) , the chandra deep field south 1@xmath15 error tie @xcite ( @xmath78 ) , the hellas - bepposax data from @xcite ( @xmath84 ) , the elais s1 ( @xmath86 ) @xcite and the 100 ks of the lockman hole ( @xmath88 ) @xcite .
the overlayed black - dashed line represents the logn - logs predicted by the model of @xcite .
the source number counts are plotted scaled by s@xmath75 in order to highlight deviation from the euclidean behavior . ] for the 0.52 kev , 210 kev and 510 kev energy ranges , respectively .
with such a representation , the deviations from the euclidean slope are clearly evident as well as the flattening of the counts towards faint fluxes .
source counts are compared with the findings of other deep and shallow surveys collected from the literature .
the plotted reference results were selected in order to sample a flux range as wide as possible and at the same time to keep the plots as clear as possible . as discussed in the previous section ,
the sky coverage @xmath56 was derived from realistic monte carlo simulations and therefore no further correction for the eddington bias is required . in order to parameterize our relations
, we performed a maximum likelihood fit to the unbinned differential counts .
we assumed a broken power - law model for the 0.52 kev and 210 kev bands : @xmath89 where @xmath90 is the normalization , @xmath91 is the bright end slope , @xmath92 the faint end slope , @xmath93 the break flux , and @xmath44 the flux in units of 10@xmath57 erg @xmath6 s@xmath7 .
notice that using the maximum likelihood method , the fit is not dependent on the data binning and therefore we can make full use of the whole dataset . moreover , the normalization @xmath94 is not a parameter of the fit , but it is obtained by imposing the condition that the number of expected sources from the best fit model is equal to the total observed number . in the 0.52 kev energy band
the best fit parameters are @xmath91=2.60@xmath95 , @xmath92=1.65@xmath96 , @xmath93=1.55@xmath97 @xmath410@xmath57 erg @xmath6 s@xmath7 and @xmath94=123 . translating this value of the normalization to that for the cumulative distribution at 2@xmath410@xmath8 erg @xmath6 s@xmath7 , which is usually used in the literature for _ chandra _ surveys
, we obtain @xmath98450 which is fully consistent with most of previous works where a fit result is presented @xcite , but significantly lower than that found in the clasxs survey @xcite . in the 210 kev band the best fit parameters
are @xmath91=2.43@xmath99 , @xmath92=1.59@xmath100 , @xmath93=1.02@xmath101 @xmath410@xmath57 erg @xmath6 s@xmath7 and @xmath94=266 .
the latest value translates into @xmath981250 .
also in this band , our results are in agreement with previous surveys within 1@xmath15 , with the exception of the clasxs survey which is @xmath030% higher in this band @xcite . in the 510 kev energy bands , where the differential counts do not show any evidence for a break in the sampled flux range , we assumed a single power - law model of the form : @xmath102 for which the best fit parameters are found to be @xmath94=102 and @xmath91=2.36@xmath103 . in the soft 0.5 - 2 kev ( fig .
8) energy range a visual inspection of the various datasets suggests a remarkably good agreement between xmm - cosmos and literature data .
in the 210 kev band the xmm - cosmos counts bridge the gap between deep field observations @xcite and shallower large area bepposax @xcite and xmm_newton _
surveys @xcite . at relatively bright fluxes ( @xmath104 erg @xmath6 s@xmath7 ) the xmm - cosmos logn
logs nicely matches previous measurements , though providing a much more robust estimate of the source counts thanks to the much smaller statistical errors .
a major step forward in the determination of x ray source counts is achieved in the 510 kev band , where the previously existing data from different surveys show very significant differences .
thanks to xmm
cosmos , a solid measure of the hard - x logn logs in the flux interval 10@xmath57
10@xmath105 erg @xmath6 s@xmath7 is obtained for the first time . from fig .
[ fig : pippo ] we notice that the normalization of the xmm
cosmos logn
logs is slightly higher than ( @xmath010% ) , although consistent at 1@xmath15 with , that measured by _ chandra _ while , elais s1 @xcite source counts are 30% lower .
however , in the overlapping flux range the latter is characterized by large errors due to the small number of relatively bright sources in the _ chandra _ deep fields .
interestingly enough , the xmm
cosmos counts match nicely , with smaller errors , those of the wide area hellas2xmm survey ( baldi et al .
2002 ) , while the pioneering measurements of bepposax @xcite are systematically higher than the counts from xmm - cosmos .
one of the main aims of xmm - cosmos with its large and medium deep coverage , is to provide a solid census of the x - ray source population to be compared with observations and models of agn evolution .
according to recent synthesis models ( see e.g. ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) a high fraction of heavily obscured agn is necessary to explain the spectral shape and the intensity of the x - ray background ( xrb ) .
we examine therefore which fraction of xrb is resolved into discrete sources in our survey . as a first test , we computed the flux which xmm - cosmos itself resolves into discrete sources by summing their fluxes weighted on the sky coverage in the 0.52 kev , 210 kev and 5 - 10 kev energy bands .
as in @xcite we used as reference value of the normalization at 1 kev of the xrb spectrum that of @xcite which assumes that the spectral shape in the 1 - 10 kev band is a power - law with spectral index @xmath19=1.4 and a normalization at 1kev of 11.6 kev @xmath6 s@xmath7 kev@xmath7 . the latter value corresponds to a flux of 0.80 , 2.31 and 1.27@xmath106 erg @xmath6 s@xmath7 deg@xmath107 in the 0.52 kev , 210 kev and 5 - 10 kev energy bands , respectively . in the 0.52 kev band
we measure a contribution of the sources to the xrb which corresponds to a normalization at 1 kev of 0.49@xmath108 erg @xmath6 s@xmath7 deg@xmath107 .
the corresponding values in the 210 kev and 510 kev bands are 0.92@xmath109 erg @xmath6 s@xmath7 deg@xmath107 and 0.28@xmath110 erg @xmath6 s@xmath7 deg@xmath107 .
therefore xmm - cosmos resolves by itself @xmath065@xmath17 , @xmath040@xmath17 and @xmath022@xmath17 of the xrb into discrete sources in the 0.52 kev , 210 kev and 510 kev energy bands , respectively .
it is worth noticing that the flux measured by @xcite is the highest measured in literature in the 110 kev energy range ( see e.g. * ? ? ?
* for a complete collection ) .
it is also worth noticing that we computed the fraction of resolved xrb by assuming that all the sources have the same spectrum .
therefore , in our estimate , effects due to the broad absorption and spectral index distributions of agns are not included . in figs .
[ fig : logn1 ] , [ fig : logn2 ] and [ fig : pippo ] we compared our logn - logs to those predicted by the recent xrb model of @xcite .
this model makes use of the most recent observational constraints on the agn populations and includes a conspicuous fraction of compton thick agn which , however , are not expected to significantly contribute to the xmm - cosmos counts . in the 0.52 kev band
a direct comparison of our data with the model shows a 1@xmath15 agreement at the bright end . at the faint
end the model predicts a slightly higher normalization when compared to most of the plotted data , including ours .
a similar behavior is observed in the 210 kev band .
it is worth noticing that in the model the average unabsorbed power - law spectral index of the sources is @xmath341.8 in the flux interval sampled by xmm - cosmos ( see fig .
19 in @xcite ) .
since in our data analysis we assumed @xmath111=1.7 we expect in this band a slight ( @xmath010% ) underestimation of the fluxes when compared to those of the model .
this effect is almost negligible ( i.e. @xmath165% ) in the other bands investigated here . by integrating our best fit 210 kev logn - logs between infinite and zero , and assuming that the slope of the `` real '' logn - logs remains constant down to low fluxes
, we estimated the total contribution of agns to the xrb .
we predict a total flux of agns in the xrb of 1.25@xmath106 erg @xmath6 s@xmath7 deg@xmath107 .
this value is @xmath040% lower than that measured by @xcite , and @xmath0 25% smaller than those obtained by integration the model logn - logs of @xcite which predicts a flux of @xmath01.66@xmath106 erg @xmath6 s@xmath7 deg@xmath107 .
this discrepancy between our predicted flux and that of @xcite , could arise by the fact that , in our measurement , we consider that all the sources have the same spectrum and from statistical uncertainty of the logn - logs parameters . by assuming an average spectral index @xmath111=1.4 for all our sources , we obtain a value for the total flux of the agns of @xmath01.48@xmath106 erg @xmath6 s@xmath7 deg@xmath107 which is consistent , within the statistical uncertanties , with the prediction of the model . considering the total flux of the xrb predicted by the model and our estimate from the logn - logs distributions , in the 210 kev band ,
xmm - cosmos resolves the @xmath05565% of the total flux of the xrb into discrete sources .
it is interesting to observe how in the 510 kev band our data are in good agreement with the prediction of the model .
this result is particularly important since in this band it is expected the major contribution of highly absorbed agn , which are an important ingredient of the xrb models .
a detailed analysis of the spectral properties of the brightest x - ray sources in xmm - cosmos is presented by @xcite .
the amplitude of source counts distributions varies significantly among different surveys ( see e.g. * ? ? ?
* ; * ? ? ?
* and references therein ) .
this `` sample variance '' , can be explained as a combination of poissonian variations and effects due to the clustering of sources @xcite .
the variance of counts in cells for sources which are angularly correlated can be obtained with : @xmath112 where @xmath113 is the mean density of objects in the sky , @xmath56 is the cell size , and @xmath114 is the angular two point correlation function .
the first term of eq . [ cvar ] is the poissonian variance and the second term is introduced by the large scale structure . in order to determine whether the differences observed in the source counts of different surveys
could arise from the clustering of x - ray sources , we estimated the amplitude of the fluctuations from our data , by producing subsamples of our survey with areas comparable to those of . e.g. , _ chandra _ surveys .
+ the xmm - cosmos field and the monte carlo sample fields of section 4 were divided in 4,9,16 and 25 square boxes . making use of the 0.52 kev energy band data , we computed for each subfield , the ratio of the number of real sources to the number of random sources . in order to prevent incompleteness artifacts , we conservatively cut the limiting flux of the random and data sample to 5@xmath410@xmath8 erg @xmath6 s@xmath7 . at this flux
our survey is complete over the entire area . in order to avoid artifacts introduced by the missing pointings in the external part of the field of view
, we concentrated our analysis to the central 80@xmath11580@xmath116 .
expected fluctuation . ] in fig.[fig : varianza ] we plot the ratio of the data to the random sample as a function of the size of the cells under investigation .
the measured fractional standard deviations of the sample is reported in table [ tab : fit ] .
using eq .
[ cvar ] we computed the expected amplitude of source counts fluctuations with @xmath114 taken from @xcite .
they computed the x - ray two point correlation function in the xmm - cosmos field and detected clustering signal with angular correlation length @xmath117 @xmath01.9@xmath20 @xmath00.8@xmath20 and @xmath06@xmath20 in the 0.52 kev , 24.5 kev and 4.510 kev band , respectively .
the observed slope is @xmath118=1.8 in all the energy bands .
the predicted fractional standard deviations are therefore 0.13 , 0.19 , 0.23 and 0.28 on scales of 0.44 deg@xmath1 , 0.19 deg@xmath1 , 0.11 deg@xmath1 and 0.07 deg@xmath1 , respectively .
these values are in good agreement with those observed in the sub - samples of our dataset as shown by the value of the fitted @xmath119 to the counts in cell fluctuations ( see table [ tab : fit ] ) . as shown in table [
tab : fit ] , at this limiting flux and on the areas considered here the main contribution to the source counts fluctuations is from the poissonian noise . at the flux limit assumed here ,
the ratio @xmath120/@xmath121 increases from @xmath00.5 on the smallest scale ( 16 x 16 arcmin ) to @xmath00.85 on the largest scale ( 40 x 40 arcmin ) .
this ratio scales as : @xmath122 where @xmath113 deg@xmath107 is the surface density of the sources , @xmath117 ( deg ) is the angular correlation length and @xmath123 ( deg ) is the size of the cell . in order to estimate at which flux limit fluctuations introduced by the large scale structure are predominant , we estimate that @xmath120/@xmath121 would be @xmath01 on the smallest scale , corresponding to a chandra acis field of view ( 16 x 16 arcmin ) at a surface density of the order of @xmath0900 deg@xmath107 , corresponding to a 0.5 - 2 kev flux s@xmath0 8@xmath410@xmath5 erg @xmath6 s@xmath7 . at even fainter fluxes the dominant contribution to the total expected source counts fluctuations on this area ( @xmath124 ) comes from the large scale structure ,
therefore the contribution of statistical fluctuations becomes less important . with the same procedure , we can estimate the total expected fluctuations ( @xmath124 ) and the relative importance of @xmath121 and @xmath120 also for the hard band ( 5 - 10 kev ) , even if in this band we do not have enough statistics to divide our field in sub - samples . using the formal best fit for @xmath117 in this band found by miyaji et al .
( 2007 ) ( @xmath117=6@xmath20 ) , we find that at the faintest 5 - 10 kev flux ( s@xmath010@xmath57 erg @xmath6 s@xmath7 ) sampled by the xmm - cosmos survey ( @xmath125 110 deg@xmath107 ) the ratio @xmath120/@xmath121 is smaller than one on all the four scales here analyzed , with a total expected standard deviation of the fluctuations ranging from @xmath00.20 on the largest scale to @xmath00.40 on the smallest scale .
these values for @xmath124 are significantly larger than those shown in table 3 for the soft band , because in the hard band the surface density of sources is lower and the angular correlation length is higher than in the soft band .
this analysis is at least qualitatively consistent with figures 8 and 10 , which show a significantly larger dispersion in the data from different surveys in the hard band than in the soft band .
moreover , the results here discussed are also consistent with the observed fluctuations in the deep chandra fields ( see , for example , bauer et al .
large area , moderately deep surveys like xmm - cosmos are needed to overcome the problem of low counting statistics , typical of deep pencil beam surveys , and , at the same time , to provide a robust estimate of the effect of large scale structure on observed source counts . as a final consideration , we tried to compute the expected intrinsic variance of xmm - cosmos .
this estimate must be made with care since we have only one sample on this scale .
assuming that the angular correlation function of @xcite was universal , the residual uncertainties on the source counts are estimated to be @xmath165 - 6% in the 0.52 kev and 210 kev bands , and of the order of the 10% in the 510 kev band .
cccccc 40@xmath11540@xmath116 & 0.09@xmath126 & 0.10 & 0.09 & 0.13 & 4.21/3 + 26@xmath11526@xmath116 & 0.20@xmath96 & 0.15 & 0.10 & 0.19 & 8.93/8 + 20@xmath11520@xmath116 & 0.21@xmath126 & 0.20 & 0.11 & 0.23 & 16.63/15 + 16 @xmath11516@xmath116 & 0.24@xmath127 & 0.25 & 0.12 & 0.28 & 25.15/24 .
the data analysis of the first run of observations of the xmm-_newton _ cosmos wide field survey has been presented .
a total of 1390 point - like sources are detected on a contiguous area of about 2 deg@xmath1 down to fluxes of 7.2@xmath410@xmath5 erg @xmath6 s@xmath7 , 4.0@xmath410@xmath8 erg @xmath6 s@xmath7 and 9.7@xmath410@xmath8 erg @xmath6 s@xmath7 in the 0.52 kev , 210 kev and 510 kev energy bands , respectively .
the detection procedure was tested through monte carlo simulations which confirmed the high level of accuracy in the determination of the source properties ( aperture photometry and positioning ) and allowed us to keep statistical biases under control .
a robust estimate of x ray source counts at both soft and hard energies , obtained thanks to the large number of sources detected in the xmm cosmos survey , is presented in this paper .
the differential logn
logs was fitted with a broken power - law model in the 0.52 kev and 210 kev energy bands , and with a single power - law model in the 510 kev energy band . in the soft 0.52 kev band ,
already extensively covered by rosat , xmm_newton _ and _ chandra _ surveys over a range of fluxes encompassing those sampled by the cosmos survey , our results are in excellent agreement with previous analysis ( see fig [ fig : logn1 ] ) , providing an independent evidence of the validity of our data analysis procedure .
the large number of x ray sources of the cosmos survey allowed us to constrain with unprecedented accuracy the logn - logs parameters in the 210 and 510 kev energy ranges over a range of fluxes which were previously poorly constrained .
most importantly , in the hard 510 kev band , we were able to fill the gap between the deep _ chandra _ surveys in the cdfs and cdfn and shallower large area surveys .
the deviations from other surveys , which are , however , less than 30% , have been explained in terms of low counting statistics of pencil beam surveys , and partially by the effect of large scale structure .
the major step forward in the determination of hard x ray source counts achieved thanks to the xmm cosmos survey will provide an important reference point for the study of the agn demography and evolution especially with applications to obscured agn .
more specifically , the evolutionary properties of the obscured agn can be tightly constrained , since they are indeed very sensitive , according to the most recent model of the x ray background ( gilli , comastri & hasinger 2006 ) , to the shape of the hard x ray source counts around the break flux , which is precisely where the cosmos data play a key role . in this context
, we compared our results to the most recent predictions of the model by @xcite , finding a remarkable agreement between data and model .
the second pass of the xmm_newton _
observations in the cosmos field ( 600 ks ) has already started , and is expected to significantly increase the total number of x ray sources .
the results of the full xmm cosmos survey including the complete ( ao3+ao4 ) source catalogue will be the subject of a future paper . it is anticipated that , when completed , the xmm cosmos survey will provide a number of x ray sources over a large enough contiguous area and down to moderately deep fluxes that it will make possible the study of agn evolution and their connection with the large scale structure in which they reside with unprecedented detail . the cosmos field has been granted 1.8 msec observation with _
chandra _ in its central square degree ( c - cosmos , p.i .
martin elvis ) .
the joint chandra and xmm_newton _ observation will provide an unprecedented lab for agn physics .
this work is based on observations obtained with xmm_newton _ , an esa science mission with instruments and contributions directly funded by esa member states and the us ( nasa ) . in germany , the xmm_newton _ project is supported by the bundesministerium fr bildung und forschung / deutsches zentrum fr luft und raumfahrt , the max
planck society , and the heidenhain stiftung . in italy ,
the xmm - cosmos project is supported by inaf and miur under grants prin/270/2003 and cofin-03 - 02 - 23 and by asi under grant asi / inaf i/023/05/0 . in the usa
the xmm - cosmos project is supported by nasa nng04gg40 g and nng06gd77 g .
this work is based in part on observations obtained with megaprime / megacam , a joint project of cfht and cea / dapnia , at the canada - france - hawaii telescope ( cfht ) , and on data products produced at terapix and the canadian astronomy data centre .
we kindly acknowledge the anonymous referee for his / her valuable comments and suggestions .
we gratefully acknowledge the contributions of the entire cosmos collaboration consisting of more than 70 scientists .
more information on the cosmos survey is available at ` http://www.astro.caltech.edu/~cosmos ` .
baldi , a. , molendi , s. , comastri , a. , fiore , f. , matt , g. , & vignali , c. 2002 , , 564 , 190 bauer , f. e. , alexander , d. m. , brandt , w. n. , schneider , d. p. , treister , e. , hornschemeier , a. e. , & garmire , g. p. 2004
, , 128 , 2048 brandt , w. n. , & hasinger , g. 2005 , , 43 , 827 brunner , a. 2006 , , submitted brusa , m. 2007 , , this volume cappelluti , n. , cappi , m. , dadina , m. , malaguti , g. , branchesi , m. , delia , v. , & palumbo , g. g. c. 2005 , , 430 , 39 cappelluti , n. , boehringer , h. , schuecker , p. , pierpaoli , e. , mullis , c. r. , gioia , i. m. , & henry , j. p. 2006
, accepted by , arxiv astrophysics e - prints , arxiv : astro - ph/0611553 cappi , m. , et al .
2001 , , 548 , 624 cavaliere , a. , & fusco - femiano , r. 1976 , , 49 , 137 comastri , a. , setti , g. , zamorani , g. , & hasinger , g. 1995 , , 296 , 1 dickey , j. m. , & lockman , f. j. 1990 , , 28 , 215 della ceca , r. , et al .
2004 , , 428 , 383 de luca , a. , & molendi , s. 2004 , , 419 , 837 eddington , a. s. , sir 1940 , , 100 , 354 finoguenov , a. 2007 , , this volume fiore , f. , et al .
2001 , , 327 , 771 fiore , f. , et al .
2003 , , 409 , 79 giacconi , r. , & zamorani , g. 1987 , , 313 , 20 giacconi , r. , et al .
2001 , , 551 , 624 gilli , r. , et al .
2003 , , 592 , 721 gilli , r. , comastri , a. & hasinger , g. 2006 , a&a , in press , astro - ph/0610939 giommi , p. , perri , m. , & fiore , f. 2000 , , 362 , 799 green , p. j. , et al .
2004 , , 150 , 43 hasinger , g. , burg , r. , giacconi , r. , hartner , g. , schmidt , m. , trumper , j. , & zamorani , g. 1993 , , 275 , 1 hasinger , g. , burg , r. , giacconi , r. , schmidt , m. , trumper , j. , & zamorani , g. 1998 , , 329 , 482 hasinger , g. , et al .
2001 , , 365 , l45 hasinger , g. , miyaji , t. , & schmidt , m. 2005 , , 441 , 417 hasinger , g. 2007 , , this volume henry , j. p. , & briel , u. g. 1991 , , 246 , l14 johnson , o. , best , p. n. , & almaini , o. 2003 , , 343 , 924 kenter , a. , et al .
2005 , , 161 , 9 kim , d .- w . , et al . 2004 , , 600 , 59 la franca , f. , et al . 2005 , , 635 , 864 lehmer , b. d. , et al .
2005 , , 161 , 21 loaring , n. s. , et al .
2005 , , 362 , 1371 lumb , d. h. , warwick , r. s. , page , m. , & de luca , a. 2002 , , 389 , 93 maccacaro , t. , gioia , i. m. , wolter , a. , zamorani , g. , & stocke , j. t. 1988 , , 326 , 680 mainieri , v. , 2007 , , this volume mc cracken , 2007 , , this volume miyaji , t. , hasinger , g. , & schmidt , m. 2000 , , 353 , 25 miyaji , t. , 2007 , , this volume moretti , a. , campana , s. , lazzati , d. , & tagliaferri , g. 2003 , , 588 , 696 murray , s. s. , et al .
2005 , , 161 , 1 mushotzky , r. f. , cowie , l. l. , barger , a. j. , & arnaud , k. a. 2000 , , 404 , 459 nandra , k. , et al . 2005 , , 356 , 568 peebles , p. j. e. 1980 , the large - scale structure of the universe ( princeton , n.j .
, princeton university press ) puccetti et al .
2006 , , 457 , 501 rosati , p. , et al .
2002 , , 566 , 667 ruderman , j. t. , & ebeling , h. 2005 , , 623 , l81 schmitt , j. h. m. m. , & maccacaro , t. 1986 , , 310 , 334 scoville , n.z .
et al . , 2007 , apjs this volume spergel , d. n. , et al .
2003 , , 148 , 175 tozzi , p. , et al .
2006 , , 451 , 457 ueda , y. , et al . 1999 , , 518 , 656 ueda , y. , akiyama , m. , ohta , k. , & miyaji , t. 2003 , , 598 , 886 virani , s. n. , treister , e. , urry , c. m. , & gawiser , e. 2006 , , 131 , 2373 yang , y. , mushotzky , r. f. , barger , a. j. , cowie , l. l. , sanders , d. b. , & steffen , a. t. 2003 , , 585 , l85 yang , y. , mushotzky , r. f. , steffen , a. t. , barger , a. j. , & cowie , l. l. 2004 , , 128 , 1501 yang , y. , mushotzky , r. f. , barger , a. j. , & cowie , l. l. 2006 , , 645 , 68 worsley , m. a. , et al .
2005 , , 357 , 1281 | we present the data analysis and the x ray source counts for the first season of xmm_newton _ observations in the cosmos field .
the survey covers @xmath02 deg@xmath1 within the region of sky bounded by @xmath2 ; @xmath3 with a total net integration time of 504 ks .
a maximum likelihood source detection was performed in the 0.52 kev , 24.5 kev and 4.510 kev energy bands and 1390 point - like sources were detected in at least one band .
detailed monte carlo simulations were performed to fully test the source detection method and to derive the sky coverage to be used in the computation of the logn - logs relations .
these relations have been then derived in the 0.52 kev , 210 kev and 510 kev energy bands , down to flux limits of 7.2@xmath410@xmath5 erg @xmath6 s@xmath7 , 4.0@xmath410@xmath8 erg @xmath6 s@xmath7 and 9.7@xmath410@xmath8 erg @xmath6 s@xmath7 , respectively .
thanks to the large number of sources detected in the cosmos survey , the logn - logs curves are tightly constrained over a range of fluxes which were poorly covered by previous surveys , especially in the 210 and 510 kev bands . the 0.52 kev and 210 kev differential logn - logs were fitted with a broken power - law model which revealed a euclidean slope ( @xmath92.5 ) at the bright end and a flatter slope ( @xmath91.5 ) at faint fluxes . in the 510 kev energy band
a single power - law provides an acceptable fit to the observed source counts with a slope @xmath92.4 .
a comparison with the results of previous surveys shows good agreement in all the energy bands under investigation in the overlapping flux range
. we also notice a remarkable agreement between our logn - logs relations and the most recent model of the xrb .
the slightly different normalizations observed in the source counts of cosmos and previous surveys can be largely explained as a combination of low counting statistics and cosmic variance introduced by the large scale structure . |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.